Patients or volunteers? The impact of motivation for trial participation on the efficacy of patient decision Aids: a secondary analysis of a Cochrane systematic review.

ORIGINAL ARTICLE

Patients or Volunteers? The Impact of Motivation for Trial Participation on the Efficacy of Patient Decision Aids: A Secondary Analysis of a Cochrane Systematic Review James G. Brown, MSc, Kerry E. Joyce, PhD, Dawn Stacey, PhD, Richard G. Thomson, MD

Background. Efficacy of patient decision aids (PtDAs) may be influenced by trial participants’ identity either as patients seeking to benefit personally from involvement or as volunteers supporting the research effort. Aim. To determine if study characteristics indicative of participants’ trial identity might influence PtDA efficacy. Methods. We undertook exploratory subgroup meta-analysis of the 2011 Cochrane review of PtDAs, including trials that compared PtDA with usual care for treatment decisions. We extracted data on whether participants initiated the care pathway, setting, practitioner interactions, and 6 outcome variables (knowledge, risk perception, decisional conflict, feeling informed, feeling clear about values, and participation). The main subgroup analysis categorized trials as ‘‘volunteerism’’ or ‘‘patienthood’’ on the basis of whether participants initiated the care pathway. A supplementary subgroup analysis categorized trials on the basis of whether any volunteerism factors were present (participants had not initiated the care pathway, had attended a research setting, or had

Received 12 November 2013 from the Institute of Health & Society, Newcastle University, Newcastle upon Tyne, United Kingdom (JGB, KEJ, RDT), and Ottawa Hospital Research Institute, University of Ottawa, Ottawa, Canada (DS). This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Revision accepted for publication 12 February 2015. Address correspondence to James Brown, Newcastle University, Institute of Health & Society, Baddiley Clark Building, Richardson Road, Newcastle upon Tyne, NE2 4AX, UK; telephone: +44 (0)191 208 7187; fax: +44 (0)191 222 8211; e-mail: [email protected]. Ó The Author(s) 2015 Reprints and permission: http://www.sagepub.com/journalsPermissions.nav DOI: 10.1177/0272989X15579172

a face-to-face interaction with a researcher). Results. Twenty-nine trials were included. Compared with volunteerism trials, pooled effect sizes were higher in patienthood trials (where participants initiated the care pathway) for knowledge, decisional conflict, feeling informed, feeling clear, and participation. The subgroup difference was statistically significant for knowledge only (P = 0.03). When trials were compared on the basis of whether volunteerism factors were present, knowledge was significantly greater in patienthood trials (P \ 0.001), but there was otherwise no consistent pattern of differences in effects across outcomes. Conclusions. There is a tendency toward greater PtDA efficacy in trials in which participants initiate the pathway of care. Knowledge acquisition appears to be greater in trials where participants are predominantly patients rather than volunteers. Key words: decision making; choice behavior; decision aids; decision support techniques; patient preference; environment; motivation; patient participation. (Med Decis Making 2015;35:419–435)

P

atient decision aids (PtDAs) are tools that facilitate shared decision making by providing evidence-based information on options and all relevant associated outcomes, helping patients to clarify and communicate the personal value they associate with different features of the options and preparing patients for participation in health care decisions.1 The Cochrane systematic review of PtDAs found that, compared with usual care, PtDAs improve knowledge, risk perception of outcomes, and patient-practitioner communication and patient participation, as well as reduce decisional conflict.2 PtDAs may also have a role in preventing use of options that informed patients do not value, without

MEDICAL DECISION MAKING/MAY 2015

Downloaded from mdm.sagepub.com at GEORGETOWN UNIV MED CTR on May 26, 2015

419

BROWN AND OTHERS

adversely affecting health outcomes. However, for most outcomes, there was unexplained statistical heterogeneity in the analyses. Such study-level variation may be artifactual—due to methodological differences between studies—or real owing to differences in study population, intervention, or control.3 For studies of complex interventions such as PtDAs, true study-level variation may be caused by differences in study context.4 Several studies have identified contextual characteristics that may have an influence on the trial efficacy of PtDAs.5–8 One qualitative study in particular has raised questions about the impact of the motivations of participants to take part in research on trial outcomes.9 This study was embedded within a UK trial of a PtDA to help people with nonvalvular atrial fibrillation make decisions about using warfarin or aspirin for stroke prevention.10 The qualitative study found that participants made attributions about their trial identity that fell on a continuum between ‘‘volunteerism,’’ where they identified themselves ‘‘contributing to the production of useful clinical knowledge as a matter of civic responsibility,’’ and ‘‘patienthood,’’ where they ‘‘saw themselves as people receiving personally relevant, individualized medical advice and information through the medium of the trial.’’ In the trial evaluating the PtDA, there were no significant differences between the PtDA group and the control group in decisional conflict or knowledge at 3 months.10 This may have been because several contextual characteristics of the trial—the inclusion of prevalent cases in the sample (in which participants were being asked to reevaluate their decision), a research setting remote from routine practice, and seeing a research doctor who was not the participants’ usual doctor—meant that participants were operating more as volunteers than as patients. If such volunteerism meant that participants were not seeing the PtDA as helping them to make real decisions that they wanted and needed to make, they may have been less receptive to the potential benefits of PtDAs. Trial identity was not specifically measured within the trial. Consequently, an individual-level subgroup analysis to compare efficacy of the PtDA between ‘‘volunteers’’ and ‘‘patients’’ within the trial itself was not possible. However, if other trials can be similarly categorized as having participants who were operating predominantly as volunteers or as patients, subgroup analysis could be undertaken at the trial level.11–13 The aim of this study was to determine if differences between PtDA trials in contextual characteristics related to participants’ trial identity might influence the trial efficacy of PtDAs. See Box 1.

420

METHODS Study Design We undertook exploratory, post hoc subgroup meta-analyses of randomized controlled trials included in the 2011 Cochrane systematic review of PtDAs.14 Detailed methods used in the Cochrane review itself are not repeated here, particularly the literature search and study selection criteria. The original review can also be consulted for further information on each study, including the participants, intervention and control characteristics, risk of bias assessments, and other elements of setting not described here.14 Study Selection Eligible studies were those included in the 2011 Cochrane review and that compared PtDAs with usual care (control) for people making treatment decisions. We excluded trials comparing more detailed with simpler PtDAs and trials that involved decisions about screening. Study-Level Moderator Variables To categorize trials with respect to participants’ trial identity, we sought study characteristics that would indicate participants acting as patients (seeking to benefit personally from involvement) or volunteers (consenting in order to support the research effort for altruistic reasons). The chosen moderator variables (effect modifiers) were initiation of pathway of care (whether participants were seeking help for a problem or condition), trial setting (whether in a routine clinical practice or in a research setting remote from routine practice, or both), and face-toface interactions with practitioners (with their responsible practitioner or with a research practitioner, or both). The moderator variables were operationalized into questions with categorical answers (yes, no, mixed, or unclear) for use in data extraction and analysis (see Box 2). Because a small number of studies did not report practitioner interactions, an assumption was made that trial participants had a face-to-face interaction with their responsible practitioner at a routine practice setting if 1 month or longer had passed between using the PtDA and outcome measurement. We developed a data extraction form for moderator variables based on the principles set out in the Cochrane Handbook for Systematic Reviews.15 It



PATIENTS OR VOLUNTEERS?

Box 1 Definitions of Key Concepts Trial identity: A trial participant’s motivation for participation in the trial sits on a continuum between benefiting personally (patienthood) and wishing to contribute to the research effort (volunteerism). Patienthood: A category or division of trial identity in which participants wish to benefit personally from participation in the trial by ‘‘receiving personally relevant, individualized medical advice and information through the medium of the trial.’’ Volunteerism: A category or division of trial identity in which participants are motivated to participate for altruistic reasons. They are ‘‘contributing to the production of useful clinical knowledge as a matter of civic responsibility.’’ Definitions derived from Heaven B, Murtagh M, Rapley T, et al. Patients or research subjects? A qualitative study of participation in a randomised controlled trial of a complex intervention. Patient Educ Couns. 2006;62(2):260–70.

Box 2

Questions about Moderator Variables in Data Collection Form

1. Initiation of pathway of care Did the trial participants initiate the pathway of care because they were seeking help for a problem or condition? Yes is indicative of patienthood. No is indicative of volunteerism. 2. Setting: routine practice Did the trial participants attend a routine clinical practice setting (defined as a clinic or other provider facility that the patient would have attended regardless of whether he or she was taking part in a research trial) at any time between randomization and outcome measurement? Yes is indicative of patienthood. No is indicative of volunteerism. 3. Setting: research practice Did the trial participants attend a research setting remote from or external to a routine clinical practice setting at any time between randomization and outcome measurement? Yes is indicative of volunteerism. No is indicative of patienthood. 4. Face-to-face interaction: responsible practitioner Did the trial participants have a face-to-face interaction with their responsible practitioner (defined as a health care practitioner whom the patient would have consulted regardless of whether he or she was taking part in a research trial) at any time between randomization and outcome measurement? Yes is indicative of patienthood. No is indicative of volunteerism. 5. Face-to-face interaction: research practitioner Did the trial participants have a face-to-face interaction with a research practitioner (who was not also their responsible practitioner) at any time between randomization and outcome measurement? Yes is indicative of volunteerism. No is indicative of patienthood. See Box 1 for definitions of patienthood and volunteerism.

was pilot-tested on a sample of 10 trials (by JB) and modified in response to difficulties observed in operationalizing the moderator variables. Data extraction was undertaken by 2 investigators (JB, KJ) independently, and disagreements were resolved by discussion. If there was no consensus, a third investigator (RT) was asked to arbitrate. Outcome Variables Data were derived from the 2011 Cochrane review for each trial on selected outcome variables, including knowledge, accurate risk perceptions, total decisional conflict (decisional conflict scale, DCS), feeling informed (uninformed subscale of the DCS), feeling clear about values (unclear subscale of the DCS), and participation in decision making.14 These outcomes were chosen on the basis of being International Patient Decision Aid Standard (IPDAS) criteria for evaluating the effectiveness of PtDAs16 and primary outcomes in the Cochrane review,14 while having enough trials to undertake subgroup analysis (a minimum of 6 trials).17 Analysis We conducted subgroup meta-analyses using Cochrane Review Manager Version 5.1.18 Effect measures and methods of analysis mirrored those used in the Cochrane review. For continuous data, mean differences were analyzed using the inverse variance method, and for dichotomous outcomes, pooled relative risks were calculated using the Mantel-Haenszel method. We employed a random-effects model and calculated an I2 test of statistical heterogeneity for each subgroup and a test of subgroup differences. Unadjusted P values were reported with cautious interpretation of the results as opposed to using a correction for multiple comparisons19; the null hypothesis of no difference in efficacy between subgroups was rejected only if there was a consistent, statistically significant difference across most of the included outcomes variables. If heterogeneity (I2) within both subgroups reduced to less than 25%,20 this was seen as further corroboration for accepting the alternative hypothesis but not a requirement. We did not attempt to state the magnitude of (difference in) effect size for impact of trial identity that would be considered meaningful because this was an exploratory study. Main subgroup analysis—initiation of pathway of care The main subgroup analysis compared trials in which participants initiated the pathway of care

ORIGINAL ARTICLE

421


BROWN AND OTHERS

leading to the decision, because they were clearly seeking help for the index problem or condition, to trials in which participants did not. This study characteristic was considered likely to have the greatest influence on trial identity because it would have shaped participants’ motivations for participation prior to consent. The subgroup of trials in which all participants initiated the pathway of care was labeled patienthood. The subgroup of trials in which some or all participants did not initiate the pathway of care was labeled volunteerism. Trials were excluded from the analysis if it was unclear whether participants initiated the pathway of care. Supplementary subgroup analysis—any volunteerism factor A supplementary subgroup analysis was undertaken to incorporate the study characteristics of trial setting and practitioner face-to-face interactions, as well as whether participants initiated the pathway of care. In this subgroup analysis, a comparison was made between trials with any volunteerism factor present (labeled volunteerism) and those with no volunteerism factors present (labeled patienthood) (see Figure 1). A volunteerism factor was defined as any of the following: a) some or all of the participants did not initiate the pathway of care leading to the decision (not clearly seeking help for a problem or condition) when they were enrolled into the trial; b) trial participants (in the PtDA arm) attended a research setting remote from, or external to, routine clinical practice at some time between randomization and outcome measurement; or c) trial participants (in the PtDA arm) had a face-to-face interaction with a research practitioner (who was not also their responsible practitioner) at some time between randomization and outcome measurement. If trial participants attended both a research setting and a routine practice setting or had a face-to-face interaction with their responsible practitioner and a research practitioner, the trial was allocated to the volunteerism subgroup. Methods of subgroup allocation are illustrated in Figure 1. The method of subgroup allocation (see Figure 1) was influenced by 2 premises. First, we considered that whether participants initiated the pathway of care would have greater influence on trial identity than setting and practitioner interactions, because this variable would exert its influence prior to consent. Second, we considered that any face-to-face interaction would increase the impact of a PtDA compared with no face-to-face interaction. Therefore, we excluded studies in which some or all participants

422

did not have a face-to-face practitioner interaction except for those in which some or all participants in the trial had not initiated the pathway of care (they were not seeking help for a problem or condition). Two studies were consequently included in which it was unclear whether participants had a face-toface interaction with a practitioner between randomization and outcome measurement but where participants did not initiate the pathway of care.21,22 Direct comparisons between trials for the setting and interaction variables were not performed for 2 reasons. First, in 45% of trials, participants attended both routine practice and research settings or had face-to-face interactions with both their responsible practitioner and a research practitioner, or the answer was unclear, or there was a mix of participants (e.g., participants had a choice of settings). Second, of 14 possible comparisons, none had more than 2 trials in the volunteerism subgroup, and 5 comparisons had fewer than 6 trials considered a priori to be the minimum number needed to undertake subgroup analysis. Sensitivity analysis In both the main analysis and the supplementary analysis, trials with a mix of participants, in which some participants initiated the pathway of care and some did not, were allocated to the ‘‘volunteerism’’ subgroup. Sensitivity analysis was undertaken to reanalyze the results when studies with a mix of participants were excluded or, in the supplementary analysis, when such a mix of participants was no longer considered a ‘‘volunteerism factor’’ such that trials were only allocated to the volunteerism subgroup if another volunteerism factor was present. Funding This research received no specific grant from any funding agency in the public, commercial, or notfor-profit sectors. RESULTS Results of Data Extraction (Moderator Variables) Of 86 trials included in the 2011 Cochrane review, a total of 54 trials pertained to treatment decisions, of which 29 trials were included in the subgroup analyses (see Figure 2).5,8,10,21–46 A total of 32 trials were excluded because they pertained to screening decisions, and 25 trials were excluded because either they compared detailed with simpler PtDAs or they presented data that could not be pooled. No further




Did some or all trial participants initiate the pathway of care because they were seeking help for a problem or condition?

Yes

Unclear

No

Did all trial participants have a face-to-face interaction with a practitioner at any time between randomization and outcome measurement?

Yes

No/unclear

Did all trial participants either (a) attend a research setting remote from or external to a routine clinical practice setting, or (b) have a face-toface interaction with a research practitioner (who was not also their responsible practitioner), at any time between randomization and outcome measurement?

Yes

No Exclude

Volunteerism subgroup

Patienthood subgroup

Figure 1 Supplementary subgroup analysis (any volunteerism factor): subgroup allocation.

trials were excluded from the main subgroup analysis (initiation of pathway of care), but 4 trials were excluded from the supplementary subgroup analysis (any volunteerism factor) because some or all participants did not have a face-to-face practitioner interaction.24,32,35,42 Table 1 indicates the decision topic for each included study. For the 29 trials included in the subgroup analyses, there were 15 categories of trials (see Table 2). In 5 studies, the answer was ‘‘unclear’’ for 1 or more data extractions questions (see Box 2 and Table 2).21,22,30,35,37 None of these studies were excluded from the analyses on this basis. In 3 of 5 studies, the answer to 2 or more questions was ‘‘unclear.’’21,22,30 The only question for

which the answer was not ‘‘unclear’’ (as judged by both reviewers) in any study was ‘‘Did the trial participants initiate the pathway of care because they were seeking help for a problem or condition?’’ One trial fell into more than one category because outcomes were measured at different times during follow-up.40 In this trial of a PtDA to help women make decisions about treatment of menorrhagia, knowledge was measured at 6 months, whereas decisional conflict was assessed at 2 weeks. Participants would almost certainly have seen their responsible practitioner prior to knowledge being measured, but it was assumed that this would not have occurred prior to the measurement of decisional conflict.

ORIGINAL ARTICLE

423


BROWN AND OTHERS

86 PtDA trials included in Cochrane review

29 PtDA trials included in subgroup analyses

Main subgroup analysis (initiation of pathway of care): 29 trials

Supplementary subgroup analysis (any volunteerism factor): 25 trials

Knowledge

17 trials Excluded: 0

14 trials Excluded: 3 trials

Accurate risk perceptions


8 trials Excluded: 1 trial

Total decisional conflict



Feeling informed



Feeling clear about values



Participation in decision making


8 trials Excluded: 1 trial

Outcomes

Excluded: 32 trials of screening and 25 trials only comparing detailed to simpler PtDAs or only presenting data that could not be pooled for the outcomes of interest

Figure 2 Flow diagram of progress of trials of patient decision aids (PtDAs) from Cochrane review to subgroup analyses.

Agreement between reviewers was excellent (k . 0.75) for 3 of 5 questions pertaining to moderator variables. Agreement was good for the question on whether participants had a face-to-face interaction with their responsible practitioner (k = 0.65) and fair for the questions on whether the participants attended a routine practice setting (k = 0.57). After discussion, there was consensus for all but one trial for which a third reviewer was asked to arbitrate.35 Results of Subgroup Analyses Main subgroup analysis—initiation of pathway of care For most outcomes (knowledge, total decisional conflict, feeling unclear, feeling uninformed, and

424

participation), pooled effect sizes (favoring PtDAs over usual care) were higher in patienthood trials where all participants initiated the care pathway than in volunteerism trials (see Table 3). The subgroup difference was significant only for the outcome of knowledge: mean difference 15.97% (95% confidence interval [CI], 12.29 to 19.64; 15 trials) in the patienthood subgroup, compared with 5.21% (95% CI, 23.49 to 13.91; 2 trials) in the volunteerism subgroup (P = 0.03) (see Figure 3). Statistical heterogeneity remained high in both subgroups. No significant subgroup differences were found for the 5 other outcome variables. In the sensitivity analysis, the subgroup difference for the outcome of total decisional conflict became statistically significant when a trial (of a PtDA on




Table 1

Topic or Decision in Included Patient Decision Aid Trials

Study First Author/Year/Reference

Topic/Decision

Auvinen 200423 Barry 199724 Bernstein 199825 Davison 199726 Dodin 200127 Dunn 199828 Johnson 200629 Kasper 200830 Laupacis 200631 Le´gare´ 200832 Man-Son-Hing 19995 McAlister 200521 McBride 200222 Montgomery 200333 Montgomery 200734 Morgan 200035 Mullan 200936 Murray 200138 Murray 200137 Nassar 200746 O’Connor 199839 Protheroe 200740 Shorten 200541 Thomson 200710 Vandemheen 200942 Vodermaier 20098 Whelan 200343 Whelan 200444 Wong 200645

Prostate cancer treatment Benign prostate disease treatment Revascularization for ischemic heart disease Prostate cancer treatment Hormone replacement therapy Infant vaccination schedule Endodontic treatment Immunotherapy for multiple sclerosis Preoperative autologous blood donation Natural health products for menopausal symptoms Atrial fibrillation treatment Atrial fibrillation treatment Hormone replacement therapy Hypertension treatment Birthing options after previous caesarean Revascularization for ischemic heart disease Diabetes treatment Benign prostate disease treatment Hormone replacement therapy Management of breech presentation Hormone replacement therapy Menorrhagia treatment Birthing options after previous caesarean Atrial fibrillation treatment Lung transplant in cystic fibrosis Breast cancer surgery Breast cancer chemotherapy Breast cancer surgery Pregnancy termination method

hormone replacement therapy)37 in the volunteerism subgroup that had a mix of participants (some of whom had initiated the pathway of care and others who had not) was either excluded from the analysis (P = 0.001) or reallocated to the patienthood subgroup (P\0.001). The results were not substantially altered for any other outcomes. Supplementary subgroup analysis—any volunteerism factor The pattern of differences in effect between trials with or without volunteerism factors was inconsistent across outcomes (see Table 4). Mean difference in knowledge (favoring PtDA over usual care) was significantly higher in the patienthood subgroup (18.75%; 95% CI, 12.06–25.43; 6 trials) than in the volunteerism subgroup (9.88%; 95% CI, 6.94–12.82; 8 trials; test for subgroup differences, P = 0.02). Statistical heterogeneity remained high in both subgroups (see Figure 4). For the outcome of feeling uninformed, the magnitude of the mean difference (suggesting greater PtDA

efficacy) was significantly greater in the volunteerism subgroup (29.16; 95% CI, 213.46 to 24.87; 8 trials) than in the patienthood subgroup (23.58; 95% CI, 26.75 to 20.40; 2 trials; test for subgroup differences, P = 0.04) (see Figure 5). Statistical heterogeneity remained high in the volunteerism subgroup (85%). No significant subgroup differences were observed for the outcome of total decisional conflict. The supplementary subgroup analysis either could not be performed, or the results would be of limited value, for the outcomes of feeling unclear about values, accurate risk perceptions, and participation in decision making because all but one or fewer trials were allocated to a single subgroup. The sensitivity analysis did not result in any substantial change in the results. DISCUSSION For almost all outcomes, PtDA efficacy was higher in trials in which participants initiated the pathway of care (which we assumed indicated that they were

ORIGINAL ARTICLE

425


BROWN AND OTHERS

Table 2 Categories and Subgroup Allocation of Trials Following Data Extraction for Moderator Variables Questiona Category

1. Participant 2. Routine Practice 3. Research 4. Responsible 5. Research Initiated Setting Setting Practitioner Practitioner

1 2 3 4 5 6 7 8

No No No Mixed Mixed Mixed Yes Yes

Unclear Unclear Yes No Unclear Yes No No

No Unclear No Yes Unclear Yes Yes No

Unclear Unclear Yes No Unclear Yes No No

No Unclear No Yes Unclear Unclear Yes Yes

9 10 11

Yes Yes Yes

Mixed Yes Yes

Mixed Yes No

No Yes Yes

Yes Yes Yes

12

Yes

No

No

No

No

13 14 15

Yes Yes Yes

Yes Yes Yes

No No No

Mixed No Yes

Unclear No No

Study First Author and Year (Outcomes) 21

McAlister 2005 McBride 200222 Man-Son-Hing 19995 Thomson 200710 Kasper 200830 Murray 200137 Dodin 200127 O’Connor 1998,39 Protheroe 2007 (DCS)40 Montgomery 200333 Murray 200138 Davison 1997,26 Laupacis 2006,31 Montgomery 2007,34 Nassar 2007,46 Protheroe 2007 (K),40 Vodermaier 2009,8 Whelan 200343 Le´gare´ 2008,32 Vandemheen 200942 Morgan 200035 Barry 199724 Auvinen 2004,23 Bernstein 1998,25 Dunn 1998,28 Johnson 2006,29 Mullan 2009,36 Shorten 2005,41 Whelan 2004,44 Wong 200645

Subgroup Analysis A

B

V V V V V V P P

V V V V V V V V

P P P

V V V

P

X

P P P

X X P

A, initiation of pathway of care (subgroup analysis); B, any volunteerism factor (subgroup analysis); V, volunteerism subgroup; P, patienthood subgroup; X, excluded; DCS, decisional conflict scale; K, knowledge. a. See Box 2 for key to questions.

seeking help for a problem or condition). Knowledge acquisition associated with PtDA use was statistically significantly greater in ‘‘patienthood’’ trials than in ‘‘volunteerism’’ trials in both subgroup analyses. However, there was considerable variation in results across outcomes when trials with no volunteerism factors were compared with those with one or more volunteerism factors, and 3 subgroup analyses could not be performed because of too few studies. Statistically significant differences favoring patienthood over volunteerism were not observed consistently across outcomes in either subgroup analysis. These findings suggest that, in PtDA trials, whether participants initiate the pathway of care may be an important effect modifier, and participants’ trial identity may exert an important influence

426

on knowledge acquisition but not necessarily on other outcomes. The influence specifically of setting and the context of interactions with practitioners remains unclear. We speculated that the influence of participant initiation of the pathway of care was mediated through its effect on engagement and motivation in decision making. Adult learning literature suggests that adults are more likely to learn when engaged and motivated to do so.11,12 Indeed, motivation to seek and engage with information has been found to be a key influence on information exchange and shared decision making in health care consultations.13 In a study comparing the results of 2 randomized controlled trials of a web-based decision support for perimenopausal and postmenopausal women making decisions about



427


10 3 11

Patienthood Volunteerism Unclear values subscale

826 745

1243 860 1571

1606 880 2103

2050 884 2486

887 1156 2934

2691 382 2043

3073

0.59 (0.42, 0.83) 0.73 (0.46, 1.16)

27.09 (211.55, 22.62) 23.03 (24.91, 21.14) 0.61 (0.47, 0.78)

28.80 (213.22, 24.38) 25.54 (27.49, 23.59) 25.75 (28.65, 22.85)

27.34 (29.77, 24.91) 23.74 (26.59, 20.90) 28.06 (211.08, 25.04)

1.59 (1.33, 1.89) 2.05 (1.27, 3.32) 26.55 (28.63, 24.46)

15.97 (12.29, 19.64) 5.21 (23.49, 13.91) 1.77 (1.45, 2.15)

14.58 (10.92, 18.24)

Effect Estimate (95% CI)

44 0

90 0 23

82 0 86

68 63 78

39 86 73

82 86 70

86

P = 0.47

P = 0.10

P = 0.19

P = 0.06

P = 0.33

P = 0.03

Heterogeneity Test for Subgroup (I2), % Differences (x2)

IV, inverse variance; Random, random effects; CI, confidence interval; M-H, Mantel-Haenszel; PtDA, patient decision aid. a. Practitioner-controlled decision making.

6 3

13 3 13

Patienthood Volunteerism Uninformed subscale

Patienthood Volunteerism

6 3 16

Patienthood Volunteerism Total decisional conflict score

8 3 9

15 2 9

Patienthood Volunteerism Accurate risk perceptions

Patienthood Volunteerism Participationa

17

Studies Participants (n) (n)

Knowledge

Outcome or Subgroup

Risk ratio (M-H, random, 95% CI)

Mean difference (IV, random, 95% CI)




Mean difference (IV, Random, 95% CI)

Statistical Methods

Patienthood

Patienthood

Patienthood

Patienthood

Volunteerism

Patienthood

Subgroup with Higher PtDA Efficacy

Table 3 Results of Main Subgroup Analysis: Whether Participants Initiated the Pathway of Care (29 Trials)

BROWN AND OTHERS

Figure 3 Main subgroup analysis (initiation of pathway of care): knowledge.

Figure 4 Supplementary subgroup analysis (any volunteerism factor): knowledge.

428



429


1 7

Patienthood Volunteerism

203 1188

367 1706 1311

473 2048 2073

201 1693 2521

1052 1465 1864

2517

0.41 (0.30, 0.57) 0.72 (0.52, 0.99)

23.58 (26.75, 20.40) 29.16 (213.46, 24.87) 0.59 (0.44, 0.79)

24.95 (27.48, 22.42) 27.50 (210.41, 24.58) 27.87 (211.39, 24.36)

1.34 (1.10, 1.63) 1.80 (1.43, 2.27) 26.75 (29.09, 24.42)

18.75 (12.06, 25.43) 9.88 (6.94, 12.82) 1.72 (1.40, 2.11)

13.71 (9.75, 17.67)

Effect Estimate (95% CI)

NA 0

0 85 23

0 82 82

NA 67 77

87 63 70

87

P = 0.02

P = 0.04

P = 0.20

P = 0.05

P = 0.02

Heterogeneity Test for Subgroup Differences (x2) (I 2), %






Statistical Method

IV, inverse variance; Random, random effects; CI, confidence interval; M-H, Mantel-Haenszel; NA, not applicable; PtDA, patient decision aid. a. Practitioner-controlled decision making.

2 8 8

3 10 10

Patienthood Volunteerism Uninformed subscale

Patienthood Volunteerism Participationa

1 7 13

Patienthood Volunteerism Total decisional conflict score

6 8 8

14

Studies Participants (n) (n)

Patienthood

Volunteerism

Volunteerism

Volunteerism

Patienthood

Subgroup with Higher PtDA Efficacy

Results of Supplementary Subgroup Analysis: Whether a Volunteerism Factor Was Present (25 Trials)

Patienthood Volunteerism Accurate risk perceptions

Knowledge

Outcome or Subgroup

Table 4

BROWN AND OTHERS

Figure 5 Supplementary subgroup analysis (any volunteerism factor): uninformed subscale of decision conflict score.

hormone replacement, there was some suggestion that the efficacy of the decision support related to whether women were actively trying to make a decision and the urgency of that decision.7 Another study found that some outcomes related to decision making are affected by whether patients with cancer were actively seeking out information.47 There are 2 reasons why this is an important finding. The first is the impact on the generalizability (external validity) of the findings from trials of PtDAs. If volunteers are less motivated to make decisions or seek out information, or less engaged in shared decision making, PtDA trials that include volunteers as opposed to patients may underestimate the true effectiveness of PtDA in practice. This is also important because of the health literacy and equity implications associated with the use of decision support tools. We know that outcomes related to decision making are influenced by patient factors, including age, sex, and education.48–50 If certain groups are less motivated or engaged in decision making (e.g., more socioeconomically disadvantaged groups), individuals in these groups may be less willing to use PtDAs and less likely to experience the benefits compared with more motivated or engaged groups. This would suggest that widespread implementation of shared decision making and PtDAs could increase health inequalities. However, people from more disadvantaged groups can be informed

430

and engaged in health care decisions provided they have access to well-designed tools and well-trained staff and have most to gain from PtDAs.51 PtDAs may be more effective in those with lower knowledge.52 A recent systematic review also found that interventions designed to support shared decision making may reduce health inequalities.53 Another study has shown that PtDAs can narrow the gap between different racial groups considering surgery for osteoarthritis of the knee.54 Yet, McCaffery et al.55 reported that few existing PtDAs consider the impact of health literacy in their design, and the effects of those PtDAs designed for low health literacy groups remain poorly understood. Only 3 of 97 trials reviewed explicitly considered the needs of low health literacy individuals; those that did demonstrated increases in both knowledge and informed choice. Although we did not set out to examine the equity implications of PtDAs, our study draws attention to the importance of contextual factors on the efficacy of PtDAs and thus reinforces the need to monitor the impact of policies to implement shared decision making and PtDAs in practice on inequalities. The greater influence of patienthood on knowledge acquisition, compared with other outcomes, may also be explained by ‘‘patients’’ having greater engagement and motivation in using a PtDA compared with ‘‘volunteers.’’ Adult learning theory is




more applicable to knowledge acquisition than other outcomes.11 Although it might be expected that the outcome of feeling informed would be influenced in the same way as knowledge, perceptions of knowledge do not correlate well with scores on knowledge tests.56 Furthermore, volunteers are less likely to need to feel informed compared with patients who may feel a stronger need to be informed. Along with participation, decisional conflict, and feeling clear about values, feeling informed is a process variable, whereas knowledge is an outcome variable.14 Process variables represent perceptions and feelings.57 The perception of participation may be related to prior experiences of participation in health care decisions.50 Perceptions are intimately related to expectations; poorer perceived outcomes occur when expectations are not met.58 We might reasonably envisage higher expectations of the benefits of PtDAs in trials whose participants identify themselves primarily as patients—and directly ‘‘help seeking’’—as opposed to volunteers.9 Risk perception also differs in nature from knowledge acquisition. Both patients and professionals are known to have considerable difficulty in processing and accurately evaluating probabilities.59 It may be that research staff in trials whose participants attend a research setting (remote from routine practice) or interact with a research practitioner face-to-face are more skilled in risk communication than nonresearch clinical staff. Volunteers may also be more educated than patients and therefore more accurately score probabilities. The lack of statistically significant differences favoring patienthood over volunteerism for many variables assessed may partly be explained by a lack of power owing to the relatively small number of eligible trials for each outcome variable. In addition, the influence of participants’ trial identity on PtDA efficacy may have been obscured by the presence of many other contextual factors acting as effect modifiers and their unequal distribution in trials between subgroups. This is suggested by the persistence of statistical heterogeneity within subgroups. Another study that used subanalysis to investigate heterogeneity found no evidence that the type of control intervention is an important effect modifier; however, baseline knowledge of outcome probabilities was found to be an important variable in explaining heterogeneity, with greater effects when baseline knowledge is lower.52 Additional contextual characteristics that have been suggested as having an influence on the trial efficacy of PtDAs include prior health care received (resulting in a ceiling effect)5 and the degree

of trust in the attending clinician.8 Other contextual factors that are likely to be important include whether the decision is permanent (e.g., surgery) or more temporary (e.g., medication), timing of the use of the PtDA (in relation to a consultation or decision) or timing of outcome measurement, whether the setting is primary or secondary care, or whether the practitioner is a doctor, nurse, or other health care professional. A possible explanation for the lack of consistent subgroup differences in favor of patienthood trials in the supplementary subgroup analysis is that the presence of a single volunteerism factor, such as research setting or a face-to-face interaction with a research practitioner, may not exert sufficient influence on the perceived trial identity of participants for a difference in outcomes to be observed, particularly when it is ‘‘competing’’ with the presence of other patienthood factors. Indeed, labeling any one trial as ‘‘volunteerism’’ or ‘‘patienthood’’ is likely to oversimplify the complex interplay of factors that influence participants’ trial identity. There is likely to be marked variation in trial identity between participants within each trial, as well as fluidity of trial identity over time within participants during the trial.9 Participation in any randomized controlled trial necessitates some degree of volunteerism. The conceptualization of trial identity as a dichotomous variable, at the level of either the participant or the trial, is therefore an oversimplification. Heaven and colleagues9 identified most participants as being on a continuum between patienthood and volunteerism. Other research that has explored reasons for trial participation has generally found that altruism and selfinterest combine to act as motivators for trial participation.60,61 Ideally, the influence of participants’ perception of trial identity on efficacy would be determined by directly measuring this variable during the trial and looking for any interaction with treatment effects. This would be akin to a fully randomized preference trial.62 However, there would be problems with this approach. First, although tools to assess the personality trait of altruism have been developed,63,64 they have not been externally validated and may not be suitably responsive to change.61,65 Second, trial identity is a shifting concept during a trial requiring periodical measurement.9 One solution may be to actually undertake a patient preference trial, using level of preference for the use of a PtDA as a proxy measure of trial identity. The ethical requirement to ensure informed consent when enrolling participants in a randomized

ORIGINAL ARTICLE

431


BROWN AND OTHERS

controlled trial (in which allocation preferences can play no part) may ‘‘intensify the tendency of participants to view clinical research studies as removed from routine care’’ and so dilute the efficacy of an intervention.9 This is likely to be particularly relevant to trials of complex interventions such as PtDAs, where clinical encounters are being simulated. Patient preference trials alongside randomized controlled trials may again offer a solution. This study had a number of strengths. Although exploratory in nature, this is new research that (to our knowledge) has not previously been undertaken. Furthermore, it linked phenomenological and positivist approaches to scientific enquiry.66 The hypothesis for this study was (in part) generated by a qualitative study evaluating the processes by which participants made treatment decisions within a randomized controlled trial.9 This hypothesis was then tested empirically in this study using quantitative analysis. Agreement between reviewers when extracting data about contextual characteristics was generally good. Limitations included the large number of comparisons increasing the risk of type I error and the relatively small numbers of trials for each outcome, so increasing the risk of type II error.15 Another important limitation was the use of subgroup analysis by trial characteristics and its inherent risk of bias due to confounding and the presence of effect modifiers besides those of interest.15,67 It is unlikely that meta-regression would have been able to adequately address this issue. Indeed, the study findings serve to illustrate the difficulties in unpacking the multiple and varied aspects of context that might influence effectiveness of PtDAs in practice. Although agreement between reviewers was good, we are cognizant that the evidence to support the use of measures of trial identity was based primarily on face validity, and definitions applied in this study could be conceptualized differently. There may have been classification error due to missing data and the assumptions used in data extraction and classification; this is a clear limitation of our study. For example, it was necessary to make an assumption that trial participants were likely to have had a faceto-face interaction with their responsible practitioner at a routine practice setting if 1 month or longer had passed after using the PtDA. Whether this took place would have been dependent on the nature of the treatment decision. However, this assumption actually made little difference to the study findings: the results of only 4 (of 29) trials were affected by this assumption, and in all 4 trials, outcome measurement was either within 2 weeks of PtDA use or 3 to 12

432

months later.24,25,32,40 This issue further reinforces the contextual nature of shared decision making, which does not occur in isolation but over repeated encounters involving a long-term relationship between patient and provider.68 A similar limitation is that measurement of the patienthood and volunteerism constructs relied on information in the published paper that was not intended for this purpose and therefore may not have been robust. It may also have been of value to have extracted data on follow-up rates to explore whether the impact of trial identity on efficacy may in part be explained by differences in rates of follow-up between patienthood and volunteerism trials. Similarly, we did not compare the risk of bias between studies; this may have been important if trial identity had an impact on placebo effect. However, for almost 80% of studies in the Cochrane review on which we based this subgroup analysis, it was unclear if there was blinding of participants, personnel, or outcome assessors.14 We are also aware that the permanence of the decision is likely to be another important contextual factor that should be addressed in future studies. The Cochrane review on which this study was based has subsequently been updated, but we have no reason to believe that the additional studies included in the update would change our findings. However, as more studies evaluating PtDAs as part of the process of care become available, we may be better able to reconduct this study.

CONCLUSIONS Although this study was unable to determine definitively whether study context relating to participants’ trial identity influences the efficacy of PtDAs, a tendency was observed for increased efficacy in ‘‘patienthood’’ trials in which participants initiated the pathway of care leading to the decision because they were clearly seeking help for a problem or condition, compared with ‘‘volunteerism’’ trials whose participants did not. It remains unclear if differences in trial setting, or whether participants see a research practitioner or their responsible practitioner, affect PtDA efficacy, although they may have an impact on knowledge acquisition if not other outcomes. These results provide further evidence that patient motivation and engagement in decision making are likely to be important for PtDAs to be effective in practice, particularly to increase knowledge. This is




likely to be important when considering barriers and facilitators to implementation of shared decision making, as well as health equity implications. Those implementing shared decision making should monitor for unequal uptake of, or outcomes from, PtDAs between groups. Designers of future PtDA trials should consider whether any contextual factors of the trial affecting motivation for participation or engagement in decision making may affect the generalizability of the findings. A preference trial alongside a randomized controlled trial could be explored. Further consensus is needed to develop a taxonomy of contextual factors for PtDA trials (and their measurement). Future empirical research is recommended to understand the impact of contextual factors in PtDA trials, particularly those related to motivation and engagement in decision making.

REFERENCES 1. Elwyn G, O’Connor A, Stacey D, et al. Developing a quality criteria framework for patient decision aids: online international Delphi consensus process. BMJ. 2006;333(7565):417–24. 2. Stacey D, Le´gare´ F, Col N, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev. 2014(1):CD001431. 3. Glasziou PP, Sanders SL. Investigating causes of heterogeneity in systematic reviews. Stat Med. 2002;21(11):1503–11. 4. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M. Developing and evaluating complex interventions: new guidance. Medical Research Council, 29 September 2008. Available from: www.mrc.ac.uk/complexinterventionsguidance. Accessed 27 March 2014. 5. Man-Son-Hing M, Laupacis A, O’Connor AM, et al. A patient decision aid regarding antithrombotic therapy for stroke prevention in atrial fibrillation. JAMA. 1999;282(8):737–43. 6. Goel V, Sawka CA, Thiel EC, Gort EH, O’Connor AM. Randomized trial of a patient decision aid for choice of surgical treatment for breast cancer. Med Decis Making. 2001;21(1):1–6. 7. Saver B, Gustafson D, Taylor T, et al. A tale of two studies: The importance of setting, subjects and context in two randomized, controlled trials of a web-based decision support for perimenopausal and postmenopausal health decisions. Patient Educ Couns. 2007;66:211–22. 8. Vodermaier A, Caspari C, Koehm J, Kahlert S, Ditsch N, Untch M. Contextual factors in shared decision making: a randomised controlled trial in women with a strong suspicion of breast cancer. Br J Cancer. 2009;100:590–7.

patients with atrial fibrillation: randomised controlled trial. Qual Saf Health Care. 2007;16(3):216–23. 11. Knowles M. The Adult Learner: The Definitive Classic in Adult Education and Human Resource Development. 7th ed. Burlington, MA: Elsevier; 2011. 12. Kreuter MW, Wray RJ. Tailored and targeted communication: strategies for enhancing information relevance. Am J Health Behav. 2003;27(Suppl 3):S227–S32. 13. Edwards M, Davies M, Edwards A. What are the external influences on information exchange and shared decision-making in healthcare consultations? A meta-synthesis of the literature. Patient Educ Couns. 2009;75:37–52. 14. Stacey D, Bennett C, Barry M, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev. 2011;(10):CD001431. 15. Higgins J, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available from: www.cochranehandbook.org. Accessed 12 November 2013. 16. International Patient Decision Aid Standards (IPDAS) Collaboration. IPDAS 2005: Criteria for Judging the Quality of Patient Decision Aids. Available from: http://ipdas.ohri.ca/IPDAS_ checklist.pdf. Accessed 12 November 2013. 17. Fu R, Gartlehner G, Grant M, et al. Conducting Quantitative Synthesis When Comparing Medical Interventions: AHRQ and the Effective Health Care Program. Rockville, MD: Agency for Healthcare Research and Quality. Available from: http://effective healthcare.ahrq.gov/. Accessed 12 November 2013. 18. Review Manager (RevMan) [Computer program]. Version 5.1. Copenhagen, Denmark: The Nordic Cochrane Centre, The Cochrane Collaboration; 2011. 19. Machin D, Campbell M, Walters S. Medical Statistics. 4th ed. Chichester, UK: Wiley; 2007. 20. Hatala R, Keitz S, Wyer P, Guyatt G. Tips for learners of evidence-based medicine: 4. Assessing heterogeneity of primary studies in systematic reviews and whether to combine their results. CMAJ. 2005;172(5):661–5. 21. McAlister FA, Man-Son-Hing M, Straus SE, et al. Impact of a patient decision aid on care among patients with nonvalvular atrial fibrillation: a cluster randomized trial. Can Med Assoc J. 2005;173(5):496–501. 22. McBride CM, Bastian LA, Halabi S, et al. A tailored intervention to aid decisionmaking about hormone replacement theory. Am J Public Health. 2002;92(7):1112–4. 23. Auvinen A, Hakama M, Ala-Opas M, et al. A randomized trial of choice of treatment in prostate cancer: the effect of intervention on the treatment chosen. BJU Int. 2004;93(1):52–6. 24. Barry M, Cherkin D, Chang Y, Fowler F, Skates S. A randomized trial of a multimedia shared decision-making program for men facing a treatment decision for benign prostatic hyperplasia. Dis Manag Clin Outcomes. 1997;1(1):5–14.

9. Heaven B, Murtagh M, Rapley T, et al. Patients or research subjects? A qualitative study of participation in a randomised controlled trial of a complex intervention. Patient Educ Couns. 2006;62(2):260–70.

25. Bernstein SJ, Skarupski KA, Grayson CE, Starling MR, Bates ER, Eagle KA. A randomized controlled trial of information-giving to patients referred for coronary angiography: effects on outcomes of care. Health Expectations. 1998;1(1):50–61.

10. Thomson RG, Eccles MP, Steen IN, et al. A patient decision aid to support shared decision-making on anti-thrombotic treatment of

26. Davison B, Degner L. Empowerment of men newly diagnosed with prostate cancer. Cancer Nurs. 1997;20(3):187–96.

ORIGINAL ARTICLE

433


BROWN AND OTHERS

27. Dodin S, Le´gare´ F, Daudelin G, Tetroe J, O’Connor A. Making a decision about hormone replacement therapy: a randomized controlled trial. Can Fam Physician. 2001;47(8):1586–93. 28. Dunn RA, Shenouda PE, Martin DR, Schultz AJ. Videotape increases parent knowledge about poliovirus vaccines and choices of polio vaccination schedules. Pediatrics. 1998;102(2):e26. 29. Johnson BR, Schwartz A, Goldberg J, Koerber A. A chairside aid for shared decision making in dentistry: a randomized controlled trial. J Dent Educ. 2006;70(2):133–41. 30. Kasper J, Ko¨pke S, Mu¨hlhauser I, Nu¨bling M, Heesen C. Informed shared decision making about immunotherapy for patients with multiple sclerosis (ISDIMS): a randomized controlled trial. Eur J Neurol. 2008;15(12):1345–52. 31. Laupacis A, O’Connor AM, Drake ER, et al. A decision aid for autologous pre-donation in cardiac surgery—a randomized trial. Patient Educ Couns. 2006;61(3):458–66. 32. Le´gare´ F, Dodin S, Stacey D, LeBlanc A, Tapp S. Patient decision aid on natural health products for menopausal symptoms: randomized controlled trial. Menopause Int. 2008;14(3):105–10. 33. Montgomery A, Fahey T, Peters TJ. A factorial randomised controlled trial of decision analysis and an information video plus leaflet for newly diagnosed hypertensive patients. Br J Gen Pract. 2003;53:446–53. 34. Montgomery AA, Emmett CL, Fahey T, et al. Two decision aids for mode of delivery among women with previous caesarean section: randomised controlled trial. BMJ. 2007;334(7607):1305. 35. Morgan MW, Deber RB, Llewellyn-Thomas HA, et al. Randomized, controlled trial of an interactive videodisc decision aid for patients with ischemic heart disease. J Gen Intern Med. 2000; 15(10):685–93. 36. Mullan RJ, Montori VM, Shah ND, et al. The diabetes mellitus medication choice decision aid: a randomized trial. Arch Intern Med. 2009;169(17):1560–8. 37. Murray E, Davis H, Tai SS, Coulter A, Gray A, Haines A. Randomised controlled trial of an interactive multimedia decision aid on hormone replacement therapy in primary care. BMJ. 2001; 323(7311):490. 38. Murray E, Davis H, Tai SS, Coulter A, Gray A, Haines A. Randomised controlled trial of an interactive multimedia decision aid on benign prostatic hypertrophy in primary care. BMJ. 2001; 323(7311):493. 39. O’Connor AM, Tugwell P, Wells GA, et al. Randomized trial of a portable, self-administered decision aid for postmenopausal women considering long-term preventive hormone therapy. Med Decis Making. 1998;18(3):295–303. 40. Protheroe J, Bower P, Chew-Graham C, Peters TJ, Fahey T. Effectiveness of a computerized decision aid in primary care on decision making and quality of life in menorrhagia: results of the MENTIP randomized controlled trial. Med Decis Making. 2007; 27(5):575–84.

43. Whelan T, Sawka C, Levine M, et al. Helping patients make informed choices: a randomized trial of a decision aid for adjuvant chemotherapy in lymph node–negative breast cancer. J Natl Cancer Inst. 2003;95(8):581–7. 44. Whelan T, Levine M, Willan A, et al. Effect of a decision aid on knowledge and treatment decision making for breast cancer surgery. JAMA. 2004;292(4):435–41. 45. Wong SSM, Thornton JG, Gbolade B, Bekker HL. A randomised controlled trial of a decision-aid leaflet to facilitate women’s choice between pregnancy termination methods. Int J Obstet Gynaecol. 2006;113(6):688–94. 46. Nassar N, Roberts CL, Raynes-Greenow CH, Barratt A, Peat B. Decision Aid for Breech Presentation Trial C: evaluation of a decision aid for women with breech presentation at term: a randomised controlled trial. Int J Obstet Gynaecol. 2007;114(3):325–33. 47. Martinez L, Schwartz JS, Freres D, Fraze T, Hornik RC. Patient-clinician information engagement increases treatment decision satisfaction among cancer patients through feeling of being informed. Patient Educ Couns. 2009;77(3):384–90. 48. Coulter A, Jenkinson C. European patients’ views on the responsiveness of health systems and healthcare providers. Eur J Public Health. 2005;15(4):355–60. 49. Coulter A. What’s Happening around the World? 2nd ed. Oxford, UK: Oxford University Press; 2009. 50. Davis RE, Jacklin R, Sevdalis N, Vincent CA. Patient involvement in patient safety: what factors influence patient participation and engagement? Health Expectations. 2007;10(3):259–67. 51. Coulter A, Collins A. Making Shared Decision Making a Reality: No Decision about Me, without Me. London: King’s Fund 2011. Available from: http://www.kingsfund.org.uk/sites/files/kf/ Making-shared-decision-making-a-reality-paper-Angela-CoulterAlf-Collins-July-2011.pdf. Accessed 12 November 2013. 52. Gentles S, Stacey D, Bennett C, Alshurafa M, Walter S. Factors explaining the heterogeneity of effects of patient decision aids on knowledge of outcome probabilities: a systematic review sub-anal ysis. Syst Rev. 2013;2:95. 53. Durand M-A, Carpenter L, Dolan H, et al. Do interventions designed to support shared decision-making reduce health inequalities? A systematic review and meta-analysis. PLoS ONE. 2014;9(4):e94670. 54. Weng HH, Kaplan RM, Boscardin WJ, et al. Development of a decision aid to address racial disparities in utilization of knee replacement surgery. Arthritis Care Res. 2007;57(4):568–75. 55. McCaffery KJ, Holmes-Rovner M, Smith SK, et al. Addressing health literacy in patient decision aids. BMC Med Inform Decis Making. 2013;13(Suppl 2):S10. 56. Sepucha KR, Fagerlin A, Couper MP, Levin CA, Singer E, Zikmund-Fisher BJ. How does feeling informed relate to being informed? The DECISIONS survey. Med Decis Making. 2010; 30(5, Suppl):77S–84S.

41. Shorten A, Shorten B, Keogh J, West S, Morris J. Making choices for childbirth: a randomized controlled trial of a decision-aid for informed birth after cesarean. Birth. 2005;32(4):252–61.

57. Sepucha K, Mulley A. A Practical Approach to Measuring the Quality of Preference-Sensitive Decisions. Oxford, UK: Oxford University Press; 2009.

42. Vandemheen KL, O’Connor A, Bell SC, et al. Randomized trial of a decision aid for patients with cystic fibrosis considering lung transplantation. Am J Respir Crit Care Med. 2009;180(8):761–8.

58. Conway T, Willcocks S. The Role of Expectations in the Perception of Health Care Quality: Developing a Conceptual Model. Int J Health Care Qual Assurance. 1997;10(3):131–40.

434




59. O’Connor A, Llewellyn-Thomas H, Stacey D. IPDAS Collaboration Background Document: International Patient Decision Aid Standards (IPDAS) Collaboration, 17 February 2005. Available from: http://ipdas.ohri.ca/IPDAS_Background.pdf. Accessed 12 November 2013.

64. Lee D, Lee J, Kang C. Development and validation of an altruism scale for adults. Psychol Rep. 2003;92(2):555–61.

60. McCann SK, Campbell MK, Entwistle VA. Reasons for participating in randomised controlled trials: conditional altruism and considerations for self. Trials. 2010;11:31.

65. Rojavin MA, Downs P, Shetzline MA, Chilingerian R, Cohard-Radice M. Factors motivating dyspepsia patients to enter clinical research. Contemp Clin Trials. 2006;27(2):103–11. 66. Bowling A. Research Methods in Health: Investigating Health and Health Services. 2nd ed. Buckingham, UK: Open University Press; 2002.

61. Fayter D, McDaid C, Eastwood A. A systematic review highlights threats to validity in studies of barriers to cancer trial participation. J Clin Epidemiol. 2007;60(10):990–1001.

67. Smith GD, Egger M. Going beyond the Grand Mean: Subgroup Analysis in Meta-Analysis of Randomised Trials. 2nd ed. London: Wiley-Blackwell/BMJ Books; 2008.

62. Torgerson D, Sibbald B. Understanding controlled trials: What is a patient preference trial? BMJ. 1998;316(7128):360.

68. Matthias MS, Salyers MP, Frankel RM. Re-thinking shared decision-making: context matters. Patient Educ Couns. 2013; 91(2):176–9.

63. Sawyer J. The Altruism Scale: a measure of co-operative, individualistic, and competitive interpersonal orientation. Am J Sociol. 1966;71(4):407–16.

ORIGINAL ARTICLE

435


The cost-effectiveness of patient decision aids: A systematic review.

Impact of sociodemographic patient characteristics on the efficacy of decision AIDS: a patient-level meta-analysis of 7 randomized trials.

Factors explaining the heterogeneity of effects of patient decision aids on knowledge of outcome probabilities: a systematic review sub-analysis.

A systematic development process for patient decision aids.

Cochrane Corner: What is the clinical impact of oxygen therapy for acute myocardial infarction? Evaluation of a Cochrane systematic review.

The impact of DECISION+2 on patient intention to engage in shared decision making: secondary analysis of a multicentre clustered randomized trial.

[Treatment of nonmetastatic prostate cancer: a systematic review of interactive, personalized patient decision aids].

What information is used in treatment decision aids? A systematic review of the types of evidence populating health decision aids.

Validity of a Cochrane Systematic Review and meta-analysis for determining the safety of vitamin E.

Medical versus surgical methods of early abortion: protocol for a systematic review and environmental scan of patient decision aids.

A systematic review of decision aids that facilitate elements of shared decision-making in chronic illnesses: a review protocol.

The impact of electronic decision support on transfusion practice: a systematic review.

Conclusiveness of the Cochrane reviews in nutrition: a systematic analysis.

Efficacy of patient-initiated follow-up clinics in secondary care: a systematic review.

The impact of Cochrane Systematic Reviews: a mixed method evaluation of outputs from Cochrane Review Groups supported by the UK National Institute for Health Research.

Undetermined impact of patient decision support interventions on healthcare costs and savings: systematic review.

The impact of a smoke-free psychiatric hospitalization on patient smoking outcomes: a systematic review.

Decision aids to help older people make health decisions: a systematic review and meta-analysis.

Video decision aids to assist with advance care planning: a systematic review and meta-analysis.

Physical activity and sport participation: A systematic review of the impact of fatherhood.

Cochrane Corner: evaluation of the Cochrane systematic review "antiplatelet agents for chronic kidney disease".

AIDS: a systematic review.

Interventions for tophi in gout: a Cochrane systematic literature review.

Impact of patients' psychiatric hospitalization on caregivers: a systematic review.