UBC Faculty Research and Publications

Diagnosis codes on extended care separations included in the discharge abstract data : a summary report McGrail, Kimberlyn, 1966- Jul 31, 2000

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

Item Metadata

Download

Media
52383-McGrail_Kimberlyn_Diagnosis_codes.pdf [ 62.26kB ]
Metadata
JSON: 52383-1.0048484.json
JSON-LD: 52383-1.0048484-ld.json
RDF/XML (Pretty): 52383-1.0048484-rdf.xml
RDF/JSON: 52383-1.0048484-rdf.json
Turtle: 52383-1.0048484-turtle.txt
N-Triples: 52383-1.0048484-rdf-ntriples.txt
Original Record: 52383-1.0048484-source.json
Full Text
52383-1.0048484-fulltext.txt
Citation
52383-1.0048484.ris

Full Text

     Diagnosis Codes On Extended Care  Separations Included In The Discharge  Abstract Data:  A Summary Report    Prepared on behalf of the Ministry of Health OPIC by:  Kimberlyn McGrail   Health Information Development Unit  HIDU  00:02 July, 2000   The University of British ColumbiaHealth Information Development Unit Centre for Health Services and Policy Research   2 Introduction  Information about individuals receiving extended care services through the British Columbia continuing care system is currently captured by two different data sources – the Continuing Care Information Management System (CCIMS) and the Discharge Abstract Database (DAD, or hospital separations).  A previous analysis suggested that there was relatively good agreement between records of care in these two sources.1  There is, however, a group of people who appear in the DAD and not the CCIMS.  Some of this mismatch is explained by different data capture rules, including the recording of care for children (not the responsibility of CCIMS) as well as respite and hospice care.  A small mismatch between the DAD and CCIMS remains, however, even after removing these special cases.  In other words, there is a small number of records of extended care service provision that appear to be missing from CCIMS.  Leaving this (perhaps minor) issue aside, questions have been raised about the utility of continuing what is essentially duplicate recording of care.  The decision about whether reporting of extended care could reasonably be eliminated from the DAD depends, in part, on what information would be lost in doing so.  The hospital separations record system is set up to capture a great deal of information about a person's stay in an acute or extended care facility, including up to sixteen diagnosis codes.  The CCIMS does not have parallel information for either home- or facility-based clients.  To the extent that diagnostic information is at least potentially useful for operational, planning, or research purposes, the discontinuation of extended care reporting on the DAD is a potential loss.  The magnitude of the loss, however, is affected by the breadth of information covered in the diagnoses recorded for these patients.  If only one diagnosis is recorded, and the extended care separation occurs only at death, then the data captured through the hospital data ought to be identical - or nearly so - to the ICD9 codes recorded on the Vital Statistics deaths file.  The purpose of this analysis was to provide a description of the ICD9 codes recorded on hospital separations abstracts for individuals receiving extended care services. Of interest was: • the distribution of the number of diagnoses recorded, from 1 to 16; • the distribution of the diagnoses recorded, at the 3-digit level, by type of diagnosis; • for extended care separations that are the result of death, the match between hospital and Vital Statistics diagnoses.  A match will be determined both comparing only the ‘most responsible diagnosis’ to the Vital Statistics diagnosis, and any diagnosis to Vital Statistics diagnosis.  The data used for this analysis were extended care separations occurring in 1998 identified in the 1997/98 and 1998/99 DAD, and 1998 Vital Statistics deaths data.  Hospital separations data were provided through the BC Linked Health Data set.  Underlying cause of death codes were provided by the BC Vital Statistics Agency.                                                   1 See McGrail KM and McCashin B (2000) A comparison of extended care records across two program areas. Health Information Development Unit, Centre for Health Services and Policy Research, HIDU 2000:01. Health Information Development Unit Centre for Health Services and Policy Research   3 Methods  All separation records with level of care ‘E’, a separation date during calendar 1998 and a linkage PHN (scrambled) were included.  Death registration numbers were identified for people who used extended care and died during 1998.  These were provided to the BC Vital Statistics Agency, which returned ICD9 coding for underlying cause of death.  Comparisons of diagnosis codes were made only at the three-digit level.  Results Descriptive information There were 7,389 extended care separation during 1998.  Nearly 99% of these (7,305) were linked and included in the analyses.  Of these, 4,106 were records indicating separation from extended care alive, and 3,199 were separations following death.  Table 1: Distribution of total days in extended care for separations ending in 1998  Percentile Number of days in EC 10% 12 20% 24 30% 42 40% 82 50% 166 60% 296 70% 488 80% 802 90% 1376 100% 6680  These 7,305 separations record care provided to 5,907 individuals.  The number of separations per individual ranged from 1 (for 5,133 individuals) to 34 (for 1 individual).  Forty-five people had more than 5 separations from extended care during the year.  The number of days spent in extended care for separations ending in 1998 ranged from 0 (10 people) to 6680 (1 person).  The median number of days in extended care was 166 days (Table 1).  Considering only the last separation for each individual during 1998, only slightly more than half (54.1%) were separations due to death.  The number of ICD9 codes recorded on each separation is shown in Figure 1.  Only 6.6% of records have just one ICD9 code; 37.3% have more than five ICD9 codes recorded. Health Information Development Unit Centre for Health Services and Policy Research   4 ICD code frequency Each diagnosis code has an associated ‘type’ which specifies whether the code was the major reason for the hospital stay, was present before admission, arose after admission, or was present but did not contribute to the length of stay (see Appendix for full definitions).  Frequencies of ICD codes by type of diagnosis and overall are provided in Tables 2-6.  The top ten most frequently occurring ‘most responsible diagnosis’ codes account for two-thirds of the codes of this type (Table 2).  The two most commonly occurring diagnoses – dementia and ‘household circumstances’ – account for the majority of length of stay in more than one-third of all extended care separations.  Figure 1: Distribution of number of ICD9 codes recorded on EC separations, 1998036912151 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16Number of ICD9 codes recordedPercent of total separations   As might be expected, there is a far wider range of codes found in the ‘pre-admit comorbidity’ diagnoses (Table 3).  The ten most common ICD9 codes found with this diagnosis type account for about one third of all occurrences.  As with the ‘most responsible diagnosis’, dementia tops the list.  One quarter of all post-admit comorbidities are conditions of the urethra or urinary tract and pneumonia (Table 4).  Secondary diagnoses tend to be chronic health conditions such as hypertension and diabetes (Table 5).  Overall, dementia remains the most common diagnosis (Table 6). Health Information Development Unit Centre for Health Services and Policy Research   5 Comparison of diagnoses  Vital Statistics records of death contain one diagnosis indicating the underlying cause of death (UCOD).  This code is meant to capture “the disease which triggered the chain of events leading directly to the death or the description of the accident or violence that produced the fatality.”2  The question of interest for this report was how often that underlying cause of death code was one of the diagnoses coded in the DAD.  Analysis was limited to the 3,137 individuals whose extended care separation ended in death, and who could be matched to a Vital Statistics death record with a non-blank UCOD.  The former criterion was applied because it is these records that have the greatest likelihood of recording diagnostic information similar to that found on the death record.  Nine of the ten most frequently occurring UCOD codes (Table 7) are also found in at least one of the top ten lists from the DAD (the exception is ‘chronic airways obstruction’, accounting for 3.3% of deaths).  The top ten lists for the UCOD and the most responsible diagnosis field are, however, quite different.  As might be expected from the WHO definition, three of the four most frequent UCOD codes are acute in nature, including cerebrovascular disease, pneumonia and myocardial infarction.  The hospital separations diagnoses, in contrast, are far more chronic and/or social in nature, reflecting the fact that the primary reason for a (usually) relatively long stay in an extended care bed is more likely to be a chronic than an acute condition.  At the individual level, only 330 of the total 3,137 hospital separation records (10.5%) contain a match to the UCOD, comparing the first three digits of the codes.  Only 28% of these are ‘most responsible’ diagnoses.  This means that only one in every thirty-four UCODs and most responsible diagnoses from the DAD are the same.  Even fewer matches were found as pre-admit comorbidity, post-admit comorbidity or secondary diagnoses.   Conclusion  There are at least three findings that suggest ICD9 coding for extended care separations in the DAD may be useful.  First, in most cases there are multiple codes found for each separation record, and there is a reasonably wide variety of codes used.  This implies that coders are not resorting to automatic, generic ICD9 codes for patients in extended care beds.  Second, nearly half of the individuals for whom diagnostic information is available have been discharged from extended care alive.  This means the diagnostic information is not only available after death, with limited use for operational and/or planning purposes.  Finally, the codes found in the Vital Statistics deaths data are only recorded on the hospital separations file in one out of every ten cases.  Clearly the ultimate cause of death is not often a reason (or at least a major reason) for the extended care stay.                                                   2 Quoted from the glossary of the Vital Statistics Agency 1998 Annual Report, Selected vital statistics and health status indicators.  That publication attributes the definition to the World Health Organization (1977). Health Information Development Unit Centre for Health Services and Policy Research   6 The results of this analysis cannot supply a definitive recommendation regarding the continuation of extended care as a level of care on the hospital separations file.  There are other factors – such as duplicate reporting, ‘missing’ extended care information in this data source, and the utility of diagnostic vs. ‘functional’ information for this patient group – that also should be considered.  At present, however, the DAD represents the only source of diagnostic information for extended care recipients.  Since there is a depth and breadth of diagnosis codes recorded, their elimination would represent a loss of potentially useful information. Health Information Development Unit Centre for Health Services and Policy Research   7  Table 2:  Frequency of ICD9 codes – most responsible diagnosis ICD9 codes Description Frequency Percent 290 Senile and presenile psychotic organic conditions 1326 18.2 V60  Housing, household and economic circumstances 1269 17.4 438 Late effects of cerebrovascular disease 576 7.9 V57 Care involving use of rehabilitation procedures 455 6.2 331 Other cerebral degenerations 364 5.0 342 Hemiplegia 295 4.0 332 Parkinson's disease 235 3.2 797 Senility without mention of psychosis 171 2.3 294 Other organic psychotic conditions (chronic) 163 2.2 V58 Other and unspecified aftercare 158 2.2    Table 3:  Frequency of ICD9 codes - pre-admit comorbidity      ICD9 codes Description Frequency Percent 290 Senile and presenile psychotic organic conditions 525 4.9 250 Diabetes mellitus 444 4.1 428 Heart failure 412 3.8 438 Late effects of cerebrovascular disease 343 3.2 715 Osteoarthrosis and allied disorders 337 3.1 342 Hemiplegia 330 3.1 733 Other disorders of bone and cartilage 277 2.6 311 Depressive disorder, not elsewhere classified 276 2.6 331 Other cerebral degenerations 262 2.4 401 Essential hypertension 262 2.4  Health Information Development Unit Centre for Health Services and Policy Research   8  Table 4:  Frequency of ICD9 codes - post-admit comorbidity      ICD9 codes Description Frequency Percent 599 Other disorders of urethra and urinary tract 544 12.8 486 Pneumonia, organism unspecified 479 11.3 507 Pneumonitis due to solids and liquids 198 4.7 436 Acute but ill-defined cerebrovascular disease 187 4.4 707 Chronic ulcer of skin 141 3.3 428 Heart failure 136 3.2 799 Other ill-defined and unknown causes of morbidity and mortality 113 2.7 410 Acute myocardial infarction 112 2.6 485 Bronchopneumonia, organism unspecified 99 2.3 780 General symptoms 82 1.9  Table 5:  Frequency of ICD9 codes - secondary       ICD9 codes Description Frequency Percent 401 Essential hypertension 865 6.1 438 Late effects of cerebrovascular disease 713 5.0 041 Bacterial infection in conditions classified elsewhere and of unspecified site 651 4.6 250 Diabetes mellitus 608 4.3 V10 Personal history of malignant neoplasm 432 3.1 414 Other forms of chronic ischemic heart disease 394 2.8 733 Other disorders of bone and cartilage 321 2.3 599 Other disorders of urethra and urinary tract 313 2.2 V04 Need for prophylactic vaccination and innoculation against certain viral diseases 275 1.9 427 Cardiac dysrhythmias 266 1.9  Health Information Development Unit Centre for Health Services and Policy Research   9  Table 6:  Frequency of ICD9 codes - overall       ICD9 codes Description Frequency Percent 290 Senile and presenile psychotic organic conditions 1977 5.3 438 Late effects of cerebrovascular disease 1632 4.4 V60  Housing, household and economic circumsances 1328 3.6 401 Essential hypertension 1137 3.0 250 Diabetes mellitus 1130 3.0 599 Other disorders of urethra and urinary tract 1017 2.7 428 Heart failure 936 2.5 331 Other cerebral degenerations 783 2.1 342 Hemiplegia 752 2.0 041 Bacterial infection in conditions classified elsewhere and of unspecified site 690 1.8  Table7:  Frequency of UCOD codes, Vital Statistics deaths records      ICD9 codes Description Frequency Percent 436 Acute but ill-defined cerebrovascular disease 304 9.7 486 Pneumonia, organism unspecified 256 8.2 414 Other forms of chronic ischemic heart disease 246 7.8 410 Acute myocardial infarction 187 6.0 290 Senile and presenile psychotic organic conditions 141 4.5 428 Heart failure 116 3.7 331 Other cerebral degenerations 115 3.7 298 Other non-organic psychoses 109 3.5 496 Chronic airways obstruction, not elsewhere classified 102 3.3 250 Diabetes mellitus 98 3.1 Health Information Development Unit Centre for Health Services and Policy Research   10 Appendix – Definition of diagnosis types  Diagnosis type Description Most responsible The diagnosis considered by the physician to be the most responsible for the patient’s stay in hospital Pre-admit comorbidity A diagnosis that describes another important condition of the patient which usually has a significant influences on the patient’s length of stay Post-admit comorbidity A diagnosis that describes a condition arising after the beginning of hospital observation and/or treatment and has an influence on the length of stay Secondary A diagnosis that describes a condition for which a patient may or may not have received treatment but did not significantly contribute to the patient’s length of stay in hospital E-Code A diagnosis that supplements any ICD9 codes between 800.0 and 999.9 inclusive. To be coded immediately following the respective 800-999 codes. Drug codes (960-989.9) are not required to have an E-code diagnosis. Source: BC Ministry of Health, Annotated Specifications for Patient Hospitalization Data, April 1, 1997-March 31, 1998.  Note: There are other diagnosis types, but this list represents those that are required to be submitted.       Acknowledgements: Thanks to Brian McCashin and Stephen Lee for assistance with data preparation and analysis and to Jim Cruickshank for comments on a draft.  This work was supported by the Ministry of Health.  Data were kindly provided by the BC Ministry of Health, through the BC Linked Health Data set, and by the BC Vital Statistics Agency. 

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:
https://iiif.library.ubc.ca/presentation/dsp.52383.1-0048484/manifest

Comment

Related Items