UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Coronary heart disease risk factor methodology and prevention/intervention strategies Jones, Wayne Nelson 1982

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1982_A6_7 J67.pdf [ 7.47MB ]
Metadata
JSON: 831-1.0095395.json
JSON-LD: 831-1.0095395-ld.json
RDF/XML (Pretty): 831-1.0095395-rdf.xml
RDF/JSON: 831-1.0095395-rdf.json
Turtle: 831-1.0095395-turtle.txt
N-Triples: 831-1.0095395-rdf-ntriples.txt
Original Record: 831-1.0095395-source.json
Full Text
831-1.0095395-fulltext.txt
Citation
831-1.0095395.ris

Full Text

e,). CORONARY HEART DISEASE RISK FACTOR METHODOLOGY AND PREVENTION/INTERVENTION STRATEGIES by WAYNE NELSON JONES B . S c , The U n i v e r s i t y of B r i t i s h Co lumbia , 1982 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n THE FACULTY OF GRADUATE STUDIES (Heal th S e r v i c e s P l a n n i n g Programme) We accept t h i s t h e s i s as conforming to the r e q u i r e d standard THE UNIVERSITY OF BRITISH COLUMBIA October , 1982 @ Wayne Nelson Jones , 1982 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the head o f my department o r by h i s o r her r e p r e s e n t a t i v e s . I t i s understood t h a t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n p e r m i s s i o n . Department of Jzp/ Penrose y The U n i v e r s i t y of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date 7 i i A b s t r a c t Coronary heart d i s e a s e has been a s e r i o u s h e a l t h problem f o r the l a s t 50+ years. Due to the magnitude of t h i s problem, s u c c e s s f u l p r e v e n t i o n and/or i n t e r v e n t i o n s t r a t e g i e s are g r e a t l y needed. One p o t e n t i a l l y s u c c e s s f u l i n t e r v e n t i o n or p r e v e n t i o n s t r a t e g y i s the r e d u c t i o n of r i s k f a c t o r s f o r coronary heart d i s e a s e . Since 1967 the primary way of d e s c r i b i n g r i s k f a c t o r s f o r the development of coronary heart d i s e a s e has been through the use of l o g i s t i c a n a l y s i s . While t h i s model has produced v a l u a b l e r e s u l t s , c e r t a i n problems must be overcome before a f e a s i b l e i n t e r v e n t i o n or p r e v e n t i o n program c o u l d be developed. P r o s p e c t i v e s t u d i e s examining the development of coronary heart d i s e a s e were s t a r t e d in the 1950's. Four main r i s k f a c t o r s were i d e n t i f i e d in these s t u d i e s : i n c r e a s i n g age, serum c h o l e s t e r o l l e v e l , e l e v a t e d blood p r e s s u r e , and smoking. However, a l a r g e number of other v a r i a b l e s i n one or another of these s t u d i e s were a l s o found to be r i s k f a c t o r s , and other methods of a n a l y s i s were needed. While the l o g i s t i c model was f i r s t used to analyze s e l e c t e d r e s u l t s of a-coronary heart disease p r o s p e c t i v e study in 1961, widespread use of the model d i d not begin u n t i l a f t e r 1967. Since then, most p r o s p e c t i v e s t u d i e s have i n c l u d e d some form of l o g i s t i c a n a l y s i s of t h e i r data. These analyses have c o n s i s t e n t l y found smoking, serum c h o l e s t e r o l , blood pressure, and age to be s t a t i s t i c a l l y s i g n i f i c a n t r i s k f a c t o r s f o r the development of coronary heart d i s e a s e . T h i s paper examines the model of coronary heart d i s e a s e that has developed out of the use of the l o g i s t i c f u n c t i o n as a method of a n a l y z i n g the data c o l l e c t e d i n p r o s p e c t i v e coronary heart d i s e a s e s t u d i e s . The focus of t h i s t h e s i s i s on the u s e f u l n e s s of the model as a t o o l i n the development of coronary heart d i s e a s e p r e v e n t i o n or i n t e r v e n t i o n programs that might be c r e a t e d in the near f u t u r e . Demonstration of a c a u s a l r e l a t i o n s h i p between a r i s k f a c t o r and the development of coronary heart d i s e a s e has not been explored in as much d e t a i l . Only two of the p r o s p e c t i v e s t u d i e s that are attempting to lower r i s k f a c t o r s have been completed. One study was not able to show a s t a t i s t i c a l l y s i g n i f i c a n t improvement in m o r t a l i t y r a t e s f o r the experimental area. The other study i n d i c a t e s that lowering an i n d i v i d u a l ' s serum c h o l e s t e r o l l e v e l lowered the coronary heart d i s e a s e m o r b i d i t y r a t e . Other s t u d i e s should be completed in the near f u t u r e , and t h i s i s s u e may become c l a r i f i e d . Some of the p o t e n t i a l problem areas that e x i s t for any pr e v e n t i o n or i n t e r v e n t i o n program that i s based on a l o g i s t i c model in c l u d e s e l e c t i o n of the a p p r o p r i a t e equation, s e l e c t i o n of s p e c i f i c d e t a i l s of an i n t e r v e n t i o n or p r e v e n t i o n s t r a t e g y , and s e l e c t i o n of the t a r g e t l e v e l s of the r i s k f a c t o r s . Once these problems are d e a l t with a v i a b l e i n t e r v e n t i o n or p r e v e n t i o n program c o u l d be developed. V Table of Contents A b s t r a c t ; i i 1 I n t r o d u c t i o n 1 2 Risk And C a u s a l i t y 7 3 H i s t o r i c a l Development 17 I The American P r o s p e c t i v e S t u d i e s 23 A The Albany Study 24 B The Los Angeles Study 24 C The Chicago Study 25 D The Minnesota Study 26 E The Western C o l l a b o r a t i v e Study 27 F The Framingham Study 27 II Serum C h o l e s t e r o l F i n d i n g s 28 III Blood Pressure F i n d i n g s 41 IV C i g a r e t t e Smoking F i n d i n g s 49 V A d d i t i o n a l Risk F a c t o r s 56 4 M u l t i p l e L o g i s t i c A n a l y s i s 60 I M u l t i v a r i a t e A n a l y s i s 60 II The L o g i s t i c Equation 65 III E a r l y Attempts At A s s e s s i n g M u l t i p l e Risk 69 IV E a r l y Use Of The M u l t i p l e L o g i s t i c Equation 73 V General Comments 75 5 Risk F a c t o r s And The L o g i s t i c Model 80 I Risk F a c t o r s From The L o g i s t i c Model 81 II Tests Of The L o g i s t i c Model 96 A Examination Of The L o g i s t i c P r e d i c t i o n s 96 B V a l i d a t i o n On Other Subjects 102 C Attempts To A l t e r Risk F a c t o r s 106 i . The North K a r e l i a P r o j e c t 107 i i . The Oslo Study 116 i i i . The M u l t i p l e Risk F a c t o r I n t e r v e n t i o n T r i a l 118 6 P r e v e n t i o n / I n t e r v e n t i o n S t r a t e g i e s And The L o g i s t i c Model 123 I P o t e n t i a l Problem Areas With The A p p l i c a t i o n Of The L o g i s t i c Model 126 A S e l e c t i o n Of The A p p r o p r i a t e Equation 126 B S e l e c t i o n Of The A p p r o p r i a t e I n t e r v e n t i o n S t r a t e g y 129 C S e l e c t i o n Of The A p p r o p r i a t e Prevention S t r a t e g y 132 D S e l e c t i o n Of The Target L e v e l s 135 II P o t e n t i a l D i f f i c u l t i e s That Might A r i s e With The I n t r o d u c t i o n Of Programs 137 III A d d i t i o n a l Research Needs Of The L o g i s t i c Model 142 7 C o n c l u s i o n s 147 References 151 A d d i t i o n a l B i b l i o g r a p h y 157 v i L i s t of Tables 2.1 A d d i t i o n a l Risk F a c t o r s For Coronary Heart Disease 9 3.1 Number Of New Events C l a s s i f i e d By The Number Of I n d i v i d u a l s E i t h e r Above The F i f t i e t h Or Seventy-f i f t h P e r c e n t i l e , By S i t e Of Centre 33 3.2 M o r b i d i t y Rate By Serum C h o l e s t e r o l C a t e g o r i z a t i o n , Albany Study A f t e r 44 Months Of Follow-up 36 3.3 M o r b i d i t y Rate By Serum C h o l e s t e r o l C a t e g o r i z a t i o n , Chicago Study A f t e r 4.42 Years Of Follow-up 37 3.4 M o r b i d i t y Rate By Serum C h o l e s t e r o l C a t e g o r i z a t i o n , Los Angeles Study A f t e r 10 Years Of Follow-up 38 3.5 M o r b i d i t y Rate By Serum C h o l e s t e r o l C a t e g o r i z a t i o n , Western C o l l a b o r a t i v e Study A f t e r 4 1/2 Years Of Follow-up 39 3.6 M o r b i d i t y Rate By Serum C h o l e s t e r o l C a t e g o r i z a t i o n , Minnesota Study A f t e r 15 Years Of Follow-up 39 3.7 M o r b i d i t y Rate By Serum C h o l e s t e r o l C a t e g o r i z a t i o n , Framingham Study 40 3.8 M o r b i d i t y Rate By D i a s t o l i c Blood Pressure C a t e g o r i z a t i o n , Albany Study A f t e r 44 Months Of Follow-up 45 3.9 M o r b i d i t y Rate By Blood Pressure C a t e g o r i z a t i o n , Los Angeles Study A f t e r 10 Years Of Follow-up 46 3.10 M o r b i d i t y Rate By Blood Pressure C a t e g o r i z a t i o n , Western C o l l a b o r a t i v e Study A f t e r 4 1/2 Years Of Follow-up 47 3.11 M o r b i d i t y Rate By Blood Pressure C a t e g o r i z a t i o n , Framingham Study A f t e r 18 Years Follow-up 48 3.12 M o r b i d i t y Rate By Smoking C a t e g o r i z a t i o n , Chicago Study A f t e r 4.4 Years Of Follow-up 53 3.13 M o r b i d i t y Rate By Smoking C a t e g o r i z a t i o n , Albany Study A f t e r 44 Months Of Follow-up 53 3.14 M o r b i d i t y Rate By Smoking C a t e g o r i z a t i o n , Western C o l l a b o r a t i v e Study A f t e r 4.5 Years Of Follow-up. . 54 3.15 M o r b i d i t y Rate By Smoking C a t e g o r i z a t i o n , Minnesota Study A f t e r 20 Years Follow-up 54 3.16 M o r b i d i t y Rate By Smoking C a t e g o r i z a t i o n , Framingham Study A f t e r 18 Years Of Follow-up 55 5.1 Summary Of The L o g i s t i c C o e f f i c i e n t s Reported In A V a r i e t y Of Studies 82 5.2 Summary Of The L o g i s t i c C o e f f i c i e n t s Reported By H a l p e r i n , Blackwelder And V e r t e r (1971) 91 5.3 R e s u l t s Of A X 2 G o o d n e s s - o f - f i t Test Comparing The P r e d i c t i v e A b i l i t y Of L o g i s t i c Equations C o n t a i n i n g V a r i o u s Numbers Of V a r i a b l e s 98 5.4 Number Of Subjects P a r t i c i p a t i n g In The North K a r e l i a Study 109 5.5 Prevalence Of Smoking In North K a r e l i a And C o n t r o l Area 110 5.6 Serum C h o l e s t e r o l C o n c e n t r a t i o n s In North K a r e l i a And C o n t r o l Area 111 7 Comparison Of The Age-adjusted Annual Incidence Of AMI For I n d i v i d u a l s Aged 30-64 113 8 M o r t a l i t y Rates (/1000) For I n d i v i d u a l s Aged 30-64 In 1977 114 9 M o r b i d i t y And M o r t a l i t y R e s u l t s Found In The Oslo Study 119 1 P r e d i c t e d And Observed Number Of Coronary Events As De s c r i b e d By The P o o l i n g P r o j e c t Research Group. ..138 2 S e l e c t i o n Of High Risk S u b j e c t s For Further I n t e r v e n t i o n , And The Observed Number Of Coronary Events As Reported By The P o o l i n g P r o j e c t Research Group (1978) 139 v i i i L i s t of F i g u r e s 3.1 Coronary Heart Disease M o r b i d i t y Rates As A Func t i o n Of Serum C h o l e s t e r o l L e v e l s 42 3.2A Coronary Heart Disease M o r b i d i t y Rates By L e v e l Of S y s t o l i c Blood Pressure 50 3.2B Coronary Heart Disease M o r b i d i t y Rates By L e v e l Of D i a s t o l i c Blood Pressure 51 3.3 Coronary Heart Disease M o r b i d i d t y Rates By The Number Of C i g a r e t t e s Smoked Per Day 57 5.1 P r e d i c t e d Number Of Cases Of Coronary Heart Disease Per D e c i l e Of Risk For The Equations D e s c r i b e d In Table 5.3 101 Acknowledgements I would l i k e to thank my Chairman, Dr. John M i l sum f o r the v a l u a b l e help and i n s i g h t s prov ided throughout the development and w r i t i n g of t h i s t h e s i s . In a d d i t i o n , the c o n t r i b u t i o n s of my o ther t h e s i s committee members, Dr. Annette Stark and Dr. Simon R a b k i n , helped r e f i n e and complete t h i s work. 1 CHAPTER 1  I n t r o d u c t i o n Coronary heart d i s e a s e has been the number one cause of death among middle aged males in the Western World f o r the l a s t few decades. As such, i t has a t t r a c t e d a great deal of medical r e s e a r c h i n t e r e s t . Most of t h i s r e s e a r c h has focused on the e t i o l o g y of the d i s e a s e , and enormous s t r i d e s have been made i n the understanding of the course and treatment of coronary heart d i s e a s e . With t h i s i n c r e a s e d understanding has come c a l l s f o r the c r e a t i o n of p r e v e n t i v e programs aimed at a r r e s t i n g the development of coronary heart d i s e a s e . In a d d i t i o n , attempts are being made to develop methods of i d e n t i f y i n g high r i s k i n d i v i d u a l s , i n order to intervene before the coronary event, and thus avoid a p o s s i b l e death or i l l n e s s . I n t e r v e n t i o n attempts are even more app e a l i n g when the c o s t s a s s o c i a t e d with the treatment of coronary p a t i e n t s are c o n s i d e r e d . From the p e r s p e c t i v e of p u b l i c h e a l t h , p r e v e n t i o n and/or i n t e r v e n t i o n are the only s t r a t e g i e s a v a i l a b l e f o r combating coronary heart d i sease. Breslow (1978a) has d e s c r i b e d two approaches to the c o n t r o l of c h r o n i c d i s e a s e s through r i s k f a c t o r i n t e r v e n t i o n . The f i r s t i s by d i s c o v e r i n g and q u a n t i f y i n g the r e l a t i o n s h i p between c e r t a i n c h a r a c t e r i s t i c s and the development of the d i s e a s e , and then changing these 2 c h a r a c t e r i s t i c s to avoid the d i s e a s e process. The second i n v o l v e s i d e n t i f y i n g the d i s e a s e process at an e a r l y stage, and then t r e a t i n g the d i s e a s e to a v o i d f u r t h e r p r o g r e s s i o n . These two approaches are commonly r e f e r r e d to as pr e v e n t i o n and i n t e r v e n t i o n . (Note that the terms primary and secondary i n t e r v e n t i o n are sometimes used.) T h i s t h e s i s i s not concerned with which approach would be the most e f f e c t i v e . I t i s premature to attempt to examine t h i s i s s u e , given the present s t a t e of understanding about coronary heart d i s e a s e . Both approaches r e q u i r e a d e t a i l e d understanding of the coronary heart d i s e a s e process i n order to be e f f e c t i v e . T h i s paper examines one model of coronary heart d i s e a s e that both has been developed on the b a s i s of res e a r c h and i t s e l f has been used to j u s t i f y the design of c e r t a i n r e s e a r c h s t u d i e s . The major poin t of i n t e r e s t i s i n examining the v a l i d i t y of t h i s model, and i t s impact on prev e n t i o n and i n t e r v e n t i o n s t r a t e g i e s . I t would appear to be an a p p r o p r i a t e time to examine the area of coronary heart d i s e a s e r i s k f a c t o r s . P u b l i c h e a l t h i n t e r v e n t i o n has had numerous successes (e.g. the e r a d i c a t i o n of smallpox; measles i n n o c u l a t i o n ) . Once a s u f f i c i e n t understanding of the problem was obtained and the a p p r o p r i a t e i n n o c u l a t i o n found the danger of the d i s e a s e was v i r t u a l l y e l i m i n a t e d . While much of t h i s success was dependent on l a b o r a t o r y r e s e a r c h , e p i d e m i o l o g i c a l research d i d p lay a r o l e . Some f e e l t h a t , based on the present s t a t e 3 of knowledge, i t i s p o s s i b l e to s u c c e s s f u l l y i n t e r v e n e i n the course of coronary heart d i s e a s e . Biorck (1975) has presented a l i s t of p r e r e q u i s i t e s f o r primary p r e v e n t i o n . He s t a t e s that f o r a p r e v e n t i v e program to be f e a s i b l e , the f o l l o w i n g c o n d i t i o n s must be p r e s e n t : 1. The mechanism of the d i s e a s e must be known and methods to i n t e r f e r e with them be ava i l a b l e . 2. Simple methods to i d e n t i f y " s u s c e p t i b l e s " ("risk f a c t o r s " , " i n d i c e s " ) must be a v a i l a b l e . 3. The i d e n t i f i c a t i o n should have a reasonable degree of s e l e c t i v i t y and s p e c i f i c i t y . (?) 1 4. There should be a reasonable t i m e - r e l a t i o n s h i p between recommended a c t i o n and p a l p a b l e r e s u l t s . 5. " S u s c e p t i b l e s " must be w i l l i n g and able to cooperate by changing t h e i r mode of l i f e (or t a k i n g " p r o p h y l a c t i c " drugs), or 6. A u t h o r i t i e s (governments) must be w i l l i n g and able to impose such changes upon " s u s c e p t i b l e s " or the p o p u l a t i o n at l a r g e . (p. 58) To date there has been no examination of the f e a s i b i l i t y of i n t e r v e n t i o n using t h i s kind of o v e r a l l p e r s p e c t i v e . D espite the lack of an o v e r a l l a n a l y s i s , the problem of coronary heart d i s e a s e i s so l a r g e that c a l l s f o r an i n t e r v e n t i o n of some s o r t are being made. For example, Breslow (1978b) s t a t e s : There i s a r a p i d l y growing s c i e n t i f i c and p r o f e s s i o n a l concensus that r i s k f a c t o r i n t e r v e n t i o n should be i n t e n s i v e l y explored as a 1 I t i s unclear what i s meant by the q u e s t i o n mark at the end of t h i s c o n d i t i o n . From the context of the paper i t would appear that Biorck does not know what the reasonable degree would be. 4 way of ma i n t a i n i n g and improving h e a l t h d u r i n g the present e r a . (p. 449) In the same paper he l a t e r s t a t e s that while knowledge i s not complete i n s e v e r a l areas i t i s time to "assemble the p e r t i n e n t data and base immediate a c t i o n on prudent i n t e r p r e t a t i o n of the data." (p. 456). At l e a s t one l a r g e s c a l e i n t e r v e n t i o n t r i a l has been completed, and others are in the process of being designed or completed. T h i s has set up f o r c e s pushing for the development of some i n t e r v e n t i o n s t r a t e g y , without a thorough understanding of how the i n t e r v e n t i o n might work at the p u b l i c h e a l t h or p h y s i o l o g i c a l l e v e l . The l a r g e s t problem f a c i n g the development of a p r e v e n t i v e program i s determining the a p p r o p r i a t e i n t e r v e n t i o n . Despite the inc r e a s e i n knowledge over the l a s t t h i r t y years, i t does not yet appear p r a c t i c a l to design an i n t e r v e n t i o n s t r a t e g y that c o u l d be used s u c c e s s f u l l y on a la r g e s c a l e b a s i s . The main source of d i f f i c u l t y i s that coronary heart d i s e a s e appears to be a degenerative (rather than a v i r a l ) d i s e a s e that develops over a time span of decades. Thus f o r any i n t e r v e n t i o n s t r a t e g y to be e f f e c t i v e the i n t e r v e n t i o n must e i t h e r be able to stop the d i s e a s e process when i t i s a p p l i e d , or at l e a s t i d e n t i f y only those i n d i v i d u a l s who need f u r t h e r medical a t t e n t i o n . Large s c a l e i n t e r v e n t i o n s t r a t e g i e s are l i k e l y to be expensive, and before they are begun in earnest i t seems worthwhile to examine the data and models on which 5 such attempts w i l l be based. The key to the development of e i t h e r i n t e r v e n t i o n s t r a t e g y l i e s i n the a b i l i t y to i d e n t i f y those f a c t o r s that put the i n d i v i d u a l at r i s k . Before any i n t e r v e n t i o n s t r a t e g y can be s u c c e s s f u l the v a r i a b l e s that c o n s t i t u t e r i s k f a c t o r s must be d e s c r i b e d . Much of the e p i d e m i o l o g i c a l r e s e a r c h i n the area of coronary heart d i s e a s e has been concerned with t h i s problem. The degenerative nature of coronary heart d i s e a s e makes t h i s problem complex, as i t i s not always c l e a r when an i n d i v i d u a l should be d e s c r i b e d as being at r i s k . The r e l a t i v e l y slow development of the d i s e a s e has a l s o meant that r e s e a r c h s t u d i e s are long and expensive undertakings. Thus much of the l i t e r a t u r e i s s t i l l e x p l o r a t o r y i n nature, and as i s normally the case with e x p l o r a t o r y s t u d i e s , a l a r g e number of v a r i a b l e s have been examined, o f t e n without a s p e c i f i c hypothesis i n mind. The a n a l y s i s of a l a r g e number of v a r i a b l e s presents c e r t a i n i n t e r p r e t i v e and s t a t i s t i c a l problems. The s t a t i s t i c a l methods used to analyze such data bases o f t e n make assumptions about how the v a r i a b l e s i n t e r a c t (that i s , i t w i l l r e s t r i c t them e i t h e r to no i n t e r a c t i o n or to i n t e r a c t i o n s of c e r t a i n s p e c i f i e d t y p e s ) . The v a l i d i t y of the c o n c l u s i o n s drawn from such s t u d i e s thus o f t e n r e s t s on the a p p r o p r i a t e n e s s of the model being used to analyze the data. Most of the proposed i n t e r v e n t i o n s t r a t e g i e s do not 6 appear to i n c l u d e p o s s i b l e s t a t i s t i c a l i n t e r a c t i o n s between r i s k f a c t o r s i n the i n t e r v e n t i o n p l a n , or assume that a lowering of one w i l l l e a d to a general lowering of a l l r i s k f a c t o r s (some of the l i f e s t y l e i n t e r v e n t i o n approaches). These assumptions a r i s e , i n p a r t , from t h e o r e t i c a l c o n s i d e r a t i o n s , but they are a l s o r e f l e c t e d i n the models used to analyze the s t u d i e s from which these plans are based. However, other models may be more a p p r o p r i a t e , and these might c o n c e i v a b l y suggest r a d i c a l l y d i f f e r e n t i n t e r v e n t i o n s t r a t e g i e s . Before t h i s p o s s i b i l i t y can be examined the nature of " r i s k " must be looked a t . Just what c o n s t i t u t e s r i s k and being at r i s k i s an important, but o f t e n glossed-over, problem. As w i l l be seen, the lack of c o n s i d e r a t i o n of t h i s problem, combined with a major s h i f t in the method of data a n a l y s i s that took p l a c e a f t e r 1967 has played a l a r g e r o l e in the present day understanding of the r i s k f a c t o r s a s s o c i a t e d with the development of coronary heart d i s e a s e . 7 CHAPTER 2 Risk and C a u s a l i t y Given the amount of re s e a r c h and s c i e n t i f i c d i s c u s s i o n on the r i s k f a c t o r s f o r d i s e a s e s i n g e n e r a l , one would assume that the phrases "at r i s k " and " r i s k f a c t o r " have been w e l l d e f i n e d . However, Grundy (1973) has p o i n t e d out that the use of these phrases has changed over time. In 1973 "at r i s k " had come to mean: that an i n d i v i d u a l or community has an i n c r e a s e d p r o b a b i l i t y of developing a c o n d i t i o n or d e t e r i o r a t i o n in a c o n d i t i o n , when compared with the parent p o p u l a t i o n . (p. 1489) Once t h i s i n c r e a s e d p r o b a b i l i t y i s determined, attempts are made to f i n d how t h i s i n d i v i d u a l or community d i f f e r s ( i n which f a c t o r s they d i f f e r ) from the parent p o p u l a t i o n . F a c t o r s that have been shown to be s t a t i s t i c a l l y a s s o c i a t e d with the c o n d i t i o n are seen as r i s k f a c t o r s f o r the c o n d i t i o n . When a s p e c i f i c i n d i v i d u a l i s e i t h e r higher or lower (depending on the nature of the a s s o c i a t i o n ) than the average of the parent p o p u l a t i o n f o r one f a c t o r , he or she i s s a i d to be "at r i s k " . Note that t h i s statement does not make ref e r e n c e to the i n c i d e n c e or prevalence of the c o n d i t i o n , s i n c e the concept i s independent of the type of d i sease. A d i f f i c u l t problem a s s o c i a t e d with the determination of r i s k f a c t o r s i s d e c i d i n g which v a r i a b l e s should be looked 8 at as p o t e n t i a l r i s k i n d i c a t o r s . Seldom can a d i s e a s e be a t t r i b u t e d to one c a u s a l f a c t o r . Often a wide v a r i e t y of f a c t o r s i n f l u e n c e the d i s e a s e process, i n c l u d i n g host and environmental f a c t o r s that are not d i r e c t l y r e l a t e d to the e t i o l o g y of the d i s e a s e . For example, the degree of crowding i n f l u e n c e s an i n d i v i d u a l ' s chance of coming i n t o c ontact with an i n f e c t i o u s d i s e a s e , even though crowding i t s e l f may not i n f l u e n c e the course of the i n f e c t i o n once con t a c t e d . With a l i t t l e imagination i t i s easy to come up with hundreds of dimensions upon which human beings vary. I t i s ob v i o u s l y impossible to examine each p o s s i b l e dimension. A method that has been used to l i m i t the search i s the use of pro g n o s t i c c a t e g o r i e s ( H a l l & Zwemer, 1979). P r o g n o s t i c c a t e g o r i e s are c l i n i c a l l y recognized p r e c u r s o r s or suspected p r e c u r s o r s to the d i s e a s e , u s u a l l y d r a f t e d by expert p a n e l s . It should be noted that the d e l i n e a t i o n of these p r o g n o s t i c c a t e g o r i e s i s o f t e n based on previous r e s e a r c h f i n d i n g s and case s t u d i e s . Those c a t e g o r i e s that are found to be s t a t i s t i c a l l y a s s o c i a t e d with the dis e a s e are seen as r i s k f a c t o r s . No s i n g l e set of v a r i a b l e s has been accepted as r i s k f a c t o r s f o r coronary heart d i s e a s e : the l i s t v a r i e s from source to source. However, three f a c t o r s have come to be accepted as being importantly a s s o c i a t e d with coronary heart d i s e a s e . Most r e s e a r c h e r s and c l i n i c i a n s would agree that 9 high c i g a r e t t e consumption, high blood p r e s s u r e , and high l e v e l s of c h o l e s t e r o l are r i s k f a c t o r s f o r coronary heart d i s e a s e 1 . B i o r c k ' s (1975) l i s t of r i s k f a c t o r s f o r coronary heart d i s e a s e appears r e p r e s e n t a t i v e of the a d d i t i o n a l r i s k f a c t o r s r e p o r t e d i n at l e a s t some s t u d i e s (see Table 2.1). Table 2.1 A d d i t i o n a l r i s k f a c t o r s f o r coronary heart d i s e a s e (Biorck, 1975) Disease. (from B i o r c k , 1975) parent's e a r l y death "type A" p e r s o n a l i t y high t r i g l y c e r i d e s p h y s i c a l i n a c t i v i t y high hemocrit overweight high E.S.R. high a l c o h o l consumption high u r i c a c i d accumulated " l i f e changes" di a b e t e s chest pain; dyspnea s o f t tap water In a d d i t i o n , some would i n c l u d e male gender and age as r i s k f a c t o r s , s i n c e the in c i d e n c e r a t e f o r coronary heart d i s e a s e i s higher f o r males than females, and i n c r e a s e s with age. The above d e s c r i p t i o n of the concept of r i s k i s s t r a i g h t f o r w a r d , although somewhat t r i v i a l . If the e x i s t e n c e of some r i s k f a c t o r s f o r coronary heart disease i s accepted, i t would seem l o g i c a l to conclude that some i n d i v i d u a l s are more "at r i s k " than o t h e r s . E f f o r t s have been made to q u a n t i f y the degree of r i s k . T h i s task i s complicated by the v a r i o u s ways r i s k f a c t o r s can be measured. Some are o b v i o u s l y dichotomous, f o r example, sex. Others are, at l e a s t p o t e n t i a l l y , continuous (e.g. serum 1 As w i l l be seen below, t h i s i s an o v e r s i m p l i f i c a t i o n , e s p e c i a l l y i n the case of serum c h o l e s t e r o l . 10 c h o l e s t e r o l l e v e l ) . I t i s with v a r i a b l e s of the l a t t e r type that v a r y i n g degrees of r i s k seem a p p r o p r i a t e . The degree of r i s k a ssigned to a p a r t i c u l a r l e v e l of a r i s k f a c t o r depends on how much t h i s l e v e l d e v i a t e s from the p o p u l a t i o n average. Va r i o u s methods have been proposed f o r c a l c u l a t i n g the amount of r i s k a s s o c i a t e d with a c e r t a i n l e v e l of a r i s k f a c t o r , and the method of H a l l and Zwemer (1979) w i l l be d e s c r i b e d here for i l l u s t r a t i v e purpose's 1. They c o n t r o l for the e f f e c t s of age, sex, and race by comparing groups that c o n s i s t e d of the same general age, sex and r a c i a l c l a s s . They s t a t e that the r i s k f a c t o r f o r a p a r t i c u l a r p r o g n o s t i c c h a r a c t e r i s t i c and event i s the "odds r a t i o " f o r the event [ p / ( l - p ) where p i s the p r o b a b i l i t y of the occurrence of the event] in the members of the race-sex-age group with a p a r t i c u l a r p r o g n o s t i c c h a r a c t e r i s t i c d i v i d e d by the r i s k of the same event [ p / ( l - p ) ] in a l l members of the race-sex-age group. (p. 9) I t should be noted that H a l l and Zwemer reserve the term r i s k f a c t o r f o r the number that d e s c r i b e s the r i s k a s s o c i a t e d with a p a r t i c u l a r , p r o g n o s t i c category. T h i s seems n e e d l e s s l y c o n f u s i n g , given the more widely accepted meaning of the term " r i s k f a c t o r " , and in t h i s paper the c a l c u l a t e d number (by whatever method) w i l l be r e f e r r e d to as the weight attached to a p a r t i c u l a r r i s k f a c t o r . While the d e s c r i p t i o n s of "at r i s k " and " r i s k f a c t o r " are r e l a t i v e l y simple, attempts to use the concepts can lead 1 Methods used to c a l c u l a t e coronary heart d i s e a s e r i s k w i l l be d e s c r i b e d l a t e r . 11 to s i t u a t i o n s that can q u i c k l y become q u i t e complex. For example, not a l l p o t e n t i a l r i s k f a c t o r s can be measured with the same degree of accuracy, and thus some measures w i l l be o r d i n a l , others continuous, and some even c a t e g o r i c a l . The i n a b i l i t y to measure a v a r i a b l e with s u f f i c i e n t p r e c i s i o n can o f t e n l e a d to a s t a t i s t i c a l u n d e r p r e d i c t i o n of the importance of the v a r i a b l e . T h i s i s e s p e c i a l l y true when continuous v a r i a b l e s are reduced to some s o r t of o r d i n a l c l a s s i f i c a t i o n . In a d d i t i o n , i n d i v i d u a l s are o f t e n at r i s k on more than one v a r i a b l e , and t h i s c r e a t e s the problem of how to combine these v a r i a b l e s so as to f i n d a measure of o v e r a l l r i s k . A major issue a f t e r f i n d i n g a r e l a t i o n s h i p between a r i s k f a c t o r and coronary heart d i s e a s e i s that of c a u s a l i t y . Grundy (1973) noted the e x p r e s s i o n " r i s k f a c t o r " i s sometimes adopted as a l a b e l f o r any f a c t o r which has a s i g n i f i c a n t s t a t i s t i c a l a s s o c i a t i o n with the development of the d i s e a s e , and i t i s used to imply c a u s a l i t y , although ... some r i s k f a c t o r s may be m a n i f e s t a t i o n s of the d i s e a s e process r a t h e r than antecedent v a r i a b l e s , and o t h e r s , ... are e p i d e m i o l o g i c a l measurements f o r d e s c r i b i n g p o p u l a t i o n s a f f e c t e d r a t h e r than c a u s a l f a c t o r s , (p. 1489) The i m p l i c a t i o n of c a u s a l i t y can le a d to a great deal of misunderstanding. In some cases a r i s k f a c t o r i s c l e a r l y not c a u s a l , f o r example being male leads to an in c r e a s e d r i s k of coronary heart d i s e a s e , but i t i s u n l i k e l y that many would s t a t e that being male causes coronary heart d i s e a s e . Other s i t u a t i o n s are not so c l e a r . For example, high blood 1 2 pressure i s c l e a r l y a r i s k f a c t o r . The i m p l i c a t i o n that i t i s a c a u s a l f a c t o r has important r a m i f i c a t i o n s f o r any i n t e r v e n t i o n program. I f the r e l a t i o n s h i p i s c a u s a l , then the r e d u c t i o n of blood pressure w i l l l ead to a r e d u c t i o n of coronary heart d i s e a s e . I f not, then such a r e d u c t i o n w i l l not n e c e s s a r i l y have any b e n e f i c i a l e f f e c t (at l e a s t f o r coronary heart d i s e a s e ) . A r i s k f a c t o r need not be a c a u s a l f a c t o r i n the development of a d i s e a s e . The chances of f i n d i n g noncausal r i s k f a c t o r s would a l s o seem higher with degenerative d i s e a s e s that have a long time course, as there are l i k e l y many symptomatic stages along the route to the f i n a l event that has been named the outcome of the d i s e a s e process. V a r i a b l e s that are not i n the c a u s a l chain l e a d i n g to the outcome i n q u e s t i o n can s t i l l serve as r i s k f a c t o r s , s i n c e a l l that i s necessary i s that they be s t a t i s t i c a l l y a s s o c i a t e d with one of the elements of the c h a i n . T h i s means that i t i s p o s s i b l e to s t a t e that an i n d i v i d u a l with l e v e l y of blood pressure has a p r o b a b i l i t y p of developing coronary heart d i s e a s e without i t n e c e s s a r i l y f o l l o w i n g that i f blood pressure was m e d i c a l l y reduced by (y-x) a decrease in p would be observed. T h i s s i t u a t i o n can e x i s t because a r i s k f a c t o r need only e x h i b i t s t a t i s t i c a l a s s o c i a t i o n with the outcome. If i t i s simply an e a r l i e r outcome of the di s e a s e process i t w i l l e x h i b i t a c o r r e l a t i o n with the outcome ( i n our case, coronary heart d i s e a s e ) . Changing an 1 3 e a r l i e r outcome to a m e d i c a l l y a c c e p t a b l e l e v e l may not e f f e c t the coronary heart d i s e a s e process in any way. I t i s w e l l known that a s t a t i s t i c a l a s s o c i a t i o n alone does not demonstrate the d i r e c t i o n of any c a u s a t i o n , or even that a c a u s a l r e l a t i o n s h i p e x i s t s , but i t has n e v e r t h e l e s s been amply demonstrated how e a s i l y one can f a l l i n t o the t r a p of assuming c a u s a l i t y on the b a s i s of such an o b s e r v a t i o n . T h e o r e t i c a l l y , there are three general types of e x p l a n a t i o n s f o r the o b s e r v a t i o n that A i s s t a t i s t i c a l l y a s s o c i a t e d with B. I t c o u l d be that A i s part of the c a u s a l chain that r e s u l t s i n v a r i a t i o n in B, that B c o u l d cause A, or that some unknown f a c t o r C c o u l d cause both A and B. Temporal c o n s i d e r a t i o n s o f t e n can e l i m i n a t e one of the p o s s i b i l i t i e s : i f i t i s c o n s i s t e n t l y observed that event B f o l l o w s event A in time, then i t i s hard to argue that B causes A. But t h i s s t i l l l eaves two p o s s i b l e processes, and u s u a l l y these can only be d i f f e r e n t i a t e d on the b a s i s of experimentation. If A i s part of the c a u s a l c h a i n of B, then manipulation of A w i l l r e s u l t in changes in B. If d i r e c t l y manipulating A does not l e a d to changes in B then the hypothesis that a f a c t o r C e x i s t s seems more tenable (assuming that i t has a l r e a d y been determined that B does not cause A). There are numerous d i s c u s s i o n s on what i t takes to e s t a b l i s h c a u s a l i t y (Weiss, 1981; MacMahon and Pugh, 1970). One p a r t i c u l a r i l y v a l u a b l e d e s c r i p t i o n i s presented by 1 4 M o s t e l l e r and Tukey (1977). T h e i r book i s concerned with r e g r e s s i o n a n a l y s i s (the primary s t a t i s t i c a l method used to d e r i v e r i s k f a c t o r c o e f f i c i e n t s ) , and they p o i n t out that three " s o r t s of ideas" are u s u a l l y r e q u i r e d to support the notion of cause: 1. Con s i s t e n c y . That when other t h i n g s are equal in the p o p u l a t i o n we examine, the r e l a t i o n between x and y i s c o n s i s t e n t across p o p u l a t i o n s i n d i r e c t i o n — perhaps even in amount. 2. Responsiveness. That i f we can interve n e and change x for some i n d i v i d u a l s , t h e i r y's w i l l respond a c c o r d i n g l y . 3. A mechanism. That there i s a mechanism, which someone might sometime understand through which "cause" i s r e l a t e d , o f t e n step by step, with the " e f f e c t " -- the so r t of mechanism where, at each step, i t would be n a t u r a l to say " t h i s causes t h a t " , (p. 260-261) Only the f i r s t c r i t e r i o n can be met from s t r i c t l y o b s e r v a t i o n a l data. The second u s u a l l y can only be met through experimentation of some kind (and t h i s may be hard to do). The f i n a l c r i t e r i o n r e p r e s e n t s the goal of theory development. However only the f i r s t c r i t e r i o n need be met for a v a r i a b l e to be c a l l e d a r i s k f a c t o r . The above d i s c u s s i o n i s not meant to imply that the major goal of r i s k e v a l u a t i o n must be the establishment of c a u s a l i t y . The importance of e s t a b l i s h i n g c a u s a l i t y depends on the goals of the i n v e s t i g a t o r . From the p e r s p e c t i v e of sc i e n c e , t h i s i s the most important goal of a l l , and the s c i e n t i f i c community i s very demanding in terms of the amount of evidence r e q u i r e d to demonstrate c a u s a l i t y . 15 E p i d e m i o l o g i s t s , when t r y i n g to demonstrate c a u s a l i t y , have to meet these demands. However, when developing p r e v e n t i v e programs, e s p e c i a l l y where an i n t e r v e n t i o n technique e x i s t s , concern about c a u s a l i t y may not be as g r e a t . MacMahon and Pugh (1970) p o i n t out that the p r a c t i c a l s i g n i f i c a n c e of c a u s a l a s s o c i a t i o n s i n the development of p r e v e n t i o n programs does not n e c e s s a r i l y depend on the degree of d i r e c t n e s s . F i r s t , more d i r e c t a s s o c i a t i o n s may not yet have been i d e n t i f i e d and so there may be no c h o i c e but to make use of o b v i o u s l y i n d i r e c t a s s o c i a t i o n s i n p r e v e n t i v e programs. ... Second, more d i r e c t causes, although known, may not be s u s c e p t i b l e to economic a l t e r a t i o n , whereas the i n d i r e c t ones may be. (p. 23) T h i s can be extended f u r t h e r . In slow, degenerative d i s e a s e s u s e f u l i n t e r v e n t i o n s may e x i s t , but these i n t e r v e n t i o n s may be of no use in h e l p i n g to i d e n t i f y those i n d i v i d u a l s who are at r i s k . In t h i s s i t u a t i o n whatever measures provide the most r e l i a b l e p r e d i c t i o n of l a t e r d i s e a s e s t a t u s w i l l be the most u s e f u l r i s k f a c t o r s , no matter i f they are part of the c a u s a l chain or i f simply a s s o c i a t e d with some part of the c h a i n . The goals of those who have done or are p r e s e n t l y doing rese a r c h i n the area of coronary heart d i s e a s e o f t e n d i f f e r . For example, attempts are being made to e s t a b l i s h c a u s a l r e l a t i o n s h i p s between v a r i a b l e s , to develop methods of i d e n t i f y i n g high r i s k i n d i v i d u a l s , and to f i n d the most e f f e c t i v e medical i n t e r v e n t i o n . Researchers w i l l use whatever rese a r c h i s a v a i l a b l e to achieve t h e i r p a r t i c u l a r g o a l . At present there does not appear to be any d e f i n i t i v e 16 study, or groups of s t u d i e s , that achieve a l l these g o a l s . The focus of much of the e p i d e m i o l o g i c a l coronary heart d i s e a s e r e s e a r c h i s on the i d e n t i f i c a t i o n of r i s k f a c t o r s . With the i n t r o d u c t i o n of the l o g i s t i c model as a data a n a l y s i s technique r e s e a r c h e r s had a powerful method of i d e n t i f y i n g "important" r i s k f a c t o r s . I t i s b e l i e v e d that the d i s c u s s i o n of the c a u s a l f a c t o r s f o r coronary heart d i s e a s e has been g r e a t l y i n f l u e n c e d by the i m p l i e d c a u s a l i t y of the term r i s k f a c t o r , and the i m p l i e d c a u s a l i t y of the r e s u l t s of the m u l t i p l e l o g i s t i c a n a l y s e s . Thus i t seems necessary to look at the evidence that l i n k s r i s k f a c t o r s to the development of coronary heart d i s e a s e , both in terms of the r e l i a b i l i t y and c o n s i s t e n c y of the f i n d i n g s , and i n terms of the c a u s a l r e l a t i o n s h i p s that might e x i s t between the r i s k f a c t o r and the outcome. 1 7 CHAPTER 3 H i s t o r i c a l Development E p i d e m i o l o g i c a l r e s e a r c h on the e t i o l o g y of coronary heart d i s e a s e has not had a long h i s t o r y . L e i b o w i t z (1970) has d e s c r i b e d the h i s t o r i c a l development of medical awareness and d i a g n o s i s of coronary heart d i s e a s e . I t appears that at l e a s t some i n d i v i d u a l s i n v o l v e d in the p r a c t i c e of medicine have been aware of cases of sudden death a s s o c i a t e d with sharp chest pain s i n c e the time of an c i e n t Egypt. O b s t r u c t i o n of the a r t e r i e s was suspected as a cause of death due to "pain i n the he a r t " i n the 1700's. ( I t was not u n t i l the twent i e t h century that the phrase coronary heart d i s e a s e came to be used. Before t h i s , names for the dis e a s e v a r i e d , with r e f e r e n c e o f t e n being made to di s e a s e s of the heart or c h e s t . From the case d e s c r i p t i o n s of the i n c i d e n t s , most of these would now be l a b l e d some form of coronary heart d i s e a s e . ) Reports w r i t t e n in 1649 and 1775 noted that anger and strong emotion were f a c t o r s which o f t e n preceded sudden death that came with chest p a i n . The term "angina p e c t o r i s " was f i r s t used i n the eightee n t h century. In the nine t e e n t h century the word " f a t t y " was used to d e s c r i b e the a r t e r i a l degeneration of a p a t i e n t who, from the d e s c r i p t i o n of the case, d i e d from some form of heart d i s e a s e . Despite t h i s awareness, the medical p r o f e s s i o n as a 18 whole was slow to recognize coronary heart d i s e a s e as a t r e a t a b l e medical problem. Some conception of r i s k f a c t o r s f o r coronary heart disease e x i s t e d as e a r l y as 1842, as shown in a paper by P.M. Latham 1. Latham d e s c r i b e d an i n c i d e n t that took plac e between a p a t i e n t (Thomas Arnold, a w e l l known and respected e d u c a t i o n a l i s t ) and h i s medical attendant (Dr. Buchwell). At the time A r n o l d was one day short of 47 years o l d and a p p a r e n t l y i n good h e a l t h . Dr. Buchwell was summoned at 7:00 am., with A r n o l d complaining of a severe chest pain that had s t a r t e d around 5:00 am. The f o l l o w i n g c o n v e r s a t i o n r e p o r t e d l y took place (the n a r r a t o r i s Dr. Buchwell): I then asked him, i f any of h i s f a m i l y ever had any d i s e a s e of the chest? "Yes, my f a t h e r had; he d i e d of i t . " He i n q u i r e d i f d i s e a s e of the heart was suddenly f a t a l ? I answered that i t was. "Was i t a common d i s e a s e ? " I s a i d not common. "Where do you f i n d i t most?" "In l a r g e towns I t h i n k . " "Why?" "Perhaps from a n x i e t y and eager competition among the higher and intemperence among the lower c l a s s e s . " ( L e i b o w i t z , 1970, p. 120 ) 2 I t i s not meant to suggest that the medical p r o f e s s i o n was aware of the d i a g n o s i s of coronary heart d i s e a s e and knew the r i s k f a c t o r s a s s o c i a t e d with i t in 1842. At t h i s time only p h y s i c i a n s e s p e c i a l l y i n t e r e s t e d in the strange cases of sudden death were aware of the symptoms of coronary heart d i s e a s e , and they were not able to diagnose or t r e a t i t with 1 T h i s paper i s reproduced in L e i b o w i t z (1970), and the quote i s taken from t h e r e . 2 A r n o l d d i e d about 1/2 hour l a t e r . 19 any degree of success. I t was not u n t i l the 1920's that a l l p h y s i c i a n s were aware of, and c o u l d diagnose, coronary heart d i s e a s e ( L e i b o w i t z , 1970). However, i t i s i n t e r e s t i n g t h at common o b s e r v a t i o n l e d some p h y s i c i a n s to suspect f a c t o r s that are seen by some as r i s k f a c t o r s today ( f a m i l y h i s t o r y of coronary heart d i s e a s e , s t r e s s f u l l i v i n g or a l c o h o l consumption). L e i b o w i t z a l s o d e s c r i b e d the development of the r e c o g n i t i o n of other symptoms i n the e a r l y 1900's. Low blood pressure was n o t i c e d to be a s s o c i a t e d with p a t i e n t s s u f f e r i n g a coronary i n the 1910's. Animal experimentation s t a r t e d i n the 1800's. I t i n i t i a l l y c o n s i s t e d of an examination of the e f f e c t of t y i n g o f f a r t e r i e s on the a c t i o n of the heart, but in 1908 i t was demonstrated that animals fed a d i e t of milk and egg yolk developed severe a r t e r i o s c l e r o s i s . In 1913 i t was re p o r t e d that the a d d i t i o n of pure c h o l e s t e r o l to r a b b i t food would a l s o r e s u l t in a r t e r i o s c l e r o s i s . (Leibowitz r e p o r t s that i t was not u n t i l 1946 that i t was d i s c o v e r e d that these r e s u l t s may not apply to higher mammals.) The above d e s c r i p t i o n i s meant to demonstrate t h a t , while i t of t e n appears that coronary heart d i s e a s e "exploded" on the Western World in the 1920's, the problem has e x i s t e d for a longer p e r i o d of time. Medicine i t s e l f was not t o t a l l y unaware of the problem, but medical r e s e a r c h was focused on many other q u e s t i o n s . The in f o r m a t i o n 20 a v a i l a b l e was, admittedly, scanty, c o n s i s t i n g mainly of s u b j e c t i v e c l i n i c a l r e p o r t s , but i t d i d give e p i d e m i o l o g i s t s a h i n t as to some of the important v a r i a b l e s when resea r c h s t a r t e d i n t h i s area. There does not seem to be one p a r t i c u l a r p o i n t in time when v a r i a b l e s a s s o c i a t e d with coronary heart disease became seen as r i s k f a c t o r s . Rather, the acceptance has been a gradual process, with those authors i n v o l v e d i n the i n i t i a l l a r g e s c a l e s t u d i e s a c c e p t i n g both t h e i r r e s u l t s and the r e s u l t s of e a r l y c l i n i c a l s t u d i e s as i n d i c a t i n g the presence of r i s k f a c t o r s , and then with l a t e r r e s e a r c h e r s r e f e r r i n g to these v a r i a b l e s i n t h i s manner. As e a r l y as 1962 i t was p o s s i b l e to s t a t e , i n a p u b l i s h e d paper, that an a s s o c i a t i o n was known to e x i s t between the v a r i a b l e s serum c h o l e s t e r o l , blood pressure and the development of coronary heart d i s e a s e without c i t i n g any p r e v i o u s r e s e a r c h (see C o r n f i e l d , 1962). Some of the r e s e a r c h e r s who p u b l i s h e d r e s u l t s p r i o r to 1960 seem much more c a u t i o u s i n the i n t e r p r e t a t i o n of t h e i r r e s u l t s than do present day i n t e r p r e t a t i o n s of the same s t u d i e s . The p r e f e r r e d method of data a n a l y s i s f o r p r o s p e c t i v e coronary heart d i s e a s e s t u d i e s underwent a major r e v i s i o n with the i n t r o d u c t i o n of the l o g i s t i c f u n c t i o n and i t s m u l t i v a r i a t e approach (Walker and Duncan, 1967; T r u e t t , C o r n f i e l d , and Kannel, 1967). T h i s f u n c t i o n can be used as a convenient c l a s s i f i c a t i o n c r i t e r i o n f o r o r g a n i z i n g the 21 i n f o r m a t i o n coming out of these s t u d i e s . Those r e p o r t s that used the more t r a d i t i o n a l e p i d e m i o l o g i c a l approach to a n a l y s i s ( u s u a l l y p u b l i s h e d before or around 1967) can be seen as f a l l i n g i n t o the h i s t o r i c a l s e c t i o n of the development of the r i s k f a c t o r s . The i n t r o d u c t i o n of the l o g i s t i c model r e s u l t e d i n some major changes in the way coronary heart disease r i s k f a c t o r s were determined, and thus these s t u d i e s are presented i n a separate s e c t i o n . Compared to the enormous number of p u b l i s h e d a r t i c l e s d e a l i n g with coronary heart d i s e a s e , the a c t u a l number of l a r g e s c a l e p r o s p e c t i v e or r e t r o s p e c t i v e s t u d i e s i s r e l a t i v e l y s m a l l . The long time course for the development of the disease i s one reason for the small number of s t u d i e s , s i n c e r e s e a r c h i s simply very expensive and time consuming. However, most of these s t u d i e s have produced many i n d i v i d u a l r e p o r t s d e s c r i b i n g the r e l a t i o n s h i p of e i t h e r one v a r i a b l e , or a group of s i m i l a r v a r i a b l e s , to the development of coronary heart d i s e a s e . Reports of the e f f e c t s of d i f f e r e n t types of v a r i a b l e s o f t e n appear in a v a r i e t y of j o u r n a l s under a v a r i e t y of d i f f e r e n t a u t h o r s 1 . In a d d i t i o n , r e p o r t s d e s c r i b i n g the r e s u l t s obtained up to the date of w r i t i n g the paper f o r p u b l i c a t i o n are o f t e n r e l e a s e d by those i n v o l v e d i n the s t u d i e s . That i s , when s u f f i c i e n t data e x i s t s f o r an i n i t i a l a n a l y s i s to be made, a 1 For example, the Framingham study has produced 169 p u b l i c a t i o n s and over 30 r e p o r t s between the years 1951 and 1978 (Dawber, 1980). 22 report d e s c r i b i n g these r e s u l t s i s r e l e a s e d . At a l a t e r p o i n t i n time, when enough a d d i t i o n a l data has been c o l l e c t e d , a second r e p o r t i s r e l e a s e d , c o n t a i n i n g r e s u l t s c a l c u l a t e d from the new data and the p r e v i o u s l y analyzed data s e t . Given the magnitude of the problem of coronary heart d i s e a s e the attempt to get r e s u l t s out as soon as p o s s i b l e i s understandable, but t h i s c r e a t e s c e r t a i n problems f o r reviewers of the l i t e r a t u r e 1 . Due to t h i s m u l t i p l e r e p o r t i n g the l i t e r a t u r e reviewed here w i l l be d i v i d e d i n t o s e c t i o n s . In t h i s s e c t i o n i n t e r e s t l i e s p r i m a r i l y with the e a r l y evidence su p p o r t i n g the e x i s t e n c e of c e r t a i n r i s k f a c t o r s f o r coronary heart d i s e a s e . A t t e n t i o n w i l l p r i m a r i l y be focused on American p r o s p e c t i v e s t u d i e s s t a r t e d i n the 1950's, but* r e t r o s p e c t i v e s t u d i e s w i l l be brought i n t o the d i s c u s s i o n where r e l e v a n t . In t h i s s e c t i o n the r i s k f a c t o r s w i l l be examined one-by-one, thus s p e c i f i c s t u d i e s w i l l be r e f e r r e d to each time they have r e s u l t s p e r t a i n i n g to the v a r i a b l e being d i s c u s s e d . In a l a t e r s e c t i o n the r e s u l t s of s t u d i e s using a m u l t i v a r i a t e approach to data a n a l y s i s w i l l be presented. Given the l a r g e number of p u b l i c a t i o n s , the s e l e c t i o n of papers that g i v e an o v e r a l l d e s c r i p t i o n of the study i s sometimes d i f f i c u l t . An attempt has been made to c i t e the 1 T h i s a l s o has the e f f e c t of making i t appear that more i s known about coronary heart d i s e a s e than i s a c t u a l l y the case. Often i t i s not c l e a r that a l a r g e number of p u b l i c a t i o n s are a l l drawing on the same data s e t . 23 l a s t major p u b l i c a t i o n r e l e a s e d by the i n v e s t i g a t o r s of each study where: 1. i t r e p o r t e d the r e s u l t s of most of the v a r i a b l e s examined in the study, and 2. i t was not p r i m a r i l y concerned with the m u l t i v a r i a t e a n a l y s i s of the r e s u l t s . I The American P r o s p e c t i v e S t u d i e s Six major p r o s p e c t i v e s t u d i e s of coronary heart d i s e a s e were i n i t i a t e d i n the United S t a t e s i n the l a t e 1950's and e a r l y 1960's: the Framingham study (Dawber, Meadors and Moore, 1951) 1, the Albany study ( H i l l i b o e , James and Doyle, 1954), the Los Angeles study (Chapman, Goerke, Dixon, Loveland and P h i l l i p s , 1957), the Chicago s t u d y 2 (Paul, Lepper, Phelan, Dupertuis, MacMillan, McKean and Park, 1963), the Minnesota study (Keys, T a y l o r , Blackburn, Brozek, Anderson and Simonson, 1963) and the Western C o l l a b o r a t i v e study (Rosenman, Friedman, Straus, Wurm, Kos i t c k e k , Hahn and Worthessen, 1964). Since the Framingham study has the longest and most d e t a i l e d a n a l y s i s of the s i x , two sets of i t s r e s u l t s w i l l be presented here; those of a four year 1 These r e f e r e n c e s are f o r the o r i g i n a l p u b l i c a t i o n s that d e s c r i b e the procedure and r a t i o n a l e used i n each study. 2 Often r e f e r e d to as the Chicago Western E l e c t r i c study so as to d i s t i n g u i s h i t from the Chicago Gas study ran around the same time. T h i s l a t t e r study i s not reviewed here. 24 follow-up (which correspond to the general l e n g t h of the other s t u d i e s ) and those of an 18 year follow-up, which w i l l serve as an i n d i c a t o r of the " s t a b i l i t y " of the r e s u l t s r e p o r t e d from these s t u d i e s . A The Albany Study R e s u l t s from the Albany study can be found i n Doyle, H e s l i n , H i l l i b o e and Formel (1959). T h i s study s t a r t e d i n 1953 and examined 1913 men between the ages of 39 and 55. The s u b j e c t s were male New York State c i v i l s e r v i c e employees l i v i n g i n the Albany area. At the i n i t i a l examination they found 70 i n d i v i d u a l s who showed evidence of coronary heart d i s e a s e (a prevalence of 36.6/1000 pe r s o n s ) . These were dropped from the study p o p u l a t i o n . During the 44 months between the f i r s t and f o u r t h examination 57 cases of coronary heart d i s e a s e were observed, an annual i n c i d e n c e rate of 8.4/1000 persons. B The Los Angeles Study Chapman and Massey (1964) present the r e s u l t s of ten years of follow-up data from the Los Angeles heart study. The study sample c o n s i s t e d of 2252 persons (1859 male and 393 female) randomly s e l e c t e d w i t h i n s t r a t a (the s t r a t i f i c a t i o n was by age, sex, race and job c l a s s i f i c a t i o n ) from the 20,199 c i v i l s e r v i c e employees of the c i t y of Los 25 Angeles. In order to ob t a i n the necessary number of s u b j e c t s i n each s t r a t a a second name was drawn from the same s t r a t a f o r any i n d i v i d u a l who re f u s e d to p a r t i c i p a t e . The f i n d i n g s r e p o r t e d i n t h i s s e c t i o n are f o r the 1552 white males (age range of 21-70) i n the sample. During the i n i t i a l examination 49 cases of coronary heart d i s e a s e were diagnosed (a prevalence of 31.6/1000 persons), reducing the sample to 1503. A f t e r ten years of follow-up the study had observed 78 deaths from coronary heart d i s e a s e , a death rate of 5.0/1000 per year. (The m o r t a l i t y rate was 11.6/1000 f o r deaths from a l l causes.) Over the ten years 177 cases of coronary heart d i s e a s e were diagnosed, a m o r b i d i t y rate of 11.8/1000 per year 1. C The Chicago Study Both the i n i t i a l d e s c r i p t i o n and summary of the major r e s u l t s of the Chicago study can be found in Paul et a l . (1963). The p o p u l a t i o n s t u d i e d c o n s i s t e d of employees of the Hawthorne Works of the Western E l e c t r i c Company. A t o t a l of 2080 men agreed to p a r t i c i p a t e (67% of the men who were approached). The i n i t i a l examination d i s c o v e r e d 44 i n d i v i d u a l s who showed some evidence of coronary heart d i s e a s e (a prevalence of 21.2/1000 persons), and 47 men were l o s t to follow-up (these were dropped from the a n a l y s i s ) , 1 The authors report a r a t e of 10.6/1000. No e x p l a n a t i o n i s given f o r t h i s presumably i n c o r r e c t value. 26 l e a v i n g a sample s i z e of 1989. A f t e r four years and f i v e months of follow-up 88 cases of coronary heart d i s e a s e had been observed, a m o r b i d i t y r a t e of 10.0/1000 persons per year. The m o r t a l i t y r a t e (deaths from a l l causes) was report e d as 4.3/1000 per year (38 de a t h s ) . D The Minnesota Study A major d e s c r i p t i o n of t h i s study, along with the r e s u l t s a f t e r 15 years of follow-up, can be found i n Keys et a l . (1963). T h i s study focused on p r o f e s s i o n a l men between the ages of 45 and 55. Not a l l of the p a r t i c i p a n t s were randomly s e l e c t e d from the t a r g e t p o p u l a t i o n ; some were s p e c i f i c a l l y s e l e c t e d because of c e r t a i n c h a r a c t e r i s t i c s of p a r t i c u l a r i n t e r e s t (e.g. very obese). Three hundred men were o r i g i n a l l y examined, with 23 being e l i m i n a t e d f o r a v a r i e t y of medical reasons, i n c l u d i n g the presence of coronary heart disease (how many of those e l i m i n a t e d had coronary disease was not r e p o r t e d ) . Some i n d i v i d u a l s dropped out a f t e r the f i r s t examination, l e a v i n g a sample s i z e of 281 men. A f t e r 15 years of follow-up 17 deaths from coronary heart disease had been observed, an annual rate of 4.0/1000 men (32 deaths in t o t a l were observed, a rate of 7.6/1000). A t o t a l of 48 cases of coronary heart d i s e a s e were observed (an annual rate of 11.4/1000), made up of 32 d e f i n i t e cases and 16 p o s s i b l e cases. 27 E The Western C o l l a b o r a t i v e Study T h i s study obtained s u b j e c t s from 11 d i f f e r e n t business e s t a b l i s h m e n t s ; nine i n the San F r a n c i s c o - O a k l a n d area, two from Los Angeles (Rosenman, et a l . , 1964). The men were between the ages of 39 and 59. Each p o t e n t i a l s u b j e c t (an employee of one of the businesses) was i n v i t e d , by l e t t e r , to j o i n the study i f he was f r e e of coronary heart d i s e a s e or other s e r i o u s i l l n e s s e s . D e s p i t e the s c r e e n i n g l e t t e r , 113 of the 3524 men screened showed some symptoms of coronary heart disease (a prevalence of 32.1/1000) and were dropped from the study p o p u l a t i o n . In a d d i t i o n , 78 were o u t s i d e the age range of i n t e r e s t . Over the course of the study 45 s u b j e c t s were l o s t to follow-up, and 106 were l o s t when one f i r m withdrew, l e a v i n g a t o t a l sample of 3182. A f t e r four and one-half years of follow-up 133 cases of coronary heart disease had been observed, an annual i n c i d e n c e of 9.3/1000 men. F The Framingham Study The e a r l y r e s u l t s from the Framingham study can be found in Dawber, Moore and Mann (1957), and the r e s u l t s of 18 years of follow-up i n S h u r t l e f f (1974). The sampling procedure i n v o l v e d randomly s e l e c t i n g 2/3 of the a d u l t p o p u l a t i o n (between the ages 29 and 62) of the town of Framingham, Mass. Of the 6,510 people contacted 4469 agreed 28 to p a r t i c i p a t e (2024 men and 2445 women). During the i n i t i a l examination 48 men (prevalence of 23.7/1000) and 28 women (prevalence of 11.5/1000) were found to have some form of coronary heart d i s e a s e . The remaining sample was supplemented by 734 dis e a s e f r e e v o l u n t e e r s , b r i n g i n g the t o t a l sample up to 5127 (2283 men and 2844 women). A f t e r four years of follow-up the study had observed 48 cases of coronary heart d i s e a s e among the males, an average y e a r l y r a t e of 5.9/1000. Twenty-eight cases were observed among the females (a rate of 2.9/1000). A f t e r 18 years of f o l l o w -up 383 cases had been observed among the men, a r a t e of 9.4/1000 1. Among the women 244 cases had developed (4.8/1000). 11 Serum C h o l e s t e r o l F i n d i n g s The suspected r e l a t i o n s h i p between serum c h o l e s t e r o l and coronary heart d i s e a s e has gone through s e v e r a l stages dur i n g i t s h i s t o r y as a p o t e n t i a l r i s k f a c t o r . The i n i t i a l i n t e r e s t in t h i s v a r i a b l e arose out of the animal s t u d i e s mentioned p r e v i o u s l y . Some r e s e a r c h e r s i n t e r p r e t e d these r e s u l t s as i n d i c a t i n g that serum c h o l e s t e r o l might be the 1 Note that t h i s f i g u r e i s not based on person years of follow-up, a much b e t t e r estimate given the long p e r i o d of the study. Since i t i s not c l e a r that the e a r l y s t u d i e s used person years as the measure of follow-up t h i s r e s u l t has been c a l c u l a t e d on years of follow-up. Thus t h i s f i g u r e should be used simply f o r comparison purposes. The rate reported by S h u r t l e f f i s 14.7/1000 person years. 29 cause of coronary heart d i s e a s e . Thus i t appeared that the r i s k of coronary heart d i s e a s e c o u l d be reduced through a l t e r a t i o n of i n d i v i d u a l d i e t p r a c t i c e s . T h i s l e d to i n v e s t i g a t i o n of two i s s u e s ; the r e l a t i o n s h i p between human c h o l e s t e r o l l e v e l s and development of heart d i s e a s e , and the r e l a t i o n s h i p between d i e t a r y c h o l e s t e r o l and blood c h o l e s t e r o l l e v e l s . I n t e r e s t here l i e s p r i m a r i l y with the f i r s t q u e s t i o n , but i t should be noted that the second q u e s t i o n appears to be s e t t l e d . D i e t a r y c h o l e s t e r o l can i n f l u e n c e blood c h o l e s t e r o l l e v e l s , but other d i e t a r y f a t s must a l s o be present. C h o l e s t e r o l alone i s not s u f f i c i e n t to r a i s e c h o l e s t e r o l l e v e l s in man (Levy and E r n s t , 1973) 1 . In humans there are two sources of c h o l e s t e r o l (hence of l i p o p r o t e i n s ) : that taken in from an exogenous source ( d i e t ) and that produced i n t e r n a l l y (endogenously produced i n the l i v e r and i n t e s t i n e ) . Exogenous c h o l e s t e r o l w i l l i n h i b i t the pro d u c t i o n of endogenous c h o l e s t e r o l . However, the endogenous system cannot, by i t s e l f , produce enough c h o l e s t e r o l to meet the body's needs, so that i f exogenous intake f a l l s below 300 mg/day the plasma c h o l e s t e r o l l e v e l s w i l l decrease. C h o l e s t e r o l l e v e l s appear to inc r e a s e with age, at l e a s t in Western c u l t u r e s . Thus i t may be impossible to s t a t e a u n i v e r s a l c h o l e s t e r o l r i s k l e v e l ; i n s t e a d the l e v e l s may have to be age s p e c i f i c . Human 1 T h i s i s the source f o r the biochemical d e s c r i p t i o n that f o l l o w s . 30 c h o l e s t e r o l l e v e l s a l s o appear to be a f f e c t e d by strong emotion and s t r e s s f u l s i t u a t i o n s . One important area that i s sometimes ne g l e c t e d i n the d i s c u s s i o n of r i s k f a c t o r s i s the s t a b i l i t y of the measure over time. T h i s i s e s p e c i a l l y important when i n t e r e s t l i e s with c h r o n i c d i s e a s e s , f o r i f the measured r i s k f a c t o r s prove to be v a r i a b l e over time, then p r e d i c t i o n i s almost i m p o s s i b l e . Gordon and S h u r t l e f f (1973) have examined the s t a b i l i t y of s e l e c t e d measures taken in the Framingham study. For serum c h o l e s t e r o l , the c o r r e l a t i o n between the f i r s t measure taken, and that taken two years l a t e r was .732 ( f o r men) and .744 ( f o r women). The c o r r e l a t i o n between the f i r s t measure and that taken 18 years l a t e r decreased to .548 and .547 r e s p e c t i v e l y . Given a time span of 18 years between measures, a c o r r e l a t i o n of .55 seems to imply a reasonably s t a b l e measure. The amount that serum c h o l e s t e r o l l e v e l s i n c r e a s e over time was a l s o examined. They found that the modal v a r i a t i o n between the f i r s t and eighteen year measure ( i . e . the d i f f e r e n c e between each i n d i v i d u a l ' s two measures) was in the range 60-69 mg/100ml-. Thus i t would appear that one c o u l d expect an i n d i v i d u a l ' s c h o l e s t e r o l measure to remain s t a b l e r e l a t i v e to other members of the p o p u l a t i o n , and to i n c r e a s e (along with everyone e l s e ' s ) as he or she grew o l d e r . I t i s not c l e a r i f the i n c r e a s e of serum c h o l e s t e r o l with age i s part of the n a t u r a l aging process or r e f l e c t s 31 the accumulated e f f e c t s of a s p e c i f i c type of l i f e s t y l e . No comparative s t u d i e s appear to e x i s t that examine the changes in c h o l e s t e r o l l e v e l s over time i n d i f f e r e n t c u l t u r e s . T h i s q u e s t i o n has important consequences f o r the qu e s t i o n of the ca u s a l r e l a t i o n s h i p between serum c h o l e s t e r o l l e v e l s and coronary heart d i s e a s e , f o r i f i t i s part of the aging process i t might be s p u r i o u s l y c o r r e l a t e d with coronary heart d i s e a s e . T h i s problem does not, however, n e c e s s a r i l y e f f e c t the p r e d i c t i v e a b i l i t y of the v a r i a b l e . As w i l l be seen below, i n i t i a l high l e v e l s of serum c h o l e s t e r o l are a s s o c i a t e d with the l a t e r development of coronary heart d i s e a s e . Since the increase with age i s r e l a t i v e l y constant f o r the general p o p u l a t i o n i t does not matter (from the p e r s p e c t i v e of p r e d i c t i o n ) i f the higher l e v e l s of some i n d i v i d u a l s r e f l e c t the f a c t that they are b i o l o g i c a l l y " o l d e r " than other members of t h e i r c u l t u r e (and thus have higher c h o l e s t e r o l l e v e l s than t h e i r c h r o n o l o g i c a l age group) or i f the l e v e l i s due to l i f e s t y l e (and i s a c t u a l l y causing the development of coronary heart d i s e a s e ) . T h i s o b v i o u s l y does matter to those concerned with p r e v e n t i o n . One of the e a r l i e s t attempts to assess the p r e d i c t i v e a b i l i t y of c h o l e s t e r o l measures for the l a t e r development of coronary heart d i s e a s e was made by the T e c h n i c a l Group of the Committee of L i p o p r o t e i n s and A t h e r o s c l e r o s i s (1956). T h i s m u l t i c e n t r e t r i a l managed to c o l l e c t about 15,000 serum l i p i d measures over a three year p e r i o d . The study was 32 p r i m a r i l y i n t e r e s t e d i n the development of coronary heart d i s e a s e i n men between the ages of 49 and 59. Four thousand nine hundred and fourteen men were w i t h i n t h i s age range, were judged c l i n i c a l l y normal, and r e c e i v e d a c l i n i c a l f ollow-up e i t h e r one or two years a f t e r the i n i t i a l examination. In the judgment of an independent review committee, 82 of these i n d i v i d u a l s had developed c l i n i c a l symptoms of a t h e r o s c l e r o s i s . If an average of 1.5 years follow-up i s assumed, t h i s produces a m o r b i d i t y r a t e of 11.1/1000 persons per year. S i x t y - f i v e of the 82 cases were c l a s s i f i e d as d e f i n i t e new events, a rate of 8.8/1000. Since t h i s was a m u l t i c e n t r e t r i a l i t i s i n t e r e s t i n g to note that a wide v a r i a t i o n was observed in the mean c h o l e s t e r o l readings obtained from each c e n t r e . In order to assess the p r e d i c t i v e a b i l i t y of c h o l e s t e r o l they converted the c h o l e s t e r o l readings i n t o p e r c e n t i l e s , and examined the number of events in the group above the f i f t i e t h p e r c e n t i l e , and above the s e v e n t y - f i f t h . Table 3.1 presents these r e s u l t s . Since the mean c h o l e s t e r o l reading v a r i e d from centr e to ce n t r e the r e s u l t s have a l s o been d e s c r i b e d f o r each c e n t r e . If c h o l e s t e r o l was a good p r e d i c t o r , then as the sample was reduced to higher p e r c e n t i l e l e v e l s the m o r b i d i t y rate due to coronary heart d i s e a s e should i n c r e a s e . The only c e n t r e to f o l l o w t h i s trend was P i t t s b u r g . I t should be noted that the 41 cases above the f i f t i e t h p e r c e n t i l e i s the 33 Table 3.1 Number of new events c l a s s i f i e d by the number of i n d i v i d u a l s e i t h e r above the f i f t i e t h or s e v e n t y - f i f t h p e r c e n t i l e , by s i t e of c e n t r e . Adapted from Tables 11 and 14, T e c h n i c a l Group (1956). Centre T o t a l CHD cases # Men Observ Rate /1 000 # CHD Choi > 50% Rate /I 000 # CHD Choi > 75% Rate /1 000 A l l 57 3985 9.5 41 13.7 20 13.4 C l e v . 6 475 8.4 4 11.2 2 11.2 Don. 1 1 7 1 1 63 9.7 1 2 13.8 4 9.2 Don. 2 1 1 541 13.6 7 17.3 2 9.9 Harv. 8 799 6.7 6 10.0 .3 10.0 P i t t . 1 5 1 007 9.9 1 2 15.9 9 23.8 only r a t e that i s s t a t i s t i c a l l y s i g n i f i c a n t l y d i f f e r e n t from the expected r a t e . The members of the committee were s p l i t on how to i n t e r p r e t these r e s u l t s . One group f e l t that i t was p o s s i b l e to i d e n t i f y and p r e d i c t high r i s k i n d i v i d u a l s on the b a s i s of a c h o l e s t e r o l index. The others thought p r e d i c t i o n was not p o s s i b l e on the b a s i s of c h o l e s t e r o l or l i p i d measures. The problem, as they saw i t , was that too many f a l s e p o s i t i v e s were produced f o r t h i s to be a u s e f u l p r e d i c t i v e t o o l . While there are many methodological q u e s t i o n s that c o u l d be r a i s e d about t h i s study, i t seems most u s e f u l to conclude that i t was able to show some in c r e a s e d r i s k a s s o c i a t e d with i n c r e a s e d l e v e l s of c h o l e s t e r o l , but s i n c e those above the s e v e n t y - f i f t h p e r c e n t i l e f a r e d no worse than those above the f i f t i e t h , the r e l a t i o n s h i p i s probably not l i n e a r . One study alone cannot 34 prove the u s e f u l n e s s of a p r e d i c t i v e measure, and concern over t h i s i s s u e was perhaps premature. Brunner and Lobl (1958) s t u d i e d , r e t r o s p e c t i v e l y , the c h o l e s t e r o l l e v e l s of 74 male p a t i e n t s who had experienced a myocardial i n f a r c t i o n . Blood samples of the p a t i e n t s were taken at l e a s t 20 days a f t e r the coronary event, although i n some cases the time span was much longe r . They compared these readings with those obtained from f i v e other groups of I s r a e l i men. (Three of the groups were from e t h n i c p o p u l a t i o n s known to have d i f f e r e n t types of l i p i d p a t t e r n s , one of which was a group of Yemenites, who were markedly d i f f e r e n t i n l i f e s t y l e , d i e t a r y p r a c t i c e s , and much lower i n r a t e s of coronary heart d i s e a s e . The remaining two groups c o n s i s t e d of a group of s u r g i c a l p a t i e n t s , and a group of p h y s i c i a n s at the h o s p i t a l . ) Of i n t e r e s t here i s t h e i r a n a l y s i s of the d i s t r i b u t i o n of the l i p o p r o t e i n measures f o r the s i x groups. Based on the r e s u l t s of pr e v i o u s research, they f e l t that 250 mg/100ml would be the upper bound of a safe c h o l e s t e r o l l e v e l , and that the p o p u l a t i o n average would be 245mg/l00ml. However, they found that only 59.3% (44 cases) of the myocardial i n f a r c t i o n group was above 250mg/l00ml. In f a c t , some of the comparison group averages d i f f e r e d more from the expected average than d i d the coronary p a t i e n t s . T h e i r r e s u l t s d i d i n d i c a t e that the l e v e l of beta c h o l e s t e r o l might be important. Only 8 of the coronary p a t i e n t s (10.8%) had a beta c h o l e s t e r o l l e v e l below 35 180mg/l00ml, while none of the Yemenites had a value higher than 180. S i m i l a r evidence of the importance of l i p o p r o t e i n s compared to simple c h o l e s t e r o l readings was p r o v i d e d by Lyon, Yankley, Gofman and Strisower (1956). Part of t h e i r data was d e r i v e d from a follow-up of 470 p a t i e n t s who had s u f f e r e d some form of coronary heart d i s e a s e . The authors had developed an anthrogenic index (based on a weighting of the v a r i o u s l i p o p r o t e i n s l e v e l s of the i n d i v i d u a l ) , and compared i t s p r e d i c t i v e behavior to that of c h o l e s t e r o l measurement alone. They found that the c h o l e s t e r o l p r e d i c t i o n and that based on t h e i r index d i d not always agree. Only 51.4% of the myocardial i n f a r c t i o n p a t i e n t s (70 of the 470 p a t i e n t s were d e s c r i b e d as d e f i n i t e MI) had 'a c h o l e s t e r o l reading higher than that of a matched c o n t r o l mean, while 75.7% of the p a t i e n t s had an anthrogenic measure higher than the c o n t r o l group's mean. While these r e s u l t s do seem to i n d i c a t e that a simple c h o l e s t e r o l measure would not do as w e l l as a measure based on l i p o p r o t e i n s , t h i s c o n c l u s i o n must be tempered with the o b s e r v a t i o n that the p r e d i c t i v e a b i l i t y of c h o l e s t e r o l i n t h i s p a r t i c u l a r study (51.4% i s at the chance l e v e l ) i s much lower than that normally found with c h o l e s t e r o l . It i s d i f f i c u l t to i n t e r p r e t the r e s u l t s of these kind of r e t r o s p e c t i v e s t u d i e s . It i s never c l e a r how much the coronary event i t s e l f might i n f l u e n c e measurements taken 36 a f t e r the event. S t u d i e s of t h i s s o r t do not seem to have i n f l u e n c e d the design of the p r o s p e c t i v e s t u d i e s d i s c u s s e d here. However, they are one of the e a r l y i n d i c a t o r s that the c h o l e s t e r o l - coronary heart d i s e a s e r e l a t i o n s h i p might be complex. The Albany study (see Doyle et a l . , 1959) was able to demonstrate a r e l a t i o n s h i p between serum c h o l e s t e r o l and coronary heart d i s e a s e . At the i n i t i a l s c r e e n i n g the average c h o l e s t e r o l l e v e l of t h e i r sample was 225 mg/100ml. Table 3.2 presents t h e i r f i n d i n g s r e l a t i n g l e v e l of serum c h o l e s t e r o l at screen and the m o r b i d i t y experience a f t e r 44 months of follow-up. Table 3.2 M o r b i d i t y rate by serum c h o l e s t e r o l c a t e g o r i z a t i o n , Albany study a f t e r 44 months of f o l l o w -up. (Adapted from Table 8, Doyle et a l . , 1959 Serum C h o l e s t e r o l (mg/100ml) Number at Risk Number of CHD cases Annual rate / I 000 men < 200 457 1 0 6.0 200-274 995 23 6.3 > 274 209 1 6 20.9 mis s i n g 182 8 12.0 The authors reported that over the 44 months they observed l i t t l e change in the i n d i v i d u a l ' s serum c h o l e s t e r o l r e a dings. For 115 of the s u b j e c t s extensive blood l i p i d measurements were taken. No " s t r i k i n g r e l a t i o n s h i p " between the l i p i d measurements and coronary heart d i s e a s e was found. 37 A s i m i l a r set of r e s u l t s was found by Paul et a l . (1963). However, they found that while coronary heart d i s e a s e r a t e s r i s e with c h o l e s t e r o l l e v e l s , there might be an upper l i m i t to t h i s e f f e c t (see Table 3.3). The mean c h o l e s t e r o l value f o r the sample at the f i r s t examination was 248mg/l00ml. The authors a l s o d i v i d e d the coronary heart d i s e a s e cases i n t o c a t e g o r i e s of i n c r e a s i n g s e v e r i t y , and found that the mean c h o l e s t e r o l value i n each category i n c r e a s e d with the s e v e r i t y of the problem. Table 3.3 M o r b i d i t y r a t e by serum c h o l e s t e r o l c a t e g o r i z a t i o n , Chicago study a f t e r 4.42 years of follow-up. (Adapted from Table 5, Paul et a l . , 1963). Serum C h o l e s t e r o l (mg/100ml) Number at Risk Number of CHD cases Annual rate /1000 men 100-204 365 6 3.7 205-239 539 19 7.8 240-256 282 1 3 10.4 257-274 212 1 5 16.0 275-309 271 19 15.9 > 309 231 1 6 15.7 Note: 181 c h o l e s t e r o l values were missing S l i g h t l y d i f f e r e n t r e s u l t s were found in the Los Angeles study. Chapman and Massey (1964) found that the mor b i d i t y r a t e f o r coronary heart d i s e a s e rose with serum c h o l e s t e r o l , but they d i d not f i n d any l e v e l i n g o f f at high 38 c h o l e s t e r o l r a t e s (see Table 3.4). Table 3.4 M o r b i d i t y r a t e by serum c h o l e s t e r o l c a t e g o r i z a t i o n , Los Angeles study a f t e r 10 years of follow-up. (Adapted from Table 3, Chapman and Massey, 1964) . Serum C h o l e s t e r o l (mg/100ml) Number at Risk Number of CHD cases Annual rate /I 000 men < 210 1 39 6 4.3 210-269 418 34 8.1 270-389 840 1 1 7 13.4 > 389 1 02 20 19.6 They a l s o attempted to f i n d a r e l a t i o n s h i p between the s e v e r i t y of the coronary d i s e a s e and the c h o l e s t e r o l l e v e l . No d i f f e r e n c e was found i n the mean c h o l e s t e r o l l e v e l s of those i n d i v i d u a l s who developed angina (the l e a s t severe form of coronary heart d i s e a s e ) , but the l e v e l s were higher for myocardial i n f a r c t i o n , and higher s t i l l f o r sudden death. The authors concluded from t h e i r r e s u l t s that the a s s o c i a t i o n between serum c h o l e s t e r o l and myocardial i n f a r c t i o n i s the reason f o r the r e l a t i o n s h i p between c h o l e s t e r o l and coronary heart d i s e a s e . Both Rosenman et a l . (1970) and Keys et a l . (1963) found i n c r e a s i n g r i s k with i n c r e a s i n g c h o l e s t e r o l . Table 3.5 and Table 3.6 present t h e i r f i n d i n g s . The r e s u l t s of the Framingham study f o l l o w the same general t r e n d . In 1957 they found that higher c h o l e s t e r o l 39 Table 3.5 M o r b i d i t y r a t e by serum c h o l e s t e r o l c a t e g o r i z a t i o n , Western C o l l a b o r a t i v e study a f t e r 4 1/2 years of follow-up. (Adapted from Table 2, Rosenman et al.,1970). Serum C h o l e s t e r o l (mg/100ml) Number at Risk Number of CHD cases Annual rate /I 000 men < 220 1 468 29 4.4 220-259 1 063 53 11.1 > 259 651 51 17.4 Table 3.6 M o r b i d i t y rate by serum c h o l e s t e r o l c a t e g o r i z a t i o n , Minnesota study a f t e r 15 years of follow-up. (Adapted from Table 5, Keys et a l . , 1963). Serum C h o l e s t e r o l (mg/100ml) Number at Risk Number of CHD cases Annual rate /1000 men < 200 85 4 3. 1 200-2 1 9 55 3 3.6 220-239 59 7 7.9 240-259 43 8 12.4 > 259 39 1 0 17.1 l e d to higher m o r b i d i t y r a t e s , and the same general f i n d i n g s were a l s o found i n 1974. Table 3.7 presents these r e s u l t s . I t i s c l e a r that there i s i n c r e a s i n g r i s k with i n c r e a s i n g l e v e l s of c h o l e s t e r o l . However, the r a t e s observed in t h i s study are higher than those observed in the other s t u d i e s . As can be seen from F i g u r e 3.1, the high r a t e s from the Framingham study were present at both 4 years and 18 ye a r s . O v e r a l l , these r e s u l t s i n d i c a t e there i s some 40 Table 3.7 M o r b i d i t y r a t e by serum c h o l e s t e r o l c a t e g o r i z a t i o n , Framingham study. See Dawber et a l . (1957) f o r four year follow-up, and S h u r t l e f f (1974) f o r 18 year r e s u l t s . A. Framingham: 4 years of follow-up Serum C h o l e s t e r o l . (mg/100ml) Number at Risk Number of CHD cases Annual r a t e /1000 men < 225 445 18 10.1 225-259 265 21 11.3 > 259 172 21 30.5 B. Framingham: 18 years of follow-up Serum C h o l e s t e r o l (mg/100ml) Person years of Risk Number of CHD cases Annual r a t e /I 000 men 096-204 5842 56 9.6 205-234 7516 1 1 7 15.6 235-264 6322 85 13.4 265-294 3522 58 16.5 295-1124 1862 52 27.9 missing 966 1 5 15.5 r e l a t i o n s h i p between the chance of developing coronary heart d i s e a s e and serum c h o l e s t e r o l l e v e l s . T h i s has been demonstrated without s t a n d a r d i z i n g the v a r i o u s s t u d i e s in terms of age. Since the ages examined i n the s t u d i e s v a r i e d , l i t t l e importance i s p l a c e d on the v a r i a b i l i t y of the r e s u l t s shown in F i g u r e 3.1. In a l l cases the general trend i s the same. However, the r e l a t i o n s h i p i s not strong enough to be an i n d i v i d u a l p r e d i c t i v e t o o l : not even most of those with high c h o l e s t e r o l l e v e l s develop coronary heart d i s e a s e , while some of those with low c h o l e s t e r o l do develop 41 coronary heart d i s e a s e . Any r e l a t i o n s h i p between c h o l e s t e r o l and coronary heart d i s e a s e i s o b v i o u s l y complex. I l l Blood Pressure F i n d i n g s E l e v a t e d blood pressure has been a s s o c i a t e d with the development of coronary heart d i s e a s e s i n c e the 1940's. I t has been estimated that h y p e r t e n s i o n (dangerously high blood pressure) a f f e c t s between 10 and 30 percent of the p o p u l a t i o n of the United S t a t e s (Weinstein and Stason, 1976). However, the d i v i d i n g l i n e between h y p e r t e n s i v e and normal i s not simple, and e s t i m a t e s of the prevalence of hypertension can vary a great deal with s l i g h t l y d i f f e r e n t d e f i n i t i o n s of the group boundaries. The World Health O r g a n i z a t i o n has attempted to develop a d e f i n i t i o n of h y p e r t e n s i o n . In 1 9 5 9 they proposed the f o l l o w i n g d e f i n i t i o n (see Dawber, 1980): Normotension -- both p r e s s u r e s recorded by the p h y s i c i a n are l e s s than 140/90 mm Hg; Hypertension -- both p r e s s u r e s recorded by the p h y s i c i a n are 160/95 and above; B o r d e r l i n e a l l p r e s s u r e s recorded are above 139/89 but below 160/95. (p. 80) However, c e r t a i n i n d i v i d u a l s d i d not f i t i n t o these c a t e g o r i e s , and in 1962 the World H e a l t h O r g a n i z a t i o n amended the b o r d e r l i n e c l a s s i f i c a t i o n to i n c l u d e i n d i v i d u a l s whose s y s t o l i c blood pressure was between 140 and 159 mm Hg and d i a s t o l i c pressure was below 95, or those whose d i a s t o l i c pressure was between 90 and 94 and s y s t o l i c was Figure 3.1 Coronary heart disease morbidity r a t e s as a fu n c t i o n of serum c h o l e s t e r o l l e v e l s . 42 30-M o r b 25-• i d • i t 20-y R a t 15-e • 1 10-0 0 0 5" T--T fe i r , T • i r o - o y o I o - ^ - O — o — o — O - J I o %\ _^ » — • — » —• Framingham (4 yrs) ----- Framingham (18 yrs) - . . . . Minnesota -x-X-Western Collaborative -o-o-Los Angeles -A-A-Chicago Albany 100 S e r um mg 200 300 C h o i e s t e / 1 00 m l . 400 ro I 43 l e s s than 160. Hypertensive i n d i v i d u a l s need only have a s y s t o l i c pressure above 159 or a d i a s t o l i c above 94 (Chapman and Massey, 1964). An understanding of these d e f i n i t i o n a l changes i s important because r e s e a r c h e r s have the unfortunate h a b i t of c l a s s i f y i n g t h e i r s u b j e c t s as hy p e r t e n s i v e , b o r d e r l i n e , or normal, and only c l o s e examination of the text w i l l i n d i c a t e e x a c t l y which d e f i n i t i o n i s being used. An i n d i v i d u a l ' s blood pressure i t s e l f i s not a s t a b l e q u a n t i t y . Blood pressure i n c r e a s e s with age, shows d a i l y and seasonal v a r i a t i o n , and decreases with m u l t i p l e measures d u r i n g one c l i n i c a l measurement (Birkenhager and Schalekamp, 1976). The decrease during a c l i n i c a l s e s s i o n i s thought to r e f l e c t s u b j e c t s ' r e d u c t i o n i n s t r e s s as they adapt to the novel s i t u a t i o n . D a i l y v a r i a t i o n in pressure can be q u i t e s u b s t a n t i a l , f a l l i n g to a low of 90/50 mm Hg du r i n g s l e e p f o r normotensives and l a b i l e h y p e r t e n s i v e s . For other h y p e r t e n s i v e s the v a r i a t i o n i s much l e s s . F i n a l l y , d u r i n g the summer blood pressure i s lower than at other times of the year (Paul et a l . , 1963). High blood pressure i t s e l f i s of t e n not co n s i d e r e d a d i s e a s e . Rather, i t i s seen as a c o n t r i b u t i n g f a c t o r to many c a r d i o v a s c u l a r d i s e a s e s . Except f o r a m i n o r i t y of c l i n i c a l cases the e t i o l o g y of high blood pressure i s unknown, although there i s b e l i e v e d to be both a genetic and environmental component to the problem. E f f e c t i v e 44 treatments f o r hypertension e x i s t , although the drug treatments q u i t e o f t e n have severe s i d e e f f e c t s . Most p r o s p e c t i v e s t u d i e s have analyzed both the s y s t o l i c and d i a s t o l i c readings f o r t h e i r r e l a t i o n s h i p to coronary heart d i s e a s e . There i s no simple method f o r combining the two measures i n t o one o v e r a l l q u a n t i t a t i v e s c o r e . Thus the a n a l y s i s must e i t h e r be made s e p a r a t e l y f o r the two readings, or the s u b j e c t s must be c l a s s i f i e d i n t o groups on the b a s i s of both s c o r e s . The c o r r e l a t i o n between s y s t o l i c and d i a s t o l i c measures f o r the i n d i v i d u a l i s s u r p r i s i n g l y low. In a study that examined repeated measures of s y s t o l i c and d i a s t o l i c blood pressure readings taken over a short time p e r i o d , the average c o r r e l a t i o n per i n d i v i d u a l was only .57 (Hart, 1980). T h i s seems to imply that d i f f e r e n t c o n c l u s i o n s c o u l d be reached depending on the measure used. It should a l s o be noted that t h i s lack of c o r r e l a t i o n presents c e r t a i n problems i n i n t e r p r e t a t i o n f o r the a n a l y s i s of the r e l a t i o n s h i p between blood pressure and coronary heart d i s e a s e . C a l c u l a t e d s i g n i f i c a n c e l e v e l s w i l l have l i t t l e meaning i f the i n v e s t i g a t o r examines the r e l a t i o n s h i p between each blood pressure measure and coronary heart d i s e a s e , s e l e c t s the blood pressure measure ( s y s t o l i c or d i a s t o l i c ) that shows a s i g n i f i c a n t r e l a t i o n s h i p to coronary heart d i s e a s e , and then concludes that blood pressure i s r e l a t e d to the development of coronary heart d i s e a s e . The 45 c o n c l u s i o n should be r e s t r i c t e d to the s p e c i f i c type of blood pressure measure taken. Doyle et a l . (1959) report only r e s u l t s f o r d i a s t o l i c blood pressure ( s y s t o l i c blood pressure i s not d i s c u s s e d in t h i s or t h e i r p r e v i o u s report (Doyle et a l . , 1 9 5 7 ) ) . They found i n c r e a s i n g r i s k with i n c r e a s i n g l e v e l s of blood pressure (Table 3.8). Note that i n t h i s c l a s s i f i c a t i o n the range 90-109 c o n t a i n s both i n d i v i d u a l s that might be co n s i d e r e d e i t h e r b o r d e r l i n e or d e f i n i t e h y p e r t e n s i v e s . Table 3.8 M o r b i d i t y r a t e by d i a s t o l i c blood pressure c a t e g o r i z a t i o n , Albany study a f t e r 44 months of f o l l o w -up. (Adapted from Table 6, Doyle et a l . , 1959). Blood Pressure (mm Hg) Number at Risk Number of CHD cases Annual rate /1000 men < 90 1 1 55 33 7.8 90-109 565 18 8.9 > 109 1 1 3 6 14.5 missing 1 0 -The Los Angeles study (Chapman and Massey, 1964) used the r e v i s e d WHO standards of 1962 to c l a s s i f y t h e i r s u b j e c t s . Table 3.9 summarizes t h e i r r e s u l t s . Here too i s some evidence of i n c r e a s i n g r i s k with higher l e v e l s of blood p r e s s u r e . Paul et a l . (1963) reported s i m i l a r f i n d i n g s , s t a t i n g that the mean blood pressure of those who had not developed coronary heart d i s e a s e was 135/87 mm Hg., while that of those who s u f f e r e d some form of coronary heart 46 d i s e a s e had a mean reading of 146/92. Both the s y s t o l i c and d i a s t o l i c d i f f e r e n c e s are s t a t i s t i c a l l y s i g n i f i c a n t . Table 3.9 M o r b i d i t y r a t e by blood pressure c a t e g o r i z a t i o n , Los Angeles study a f t e r 10 years of follow-up. (Adapted from Table 5, Chapman and Massey, 1964). Blood Pressure Number at Risk Number of CHD cases Annual rate /I 000 men normotensive 1 029 95 9.2 border1ine 325 46 14.2 h y p e r t e n s i v e 1 45 36 24.8 The r e s u l t s of Keys et a l . (1963) are presented i n terms of the number of cases of coronary heart d i s e a s e i n each q u a r t i l e of the d i s t r i b u t i o n of blood pressure r e a d i n g s . T h i s makes comparison of the r e s u l t s with other s t u d i e s d i f f i c u l t . They concluded that there was l e s s r i s k f o r those in the bottom q u a r t i l e of the d i s t r i b u t i o n , but the f i n d i n g that the group in the 25-50 p e r c e n t i l e range had a higher coronary heart d i s e a s e r a t e than those i n the 50-75 range makes i t d i f f i c u l t to conclude that there i s any o v e r a l l t r e n d . The r e s u l t s from Rosenman et a l . (1970) are presented in Table 3.10. The Framingham r e s u l t s are produced i n Table 3.11. I n i t i a l l y the Framingham study d i d not c o l l e c t blood pressure readings from a l l p a r t i c i p a n t s , so the r e s u l t s of Dawber et a l . (1957) are not presented. From Table 3.11A i t 47 Table 3.10 M o r b i d i t y r a t e by blood pressure c a t e g o r i z a t i o n , Western C o l l a b o r a t i v e study a f t e r 4 1/2 years of follow-up. (Adapted from Table 2, Rosenman et a l . , 1970). A. S y s t o l i c Pressure Blood Pressure (mm Hg) Number at Risk Number of CHD cases Annual r a t e / I 000 men < 120 774 19 5.5 120-159 2278 93 9.1 > 159 1 30 21 35.9 B. D i a s t o l i c Pressure Blood Pressure (mm Hg) Number at Risk Number of CHD cases Annual r a t e /1000 men < 95 2883 1 02 7.9 95-99 1 1 4 6 11.7 > 99 185 25 30.0 can be seen that higher l e v e l s of s y s t o l i c blood pressure are a s s o c i a t e d with higher r a t e s of coronary heart d i s e a s e . The same i s g e n e r a l l y true f o r d i a s t o l i c pressure (Table 3.11B), but the trend i s not n e a r l y as c o n s i s t e n t . F i g u r e s 3.2A and 3.2B attempt to p l o t these r e s u l t s in the same manner as Fig u r e 3.1. Note that the same data have been p l o t t e d i n Fi g u r e s 3.2A and 3.2B f o r those s t u d i e s that only c l a s s i f i e d t h e i r s u b j e c t s by hyp e r t e n s i v e c a t e g o r i e s . Thus i f h y p e r t e n s i v e s were c l a s s i f i e d as s u b j e c t s with blood pressures over 140/90 mm Hg, then the death r a t e a s s o c i a t e d with these s u b j e c t s would be p l o t t e d as s t a r t i n g at 140 in Fi g u r e 3.2A and at 90 in F i g u r e 3.2B. While the trends are the same i n both graphs, o v e r a l l i t would appear that the 48 a s s o c i a t i o n of coronary heart d i s e a s e with s y s t o l i c blood pressure i s more pronounced than that of d i a s t o l i c blood pressure and coronary heart d i s e a s e . Table 3.11 M o r b i d i t y rate by blood pressure c a t e g o r i z a t i o n , Framingham study a f t e r 18 years follow-up. (Adapted from Tables 1-1 and 1-2, S h u r t l e f f , 1974). A. S y s t o l i c blood pressure Blood Pressure (mm Hg) Person years of r i s k Number of CHD cases Annual r a t e /1000 men 074-109 1512 9 6.0 110-119 3700 29 7.8 120-129 5120 55 10.7 130-139 5292 62 11.7 140-149 3902 71 18.2 150-159 2592 50 19.3 160-169 1 562 35 22.4 170-179 950 29 30.5 180-189 604 1 5 24.8 190-300 796 28 35.2 cont. 49 B. D i a s t o l i c blood pressure Blood Pressure (mm Hg) Person years of r i s k Number of CHD cases Annual r a t e /I 000 men 020-069 1 920 20 10.4 070-074 3512 47 13.4 075-079 2666 29 10.9 080-084 61 68 66 10.7 085-089 31 60 30 9.5 090-094 3632 76 20.9 095-099 1638 32 19.5 100-104 1760 42 23.9 105-109 488 1 0 20.5 110-160 1 086 31 28.5 IV C i g a r e t t e Smoking F i n d i n g s C i g a r e t t e smoking was the l a s t of the major r i s k f a c t o r s f o r coronary heart d i s e a s e to be r e c o g n i z e d . Even by 1960 t h i s issue had not begun to be r e s o l v e d , with some s t u d i e s r e p o r t i n g no r e l a t i o n s h i p between tobacco consumption and coronary heart d i s e a s e , while others found one f o r c i g a r e t t e s but not f o r pipe and c i g a r smokers (Dawber, 1960). Yet the e f f e c t s of n i c o t i n e on the c a r d i o v a s c u l a r system were w e l l known, and some re s e a r c h e r s continued to t h e o r i z e that there should be a r e l a t i o n s h i p of F i g u r e 3.2A Coronary heart d i s e a s e m o r b i d i t y r a t e s by l e v e l of s y s t o l i c blood p r e s s u r e . M o r b i d i t y R a t e 1 0 0 0 35-30 25H 20H 10-5-x I I x i | * X r k • 0 i - o i - - i - o - O -f I I 4» -4 r l . JL' Framingham (18 yrs) - X - / - Western Collaborative -o - o- Los Angeles 100 150 200 S y s t o I i c mm. P r e s s u r e Hg. 51 F i g u r e 3.2B Coronary heart d i s e a s e m o r b i d i t y r a t e s by l e v e l of d i a s t o l i c blood p r e s s u r e . M o r b i d i t y R a t e 1 0 0 0 35 30 25^ 20 15-ID-S' ¥ X I X I X I I. > O— Of-O-O -> I I I * r O - O-O-O- 8"-"Zr 4-JL=_Xr-JL= Ki- l-j "T ll I r ? f -> J Framingham (18 yrs) - X - X - Western Collaborative - o - e> - Los Angeles - Albany -th 60 80 100 120 D i a s t o l i c P r e s s u r e mm. H g . 52 some s o r t (Davis, 1960). E v a l u a t i o n of t h i s p o s s i b i l i t y was hampered by the f a c t that not a l l s t u d i e s i n c l u d e d an examination of the e f f e c t s of smoking i n t h e i r experimental p r o t o c o l . Even i f i t was i n c l u d e d , the data had to r e l y on s e l f r e p o r t s , and the r e l i a b i l i t y of the measures was not checked. Thus these data are probably e r r o r prone, and i t i s not unusual to see very coarse smoking c a t e g o r i e s r e p o r t e d . Where f i n e r d i v i s i o n s are used the q u e s t i o n a b l e r e l i a b i l i t y of s e l f r e p o r t s probably makes r e s u l t s more suspect than the previous r i s k f a c t o r measurements. Paul et a l . (1963) r e p o r t e d an e f f e c t f o r c i g a r e t t e smokers but not f o r c i g a r or pipe smokers. Table 3.12 summarizes t h e i r r e s u l t s . The "amount smoked" rep r e s e n t s s e l f r e p o r t s of the d a i l y amount smoked f o r most of the s u b j e c t ' s a d u l t l i f e . They a l s o report that those who smoked more than 1/2 a pack of day developed more severe forms of coronary heart d i s e a s e than those who d i d not smoke or smoked l e s s than 1/2 pack a day. Doyle et a l . (1959) found a s i m i l a r set of r e s u l t s (see Table 3.13). Note that the r a t e s for c i g a r e t t e smokers are not s t a t i s t i c a l l y s i g n i f i c a n t from each other, and the authors concluded that i t was too e a r l y to determine i f a r e l a t i o n s h i p e x i s t e d . The f i n d i n g s of Rosenman et a l . (1970) are presented i n Table 3.14. They show a r i s i n g t r end with i n c r e a s e d consumpt i o n . 53 Table 3.12 M o r b i d i t y r a t e by smoking c a t e g o r i z a t i o n , Chicago study a f t e r 4.4 years of follow-up. (Adapted from Table 8, Paul et a l . , 1963). Amount Smoked d a i l y Number at Risk Number of CHD cases Annual r a t e /1000 men 0 609 20 7.4 1-7 127 2 3.6 8-12 204 8 8.9 13-17 220 5 5.1 1 8-22 577 41 16.1 23-27 38 3 17.4 < 27 1 1 5 8 15.7 Table 3.13 M o r b i d i t y r a t e by smoking c a t e g o r i z a t i o n , Albany study a f t e r 44 months of follow-up. (Adapted from Table 7, Doyle et a l . , 1959). Amount Smoked da i l y Number at Risk Number of CHD cases Annual rate /1000 men 0 327 7 5.8 c i g a r s / p i p e s 256 4 4.3 1-14 1 72 7 11.1 1 5-34 717 22 8.4 < 34 1 58 8 12.1 q u i t 202 9 12.1 The f i n d i n g s of the Minnesota study are repo r t e d in Keys, T a y l o r , Blackburn, Brozek, Anderson and Simonson (1971). Table 3.15 summarizes t h e i r r e s u l t s (note that t h i s report i s based on 20 years of f o l l o w - u p ) . 54 Table 3.14 M o r b i d i t y r a t e by smoking c a t e g o r i z a t i o n , Western C o l l a b o r a t i v e study a f t e r 4.5 years of follow-up. (Adapted from Table 1, Rosenman et a l . , 1970). Amount . Smoked d a i l y Number at Risk Number of CHD cases Annual r a t e /1000 men 0 1673 49 6.5 1-15 321 1 1 7.6 1 6-25 602 34 12.6 > 25 586 39 14.8 Table 3.15 M o r b i d i t y rate by smoking c a t e g o r i z a t i o n , Minnesota study a f t e r 20 years follow-up. (Adapted from Table 2, Keys et a l . , 1971). Amount Smoked d a i l y Number at Risk Number of CHD cases Annual rate /1000 men 0 82 10 6. 1 1-9 62 5 4.0 10-19 41 8 9.8 > 19 53 1 5 14.2 q u i t 39 1 4 17.9 From the e a r l y Framingham r e s u l t s Dawber et a l . (1957) concluded that i t was too e a r l y to determine i f an e f f e c t e x i s t e d . The l a t e r r e s u l t s (shown in Table 3.16) do i n d i c a t e that a r e l a t i o n s h i p e x i s t s , but i t i s not as strong as f o r the other r i s k f a c t o r s . In summary, F i g u r e 3.3 g i v e s an i n d i c a t i o n of the v a r i a b i l i t y of the r e s u l t s presented i n t h i s s e c t i o n . The e a r l y r e s u l t s f o r smoking were not n e a r l y as strong as f o r those of c h o l e s t e r o l and blood p r e s s u r e . 55 Table 3.16 M o r b i d i t y r a t e by smoking c a t e g o r i z a t i o n , Framingham study a f t e r 18 years of follow-up. (Adapted from Table 1-9, S h u r t l e f f , 1974). Amount Smoked d a i l y Person years of Risk Number of CHD cases Annual rate /I 000 men 0 11,552 1 45 12.6 1-19 3938 54 13.7 20 5324 94 17.7 > 20 5096 87 17.1 It seems l i k e l y that smoking had not been a s s o c i a t e d with the development of other d i s e a s e s , and that i t was known to have an adverse e f f e c t on the c a r d i o v a s c u l a r system, the r e l a t i v e l y small r e l a t i o n s h i p between smoking and coronary heart d i s e a s e would have r e s u l t e d i n l i t t l e a d d i t i o n a l a t t e n t i o n being p a i d to t h i s v a r i a b l e . Note that the strong a s s o c i a t i o n which smoking has with other d i s e a s e s would tend to mask i t s e f f e c t on the development of coronary heart d i s e a s e , s i n c e i t i s p o s s i b l e that before the negative e f f e c t s of smoking on the c a r d i o v a s c u l a r system become apparent some heavy smokers w i l l have developed some other d i s e a s e . I t i s d i f f i c u l t to i n t e r p r e t the g e n e r a l l y equal or higher m o r b i d i t y r a t e of nonsmokers to those i n the lowest smoking category. T h i s issue was not examined in any of the s t u d i e s , except to note that only a small follow-up time had been observed. However, the general c o n s i s t e n c y of t h i s f i n d i n g a c r o s s s t u d i e s seems to imply that the e f f e c t i s l e g i t i m a t e . T h i s may mean that there i s some s o r t of 56 t h r e s h o l d e f f e c t f o r the e f f e c t s of smoking on coronary heart d i s e a s e . V A d d i t i o n a l Risk F a c t o r s A wide v a r i e t y of other r i s k f a c t o r s were examined i n these and other s t u d i e s . In an e x t e n s i v e review of 35 r i s k f a c t o r s that were suspected to be a s s o c i a t e d with coronary heart d i s e a s e Simborg (1970) concluded that there were 5 primary r i s k f a c t o r s ( h y p e r c h o l e s t e r o l e m i a , hypertension, smoking, sedentary l i v i n g , and o b e s i t y ) and 6 secondary r i s k f a c t o r s (EKG a b n o r m a l i t i e s , p e r s o n a l i t y f a c t o r s , EKG changes in response to e x e r c i s e , a f a m i l y h i s t o r y of coronary heart d i s e a s e , d i a b e t e s m e l l i t u s , and l i p o p r o t e i n e l e c t r o p h o r e s i s a b n o r m a l i t y ) . The assignment of a v a r i a b l e to a primary and secondary category was done more on the b a s i s of the s t r e n g t h of a s s o c i a t i o n of the v a r i a b l e with coronary heart disease and the amount of i n v e s t i g a t i o n that p a r t i c u l a r r i s k f a c t o r had r e c e i v e d . I t was not meant to d e s c r i b e the c a u s a l r e l a t i o n s h i p s between the v a r i a b l e s . Three of Simborg's primary r i s k f a c t o r s have been d e s c r i b e d above. The evidence presented by Simborg for the other two (sedentary l i v i n g and o b e s i t y ) has come from r e t r o s p e c t i v e rather than p r o s p e c t i v e s t u d i e s . T h i s i t s e l f i s not a reason to e l i m i n a t e these v a r i a b l e s as r i s k f a c t o r s . However, the f a i l u r e to f i n d an e f f e c t due to these v a r i a b l e s i n the m a j o r i t y of the p r o s p e c t i v e s t u d i e s 57 Figure 3.3 Coronary heart disease morbididty rates by the number of cigarettes smoked per day. t e 1 0 0 0 25A M o r b i d 2(H I t y S-r-A-A:T -> I A A 151 fit • - A - A - A - > f - x - x - x - x - x - » 10 r I I * f -XAx«-X -X—* ! A . A I I . I 1 ir v j ? • x-tx-*-*-x-* - • 5J I-A I I A . t ; - A - . i A I Framingham (18 yrs) - • - • -Minnesota - x - x - Western Collaborative -A - A - Chicago Albany ' 10 ' 20 ' 3'0 ' Cigarettes per Day 58 i s d i s a p p o i n t i n g , as these two v a r i a b l e s would be the e a s i e s t around which to p l a n an i n t e r v e n t i o n . I t may be that measurement problems are the primary reason f o r the f a i l u r e to f i n d a l a r g e e f f e c t f o r these v a r i a b l e s i n p r o s p e c t i v e s t u d i e s . For example, sedentary l i v i n g i s u s u a l l y d e f i n e d as the type of work the i n d i v i d u a l does, and the c a t e g o r i e s can be so broad that they do not take i n t o account v a r i a t i o n s in the amount of labor w i t h i n the job c l a s s i f i c a t i o n (Simborg, 1970). In a l a t e r review, G e r t l e r and White (1976) l i s t e d 44 v a r i a b l e s that were examined in one or more of the s t u d i e s d e s c r i b e d above. Some of the v a r i a b l e s appear to be e l i m i n a t e d as p o s s i b l e r i s k f a c t o r s , f o r example f i v e of the s i x s t u d i e s examined the e f f e c t s of x-ray exposure, and none found a r e l a t i o n s h i p with l a t e r development of coronary heart d i s e a s e . Other cases are not so c l e a r . For example, in regard to f a m i l y h i s t o r y of coronary heart d i s e a s e and l a t e r development of coronary d i s e a s e by the s u b j e c t , three s t u d i e s found a s i g n i f i c a n t r e l a t i o n s h i p , while two d i d not. F i n a l l y , in some cases only one study examined the e f f e c t s of a v a r i a b l e . It i s unclear how t h i s type of f i n d i n g should be i n t e r p r e t e d , s i n c e i t i s not known i f t h i s v a r i a b l e was c o r r e l a t e d to other more g e n e r a l l y accepted r i s k f a c t o r s examined in the study. For example, c o f f e e intake was found to be r e l a t e d to the development of coronary heart d i s e a s e in the Chicago study and i t i s not 59 c l e a r i f c o f f e e should be seen as a p o s s i b l e r i s k f a c t o r or i f i t i s more l i k e l y a s s o c i a t e d with coronary heart d i s e a s e because of a c o r r e l a t i o n with smoking (or some other v a r i a b l e ) . I t i s at t h i s p o i n t that the u s e f u l n e s s of the r e s u l t s of these s t u d i e s begins to break down. The number of v a r i a b l e s , the unknown nature of t h e i r i n t e r r e l a t i o n s h i p s , and the l a r g e number of c r o s s - c l a s s i f i c a t i o n s r e q u i r e d to examine every p o s s i b l e hypothesis about the nature of t h e i r r e l a t i o n s h i p with coronary heart d i s e a s e seemed to suggest that c l a r i f i c a t i o n of the is s u e c o u l d not be achieved. I t was at t h i s p o i n t that i n t e r e s t s h i f t e d to the m u l t i v a r i a t e a n a l y s i s of these r i s k f a c t o r s . 60 CHAPTER 4  M u l t i p l e L o g i s t i c A n a l y s i s I M u l t i v a r i a t e A n a l y s i s When data are c o l l e c t e d on s e v e r a l dimensions of an i n d i v i d u a l (e.g. blood pres s u r e , serum c h o l e s t e r o l l e v e l , e t c . ) the researcher i s faced with the ch o i c e of a n a l y z i n g the dimensions s e p a r a t e l y (a s e r i e s of u n i v a r i a t e a n a l y s e s ) , a l l together (an o v e r a l l m u l t i v a r i a t e a n a l y s i s ) , or i n some combination of both methods. The c h o i c e i s important, since a s e r i e s of u n i v a r i a t e analyses would o f t e n not r e s u l t i n the same i n t e r p r e t a t i o n of the data as would be made from a m u l t i v a r i a t e a n a l y s i s . How t h i s can happen can be e a s i l y i l l u s t r a t e d . Consider a v a r i a b l e 'A' that causes, by i t s e l f , e f f e c t 'B'. Now assume that 'A' and a d i f f e r e n t v a r i a b l e 'C are h i g h l y c o r r e l a t e d (say r=.90). A strong r e l a t i o n s h i p would be found between 'C and 'B', and between 'A' and 'B' i f the v a r i a b l e s were examined s e p a r a t e l y . If the v a r i a b l e s were put together i n t o a m u l t i v a r i a t e equation only one of them (most l i k e l y 'A', but t h i s does not n e c e s s a r i l y have to happen) would be found s i g n i f i c a n t . The m u l t i v a r i a t e approach has s e v e r a l advantages over u n i v a r i a t e a n a l y s e s . One i s that the researcher can determine how a l l the v a r i a b l e s i n f l u e n c e the outcome of 61 i n t e r e s t , yet be p r o t e c t e d from the i n t e r p r e t a t i o n of type I e r r o r s that i n e v i t a b l y r e s u l t from many u n i v a r i a t e a n a l y s e s . Secondly, i t i s sometimes the case that several, v a r i a b l e s each show some a s s o c i a t i o n with the outcome when analyzed alone, and are known to be h i g h l y c o r r e l a t e d with each other. Only a m u l t i v a r i a t e a n a l y s i s can so r t out which of the v a r i a b l e s i s most h i g h l y c o r r e l a t e d with the outcome, and which show some a s s o c i a t i o n because of t h e i r c o r r e l a t i o n with some other v a r i a b l e . F i n a l l y , i t i s rare that there i s only one c a u s a l v a r i a b l e i n any complex system, and a m u l t i v a r i a t e approach i s sometimes able to i n d i c a t e how the v a r i a b l e s i n t e r a c t with each o t h e r . There are, however, some drawbacks to a m u l t i v a r i a t e approach. O c c a s i o n a l l y a noncausal v a r i a b l e can enter i n t o an equation because of a high c o r r e l a t i o n with a c a u s a l v a r i a b l e . In the i l l u s t r a t i o n given above, i t c o u l d happen t h a t , in some samples, e r r o r s of measurement (which are always present) c o u l d r e s u l t in f a c t o r 'C appearing in the m u l t i v a r i a t e equation rather than 'A'. While t h i s would r a r e l y happen the p o s s i b i l i t y cannot, on the b a s i s of one sample, be e l i m i n a t e d . A second problem i s that the r e s u l t s can appear to change when d i f f e r e n t subsets of the t o t a l set of v a r i a b l e s are entered i n t o the a n a l y s i s . The magnitude of the r e g r e s s i o n weights are dependent on which v a r i a b l e s are entered i n t o the equation, and thus they cannot be i n t e r p r e t e d m e a n i n g f u l l y . Only when the v a r i a b l e s have a l l 62 been measured i n the same u n i t s or transformed to the same s c a l e do the r e g r e s s i o n weights have meaning. A t h i r d problem i s that i n t e r p r e t a t i o n of the r e s u l t s can o f t e n be very complicated. O c c a s i o n a l l y a v a r i a b l e that should, a c c o r d i n g to both theory and u n i v a r i a t e r e s u l t s , be a s s o c i a t e d with the outcome w i l l not have a c o e f f i c i e n t s i g n i f i c a n t l y d i f f e r e n t from zero i n the m u l t i v a r i a t e a n a l y s i s . I t i s o f t e n not c l e a r i f the theory i s in a c c u r a t e , or i f the true a s s o c i a t i o n i s being masked by other, l e s s important, v a r i a b l e s i n the equation. T h i s can be overcome by ensuring that t h i s v a r i a b l e i s entered i n t o the equation f i r s t , but without s u f f i c i e n t i n i t i a l knowledge an important a s s o c i a t i o n c o u l d be missed. F i n a l l y , the s e l e c t i o n of an a p p r o p r i a t e s t a t i s t i c a l model must be made. Th i s i s more important than i s o f t e n r e a l i z e d , f o r i f an i n t e r a c t i o n e f f e c t i n v o l v i n g a v a r i a b l e i s i n c l u d e d i t can mask the main e f f e c t of t h i s v a r i a b l e ( M o s t e l l e r & Tukey, 1977). Most models however do not i n c l u d e i n t e r a c t i o n e f f e c t s , and there i s then the danger that important i n t e r a c t i o n s can be overlooked. The simplest form of a m u l t i v a r i a t e a n a l y s i s would be to examine a l i n e a r r e l a t i o n s h i p between the v a r i a b l e s ( r i s k f a c t o r s ) and the response (here CHD). The parameters f o r such an equation c o u l d be determined by m u l t i p l e l i n e a r r e g r e s s i o n , which would produce an equation of the form: 63 (4.1) Y = a + B X + . . . + B X + . . . + B X 1 1 i i k k where: Y i s the outcome measure, B i s the i ' t h r e g r e s s i o n c o e f f i c i e n t i X i s the i ' t h independent v a r i a b l e i Thus the outcome can be i n t e r p r e t e d as the combination of s e v e r a l v a r i a b l e s (X's), that when summed with the weights s p e c i f i e d by the c o e f f i c i e n t s (B's) w i l l produce a response of magnitude Y. In the case of coronary heart disease the response v a r i a b l e (Y) i s not continuous ( e i t h e r the i n d i v i d u a l has or does not have coronary heart d i s e a s e ) . Rather, the response i s dichotomous, and i s u s u a l l y represented as 0 or 1, with 0 meaning " i s fre e of coronary heart d i s e a s e " and 1 meaning "has coronary heart d i s e a s e " . In order to c o n c e p t u a l l y f i t dichotomous data i n t o a continuous model, the O's and 1's can be seen as r e p r e s e n t i n g o b s e r v a t i o n s that r e f l e c t the i n d i v i d u a l s ' p r o b a b i l i t y of dev e l o p i n g coronary heart d i s e a s e . Thus the Y's c a l c u l a t e d in Eq. 4.1 c o u l d be grouped, with those near one being c l a s s e d as i n d i v i d u a l s who should develop coronary heart d i s e a s e , and those near zero as i n d i v i d u a l s who should not. From examination of the B's that make up the equation i t i s p o s s i b l e to determine which of the X's i n f l u e n c e the p r o b a b i l i t y of developing coronary heart d i s e a s e . L i n e a r m u l t i p l e r e g r e s s i o n has c e r t a i n draw-backs i f i t 64 i s a p p l i e d to a s i t u a t i o n where the p r o b a b i l i t y of an event i s of i n t e r e s t . Perhaps most annoying i s that i t o c c a s i o n a l l y produces i n d i v i d u a l p r o b a b i l i t i e s g r e a t e r than 1.0 or l e s s than 0.0. This can happen because the equation i s dependent on the sample from which i t was d e r i v e d . I f the sample does not c o n t a i n the f u l l range of values f o r each X that c o u l d be expected i n the p o p u l a t i o n , the equation can e a s i l y produce n o n s e n s i c a l r e s u l t s when a p p l i e d to i n d i v i d u a l s not in the o r i g i n a l sample who have X values that l i e near the p o p u l a t i o n extremes. Secondly, a l i n e a r model does not r e a l l y make b i o l o g i c a l "sense", i t i m p l i e s higher exposure always r e s u l t s i n higher r i s k . There u s u a l l y i s some upper l i m i t to a c a u s a l exposure that w i l l guarantee a 100% p r o b a b i l i t y , so that any exposure above t h i s l i m i t does not e n t a i l a d d i t i o n a l r i s k . A u s e f u l way to view the r e l a t i o n s h i p between r i s k f a c t o r s and coronary heart d i s e a s e i s to view coronary heart d i s e a s e as the r e s u l t of exposure to s e v e r a l d i f f e r e n t doses of v a r i o u s r i s k f a c t o r s . What i s then being searched f o r i s the o v e r a l l dose-response r e l a t i o n s h i p that best p r e d i c t s the outcome, in t h i s case coronary heart d i s e a s e . One method of o b t a i n i n g such an equation i s through the use of the l o g i s t i c model. 65 11 The L o g i s t i c Equation If a dose-response analogy i s used those with a b i o l o g i c a l p e r s p e c t i v e would argue f o r a model where l i t t l e r i s k i s a s s o c i a t e d with minimal exposure. Small i n c r e a s e s in exposure should l e a d to l i t t l e i n c r eased r i s k , s i n c e the b i o l o g i c a l system can i t s e l f respond to e l i m i n a t e the " t h r e a t " . However, as the dose begins to i n c r e a s e , there comes a poin t where i t overcomes the normal defenses of the organism, and a measurable r i s k of developing the d i s e a s e i s pre s e n t . T h i s r i s k w i l l continue to r i s e as the dose i n c r e a s e s u n t i l there i s a high p r o b a b i l i t y that almost a l l the organisms exposed to a dose of a c e r t a i n magnitude w i l l develop the outcome. However, some organisms are l i k e l y to be extremely r e s i s t a n t to the exposure. Much higher exposures w i l l be needed f o r these h i g h l y r e s i s t a n t organisms to develop the outcome, and the r e l a t i o n s h i p between exposure and r i s k w i l l approach an asymptote. F i n a l l y , there comes a p o i n t where the p r o b a b i l i t y i s f o r a l l purposes 1.0 (when a l l of the sample develops the outcome), and f u r t h e r exposure can have no e f f e c t . T h i s p e r s p e c t i v e suggests that a sigmoid curve r a t h e r than a s t r a i g h t l i n e i s the more a p p r o p r i a t e dose-response r e l a t i o n s h i p . Such a curve i s produced by the m u l t i p l e l o g i s t i c f u n c t i o n . I t produces a cumulative p r o b a b i l i t y d e n s i t y f u n c t i o n that s t a r t s o f f near zero ( i t approaches 0.0 66 a s y m p t o t i c a l l y ) , r i s e s i n i t i a l l y as an e x p o n e n t i a l , then f l a t t e n s o f f and approaches 1.0 as a negative e x p o n e n t i a l . The l o g i s t i c curve i s a l s o very easy to manipulate mathematically, and ,easy to t r a n s l a t e i n t o computer programs. The general form of the l o g i s t i c d i s t r i b u t i o n i s : 1 (4.2) Y = -T 1 + exp{-BX'} where B i s a ve c t o r of r e g r e s s i o n c o e f f i c i e n t s and X' a vecto r of v a r i a b l e s . The value of Y ranges between 0.0 and 1.0. Two methods of e s t i m a t i n g the parameters of the l o g i s t i c f u n c t i o n e x i s t (Gordon, 1974). The f i r s t method developed was an ada p t a t i o n of the d i s c r i m i n a n t a n a l y s i s approach ( C o r n f i e l d , Gordon and Smith, 1961). The second method (Walker and Duncan, 1967) uses a maximum l i k e l i h o o d approach, and i s d e r i v e d from the general dose-response r e l a t i o n s h i p that i s assumed to e x i s t between a r i s k f a c t o r and outcome. The two procedures g e n e r a l l y give s i m i l a r , but not e x a c t l y the same r e s u l t s . The method developed by C o r n f i e l d et a l . s t a r t s with an assumption that u n d e r l i e s any d i s c r i m i n a n t a n a l y s i s : that fo r a random v a r i a b l e x (a r i s k f a c t o r ) a p r o b a b i l i t y d e n s i t y f u n c t i o n f ( x ) or g(x) e x i s t s , depending on whether the outcome (here coronary heart disease) i s present ( f ( x ) ) or not present ( g ( x ) ) . They show that i f the x v a r i a b l e i s changed from a continuous v a r i a b l e to a c a t e g o r i c a l one (with boundaries x and dx), the p r o b a b i l i t y (p(x)) that an 67 i n d i v i d u a l w i l l be w i t h i n the range x and dx, and develop the outcome i s : (4.3) p(x) = 1/(1 + q ( g ( x ) ) / p ( f ( x ) ) ) . Here p i s the p r o b a b i l i t y the outcome w i l l occur at dose l e v e l x, and q=1-p. C o r n f i e l d et a l . then make two assumptions. The f i r s t i s t h a t f ( x ) and g(x) are normal d e n s i t y f u n c t i o n s with d i f f e r e n t means and v a r i a n c e s . S u b s t i t u t i n g these formulae i n t o (4.3) produces a dose-response curve that r e v e r s e s at a p o i n t dependent on the means and v a r i a n c e s of the two d i s t r i b u t i o n s . They then make the assumption that the dose-response curve should be monotonic ( f o r the reasons o u t l i n e d above). Such a curve w i l l e x i s t i f the v a r i a n c e s of f ( x ) and g(x) are eq u a l . Under t h i s c o n d i t i o n Eq. (4.3) can be shown to be equal t o : 1 (4.4) q/p(exp{-(u -u )/o 2 ( X - ( u +u )/2)}) f g f g where u = mean of f( x ) or g(x) i o 2 = common v a r i a n c e Some a l g e b r a i c manipulation leads to (4.5): 1 (4.5) P = 1 + exp{-(a + BX)} where: B = (u '+ u )/o 2 a = -ln(q/p) - B(-(u + u )/2) f 9 Equation 4.5 i s the same as (4.2) except t h a t a constant (a) 68 has been added to the exponent. C o r n f i e l d et a l . then g e n e r a l i z e d (4.5) to the m u l t i v a r i a t e case, where the X i s rep l a c e d by a vecto r of X's, and the B by a v e c t o r of c o e f f i c i e n t s . The m u l t i v a r i a t e e q u i v a l e n t of (4.3) i s : 1 (4.6) P = (q(g(X ,X ,...,X ))) 1 2 k (p( f ( X ,X ,...,X ))) 1 2 k Assumptions s i m i l a r to those made i n the u n i v a r i a t e case must be made i n order to develop a s u i t a b l e m u l t i v a r i a t e model f o r coronary heart d i s e a s e . If i t i s assumed that the f u n c t i o n s f and g are m u l t i v a r i a t e normal, Equation 4.7 r e s u l t s : « (4.7) P = 1/(1 + k(exp{-l/2(Q - Q )}) 0 1 where Q =F£ a (X -u ) (X -u ) t = 1 ,2 t L j i j t i i t j j t and a are the elements of the i n v e r s e of the i j t v a r i a n c e c o v a r i a n c e matrix. Equation (4.7) i s monotonic i f i t i s assumed that a'L\_x~aLtO f o r a l l i . C o r n f i e l d , Gordon and Smith (1961) make the a d d i t i o n a l assumption that ajj^ = aijO ^ o r a ^ * n o t e < 3 u a l 3> t h i s has the e f f e c t of e l i m i n a t i n g any i n t e r a c t i o n terms i n the model. These assumptions l e a d to the f o l l o w i n g m u l t i v a r i a t e l o g i s t i c e quation: 69 (4.8) P =  ( 1 + exp{-(a + X B X ) }) i i where: B = a (u -u ) i j i j j i jo a = -ln(q/p) - ( 1 / 2 ) £ £ a (u u i j i j i i J 1 u u ) iO jO The e s t i m a t i o n procedure* of Walker and Duncan (1967) was developed from a d i f f e r e n t set of assumptions. They simply assumed that the m u l t i p l e l o g i s t i c f u n c t i o n d e s c r i b e s the dose-response r e l a t i o n s h i p between the r i s k f a c t o r s and outcome and then develop a l e a s t squares approach to estimate the parameters of the model. They d e s c r i b e an i t e r a t i v e method to estimate the B parameters that w i l l p r o vide a s a t i s f a c t o r y f i t of the observed number of cases of the outcome using the model they assume to be c o r r e c t . T h e i r procedure has the advantage that the expected number of deaths equals that of the observed number (which does not n e c e s s a r i l y happen with the d i s c r i m i n a t i o n approach). I l l E a r l y Attempts at A s s e s s i n g M u l t i p l e Risk The s t u d i e s d e s c r i b e d i n the p r e v i o u s chapter were able to demonstrate that high l e v e l s of some c h a r a c t e r i s t i c s are a s s o c i a t e d with the l a t e r development of coronary heart d i s e a s e . These r e s e a r c h e r s were a l s o i n t e r e s t e d i n the e f f e c t s of being at r i s k on two or more v a r i a b l e s . 70 Examination of such cases c o u l d answer s e v e r a l q u e s t i o n s , among them: 1. are a l l the r i s k f a c t o r s independent of each other, or do some v a r i a b l e s appear to be r i s k f a c t o r s only because they are c o r r e l a t e d with other f a c t o r s ? 2. what i s the t o t a l r i s k a s s o c i a t e d with being at r i s k on more than one f a c t o r ? 3. do the r i s k f a c t o r s i n t e r a c t ; that i s , i s the r i s k f o r an i n d i v i d u a l who i s at r i s k on two f a c t o r s higher than that which would be p r e d i c t e d on the b a s i s of the r i s k a s s o c i a t e d with h i s l e v e l s on each f a c t o r alone? Some mention should be made about the use of the word i n t e r a c t i o n . In the medical l i t e r a t u r e the words synergism and i n t e r a c t ion tend to be used synonomously. However, d i f f e r e n t authors appear to mean d i f f e r e n t t h i n g s when they use these terms. A v a r i e t y of models can be used to d e s c r i b e a s y n e r g i s t i c r e l a t i o n s h i p , and of t e n j u s t what i s being d e s c r i b e d i s not c l e a r . With the l o g i s t i c equation, the r e g r e s s i o n approach used in the two methods of es t i m a t i n g the l o g i s t i c parameters d e s c r i b e d above means that i n t e r a c t i o n s w i l l most o f t e n represent the e f f e c t s of a v a r i a b l e d e r i v e d from the product of two suspected r i s k f a c t o r s . There are s e v e r a l s t a t i s t i c a l methods that c o u l d be 71 used to examine the p o s s i b l e interdependance or i n t e r a c t i o n of two v a r i a b l e s . One of the i n i t i a l steps would seem to be the examination of the c o r r e l a t i o n between these v a r i a b l e s . If two r i s k f a c t o r s are h i g h l y c o r r e l a t e d with each other i t i s p o s s i b l e that the r e l a t i o n s h i p of one to coronary heart d i s e a s e i s s p u r i o u s . However, i t i s r a r e that the c o r r e l a t i o n matrix of the r i s k f a c t o r s i s presented or mentioned in a paper d e s c r i b i n g the r e s u l t s of a coronary heart d i s e a s e study. When t h i s has been done the c o r r e l a t i o n s tend to be moderate to low, sometimes even among v a r i a b l e s thought to be h i g h l y c o r r e l a t e d with each other (see the p revious d i s c u s s i o n on the r e l a t i o n s h i p between s y s t o l i c and d i a s t o l i c blood p r e s s u r e ) . A second method i n v o l v e s c a l c u l a t i o n of the r i s k of d e v eloping coronary heart d i s e a s e at s p e c i f i c combinations of l e v e l s of r i s k f a c t o r s . T h i s i s u s u a l l y done by p r e s e n t i n g the r e s u l t s i n the form of two-way t a b l e s , and then comparing the r i s k a s s o c i a t e d with the combinations of i n t e r e s t with the r i s k a s s o c i a t e d with those at r i s k on one of the a t t r i b u t e s but not on the o t h e r ( s ) . The f i n a l method to be mentioned here i s sometimes used to check for p o s s i b l e i n t e r a c t i o n s of the v a r i o u s f a c t o r s . It c o n t r o l s f o r a f a c t o r by examining the r e l a t i o n s h i p between the other f a c t o r s and the development of coronary heart d i s e a s e at v a r i o u s l e v e l s of a c o n t r o l l e d f a c t o r . If two f a c t o r s do not i n t e r a c t , then the r e l a t i o n s h i p between 72 the f a c t o r not being c o n t r o l l e d f o r and the development of coronary heart d i s e a s e w i l l be the same at a l l l e v e l s of s t r a t i f i c a t i o n of the f a c t o r . The l a s t two methods are examples of a t r a d i t i o n a l e p i d e m i o l o g i c a l approach to the a n a l y s i s . They represent an easy c a l c u l a t i o n method by which the e f f e c t s of v a r i a b l e s can be h e l d constant, and an idea of the i n t e r a c t i o n s between these v a r i a b l e s can be obtained. The major drawback of these methods i s that when the number of v a r i a b l e s to be examined i s l a r g e compared to the number of o b s e r v a t i o n s , t a b l e s c o n t a i n i n g c e l l s with very few s u b j e c t s r e s u l t . T h i s produces s t a t i s t i c a l c o n c l u s i o n s f o r which there i s a l a r g e degree of u n c e r t a i n t y . The only way to get overcome t h i s problem i s to make the c a t e g o r i e s broader so as to i n c l u d e more s u b j e c t s . But t h i s can l e a d to c a t e g o r i e s that have l i t t l e p r a c t i c a l or t h e o r e t i c a l v a l u e . The r e s e a r c h reviewed in Chapter 3 o c c a s i o n a l l y examined the e f f e c t s of being at r i s k on more than one v a r i a b l e (see, f o r example, Dawber et a l . , 1957). The major problem with these analyses i s that the outcome of g r e a t e s t i n t e r e s t was death from coronary heart d i s e a s e , and there j u s t were not enough events (deaths) to make these analyses meaningful. To avo i d t h i s , broad c a t a g o r i e s of the independent v a r i a b l e s were used. However, analyses that combine b o r d e r l i n e and h y p e r t e n s i v e i n d i v i d u a l s who smoked for comparison with normotensives who d i d not smoke 73 (e.g. Dawber et a l . , 1957) are not very i n f o r m a t i v e . In ge n e r a l , these analyses show i n c r e a s e d r i s k with being at r i s k on more than one v a r i a b l e , but the nature of the r e l a t i o n s h i p i s f a r from c l e a r . IV E a r l y Use of the M u l t i p l e L o g i s t i c Equation The f i r s t attempt at a n a l y z i n g the data from a p r o s p e c t i v e coronary heart d i s e a s e study using a l o g i s t i c model was made by C o r n f i e l d , Gordon and Smith (1961; C o r n f i e l d , 1962). They analyzed the r e s u l t s of the Framingham study a f t e r s i x years of follow-up, c o n c e n t r a t i n g on the r e l a t i o n s h i p between the r i s k f a c t o r s serum c h o l e s t e r o l and blood pressure and the development of coronary heart d i s e a s e . At that time the authors b e l i e v e d i t was necessary to transform the independent v a r i a b l e s (serum c h o l e s t e r o l and s y s t o l i c blood pressure) to d i s t r i b u t i o n s that would approach n o r m a l i t y . T h i s was achieved by t a k i n g the log (to the base 10) of serum c h o l e s t e r o l and the l o g (to the base 10) of the s y s t o l i c blood pressure minus 75 ( i . e . log(SBP-75)). The l o g i s t i c equation was: -1 P = (1+exp{-(23.13+6.14(log(chol.))+3.29(log(SBP-75)))}) Both beta weights ( c o e f f i c i e n t s ) are s i g n i f i c a n t l y d i f f e r e n t from zero. C o r n f i e l d (1962) po i n t e d out that t h i s r e s u l t i s 74 incompatible with the notion that there i s some c r i t i c a l value of c h o l e s t e r o l or blood pressure that d i s t i n g u i s h e s between those who are at r i s k f o r coronary heart d i s e a s e and those who are not. He a l s o emphasized that the nature of the model i s such that f o r an i n d i v i d u a l who i s moderately high on e i t h e r one of these r i s k f a c t o r s an in c r e a s e or decrease of (x u n i t s ) of t h i s r i s k f a c t o r w i l l r e s u l t i n a gr e a t e r change i n P than f o r an i n d i v i d u a l at e i t h e r extreme of t h i s r i s k f a c t o r ' s d i s t r i b u t i o n . Two l a t e r papers had a much l a r g e r impact on the way suspected coronary heart d i s e a s e r i s k f a c t o r s were analyzed. T r u e t t , C o r n f i e l d and Kannel (1967) examined the Framingham r e s u l t s a f t e r 12 years of follow-up ( t h e i r r e s u l t s are presented in Table 5.1). T h e i r i n t e r e s t was p a r t l y m e t hodological, and one of t h e i r c o n c l u s i o n s was that a normal d i s t r i b u t i o n was not necessary f o r v a r i a b l e s to be used in t h i s type of a n a l y s i s (even dichotomous v a r i a b l e s c o u l d be i n c l u d e d ) . They a l s o concluded that t h i s type of approach provided more i n f o r m a t i o n about the r e l a t i v e importance of v a r i o u s r i s k f a c t o r s than other methods used at that time. Walker and Duncan (1967) a l s o i n c l u d e d an a n a l y s i s of the Framingham data i n the paper d e s c r i b i n g t h e i r method of a n a l y s i s (see Table 5.1). They a l s o b e l i e v e d dichotomous v a r i a b l e s c o u l d be i n c l u d e d i n the a n a l y s i s (they i n c l u d e d sex as a f a c t o r ) , and that t h e i r r e s u l t s p rovided u s e f u l i n f o r m a t i o n about v a r i o u s r i s k 75 f a c t o r s . A f t e r p u b l i c a t i o n of these two papers more re s e a r c h e r s began i n t e r p r e t i n g f i n d i n g s using the l o g i s t i c model. V General Comments The m u l t i p l e l o g i s t i c f u n c t i o n appears to be a u s e f u l way in which a l a r g e number of suspected r i s k f a c t o r s can be analyzed s i m u l t a n e o u s l y . T h i s a n a l y s i s i s able to a v o i d the problem of small c e l l s i z e that c h a r a c t e r i z e s a t r a d i t i o n a l e p i d e m i o l o g i c a l or s t a t i s t i c a l a n a l y s i s of t h i s type of data, and i s p o t e n t i a l l y able to d e s c r i b e some i n t e r r e l a t i o n s h i p s among the r i s k f a c t o r s that would be almost impossible to determine by u n i v a r i a t e methods. In a d d i t i o n , i t provides the researcher with a ready made (and i n t u i t i v e l y l o g i c a l ) model of how r i s k f a c t o r s might a f f e c t a process l i k e the development of coronary heart d i s e a s e . F i n a l l y , i t provides a p r e d i c t i o n of the p r o b a b i l i t y of an i n d i v i d u a l developing coronary heart d i s e a s e . While i t i s u n l i k e l y that any researcher would c l a i m that the p r o b a b i l i t i e s c a l c u l a t e d by t h i s type of f u n c t i o n represent the i n d i v i d u a l ' s exact p r o b a b i l i t y of developing coronary heart d i s e a s e , the u n d e r l y i n g b e l i e f would l i k e l y be that with s u f f i c i e n t understanding of the process and with i n c l u s i o n of the c o r r e c t v a r i a b l e s the r e a l p r o b a b i l i t y c o u l d be c l o s e l y approximated. However, the problem seldom d i s c u s s e d , or i f so only c a s u a l l y mentioned, i s the 76 i n t e r p r e t a t i o n of these r e s u l t s . A m u l t i p l e l o g i s t i c a n a l y s i s i s c o n c e p t u a l l y the same as m u l t i p l e l i n e a r r e g r e s s i o n or l i n e a r d i s c r i m i n a n t a n a l y s i s , so that the problems of i n t e r p r e t i n g l o g i s t i c c o e f f i c i e n t s are the same as the problems which e x i s t with the i n t e r p r e t a t i o n of other r e g r e s s i o n c o e f f i c i e n t s . There are s e v e r a l c o n s i d e r a t i o n s that must be kept in mind when i n t e r p r e t i n g r e g r e s s i o n c o e f f i c i e n t s of any s o r t . Perhaps the most important i s that the s i z e of the c o e f f i c i e n t (assuming the a n a l y s i s was done on v a r i a b l e s expressed i n n a t u r a l u n i t s , as most coronary heart d i s e a s e analyses are) has no meaning by i t s e l f . The magnitude of t h i s number depends on s e v e r a l f a c t o r s : which other v a r i a b l e s are i n the equation; the i n t e r r e l a t i o n s h i p among these v a r i a b l e s ; and even the method by which the coronary heart d i s e a s e events and nonevents are coded. A l l that can m e a n i n g f u l l y be s a i d about these v a r i a b l e s i s whether or not the confidence i n t e r v a l s of the c o e f f i c i e n t i n c l u d e s 0.0. T h i s leads to c e r t a i n p r e d i c t i v e problems i f i n t e r v e n t i o n i s c o n s i d e r e d on the b a s i s of such r e s u l t s . The simplest way to. p r e d i c t the r e s u l t s of an i n t e r v e n t i o n (based on the assumption that the model i s c o r r e c t ) would be to reduce a v a r i a b l e by amount 'x', and r e c a l c u l a t e the p r o b a b i l i t y of developing coronary heart d i s e a s e . The d i f f e r e n c e between t h i s and the o r i g i n a l p r o b a b i l i t y would be the expected b e n e f i t of the i n t e r v e n t i o n . Thus i t would 77 be p o s s i b l e to determine which i n t e r v e n t i o n s t r a t e g y would maximize the expected gain given the resources a v a i l a b l e . However, t h i s method, i f attempted at a l l , must be used with a great deal of c a u t i o n . Simply comparing the e f f e c t s of dec r e a s i n g a v a r i a b l e by a s p e c i f i e d amount i s not s u f f i c i e n t , because most analyses are done with the v a r i a b l e s i n t h e i r o r i g i n a l u n i t s , and a change of 5 u n i t s of c h o l e s t e r o l i s not the same as a change of 5 u n i t s of s y s t o l i c blood p r e s s u r e . Instead, i t may be b e t t e r to make 'x' a value that r e f l e c t s some f u n c t i o n of the v a r i a n c e of the d i s t r i b u t i o n , f o r example one standard d e v i a t i o n (Rosenman, et a l . , 1976). As mentioned e a r l i e r , two methods e x i s t for the c a l c u l a t i o n of the l o g i s t i c c o e f f i c i e n t s . H a l p e r i n , Blackwelder and V e r t e r (1971) compared the r e l a t i v e m e rits of these two methods. The method of Walker and Duncan does not r e q u i r e an assumption that the r i s k f a c t o r s are d i s t r i b u t e d normally (as does that of C o r n f i e l d ' s ) , but i t does r e q u i r e an i t e r a t i v e s o l u t i o n ( C o r n f i e l d ' s does n o t ) . Thus i f the s o l u t i o n does not converge q u i c k l y the Walker and Duncan method can prove very expensive to compute. H a l p e r i n et a l . showed that the Walker and Duncan method w i l l converge a s y m p t o t i c a l l y to the r e a l values of the constant (a) and the l o g i s t i c c o e f f i c i e n t s ( b ' s ) . In c o n t r a s t , only i f the p o p u l a t i o n value of b i s zero w i l l C o r n f i e l d ' s d i s c r i m i n a n t a n a l y s i s approach the true value of 78 'b'. In other cases i t converges to c o e f f i c i e n t s that are dependent on the p r o p o r t i o n of the p o p u l a t i o n that develops the outcome (here coronary heart d i s e a s e ) . They concluded that the d i s c r i m i n a n t a n a l y s i s approach w i l l l e a d to estimates that are o c c a s i o n a l l y i n e r r o r . When the same set of data were analyzed, d i f f e r e n c e s between the c o e f f i c i e n t s d e r i v e d from each method were found for both continuous and c a t e g o r i c a l v a r i a b l e s . They c o u l d not determine any method by which i t c o u l d be p r e d i c t e d that the d i s c r i m i n a n t c o e f f i c i e n t s would be i n e r r o r . In g e n e r a l , they concluded the d i s c r i m i n a n t approach w i l l work reasonably w e l l i f the purpose of the a n a l y s i s i s to estimate the r i s k f a c t o r s that are important. However, i f i n t e r e s t l i e s in the magnitude of the c o e f f i c i e n t s then t h i s method can give m i s l e a d i n g r e s u l t s . The primary advantage of the d i s c r i m i n a n t procedure, then, i s that i t took h a l f as long to compute. The importance of a c h i e v i n g unbiased estimates of the c o e f f i c i e n t s i s that the odds r a t i o can be d i r e c t l y c a l c u l a t e d from the l o g i s t i c equation ( F l e i s s , 1981). The e x p r e s s i o n e x p { c o e f f i c i e n t } estimates the odds r a t i o a s s o c i a t e d with a change of one u n i t of a v a r i a b l e . At l e a s t one attempt has been made to d e s c r i b e the e f f e c t s of a l t e r i n g r i s k f a c t o r s by c a l c u l a t i n g t h i s value (see Kannel, 1976). If the equation i s to be used in t h i s way, then the need for good estimates of the c o e f f i c i e n t s becomes important. 79 The next chapter examines the r e s u l t s obtained by using the m u l t i p l e l o g i s t i c e quation. From the preceding d i s c u s s i o n i t should be c l e a r that i n i t i a l i n t e r e s t does not l i e with determining e i t h e r the " c o r r e c t " value of the c o e f f i c i e n t s , or the best model to f i t the coronary heart d i s e a s e process. The f i r s t " t e s t " of i n t e r e s t i s how c o n s i s t e n t across s t u d i e s are the r e s u l t s . If a r i s k f a c t o r c o n s i s t e n t l y enters any equation, no matter which sample i t was measured from, or what other v a r i a b l e s were entered i n t o the equation, then i t i s l i k e l y an important r i s k f a c t o r . 80 C H A P T E R 5 Risk F a c t o r s and the L o g i s t i c Model In the s t u d i e s d i s c u s s e d above, and those to.be d i s c u s s e d below, there i s much v a r i a t i o n i n the v a r i a b l e s examined, the type of measurement used to r e c o r d the v a r i a b l e s , and i n the s t a t i s t i c a l techniques used to analyze the data. T h i s makes d i r e c t comparison of the r e s u l t s of the s t u d i e s d i f f i c u l t . For example, seemingly minor d i f f e r e n c e s such as a d i f f e r e n t way of coding the amount smoked c o u l d a l t e r the apparent magnitude of the i n f l u e n c e of smoking. T h i s s e c t i o n w i l l present a summary of the r e s u l t s obtained from l o g i s t i c a n a l y s i s of coronary heart d i s e a s e s t u d i e s . Given the v a r i e t y of v a l i d approaches that can be used i n a p r o s p e c t i v e study, the v a r i o u s l i m i t a t i o n s imposed on experimenters (e.g. lack of funds, lack of manpower), and the f a c t that most of these s t u d i e s set out to f i n d the important v a r i a b l e s that lead to the development of coronary heart d i s e a s e , i t would seem f r u i t l e s s to c r i t i c a l l y evaluate each study, f i n d the "best" study, and use these r e s u l t s to d e f i n e the r i s k f a c t o r s f o r coronary heart d i s e a s e . Almost any study can be c r i t i c i z e d , and the simple l o g i s t i c s of c a r r y i n g out a l a r g e s c a l e p r o s p e c t i v e study leave experimenters open to many p o s s i b l e c r i t i c i s m s . The approach used here w i l l be somewhat in keeping with that of Smith, Glass and M i l l e r (1980), that i s , i n t e r e s t w i l l 81 focus on the c o n s i s t e n c y of the r e s u l t s from study to study, i r r e s p e c t i v e of the v a r i a b l e s examined in the study or the methodological problems of the study. The i n i t i a l concern i s with which of the v a r i a b l e s examined in the study have c o e f f i c i e n t s with 95% confidence i n t e r v a l s that do not i n c l u d e z e r o . Those v a r i a b l e s that c o n s i s t e n t l y show up as s i g n i f i c a n t ( i n the manner j u s t d e f i n e d ) are l i k e l y to be important r i s k f a c t o r s f o r coronary heart d i s e a s e . I Risk F a c t o r s from the L o g i s t i c Model Table 5.1 presents a summary of the c o e f f i c i e n t s found in a v a r i e t y of p u b l i c a t i o n s that r e p o r t e d a l o g i s t i c a n a l y s i s of the r e s u l t s of a p r o s p e c t i v e coronary heart d i s e a s e study. Note that u n l i k e the r e s u l t s presented in Chapter 3, the r e s u l t s . i n Table 5.1 are not r e s t r i c t e d to American s t u d i e s (although these c o n s t i t u t e the m a j o r i t y ) . Those c o e f f i c i e n t s that are s i g n i f i c a n t l y d i f f e r e n t from zero are marked with an a s t e r i s k . It should be noted that while 29 equations are presented in the t a b l e , only 12 independent s t u d i e s are represented in the t a b l e . T h i s l i s t i s not meant to be exhaustive, as some p u b l i c a t i o n s i n c l u d e d l a r g e numbers of equations of v a r i o u s types. Thus Table 5.1 i s made up of those equations that i n c l u d e d new v a r i a b l e s i n the a n a l y s i s , or showed the e f f e c t s of i n c l u d i n g or dropping s t o n g l y suspected coronary heart d i s e a s e r i s k f a c t o r s . 82 Table 5.1 Summary of the l o g i s t i c c o e f f i c i e n t s r e p o r t e d i n a v a r i e t y of s t u d i e s . ( C o e f f i c i e n t s s i g n i f i c a n t l y d i f f e r e n t from 0.0 are f o l l o w e d by an *.) V a r i a b l e Type Equation Number (de ; t a i l s at : foot oi ! the pac je) 1 2 3 4 5 Constant -13.997 -12.812 -11.747 -11.790 -14.654 Age .081* .058 .079* . 053 .090* E x e r c i s e (0,1) -.190 -.349 -.348 Work a c t . (0,1) -.247 -.162 Schooling ( c a t . ) -.163 -.235 -.'146 -.214 H i s t o r y CHD .348 .294 .356 Diabetes .359 .231 Smoking (cat) .211* .277* . 223* .285* .264* Body Mass .045 .064 .071 .01 1 ECG abnorm. -.431 .042 -.389 Corneal Arcus .469* -.105 .458* Hematocrit .036 .018 .035 S y s t o l i c B.P. .020* .015 .017* .020* .016* D i a s t o l i c B.P. -.010 .011 C h o l e s t e r o l .011* .008* .010* .008* .012* Behavior type .616* .756* .628* .774* T r i g l y c e r i d e s - .001 -.000 LVH - ECG Glucose I n t o l . Dyspnea A l c o h o l B i r t h Place Sex Height Age x Age Choles. x Age Beta/alpha r a t i o .112 .000 .112 DBP x age Wt x HtxHt x Ag Equations 1-5 are from the Western C o l l a b o r a t i v e study and are reported i n Brand et a l . , 1976. -equation 1 i s f o r men aged 39-49 -equation 2 i s f o r men aged 50-59 -equation 3 i s f o r men aged 39-49 -equation 4 i s f o r men aged 50-59 -equation 5 i s f o r men aged 40-49 83 Table 5.1 cont. V a r i a b l e Type Equation Number (de ' t a i l s at . foot oi the par. >e) 6 7 8 9 10 Constant -13.257 -19.453 -9.795 -11.888 -3.427 Age . 1 22* .735* .033* .066* E x e r c i s e (0,1) Work a c t . (0,1) -.359* Sch o o l i n g (cat.) H i s t o r y CHD Diabetes Smoking (cat) .422* .439* . 1 99* .065* .503* Body Mass .026* .039 .026 ECG abnorm. .721 Corneal Arcus Hematocrit -.003 . 1 26 S y s t o l i c B.P. .007 .012* .019* .019* .381* D i a s t o l i c B.P. C h o l e s t e r o l .071* .027* .004* .008* .566* Behavior type T r i g l y c e r i d e s -.055 LVH - ECG .522 Glucose I n t o l . .231 Dyspnea .361* A l c o h o l .356* B i r t h Place -.031 Sex Height Age x Age -.002* Choles. x Age -.001* Beta/alpha r a t i o DBP x age Wt x HtxHt x Ag Equation 6 i s from H a l p e r i n et a l . , 1971 -men 40-49 from the Framingham study Equation 7 i s from McGee, 1973 -men aged 35-74 fron the Framingham study Equation 8 and 9 are from Keys et a l . , 1972 -equation 8 i s f o r R a i l r o a d men aged 40-59 -equation 9 i s f o r European men aged 40-59 Equation 10 i s from Welhelmesen et a l . , 1973 -Swedish men born i n 1913 84 Table 5.1 cont V a r i a b l e Type Equation Number ( d e t a i l s at foot of the page) 1 1 1 2 1 3 1 4 15 Constant Age E x e r c i s e (0,1) Work a c t . (0,1) Schooling (cat.) H i s t o r y CHD Diabetes Smoking (cat) Body Mass ECG abnorm. Corneal Arcus Hematocrit S y s t o l i c B.P. D i a s t o l i c B.P. C h o l e s t e r o l Behavior type Tr i g l y c e r ides LVH - ECG Glucose I n t o l . Dyspnea A l c o h o l B i r t h Place Sex Height Age x Age Choles. x Age Beta/alpha r a t i o DBP x age Wt x HtxHt x Ag -3.271 663* 401 * 504* -5.790 677* 770* .012* .029* -5.228 2.263* 01 1 10.900 .071* .361* .014* 1 .406* -.084 .017* .011* -5.370 .081* -.359* 1.359* .854* .009* .006 .007* -.059 -1.588* .053 Equation 11 i s from Wilhelmesen et e l . , 1973 -Swedish men born in 1913 Equation 12 and 13 i s from Kleinbaum et a l . , 1971 -equation 12 i s f o r white males -equation 13 i s f o r black males Equation 14 i s from T r u e t t et a l . , 1967 -men aged 30-62 from Framingham Equation 15 i s from Walker and Duncan, 1967 -men and women aged 29-62 from Framingham 85 Table 5.1 cont V a r i a b l e Type Equation Number ( d e t a i l s at foot of the page) 1 6 1 7 18 19 20 Constant Age E x e r c i s e (0,1) Work a c t . (0,1) Sc h o o l i n g (cat.) H i s t o r y CHD Diabetes Smoking (cat) Body Mass ECG abnorm. Corneal Arcus Hematocrit S y s t o l i c B.P. D i a s t o l i c B.P. C h o l e s t e r o l Behavior type Tr i g l y c e r ides LVH - ECG Glucose I n t o l . Dyspnea A l c o h o l B i r t h Place Sex Height Age x Age Choles. x Age Beta/alpha r a t i o DBP x age Wt x HtxHt x Ag •6.653 1 64' 017* 007* -6.441 1 63' 025* 007* -6.664 188* 013* 008* •6.869 . 1 93* .022* .008* -7.263 .331 .019* .005 Equations 16-20 are from McGee and Gordon, 1976 -equations 16 and 17 are from the Framingham study -equations 18 and 19 are from the Albany study -equation 20 i s from the Chicago Gas study 86 Table 5.1 cont V a r i a b l e Type Equation Number ( d e t a i l s at foot of the page) 21 22 23 24 25 Constant Age E x e r c i s e (0,1) Work a c t . (0,1) Scho o l i n g ( c a t . ) H i s t o r y CHD Diabetes . Smoking (cat) Body Mass ECG abnorm. Corneal Arcus Hematocrit S y s t o l i c B.P. D i a s t o l i c B.P. C h o l e s t e r o l Behavior type T r i g l y c e r i d e s LVH - ECG Glucose I n t o l . Dyspnea A l c o h o l B i r t h Place Sex Height Age x Age Choles. x Age Beta/alpha r a t i o DBP x age Wt x HtxHt x Ag -7.865 -6.165 •6.607 12.546 -12.026 342' 1 33 1 39* 253 260 037* 005 012 006* 024* 006* 027* 020* 040* 019* Equations 21-25 are from McGee and Gordon, 1976 -equation 21 i s from the Chicago Gas study -equations 22 and 23 are from the Western E l e c t r i c study -equations 24 and 25 are from the Tecumseh study 87 Table 5.1 cont. V a r i a b l e Type Equation Number ( d e t a i l s at foot of the page) 26 27 28 29 Constant Age Exerc i s e (0,1) Work a c t . (0,1) Schooling ( c a t . ) H i s t o r y CHD Diabetes Smoking (cat) Body Mass ECG abnorm. Corneal Arcus Hematocrit S y s t o l i c B.P. D i a s t o l i c B.P. C h o l e s t e r o l Behavior type T r i g l y c e r i d e s LVH - ECG Glucose I n t o l . Dyspnea A l c o h o l B i r t h Place Sex Height Age x Age Choles. x Age Beta/alpha r a t i o DBP x age Wt x HtxHt x Ag 10.707 .068* -8.11 -7.27 .089* 388* 004 441 * .684* 029* 006* .026* .002 025' 646* 008 036* Equation 26 i s from P o o l i n g P r o j e c t Report Group, 1978 -white males from a v a r i e t y of s t u d i e s , aged 40-59 Equations 27-28 are from S c h r o l l and Larsen, 1981 -equation 27 i s f o r Danish men aged 50 -equation 28 i s f o r Danish men aged 40-59 Equation 29 i s from Rabkin et a l . , 1977 -Canadian males between the ages 15 and 64 at the s t a r t of the study. 88 Some of the major c o n c l u s i o n s that come out of an examination of t h i s t a b l e i n c l u d e : 1. For some v a r i a b l e s that are present i n most of the analyses the c o e f f i c i e n t s are almost always s i g n i f i c a n t , and thus these v a r i a b l e s are l i k e l y very important. Serum c h o l e s t e r o l ( a l l but 2 e q u a t i o n s ) , smoking ( a l l but 1 eq u a t i o n ) , and s y s t o l i c blood pressure ( a l l but 4 equations) f a l l i n t o t h i s category. 2. The v a r i a b l e age, while not i n c l u d e d in many of the ana l y s e s , i s o f t e n s i g n i f i c a n t . In many cases the age range of the s u b j e c t s was r e s t r i c t e d by the study design, and thus i t seems l i k e l y that the v a r i a b l e would always be s i g n i f i c a n t with a l a r g e enough age range. 3. The v a r i a b l e "behavior type" i s s i g n i f i c a n t i n those four equations i n which i t i s entered. However, t h i s v a r i a b l e was only measured in one study, and thus i t i s impossible to determine i f i t i s important i n i t s own r i g h t , or i s unique to the sample that was s t u d i e d . 4. The v a r i a b l e "body mass" was measured i n f i v e d i f f e r e n t samples, but was only s i g n i f i c a n t in two. T h i s c o u l d imply that the importance of t h i s v a r i a b l e , when compared to other v a r i a b l e s , i s q u e s t i o n a b l e . 5. The v a r i a b l e "job a c t i v i t y " was measured i n two samples (three e q u a t i o n s ) , and was s i g n i f i c a n t i n one. It i s unclear i f sampling v a r i a t i o n or random e r r o r c o u l d 89 account f o r t h i s f i n d i n g , or i f d i f f e r e n c e s i n the job a c t i v i t i e s of the s u b j e c t s i n the two s t u d i e s might be a b e t t e r e x p l a n a t i o n . From Table 5.1 i t i s c l e a r that the magnitude of a c o e f f i c i e n t f o r any one v a r i a b l e v a r i e s from equation to equation. There are two p o s s i b l e sources of v a r i a t i o n f o r a c o e f f i c i e n t : e i t h e r the c o e f f i c i e n t c o u l d change because d i f f e r e n t v a r i a b l e s have been entered i n t o the equation, or d i f f e r e n c e s in the samples used i n the two s t u d i e s c o u l d account f o r the d i f f e r e n c e . While no attempt should be made to i n t e r p r e t the magnitude of the c o e f f i c e n t s i n Table 5.1, the amount of v a r i a b i l i t y between c o e f f i c i e n t s of the same v a r i a b l e i s of some i n t e r e s t . In g e n e r a l , v a r i a b l e s that are not c a u s a l l y r e l a t e d to the outcome of i n t e r e s t , or are h i g h l y c o r r e l a t e d with many other v a r i a b l e s a s s o c i a t e d with the outcome w i l l have c o e f f i c i e n t s that show a great deal of v a r i a b i l i t y when other v a r i a b l e s are entered i n t o the equation ( M o s t e l l e r and Tukey, 1977). V a r i a b l e s that are c l o s e to the outcome in the c a u s a l c h a i n , and not i n t e r c o r r e l a t e d to a l a r g e degree with other v a r i a b l e s , w i l l tend to show grea t e r s t a b i l i t y when unimportant e x t r a v a r i a b l e s are added i n t o the equation. From Table 5.1 i t i s d i f f i c u l t to estimate the extent to which the c o e f f i c i e n t s vary with the a d d i t i o n of e x t r a v a r i a b l e s (note that some of the equations, even though taken from the same study, o f t e n use s u b j e c t s of d i f f e r e n t 90 ages). Some idea of the v a r i a b i l i t y of the c o e f f i c i e n t s of a v a r i a b l e entered i n t o a number of equations with d i f f e r e n t s e t s of v a r i a b l e s ( h o l d i n g the study p o p u l a t i o n constant) can be found from H a l p e r i n , Blackwelder and V e r t e r (1971). In t h i s study a number of l o g i s t i c equations using d i f f e r e n t subsets of v a r i a b l e s were examined, but the data used to c a l c u l a t e each equation were from the same set of s u b j e c t s . Table 5.2 presents the r e s u l t s . The s u b j e c t s were 40-49 year o l d men who had been followed f o r 12 years i n the Framingham study. By comparing the v a r i a b i l i t y a s s o c i a t e d with the c o e f f i c i e n t s i n Table 5.1 with that found i n Table 5.2 i t can be seen that the c o e f f i c i e n t s do not vary much more between samples than they do w i t h i n samples. T h i s , combined with the c o n s i s t a n t s i g n i f i c a n c e of some c o e f f i c i e n t s , may be seen as an i n d i c a t i o n that the v a r i a b l e s noted above are s t r o n g l y a s s o c i a t e d with the development of coronary heart d i s e a s e . Thus, even i f w i t h i n each i n d i v i d u a l study the l o g i s t i c c o e f f i c i e n t of a given v a r i a b l e has a l a r g e 95% confidence i n t e r v a l , the c o n s i s t e n c y of t h i s f i n d i n g i n d i c a t e s an important r i s k f a c t o r . In a d d i t i o n , from Table 5.1 i t does not appear that r i s k f a c t o r s vary a great deal from sample to sample. 91 Table 5.2 Summary of the l o g i s t i c c o e f f i c i e n t s r e p o r t e d by H a l p e r i n , Blackwelder and V e r t e r (1971). V a r i a b l e Type Equat :ion Numt >er 1 2 3 4 5 Constant Age Smoking(0,1,2,3) Smoking(0,1) R e l a t i v e wt. ECG Hemoglobin S y s t o l i c B.P. C h o l e s t e r o l C h o l e s t e r o l ( 0 , 1 ) -13.253 . 1 22* .422* .026* .721 -.001 .007 .007* -6.118 1 .090* .857* .014* .006* -6.330 1.078* .016* .006* -4.449 1.055* 1 .053* .007* -5.428 .014* .006* V a r i a b l e Type Equat :ion Numt ser 6 7 8 9 10 Constant Age Smoking(0,1,2,3) Smoking(0,1) R e l a t i v e wt. ECG Hemoglobin S y s t o l i c B.P. C h o l e s t e r o l C h o l e s t e r o l ( 0 , 1 ) -4.307 1.025* .007* -3.665 . 007* -4.306 .017* -3.120 1.089* 1.097* . 927* -2.891 1.094* 1.022* 92 The r e s u l t s in Table 5.1 do not i n d i c a t e the presence of any new r i s k f a c t o r s . Those that appear to be s i g n i f i c a n t from an examination of Table 5.1 were a l s o found to be s i g n i f i c a n t i n u n i v a r i a t e analyses (as would be expected). The l i s t of p o t e n t i a l r i s k f a c t o r s that c o u l d be d e r i v e d from Table 5.1 i s , however, much s h o r t e r than what c o u l d be d e r i v e d from the u n i v a r i a t e r e s u l t s . Table 5.1 i n d i c a t e s that serum c h o l e s t e r o l and c i g a r e t t e smoking are r i s k f a c t o r s f o r coronary heart d i s e a s e . Age and s y s t o l i c blood pressure a l s o show up as r i s k f a c t o r s , although some qu e s t i o n s can be r a i s e d about the u s e f u l n e s s of s y s t o l i c blood pressure as a r i s k f a c t o r f o r coronary heart d i s e a s e 1 . Some h i g h l y suspected r i s k f a c t o r s are n o t i c e a b l y absent from Table 5.1. For example, the suspected b e n e f i t s of e x e r c i s e are absent, as i s the d e t r i m e n t a l e f f e c t of o b e s i t y . The obvious e x p l a n a t i o n i s that the v a r i a b l e s were poorly' measured. Yet c o n f i r m a t i o n of smoking as a r i s k f a c t o r , which was o f t e n c a t e g o r i z e d as d i d or d i d not smoke, would seem to i n d i c a t e that t h i s type of e x p l a n a t i o n i s s i m p l i s t i c . An a l t e r n a t i v e e x p l a n a t i o n f o r the f a i l u r e to c o n s i s t e n t l y f i n d s i g n i f i c a n t c o e f f i c i e n t s f o r body mass has been provided by Rabkin et a l . (1977). T h e i r f i n d i n g s i n d i c a t e that body mass was an important r i s k f a c t o r f o r younger men (that i s , those who were l e s s than 35 years of 1 T h i s w i l l be d i s c u s s e d i n g r e a t e r d e t a i l i n Chapter 6. 93 age at the s t a r t of the study) who were followed f o r a long p e r i o d of time (26 years i n the case.of t h i s s t u d y ) . No e f f e c t of body mass was found f o r men who were over 40 years of age at the s t a r t of the study. Most of the s t u d i e s i n Table 5.1 concentrated on o l d e r s u b j e c t s and i n v o l v e d a s h o r t e r follow-up p e r i o d , thus the e f f e c t on younger men c o u l d e a s i l y be missed. These r e s u l t s would seem to i n d i c a t e a f a i r l y c l e a r course of a c t i o n ; to concentrate only on those r i s k f a c t o r s that are c o n s i s t e n t l y s i g n i f i c a n t . In t h i s manner one c o u l d be f a i r l y sure that any r i s k f a c t o r being a l t e r e d was important. However, before a c c e p t i n g these r e s u l t s , i t i s necessary to examine the v a l i d i t y of the coronary heart d i s e a s e model presented by l o g i s t i c a n a l y s i s in order to see i f the suggestions f o r i n t e r v e n t i o n make "sense". I d e a l l y , the value (or behavior) of a model should be t e s t e d a g a i n s t some other model that attempts to d e s c r i b e the same process, but no other model of t h i s s o r t appears to e x i s t . In other cases, a g o o d n e s s - o f - f i t t e s t might be a s u f f i c i e n t check of the model, but here the model p r e d i c t s p r o b a b i l i t i e s of dying, not who i s and who i s not going to d i e . Thus there i s no exact way to t e s t the "goodness" of the models f i t to the observed data. An apprpximation of the g o o d n e s s - o f - f i t t e s t has been developed ( T r u e t t , C o r n f i e l d and Kannel, 1967). This t e s t uses the c a l c u l a t e d p r o b a b i l i t y of dying f o r each subject 94 (obtained from the l o g i s t i c e q u a t i o n ) , and rank orders the s u b j e c t s a c c o r d i n g to t h i s p r o b a b i l i t y . T h i s d i s t r i b u t i o n i s then d i v i d e d i n t o s e c t i o n s ( u s u a l l y d e c i l e s ) , and the sum of the i n d i v i d u a l p r o b a b i l i t i e s w i t h i n each d e c i l e i s o b t ained. T h i s sum i s then compared to the number of coronary heart d i s e a s e deaths observed among the s u b j e c t s in that d e c i l e . An i d e a l f i t would be obtained when the number of observed deaths matches the number of p r e d i c t e d deaths (the sum of the p r o b a b i l i t i e s ) i n each d e c i l e . Such an a n a l y s i s , when i n c l u d e d with the r e s u l t s , u s u a l l y i n d i c a t e s that the f i t of the l o g i s t i c f u n c t i o n i s "good". However, c e r t a i n f a c t o r s should be kept in mind when support for the model i s presented in t h i s manner. The maximum l i l e l i h o o d a l g o r i t h m that i s most oft*en used to c a l c u l a t e the l o g i s t i c equation has the c o n s t r a i n t that the p r e d i c t e d number of deaths ( i . e . the t o t a l sum of the p r o b a b i l i t i e s ) must equal the t o t a l number of observed deaths. If the data c o n t a i n a few i n d i v i d u a l s who have low r i s k f a c t o r scores but develop coronary heart d i s e a s e , the equation w i l l a s s i g n a low p r o b a b i l i t y f o r s i m i l a r i n d i v i d u a l s . Thus the sum of the p r o b a b i l i t i e s f o r these low r i s k i n d i v i d u a l s should be f a i r l y c l o s e to the observed t o t a l . As pointed out by T r u e t t et a l . "Any normally d i s t r i b u t e d compound of r i s k f a c t o r s 1 can be used to obtain 1 The c e n t r a l l i m i t theorem suggests that such compounds w i l l tend to be normally d i s t r i b u t e d (Truett et a l . , 1967). 95 a r i s k f u n c t i o n l e a d i n g to agreement between o b s e r v a t i o n and e x p e c t a t i o n . " (1967, P. 521-522). What i s d e s i r e d i s the f u n c t i o n that leads to the ste e p e s t r i s k g r a d i e n t . The g o o d n e s s - o f - f i t approach can l e a d to some m i s l e a d i n g c o n c l u s i o n s i f the r i s k g r a d i e n t i s not a l s o examined. F i n a l l y , i t should be noted that the p r a c t i c e of d i v i d i n g the d i s t r i b u t i o n up i n t o d e c i l e s i s simply a convention. Other d i v i s i o n s would change the apparent f i t df the model. For example, i f the d i s t r i b u t i o n was d i v i d e d i n t o f i f t h s the f i t might be expected to improve, but i f d i v i d e d i n t o twentieths i t might decrease d r a m a t i c a l l y . There i s , then, no s t a t i s t i c a l way of t e s t i n g the v a l i d i t y of the model. The one method that has been suggested should i n d i c a t e a good f i t . I t should be noted that those who developed the t e s t suggested i t was s u i t a b l e only as a method of comparing the f i t of v a r i o u s equations. However, as w i l l be shown below, even these r e s u l t s can be mi s l e a d i n g . More i n d i r e c t methods of t e s t i n g the model have to be used. S e v e r a l a l t e r n a t i v e s are p o s s i b l e . One method i s to use the parameters d e r i v e d from one set of s u b j e c t s to p r e d i c t the i n d i v i d u a l r i s k of s u b j e c t s taken from a d i f f e r e n t sample. While the same g o o d n e s s - o f - f i t t e s t i s used, the parameters are not being t e s t e d on the data from which the model was d e r i v e d . In a d d i t i o n , the f i t of the equation used on the second set of s u b j e c t s c o u l d be 96 compared to that obtained using the data from which the equation was d e r i v e d . Thus some idea of how much i s " l o s t " when the equation i s used to p r e d i c t coronary heart d i s e a s e r i s k i n another p o p u l a t i o n can be obtained. A second method c o n s i s t s of examining p r e d i c t i o n s of the model in d e t a i l . T h i s approach i s , e s s e n t i a l l y , to assume the model i s c o r r e c t , and to determine i f the p r e d i c t i o n s seem reasonable. If they are not reasonable, then the model i t s e l f may be suspect. A t h i r d method i s to t e s t the model by attempting to a l t e r s p e c i f i c r i s k f a c t o r s and to look f o r changes in the r a t e s of coronary heart d i s e a s e . 11 T e s t s of the L o g i s t i c Model A Examination of the L o g i s t i c P r e d i c t i o n s An examination of the model's p r e d i c t i o n s i s made d i f f i c u l t by the l a r g e number of equations that have been p u b l i s h e d (see Table 5.1). Since there does not yet appear to be any " c o r r e c t " equation, the set of r e s u l t s produced by H a l p e r i n , Blackwelder and V e r t e r (1971), (see Table 5.2) w i l l be used here. In t h e i r paper the authors do not present the means and standard d e v i a t i o n s of the sample from which the equations were d e r i v e d . Thus the means and standard d e v i a t i o n s used below have been taken from a 97 d i f f e r e n t source (McGee and Gordon, 1976). These data were obtained from the same p o p u l a t i o n as were the equations (the Framingham sample), but the exact number of s u b j e c t s may not be the same. However, as the purpose of the f o l l o w i n g a n a l y s i s i s demonstrative rather than to make a c t u a l p r e d i c t i o n s t h i s problem should be of l i t t l e present concern. From Table 5.2 i t can be seen that the l o g i s t i c equation d e r i v e d from the r i s k f a c t o r serum c h o l e s t e r o l (#7) i s : -1 (5.1) P = (1 + exp{-(-3.6648 + . 0 0 7 2 ( c h o l e s t e r o l ) ) } ) , and that f o r s y s t o l i c blood pressure alone (#8) i s : -1 (5.2) P = (1 + exp{-(-4.3064 + .0167(S.B.P.))}). The equation f o r both v a r i a b l e s (#5) i s : -1 (5.3) P = (l+exp{-(-5.428l+.0062(chol)+.0l44(S.B.P.))}). A l l of the c o e f f i c i e n t s in equations 5.1, 5.2 and 5.3 are s t a t i s t i c a l l y s i g n i f i c a n t . One q u e s t i o n of i n t e r e s t i s which of the v a r i a b l e s (serum c h o l e s t e r o l or s y s t o l i c blood pressure) i s more important. That i s , i f i t was only p o s s i b l e to change one v a r i a b l e , which would best be focused on? Here i t w i l l be assumed that each can be changed with equal e f f o r t . The f i r s t attempt at a s s e s s i n g which v a r i a b l e i s more important might deal with which of the equations best p r e d i c t s coronary heart d i s e a s e events. H a l p e r i n et 98 a l . c a l c u l a t e d the p r e d i c t e d and observed number of deaths i n d e c i l e s of r i s k f o r equations 5.1, 5.2 and 5.3. T h e i r method of comparison was through the use of the r e g r e s s i o n c o e f f i c i e n t s of the observed on the expected ( p r e d i c t e d ) v a l u e s . A l l the r e g r e s s i o n c o e f f i c i e n t s are near 1.0, but with only 10 data p o i n t s upon which to do a r e g r e s s i o n a n a l y s i s , i n t e r p r e t a t i o n of the d i f f e r e n c e s i n the r e g r e s s i o n values i s d i f f i c u l t . The c h i square goodness-of-f i t t e s t i s more of t e n used f o r t h i s type of s i t u a t i o n , and these r e s u l t s are presented i n Table 5.3. Note that Table 5.3 i n c l u d e s r e s u l t s from other equations d e s c r i b e d in Table Table 5.3 R e s u l t s of a X 2 g o o d n e s s - o f - f i t t e s t comparing the p r e d i c t i v e a b i l i t y of l o g i s t i c equations c o n t a i n i n g v a r i o u s numbers of v a r i a b l e s . (equations from H a l p e r i n , Blackwelder and V e r t e r , 1971). Number of V a r i a b l e Names Chi Var i a b l e s Square 1 c h o l e s t e r o l 6.681 1 s y s t o l i c blood pressure 3.458 2 c h o l e s t e r o l / s y s t o l i c b.p. 5.936 2 c h o l e s t e r o l / s m o k i n g 4.705 3 c h o l . / s y s t o l i c b.p./smoking 6.279 7 c h o l . / s y s . b.p./age/ECG/ r e l a t i v e weight/smoking/ hemoglobin 6.779 From t h i s t a b l e i t would appear that the model which c o n t a i n s only s y s t o l i c blood pressure best f i t s the data. T h i s c o n c l u s i o n i s not in keeping with the r e s u l t s d i s p l a y e d in Table 5.1, s i n c e there s y s t o l i c blood pressure d i d not 99 seem to be the most important r i s k f a c t o r f o r coronary heart d i s e a s e . An e x p l a n a t i o n f o r t h i s d i screpancy l i e s i n the way the model was j u s t v a l i d a t e d . An examination of F i g u r e 5.1 shows that the equations using only one r i s k f a c t o r (5.1 and 5.2) r i s e more slowly than those with more than one r i s k f a c t o r . The t h e o r e t i c a l j u s t i f i c a t i o n f o r t h i s model ( d e s c r i b e d i n Chapter 4) suggests that, in the case where every v a r i a b l e that e f f e c t e d coronary heart d i s e a s e was i n c l u d e d i n the model, a l l the cases would be i n the tenth (some p o s s i b l y i n the n i n t h ) d e c i l e of r i s k . Thus the l o g i c used to develop t h i s type of a n a l y s i s argues that equations with the steeper slope should be p r e f e r r e d , even i f t h e i r g o o d n e s s - o f - f i t r e s u l t i s not as "good" as other equations. If an equation has a steeper slope than another equation t h i s means the f i r s t equation w i l l have more of the i n d i v i d u a l s who a c t u a l l y develop coronary .heart d i s e a s e i n the upper r i s k d e c i l e s . The c h i square g o o d n e s s - o f - f i t value may not be as low as the second equation's because of the extremely small expected and observed number of cases in the lower d e c i l e s . F i g u r e 5.1 i n d i c a t e s there i s l i t t l e to chose between r e s u l t s based on the s i n g l e r i s k f a c t o r s of serum c h o l e s t e r o l and s y s t o l i c blood p r e s s u r e . I t a l s o shows l i t t l e i s gained from combining the two v a r i a b l e s , as t h i s slope i s not much d i f f e r e n t from that of the two s i n g l e v a r i a b l e s . The slope of the equations using the other two 100 v a r i a b l e combination ( c h o l e s t e r o l and smoking) 1 i s g r e a t e r than that of c h o l e s t e r o l and blood pressure, and the a d d i t i o n of more v a r i a b l e s c o n t i n u e s to improve the r e s u l t s . Most other r e s e a r c h e r s who use l o g i s t i c equations t e s t the p r e d i c t i v e a b i l i t y of t h e i r equations using d e c i l e s (or some other grouping of p e r c e n t i l e s ) of r i s k . The f i t s are u s u a l l y d e s c r i b e d as good, but there does not appear to have been any reported attempt to check the f i t i n the manner d e s c r i b e d i n F i g u r e 5.1. Thus, attempts to measure the adequacy of the model through g o o d n e s s - o f - f i t t e s t s can be m i s l e a d i n g . No p u b l i s h e d work has attempted to develop a model using the r i s k g r a d i e n t as the s e l e c t i o n c r i t e r i o n . Such a model would have to be developed step-wise, that i s , the v a r i a b l e that produced the steepest g r a d i e n t by i t s e l f would be entered f i r s t , then the v a r i a b l e that most i n c r e a s e d the g r a d i e n t would be i n c l u d e d second, and so on u n t i l the a d d i t i o n of e x t r a v a r i a b l e s no longer i n c r e a s e d the g r a d i e n t . T h i s approach might h e l p to d i f f e r e n t i a t e between v a r i a b l e s and t h e i r i n f l u e n c e on coronary heart d i s e a s e to a more u s e f u l degree. However, these r e s u l t s would only be of use i f the t h e o r e t i c a l base of the model has v a l i d i t y . 1 U n f o r t u n a t e l y , an equation f o r s y s t o l i c blood pressure and smoking was not produced. 101 F i g u r e 5.1 P r e d i c t e d number of c a s e s of c o r o n a r y h e a r t d i s e a s e per d e c i l e of r i s k f o r the e q u a t i o n s d e s c r i b e d i n T a b l e 5.3 25 H P r e d I c t e d D e a t h s 201 15H 1(H 5H Cholesterol Systolic B.P. Choi. + S.B.P. Choi. + Smoking Choi. + S.B.P. + Smk. Choi. + S.B.P. + Smk. + Age + ECG + Relative wt + Hemoglobin 2 ~6 8 — i — 10 Deci les of Risk 102 B V a l i d a t i o n on Other Subjects Two s t u d i e s have attempted to v a l i d a t e r e s u l t s on a d i f f e r e n t p o p u l a t i o n (McGee and Gordon, 1976; Keys et a l . , 1972). Keys et a l . obtained data f o r three groups of men between the ages of 40 and 59: 2,404 U.S. R a i l r o a d workers, 8,728 European men, and data on 6,221 men from the U.S. P o o l i n g P r o j e c t 1 . The suspected r i s k f a c t o r s measured were age, s y s t o l i c blood p r e s s u r e , serum c h o l e s t e r o l , smoking h a b i t s , and a body mass index. The l o g i s t i c equations d e r i v e d from the R a i l r o a d and European data are presented in Table 5.1 (the P o o l i n g P r o j e c t data were used only f o r v a l i d a t i o n ) . Of i n t e r e s t i s the authors' comparison of the p r e d i c t e d number of cases of coronary heart d i s e a s e among the R a i l r o a d men (using c o e f f i c i e n t s d e r i v e d from the European men) with the a c t u a l number of cases observed among the R a i l r o a d men. They a l s o compared the p r e d i c t e d number of European cases (from the R a i l r o a d equation) with the observed number of European cases. The p r e d i c t e d number of cases i n each d e c i l e was h i g h l y c o r r e l a t e d with the observed number, but the equations over (or under) p r e d i c t . The R a i l r o a d equation p r e d i c t e d a t o t a l of 664 European cases of coronary heart d i s e a s e , while only 405 cases were a c t u a l l y observed. The authors d i s t i n g u i s h e d between simple reported cases of coronary heart disease and 1 T h i s sample was c r e a t e d by p o o l i n g the samples of s e v e r a l of the s t u d i e s d i s c u s s e d i n Chapter 3. 103 hard coronary heart d i s e a s e cases (cases with more ev i d e n c e ) , but t h i s d i s t i n c t i o n d i d not i n f l u e n c e the r e s u l t s to a great degree: the R a i l r o a d equation p r e d i c t e d 238 cases of hard coronary heart d i s e a s e , while only 136 cases were observed. The European equation u n d e r p r e d i c t e d the R a i l r o a d r e s u l t s : 140 cases were p r e d i c t e d (48 hard cases) while 210 were observed (78 hard c a s e s ) . In a d d i t i o n to these p r e d i c t i o n s , the r e s e a r c h e r s a l s o c a l c u l a t e d the p r e d i c t e d number of coronary heart d i s e a s e cases i n the P o o l i n g P r o j e c t sample, using the equations d e r i v e d from the "hard" data of the European and R a i l r o a d men. The R a i l r o a d equation p r e d i c t e d 158 cases, the European equation p r e d i c t e d 104. Both p r e d i c t i o n s were l e s s than the 223 observed cases. It i s of i n t e r e s t that p r e d i c t i o n s based on these two equations, both f o r the other set of s u b j e c t s and f o r the P o o l i n g P r o j e c t ' s s u b j e c t s , are so d i f f e r e n t . These d i f f e r e n c e s are not easy to e x p l a i n . One p o s s i b l e e x p l a n a t i o n l i e s i n the d i f f e r e n t c u l t u r a l backgrounds of the s u b j e c t s . The European men are, of course, from a s u b s t a n t i a l l y d i f f e r e n t c u l t u r e , with d i f f e r e n t c u l t u r a l h a b i t s . C u l t u r a l d i f f e r e n c e s are o f t e n used to e x p l a i n v a r i a t i o n s i n the observed worldwide coronary heart d i s e a s e death r a t e s . T h i s i s an e x p l a n a t i o n that the authors appear to p r e f e r , s i n c e they e x p l a i n the u n d e r p r e d i c t i o n of the Railway mens' death rate by the European equation as being a 1 04 f u n c t i o n of the d i f f e r e n c e s between the two c u l t u r e s . T h i s f a i l s , however, to e x p l a i n the f a i l u r e of the Railway equation to p r e d i c t the P o o l i n g P r o j e c t death r a t e to an acc e p t a b l e degree of accuracy. The authors e x p l a i n t h i s f a i l u r e i n terms of the d i f f e r e n c e s i n d i a g n o s t i c c r i t e r i a used between t h e i r study and that of the P o o l i n g P r o j e c t . The d i a g n o s t i c c r i t e r i a f o r the P o o l i n g P r o j e c t was d i f f e r e n t from that used by Keys et a l . , and t h i s i s one p o s s i b l e reason f o r the d i f f e r e n c e s . Support f o r t h i s e x p l a n a t i o n comes from the o b s e r v a t i o n that i f the death r a t e s f o r the Railway men are c a l c u l a t e d , a rate of 6.5/1000 i s found (78 deaths among 2404 men over 5 y e a r s ) , while the ra t e for the P o o l i n g P r o j e c t i s 7.2/1000 per year. If the weighted average (10.1/1000 per year) of the s i x p r o s p e c t i v e s t u d i e s reviewed i n Chapter 3 i s used as a North American average f o r these kind of s t u d i e s , then the death experience of the Railway men i s much lower than would be expected. T h i s can be seen as implying a s t r i c t e r d i a g n o s t i c c r i t e r i a than i s normally a p p l i e d to coronary heart d i s e a s e s t u d i e s , or i t may mean that the Railway men d i f f e r e d from the normal p o p u l a t i o n i n some other way (note that they were a l l working). If the c r o s s - c u l t u r a l e x p l a n a t i o n i s c o r r e c t , the c o r r e l a t i o n s between the observed and p r e d i c t e d number of deaths suggest the same r i s k f a c t o r s are important i n each c u l t u r e , but one c u l t u r e i s e f f e c t e d by a c e r t a i n l e v e l of a r i s k f a c t o r to a grea t e r degree. 105 The second major attempt at c r o s s - v a l i d a t i o n was r e p o r t e d by McGee and Gordon (1976). Here they were i n t e r e s t e d i n comparing p r e d i c t e d and observed death r a t e s f o r a number of s t u d i e s that developed a l o g i s t i c equation from t h e i r data. T h e i r method was to compare the adequacy of death r a t e s p r e d i c t e d from the l o g i s t i c equation (as developed from the study's data) with p r e d i c t e d death experience d e r i v e d from the use of the Framingham c o e f f i c i e n t s on these data. They concluded that the Framingham experience was not d i f f e r e n t from that of the other s t u d i e s , as the p r e d i c t i o n s from the Framingham equation were almost as good as those from equations d e r i v e d from the a c t u a l data. The problems with using a g o o d n e s s - o f - f i t t e s t approach were d i s c u s s e d above. However, an examination of p l o t s comparing the r i s k g r a d i e n t d e r i v e d from the Framingham p r e d i c t i o n s and that of the p r e d i c t i o n s d e r i v e d from the study's equation showed no c o n s i s t e n t trend f o r the Framingham g r a d i e n t to be steeper or more shallow. The Framingham model a l s o d i d not c o n s i s t e n t l y over- or under-p r e d i c t the number of deaths. It o v e r - p r e d i c t e d the number of deaths i n the Albany study (by 6.5%), Chicago Western E l e c t r i c (by' 19.5%) and Tecumeseh (by 26.0%), and u n d e r p r e d i c t e d the Chicago Gas Study by 4.4%. Given the r e l a t i v e l y small sample s i z e s of these s t u d i e s , these percentages would seem to r e f l e c t r e l a t i v e l y good p r e d i c t i o n 106 by t h i s equation. These two s t u d i e s give s l i g h t l y c o n f l i c t i n g r e s u l t s . The Keys et a l . study (1972) seems to have i n d i c a t e d the p o s s i b i l i t y that l a r g e s c a l e d i f f e r e n c e s can be expected between the p r e d i c t i v e a b i l i t y of v a r i o u s equations. Yet the r e s u l t s of the McGee and Gordon (1976) study suggests that f a i r l y good r e s u l t s can be expected by a p p l y i n g an equation d e r i v e d from one p o p u l a t i o n to data from a d i f f e r e n t sample. Perhaps these two f i n d i n g s can be r e c o n c i l e d by the f a c t that one study examined c u l t u r a l d i f f e r e n c e s while the other looked at v a r i a t i o n w i t h i n a c u l t u r e . However, t h i s e x p l a n a t i o n i s post-hoc. C Attempts to A l t e r Risk F a c t o r s One of the most c o n v i n c i n g methods by which c a u s a l i t y can be demonstrated i s i n the case where f a c t o r A i s a l t e r e d and a change in the outcome i s then observed. To date, no pure experiments have been performed i n the area of coronary heart d i s e a s e r i s k f a c t o r s , and e t h i c a l c o n s i d e r a t i o n s would l e a d one to hope that such a study would never be done. There are, however, some quas i - e x p e r i m e n t a l designs that have examined (or are in the process of examining) the e f f e c t of a l t e r a t i o n s of r i s k f a c t o r s and subsequent changes in the development of coronary heart d i s e a s e . R e s u l t s from two such s t u d i e s have been r e p o r t e d , although some p r e l i m i n a r y data have been r e l e a s e d by other s t u d i e s . 107 i . The North K a r e l i a P r o j e c t T h i s i s the f i r s t p r o s p e c t i v e coronary heart d i s e a s e r i s k f a c t o r i n t e r v e n t i o n t r i a l to be completed. The m o r t a l i t y and m o r b i d i t y r a t e s f o r coronary heart d i s e a s e vary widely i n F i n l a n d , and the h i g h e s t r a t e i n the 1960's was i n North K a r e l i a (a mainly r u r a l area l o c a t e d i n Eastern F i n l a n d ) . The North K a r e l i a P r o j e c t (Puska, 1973) was s t a r t e d i n 1972 with the aim of reducing the l e v e l s of coronary heart d i s e a s e m o r b i d i t y and m o r t a l i t y in North K a r e l i a . The program i n v o l v e d comprehensive p r e v e n t i o n and i n t e r v e n t i o n s t r a t e g i e s , and was c a r r i e d out w i t h i n the s t r u c t u r e of the h e a l t h and s o c i a l s e r v i c e departments that e x i s t e d i n the community. e The program d e f i n e d i t s o b j e c t i v e in the f o l l o w i n g way (Puska et a l . , 1979): Main o b j e c t i v e : to reduce m o r t a l i t y and m o r b i d i t y , e s p e c i a l l y from CVD, with s p e c i a l r e f e r e n c e to middle-aged men. Intermediate o b j e c t i v e s : to reduce the known c a r d i o v a s c u l a r r i s k f a c t o r s -that i s , smoking, serum c h o l e s t e r o l c o n c e n t r a t i o n , and blood pressure - and to promote e a r l y d e t e c t i o n , treatment, and r e h a b i l i t a t i o n i n people with CHD. N a t i o n a l o b j e c t i v e : to t e s t the f e a s i b i l i t y and e f f e c t of t h i s approach and to provide t e s t e d methods and programmes f o r nationwide use i n connection with the c o n t r o l of CHD and other h e a l t h problems. (p. 1173) The r e s e a r c h e r s s t a t e they s e l e c t e d the three r i s k f a c t o r s to be a l t e r e d on the b a s i s of p u b l i s h e d r e s e a r c h on coronary heart d i s e a s e r i s k f a c t o r s . Note that these r i s k f a c t o r s are the ones which appear, in Table 5.1, to be the most 108 important. The study i t s e l f i n v o l v e d an i n i t i a l b a s e l i n e survey c a r r i e d out i n 1972 i n North K a r e l i a and Kuopis (a county used as a c o n t r o l , i t i s a r u r a l area s i m i l a r t o North K a r e l i a and a l s o i n Eastern F i n l a n d ) . A r e p r e s e n t a t i v e random sample of 6.6% of men and women born between the years 1913-1947 (thus aged 25-59 in 1972) was drawn from the p o p u l a t i o n of each county. F i v e years l a t e r a second sample of the same s i z e ( c a l l e d the t e r m i n a l sample) was drawn from the same two areas. T h i s sample a l s o c o n s i s t e d of men and women born between the years 1913 and 1947 (and thus aged 30-64 i n 1975). It should be noted that u n l i k e many l o n g i t u d i n a l s t u d i e s , the same sample was not fol l o w e d over the f i v e years of the study. T h i s procedure was used i n an attempt to assess the e f f e c t i v e n e s s of the i n t e r v e n t i o n on the whole p o p u l a t i o n . T h i s avoids the problem of an i n f l a t e d e f f e c t due to measuring the same s u b j e c t s twice (the Hawthorne e f f e c t ) but r e s u l t s i n a s i t u a t i o n where the magnitude of i n d i v i d u a l changes cannot be estimated. There are two areas of i n t e r e s t with t h i s study. If the i n t e r v e n t i o n was s u c c e s s f u l , that i s i f the r i s k f a c t o r s s e l e c t e d by the experimenters were decreased i n North K a r e l i a but not i n the c o n t r o l area, then t h i s p r o j e c t can be viewed as an experiment in which the independent v a r i a b l e s are the r i s k f a c t o r s , and the dependent v a r i a b l e s are the m o r t a l i t y or mo r b i d i t y experience of the two 109 communities. Secondly, Puska et a l . (1979) have presented a t a b l e of estimated coronary heart d i s e a s e r i s k based on an as yet unpublished l o g i s t i c a n a l y s i s of a f i v e year f o l l o w -up of the o r i g i n a l sample f o r acute myocardial i n f a r c t i o n . These r e s u l t s can be compared to the o v e r a l l p o p u l a t i o n trends they measured, and gi v e an i n d i c a t i o n of the v a l i d i t y of the l o g i s t i c model. I t would appear that the design of the study i s adequate, with a l a r g e enough sample s i z e and p a r t i c i p a t i o n r a t e to make the measures meaningful. The p a r t i c i p a t i o n r a t e and sample s i z e i s about the same in the c o n t r o l area in the s e l e c t e d age groupings (see Table 5.4). Table 5.4 Number of s u b j e c t s p a r t i c i p a t i n g i n the North K a r e l i a Study. (Puska, et a l . , 1979). 1972 < Survey 1 977 < Survey North K a r e l i a C o n t r o l Area North K a r e l i a C o n t r o l Area # of men % p a r t i c i p a t i o n 2228 92.6 31 59 90.2 2002 86.4 21 98 90.4 # of women % p a r t i c i p a t i o n 2307 95.2 3246 92.5 2121 90.6 31 58 91 .8 t o t a l # % p a r t i c i p a t i o n 4535 93.9 6405 91.4 41 23 88.5 6076 90.7 Of the three t a r g e t e d r i s k f a c t o r s , a f t e r f i v e years of follow-up two showed d i f f e r e n c e s between the c o n t r o l and experimental areas, while smoking d i d not. Although the percentage of i n d i v i d u a l s smoking in North K a r e l i a decreased 110 over the f i v e years of the study, a s i m i l a r decrease was found in the c o n t r o l area (see Table 5.5). In regard to the amount smoked by men, there had been a s i g n i f i c a n t l y higher mean number of c i g a r e t t e s smoked per day i n North K a r e l i a at the s t a r t of the study (North Karelia=9.9, c o n t r o l = 8 . 9 ) . By 1977 t h i s d i f f e r e n c e had disappeared, with both groups showing an average of 8.1 c i g a r e t t e s smoked per day. No d i f f e r e n c e s were found among the women smokers at e i t h e r t ime. Table 5.5 Prevalence of smoking i n North K a r e l i a and c o n t r o l area. (Puska et a l . , 1979) ( i n percent of t o t a l populat i o n ) . 1 972 < Survey 1977 < Survey North K a r e l i a C o n t r o l Area North K a r e l i a C o n t r o l Area Men Women 52.2 11.7 50.9 13.1 43.2 8.9 43.3 11.1 For the r i s k f a c t o r serum c h o l e s t e r o l , the 1972 survey found that both the men and women of North K a r e l i a had s i g n i f i c a n t l y higher v a l u e s than the c o n t r o l area. By 1977 there was no longer any major d i f f e r e n c e between the two areas. The only d i f f e r e n c e was for men between the ages of 50 and 59, where the North K a r e l i a men had a s t a t i s t i c a l l y s i g n i f i c a n t lower serum c h o l e s t e r o l l e v e l than those of the same age i n the c o n t r o l area. From Table 5.6 i t can be seen that the change i n serum c h o l e s t e r o l l e v e l s i n the two areas was due to a decrease i n serum c h o l e s t e r o l l e v e l s in the 111 North K a r e l i a p o p u l a t i o n , s i n c e the c o n t r o l l e v e l s remained e s s e n t i a l l y the same over the f i v e y e ars. Table 5.6 Serum c h o l e s t e r o l c o n c e n t r a t i o n s i n North K a r e l i a and c o n t r o l a r ea. (Puska et a l . , 1979) ( i n mmol/1iter). 1972 < Survey 1977 i Survey North K a r e l i a C o n t r o l Area North K a r e l i a C o n t r o l Area Men Women 6.96 6.86 6.73 6.70 6.70 6.68 6.75 6.60 Both s y s t o l i c and d i a s t o l i c blood pressure were recorded. The 1972 North K a r e l i a l e v e l s of s y s t o l i c blood pressure were s t a t i s t i c a l l y s i g n i f i c a n t l y higher than that of the c o n t r o l area (men: 147.3 mm Hg vs 145.0, women: 149.4 vs 144.1). By 1977 t h i s p a t t e r n had reversed i t s e l f , and the North K a r e l i a region had the lower s y s t o l i c blood pressure (men: 143.9 mm Hg vs 146.8, women: 143.5 vs 145.4). The p a t t e r n was d i f f e r e n t f o r d i a s t o l i c blood p r e s s u r e , i n 1972 the l e v e l f o r North K a r e l i a men was al r e a d y lower (90.8 mm Hg vs 92.4) and by 1977 t h i s d i f f e r e n c e had in c r e a s e d to 4.2 mm Hg (88.6 vs 92.8). In 1972 there was no d i f f e r e n c e f o r the women (North K a r e l i a : 90.7, c o n t r o l : 90.0), but by 1977 the North K a r e l i a women's ra t e was lower by a s t a t i s t i c a l l y s i g n i f i c a n t amount (86.8 vs 89.5). It seems c l e a r the i n t e r v e n t i o n used i n North K a r e l i a was able to lower some of the r i s k f a c t o r s . S p e c i f i c a l l y , r e d u c t i o n s i n blood pressure and serum c h o l e s t e r o l l e v e l s 1 12 were only observed i n North K a r e l i a . However, t h i s may not be true f o r smoking, as the c o n t r o l area a l s o showed a s u b s t a n t i a l decrease. These r e s u l t s , when combined with the hypothesis that a r e d u c t i o n i n any r i s k f a c t o r w i l l r e s u l t i n a r e d u c t i o n i n coronary heart disease l e a d to the f o l l o w i n g p r e d i c t i o n s : (1) The smoking changes i n d i c a t e there should be a r e d u c t i o n i n coronary heart d i s e a s e f o r both communities. (2) The serum c h o l e s t e r o l changes i n d i c a t e that the North K a r e l i a men should show a decrease i n coronary heart d i s e a s e , while the c o n t r o l men should not have any such decrease. Both groups of women should show a decrease, but the North K a r e l i a change should be g r e a t e r . (3) The blood pressure changes i n d i c a t e that North K a r e l i a should have a r e d u c t i o n in coronary heart d i s e a s e , but no change should be observed i n the c o n t r o l area. The changes in the m o r t a l i t y and m o r b i d i t y r a t e s i n North K a r e l i a during the years 1972-77 were repo r t e d by Salonen, Puska and Mustaniemi (1979). These changes were monitored through the establishment of a r e g i s t r y of myocardial i n f a r c t i o n , a r e g i s t r y of c e r e b r a l s t r o k e , and the c o l l e c t i o n of death r e c o r d s . U n f o r t u n a t e l y , the same c o n s i d e r a t i o n was not given to the c o n t r o l area. There the r e g i s t r i e s were i n p l a c e f o r only 5 months, from May 1, 1977 to August 31, 1977. I t i s not p o s s i b l e to compare the 1 13 mo r b i d i t y experience of the two areas over the f i v e years that the study was c a r r i e d out. Salonen et a l . thus compare the morb i d i t y experience of the two regions over a f i v e month p e r i o d . They d i v i d e the repo r t e d cases i n t o four c a t e g o r i e s : d e f i n i t e AMI, p o s s i b l e AMI, no AMI, and i n s u f f i c i e n t data. However, they present only r e s u l t s f o r the comparison of a l l AMI's. T h e i r r e s u l t s are reproduced i n Table 5.7. Table 5.7 Comparison of the age-adjusted annual i n c i d e n c e of AMI f o r i n d i v i d u a l s aged 30-64. (Salonen, et a l . , 1979) ( i n events/1000 p e o p l e ) . Sex North C o n t r o l Normal K a r e l i a Area Deviate ALL AMI's Men 11.1 12.8 2.10 standard e r r o r 1 . 1 1 . 1 Women 2.3 3.8 3.94 0.5 0.6 DEFINITE AMI'S Men (6.3 5.8 0.92 standard e r r o r 0.8 0.7 Women 1 . 1 0.8 1 . 46 0.4 0.3 I n t e r p r e t a t i o n of the r e s u l t s i s d i f f i c u l t , given the d i f f e r e n t lengths of time the r e g i s t r i e s were i n p l a c e . While the t a b l e does show a s t a t i s t i c a l l y s i g n i f i c a n t d i f f e r e n c e i n the number of a l l AMI's between North K a r e l i a and the c o n t r o l area, the lack of data for the e a r l y part of the c o n t r o l area experience makes the meaning of t h i s d i f f e r e n c e hard to i n t e r p r e t . It i s p o s s i b l e that d i f f e r e n c e s in f a m i l i a r i t y with the r e g i s t e r i e s accounts f o r 1 1 4 the d i f f e r e n c e , as the r e p o r t i n g of q u e s t i o n a b l e cases i s most l i k e l y to be highest when those who do the r e p o r t i n g are not yet f a m i l i a r with the procedure. Salonen et a l . do produce t a b l e s d e s c r i b i n g the change in the m o r t a l i t y r a t e s of the two regions over the f i v e y e a rs. T h i s i n f o r m a t i o n was d e r i v e d , presumably, from the c o l l e c t i o n of death c e r t i f i c a t e s . They present a comparison of the years 1970-71 with that of 1976-77 (see Table 5.8). Table 5.8 M o r t a l i t y r a t e s (/1000) f o r i n d i v i d u a l s aged 30-64 in 1977. (Salonen, et a l . , 1979). Me jn Wor nen North C o n t r o l North C o n t r o l K a r e l i a Area K a r e l i a Area TOTAL MORTALITY 1970-71 13.8 13.6 4.8 5.0 standard e r r o r 0.9 0.7 0.5 0.4 1976-77 11.6 11.4 3.9 3.8 standard e r r o r 0.8 0.7 0.5 0.4 d i f ference 2.2 2.2 0.9 1 .2 standard e r r o r 1 . 1 1 .0 1 . 3 0.6 CVD MORTALITY 1970-71 7.7 7.,7 2.5 2.5 standard e r r o r 0.6 0.6 0.4 0.3 1976-77 6.3 5.8 1 .7 2.5 standard e r r o r 0.6 0.5 0.3 0.3 d i f ference 1 . 4 1 .9 0.8 0.9 standard e r r o r 0.9 0.7 0.5 0.4 The r e s u l t s i n Table 5.8 do not support the hypothesis that a r e d u c t i o n in r i s k f a c t o r s w i l l l ead to a r e d u c t i o n in o v e r a l l m o r t a l i t y . Nor do they support the hypothesis that the r e d u c t i o n w i l l l ead to a r e d u c t i o n in m o r t a l i t y 1 15 a s s o c i a t e d with coronary heart d i s e a s e . While the death r a t e s have f a l l e n f o r North K a r e l i a , they have f a l l e n as much, or more, i n the c o n t r o l a r ea. T h i s c o n f l i c t s with the p r e d i c t i o n s made from the observed changes i n the r i s k f a c t o r s . As o f t e n happens when a study f a i l s to r e j e c t the n u l l h y p o t h e s i s , i n t e r p r e t a t i o n of the f i n d i n g s i s d i f f i c u l t . I t i s p o s s i b l e that 5 years or l e s s i s not long enough f o r e f f e c t s of the a l t e r e d r i s k f a c t o r s to show up, and d i f f e r e n c e s in the expected d i r e c t i o n . m i g h t be found at a l a t e r date. I t i s a l s o p o s s i b l e that while the l e v e l s were reduced, they were not reduced enough to immediately a l t e r the coronary heart d i s e a s e p r o c e s s . F i n a l l y , i t i s a l s o p o s s i b l e the r e d u c t i o n i n smoking was l a r g e enough to mask re d u c t i o n s due to the other f a c t o r s . I t i s , however, c l e a r that the e f f e c t s of r i s k f a c t o r r e d u c t i o n are not as dramatic as might be hoped. A second area of i n t e r e s t in t h i s study i s the r i s k f u n c t i o n that was developed from the data gathered. The development of the f u n c t i o n was b r i e f l y mentioned by Puska et a l . (1979). The c o e f f i c i e n t s of the equation are not d e s c r i b e d , but the authors present a t a b l e (Table IX, P. 1176) comparing the estimated coronary heart d i s e a s e r i s k scores f o r North K a r e l i a and the c o n t r o l area. E x a c t l y how t h i s f u n c t i o n was d e r i v e d i s not d e s c r i b e d , but i t appears the research team followed the m o r t a l i t y experience of the 1 1 6 cohort i n t e r v i e w e d i n 1972. These data were then analyzed by the method of Walker and Duncan to ob t a i n a r i s k f u n c t i o n (using the r i s k f a c t o r s smoking, serum c h o l e s t e r o l and s y s t o l i c blood p r e s s u r e ) . T h i s f u n c t i o n was then used to c a l c u l a t e r i s k scores f o r the 1977 data obtained i n North K a r e l i a , and r i s k scores f o r the data obtained i n 1972 and 1977 from the c o n t r o l a rea. Based on t h i s f u n c t i o n , the r i s k scores f o r both men and women in North K a r e l i a were s i g n i f i c a n t l y higher than those f o r the c o n t r o l area in 1972. By 1977 the r i s k score f o r men of North K a r e l i a was s i g n i f i c a n t l y lower than that f o r men of the c o n t r o l area, while there was no s i g n i f i c a n t d i f f e r e n c e between women's scores f o r the two areas. T h i s would seem to p r e d i c t that coronary heart d i s e a s e m o r t a l i t y r a t e s f o r the men of North K a r e l i a should be lower than the rat e f o r the c o n t r o l area, while the rat e s f o r women should be the same. T h i s change was not observed in the 5 years of the study, suggesting that the p r e d i c t i o n s based on the equation were i n e r r o r . i i . The Oslo Study The e f f e c t s of a l t e r i n g d i e t and smoking were examined by Hjermann, Velve Byre, Holme, and Leren (1981) i n a f i v e year t r i a l . A t o t a l of 1232 healthy men (aged 40-49) who, on the b a s i s of t h e i r c h o l e s t e r o l l e v e l s , smoking h a b i t s , and blood pressure l e v e l s had high coronary heart d i s e a s e 1 1 7 r i s k scores, and who met other w e l l d e f i n e d c r i t e r i a , were randomly a s s i g n e d to e i t h e r an i n t e r v e n t i o n or c o n t r o l group. The i n t e r v e n t i o n c o n s i s t e d o f : -two t a l k s given to each i n d i v i d u a l s u b j e c t : -a ten to f i f t e e n minute t a l k about the r i s k f a c t o r concept, -a 30 minute t a l k that c o n s i s t e d of both d i e t a r y a d v i c e (based on the i n d i v i d u a l ' s d i e t a r y r e c o r d , body weight, serum c h o l e s t e r o l and t r i g l y c e r i d e l e v e l s , and general background) and anti-smoking ad v i c e , -one group t a l k ( i n group s i z e s of 30-40) given to the i n t e r v e n t i o n s u b j e c t s and t h e i r wives that a l s o focused on d i e t a r y and smoking i n f o r m a t i o n . The i n t e r v e n t i o n group was re-examined every s i x months. Included in t h i s re-examination were q u e s t i o n s about the s u b j e c t ' s e a t i n g and smoking h a b i t s . Each sub j e c t was a l s o presented with a graph of h i s c h o l e s t e r o l l e v e l s s i n c e the s t a r t of the study, i n d i c a t i n g how much he had lowered h i s l e v e l . The c o n t r o l s u b j e c t s were simply r e -examined every 12 months. Over the f i v e years of the study only f i v e s u b j e c t s were l o s t to follow-up. In a d d i t i o n , 17 s u b j e c t s e i t h e r refused to a t t e n d the f i n a l examination or had moved away (3 c o n t r o l s u b j e c t s and 14 i n t e r v e n t i o n s u b j e c t s ) , r e s u l t i n g in the f i n a l examination sample being reduced by only 22 s u b j e c t s . 1 18 Comparison of the demographic and biochemical measures taken d u r i n g the f i r s t examination r e v e a l e d no s t a t i s t i c a l l y s i g n i f i c a n t d i f f e r e n c e s between the c o n t r o l and i n t e r v e n t i o n groups. The authors present graphs showing the change i n the l e v e l s of the three r i s k f a c t o r s c h o l e s t e r o l , t r i g l y c e r i d e s and smoking over the course of the study. At every examination a f t e r the i n i t i a l measurements were taken the i n t e r v e n t i o n group had the lower mean l e v e l of the r i s k f a c t o r . U n f o r t u n a t e l y , s i g n i f i c a n c e t e s t s were only done for a s e l e c t e d subgroup of s u b j e c t s , so i t i s not c l e a r i f the t o t a l i n t e r v e n t i o n group had managed to lower t h e i r r i s k f a c t o r s to a degree that a t t a i n e d s t a t i s t i c a l s i g n i f i c a n c e . Table 5.9 presents s e l e c t e d m o r b i d i t y and m o r t a l i t y r e s u l t s a f t e r completion of the study. Only the t o t a l c a r d i o v a s c u l a r death r a t e s are not s i g n i f i c a n t l y d i f f e r e n t from each other. The authors a l s o c a l c u l a t e d that changes in the i n t e r v e n t i o n group's c h o l e s t e r o l l e v e l accounted f o r 60% of the observed r e d u c t i o n in m o r b i d i t y . i i i . The M u l t i p l e Risk Factor I n t e r v e n t i o n T r i a l The M u l t i p l e Risk Factor I n t e r v e n t i o n T r i a l (MRFIT) can be d e s c r i b e d as a reseach p r o j e c t aimed at reducing a s u b j e c t ' s r i s k score, and attempting to see what happens when such a r e d u c t i o n i s achieved (Zukel, Paul and Schnaper, 1 9 8 1 ) . The r i s k score f o r an i n d i v i d u a l was c a l c u l a t e d from a l o g i s t i c equation taken from r e s u l t s of the Framingham 119 Table 5.9 M o r b i d i t y and m o r t a l i t y r e s u l t s found i n the Oslo study. (Hjermann, et a l . , 1981) (adapted from Tables V and VI) I nter\ (n = fent ion = 604) Cont (n = ( : r o l 528) Number of cases Rate /I 000 Number, of cases Rate /I 000 T o t a l coronary events 1 9 31.5 36 57. 3 T o t a l c a r d i o -v a s c u l a r events 22 36.4 39 62. 1 T o t a l coronary death 3 5.0 1 1 17. 5 T o t a l c a r d i o -v a s c u l a r death 8 13.2 1 5 23. 9 study (Neaton, et a l . , 1981). The a c t u a l equation used estimates the coronary heart d i s e a s e r i s k over s i x years, and i s : -1 (5.4) P=(1+exp{-(-11.034+.009(chol)+.046(DBP)+.029(cig))}). The t a r g e t group f o r t h i s study was the group of men i n the upper 20% of the r i s k score d i s t r i b u t i o n . Health c e n t r e s were randomly assigned to e i t h e r a s p e c i a l i n t e r v e n t i o n or usual care p r o t o c o l . For e t h i c a l reasons the usual care group r e c e i v e d some form of treatment. These s u b j e c t s were informed of t h e i r above average r i s k f o r coronary heart d i s e a s e , were co n t a c t e d once every four months, and given an annual p h y s i c a l examination. The in f o r m a t i o n obtained at these examinations was given to the subj e c t and h i s pe r s o n a l p h y s i c i a n . The s p e c i a l i n t e r v e n t i o n s u b j e c t s r e c e i v e d e s s e n t i a l l y the same 120 i n f o r m a t i o n and c o n t a c t , but were a l s o i n v i t e d to a t t e n d ten i n t e n s i v e i n t e r v e n t i o n s e s s i o n s . A f t e r these s e s s i o n s were completed the s u b j e c t s ' r i s k f a c t o r s were measured every four months, and, i n a d d i t i o n , they had the y e a r l y medical exam. The r e s u l t s a f t e r four years of the study i n d i c a t e that both groups have shown a r e d u c t i o n i n the r i s k f a c t o r s (Neaton, et a l . , 1981). The r e d u c t i o n i n the usual treatment group, while not as great as that of the s p e c i a l i n t e r v e n t i o n group, were l a r g e r than the experimenters p r e d i c t e d . T h i s means t h a t , once the t r i a l i s completed, i t w i l l be more d i f f i c u l t to f i n d s t a t i s t i c a l l y s i g n i f i c a n t d i f f e r e n c e s between the s p e c i a l i n t e r v e n t i o n group and the usual i n t e r v e n t i o n group, assuming that the i n t e r v e n t i o n a c t u a l l y reduces the coronary heart disease r i s k . At t h i s time no comparison of m o r t a l i t y r a t e s has been r e l e a s e d . O v e r a l l , t h i s s e c t i o n has r a i s e d some qu e s t i o n s about the g e n e r a l i z a b i l i t y - o f r e s u l t s found using a m u l t i p l e l o g i s t i c a n a l y s i s . There a l s o appears to be some qu e s t i o n as to whether or not i t has been p o s s i b l e to demonstrate that an immediate r e d u c t i o n in coronary heart d i s e a s e i n c i d e n c e f o l l o w s a r e d u c t i o n in the r i s k as determined from these equations. The Oslo study was able to f i n d a r e d u c t i o n by c o n c e n t r a t i n g on a s p e c i f i c group of men. The North K a r e l i a p r o j e c t , while being able to lower the average r i s k f a c t o r l e v e l s of a community as a whole, d i d not f i n d a 121 s t a t i s t i c a l l y s i g n i f i c a n t d i f f e r e n c e i n m o r t a l i t y . That the r i s k f a c t o r s can be a l t e r e d has been demonstrated by these s t u d i e s (although no c a l c u l a t i o n s of the cost and e f f o r t r e q u i r e d to achieve such a r e d u c t i o n has yet been made). But immediate b e n e f i t s i n terms of re d u c t i o n of coronary heart d i s e a s e has only been demonstrated once. There i s no i n d i c a t i o n of what long term b e n e f i t s of r i s k f a c t o r r e d u c t i o n might be. F i v e years might be a long enough follow-up p e r i o d f o r most i n t e r v e n t i o n s t u d i e s , but compared to the slow, degenerative nature of the coronary heart d i s e a s e process t h i s p e r i o d i s r e l a t i v e l y s h o r t . I t i s p o s s i b l e that those who might b e n e f i t the most from r i s k f a c t o r r e d u c t i o n s are those who are not too advanced i n the dis e a s e process. Thus a much longer follow-up p e r i o d may be r e q u i r e d before e f f e c t s of the i n t e r v e n t i o n can be f u l l y estimated. Part of the reason may l i e i n the v a r i a b i l i t y of the r i s k f a c t o r c o e f f i c i e n t s demonstrated in Table 5.1. I.f these equations are h i g h l y v a r i a b l e , and l i t t l e i s known about the long term s t a b i l i t y of a r i s k f a c t o r not subject to i n t e r v e n t i o n , then both the e s t i m a t i o n of a s u b j e c t ' s a c t u a l r i s k score and the expected b e n e f i t s from the a l t e r a t i o n of a r i s k f a c t o r c o u l d be s e r i o u s l y in e r r o r . How t h i s v a r i a b i l i t y , as w e l l as the other questions r a i s e d about the r e s u l t s of the l o g i s t i c a n alyses, might a f f e c t i n t e r v e n t i o n or preve n t i o n s t r a t e g i e s i s d i s c u s s e d in the next chapter. 123 CHAPTER 6 P r e v e n t i o n / I n t e r v e n t i o n S t r a t e g i e s and the L o g i s t i c Model The preceding chapters have: given an overview of the problems that can be encountered when attempting to c a u s a l l y r e l a t e r i s k f a c t o r s and the development of coronary heart d i s e a s e ; presented a d e s c r i p t i o n of some of the major North American p r o s p e c t i v e s t u d i e s and t h e i r i n i t i a l f i n d i n g s ; d e s c r i b e d the r a t i o n a l e behind, and the s t a t i s t i c a l development of, a m u l t i v a r i a t e model (the l o g i s t i c model) that many i n v e s t i g a t o r s see as an a p p r o p r i a t e method of a n a l y z i n g the data from these s t u d i e s ; and presented the r e s u l t s obtained from the l o g i s t i c a n a l y s i s of data c o l l e c t e d from some p r o s p e c t i v e s t u d i e s and the evidence that supports (or f a i l s to support) the f i n d i n g s obtained from t h i s type of a n a l y s i s . The c o n t r i b u t i o n s t h i s model can make toward development of s t r a t e g i e s to prevent or int e r v e n e i n the coronary heart d i s e a s e process w i l l now be d i s c u s s e d . The main concern of t h i s chapter i s with coronary heart d i s e a s e . Thus only those aspects of the l o g i s t i c model that c o u l d p l a y a r o l e in the pl a n n i n g of i n t e r v e n t i o n or pr e v e n t i o n s t r a t e g i e s w i l l be d i s c u s s e d . The model i t s e l f has been put to other worthwhile uses, for example as a h e a l t h promotion t o o l , but such p o t e n t i a l uses w i l l not be co n s i d e r e d here. 124 There are at l e a s t three i n i t i a l c o n d i t i o n s that an approach as r e l a t i v e l y complex as the l o g i s t i c model would have to meet before i t c o u l d be c o n s i d e r e d a worthwhile p l a n n i n g t o o l . F i r s t of a l l , the model has to have some advantages over other, l e s s complex, approaches. Indeed, the l o g i s t i c model has many advantages over i t s co m p e t i t o r s . The t h e o r e t i c a l j u s t i f i c a t i o n of the model i s i n t u i t i v e l y reasonable, much more so than that i m p l i e d by the u n i v a r i a t e a n a l y s e s of r i s k f a c t o r s (one f a c t o r c a u s a l models), or the m u l t i p l e l i n e a r r e g r e s s i o n models used by some r e s e a r c h e r s (e.g. Burch, 1980; V l i e t s t r a , et a l . , 1980). Secondly, the model should appeal t o p o t e n t i a l u s e r s . T h i s model has the a p p e a l i n g aspect that i t s u p p l i e s a r i s k score f o r each i n d i v i d u a l that i s between 0.0 and 1.0. T h i s i s c o n s i s t e n t with the f a c t t h a t few of the coronary heart d i s e a s e deaths w i l l occur i n i n d i v i d u a l s with a r i s k score near 1.0. F i n a l l y , the l o g i s t i c model has not produced any r e s u l t s t h a t c o n t r a d i c t the main body of knowledge about the e t i o l o g y of coronary heart d i s e a s e . To date i t has presented r e s e a r c h e r s with r e s u l t s c o n s i s t e n t with t h e i r t h e o r e t i c a l and experimental p e r s p e c t i v e . Thus the model c o u l d be e a s i l y i n c o r p o r a t e d i n t o medical p r a c t i c e . As a plan n i n g t o o l , the l o g i s t i c model presents, at l e a s t s u p e r f i c i a l l y , a s t r a i g h t f o r w a r d p r e v e n t i o n or i n t e r v e n t i o n p l a n . Those r i s k f a c t o r s that have been shown to be c o n s i s t e n t l y s i g n i f i c a n t i n a v a r i e t y of s t u d i e s are 125 accepted as the prime r i s k f a c t o r s f o r coronary heart d i s e a s e . A p r e v e n t i o n s t r a t e g y would simply be aimed at keeping the i n d i v i d u a l ' s l e v e l s of these r i s k f a c t o r s as low as p o s s i b l e . If a l l the r i s k f a c t o r s had been i d e n t i f i e d , and i f no i n d i v i d u a l ever reached a high l e v e l on any of these r i s k f a c t o r s , then the r i s k of coronary heart d i s e a s e f o r the whole p o p u l a t i o n would presumably be very low, and the i n c i d e n c e of coronary heart d i s e a s e should decrease. An i n t e r v e n t i o n s t r a t e g y would s t a r t with the c r e a t i o n of a r i s k score f o r an i n d i v i d u a l (using an a p p r o p r i a t e l o g i s t i c e q u a t i o n ) , and those i n d i v i d u a l s who were at high r i s k would be subjected to f u r t h e r i n t e r v e n t i o n ( s ) aimed at reducing t h e i r r i s k . The equation i t s e l f c o u l d guide the form of i n t e r v e n t i o n . Decrease in t o t a l r i s k c o u l d be c a l c u l a t e d for v a r i o u s types of r i s k f a c t o r r e d u c t i o n p l a n s , and the s t r a t e g y t hat p r o v i d e d the g r e a t e s t o v e r a l l r e d u c t i o n c o u l d be f o l l o w e d . A b e n e f i t of t h i s approach i s that some i n d i v i d u a l s might show r i s k p a t t e r n s that normally would not come to the a t t e n t i o n of an examining p h y s i c i a n . For example, an i n d i v i d u a l might not be o u t s i d e the normal range f o r blood pressure and c h o l e s t e r o l l e v e l s , but only higher than the average. However, i f t h i s i n d i v i d u a l a l s o smoked, the l o g i s t i c equation might i n d i c a t e that together these f a c t o r s produce a s u b s t a n t i a l r i s k f o r coronary heart d i s e a s e , and a p p r o p r i a t e a c t i o n c o u l d be taken ( P o o l i n g P r o j e c t Research Group, 1978). 126 There are, however, c e r t a i n problems that would be encountered i f attempts were made to develop a plan along these l i n e s . These problems, u n t i l they are overcome, w i l l l i m i t the p o t e n t i a l u s e f u l n e s s of i n t e r v e n t i o n and prev e n t i o n s t r a t e g i e s based on t h i s model. I P o t e n t i a l Problem Areas with the A p p l i c a t i o n of the L o g i s t i c Model A S e l e c t i o n of the App r o p r i a t e Equation If a l o g i s t i c equation were used as a method to screen for high r i s k s u b j e c t s , or used to monitor the r i s k l e v e l of a p o p u l a t i o n , then a s p e c i f i c equation must be s e l e c t e d . From Table 5.1 i t i s c l e a r that the s e l e c t i o n of the a p p r o p r i a t e equation would not be simple. Some equations can immediately be e l i m i n a t e d . For example, the use of an equation d e r i v e d from a c u l t u r a l group d i f f e r e n t from the ta r g e t p o p u l a t i o n i s l i k e l y i n a p p r o p r i a t e . While d i f f e r e n t c u l t u r a l groups do not appear to present a d i f f e r e n t r i s k p r o f i l e 1 , the f i n d i n g s of Keys et a l . (1972) (that the European equation u n d e r p r e d i c t e d the American r i s k by about 1/3) suggests use of such an equation could l e a d to s e r i o u s 1 In Table 5.1 equations 9-11 and 27-28 were d e r i v e d from European p o p u l a t i o n s , and they present the same general p a t t e r n of s i g n i f i c a n t and n o n s i g n i f i c a n t c o e f f i c i e n t s as the American s t u d i e s . 1 27 e r r o r s i n p r e d i c t i o n . S i m i l a r c o n c l u s i o n s were reached by S c h r o l l and Larsen (1981). They a p p l i e d a l o g i s t i c equation d e r i v e d from the Framingham sample (equation 16 i n Table 5.1) to the 10 year coronary heart d i s e a s e r a t e of a p o p u l a t i o n of 40-59 year o l d Danish men. They found the Framingham equation o v e r p r e d i c t e d the d i s e a s e r a t e by more than 6%. T h i s was g r e a t e r than any of the e r r o r s of p r e d i c t i o n of three equations d e r i v e d from European data. The i n v e s t i g a t o r s concluded that the main d i f f e r e n c e between equations d e r i v e d from d i f f e r e n t c u l t u r e s was p r i m a r i l y in the constant term, and thus the use of such an equation would be of q u e s t i o n a b l e v a l u e . S c h r o l l and Larsen d i d , however, conclude that equations are l i k e l y t r a n s f e r a b l e w i t h i n a c u l t u r e . T h i s c o n c l u s i o n i s s i m i l a r to the c o n c l u s i o n of the P o o l i n g P r o j e c t Research Group (1978). However, the f i n d i n g s of Keys et a l . (1972) suggest some q u a l i f i c a t i o n s should be made. I t would appear that good p r e d i c t i o n should only be expected when i n t e r e s t l i e s with d i a g n o s t i c c r i t e r i a that are the same as the d i a g n o s t i c c r i t e r i a used i n the study from which the equation was d e r i v e d . T h i s i m p l i e s that once a d e c i s i o n i s made as to the nature of the t a r g e t e d problem (e.g. a l l forms of coronary heart d i s e a s e , only AMIs, etc.) an equation used to p r e d i c t t h i s outcome must be found. The Framingham study has produced a number of equations f o r d i f f e r e n t outcomes (see McGee, 1973). If the Framingham 128 d i a g n o s t i c c r i t e r i a are s u i t a b l e , an a p p r o p r i a t e equation c o u l d l i k e l y be found (assuming i n t e r e s t a l s o l a y with white ma1e s) . Another f a c t o r that must be c o n s i d e r e d i n the s e l e c t i o n of the equation i s which v a r i a b l e s to i n c l u d e , and which to ignore. The temptation might be to i n c l u d e only those v a r i a b l e s that entered s i g n i f i c a n t l y i n t o the equation ( p r o v i d i n g the measures are e a s i l y o b t a i n e d ) . While g o o d n e s s - o f - f i t r e s u l t s might imply that l i t t l e would be gained from the a d d i t i o n of other v a r i a b l e s , the a n a l y s i s presented i n Chapter 5 (see Table 5.3 and F i g u r e 5.1) i n d i c a t e s that the r i s k g r a d i e n t may be a more important s e l e c t i o n c r i t e r i o n . I t would be a d v i s a b l e to examine the r i s k g r a d i e n t of the p o t e n t i a l equations before making a s e l e c t i o n . T h i s c o n s i d e r a t i o n i s s i m i l a r to examining the s e n s i t i v i t y and s p e c i f i c i t y of a s c r e e n i n g t e s t . Since i t i s l i k e l y that only high r i s k i n d i v i d u a l s would be t a r g e t e d for i n t e r v e n t i o n , the equation with the steepest r i s k g r a d i e n t would have the highest degree of s e n s i t i v i t y and s p e c i f i c i t y . Information on the r i s k g r a d i e n t i s not u s u a l l y p u b l i s h e d , and thus i t would be necessary to o b t a i n these data from the authors of the o r i g i n a l s t u d i e s . If s u f f i c i e n t p r e c a u t i o n s are taken i t should be p o s s i b l e to f i n d an equation a p p r o p r i a t e f o r a t a r g e t e d p o p u l a t i o n . The equation might not f i t e x a c t l y , but r e s e a r c h e r s b e l i e v e that the f i t should be good enough to 1 29 make p r e d i c t i o n s meaningful. The g r e a t e s t l i m i t a t i o n i n the s e l e c t i o n of an equation i s l i k e l y to be i n the areas of d i a g n o s t i c c r i t e r i a , and in the p o p u l a t i o n s that have been su b j e c t e d to thorough a n a l y s i s (mainly white men). B S e l e c t i o n of the Appropriate I n t e r v e n t i o n S t r a t e g y The u s e f u l n e s s of a scre e n i n g program should be eva l u a t e d in two areas: the p r e d i c t i v e a b i l i t y of the scr e e n i n g t o o l , and the a v a i l a b i l i t y of a s u i t a b l e follow-up treatment plan f o r high r i s k i n d i v i d u a l s (Friedman, 1974). I t i s the l a t t e r requirement that c o u l d prove to be the g r e a t e s t stumbling block in the c r e a t i o n of an i n t e r v e n t i o n s t r a t e g y . The l o g i s t i c model suggests a s t r a i g h t f o r w a r d i n t e r v e n t i o n program: reduce (and keep down) l e v e l s of those r i s k f a c t o r s that l e a d to high r i s k . The problem with t h i s approach i s that the c l a i m can be made that there i s l i t t l e evidence that these reduced r i s k f a c t o r l e v e l s w i l l cause a r e d u c t i o n i n the coronary heart d i s e a s e i n c i d e n c e r a t e . The e x p e c t a t i o n of a r e d u c t i o n i n coronary heart d i s e a s e r i s k as a r e s u l t of a r e d u c t i o n in the l e v e l of a r i s k f a c t o r assumes a c a u s a l r e l a t i o n s h i p between the r i s k f a c t o r and the development of coronary heart d i s e a s e . In Chapter 2 i t was poin t e d out that a v a r i a b l e need not be c a u s a l l y r e l a t e d to the outcome f o r i t to be viewed as a r i s k f a c t o r . Of the three requirements f o r the establishment of cause d e s c r i b e d in.Chapter 2 (see M o s t e l l e r 130 and Tukey, 1977), two are s u f f i c i e n t l y met by the l o g i s t i c f u n c t i o n . The r e s u l t s d e s c r i b e d in Table 5.1 show a l a r g e degree of c o n s i s t e n c y (the f i r s t requirement). The model a l s o i s c o n s i s t e n t with a v a r i e t y of b i o l o g i c a l and bio c h e m i c a l t h e o r i e s on the development of coronary heart d i s e a s e (the t h i r d requirement). The second requirement i s that of responsiveness, and i t appears that t h i s has not yet been r e l i a b l y demonstrated. The experiences of the Oslo study, the North K a r e l i a p r o j e c t and MRFIT i n d i c a t e that i t i s p o s s i b l e to interve n e and reduce the average l e v e l s of r i s k f a c t o r s i n a p o p u l a t i o n . The Oslo study was able to demonstrate a r e d u c t i o n in the l e v e l s of the r i s k f a c t o r s i t examined, and a decrease in the coronary heart d i s e a s e m o r b i d i t y i n the experimental group. However, d e s p i t e a r e d u c t i o n i n r i s k f a c t o r s , North K a r e l i a , when compared to the c o n t r o l r e g i o n , d i d not show any s u b s t a n t i a l improvement in coronary heart d i s e a s e m o r t a l i t y . In a d d i t i o n , the re d u c t i o n in the l e v e l s of the r i s k f a c t o r s i n the usual treatment group of the MRFIT p r o j e c t w i l l reduce the s t a t i s t i c a l power of the r e s e a r c h e r s ' t e s t s . T h i s suggests that even i f the r i s k f a c t o r s are c a u s a l l y r e l a t e d the experimenters may have d i f f i c u l t y f i n d i n g a s t a t i s t i c a l l y s i g n i f i c a n t improvement i n t h e i r s p e c i a l i n t e r v e n t i o n g r o u p 1 . 1 A recent newspaper a r t i c l e (The Vancouver Sun, Sept. 17, 1982) r e p o r t s that the MRFIT p r o j e c t was not able to f i n d a s t a t i s t i c a l d i f f e r e n c e between the two groups. 131 T h i s p o s i t i o n has been taken by Corday and Corday (1975). They s t a t e that they b e l i e v e there i s v a l i d i t y i n the r i s k f a c t o r h y p o t h e s i s . But they a l s o c l a i m that they agree with a 1971 report of the N a t i o n a l Heart and Lung I n s t i t u t e . The authors c i t e t h i s r e p o r t ' s q u e s t i o n i n g whether enough evidence e x i s t s to warrent a c o n c l u s i o n that r e d u c t i o n of a r i s k f a c t o r w i l l stop or reverse the p r o g r e s s i o n of coronary heart d i s e a s e . Corday and Corday's a p p r a i s a l of the s i t u a t i o n r egarding r i s k f a c t o r s and development of coronary heart d i s e a s e i s : Most i n v e s t i g a t o r s agree that the r i s k f a c t o r hypothesis has provided v a l u a b l e i n f o r m a t i o n and to be s u c c e s s f u l must be a p p l i e d w e l l before the c l i n i c a l evidence becomes manifest. But once a r t e r i o s c l e r o s i s i s c l i n i c a l l y e v i d ent, there i s l i t t l e evidence the l e s i o n can be re v e r s e d . (Corday and Corday, 1975; P. 334) They then continue with a b r i e f d e s c r i p t i o n of some a l t e r n a t i v e models of coronary heart d i s e a s e . It would seem, then, i f an i n t e r v e n t i o n based on r i s k f a c t o r r e d u c t i o n was s t a r t e d today (1982), i t would be p a r t i a l l y conducted on f a i t h . It i s d i f f i c u l t to see how any harm c o u l d come to an i n d i v i d u a l who attempted to a l t e r h i s or her l i f e s t y l e so as to a t t a i n a r e d u c t i o n i n any of the r i s k f a c t o r s of serum c h o l e s t e r o l , blood pressure or smoking (assuming some degree of common sense on the part of that i n d i v i d u a l ) . However, only one study has been able to f i n d a s t a t i s t i c a l l y s i g n i f i c a n t r e d u c t i o n i n m o r b i d i t y , and thus there i s no guarantee that such a procedure would 1 32 r e s u l t in a r e a l r e d u c t i o n i n coronary heart d i s e a s e i n c i d e n c e . At present such a procedure would r e s t on the a v a i l a b l e evidence which s t r o n g l y i n d i c a t e s that such a r e d u c t i o n would take p l a c e . If the i n t e r v e n t i o n i n v o l v e d some form of treatment that i t s e l f c a r r i e d some r i s k f o r the i n d i v i d u a l ( f o r example, the lowering of blood pressure through the use of drugs) then the a d v i s a b i l i t y of such a procedure c o u l d be questioned. I t i s p o s s i b l e that l i t t l e r e a l gain i n p r o t e c t i o n from coronary heart d i s e a s e might r e s u l t , while the r i s k of other medical problems c o u l d be s u b s t a n t i a l l y i n c r e a s e d . If t h i s kind of i n t e r v e n t i o n i s needed i t might be b e t t e r to wait u n t i l a c l e a r e r p i c t u r e of the expected b e n e f i t s of r i s k f a c t o r r e d u c t i o n can be c a l c u l a t e d . Only then c o u l d i t be determined i f the r i s k s i n v o l v e d in i n t e r v e n t i o n are l e s s than the b e n e f i t s of i ntervent i on. C S e l e c t i o n of the Appropriate Prevention S t r a t e g y U n l i k e the i n t e r v e n t i o n s t r a t e g i e s , the prospects f o r p r e v e n t i o n seem somewhat b r i g h t e r . The l o g i s t i c model suggests that i f i t i s p o s s i b l e to a v o i d the l e v e l s of c h o l e s t e r o l , blood p r e s s u r e , e t c . that l e a d to high r i s k of coronary heart d i s e a s e , then the process of developing coronary heart d i s e a s e should be avoided. T h i s would be true even i f the r i s k f a c t o r s are a c t u a l l y e a r l y outcomes of the coronary heart disease process i t s e l f , r a t h e r than 1 33 c o n d i t i o n s that lead to the development of the d i s e a s e . By a v o i d i n g the development of high l e v e l s of r i s k f a c t o r s one would be somehow a l t e r i n g the d i s e a s e process, even i f the u n d e r l y i n g model (that r i s k f a c t o r s are c a u s a l l y r e l a t e d ) was wrong. There e x i s t s a l a r g e amount of l i t e r a t u r e on d i e t s , l i f e s t y l e s , e t c . , that should a v o i d the development of high l e v e l s of some r i s k f a c t o r s . C h o l e s t e r o l l e v e l s can be c o n t r o l l e d through d i e t p r a c t i c e s , and smoking i s a b e h a v i o r a l c h o i c e . The maintenance of low blood pressure l e v e l s may be more d i f f i c u l t . High blood pressure does not present any symptoms to the i n d i v i d u a l . In a d d i t i o n , u n l i k e serum c h o l e s t e r o l l e v e l s (which have been a t t r i b u t e d to c e r t a i n d i e t a r y p r a c t i c e s ) , high blood pressure does not appear to be caused by e a s i l y i d e n t i f i a b l e or a l t e r a b l e f a c t o r s . Thus without s p e c i a l emphasis the i n d i v i d u a l might not focus on the maintenance of low blood p r e s s u r e . However, should a p r e v e n t i o n program be developed that uses these s t r a t e g i e s to reduce r i s k f a c t o r s , then the experience of North K a r e l i a and MRFIT suggests the l e v e l s c o u l d be lowered. An experimental demonstration of the success of the p r e v e n t i o n program would be d i f f i c u l t . It would r e q u i r e an experimental group who, f o r most of t h e i r l i v e s , were subjected to measures aimed at keeping t h e i r r i s k f a c t o r l e v e l s w i t h i n c e r t a i n l i m i t s . If t h i s group showed a 1 34 s u b s t a n t i a l l y lower rate of coronary heart d i s e a s e than a c o n t r o l group, then the p r o j e c t would be shown to be a success. However, the apparent long time course for the development of coronary d i s e a s e s suggests such a p r o j e c t would have to s t a r t e a r l y in l i f e , and continue f o r most of the s u b j e c t ' s l i f e . I t seems u n l i k e l y that a 60+ year p r o j e c t w i l l be undertaken. An a l t e r n a t i v e s t r a t e g y might be to attempt to show that i n d i v i d u a l ' s who have s e l f -s e l e c t e d a l i f e s t y l e that should r e s u l t i n the maintenance of lower r i s k f a c t o r l e v e l s have reduced coronary heart d i s e a s e r a t e s . Such a q u a s i - e x p e r i m e n t a l design would not demonstrate that a c a u s a l r e l a t i o n s h i p e x i s t e d between the r i s k f a c t o r s and coronary heart d i s e a s e , but i t would i n d i c a t e i f the d e c i s i o n to l e a d a l i f e s t y l e that r e s u l t s in lower r i s k f a c t o r s a l s o r e s u l t s i n lower r a t e s of coronary heart d i s e a s e . Since any p r e v e n t i o n program i s l i k e l y to succeed or f a i l on the b a s i s of the i n d i v i d u a l ' s p e r s o n a l d e c i s i o n to f o l l o w such a program, such a design would allow for an e s t i m a t i o n of the expected b e n e f i t s that would be o b t ained from the d e c i s i o n to a l t e r one's l i f e s t y l e . Once these b e n e f i t s were known the "only" problem l e f t to s o l v e would be the development of an e f f e c t i v e method to convince the i n d i v i d u a l to change h i s or her l i f e s t y l e to one that has been shown to be e f f e c t i v e . 1 35 D S e l e c t i o n of the Target L e v e l s The s e l e c t i o n of i d e a l l e v e l s f o r r i s k f a c t o r s that an i n t e r v e n t i o n or pr e v e n t i o n program c o u l d aim f o r cannot be done on the b a s i s of r e s u l t s o b tained from a l o g i s t i c a n a l y s i s . From the p e r s p e c t i v e of t h i s model, the lower the l e v e l s the b e t t e r (the lowest p o s s i b l e r i s k score would be obtained when a l l the r i s k f a c t o r s were 0.0). T h i s , of course, i s a c h a r a c t e r i s t i c of the model, and would never be taken s e r i o u s l y by those using the r e s u l t s . The t a r g e t l e v e l s would have to be obtained from other sources, and the medical l i t e r a t u r e c o n t a i n s suggestions as to what the normal l i m i t s should be. Note though, that the l o g i s t i c model suggests the lower end of the range i s to be p r e f e r r e d , and t h i s may not n e c e s s a r i l y be the m e d i c a l l y p r e f e r e d v a l u e . Within a range of normal values the optimum value i s l i k e l y somewhere i n the middle of the range, and t h i s f a c t o r should a l s o be taken i n t o c o n s i d e r a t i o n when determining the t a r g e t l e v e l s . The smoothing nature of the l o g i s t i c curve may have al r e a d y c r e a t e d some m i s l e a d i n g r e s u l t s . Anderson (1978) po i n t e d out that while the l o g i s t i c equation d e r i v e d from the Framingham study suggests that a d i a s t o l i c blood pressure of 70 mm Hg should produce l e s s coronary heart d i s e a s e than 80 mm Hg, which should produce l e s s than 90 mm Hg, the a c t u a l data the equation was d e r i v e d from suggests the o p p o s i t e . If the data are examined, i t turns out that 1 36 those i n d i v i d u a l s who were recorded as having a d i a s t o l i c blood pressure of 90 mm Hg had a lower coronary heart d i s e a s e r a t e than those with a l e v e l of 80 mm Hg, who i n turn were lower than those with a l e v e l of 70 mm Hg. These d i f f e r e n c e s were not s t a t i s t i c a l l y s i g n i f i c a n t , but Anderson makes the p o i n t that such f i n d i n g s deserve to be examined f u r t h e r . For example, i t might be that t h i s p a t t e r n i s simply a r e s u l t of sampling v a r i a t i o n w i t h i n the Framingham sample. If t h i s i s the case, then the poor performance of Framingham d i a s t o l i c blood pressure when compared to s y s t o l i c blood pressure as a p r e d i c t o r of coronary heart d i s e a s e might be due to t h i s v a r i a t i o n . Another p o s s i b i l i t y i s that t h i s f i n d i n g might r e f l e c t the a c t u a l s t a t e of a f f a i r s , and thus an i n t e r v e n t i o n that recommended the decrease of d i a s t o l i c blood pressure below 90 mm Hg would be i n c r e a s i n g the r e a l r i s k f o r coronary heart d i s e a s e . The s e l e c t i o n of i d e a l l e v e l s of r i s k f a c t o r s would not be an easy task. The s e l e c t i o n would have to be made with an eye toward r e d u c t i o n of coronary heart d i s e a s e r i s k , but at the same time being sure that i t d i d not c r e a t e other h e a l t h problems. The l o g i s t i c analyses presented above provide no information on t h i s matter, and may, in at l e a s t one case, have obscured an important r e l a t i o n s h i p between coronary heart d i s e a s e and d i a s t o l i c blood pressure l e v e l . 1 37 11 P o t e n t i a l D i f f i c u l t i e s t h a t Might A r i s e with the  I n t r o d u c t i o n of Programs The g r e a t e s t advantage of any program based on a l o g i s t i c model i s that i t would provide a method of i d e n t i f y i n g i n d i v i d u a l s who may be at r i s k f o r coronary heart d i s e a s e , even though they may not have any c l i n i c a l l y s i g n i f i c a n t symptoms or r i s k f a c t o r l e v e l s . A problem with t h i s i d e n t i f i c a t i o n process i s that i t w i l l y i e l d a l a r g e number of f a l s e p o s i t i v e s , that i s , i n d i v i d u a l s who are at the high end of the r i s k spectrum and yet do not (at l e a s t over the course of the study), develop coronary heart d i s e a s e . T h i s p o i n t i s o f t e n not c o n s i d e r e d by those who advocate the i n t r o d u c t i o n of r i s k f a c t o r r e d u c t i o n programs. Some of these p o t e n t i a l problems can be demonstrated. One of the s t r o n g e s t c a l l s i n recent years f o r the i n t r o d u c t i o n of some s o r t of i n t e r v e n t i o n program has come from the P o o l i n g P r o j e c t Research Group (1978). In t h e i r report they concluded that the r e l a t i o n s h i p between r i s k f a c t o r s and premature a t h e r o s c l e r o t i c d i s e a s e i s "almost c e r t a i n l y c a u s e - a n d - e f f e c t , i . e . e t i o l o g i c , in nature" (p. 266). T h i s c o n c l u s i o n was reached by c o n s i d e r i n g a l i s t of c o n d i t i o n s necessary fo r demonstration of c a u s a l i t y (which d i d not i n c l u d e a c o n d i t i o n e q u i v a l e n t to M o s t e l l e r and Tukey's (1977) c r i t e r i o n of r e s p o n s i v e n e s s ) . They go on to s t a t e : 138 i t should be emphasized that the r e s u l t s of the P o o l i n g P r o j e c t p r e d i c t i v e a n a l y s e s 1 demonstrate i t s f i n d i n g s to be g e n e r a l i z a b l e f o r p r e d i c t i v e purposes both q u a l i t a t i v e l y and q u a n t i t a t i v e l y at l e a s t f o r middle-aged white men in the U.S.A. ... T h e i r widespread use should be c o n s i s t e n t l y and v i g o r o u s l y encouraged. ( P o o l i n g P r o j e c t Research Group, 1978, p. 266 ( t h e i r i t a l i c s ) ) T h e i r r e p o r t c o n t a i n s a l o g i s t i c equation developed from the v a r i a b l e s age, d i a s t o l i c blood p r e s s u r e , serum c h o l e s t e r o l , smoking, and r e l a t i v e weight (equation 26, Table 5.1). They a l s o present a t a b l e ( t h e i r Table 9E) that compares the expected event rate c a l c u l a t e d from t h e i r equation and the observed event r a t e obtained from the data. The equation was d e r i v e d by the method of T r u e t t et a l . (1967), using a sample of 6854 men between the ages of 40 and 59. T h e i r r e s u l t s are reproduced i n Table 6.1. Table 6.1 P r e d i c t e d and observed number of coronary events as d e s c r i b e d by the P o o l i n g P r o j e c t Research Group. (1978: adaption of Table 9E). Q u i n t i l e of Expected Risk Expected # of Events Observed # of Events V 264.6 245 IV 145.2 1 68 III 100.5 1 1 2 II 70.7 68 I 41 .0 27 A g o o d n e s s - o f - f i t t e s t r e s u l t s in a X 2 value of 11.2 (higher than the values found in Table 5.3). In a d d i t i o n , 1 Both l o g i s t i c and l i f e t a b l e methods were used in t h e i r s e c t i o n on p r e d i c t i v e a n a l y s e s . 139 i t should be noted that t h i s equation was d e r i v e d by a method that H a l p e r i n et a l . (1971) suggested c o u l d l e a d to estimated c o e f f i c i e n t s that are i n e r r o r (see Chapter 4). Even i f these p o t e n t i a l problems are ignored, t h i s equation would produce a l a r g e number of f a l s e p o s i t i v e s i f used by an i n t e r v e n t i o n program. I t seems reasonable to assume that a program which used t h i s equation might co n c e n t r a t e on those s u b j e c t s i n the highest q u i n t i l e of r i s k . Table 6.2 uses the r e s u l t s as presented by the P o o l i n g P r o j e c t Research Group (1978), and c o n t a i n s a standard two by two t a b l e comparing the expected and observed "cases" that such a program would have to contend with. Table 6.2 S e l e c t i o n of high r i s k s u b j e c t s f o r f u r t h e r i n t e r v e n t i o n , and the observed number of coronary events as reported by the P o o l i n g P r o j e c t Research Group (1978). Intervent ion Status Observed CHD Observed no CHD T o t a l Intervene 245 1 1 26 1 371 ( Q u i n t i l e V) Not - Intervene 375. 51 08 5483 (Quint. I-IV) T o t a l 620 6234 6854 s e n s i t i v i t y = 245/620 = .395 s p e c i f i c i t y = 5108/6234 = .819 true p o s i t i v e r a t e = 245/1371 = .179 If such an i n t e r v e n t i o n was used, then Table 6.2 i n d i c a t e s that i t would only be a p p l i e d to 39.5% of the a c t u a l number of cases developing coronary heart d i s e a s e . 1 40 In a d d i t i o n , 82.1% of the i n d i v i d u a l s given the i n t e r v e n t i o n would not have developed coronary heart d i s e a s e (the equation p r e d i c t s r i s k over 8.6 y e a r s ) . The reason f o r the r e l a t i v e l y poor performance when expected b e n e f i t s of the program are d e s c r i b e d i n t h i s manner i s that the p r o b a b i l i t y of an i n d i v i d u a l developing coronary heart d i s e a s e i s low. The average p r o b a b i l i t y of an i n d i v i d u a l in a q u i n t i l e can be estimated by d i v i d i n g the expected number of events in the q u i n t i l e by the number of i n d i v i d u a l s i n the q u i n t i l e . The average p r o b a b i l t y f o r an i n d i v i d u a l in q u i n t i l e V i s 264.6/1371, or .193. Thus, even i n the high r i s k group most of the i n d i v i d u a l s would not be expected to develop coronary heart d i s e a s e . The performance of the P o o l i n g P r o j e c t Research Group's equation i s not a t y p i c a l . S c h r o l l and Larsen (1981) estimated t h a t an i n t e r v e n t i o n program based on t h e i r r e s u l t s would only prevent one case f o r every s i x i n t e r v e n t i o n s . T h i s success r a t e a l s o assumes that the i n t e r v e n t i o n moved a l l high r i s k s u b j e c t s to the lowest r i s k c a t e g o r y . Thus any i n t e r v e n t i o n program would l i k e l y have a high f a l s e p o s i t i v e rate (the complement of the true p o s i t i v e r a t e ) . In a d d i t i o n , unless g e n e r a t i o n of even more f a l s e p o s i t i v e s was a c c e p t a b l e , the i n t e r v e n t i o n would miss the m a j o r i t y of the coronary heart d i s e a s e cases. Only i f an equation with a much steeper r i s k g r a d i e n t was found could 141 some of these problems be overcome. The magnitude of the coronary heart d i s e a s e problem in North America may make a high f a l s e p o s i t i v e r a t e a c c e p t a b l e , but the e x i s t e n c e of these p o t e n t i a l problem areas should not be ignored. C a r e f u l s e l e c t i o n of the l o g i s t i c equation and the d i a g n o s t i c category to be concentrated on c o u l d reduce, but not e l i m i n a t e , these problems. In Chapter 1 the c r i t e r i a f o r a s u c c e s s f u l primary p r e v e n t i o n program, as d e s c r i b e d by B i o r c k (1975), were o u t l i n e d . Using h i s o u t l i n e , from the d i s c u s s i o n i n t h i s and the p r e v i o u s Chapters i t would seem reasonable to conclude t h a t : 1. A s t r o n g l y suspected d i s e a s e mechanism e x i s t s , and there i s a l a r g e amount of evidence to suggest (and one study has demonstrated) that r i s k f a c t o r r e d u c t i o n i s a method of i n t e r f e r i n g with t h i s mechanism. 2. The l o g i s t i c method, once understood, i s a simple method of i d e n t i f y i n g " s u s c e p t i b l e s " . 3. S e n s i t i v i t y ( s e l e c t i v i t y ) and s p e c i f i c i t y e s t i m a t i o n s can be made. Whether or not the values are acceptable would depend on the cost of the i n t e r v e n t i o n . It should be noted that these c a l c u l a t i o n s are most l i k e l y to be made by comparing the p r e d i c t e d r e s u l t s with o b s e r v a t i o n s made on the sample from which the equation was d e r i v e d . T h i s means the s e n s i t i v i t y of the t e s t i s l i k e l y overestimated to an unknown degree. 142 4. The r e s u l t s from the North K a r e l i a p r o j e c t suggest t h a t , assuming there i s a c a u s a l r e l a t i o n s h i p between r i s k f a c t o r s and coronary heart d i s e a s e , the time l a g between the i n t r o d u c t i o n of the i n t e r v e n t i o n and i t s e f f e c t on coronary heart d i s e a s e i s more than 5 years. Not u n t i l the r e s u l t s of s t u d i e s that are p r e s e n t l y i n progress are known, w i l l i t be p o s s i b l e to estimate the time between i n t e r v e n t i o n and a c t u a l r i s k r e d u c t i o n . 5. Attempts to get s u b j e c t s to change t h e i r r i s k f a c t o r scores have been s u c c e s s f u l , at l e a s t i n the short run. Th i s suggests that e i t h e r an i n t e r v e n t i o n or p r e v e n t i o n program i s c l o s e to being f e a s i b l e . C e r t a i n a d d i t i o n a l r e s e a r c h questions have to be answered before a program would become t r u l y f e a s i b l e , and in f o r m a t i o n about c e r t a i n other q u e s t i o n s c o u l d lead to a much more e f f i c i e n t program. I l l A d d i t i o n a l Research Needs of the L o g i s t i c Model Before the l o g i s t i c model w i l l become a v i a b l e model on which to base an i n t e r v e n t i o n or pr e v e n t i o n program, some a d d i t i o n a l r e s e a r c h areas should be examined. One area of p a r t i c u l a r i n t e r e s t i s the p o t e n t i a l i n t e r a c t i o n of v a r i a b l e s . Most of the equations in Table 5.1 do not co n t a i n i n t e r a c t i o n terms. When an i n t e r a c t i o n i s i n c l u d e d , i t was most o f t e n s e l e c t e d because the l o g i s t i c model f a i l e d to a t t a i n a s a t i s f a c t o r y f i t of the data. Of a l l the s t u d i e s represented in Table 5.1, only Kleinbaum et 143 a l . (1971) s y s t e m a t i c a l l y examined p o s s i b l e i n t e r r e l a t i o n s h i p s between r i s k f a c t o r s . T h i s was done by i n c l u d i n g a s e r i e s of cro s s - p r o d u c t i n t e r a c t i o n terms i n the exponent of the l o g i s t i c model. The i n c l u s i o n of the same v a r i a b l e s i n a v a r i e t y of d i f f e r e n t i n t e r a c t i o n terms would make r e s u l t s of t h i s type of a n a l y s i s very s e n s i t i v e to sampling v a r i a t i o n ( M o s t e l l e r and Tukey, 1977), and, thus, the g e n e r a l i z a b i l i t y of these r e s u l t s c o u l d be questioned. Models that might be used to d e s c r i b e the form of an i n t e r a c t i o n between two r i s k f a c t o r s have been d e s c r i b e d . (Morse, 1978). Morse's approach i n v o l v e s the d e r i v a t i o n of a model that d e s c r i b e s the combined e f f e c t of two v a r i a b l e s , assuming no i n t e r a c t i o n between the v a r i a b l e s . The d e r i v e d model's p r e d i c t i o n s are then compared to the observed data. From the d i s c r e p a n c i e s that e x i s t between the p r e d i c t i o n s and o b s e r v a t i o n s , estimates can be made about the form of the synergism that might e x i s t between the two v a r i a b l e s . T h i s approach i s c o n c e p t u a l l y s i m i l a r to a method of studying i n t e r a c t i o n s suggested by K e r l i n g e r and Pedhazur (1973). They suggest comparing the o v e r a l l r e g r e s s i o n c o e f f i c i e n t obtained from a model that does not c o n t a i n an i n t e r a c t i o n term with the o v e r a l l c o e f f i c i e n t obtained with the i n t e r a c t i o n term i n c l u d e d . The d i f f e r e n c e between the two values i s the a d d i t i o n a l c o n t r i b u t i o n of the i n t e r a c t i o n term. Such an approach could e a s i l y be adapted to the l o g i s t i c model, i f the c r i t e r i o n f o r improved p r e d i c t i o n was 144 the i n c r e a s e i n the slope of the r i s k g r a d i e n t . An attempt to d e s c r i b e the i n t e r a c t i o n of two r i s k f a c t o r s as a form of synergism (or antagonism) c o u l d p o s s i b l y lead to a l t e r n a t e i n t e r v e n t i o n or p r e v e n t i o n s t r a t e g i e s . The l o g i s t i c model "sees" the a c t i o n of r i s k f a c t o r s as independent. The amount of r i s k r e d u c t i o n obtained by a decrease in a r i s k f a c t o r does depend on the l e v e l s of other r i s k f a c t o r s , but r e d u c t i o n of one f a c t o r should not a l t e r the l e v e l s of other f a c t o r s s i n c e the r i s k f a c t o r s add l i n e a r l y i n t o the exponent of the l o g i s t i c e quation. However, i t i s p o s s i b l e that the r i s k f a c t o r s i n t e r a c t with each other, and that r e d u c t i o n of one f a c t o r may be most e a s i l y a t t a i n e d when some other f a c t o r i s at a s p e c i f i c l e v e l . I t i s a l s o p o s s i b l e that v a r i a b l e s which do not enter s i g n i f i c a n t l y i n t o the l o g i s t i c equation c o u l d be the most e f f i c i e n t v a r i a b l e s at reducing r i s k . For example, i t has been suggested that e x e r c i s e c o u l d be the key to reducing a l l l e v e l s of coronary heart d i s e a s e r i s k f a c t o r s , but e x e r c i s e has not yet been found to be a r i s k f a c t o r through l o g i s t i c a n a l y s i s . However, the c a u s a l chain l e a d i n g to the development of coronary heart d i s e a s e has not been determined, and there are models that c o u l d l e a d to a s i t u a t i o n where e x e r c i s e would be the most b e n e f i c i a l i n t e r v e n t i o n , but not show up as s i g n i f i c a n t in a l o g i s t i c a n a l y s i s . Consider a s i t u a t i o n where the r i s k f a c t o r s found 1 45 through a l o g i s t i c a n a l y s i s do cause coronary heart d i s e a s e . In a d d i t i o n , assume that h i g h l e v e l s of these r i s k f a c t o r s a r i s e due to a lack of e x e r c i s e . There i s l i k e l y to be a great deal of v a r i a t i o n between i n d i v i d u a l s in t h e i r p h y s i o l o g i c a l response to e x e r c i s e . Some people w i l l respond to lower l e v e l s of p h y s i c a l e x e r t i o n than others, and coarse measures of e x e r c i s e l e v e l s w i l l not be s e n s i t i v e to t h i s i n d i v i d u a l v a r i a t i o n . Thus low l e v e l s of e x e r c i s e may not show up as a r i s k f a c t o r ( s i n c e t h i s l e v e l may be s u f f i c i e n t f o r a number of i n d i v i d u a l s ) , when, in the model o u t l i n e d above, an i n d i v i d u a l l y t a i l o r e d e x e r c i s e program would be the only way to e f f i c i e n t l y reduce the l e v e l of the r i s k f a c t o r s . Research i n t o the t o t a l c a u s a l chain of coronary heart d i s e a s e , rather than attempts to f i n d the immediately preceding f a c t o r s , would b e n e f i t the development of i n t e r v e n t i o n s t r a t e g i e s . A second area that c o u l d b e n e f i t from a d d i t i o n a l r e s e a r c h i s i n the s p e c i f i c i t y of the r i s k f a c t o r . That i s , i s a v a r i a b l e only a r i s k f a c t o r f o r coronary heart d i s e a s e , or does i t f u n c t i o n as a r i s k f a c t o r for a v a r i e t y of problems? T h i s q u e s t i o n i s of s p e c i a l i n t e r e s t to h e a l t h p l a n n e r s , as the answer to t h i s q u e s t i o n c o u l d i n f l u e n c e the s e l e c t i o n of which r i s k f a c t o r s are i n i t i a l l y a l t e r e d . For example, McGee and Gordon (1976) present a v a r i e t y of t a b l e s of l o g i s t i c analyses from the P o o l i n g P r o j e c t study. In the equations d e r i v e d from coronary heart d i s e a s e death, the 1 46 v a r i a b l e s serum c h o l e s t e r o l , smoking and s y s t o l i c blood pressure entered s i g n i f i c a n t l y i n t o most of the equations. Combining t h i s r e s u l t with the r e s u l t s of the Oslo study (Hjermann et a l . , 1981) would suggest that serum c h o l e s t e r o l l e v e l s should be the prime r i s k f a c t o r to attempt to a l t e r . However, McGee and Gordon (1976) found that f o r death from a l l causes, the v a r i a b l e s y s t o l i c blood pressure was by the f a r the best p r e d i c t o r . Smoking a l s o s i g n i f i c a n t l y entered the equations. Serum c h o l e s t e r o l was not a s i g n i f i c a n t r i s k f a c t o r f o r death from a l l causes i n any of the f i v e l o g i s t i c equations they presented. Thus i t i s p o s s i b l e that even i f an immediate r e d u c t i o n i n serum c h o l e s t e r o l l e v e l s c o u l d produce the g r e a t e s t r e d u c t i o n i n coronary heart d i s e a s e r i s k , l a r g e r gains i n o v e r a l l r i s k r e d u c t i o n c o u l d be obtained through a decrease i n the i n d i v i d u a l ' s s y s t o l i c blood p r e s s u r e . Answers to these s o r t s of qu e s t i o n s c o u l d a l t e r the form of any i n t e r v e n t i o n program. While demonstration of a c a u s a l r e l a t i o n s h i p between some r i s k f a c t o r s and the development of coronary heart d i s e a s e would immediately suggest the i n t r o d u c t i o n of some s o r t of i n t e r v e n t i o n program aimed at r i s k f a c t o r r e d u c t i o n , the e f f i c i e n c y of any such program c o u l d be impeded i f such i s s u e s were not a l s o examined. 1 47 CHAPTER 7  Conclusio n s The m u l t i p l e l o g i s t i c model has proven to be a u s e f u l way of both determining r i s k f a c t o r s f o r coronary heart d i s e a s e and c a l c u l a t i n g the i n d i v i d u a l r i s k of developing coronary heart d i s e a s e . The success of t h i s model has made i t a n a t u r a l candidate f o r use i n any coronary heart d i s e a s e p r e v e n t i o n or i n t e r v e n t i o n program. When examined from the p e r s p e c t i v e of pr e v e n t i o n or i n t e r v e n t i o n t h i s model has many appealing a s p e c t s , but any such a p p l i c a t i o n i s faced with some p o t e n t i a l drawbacks and problems. The focus of t h i s paper has been on the problems and drawbacks, not because the model i s seen as i n a p p r o p r i a t e , but because the great b e n e f i t s of the model have tended to overshadow the p o t e n t i a l problems. None of these problems appear i n s o l v a b l e , but u n t i l these i s s u e s are addressed the p o t e n t i a l b e n e f i t s of any i n t e r v e n t i o n or p r e v e n t i o n program based on t h i s model appear suspect. Some of the important p o i n t s or c o n c l u s i o n s that have been r a i s e d in v a r i o u s p a r t s of t h i s paper i n c l u d e : 1. A r i s k f a c t o r does not n e c e s s a r i l y have to be a c a u s a l f a c t o r . The term r i s k f a c t o r i s sometimes used in a manner that i m p l i e s c a u s a l i t y , but the r e g r e s s i o n or c o r r e l a t i o n methods used to d e r i v e r i s k f a c t o r s c r e a t e s the p o s s i b i l i t y that a noncausal v a r i a b l e c o u l d be 1 48 l a b e l l e d a r i s k f a c t o r . 2. U n i v a r i a t e analyses of p o t e n t i a l r i s k f a c t o r s r e s u l t e d in a long l i s t of p o s s i b l e r i s k f a c t o r s f o r coronary heart d i s e a s e . Four main r i s k f a c t o r s were u s u a l l y i d e n t i f i e d (age, serum c h o l e s t e r o l , blood pressure and smoking), of which a l l but age can be a l t e r e d . Each p r o s p e c t i v e study a l s o found other v a r i a b l e s that were a s s o c i a t e d with the development of coronary heart d i s e a s e . T h i s l a r g e number of v a r i a b l e s , when combined with the unknown nature of t h e i r i n t e r r e l a t i o n s h i p s , s e v e r l y l i m i t e d the u s e f u l n e s s of r i s k f a c t o r s d e r i v e d through u n i v a r i a t e a n a l y s e s . 3. One a p p r o p r i a t e m u l t i v a r i a t e model for r i s k f a c t o r a n a l y s i s uses the m u l t i p l e l o g i s t i c f u n c t i o n . T h i s f u n c t i o n i s c o n s i s t e n t with a b i o l o g i c a l view of an organism's response to harmful exposures, and i s a l s o c o n s i s t e n t with present t h e o r i e s on the development of coronary heart d i s e a s e . 4. Two methods e x i s t for the c a l c u l a t i o n of l o g i s t i c c o e f f i c i e n t s when the outcome v a r i a b l e i s c a t e g o r i c a l . Both methods are u s e f u l f o r i d e n t i f y i n g r i s k f a c t o r s , but the method of Walker and Duncan (1967) i s to be p r e f e r r e d i f the c o e f f i c i e n t s are to be used f o r p r e d i c t i v e purposes. 5. The same set of major r i s k f a c t o r s are found from a l o g i s t i c a n a l y s i s as were found with u n i v a r i a t e 1 49 a n a l y s e s . The l i s t of a d d i t i o n a l r i s k f a c t o r s i s s h o r t e r than that provided by u n i v a r i a t e a n a l y s e s , but there s t i l l e x i s t s some v a r i a b l e s that have been shown to s i g n i f i c a n t l y enter i n t o the l o g i s t i c equation i n one study and d i d not (or were not examined) in other s t u d i e s . The l o g i s t i c equation presents a method of measuring coronary heart d i s e a s e r i s k that i n c o r p o r a t e s i n f o r m a t i o n about a l l of the i n d i v i d u a l ' s r i s k f a c t o r s that an experimenter c o n s i d e r s to be important. V a l i d a t i o n of the l o g i s t i c p r e d i c t i o n s i s d i f f i c u l t . The usual method of determining a model's goodness-of-f i t can produce m i s l e a d i n g r e s u l t s , and thus an examination of the model's r i s k g r a d i e n t i s perhaps a more u s e f u l c r i t e r i o n . Attempts to c r o s s - v a l i d a t e the model i n d i c a t e that equations d e r i v e d from the same c u l t u r a l group f i t reasonably w e l l (as long as the d i a g n o s t i c c r i t e r i a are the same), but c r o s s - c u l t u r a l v a l i d a t i o n has been poor. Attempts to a l t e r the major r i s k f a c t o r s have proven s u c c e s s f u l , but to date only one completed p r o s p e c t i v e study has been able to show subsequent changes in m o r b i d i t y (Oslo s t u d y ) . T h i s study was able to show a g r e a t e r r e d u c t i o n in coronary heart disease m o r b i d i t y in the experimental group when compared to the c o n t r o l group. Other s t u d i e s (North K a r e l i a and MRFIT) have f a i l e d to f i n d d i f f e r e n c e s 1 50 between the experimental and c o n t r o l groups even though they have observed a decrease in the r i s k f a c t o r l e v e l s of the experimental groups. Experimental d i f f i c u l t i e s i n showing such r e d u c t i o n s are d i s c u s s e d . 8. S e v e r a l problem areas e x i s t i f attempts are made to design a p r e v e n t i o n or i n t e r v e n t i o n program on a l o g i s t i c model. These i n c l u d e : s e l e c t i o n of an a p p r o p r i a t e equation, s e l e c t i o n of the i n t e r v e n t i o n or p r e v e n t i o n s t r a t e g y (most notably in the area of t a r g e t l e v e l s f o r the r i s k f a c t o r s ) , the l a r g e number of f a l s e p o s i t i v e s that would be produced from the use of the l o g i s t i c equation, the p o s s i b i l i t y of i n t e r a c t i o n s between v a r i a b l e s not being s u f f i c i e n t l y i n c o r p o r a t e d i n t o the model, and the p o s s i b i l i t y that the l o g i s t i c model would f a i l to i n c o r p o r a t e a v a r i a b l e that was best s u i t e d f o r i n t e r v e n t i o n . While none of these problems has an easy s o l u t i o n , none appear to be u n s o l v a b l e . 151 References Anderson, T.W. Re-examination of some of the Framingham blood-pressure data. The Lancet, 1978, 2, 1139-1141. B i o r c k , G. C o n t r a s t i n g Concepts of Ischaemic Heart Disease. Stockholm: Almqvist & W i k s e l l , 1975. Birkenhager, W.H. & Schalekamp, M.A. C o n t r o l Mechanisms in  E s s e n t i a l Hypertension. New York: E l s e v i e r S c i e n t i f i c P u b l i s h i n g Co., 1976. Brand, R.J., Rosenman, R.H., S h o l t z , R.I., & Friedman, M. M u l t i v a r i a t e p r e d i c t i o n of coronary heart d i s e a s e i n the Western C o l l a b o r a t i v e Group study compared to the f i n d i n g s of the Framingham study. C i r c u l a t i o n , 1976, 53, 348-355. Breslow, L. Risk f a c t o r i n t e r v e n t i o n f o r h e a l t h maintenance. Science, 1978a, 200, 908-912. Breslow, L. Prospects f o r improving h e a l t h through reducing r i s k f a c t o r s . P r e v e n t i v e Medicine, 1978b, 7, 449-458. Brunner, D., & L o b l , K. Serum c h o l e s t e r o l , e l e c t r o p h o r e t i c l i p i d p a t t e r n , d i e t , and coronary a r t e r y d i s e a s e : A study i n coronary p a t i e n t s and i n healthy men of d i f f e r e n t o r i g i n and occupations i n I s r a e l . Annals of  I n t e r n a l Medicine, 1958, 49, 732-750. Burch, P.R.J. Ischaemic heart d i s e a s e : Epidemiology, r i s k f a c t o r s and cause. C a r d i o v a s c u l a r Research, 1980, 14, 307-338. Chapman, J.M., Georke, L.S., Dixon, W., Loveland, D.B., & P h i l l i p s , E. The c l i n i c a l s t a t u s of a p o p u l a t i o n group in Los Angeles under o b s e r v a t i o n f o r two to three years. American J o u r n a l of P u b l i c Health, 1957, Supplement to V o l . 47, 33-42. Chapman, J.M., & Massey, F . J . The i n t e r r e l a t i o n s h i p of serum c h o l e s t e r o l , h y p e r t e n s i o n , body weight, and r i s k of coronary d i s e a s e . J o u r n a l of Chronic Diseases, 1964, J_7, 933-949. Corday, E., & Corday, S.R. Prevention of heart d i s e a s e by c o n t r o l of r i s k f a c t o r s : the time has come to face the f a c t s . The American J o u r n a l of C a r d i o l o g y , 1975, 35, 1 52 330-334. C o r n f i e l d , J . J o i n t dependence of r i s k of coronary heart d i s e a s e on serum c h o l e s t e r o l and s y s t o l i c blood p r e s s u r e : A d i s c r i m i n a n t f u n c t i o n a n a l y s i s . F e d e r a t i o n  Proceedings, 1962, 2J_, 58-61. C o r n f i e l d , J . , Gordon, T., & Smith, W.W. Quantal response curves f o r e x p e r i m e n t a l l y c o n t r o l l e d v a r i a b l e s . B u l l e t i n of the I n t e r n a t i o n a l S t a t i s t i c a l I n s t i t u t e , 1 961 , 3_8, 97-1 15. Davis, F.W. Summary of r e p o r t s p e r t a i n i n g to c i g a r e t t e smoking and the c a r d i o v a s c u l a r system: C l i n i c a l data 1956-60. C i r c u l a t i o n , 1960, 22, 163-164. Dawber, T.R. Summary of recent l i t e r a t u r e r e g a r d i n g c i g a r e t t e smoking and coronary heart d i s e a s e . C i r c u l a t i o n , 1960, 22, 164-166. Dawber, T.R. The Framingham Study. The Epidemiology of  A t h e r o s c l e r o t i c Disease. Cambridge, Mass.: Harvard U n i v e r s i t y Press, 1980. Dawber, T.R., Meadors, G.F., & Moore, F.E. E p i d e m i o l o g i c a l approaches to heart d i s e a s e : The Framingham study. American J o u r n a l of P u b l i c Health, 1951, 4_[, 279-286. Dawber, T.R., Moore, F.E., & Mann, G.V. Coronary heart d i s e a s e i n the Framingham study. American J o u r n a l of  P u b l i c H e a l t h , 1957, Supplement to V o l . 47, 4-24. Doyle, J.T., H e s l i n , A.S., H i l l b o e , H.E., & Formel, P.F. E a r l y d i a g n o s i s of ischaemic heart d i s e a s e . The New  England J o u r n a l of Medicine, 1959, 261, 1096-1101. Doyle, J.T., H e s l i n , A.S., H i l l b o e , H.E., Formel, P.F., & Korns, R.F. A p r o s p e c t i v e study of degenerative c a r d i o v a s c u l a r d i s e a s e in Albany: Report of three years e x p e r i e n c e - - I . Ischaemic heart d i s e a s e . American  J o u r n a l of P u b l i c H e a l t h , 1957, Supplement to V o l . 47, 25-32. F l e i s s , J.L. S t a t i s t i c a l Methods f o r Rates and P r o p o r t i o n s . New York: John Wiley & Sons, 1981. Friedman, G.D. Primer of Epidemiology. New York: McGraw-H i l l , 1974. G e r t l e r , M.M., & White, P.D. Coronary heart d i s e a s e . O r a d e l l , New J e r s e y : Medical Economics Co., 1976. 153 Gordon, T. Hazards i n the use of the l o g i s t i c f u n c t i o n with s p e c i a l reference to data from p r o s p e c t i v e c a r d i o v a s c u l a r s t u d i e s . J o u r n a l of Chronic Diseases, 1974, 27, 97-102. Gordon, T., & S h u r t l e f f , D. Means at each examination and int e r - e x a m i n a t i o n v a r i a t i o n of s p e c i f i e d c h a r a c t e r i s t i c s : Framingham study, Exam 1 to Exam 10. S e c t i o n 29, The Framingham Study. Department of Health Education and Welfare P u b l i c a t i o n s (NIH) 74-478, N a t i o n a l I n s t i t u t e of He a l t h , 1973. Grundy, P.F. A r a t i o n a l approach to the "at r i s k " concept. Lancet, 1973, 2, 1489. H a l l , J.H., & Zwemer, J.D. P r o s p e c t i v e Medicine. I n d i a n a p o l i s : Methodist H o s p i t a l of Indiana, 1979. H a l p e r i n , M. , Blackwelder, W.C, & V e r t e r , J . I . E s t i m a t i o n of the m u l t i v a r i a t e l o g i s t i c r i s k f u n c t i o n : a comparison of the d i s c r i m i n a n t f u n c t i o n and the maximum l i k e l i h o o d approaches. J o u r n a l of Chronic Diseases, 1971, 24, 125-158. Hart, J.T. Hypertension. London: C h u r c h i l l L i v i n g s t o n e , 1 980. H i l l e b o e , H.E., James, G., & Doyle, J.T. C a r d i o v a s c u l a r Health Center. I. P r o j e c t design f o r p u b l i c h e a l t h r e s e a r c h . American J o u r n a l of P u b l i c H e a l t h , 1954, 44, 851-863. Hjermann, I., Velve Byre, K., Holme, I., & Leren, P. E f f e c t of d i e t and smoking i n t e r v e n t i o n on the i n c i d e n c e of coronary heart d i s e a s e . Lancet, 1981, 2, 1303-1310. Kannel, W.B. Some lessons in c a r d i o v a s c u l a r epidemiology from Framingham. American J o u r n a l of C a r d i o l o g y , 1976, 37, 269-282. K e r l i n g e r , F.N., & Pedhazur, E . J . M u l t i p l e Regression i n  B e h a v i o r a l Research. New York: H o l t , Rinehart and Winston, 1973. Keys, A., Aravanis, C , Blackburn, H., van Buchen, F.S.P., Buzina, R., D j o r d j e v i c , B.S., Fidanza, F., Karvonen, M.J., Menotti, A., Puddu, V., & T a y l o r , H.L. P r o b a b i l i t y of middle-aged men developing coronary heart d i s e a s e in f i v e y e a r s . C i r c u l a t i o n , 1972, 45, 815-828. Keys, A., T a y l o r , H.L., Blackburn, H., Brozek, J . , Anderson, 1 54 T.T., & Simonson, E. Coronary heart d i s e a s e among Minnesota buisness men and p r o f e s s i o n a l men followed f i f t e e n y e a r s . C i r c u l a t i o n , 1 9 6 3 , 2 8 , 3 8 1 - 3 9 5 . Keys, A., T a y l o r , H.L., Blackburn, H., Brozek, J . , Anderson, T.T., & Simonson, E. M o r t a l i t y and coronary heart d i s e a s e among men s t u d i e d f o r 23 years. A r c h i v e s of  I n t e r n a l Medicine, 1 9 7 1 , 2 1 8 , 2 0 1 - 2 1 4 . Kleinbaum, D.G., Kupper, L.L., C a s s e l , J . C , & T y r o l e r , H.A. M u l t i v a r i a t e a n a l y s i s of r i s k of coronary heart d i s e a s e i n Evans County, Georgia. A r c h i v e s of I n t e r n a l  Medicine, 1 9 7 1 , J_2j8, 9 4 3 - 9 4 8 . L e i b o w i t z , J.O. The H i s t o r y of Coron'ary Heart Disease. London: Willcome I n s t i t u t e of the H i s t o r y of Medicine, 1 9 7 0 . Levy, R.I., & E r n s t , N. D i e t , h y p e r l i p i d e m i a and a t h e r o s c l e r o s i s . In R.S. Goodhart & M.E. S h i l s (Eds.) Modern N u t r i t i o n i n H e a l t h and Disease, P h i l a d e l p h i a : Lea & F e b i g e r , 1 9 7 3 , 8 9 5 - 9 1 8 . Lyon, T.P., Yankley, A., Gofman, J.W., & S t r i s o w e r , B. L i p o p r o t e i n s and d i e t i n coronary heart d i s e a s e . C a l i f o r n i a Medicine, 1 9 5 6 , 8 4 , 3 2 5 - 3 2 8 . MacMahon, B., & Pugh, T.F. Epidemiology: P r i n c i p l e s and  Methods. Boston: L i t t l e , Brown and Company, 1 9 7 0 . McGee, D. The p r o b a b i l i t y of developing c e r t a i n • c a r d i o v a s c u l a r d i s e a s e s in e i g h t years at s p e c i f i e d values of some c h a r a c t e r i s t i c s . S e c t i o n 2 8 : The  Framingham Study. Department of Health, Education, and Welfare P u b l i c a t i o n s (NIH) 7 4 - 6 1 8 , N a t i o n a l I n s t i t u t e of H e a l t h , 1 9 7 3 . McGee, D., & Gordon, T. The r e s u l t s of the Framingham study a p p l i e d to four other U.S.-based e p i d e m i o l o g i c a l s t u d i e s of c a r d i o v a s c u l a r d i s e a s e . Sect ion 3 1: The  Framingham Study. Department of Health, Education, and Welfare P u b l i c a t i o n s (NIH) 7 6 - 1 0 8 3 , N a t i o n a l I n s t i t u t e of Health, 1 9 7 6 . Morse, P.M., Some comments on the assessment of j o i n t a c t i o n i n h e r b i c i d e mixtures. Weed Science, 1 9 7 8 , 2 6 , 5 8 - 7 1 . M o s t e l l e r , F., & Tukey, J.W. Data A n a l y s i s and Regression. Reading, Massachusetts: Addison-Wesley, 1 9 7 7 . Neaton, J.D., Broste, S., Cohen, L., Fishman, E.L., K j e l s b e g , M.O., & Schoenberg, J . The M u l t i p l e Risk 155 Factor I n t e r v e n t i o n T r i a l (MRFIT) V I I . A comparison of r i s k f a c t o r changes between the two groups. Prevent ive  Medicine, 1981, j_0, 519-543. Paul, 0., Lepper, M.H., Phelan, W.H., Dupe r t u i s , G.W., MacMillan, A., McKean, H., & Park, H. A l o n g i t u d i n a l study of coronary heart d i s e a s e . C i r c u l a t i o n , 1963, 28, 20-31. P o o l i n g P r o j e c t Research Group. R e l a t i o n s h i p of blood p r e s s u r e , serum c h o l e s t e r o l , smoking h a b i t s , r e l a t i v e weight and ECG a b n o r m a l i t i e s to inc i d e n c e of major coronary events. F i n a l Report of the P o o l i n g P r o j e c t . J o u r n a l of Chronic Diseases, 1978, 3_8, 201-306. Puska, P. The North K a r e l i a p r o j e c t : An attempt at community preve n t i o n of c a r d i o v a s c u l a r d i s e a s e . World  Health O r g a n i z a t i o n C h r o n i c l e , 1973, 27, 55-58. Puska, P., Tuomilehto, J . , Salonen, J . , Neittaanmaki, L., Maki, J . , Virtamo, J . , N i s s i n e n , A., Koskela, K., & Takalo, T. Changes i n coronary r i s k f a c t o r s during comprehensive f i v e - y e a r community programme to c o n t r o l c a r d i o v a s c u l a r d i s e a s e s (North K a r e l i a P r o j e c t ) . B r i t i s h Medical J o u r n a l , 1979, 2, 1173-1178. Rabkin, S.W., Mathewson, F.A., & Ping-Hwa, H. R e l a t i o n of body weight to development of ischemic heart d i s e a s e in a cohort of young North American men a f t e r a 26 year o b s e r v a t i o n p e r i o d : The Manitoba study. American  J o u r n a l of C a r d i o l o g y , 1977, 3_9, 452-458. Rosenman, R.H,, Brand, R.J., S h o l t z , R.I., & Friedman, M. M u l t i v a r i a t e p r e d i c t i o n of coronary heart d i s e a s e d u r i n g 8.5 year follow-up i n the Western C o l l a b o r a t i v e Group study. American J o u r n a l of C a r d i o l o g y , 1976, 37, 903-910. Rosenman, R.H,, Friedman, M. Stra u s , R., Je n k i n s , CD., Zyzanski, S.J../& Wurm, M. Coronary heart d i s e a s e i n the Western C o l l a b o r a t i v e Group Study. A follow-up experience of 4 1/2 years. J o u r n a l of Chronic Diseases , 1970, 23, 173-190. Rosenman, R.H., Friedman, M., Stra u s , R., Wurm, M., Kositchek, R., Hahn, W., & Werthessen, N.T. A p r e d i c t i v e study of coronary heart d i s e a s e . J o u r n a l of  the American Medical A s s o c i a t i o n , 1964, 189, 15-22. Salonen, J . , Puska, P., & Mustaniemi, H. Changes i n mor b i d i t y and m o r t a l i t y d u r i n g comprehensive community programme to c o n t r o l c a r d i o v a s c u l a r d i s e a s e s during 1 56 1972-7 i n North K a r e l i a . B r i t i s h Medical J o u r n a l , 1979, 2, 1 173-1178. S c h r o l l , M., & Larsen, S. A ten-year p r o s p e c t i v e study, 1964-74, of c a r d i o v a s c u l a r r i s k f a c t o r s i n men and women from the G l o s t r u p p o p u l a t i o n born i n 1914. Danish Medical B u l l e t i n , 1981, 28, 236-251. S h u r t l e f f , D. Some c h a r a c t e r i s t i c s r e l a t e d to the inc i d e n c e of c a r d i o v a s c u l a r d i s e a s e and death. S e c t i o n 30, The  Framingham Study. Department of Health Education and Welfare P u b l i c a t i o n s (NIH) 74-599, N a t i o n a l I n s t i t u t e of H e a l t h , 1974. Simborg, D.W. The s t a t u s of r i s k f a c t o r s and coronary heart d i s e a s e . J o u r n a l of Chronic Diseases, 1970, 22, 515-552. Smith, M.L., Glass, G.V., & M i l l e r , T.I. The B e n e f i t s of  Psychotherapy. B a l t i m o r e : John Hopkins U n i v e r s i t y Press, 1980. T e c h n i c a l Group of the Committee on L i p o p r o t e i n s and A t h e r o s c l e r o s i s . E v a l u a t i o n of serum l i p o p r o t e i n and c h o l e s t e r o l measurements as p r e d i c t o r s of c l i n i c a l c o m p l i c a t i o n s of a t h e r o s c l o r o s i s . C i r c u l a t i o n , 1956, j_4 , 691 -741. T r u e t t , J . , C o r n f i e l d , J . , & Kannel, W. A m u l t i v a r i a t e a n a l y s i s of the r i s k of coronary heart d i s e a s e in Framingham. J o u r n a l of Chronic Diseases, 1967, 20, 511-524. V l i e t s t r a , R.E., Frye, R.L., Kronmal, R.A., Sim, D.A., T r i s t a n i , F.E., & K i l l i p I I I , T. Risk f a c t o r s 'and angiographic coronary a r t e r y d i s e a s e : The s t a t i s t i c a l a n a l y s i s of r i s k . C i r c u l a t i o n , 1980, 62, 254-261. Walker, S.H., & Duncan, D.B. E s t i m a t i o n of the p r o b a b i l i t y of an event as a f u n c t i o n of s e v e r a l independent v a r i a b l e s . B i o m e t r i c s , 1967, 167-179. Weiss, N.S. I n f e r r i n g c a u s a l r e l a t i o n s h i p s . American  J o u r n a l of Epidemiology, 1981, 113, 487-490. Wilhelmsen, L., Wedel, H., & T i b b l i n , G. M u l t i v a r i a t e a n a l y s i s of r i s k f a c t o r s f o r coronary heart d i s e a s e . C i r c u l a t i o n , 1972, 45, 815-828. Zukel, W.J., Paul, 0., & Schnaper, H.W. The m u l t i p l e r i s k f a c t o r i n t e r v e n t i o n t r i a l (MRFIT). 1. H i s t o r i c a l p e r s p e c t i v e s . P r e v e n t i v e Medicine, 1981, j_0, 387-401. 1 57 A d d i t i o n a l B i b l i o g r a p h y B e n f a r i , R.C. The M u l t i p l e Risk I n t e r v e n t i o n T r i a l (MRFIT) I I I . The model f o r i n t e r v e n t i o n . P r e v e n t i v e Medicine, 1 981 , J_0, 426-442. Berenson, G.S. C a r d i o v a s c u l a r Risk F a c t o r s i n C h i l d r e n . Oxford: Oxford U n i v e r s i t y Press, 1980. Bi o r c k , G. Some fundemental problems i n coronary heart d i s e a s e . Acta Medica S c a n d i n a v i c a , 1960, 168, 245-268. Ciocco, A.I. I m p l i c a t i o n s f o r the b i o s t a t i s t i c i a n . American J o u r n a l of P u b l i c Health, 1957, Supplement to  V o l . 47, 58-59. C i r q u i , M.H., Barret-Connor, E., Holdbrook, M.J., A u s t i n , M., & Turner, J.D. C l u s t e r i n g of c a r d i o v a s c u l a r r i s k f a c t o r s . P r e v e n t i v e Medicine, 1980, 9, 525-533. Dawber, T.R., Kannel, W.B., Revotskie, N., Stokes, J . , Kagen, A., & Gordon, T. Some f a c t o r s a s s o c i a t e d with the development of coronary heart d i s e a s e . Six years' follow-up experience i n the Framingham study. American  J o u r n a l of P u b l i c Health, 1959, 49, 1349-1356. D o l l , R. P r e v e n t i o n : Some f u t u r e p e r s p e c t i v e s . Prevent ive  Medicine, 1978, 7, 486-497. E l m f e l d t , D., Wilhelmsen, L., Wedel, H., Vedin, A., Wilhelmsen, C , & T i b b l i n , G. Primary r i s k f a c t o r s i n p a t i e n t s with myocardial i n f a r c t i o n . American Heart  J o u r n a l , 1976, 9_1_, 412-418. E r i k s s e n , • J . , & Enger, S.C. The e f f e c t of smoking on s e l e c t e d coronary heart d i s e a s e r i s k f a c t o r s in middle-aged men. Acta Medicine Scandinavia, 1978, 203, 31-37. Goldsmith, J.R. Paths of a s s o c i a t i o n i n e p i d e m i o l o g i c a l a n a l y s i s : A p p l i c a t i o n to h e a l t h e f f e c t s of environmental exposures. I n t e r n a t i o n a l J o u r n a l of  Epidemiology, 1977, 6, 391-399. Haust, M.D., & More, R.H. Development of modern t h e o r i e s on the pathogenesis of A t h e r o s c l e r o s i s . In R.W. W i s s l e r & J.C. Geer (Eds.) The Pathogenesis of A t h e r o s c l e r o s i s . B a l t i m o r e : W i l l i a m s & W i l k i n s , 1972, 1-19. Kagan, A., Kannel, W.B., Dawber, T.R., & Revotskie, N. The coronary p r o f i l e . Annals of the New York Academy of  Scie n c e s , 1962, 97, 883-894. 158 K o l a t a , G.B. Prevention of heart d i s e a s e : c l i n i c a l t r i a l s at what cost? Science, 1975, J 9 0 , 764-765. L e l l o u c h , J . , & Rakotouao, R. E s t i m a t i o n of r i s k as a f u n c t i o n of r i s k f a c t o r s . I n t e r n a t i o n a l J o u r n a l of  Epidemiology, 1976, 5, 349-352. Levy, R.I. Progress in pr e v e n t i o n of c a r d i o v a s c u l a r d i s e a s e . P r e v e n t i v e M e c i c i n e , 1978, ]_, 464-475. Levy, R.I., & Stone, N.J. A t h e r o s c l e r o s i s : Role of l i p o p r o t e i n s . In R.W. W i s s l e r & J.C. Geer (Eds.) The  Pathogenesis of A t h e r o s c l e r o s i s . B a l t i m o r e : W i l l i a m s & W i l k i n s , 1972, 227-238. Medalie, J.H., Kahn, H.A., Neufeld, H.N., R i s s , E., & Goldbourt, U. F i v e - y e a r myocardial i n f a r c t i o n i n c i d e n c e - I I . A s s o c i a t i o n of s i n g l e v a r i a b l e s to age and b i r t h p l a c e . J o u r n a l of Chronic D i s e a s e s , 1973, -26, 329-349. Meyer, A.J., & Henderson, J.B. M u l t i p l e r i s k f a c t o r r e d u c t i o n i n the pr e v e n t i o n of c a r d i o v a s c u l a r d i s e a s e . Preventive Medicine, 1974, 3, 225-326. M i e t t i n e n , M., Turpeinen, 0 . , Karvonen, M.J., Elosou, R., & P a a v i l a i n e n , E. E f f e c t of a c h o l e s t e r o l - l o w e r i n g d i e t on m o r t a l i t y from coronary h e a r t - d i s e a s e and other causes. The Lancet, 1972, 2, 835-838. Ostrander, L.D., & Lamphiear, D.E. Coronary r i s k f a c t o r s i n a community. F i n d i n g s in Techumseh, Michigan. C i r c u l a t i o n , 1976, 53, 152-156. Paul, 0 . Prevention of c a r d i o v a s c u l a r d i s e a s e . Prevent ive  Medicine, 1973, 2, 473-479. Rahe, R.H., Bennett, L., Romo, M., S i l t a n e n , P., & Arthur, R.J. Subjects' recent l i f e changes and coronary heart disease in F i n l a n d . American J o u r n a l of P s y c h i a t r y , 1973, J_30, 1222-1226. S c o t t , D.W., Gorry, A., & Gotto, A.M. E d i t o r i a l : D i e t and coronary heart d i s e a s e : The s t a t i s t i c a l a n a l y s i s of r i s k . C i r c u l a t i o n , 1981, 63, 516-518. Sherwin, R., Kaelber, C.T., Kezdi, P., K j e l s b e r g , M.O., & Thomas, H.E. The M u l t i p l e Risk F a c t o r I n t e r v e n t i o n T r i a l (MRFIT) I I . The development of the p r o t o c o l . Preventive Medicine, 1981, J_0, 402-405. Tuomilehto, J . , Puska, P., Virtamo, J . , Neittaanmaki, L., & 159 Koskela, K. Coronary r i s k f a c t o r s and socioeconomic s t a t u s i n Eastern F i n l a n d . P r e v e n t i v e Medicine, 1978, 7, 539-549. Upton, A.C. Progress i n the p r e v e n t i o n of cancer. P r e v e n t i v e Medicine, 1978, 7, 476-485. Weinstein, M.C, & Stason, W.B. Hypertension. A P o l i c y  P e r s p e c t i v e . Cambridge, Mass: Harvard U n i v e r s i t y Press, 1976. Werko, L. Risk f a c t o r s and coronary heart d i s e a s e — f a c t s or fancy? American Heart J o u r n a l , 1976, 9J_, 87-91. ' Wong, 0 . A competing r i s k model based on the l i f e t a b l e procedure in e p i d e m i o l o g i c a l s t u d i e s . I n t e r n a t i o n a l  J o u r n a l of Epidemiology, 1977, 6, 153-159. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0095395/manifest

Comment

Related Items