UBC Faculty Research and Publications

Recognizing difficult trade-offs: values and treatment preferences for end-of-life care in a multi-site… Howard, Michelle; Bansback, Nick; Tan, Amy; Klein, Doug; Bernard, Carrie; Barwich, Doris; Dodek, Peter; Nijjar, Aman; Heyland, Daren K Dec 6, 2017

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

Item Metadata


52383-12911_2017_Article_570.pdf [ 608.27kB ]
JSON: 52383-1.0361567.json
JSON-LD: 52383-1.0361567-ld.json
RDF/XML (Pretty): 52383-1.0361567-rdf.xml
RDF/JSON: 52383-1.0361567-rdf.json
Turtle: 52383-1.0361567-turtle.txt
N-Triples: 52383-1.0361567-rdf-ntriples.txt
Original Record: 52383-1.0361567-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessRecognizing difficult trade-offs: values andtreatment preferences for end-of-life carein a multi-site survey of adult patients infamily practicesMichelle Howard1* , Nick Bansback2,3, Amy Tan4, Doug Klein5, Carrie Bernard6, Doris Barwich7, Peter Dodek3,8,Aman Nijjar9 and Daren K. Heyland10,11,12AbstractBackground: Decisions about care options and the use of life-sustaining treatments should be informed by aperson’s values and treatment preferences. The objective of this study was to examine the consistency of ratings ofthe importance of the values statements and the association between values statement ratings and the patient’sexpressed treatment preference.Methods: We conducted a multi-site survey in 20 family practices. Patients aged 50 and older self-completed aquestionnaire assessing the importance of eight values (rated 1 to 10), and indicated their preference for use of life-sustaining treatment (5 options). We compared correlations among values to a priori hypotheses based on whetherthe value related to prolonging or shortening life, and examined expected relationships between importance ofvalues and the preference option for life-sustaining treatment.Results: Eight hundred ten patients participated (92% response rate). Of 24 a priori predicted correlations amongvalues statements, 14 were statistically significant but nearly all were negligible in their magnitude and some werein the opposite direction than expected. For example, the correlation between importance of being comfortableand suffering as little as possible and the importance of living as long as possible should have been inverselycorrelated but was positively correlated (r = 0.08, p = 0.03). Correlations between importance of values items andpreference were negligible, ranging from 0.03 to 0.13.Conclusions: Patients may not recognize that trade-offs in what is most important may be needed whenconsidering the use of treatments. In the context of preparation for decision-making during serious illness, decisionaids that highlight these trade-offs and connect values to preferences more directly may be more helpful thanthose that do not.BackgroundWhen people are faced with the need to make decisionsabout health care during serious illness or near the end-of-life, they may not be adequately prepared. Mostpeople who are approaching the end of life have a pointafter which they shift their preferences from a treatmentand cure orientation and express a preference for non-invasive treatment aimed at improving quality of life andsymptoms [1]. However this may not be clear to healthcare providers. Treatments given or ordered are oftendiscordant with this preference and may be more inten-sive than desired [2–4]. Advance care planning (ACP)has been proposed as one solution to improve patient-centred care and ensure that the values and preferencesof patients are used to direct health care decisions. ACPis defined as “a process that supports adults at any ageor stage of health in understanding and sharing theirpersonal values, life goals, and preferences regarding* Correspondence: mhoward@mcmaster.ca1Department of Family Medicine, McMaster University, David Braley HealthSciences Centre, 100 Main Street West, 5th floor, Hamilton, ON L8P 1H6,CanadaFull list of author information is available at the end of the article© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Howard et al. BMC Medical Informatics and Decision Making  (2017) 17:164 DOI 10.1186/s12911-017-0570-xfuture medical care” [5]. In prospective and randomizedtrials, ACP significantly improves outcomes including in-creased likelihood that clinicians and families under-stand and comply with a patient’s wishes [6–9], reduceshospitalization at the end of life, results in less intensivetreatments at the end of life (according to patients’wishes) and increases use of hospice services [2, 10].In some western countries approximately half of adultssurveyed in the community have engaged in some aspectsof ACP such as thinking about their wishes and whowould make treatment decisions on their behalf if theywere to become mentally incapable [11, 12], howeverthese activities alone may be inadequate. Advance careplanning has been equated with the completion of ad-vance directive documents. However, the considerationand communication of values and preferences in prepar-ation for future decisions made by patients or their substi-tute decision-makers in consultation with health careproviders, are also key aspects of ACP [13]. Furthermore,advance directives have limited utility because they con-tain instructions that may not apply to the specific situ-ation of the patient when the decision is needed [14] andthey do not have legal standing in some jurisdictions.The success of ACP rests on the assumptions that pa-tients have adequately communicated their values toothers who will make specific treatment decisions ontheir behalf in the future and that those values truly re-flect what is most important to the patient at the time ofdecision-making. However research has shown that sub-stitute decision-makers are often unclear about the pa-tient’s treatment preferences [15], preferences sometimeschange over time [16] and patients’ expression of theirvalues can be inconsistent, making them unhelpful aspreparation for discussion of treatment decisions withhealth care providers [17]. A sub-optimal process ofclarifying values may contribute to difficulties with ACP.Structured interventions and decision aids have beenrecommended as tools to improve the process and out-comes of ACP [18–20]. However, a systematic reviewthat examined decision aids for end-of-life decisionsfound that patient values and treatment goals were notwell addressed in the tools, limiting the usefulness ofsome decisions aids for ACP [21].Tools for patients such as decision aids are importantin ACP because people can construct preferences onlywhen they are faced with choices [22]. Furthermore,consumer theory suggests that patients must recognizetrade-offs between features of options so that they canclarify the relative importance, or value, of each [23].One of the key components of decision-making duringserious illness is, for example, the patient’s values regard-ing maximizing quality versus quantity of life. Well-designed decision aids can support the alignment ofdecisions with values and preferences of the patient ifthey help patients recognize trade-offs and link the pa-tient’s values to preferences in a reliable and transparentway [23–25]. If values are not clarified and values andpreferences are not clearly linked, a poor decision-makingprocess may arise leading to uncertainty about decisions.As part of a survey on engagement in ACP amongpatients in primary care, we asked about values andpreferences for life-sustaining treatments using previ-ously published questions and taxonomies [26–28].The objective of this study was to examine theconsistency of ratings of the importance of the valuesstatements and the association between values state-ment ratings and the patient’s expressed treatmentpreference, to understand whether and how thesevalues statements could be useful in guiding delibera-tions with health care providers about use of life-sustaining treatments.MethodsSetting and designWe conducted a cross-sectional study in primary carepractices between October 2014 and March 2015 as partof a larger project to improve the quantity and quality ofACP discussions in primary care practices. The methodsof the main study that provided these data have been de-scribed previously [29]. The study was conducted in aconvenience sample of 20 family physicians’ practices inCanada: 13 practices in the province of Ontario, five inAlberta, and two in British Columbia. The family prac-tices were located in the same communities as membersof the research team and were invited by physicianmembers of the research team. Included practices had adefined patient population (versus episodic walk-inclinics) and provided comprehensive family medicineservices (as opposed to a focused practices such as psy-chotherapy or sports medicine). A variety of practicetypes were intentionally invited. All were located inurban centres, three were teaching clinics and 17 werecommunity-based private practices, 17 employed alliedhealth professionals, and 17 were practices with morethan one physician in the same office.ParticipantsStaff or clinicians in the family practices invited con-secutive eligible patients to speak to the research assist-ant. Eligible patients were 50 years of age and older,could read and speak/understand English, and did nothave cognitive impairment that would limit participa-tion. The referring staff member or clinician knew theage of the patient and determined whether the patientwas cognitively suitable but a formal assessment of cog-nitive impairment was not done. The research assistantmet with the patients in a private space in the clinic andexplained the study and obtained informed consent.Howard et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 2 of 9QuestionnaireWe adapted a questionnaire that was previously devel-oped and validated for use in hospitalized patients [27].The hospital version has face and content validity andgood ratings of clarity and low emotional burden [30].We modified this questionnaire for use in primary careand piloted it with 25 patients in primary care to assessits clarity, sensibility, and acceptability. Revisions weremade to improve clarity. These questions addressed thepatient’s engagement in ACP (see Additional file 1).Our definitions of values and preferences are takenfrom a consensus-based framework on end-of-life com-munication and decision-making [31]. ‘Values’ refer toan expression of a person’s overarching philosophies ormost important priorities in life (such as maximizingquality of life or time spent with family). Preferencesrefer to specific preferred options for treatments (suchas use of resuscitative treatments or not) or preferredhealth states (conditions that would be acceptable or un-acceptable) [17]. To assess the values of respondents, weused the End of Life (EOL) Values Scale [26], a multi-item questionnaire that asks the respondent to rate thelevel of importance (1 = not important;10 = extremelyimportant) of each item in a list of common values thatare relevant to EOL care (Table 3). The values question-naire was based on a previous studies on EOL values incommunity dwelling elderly individuals [26, 28], andadapted through expert input and pilot testing with ser-iously ill patients in hospital [15] and in primary care[29]. To elicit preferences for use of life-sustaining treat-ments, we used a taxonomy of different levels of the useor non-use of life-sustaining treatments (shown inTable 2). This taxonomy was developed with inputs frommedical experts and has been used extensively in ourprevious research [27, 32]. The questionnaire also in-cluded socio-demographic items and the brief clinicalfrailty scale with text description and pictographs onwhich patients were asked to rate themselves [33]. Pa-tients self-completed the questionnaire.Sample size and statistical analysisA sample size of 30-50 patients was targeted in eachfamily practice to attempt to obtain a representativesample for each practice, because the overall study wasdesigned to provide baseline data to each practice forthe purpose of quality improvement.Characteristics of patients were described as counts andpercentages for categorical variables and as means withstandard deviations and ranges for continuous variables.Median and inter-quartile ranges of the ratings on the EOLValues Scale were reported because the data were skewed.To assess the internal consistency of respondents’ statedvalues, we calculated the Pearson correlation coefficientsamong the various value statements in the EOL valuesscale. Most items in the EOL Values Scale are associatedwith either prolongation or shortening of life. We used thesame a priori hypotheses regarding which values would becorrelated and in which direction as in a previous paperthat used the same rating scales [17]. In that paper the hy-potheses were formulated by the authors based on whetherthe value statement related to prolonging or shortening oflife. For example two values statements that related toprolonging life such as wanting to live as long as possibleand the belief that life should be preserved were expectedto be strongly positively correlated. The survey in thecurrent study included an additional value statement abouthospital avoidance and two authors (MH, DKH) con-structed the a priori hypotheses for this question.Similar to the values statements wherein a high rating ofimportance suggests a desire for either prolongation orshortening of life, the treatment preference items were ori-ented to either prolongation or shortening of life. We there-fore hypothesized that higher importance rating of valuesaligned to life prolongation would be associated with treat-ment preferences involving life-prolonging measures (be-yond comfort care only), and vice versa. Correlations weredescribed in terms of size using the rule of thumb of Hinkle(0.9-1.0 = very high; 0.7-0.9 = high; 0.5-0.7 =moderate; 0.3-0.5 = low; 0.0-0.3 = negligible) [34]. For the analysis of asso-ciation between values and preferences we first used thefive categories of response related to preferences; then, tosimplify the options, we repeated the analysis after collaps-ing these responses to three categories: full measures in-cluding cardiopulmonary resuscitation (CPR), machines ormedical treatments but not CPR, and comfort care only.Patients were excluded from analysis if they answered ‘un-sure’ or did not respond to the item. Box plots were usedto illustrate the distribution of the importance of statedvalues by the preference option for life-sustaining treatmentchosen. The concordance between respondents’ statedvalues and their chosen preference for use of life sustainingtreatments was measured by Kendall’s τ-b statistic. Ken-dall’s τ-b is the excess concordance (C-(1-C)) where C isthe probability that an individual who puts a higher import-ance on a given value related to prolonging life than some-one who put a lower importance on the same value.Kendall’s τ-b ranges from 1 to −1 where 1 indicates perfectconcordance, −1 indicates perfect discordance and 0 indi-cates no association. We first calculated the statistic usingthe five categories of preference, and based on visual in-spection of the boxplots, also calculated with three categor-ies of preference where the middle categories between ‘allpossible measures including CPR’ and ‘comfort care only’(the extreme points) were collapsed to a single category.Ethics and consentEthics approval for the study was obtained from the Re-search Ethics Board of each participating institutionHoward et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 3 of 9(Hamilton Integrated Research Ethics Board [14-582],Queen’s University Health Sciences and AffiliatedTeaching Hospitals Research Ethics Board [6013436],University of British Columbia Office of Research Ethics[H14-02836], University of Alberta Health Research Eth-ics Board [Pro00051082]). Verbal informed consent wasobtained from participants.ResultsThe survey was completed by 810 patients (92% completionrate among patients approached by family practice staff).The number of patients recruited per family practice rangedfrom 23 to 140 (in one multi-physician practice approxi-mately 20 patients per physician were recruited). The meanage of patients was 66 years (range 50-95 years) and 56%were female (449/809). Most patients were Caucasian (88%;713/810). The median score on the clinical frailty scale was2 (corresponding to ‘well’).There was a ceiling effect, that is the median ratingwas 10 out of 10 for three values statements: being com-fortable and suffering as little as possible, having moretime with family, and that death is not prolonged. Thelowest rated statements were for the belief that lifeshould be preserved and avoiding hospitalization (me-dian rating = 6) (Table 1). Across all items 11% to 13% ofthe statements were answered as ‘unsure’.Of the 24 a priori predicted correlations between im-portance ratings for value statements, 14 were statisti-cally significant in the expected direction but 11 of thosewere negligible correlations (defined as r < 0.3) (Table 2).For example, we hypothesized that there would be astrongly positive correlation between the importance ofbeing comfortable and suffering as little as possible andthe importance of avoiding machines, and between theimportance of the belief that death should not be pro-longed and the importance of the belief that nature beallowed to take its course. However, the correlationswere negligible (0.24, 0.19 respectively). Of the 12 corre-lations we expected to be negative, four were in factsignificantly positive; for example there was a positivecorrelation between the importance of living as long aspossible and being comfortable and suffering as little aspossible (r = 0.08, p = 0.036). We hypothesized that thecorrelation between the importance of avoiding tubesand machines and the belief that life should be preservedwould be strongly negative, and found that it was negli-gible (−0.03, p = 0.494). Similarly, we expected the cor-relation between the importance of living as long aspossible and avoiding machines to be strongly negativeand observed no significant results (r = 0.02, p = 0.518)The largest of any correlation was (moderate): the cor-relation between living as long as possible and havingmore time with family (r = 0.64, p < 0.001).Most patients preferred the treatment option thatexcluded CPR (92%). Nearly one-third of patients pre-ferred ‘comfort measures only’ (30%), 7% preferredfull medical care without resuscitation or breathingmachines, and 36% preferred short term use of ma-chines but not CPR, switching to comfort measuresor all possible measures but without CPR (4%). Only8% preferred all possible measures including CPR,and 15% were unsure or did not answer the question(Table 3).The Kendall’s τ-b for the correlation between the im-portance rating for values and the five-option preferencequestion (1 = all measures including CPR, 5 = comfortcare only, no CPR; increasing number means less use oflife sustaining treatments) were negligible, ranging from0.03 to 0.13. The highest correlations between valuesand preferences were found for ‘nature allowed to takeits course’, ‘death not prolonged’, and ‘life should be pre-served’ (Kendall’s τ-b = 0.12, 0.13 and −0.12 respectively,all statistically significant). These correlations were inthe expected direction where a higher importance ratingof values related to not prolonging life correlated posi-tively with the choice on the preference scale and valuesassociated with prolonging life correlated negatively withthe preference item.Table 1 Rating of importance of each issue in thinking about the kinds of medical treatments wanted or not wanted nearthe end of lifeItem: How important … median(q1,q3) (n) Unsure Missing or declined % (n/N)I be comfortable and suffer as little as possible? 10(9,10) (793) 13% (100/793) 2% (15/793)I have more time with my family? 10(8,10) (785) 13% (102/785) 2% (15/785)I live as long as possible? 8(5,10) (777) 12% (97/777) 2% (15/777)I avoid being attached to machines and tubes? 9(7,10) (753) 11% (83/753) 2% (14/753)My death is not prolonged? 10(8,10) (750) 11% (81/750) 2% (14/750)Belief that nature should be allowed to take its course? 8(5,10) (760) 12% (89/760) 2% (14/760)Belief that life should be preserved? 6(3,8) (736) 11% (83/736) 2% (13/736)I respect the wishes of other family members? 7(4,10) (778) 12% (97/778) 2% (14/778)I avoid hospitalization? 6(3,9) (762) 12% (90/762) 2% (13/762)Howard et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 4 of 9On visual inspection of the figures, the mean import-ance ratings of values statements discriminated better be-tween the preference for ‘all possible measures includingCPR’ and ‘comfort care only’ (the extreme points) butthere was less variation across the other three options forcare. After collapsing to three categories for the preferencequestion: all measures including CPR, collapsing the op-tions for life-prolonging treatments without CPR, andcomfort care only, the correlations increased but remainednegligible, ranging from 0.11 to 0.29 (Fig. 1).DiscussionIn this multi-site cross-sectional survey of approximately800 adult patients in the family practice setting, patients’ratings of a set of values statements and response to aquestion about treatment preference demonstrated thatthere is a ceiling effect for some values statements, theratings of some values statements are inconsistent or il-logical when compared to each other, and the ratingsgiven to the importance of some values statements areminimally associated with the expected treatment prefer-ence. Taken together, these findings suggest that patientsare not able to discriminate values that compete with ortrade-off with each other and have trouble linking theirvalues statements to preferences for medical treatments.Patient decision aids have been proposed as a way toimprove decision-making relating to serious illness orend of life [20, 35–37] and values clarification is a cen-tral component of these decision aids [24]. Our resultssuggest that patients do not recognize that trade-offsmay be required with respect to receiving health carethat would be consistent with values. For example iftrade-offs are recognized we would expect a negativecorrelation between the importance ratings of living aslong as possible and being as comfortable and sufferingas little as possible. However both items had a medianimportance rating of 10 out of 10, and the correlationbetween them was weak. There is growing literaturesupporting the need for approaches to designing ques-tions such as ranking, constant sum, and discrete choiceexperiments that ask patients to specifically compare therelative importance of each issue [38, 39]. It is possiblethat asking patients to consider the importance of eachissue on its own, without considering any trade-offs, didnot necessarily elicit informed values from individuals.The decision theory literature suggests values only be-come informed and stable when individuals have the op-portunity to trade off one aspect with another andreflect on those trade-offs [40, 41]. Values can then beTable 3 Preferences for care if life supports were needed tostay aliveResponse item % (n)All possible measures including resuscitation (CPR) 8 (65)All possible measures, no resuscitation. 4 (35)Machines only in the short term, change focus to comfortmeasures, no resuscitation (CPR) or breathing machines.36 (294)Use full medical care, no resuscitation (CPR) or breathingmachines7 (55)Comfort measures only 30 (239)Unsure 13 (106)Missing or declined 2 (16)CPR cardiopulmonary resuscitationTable 2 Correlations between different values pertaining to decisions about life-sustaining treatmentsa b c d e f g h ia) How important is it that I be comfortable and suffer as little as possible?(shorten)1.00 0.110.00220.08↓0.03640.24↑<.00010.28↑<.00010.09↑0.0130−0.04↓0.28820.050.14530.120.0011b) How important is it that I have more time with my family? (prolong) 1.00 0.64↑<.0001−0.03↓0.3449−0.08↓0.03810.070.05210.37<.00010.43<.00010.09↓0.0129c) How important is it that I live as long as possible? (prolong) 1.00 0.02↓0.5183−0.12↓0.00090.10↓0.00660.52↑<.00010.43<.00010.13↓0.0003d) How important is it that I avoid being attached to machines and tubes?(shorten)1.00 0.46↑<.00010.16↑<.0001−0.03↓0.49430.020.55620.29↑<.0001e) How important is it that my death is not prolonged? (shorten) 1.00 0.19↑<.0001−0.14↓0.00020.030.48680.19↑<.0001f) How important is a belief that nature should be allowed to take itscourse? (shorten)1.00 0.18<.00010.15↓<.00010.14↑0.0002g) How important is the belief that life should be preserved? (prolong) 1.00 0.38<.00010.04↓0.3300h) How important is it that I respect the wishes of other family membersregarding my care? (neutral)1.00 0.17<.0001i) How important is it that I avoid hospitalization? (neutral) 1.00Association between value statements and likely influence on length of life indicated in brackets (prolong, shorten, neutral)A priori, correlations indicated ↑ were expected to be strongly positive and correlations indicated ↓ were expected to be strongly negativeP values are written under correlation coefficientHoward et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 5 of 9determined by considering the amount one is willing tosacrifice from something else [42]. There appears a rolefor these questions to not just elicit, but also help clarifypatients’ values, and help them recognize the trade-offsin end of life care, Understanding how a patient priori-tizes his/her values, may then allow the clinician toguide the patient to further prioritize decision options ifthey are faced with decisions whose outcomes will notalign with all of their values. Further research should ex-plore whether these explicit questions can overcome theproblems of ceiling effects and inconsistency amongvalues ratings that arose with our results.We used a taxonomy for preferences regarding life-sustaining treatments that provided five options, three ofwhich were for various degrees of medical care withoutCPR that would attempt to prolong life to some extent.The patterns of values statement ratings across these threemedical care options suggested that patients may not haveunderstood the differences between the nuanced treatmentapproaches because the correlations with values statementswere very low and improved when the taxonomy was col-lapsed to combine these three categories. For patients toexpress a preference for a treatment approach and appreci-ate the connection to their values, attention to languageand level of detail regarding the goals of the treatment ap-proaches are needed.The findings of this study could be helpful in designingdecision aids that elicit the values that are most strongly as-sociated with treatment preferences and therefore supportthe process of ACP because they are informative to pa-tients, their substitute decision-makers and clinicians. Forexample, our results have demonstrated that some valuesin our values questionnaire such as being comfortable andsuffering as little as possible, and having more time withfamily would not be helpful to elicit in a clinical setting be-cause of a ceiling effect in importance ratings. The valuesstatements that might have utility were the importance ofliving as long as possible, the belief that death not beFig. 1 Box-plots of relationship between importance ratings for values and preference option for life-sustaining treatment, CPR –cardiopulmonary resuscitation. 1 = Use comfort measures only with a focus on improving my quality of life and comfort. Allow naturaldeath and no artificial prolongation of life and no resuscitation. 2, 3, 4 = combined categories of: Use full medical care to prolong mylife but if my heart or my breathing stops, no resuscitation (CPR) or breathing machine; Use machines only in the short term to see ifI will get better but if my illness is prolonged, change focus to comfort measures only. If my heart stops, no resuscitation (CPR); Usemachines and all possible measures with a focus on keeping me alive but if my heart stops, no resuscitation. 5 = Use machines andall possible measures including resuscitation (CPR) with a focus on keeping me alive at all costsHoward et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 6 of 9prolonged and the belief that nature be allowed to take itscourse or that life should be preserved, because these werethe items with the greatest association with preferences forlife sustaining treatments. In this study, values and prefer-ences were assessed at one point in time and the questionof the usefulness of considering them in advance arises.Studies have reported that some patients change their pref-erences once they experience a health condition [16, 43],and some have difficulty predicting how they will feel aboutfuture health states [44]. Thus, it is important that the pa-tient’s values and preferences be revisited with their clini-cians on an ongoing basis, and in particular if healthdeteriorates. Prior values clarification exercises still haveutility because the process may help patients to be in-formed and more able to confront trade-offs so they can ar-rive at better decisions than if they are thinking about theseissues for the first time, or the patient may lose capacitymaking it necessary to rely on past expressions of values.Preferences for health care options can remain stable evenwith changing experience of health care [45], and the ma-jority of patients have stable preferences relating to end oflife decisions, especially those who have completed an ad-vance directive [16].Our finding that 13% of patients are uncertain abouttheir preferences regarding use of life-sustaining treat-ments may not be surprising given that the sample ofgenerally healthy patients with an average age of 66 maynot have given much prior consideration to these issues.However, similar findings have been reported in ser-iously ill older adults in hospital who might be expectedto have considered decisions about life-sustaining treat-ments [17]. Similar to other findings in older adults wholive in the community [7], only 8% of patients in thecurrent study expressed a preference to receive CPR, Itmay be that patients do not understand the implicationsof their preference regarding CPR [46], therefore there isan imperative to ensure high quality and timely commu-nication and decision-making. Guideline-recommendedelements of goals of care discussions that are desired bypatients in hospital, such as discussions about valueshappen infrequently [3].A strength of this study is that we surveyed patientsin multiple family practices across regions. It is a limi-tation that the values statements instrument we usedwas not designed for the purpose of this study- to as-sess consistence of values and compare values to treat-ment preferences A more constrained technique withranking of values may increase the recognition oftrade-offs. It is possible that in this population of pa-tients from primary care settings aged 66 years on aver-age and who were not acutely ill, the responses toquestions that are relevant to decisions during seriousor critical illness may differ from responses in the fu-ture when faced with the situation. In addition, familypractices and patients were not randomly sampled andwe cannot be certain of the generalizability of the studyresults. Finally, our participants were predominantlyCaucasian and had a higher level of education; resultsmay not apply to people who are not Caucasian or whohave lower educational attainment.ConclusionThe reported values and preferences related to use oflife-sustaining treatments among primary care patientsreflect considerable inconsistency and discordance. Toengage in high quality ACP, patients will need assistanceto recognize possible trade-offs among their values andto understand the relationship between their values andtheir treatment preferences. The results demonstrate aneed to improve the processes and tools to support ad-vance care planning. These efforts should specifically in-clude approaches that recognize trade-offs among valuesand link values explicitly to treatment preferences.Additional fileAdditional file 1: Questionnaire asking patients about engagement inadvance care planning. (DOCX 37 kb)AbbreviationsACP: Advance care planning; CPR: Cardiopulmonary resuscitation; EOL: Endof lifeAcknowledgementsThe authors gratefully acknowledge Neha Arora, Dawn Elston, Nola Fuller,Rebecca Heyland, Larry Mroz and Shelley Wrona for their role in datacollection.FundingThe study was funded by the Canadian Frailty Network (formerly TVN) whichis supported by the Government of Canada through the Networks ofCentres of Excellence program) and the Ontario Ministry of Health and Long-Term Care. Opinions are those of the authors and do not imply endorsementfrom the funding sources.Availability of data and materialsAll data generated or analyzed during this study are included in thispublished article.Authors’ contributionsMH, DKH contributed to the conception and design of the study. MHdrafted the article. All authors contributed to interpretation of the analyses,critical revision of the manuscript and gave final approval of the version tobe published.Ethics approval and consent to participateEthics approval for the study was obtained from the Research Ethics Boardof each participating institution (Hamilton Integrated Research Ethics Board[14-582], Queen’s University Health Sciences and Affiliated Teaching HospitalsResearch Ethics Board [6013436], University of British Columbia Office ofResearch Ethics [H14-02836], University of Alberta Health Research EthicsBoard [Pro00051082]). Verbal informed consent was obtained fromparticipants.Consent for publicationNot applicable.Howard et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 7 of 9Competing interestsThe authors declare that they have no competing interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Department of Family Medicine, McMaster University, David Braley HealthSciences Centre, 100 Main Street West, 5th floor, Hamilton, ON L8P 1H6,Canada. 2School of Public and Population Health at the University of BritishColumbia, Vancouver, BC, Canada. 3Centre for Health Evaluation andOutcome Sciences, St Paul’s Hospital, Vancouver, BC, Canada. 4Department ofFamily Medicine, University of Calgary, Calgary, Alberta, Canada. 5Departmentof Family Medicine, University of Alberta, Edmonton, Alberta, Canada.6Department of Family and Community Medicine, University of Toronto,Toronto, Canada. 7Division of Palliative Care, Department of Medicine,University of British Columbia, Vancouver, British Columbia, Canada. 8Divisionof Critical Care Medicine, St. Paul’s Hospital and University of BritishColumbia, Vancouver, British Columbia, Canada. 9General Internal Medicine,Faculty of Medicine, St. Paul’s Hospital, University of British Columbia,Vancouver, British Columbia, Canada. 10Department of Critical Care Medicine,Queen’s University, Kingston, Ontario, Canada. 11Clinical Evaluation ResearchUnit, Kingston General Hospital, Queen’s University, Kingston, Ontario,Canada. 12Department of Public Health Sciences, Queen’s University,Kingston, Ontario, Canada.Received: 15 August 2017 Accepted: 28 November 2017References1. Heyland DK, Dodek P, Rocker G, Groll D, Gafni A, Pichora D, et al. Whatmatters most in end-of-life care: perceptions of seriously ill patients andtheir family members. CMAJ. 2006;174:627–33.2. Teno JM, Fisher ES, Hamel MB, Coppola K, Dawson NV. Medical careinconsistent with patients’ treatment goals: association with 1-year Medicareresource use and survival. J Am Geriatr Soc. 2002;50:496–500.3. You JJ, Dodek P, Lamontagne F, Downar J, Sinuff T, Jiang X, et al. Whatreally matters in end-of-life discussions? Perspectives of patients in hospitalwith serious illness and their families. CMAJ. 2014;186:E679–87.4. Heyland DK, Ilan R, Jiang X, You JJ, Dodek P. The prevalence of medicalerror related to end-of-life communication in Canadian hospitals: results ofa multicentre observational study. BMJ Qual Saf. 2015;25:671–9.5. Sudore RL, Lum HD, You JJ, Hanson LC, Meier DE, Pantilat SZ, et al. Definingadvance care planning for adults: a consensus definition from amultidisciplinary Delphi panel. J Pain Symptom Manag. 2017;53:821–32.6. Hammes BJ, Rooney BL. Death and end-of-life planning in one midwesterncommunity. Arch Intern Med. 1998;158:383–90.7. Silveira MJ, Kim SY, Langa KM. Advance directives and outcomes ofsurrogate decision making before death. N Engl J Med. 2010;362:1211–8.8. Detering KM, Hancock AD, Reade MC, Silvester W. The impact of advancecare planning on end of life care in elderly patients: randomised controlledtrial. BMJ. 2010;340:c1345.9. Metzger M, Song M-K, Ward S, Chang PP-Y, Hanson LC, Lin F-C. Arandomized controlled pilot trial to improve advance care planning forLVAD patients and their surrogates. Heart Lung J Acute Crit Care. 2016;45:186–92.10. Houben CHM, Spruit MA, Groenen MTJ, Wouters EFM, Janssen DJA. Efficacyof advance care planning: a systematic review and meta-analysis. J Am MedDir Assoc. 2014;15:477–89.11. Teixeira AA, Hanvey L, Tayler C, Barwich D, Baxter S, Heyland DK. What doCanadians think of advanced care planning? Findings from an onlineopinion poll. BMJ Support Palliat Care. 2015;5:40–7.12. Fried TR, Redding CA, Robbins ML, Paiva A, O’Leary JR, Iannone L. Stages ofchange for the component behaviors of advance care planning. J AmGeriatr Soc. 2010;58:2329–36.13. McMahan RD, Knight SJ, Fried TR, Sudore RL. Advance care planningbeyond advance directives: perspectives from patients and surrogates. JPain Symptom Manag. 2013;46:355–65.14. Sudore RL, Fried TR. Redefining the “planning” in advance care planning:preparing for end-of-life decision making. Ann Intern Med. 2010;153:256.15. Abdul-Razzak A, Heyland DK, Simon J, Ghosh S, Day AG, You JJ. Patient-familyagreement on values and preferences for end-of-life values and medicaltreatment PreferencesLife-sustaining treatment: results of a multi-centerobservational study. BMJ Support Palliat Care. doi: 10.1136/bmjspcare-2016-001284. [Epub ahead of print]16. Auriemma CL, Nguyen CA, Bronheim R, Kent S, Nadiger S, Pardo D, et al.Stability of end-of-life preferences. JAMA Intern Med. 2014;174:1085.17. Heyland DK, Heyland R, Dodek P, You JJ, Sinuff T, Hiebert T, et al.Discordance between patients’ stated values and treatment preferences forend-of-life care: results of a multicentre survey. BMJ Support Palliat Care.2016;7:292–9. doi: 10.1136/bmjspcare-2015-001056.18. Butler M, Ratner E, McCreedy E, Shippee N, Kane RL. Decision aids foradvance care planning: an overview of the state of the science. Ann InternMed. 2014;161:408–18.19. Sudore RL, Knight SJ, McMahan RD, Feuz M, Farrell D, Miao Y, et al. A novelwebsite to prepare diverse older adults for decision making and advancecare planning: a pilot study. J Pain Symptom Manag. 2014;47:674–86.20. Cox CE, White DB, Abernethy AP. A universal decision support system.Addressing the decision-making needs of patients, families, and clinicians inthe setting of critical illness. Am J Respir Crit Care Med. 2014;190:366–73.21. Cardona-Morrell M, Benfatti-Olivato G, Jansen J, Turner RM, Fajardo-PulidoD, Hillman K. A systematic review of effectiveness of decision aids to assistolder patients at the end of life. Patient Educ Couns. 2017;100:425–35.22. Lichtenstein S, Slovic P, editors. The construction of preference. Cambrdige:Oxford University Press; 2006.23. Shiell A, Hawe P, Seymour J. Values and preferences are not necessarily thesame. Health Econ. 1997;6:515–8.24. Elwyn G, O’Connor A, Stacey D, Volk R, Edwards A, Coulter A, et al.Developing a quality criteria framework for patient decision aids: onlineinternational Delphi consensus process. Br Med J. 2006;333:417.25. Abhyankar P, Bekker HL, Summers BA, Velikova G. Why values elicitationtechniques enable people to make informed decisions about cancer trialparticipation. Health Expect. 2011;14(Suppl 1):20–32.26. Winter L, Parks SM. Family discord and proxy decision makers’ end-of-lifetreatment decisions. J Palliat Med. 2008;11:1109–14.27. Heyland D, Pichora D, Dodek P, Lamontagne F, You J, Barwich D, et al. Thedevelopment and validation of a questionnaire to audit advance careplanning. J Palliat Care Med. 2012;2:119. doi: 10.4172/2165-7386.1000119.28. Winter L. Patient values and preferences for end-of-life treatments: arevalues better predictors than a living will? J Palliat Med. 2013;16:362–8.29. Howard M, Bernard C, Klein D, Tan A, Slaven M, Barwich D, et al. Olderpatient engagement in advance care planning in Canadian primary carepractices: results of a multi-site survey. Can Fam Physician. 2017; in press30. Heyland DK, Pichora D, Dodek P, Lamontagne F, You JJ, Barwich D, Tayler C,Porterfield P, Simon J, Enns B. The development and validation of aquestionnaire to audit advance care planning. J Palliat Care Med. 2012;2:175–6.31. Sinuff T, Dodek P, You JJ, Barwich D, Tayler C, Downar J, et al. Improvingend-of-life communication and decision making: the development ofconceptual framework and quality indicators. J Pain Symptom Manag. 2015;49:1070–80.32. Heyland DK, Barwich D, Pichora D, Dodek P, Lamontagne F, You J, et al.Failure to engage hospitalized elderly patients and their families in advancecare planning. JAMA Intern Med. 2013;173:778.33. Rockwood K, Song X, MacKnight C, Bergman H, Hogan DB, McDowell I, etal. A global clinical measure of fitness and frailty in elderly people. CMAJ.2005;173:489–95.34. Hinkle D, Wiersma W, Jurs S. Applied statistics for the behavioral sciences.5th ed. Boston: Houghton Mifflin; 2003.35. Green MJ, Levi BH. Development of an interactive computer program foradvance care planning. Health Expect. 2009;12:60–9.36. Sudore RL, Boscardin J, Feuz MA, McMahan RD, Katen MT, Barnes DE. Effectof the PREPARE website vs an easy-to-read advance directive on advancecare planning documentation and engagement among veterans. JAMAIntern Med. 2017;177:1102–9.37. Oczkowski SJ, Chung H-O, Hanvey L, Mbuagbaw L, You JJ. Communicationtools for end-of-life decision-making in ambulatory care settings: asystematic review and meta-analysis. PLoS One. 2016;11:e0150671.38. Pignone MP, Howard K, Brenner AT, Crutchfield TM, Hawley ST, LewisCL, et al. Comparing 3 techniques for eliciting patient values fordecision making about prostate-specific antigen screening. JAMA InternMed. 2013;173:362.Howard et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 8 of 939. Bansback N, Li LC, Lynd L, Bryan S. Development and preliminary usertesting of the DCIDA (dynamic computer interactive decision application)for ‘nudging’ patients towards high quality decisions. BMC Med InformDecis Mak. 2014;14:62.40. Simon S, Howe L, Kirschenbaum H. Values clarification: a handbook of practicalstrategies for teachers and students. New York: Hart Publishing Co.; 1975.41. Fischhoff B. Values elicitation: is there anything there? In: Hechter M, NadelL, Michod R, editors. The origin of values. New York: Aldine De Gruyter;1993. p. 187–214.42. Hargreaves-Heap S, Hollis B, Lyons B, Sugden R, Weale A. The theory ofchoice: a critical guide. Oxford: Blackwell; 1992.43. Fried TR, O’Leary J, Van Ness P, Fraenkel L. Inconsistency over time in thepreferences of older persons with advanced illness for life-sustainingtreatment. J Am Geriatr Soc. 2007;55:1007–14.44. Fried TR, Byers AL, Gallo WT, Van Ness PH, Towle VR, O’Leary JR, et al.Prospective study of health status preferences and changes in preferencesover time in older adults. Arch Intern Med. 2006;166:890–5.45. San Miguel F, Ryan M, Scott A. Are preferences stable? The case of healthcare. J Econ Behav Organ. 2002;48:1–14.46 Heyland DK, Frank C, Groll D, Pichora D, Dodek P, Rocker G, et al. Understandingcardiopulmonary resuscitation decision making. Chest. 2006;130:419–28.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Howard et al. BMC Medical Informatics and Decision Making  (2017) 17:164 Page 9 of 9


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items