Medical Decision Making http://mdm.sagepub.com

Decisional Conflict in Patients and Their Physicians: A Dyadic Approach to Shared Decision Making Annie LeBlanc, David A. Kenny, Annette M. O'Connor and France Légaré Med Decis Making 2009; 29; 61 originally published online Feb 4, 2009; DOI: 10.1177/0272989X08327067 The online version of this article can be found at: http://mdm.sagepub.com/cgi/content/abstract/29/1/61

Published by: http://www.sagepublications.com

On behalf of:

Society for Medical Decision Making

Additional services and information for Medical Decision Making can be found at: Email Alerts: http://mdm.sagepub.com/cgi/alerts Subscriptions: http://mdm.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations http://mdm.sagepub.com/cgi/content/refs/29/1/61

Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

SHARED DECISION MAKING

Decisional Conflict in Patients and Their Physicians: A Dyadic Approach to Shared Decision Making Annie LeBlanc, MSc, David A. Kenny, PhD, Annette M. O’Connor, PhD, France Le´gare´, MD, PhD

Background. Decisional conflict is defined as personal uncertainty about which course of action to take when choice among competing options involves risk, regret, or challenge to personal life values. It is influenced by inadequate knowledge, unclear values, inadequate support, and the perception that an ineffective decision has been made. Until recently, it has been studied at the individual level, which ignores the interpersonal system between patients and physicians. Objective. To explore the effect of feeling uninformed, unclear values, inadequate support, and the perception that an ineffective decision has been made on one own’s outcome (actor effect) and on the other person’s outcome (partner effect). Methods. After a clinical encounter, modifiable deficits and personal uncertainty were measured in physicians and patients using the Decisional Conflict Scale. Structural equation modeling was used to measure the parameters of the Actor-Partner Interdependence Model. Results. A total of 112 dyads of physicians and patients were included in the analysis. For both

patients and physicians, 2 actor effects, unclear values (P < 0:0001) and the perception that an ineffective decision has been made (P < 0:0001), were found to be positively correlated with personal uncertainty. One partner effect, feeling uninformed (P = 0:03), was found to be negatively correlated with personal uncertainty. Conclusions. Personal uncertainty of patients and physicians is influenced not only by their respective deficits but also by the deficits of the other member of the dyad. Our results indicate that the more unclear the expression of their own values and the more they perceive that an ineffective choice had been made, the more both physicians and patients experience personal uncertainty. They also indicate that the less uninformed they feel, the more both physicians and patients experience personal uncertainty. Key words: shared decision making; relationship-centered care; dyadic decision making; physician-patient relationship; decisional conflict; Actor-Partner Interdependence Model. (Med Decis Making 2009;29:61–68)

T

of the individual rather than at the level of the dyad, thus ignoring the interpersonal system that is at play.9 12 Individuals involved in dyadic interactions, even brief ones, can, and often do, influence each other’s cognitions, emotions, and behaviors.13 Moreover, in most of the current shared decisionmaking research programs, predictor variables and outcomes assessed on the patient’s side are typically different from those assessed on the provider’s side.

he decision-making process in the clinical encounter is influenced by its inherent uncertainty and by the reaction of physicians and patients to this uncertainty.1 3 Given this context of uncertainty, decisional conflict is one of the key elements in decision making in clinical settings.4 Decisional conflict is defined as an individual perception of uncertainty about which course of action to take when choice among competing options involves risk, loss, regret, or challenge to personal life values.5 In lay terms, decisional conflict refers to the level of comfort that an individual experiences when facing a difficult decision. Decisional conflict should not be confused with the uncertainty inherent in the nature of the available scientific evidence.6 Until recently, decisional conflict in patients7 as well as in physicians8 had been studied at the level

Address correspondence to France Le´gare´, MD, PhD, Centre Hospitalier Universitaire de Que´bec, Hoˆpital St-Franc¸ois d’Assise, 10 rue de l’Espinay, Que´bec (Que´bec), Canada G1L3L5; telephone: (418)525-4437; fax: (418)525-4194; e-mail: France.Le´gare´@ mfa.ulaval.ca. DOI: 10.1177/0272989X08327067

MEDICAL DECISION MAKING/JAN–FEB 2009 Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

61

LEBLANC AND OTHERS

For example, a patient’s preferred role in decision making may be assessed as a predictor and satisfaction with the decision as an outcome. On the provider’s side, gender may be assessed as a predictor and performance in shared decision making as an outcome. Several modifiable deficits can increase or decrease the level of uncertainty perceived by individuals facing health-related decisions. Among them are 1) the perception that they have inadequate knowledge or understanding of the advantages/benefits and disadvantages/risks associated with all options, including doing nothing (i.e., feeling uninformed); 2) unclear values of the users regarding the relative desirability or importance of the benefit versus harm of the available options; 3) inadequate support the user feels, including undue pressure; and 4) the overall perception that a poor quality or ineffective decision has been made.14 Thus, combining the patient and the physician perspectives regarding the uncertainty associated with a specific decision has the potential for improving the understanding of the decisional process that occurs during a clinical encounter.8 The ActorPartner Interdependence Model (APIM) is a model for dyadic data analysis that explains the interdependence between 2 individuals. Using data from a previous implementation trial of shared decision making, the aim of the present study was to explore the effect of feeling uninformed, unclear values, inadequate support, and the perception that an ineffective decision has been made on personal uncertainty in both patients and their physicians.

Received 31 January 2008 from the Research Center, Hoˆpital SaintFranc¸ois d’Assise, Centre Hospitalier Universitaire de Que´bec, Universite´ Laval, Que´bec, Canada (AL); the Department of Psychology, University of Connecticut, Storrs, CT (DAK); the Faculty of Health Sciences, School of Nursing, University of Ottawa, Ottawa, Canada (AMO); and the Faculty of Medicine, Universite´ Laval, Que´bec, Canada (FL). AL holds a doctoral scholarship from the Canadian Institutes of Health Research (CIHR). FL conceived the main study, which provided the data that were used for this secondary analysis. She supervised AL’s graduate student project, validated the methods, and participated in the interpretation of the results and in the writing of the paper. AL conceived the analytical framework, analyzed and interpreted the data, and participated in the writing of the paper. DAK is the coinventor of the Actor-Partner Interdependence Model and thus provided expertise and support in the analysis and interpretation of results. He reviewed the paper and provided insightful comments. AMO is the author of the Ottawa Decision Support Framework and was a coauthor of the main study. She reviewed the paper and provided insightful comments. FL is its guarantor. All authors declare that they have no conflicting financial interests. Revision accepted for publication 16 June 2008.

METHODS Data Source Data originated from a before-and-after study for which the overall goal was to implement shared decision making in clinical primary care practices.11 The previous study assessed the impact of shared decision making on the agreement between physicians’ and their patients’ decisional conflict using the Decisional Conflict Scale (DCS). Briefly, at baseline and then after training sessions on shared decision making, clinical teachers and residents (n = 112) from 5 Family Practice Teaching Units (FPTUs) and their patients completed a postclinical encounter questionnaire that assessed their respective level of decisional conflict regarding the decision that had been made. Clinical teachers and residents had to be members of one of the 5 FPTUs and be involved in outpatient clinical activities in order to be eligible for the study. Patients who made a decision with their physician regarding any given condition were eligible for the study. Sociodemographic characteristics, preference in decisionmaking style,15 and decisional conflict16 data were collected using self-administered questionnaires. The study was approved by the Institutional Ethics Committee, and participants, both physicians and patients, signed an informed consent form. For the present study, we used data from the first clinical encounter for each physician in the original study. Therefore, data used for this secondary analysis resulted in a standard dyadic design of 112 dyads. In other words, each patient and each physician was a member of one and only one dyad. This limited the possibility of a learning effect on the provider’s side. Data Collected To assess decisional conflict, both physicians and patients completed the DCS after the clinical encounter. The DCS for both patients and physicians had the same items. The DCS is a multidimensional scale of 16 items divided into 5 subscales: personal uncertainty (3 items) and its modifiable deficits of feeling uninformed (3 items), unclear values (3 items), inadequate support (3 items), and perception that an ineffective choice has been made (4 items).16,17 Each item was scored on a 5-point Likert scale (1 = strongly agree to 5 = strongly disagree). Each one of the 5 subscale scores was calculated as follows: the sum of the corresponding items was computed and then divided by the number of items.

62 • MEDICAL DECISION MAKING/JAN–FEB 2009 Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

DYADIC APPROACH TO SHARED DECISION MAKING

In this present study, we did not use the total DCS score (16 items). However, the total DCS score correlates with knowledge scores, intention to accept influenza vaccine or breast cancer screening and delay in the decision to be immunized.16 Internal consistency coefficients (Cronbach a) for the total DCS score from previous studies ranged from 0.78 to 0.92 in the patient version and from 0.78 to 0.90 in the physician version of this scale.10;16 Furthermore, comparison of the respective structures of the patient and physician versions of the DCS showed that the subscales appear to have the same meaning for both physicians and patients (data available from corresponding author).

Data Analysis Descriptive data analysis of participants’ characteristics was performed using mean ± SD for continuous variables and percentage for categorical data. The APIM served as the analytical framework, as this model takes into account the interdependence between observations without losing possible valuable information about what each member contributes to the dyad.13;18;19 The APIM allows for the concurrent evaluation of a person’s predictor variables on his or her own outcome (actor effect) and on the other person’s outcome (partner effect) (Figure 1). Statistical analysis was done by means of structural equation modeling (SEM) with a maximum likelihood estimation. The dependent variable (outcome) was personal uncertainty. The predictor variables were the modifiable factors contributing to personal uncertainty (i.e., feeling uninformed, unclear values, inadequate support, and perception that an ineffective choice has been made). Results are presented as regression weights. An initial APIM model was first constructed that allowed all paths (effects) to be estimated. Then a second model was built, where all corresponding actor and partner paths for both patient and physician were set to be equal, thereby assessing how similar the effects were between physicians and patients. If the w2 difference test of the second model was not statistically different from the initial model, then this second simpler/less elaborate model was retained. Measures of model fit calculated included the w2 , the comparative fit index (CFI), and root mean square error of approximation (RMSEA).20;21 A nonsignificant w2 , a CFI ≥ 0.95, and a RMSEA value ≤ 0.06 all indicate a good model fit. Statistical analysis was done using SPSS (version 12.0, Chicago, IL) and AMOS (version 6.0, SPSS) software.

Er1 a

PREDICTOR person 1 f

d PREDICTOR person 2

1 OUTCOME person 1 e

c

b

OUTCOME person 2 1 Er2

Figure 1 The Actor-Partner Interdependence Model. The actor effects are represented by paths a and b and estimate the effect that a predictor variable of a person has on the outcome variable of that same person. Paths c and d represent partner effects that estimate the effect of the predictor of one person on the outcome of the other person in the dyad. The f curved line with doubleheaded arrows represents the correlation between the 2 predictor variables, and the e curved line with double-headed arrows represents the residual nonindependence not explained by actor and partner effects.13

RESULTS Characteristics of Participants A total of 112 physicians in family medicine (63 clinical teachers and 49 residents) and a single patient for each (n = 112) were included in the present study. Characteristics of participants are detailed in Tables 1 and 2. Scores for each of the subscales of the DCS are presented in Table 3. For both physicians and patients, all scores are under 2.5, indicating an overall low level of personal uncertainty and its modifiable factors. When asked what would be their preference in decision-making style, ‘‘sharing the decision’’ was the preferred style for 36.6% of physicians and for 34.8% of patients (Tables 1 and 2).

Actor-Partner Interdependence Model The initial model tested, which included all possible actor and partner effects, is saturated with zero degree of freedom, so the fit cannot be evaluated. Constraining corresponding actor and partner paths across both physician and patient resulted in a w2 with 8 df = 11.2, P = 0:191, which was not statistically different from the initial model (CFI = 0.994, RMSEA = 0.06). Hence our final model includes equality of corresponding paths and is

63

SHARED DECISION MAKING Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

LEBLANC AND OTHERS

Table 1 Characteristics of Physicians Characteristic

Age, y Years in practice (physicians only, n = 63) Hours per week spent in professional activities Female Status Residents Clinical teachers in family medicine Preference in decision-making style Patient alone Patient after considering my opinion Patient and myself Myself after considering the patient’s opinion Myself alone Did not answer

Table 2 Characteristics of Patients

Physicians (n = 112) a

35.5+9.3 13.9+8.5a 43.1+12.7a 75 (67)b b

49 (43.8) 63 (56.2)b

12 (10.7)b 39 (34.8)b 41 (36.6)b 17 (15.2)b 0 (0)b 3 (2.7)b

a. Data in mean+SD. b. Data in number (%).

Characteristic

Patients (n = 112)

Age, y Female Schooling University degree Secondary school or college degree Primary school degree Preference in decision-making style Myself alone Myself after considering the opinion of my physician Physician and myself Physician after considering my opinion Physician alone Did not answer

48.4+17.2a 61 (55)b 34 (30)b 61 (55)b 17 (15)b 5 (4.5)b 36 (32.1)b 39 (34.8)b 20 (17.9)b 6 (5.4)b 6 (5.4)b

a. Data in mean+SD. b. Data in number (%).

Table 3 Decisional Conflict Scale (DSC) Subscale Scores for Physicians and Patients

presented in Figure 2. The proportion of variance explained by this model (square multiple correlation) is 42% for the physician’s personal uncertainty and 47% for the patients’ personal uncertainty. Actor effects. Actor effects estimate the effect that a person’s predictor variable has on the outcome variable of that same person. In this study, it refers to whether one’s own modifiable deficits (feeling uninformed, unclear values, inadequate support, and perception that an ineffective choice has been made) have an influence on one’s own personal uncertainty. Two actor effects have a statistically significant effect on personal uncertainty (Table 4). For both physicians and patients, unclear values and perception that an ineffective choice has been made are positively related to personal uncertainty (P < 0:0001). The predictors feeling uninformed and inadequate support have no statistically significant actor effect on personal uncertainty. Partner effects. Partner effects estimate the effect of a person’s predictor on the outcome of the other person in the dyad. In this study, it refers to whether one’s own modifiable deficits (feeling uninformed, unclear values, inadequate support, and perception that an ineffective choice has been made) have an influence on the level of personal uncertainty in the

DCS Subscale Scores

Personal uncertainty Feeling uninformed Unclear about personal values Inadequate support in decision making Perception that one has made an ineffective decision

Physicians (n = 112)

Patients (n = 112)

1.85+0.74 1.69+0.53 1.78+0.58

1.46+0.62 1.46+0.57 1.40+0.46

1.89+0.57

1.60+0.64

1.71+0.47

1.34+0.43

other member of the dyad. Only one statistically significant reciprocal partner effect was observed for both patients and physicians; one’s own perception of feeling uninformed was negatively related with the personal uncertainty of the other member of the dyad (Table 4). Note that this effect is negative: the more one’s partner felt uniformed, the less personal uncertainly the person felt. No other significant partner effects were observed for the other predictors. DISCUSSION To the best of our knowledge, this study is the first to explore the effects of the modifiable deficits of personal uncertainty (i.e., feeling uninformed, unclear

64 • MEDICAL DECISION MAKING/JAN–FEB 2009 Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

DYADIC APPROACH TO SHARED DECISION MAKING

Feeling uninformed (MD)

Unclear values (MD)

P1

A1 A2

P2 Inadequate support (MD)

A3 P3

A4

Ineffective decision (MD)

Personal uncertainty (MD) Er1

P4 P1

Feeling uninformed (PT)

P2

P3 Unclear values (PT)

A1 A2

P4 A3 Inadequate support (PT)

Er2

Personal uncertainty (PT)

A4

Ineffective decision (PT)

Figure 2 The model used to test the Actor-Partner Interdependence Model for personal uncertainty and its modifiable deficits. Actor effects are represented by A and partner effects by P. The equality of actor and partner paths is represented by the arrows numbered A1 to A4 and P1 to P4. The curved lines with double-headed arrows represent the correlation between actor and partner variables, except for the curved line between Er1 and Er2, which represents residual nonindependence. MD = physicians; PT = patients; Er1 and Er2 = variance not explained by the model.

values, inadequate support, and perception that an effective decision has been made) on personal uncertainty in both patients and physicians. Hence, dyadic analysis of decisional conflict through assessment of the association between personal uncertainty and its modifiable deficits in both patients and physicians and between patients and physicians allows the identification of actor and partner effects that might have important practical implications. One of the most interesting and important findings of this study is the presence of a similar partner effect for patients and physicians. Indeed, we found that feeling uninformed had a significant negative

partner effect on personal uncertainty. These results suggest that the more informed one felt (physician or patient), the more personal uncertainty the other member of the dyad experienced. In other words, the less one’s partner feels informed, the more certain one is. Taking into consideration that increasing knowledge is a major component of shared decision-making programs (i.e., decision aids), this is of utmost importance because it suggests that intervening to increase one’s own feeling of being informed in order to decrease one’s own personal uncertainty might in fact increase personal uncertainty in the

65

SHARED DECISION MAKING Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

LEBLANC AND OTHERS

Table 4 Actor and Partner Estimates for Physicians and Patients for the Actor-Partner Interdependence Model (N = 112 dyads)

Modifiable Deficits

Feeling uninformed Actor effect Partner effect Unclear values Actor effect Partner effect Inadequate support Actor effect Partner effect Ineffective decision Actor effect Partner effect

Estimate of Regression Weighta

P Value

0.07 −0.18

0.40 0.03

0.38 0.02

<0.0001 0.82

0.04 0.01

0.94 0.91

0.72 0.15

<0.0001 0.20

Note: Personal uncertainty is the dependent variable, and its modifiable deficits are the predictor variables. Bolded values, P < 0.05. a. Because estimates of regression weight were similar for both, patients and physicians, the actor and partner paths are set to be equal (Figure 2).

other member of the dyad. However, we caution that our model is a predictive and not a causal model. Somewhat surprisingly, feeling uninformed had no impact on one’s own personal uncertainty when controlling for all actor and partner effects in the model. As suggested by previous research, assessing the impact of shared decision making on physicians will be as important as concomitantly assessing its impact on patients because there may be unintended effects.8 In other words, intervening to improve the patient’s situation and not taking into account the physician or vice versa could be associated with collateral damage affecting the other member of the dyad. Ten years ago, a group of researchers in primary care reflected on how clinicians’ consulting methods might affect patients’ outcomes in an unintended way.22 Instead of seeing resistance to change as rooted entirely in the patient, this group viewed it as stemming partly from the way clinicians talk to patients and proposed a negotiation-based framework that harnessed patients’ intrinsic motivation to make their own decisions. In summary, these researchers were tapping into the interpersonal system during clinical encounters and, more specifically, into shared decision making itself. Second, 2 actor effects were found to increase personal uncertainty in physicians and patients. Our results indicate that the more unclear the expression of their own values and the more they perceive that

an ineffective choice had been made, the more both physicians and patients experience personal uncertainty. In other words, reducing the perception of unclear values and/or that an ineffective choice had been made would reduce one’s level of personal uncertainty. Interestingly, despite the fact that physicians and patients have 2 completely different roles during the medical encounter, the magnitude of these actor effects was similar for both members of the dyad. This suggests that the decision-making process that occurs during the clinical encounter might very well be in fact a meeting of 2 experts: the health provider and the expert patient.23 Although many theorists disagree about how individuals make or ought to make judgments and decisions, most agree that the perception of how effective a choice will be and of the clarity of one’s values are essential factors in the process.24 Moreover, our analyses show that the process might be much more similar for the 2 members than expected. Last, in this study, we found there was neither an actor nor a partner effect for the predictor inadequate support on personal uncertainty. It is possible that the nature of the decisions that were made in the FPTU enrolled in the implementation study did not necessitate support from other individuals. Indeed, for both physicians and patients, all scores of the DCS subscales were under a threshold value of 2.5, indicating an overall low level of personal uncertainty and its modifiable factors. Therefore, it is possible that for decisions with a low level of personal uncertainty, adequate support from other individuals is not felt as important. Further studies will be required to assess the effect of these modifiable deficits on personal uncertainty in different types of decisions that have various levels of decisional conflict. Despite its interesting findings, limitations of the present study need to be acknowledged. First, this is a secondary analysis of a before-and-after trial, and thus the data were not intended primarily for the analysis that was performed. However, to the best of our knowledge, there are no other existing dyadic datasets that provide a better data source for exploring the effect on personal uncertainty of its modifiable deficits concomitantly in patients and their physicians.9 Furthermore, we limited data to the 112 dyads in which both physicians and patients completed the questionnaires for the first time. Keeping only one encounter per physician limited the bias that might have been caused by physicians completing this questionnaire more than once, thus creating a possible learning effect. Second, the overall low level of decisional conflict observed among

66 • MEDICAL DECISION MAKING/JAN–FEB 2009 Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

DYADIC APPROACH TO SHARED DECISION MAKING

members of dyads included in this study might have affected the results. A recently published study on the validation of the DCS in a group of French patients suggested that a low level of decisional conflict might influence the factorial nature of the DCS.25 Further investigation is needed among physicians and patients experiencing higher levels of decisional conflict. Finally, we have only a predictive and not causal model. Despite these limitations and given the need for improving our understanding of the complex mechanisms that are at play for a shared decisionmaking clinical encounter to occur, we believe our results can help improve the theoretical underpinnings, design, and methods of future studies. As anticipated in our previous work, the provision of research-based information about the effectiveness of interventions designed to inform and engage patients in health-care decisions will need to take into account a complex range of effects, among which is the way professionals and patients interact and influence each other.9,10 A major focus of shared decisionmaking programs is to increase the understanding of the risks and benefits of the proposed options for a health-care problem.26 According to our results, the perception of knowledge in both physicians and patients did not affect their own level of personal uncertainty but did affect the level of personal uncertainty of their partner, thereby creating ‘‘collateral damage’’. This implies that intervening at the level of only one member of the dyad (either the patient or the physician) may adversely affect the other member of the dyad. Wider use of a dyadic analytical framework like the APIM in future research has the potential to increase our understanding of the patient-physician interaction, leading the field of shared decision making in new and exciting directions.9 Only then will it be possible to acknowledge that the health-related clinical encounter of the 21st century is indeed a meeting of 2 experts23 who influence each other in order to find common ground.27 ACKNOWLEDGMENT We thank Mrs. Sylvie St-Jacques for her constructive comments of an earlier draft of this article and Hugh Glassco for reviewing this article.

2. Gerrity MS, Earp JA. Uncertainty and professional work: perceptions of physicians in clinical practice. Am J Sociol. 1992;97: 1022–51. 3. Gerrity MS, White KP, DeVellis RF, Dittus RS. Physicians’ reactions to uncertainty: refining the constructs and scales. Motiv Emot. 1995;19:175–91. 4. Janis IL, Mann L. Decision Making: A Psychological Analysis of Conflict, Choice, and Commitment. New York: Free Press; 1977. 5. Carpenito LJ. Decisional conflict. In: Carpenito-Moyet LJ, ed. Nursing Diagnosis: Application to Clinical Practice. Philadelphia: Lippincott Williams & Wilkins; 2000. p 312–21. 6. Nelson WL, Han PK, Fagerlin A, Stefanek M, Ubel PA. Rethinking the objectives of decision aids: a call for conceptual clarity. Med Decis Making. 2007;27:609–18. 7. O’Connor AM, Stacey D, Entwistle V, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev. 2003;2:CD001431. 8. Dolan JG. A method for evaluating health care providers’ decision making: the Provider Decision Process Assessment Instrument. Med Decis Making. 1999;19:38–41. 9. Le´gare´ F, Elwyn G, Fishbein M, et al. Translating shared decision-making into health care clinical practices: proof of concepts. Implement Sci. 2008;3:2. 10. Le´gare´ F, O’Connor A, Graham I, et al. The effect of decision aids on the agreement between women’s and physician’s decisional conflict about hormone replacement therapy. Patient Educ Couns. 2003;50:211–21. 11. Le´gare´ F, O’Connor AM, Graham ID, Wells GA, Tremblay S. Impact of the Ottawa Decision Support Framework on the agreement and the difference between patients’ and physicians’ decisional conflict. Med Decis Making. 2006;26:373–90. 12. Le´gare´ F, Tremblay S, O’Connor AM, Graham ID, Wells GA, Jacobsen MJ. Factors associated with the difference in score between women’s and doctors’ decisional conflict about hormone therapy: a multilevel regression analysis. Health Expect. 2003;6: 208–21. 13. Kenny DA, Kashy DA, Cook WL. Dyadic Data Analysis. New York: Guilford Press; 2006. 14. O’Connor AM, Tugwell P, Wells GA, et al. A decision aid for women considering hormone therapy after menopause: decision support framework and evaluation. Patient Educ Couns. 1998;33: 267–79. 15. Degner LF, Sloan JA, Venkatesh P. The Control Preferences Scale. Can J Nurs Res. 1997;29:21–43. 16. O’Connor AM. Validation of a decisional conflict scale. Med Decis Making. 1995;15:25–30. 17. Le´gare´ F, Graham I, O’Connor A, Dolan J, Be´langer-Ducharme F. Prise de de´cision partage´e: traduction et validation d’une e´chelle de confort de´cisionnel du me´decin. Pe´dagogie Me´dicale. 2003;4:216–22. 18. Kenny DA, Cook W. Partner effects in relationship research: conceptual issues, analytic difficulties, and illustrations. Pers Relatsh. 1999;6:433–48.

REFERENCES 1. Gerrity MS, DeVellis RF, Earp JA. Physicians’ reactions to uncertainty in patient care: a new measure and new insights. Med Care. 1990;28:724–36.

19. Kenny DA. Models of non-independence in dyadic research. J Soc Pers Relat. 1996;13:279–94. 20. Bentler PM. Comparative fit indexes in structural models. Psychol Bull. 1990;107:238–42.

67

SHARED DECISION MAKING Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009

LEBLANC AND OTHERS

21. Bollen KA, Long JS. Testing Structural Equation Models. Newbury Park (CA): Sage; 1993. 22. Butler C, Rollnick S, Stott N. The practitioner, the patient and resistance to change: recent ideas on compliance. CMAJ. 1996; 154:1357–62. 23. Coulter A. Paternalism or partnership? Patients have grown up-and there’s no going back. BMJ. 1999;319:719–20. 24. Pitz GF, Sachs NJ. Judgment and decision: theory and application. Ann Rev Psychol. 1984;35:139–64.

25. Mancini J, Santin G, Chabal F, Julian-Reynier C. Crosscultural validation of the Decisional Conflict Scale in a sample of French patients. Qual Life Res. 2006;15:1063–8. 26. 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;333:417. 27. Brown JB, Weston WW, Stewart MA. Patient-centred interviewing. Part II: finding common ground. Can Fam Physician. 1989;35:153–7.

68 • MEDICAL DECISION MAKING/JAN–FEB 2009 Downloaded from http://mdm.sagepub.com at CAPES on April 7, 2009