| | Measuring shared decision making in the consultation: A comparison of the OPTION and Informed Decision Making instrumentsReceived 15 May 2007; received in revised form 13 July 2007; accepted 2 September 2007. Abstract ObjectiveTo investigate the applied and conceptual relationship between two measures of shared decision making using the OPTION instrument developed in Wales and the Informed Decision Making instrument developed in Seattle, USA using audio-taped consultation data from a UK general practice population. MethodsTwelve general practitioners were recruited from 6 general practices in the southwest of England. One hundred twenty-three GP-patient consultations were audio-recorded. Audiotapes were sent off to, and rated by, respective experts in the use of the OPTION and the Informed Decision Making instruments. ResultsCompared to earlier work using the Informed Decision Making tool, consultations in this sample were shorter, had fewer decisions and tended to have a greater number of elements present. Similar to previous research using the OPTION, values using the OPTION instrument were low with two items, giving the patient opportunities to ask questions and checking patient understanding, exhibiting the most variability. Using a ‘key’ decision in each consultation as the basis for comparison, the Informed Decision Making score was not related to the overall OPTION score (Spearman's rho =0.14, p=0.13). Both instruments also predicted different ‘best’ and ‘worst’ doctors. Using a Bland–Altman plot for assessing agreement, the mean difference between the two measures was 1.11 (CI 0.66–1.56) and the limits of agreement were −3.94 to 6.16. There were several elements between the two instruments that appeared conceptually similar and correlations for these were generally higher. These were: discussing alternatives or options (Spearman's rho=0.35, p=0.0001), discussion of the patient's role in decision making (Spearman's rho=0.23, p=0.012), discussion of the pros/cons of the alternatives (Spearman's rho=0.20, p=0.024) and assessment of the patient's understanding (Spearman's rho=0.19, p=0.03). ConclusionMeasures of shared decision making are helpful in identifying those shared decision making skills which may be problematic or difficult to integrate into practice and provide a tool by which the development of skills can be assessed over time. Research may implicitly place undue value on those aspects of shared decision making which are most easily measured. Practice implicationsShared decision making tools are a useful way of capturing the presence or absence of specific shared decision making skills and changes in skills acquisition over time. However there may be limits in the extent to which the concept of shared decision making can be measured and that more easily measured skills will be emphasised to the detriment of other important shared decision making skills. 1. Introduction  There is an ever increasing range of patient-centred instruments to measure communication and consultation skills in the health professional–patient encounter [1], [2]. These range from history-taking [3], non-verbal behaviour [4], empathic processes [5], [6], or the skills needed to demonstrate evidence-based patient choice [7]. Some instruments are more appropriate for use in teaching where the ‘findings’ are used formatively to improve communication and consultations skills [5], [8]. Other observational instruments focus on interaction analysis; coding phrases or utterances of speech [9], [10], [11]. To develop an instrument which has good internal and external validity, is reliable yet easy to use (e.g. minimal training requirements) and has wide acceptance and use amongst researchers and practitioners in shared decision making is a goal for any such instrument. The common thread between tools is the aim to quantify aspects of human interaction in order to judge ‘good’ communication between (most commonly) doctor and patient. Yet there is a broad spectrum in how different instruments have been operationalised. There may be a conceptual overlap between individual items on different instruments without there being a complete overlap of items across the two instruments as a whole. For example, in patient-completed measures, an item such as ‘the doctor seems interested in me as a person’ might overlap conceptually with ‘how much would you say that this doctor cares about you as a person?’ but the former is part of a rapport sub-scale on the Medical Interview Satisfaction Scale [12] while the latter is part of an understanding the whole person sub-scale on the Patient Perception of Patient Centredness scale [13], [14]. There are also differences in terms of depth of individual elements on a particular tool. Some items focus on discrete, observable skills (competencies) such as ‘avoids directive or leading questions’ while other items are broad competences (outputs) such as ‘involves patient in deciding upon a plan’ which might include several competencies within it [15]. The other key issue in the development of instruments is the perspective adopted for the rating of consultations: patient completed instruments, observer-rated instruments or health professional completed instruments. Different instruments from different perspectives will potentially lead to conflicting views as to what constitutes a ‘good’ consultation. This paper will explicitly focus on observer-rated instruments. Shared decision making can be considered an important aspect of patient centredness [16] and the consultation process. Attempts have been made to develop a core definition of shared decision making [17] in order to develop consistent measures which can be related to patient outcome. The most widely accepted model of shared decision making draws upon the work of Charles et al. [18], [19] most recently tailored to fit the context of general practice [20]. As originally devised, this model describes shared decision making as having four characteristics: (1) both the doctor and patient are involved in the treatment decision-making process; (2) both share information with each other; (3) both take steps to participate in the decision-making process by expressing treatment preferences and (4) both the doctor and patient agree on the treatment to implement. As constructed, shared decision making has elements which overlap with the tools described above such as exploring the patient's reasons for attending, adopting a bio-psychosocial perspective, giving information and eliciting patient concerns. There are two observer-rated tools, the observing patient involvement [OPTION] instrument developed in Wales and the Informed Decision Making [IDM] instrument development in California, which take an explicitly shared decision making perspective. The OPTION scale was developed from a skills framework and consists of a set of competences [21] which include: •problem definition; •explaining legitimate choices; •portraying options and communicating risk; •conducting the decision process or its deferment. The psychometric properties of this scale have been explored and it has undergone extensive reliability testing [22], [23]. Early use of Likert scales to rate competences have evolved to magnitude scales rated from ‘0’ to ‘4’ and improved OPTION's reliability [23]. More recently OPTION has been used in a cluster randomised controlled trial to investigate the effects of interventions to improve doctors’ skills in shared decision making or in their use risk communication aids [24]. The Informed Decision Making [IDM] tool identified both the shared decision making and informed consent paradigms as relevant to its conceptual development. The authors initially identified six elements in the IDM model, drawn from the bioethics literature: •discussion of the clinical issue and nature of the decision to be made; •discussion of the alternatives; •discussion of the pros and cons of the alternatives; •discussion of the uncertainties associated with the decision; •assessment of the patient's understanding; •asking the patient to express a preference [25]. This research was followed up with a large study of 1057 doctor–patient interactions among 124 doctors [26]. This later study added the additional element of ‘discussion of the patient's role in decision making’ and found that only 9% of decisions met their definition for informed decision making. In a review of patient–doctor communication assessment instruments, 44 instruments were examined [1]. Assessment tools included those using real-time assessment by an observer, those using standardised patients to assess communication skills, assessment with video or audio taped interactions and measures using self report by doctors or patients. The authors were critical of some measures which had been used solely by their developers and for their lack of robust reliability and validity testing [1]. There have been a few studies comparing across instruments but poor concurrent validity amongst instruments has been found [27], [28], [29]. A notable exception is the Roter Interaction Analysis System (RIAS) which has been shown to have good predictive validity with other instruments [30]. One solution to the growth of communication instruments is to develop a new instrument which involves several research groups, is tested rigorously for validity and reliability and is used widely. Another approach is to undertake comparative work between two or more instruments. Two instruments, the OPTION instrument developed in Wales and the Informed Decision Making instrument developed in Seattle, USA, were compared. The reasons for comparing these two particular instruments are (a) both instruments have an explicit focus developed from a theoretical perspective of involving patients in healthcare decision making; (b) both assume a chronology for the consultation for the rating process; (c) the Informed Decision Making tool had the highest number of stages or competences in common with Elwyn's stages identified with his tool [21] and (d) the Informed Decision Making tool was identified by Elwyn as the instrument most directly comparable with his instrument [22]. This was a feasibility study to explore in an undifferentiated general practice population, the relationship between the OPTION and Informed Decision Making instruments. For brevity, the term ‘shared decision making’ will include both the shared decision making and Informed Decision Making approaches. 2. Method After obtaining ethical approval, letters were sent to 82 practice managers within three Primary Care Trusts in the southwest of England during the summer of 2003. The letters asked if the general practitioners (GPs) in the practices would be willing to participate in a study investigating GPs’ communication about medicines. The aim was to recruit between 10 and 20 GPs and, for each participating GP, to record approximately 10 patient consultations. General practitioners were provided with feedback at the end of the research if they wished. GPs agreed ‘in principle’ to having their patient consultations recorded and were provided with an audio tape-recorder in their consultation room at the start of the surgery session. Patients were recruited by the researcher (MCW) in the waiting room prior to their appointment. A range of surgery sessions were selected for recruitment and varied with regard to day of the week and time of day. To maintain the broad mix of general practice consultations, ‘single disease’ surgery sessions were avoided (e.g. hypertension clinics). Patients were recruited only if they had an appointment with that particular GP. For those who had booked appointments in advance, a letter and information sheet was sent to the patient's home prior to their appointment date, although it was evident that few patients had read the study information prior to coming in for their appointment. For patients who made an appointment on the day, a patient information sheet was given to them by the receptionist when they booked in. Patients were told to see the researcher sitting in the waiting area if they had any questions or if they were interested in participating in the research. Patients were excluded if they did not speak English, were unable to consent or appeared too unwell. Consent was obtained prior to the patient going into their consultation. Although there were initial concerns that the process of informed consent would delay the patient's appointment, this never occurred—either because the process of recruitment was abandoned when a patient was called in or, more frequently, that the GP was running late on their appointment times. Patients who consented were told to tell their GP to turn on the tape-recorder. Consenting patients also received at questionnaire at the time of recruitment to complete after their consultation and to post back to the researchers. The patient-completed questionnaire contained a number of items including the 14-item Measure of Patient Perception of Patient-Centredness (PPPC) [13], [14]; the extent to which the patient's concerns were resolved by the consultation [13]; patient satisfaction with the consultation; the patient's views on information provided in the consultation [31]; the extent to which the patient wanted to be involved in treatment decisions [32]; the Beliefs about Medicines Questionnaire [33]; the patient's anxiety level [34] and general health (the COOP/WONCA functional assessment charts) [35]. Findings from the questionnaire will not be discussed in this paper. GP-patient consultations were rated using the OPTION and IDM instruments. Each audio-recorded consultation was rated twice: tapes were sent to Wales and rated by a trained expert in OPTION and they were also sent to Stanford and rated by a trained expert in IDM. OPTION contains 12 items where each item is rated from ‘0’ to ‘4’ with ‘0’ indicating a competency which is not observed and scores of ‘1’ to ‘4’ representing increasing levels of achievement. The 12 items included on the OPTION tool is shown in Table 1. An overall OPTION score is obtained by adding the scores on each of the 12 items together (maximum score of 48) and then standardising it to have a score between 0 and 100. Although a number of conditions may be discussed in a consultation, an ‘index problem’ is determined which is the problem, over the course of the entire consultation, where the highest degree of involvement occurs [36].  | 1 | The clinician draws attention to an identified problem as one that requires a decision making process | | | 2 | The clinician states that there is more than one way to deal with the identified problem (‘equipoise’) | | | 3 | The clinician assesses the patient's preferred approach to receiving information to assist decision making (eg discussion, reading printed material, assessing graphical data, using videotapes or other media) | | | 4 | The clinician lists ‘options’, which can include the choice of ‘no action’ | | | 5 | The clinician explains the pros and cons of options to the patient (taking ‘no action’ is an option) | | | 6 | The clinician explores the patient's expectations (or ideas) about how the problem(s) are to be managed | | | 7 | The clinician explores the patient's concerns (fears) about how problem(s) are to be managed | | | 8 | The clinician checks that the patient has understood the information | | | 9 | The clinician offers the patient explicit opportunities to ask questions during the decision making process | | | 10 | The clinician elicits the patient's preferred level of involvement in decision-making | | | 11 | The clinician indicates the need for a decision making (or deferring) stage | | | 12 | The clinician indicates the need to review the decision (or deferment) | | | | |
The IDM tool has nine key elements as shown in Table 2. An IDM score was determined at the level of each consultation decision, at the level of the consultation to include all decisions in that consultation and at the level of an individual doctor to include all their consultations. Each decision in a consultation is rated for each element with a ‘0’ indicating that the element is absent, a ‘1’ indicating that the element is partially present or that some effort was made and a ‘2’ indicating that there was a complete example of the element. Scores are then summed over the nine elements to get an aggregate ‘IDM-18’ score (maximum score of 18) for each decision. A summative score for each consultation was also calculated by determining the average for each element for the total number of decisions in the consultation and determining an aggregate ‘consultation IDM-18’. For the IDM tool, every decision made in the consultation is assessed resulting in some consultations having three or more decisions being rated. For the analysis, the IDM score for the decision which corresponded to OPTION's ‘index problem’ were compared. A Bland–Altman analysis [37] was used to assess the differences between the OPTION and IDM tools when these variables were standardised (mean=0, standard deviation=1). Descriptive statistics were used to explore patterns in the data and Spearman's rho correlation coefficient was used to investigate relationships between two variables. All analyses were considered significant at a probability value of p<0.05. Data analysis was performed using STATA, version 7. 3. Results 3.1. Description of the sample Of the 82 letters sent out to practices, 6 (7%) responded and 12 GPs agreed to participate. In four practices, two GPs participated, in one practice 3 participated and in one practice one GP participated. Practices were mixed such that two practices were in an urban location, 2 were suburban, 1 in a small market town and one in a rural location. Of the six participating practices, two were small practices (3 or fewer partners), one was a medium-sized practice (4–6 partners) and there were three large practices (7 or more partners). Of the 12 participating GPs, 4 were female and 8 were current members of the Royal College of General Practitioners. The average GP age was 42. In total, 128 out of a possible 274 patients (47%) with appointments agreed to participate in the study. Recruitment took place over 28 surgery sessions each lasting between 1 and 3h. Of those who did not participate, reasons for not participating included inadequate time for the researcher to meet with all patients prior to their appointment, patient refusal (either due to the nature of their planned consultation or they felt they were too busy) or lack of fluency in English. Five patient consultations were not recorded due to technical failure or because the patient forgot to inform the doctor of their participation resulting in 123 GP-patient consultations audio-recorded. Of the 128 patient questionnaires distributed, 80 (63%) were returned. Between 9 and 12 consultations were recorded per GP. The median consultation length was 8min, 30s (95% confidence interval, 7′ 17″ to 9′19″). Consultations comprised a wide range of clinical issues including chronic disease reviews (asthma, cardiac failure, hypertension, diabetes, depression, arthritis), acute injuries (tendonitis, knee or back pain), infections (URTI, flu, sinusitis, conjunctivitis) or illnesses (IBS, dyspepsia) and pre or post surgical assessments. 3.2. Shared decision making and the individual tools Women and younger people had higher involvement scores using the Informed Decision Making scale (p=0.03 and p=0.001, respectively) but gender and age were unrelated to the OPTION score (p=0.44 and p=0.20). As consultation length increased, so did the level of patient involvement in decision making using both the OPTION (p=0.02) and IDM tool (p=0.0005). Using the IDM tool, of the 123 consultations, there were 6 consultations where there were no decisions. In the remaining 117 consultations, 36 consultations (30.7%) had one decision, 44 consultations (37.6%) had two decisions and 37 consultations (31.6%) had 3 or more decisions. A comparison with how these values relate to Braddock's earlier work [26] is shown in Table 3. | a Mean number of decisions per visit for primary care physicians. |
Table 3 shows that the UK data tended to have shorter consultations with fewer decisions. Since the original publication of Braddock's work in 1999, he has added an additional two elements to his model (elements 2 and 8 in Table 2). By comparing the common elements between Braddock's earlier work and this study (elements 1, 3, 4, 5, 6, 7, 9), it can be seen that the decisions in this study tended to have a greater number of elements present (Table 4), particularly with regard to discussing alternatives, the uncertainties associated with the decision and exploring patient preferences. | a Element number as in Table 2. bValues are given as percentages (number) per total number of decisions. |
In Braddock's research, he defined decisions meeting an ‘all basic’ criterion as those decisions meeting the following two criteria: both element 1 or 9 present and element 3 (Table 2) present. That is, consultations had to include the patient's role in decision making or their preference for role involvement, and a discussion of the clinical issue. In Braddock's earlier work, 20.4% of decisions met these minimum criteria whereas 26.1% of decisions in this sample met the ‘all basic’ criteria for minimum dialogue acceptable for a clinical decision. The mean OPTION score for the sample was 3.81 (CI=3.25–4.39). The compares favourably with Elwyn's mean OPTION score for his sample of 3.17 [23]. Mean OPTION scores by individual GP participants is shown in Table 5. Responses were rated as ‘0’ in 90% or more consultations for items 1, 2, 3, 5, 6, 7 and 10. Item 9 (‘the clinician offers the patient explicit opportunities to ask questions during the decision-making process’) had the most variability with 54% of consultations rated as ‘0’, 44% as ‘1’ and 2% as a ‘2’. Similarly, item 8 (‘the clinician checks that the patient has understood the information’) had 61% of consultations rated as a ‘0’ while item 12 (‘the clinician indicates the need to review the decision for deferment’) had 70% of consultations rated as ‘0’. No item exceeded a ‘2’ for any consultation. These findings are comparable to Elwyn's dataset [23] who had a predominance of low scores with items 8 and 9 exhibiting the most variability. | | | | GP number | Mean OPTION score (95% CI) | Number of consultations | |
|---|
| 1 | 3.22 (1.30, 5.14) | 11 | | | 2 | 1.46 (−0.41, 3.32) | 10 | | | 3 | 5.09 (1.33, 8.86) | 9 | | | 4 | 5.21 (2.72, 7.70) | 12 | | | 5 | 2.95 (1.41, 4.49) | 12 | | | 6 | 6.67 (4.36, 8.98) | 10 | | | 7 | 2.55 (0.99, 4.11) | 9 | | | 8 | 2.5 (1.32, 3.68) | 10 | | | 9 | 4.92 (3.49, 6.36) | 11 | | | 10 | 4.17 (1.19, 7.15) | 10 | | | 11 | 4.17 (2.05, 6.29) | 9 | | | 12 | 2.71 (1.13, 4.29) | 10 | | | | |
3.3. Comparisons between the OPTION and Informed Decision Making tools As the Informed Decision Making scale rated all decisions which occurred in the consultation, the score for the decision which was corresponded to the index decision using the OPTION score was determined. The Informed Decision Making score for this key decision was not related to the overall OPTION score (Spearman's rho=0.14, p=0.13). Interestingly, in a sample of only 12 GPs, the two measures also predicted different ‘best’ (OPTION GP6, IDM GP1) and ‘worst’ doctors (OPTION GP2, IDM GP5). A Bland–Altman analysis for assessing the agreement between to methods of measurement was also conducted [37] (Fig. 1). The mean difference between the two measures was 1.11 (CI 0.66–1.56) and the limits of agreement were −3.94 to 6.16 indicating that the OPTION instrument could be nearly 4 points below the Informed Decision Making instrument to just over 6 above, suggesting an unacceptable level of agreement between the two instruments. It was hypothesised that individual items on the two instruments may be related to each other. Element 4 on IDM (‘discussion of alternatives including no action’) and item 4 on OPTION (‘the clinician lists options which can include the choice of no action’) were related (Spearman's rho=0.35, p=0.0001). Element 1 on IDM (‘discussion of the patient's role in decision making’) was related to Item 10 on OPTION (‘the clinician elicits the patient's preferred level of involvement in decision making’) (Spearman's rho=0.23, p=0.012). Element 5 on IDM (‘discussion of the pros/cons of the alternatives was related to item 5 on OPTION (‘the clinician explains the pros and cons of options to the patient’) were related (Spearman's rho=0.20, p=0.024). Element 7 on IDM (‘assessment of the patient's understanding’) and Item 8 on OPTION (‘the clinician check that the patient has understood the information’) were related (Spearman's rho=0.19, p=0.03). Element 3 on IDM (‘explains the nature of the clinical issue or decision) was related to item 1 on OPTION (‘the clinician draw attention to an identified problem as one that requires a decision making process’) (Spearman's rho=−0.18, p=0.049). It was hypothesised that element 6 on IDM (‘discussion of uncertainties of alternatives’) might be related to item 2 on OPTION (‘the clinician states that there is more than one way to deal with the identified problem’) but it was not (Spearman's rho=−0.07, p=0.42). Similarly elements 2 (‘determining the context of the decisions’) and 9 (‘elicitation or acknowledgement of the patient's preference’) on IDM were not related to either item 6 (‘the clinician explore the patient's expectations or ideas about how the problems are to be managed’) or 7 (‘the clinician explores the patient's concerns about how problems are to be managed’) on OPTION. No other relationships were explored. 4. Discussion and conclusion 4.1. Discussion This study investigated the relationship between two measures of shared decision making, the OPTION scale developed in Wales and the Informed Decision Making scale developed in Seattle, USA, in 123 general practice consultations. Overall the lack of agreement between the IDM tool and OPTION was surprising. The IDM tool was identified by Elwyn as the measure closest conceptually to his OPTION tool [2] with the OPTION tool achieving a higher level of reliability in OPTION's initial development than the IDM tool [22]. The lack of agreement between OPTION and the IDM tool may have been due to the breadth of clinical decisions included in this research. The sample included clinical decisions involving true ‘equipose’ where there are legitimate treatment options as well as other ‘uncomplicated’ decisions such as initiating or reducing medication, referrals to a specialist, follow-up appointments or changing the dose of a medicine. Earlier research suggested that shared decision making might be more useful in situations where several legitimate treatment options are available and genuine medical uncertainty [22]. This is in contrast to the approach by Braddock which includes all decisions, even those involving routine laboratory tests or follow-up visits [26]. The difficulty with the former approach is that it risks making a paternalistic judgement by health professionals of the type of decision most suitable for patient involvement. The difficulty with the latter is in developing a tool which robustly covers such a broad range of decision making. Patients make decisions all the time [38]; about whether or not to start or continue with a medicine or whether they should make a new appointment. Patients are always making choices, even if only one option is presented to them. The difficulty is being able to capture such diversity in decision making with a single quantitative instrument. The findings did show that some individual items on each of the instruments were related to each other. Items which were more specific and conceptually clearer were more likely to be related (e.g. listing options in treatment) than those with less clarity and specificity (e.g. exploring patient concerns and the context of decision-making). Interestingly, two items [4 and 8 on the OPTION instrument] which appeared to have greater conceptual clarity and specificity were also those with the greatest variability in rating using the OPTION instrument. It may be easier for raters to assess and grade levels of items which are conceptually clearer and more specific. The earlier IDM data and instrument were distinct from the current dataset both culturally and in time. The original IDM data was collected from practices on the west coast of the US 10 years before the current data. This may partly explain the findings that the UK data had shorter consultations, included fewer decisions in each consultation and had more elements of decision-making from the IDM model. This may reflect a general trend towards shared decision making being integrated within routine general practice and the distinct cultural differences between US and UK. The inclusion of a broad range of decisions in the development of the IDM instrument (e.g. not just equipoise decisions) may also reflect the more consumerist nature of US health care facilitating a more wide-ranging definition of decision making than the Welsh OPTION instrument. The recent reflective paper on measuring patient-centred communication in physician–patient consultations [39] argues that measures should include the communication behaviours of individuals in the interaction, as well as the interactions between them. Specifically, measures should not focus solely on health care professionals’ behaviours as they may miss factors which can influence outcomes. In this regard the IDM instrument comprises both doctor-focused elements (‘assessment of the patient's understanding’) and interactional elements (‘discussion of alternatives’). The OPTION instrument has a declared focus on doctor behaviours with items such as ‘the clinician checks that the patient has understood the information’. The IDM instrument may also have a stronger bio-psychosocial focus with several elements (elements 2, 8 and 9) exploring the patient's context and values in comparison with OPTION's single element (item 7) exploring patient concerns or fears about how problems are to be managed. The recent paper by Makoul and Clayman [17] identified essential and ideal elements of shared decision making based upon a review of the relevant literature. The five areas of commonality between IDM and OPTION identified in this study accord with Makoul's essential and ideal elements of shared decision-making as follows: •List or discussion of options (essential). •Defines nature of clinical issue requiring a decision (essential). •Discussion of pros and cons of alternatives (essential). •Assessment of patient understanding (essential). •Discussion of patient's desired role in decision-making (ideal). In this context, Makoul defines ideal elements as those which may enhance shared decision making but may not always be relevant or necessary for shared decision making to occur. This research supports Makoul's findings in identifying these core elements although other aspects identified by Makoul (patient values/preferences, patient self-efficacy, doctor knowledge, deferring decisions and arranging follow-up) were not found when applying these two shared decision making tools on a single dataset. These findings suggest that certain aspects of shared decision making may be easier to measure and, further, that there may be limits in the extent to which we can capture, in quantifiable terms, the concept of shared decision making. Tools which try to capture the full complexity of shared decision making may not be an achievable outcome, recognising the disjunction between theoretical definitions of shared decision making and the practicalities of measurement. 4.2. Study limitations The strengths of the study are that it explored the relationship between these two measures using a single dataset of undifferentiated UK general practice consultations and used experts in the respective measures to rate the consultations. However there are limitations to the research. As this was a feasibility study, there were a small number of patients agreeing to participate, making the sample unrepresentative of all of general practice consultations and possibly biased. Although the general practitioners were unaware of the precise nature of the study, they did know that the research involved communication in the consultation. This is likely to have affected GP recruitment and, for those that were recruited, it may have affected how they behaved during tape-recorded consultations. Certain types of consultations were also less likely to be recorded, such as difficult or embarrassing consultations where communication was more likely to be problematic. This study also used raters experienced in the use of their particular instrument. While this was a reasonable approach methodologically, it would have been interesting to use an independent rater who was specifically trained in the use of both instruments (who would rate each consultation twice) or for the experienced raters to be trained in the use of the other instrument. This may have addressed some of the cultural biases or implicit assumptions made by experienced raters and, potentially, could have led to greater congruence between the two instruments. 4.3. Conclusion This research provides insight into some of the dilemmas in measuring shared decision making. On the one hand these measures provide evidence as to which key elements of shared decision making are likely to present in a consultation and those which may be more problematic in practice. Further, with repeated studies, these measures can show how the skills in shared decision making have evolved over time, identifying those which may be integrated more easily into routine practice. On the other hand, while the strategic and political drive towards quantification and measurement of skills is typically considered to be a benign and value-free process, this research raises some uncertainties. Each researcher feels the need to develop their own particular measure capturing the elements of shared decision making they consider to be most valuable. These socially constructed measures are influenced by the cultural and sociological context in which they were developed and satisfy society's need to quantify aspects of human behaviour. This can place an implicit value on those elements most easily quantified, with the most conceptually specific and reducible aspects of behaviour gaining a salience and significance by their mere ability to be easily measured. Equally it may be that we need to recognise the limitations of measurement and its ability to capture the full complexity of the shared decision making process. 4.4. Practice implications Tools for measuring shared decision making are useful to practitioners in identifying new skills they may need to acquire or to measure progress in skill development. However, they may also, through their construction and application, over-emphasise skills which are easy to measure and implicitly place an undue value on these skills. Instruments need to be seen within the wider context in which they were developed and as a lens which views and measures some aspects of shared decision making but which may be less able to capture others. In common with other protocol or structured practice approaches, there is a danger of over-emphasising structure over the uniqueness of the individual GP-patient encounter where some of the less quantifiable skills and approaches may be the more appropriate. Acknowledgements This research was supported by a grant from the Health Foundation (formerly the PPP Foundation). The views presented in this article are those of the authors and do not necessarily represent the views of the Health Foundation. The authors would like to thank all the patients and general practitioners who participated in this research. The authors also gratefully acknowledge the kind help and support of Clarence Braddock III, Sylvia Bereknyei, Glyn Elwyn and Cathy Lisles. References [1]. [1]Boon H, Stewart M. Patient–physician communication assessment instruments: 1986–1996 in review. Pat Educ Counsel. 1998;35:161–176. [2]. [2]Elwyn G, Edwards A, Mowle S, Wensing M, Wilkinson C, Kinnersley P, et al. Measuring the involvement of patients in shared decision making: a systematic review of instruments. Pat Educ Counsel. 2001;43:5–22. [3]. [3]Van Thiel J, Kraan HF, Van der Vleuten CPM. Reliability and feasibility of measuring medical interviewing skills: the revised Maastricht history-taking and advice checklist. Med Educ. 1991;25:224–229. MEDLINE |
CrossRef
[4]. [4]Rosenthal R, Hall JA, DiMatteo MR, Rogers PL, Archer D. Sensitivity to nonverbal communication: the Pons test 1999. Baltimore: The Johns Hopkins University Press; 1999;. [5]. [5]Eide H, Frankel R, Haaversen ACB, Vaupel KA, Graugaard PK, Finset A. Listening for feelings: identifying and coding empathic and potential empathic opportunities in medical dialogues. Pat Educ Counsel. 2004;54:291–297. [6]. [6]Bylund CL, Makoul G. Empathic communication and gender in the physician–patient encounter. Pat Educ Counsel. 2002;48:207–216. [7]. [7]Ford S, Schofield T, Hope T. Observing decision-making in the general practice consultation: who makes which decisions?. Health Expectations. 2006;9:130–137. MEDLINE [8]. [8]Kurtz SM, Silverman JD. The Calgary–Cambridge referenced observations guides: an aid to defining the curriculum and organising the teaching in communication training programmes. Med Educ. 1996;30:83–89. MEDLINE |
CrossRef
[9]. [9]Roter DL. The Roter method of interaction process analysis. Baltimore: The John Hopkins University, Department of Health Policy and Management; 1991;. [10]. [10]Brown JB, Stewart M, Tessier S. Assessing communication between patients and physicians: the measure of patient-centred communication (MPCC). Working Paper Series #95-2, London, Canada: The University of Western Ontario; 2001 [11]. [11]Richard C, Lussier MT. MEDICODE: an instrument to describe and evaluate exchanges on medications that occur during medical encounters. Pat Educ Counsel. 2006;64:197–206. [12]. [12]Meakin R, Weinman J. The ‘Medical Interview Satisfaction Scale’ MISS adapted for British general practice. Fam Pract. 2002;19:257–263. MEDLINE |
CrossRef
[13]. [13]Henbest R, Stewart M. Patient-centredness in the consultation. 2: Does it really make a difference?. Fam Pract. 1990;7:28–33. MEDLINE [14]. [14]Stewart M, Belle Brown J, Donner A, McWhinney IR, Oates J, Wayne Weston W, et al. The impact of patient-centred care on outcomes. J Fam Pract. 2000;49:796–804. MEDLINE [15]. [15]Makoul G. The SEGUE framework for teaching and assessing communication skills. Pat Educ Counsel. 2001;45:23–34. [16]. [16]Siriwardena AN, Campion P, Freeman A, Elwyn G. Involve the patient and pass the MRCGP: investigating shared decision making in a consulting skills examination using a validated instrument. Br J Gen Pract. 2006;56:857–862. MEDLINE [17]. [17]Makoul G, Clayman ML. An integrative model of shared decision making in medical encounters. Pat Educ Counsel. 2006;60:301–312. [18]. [18]Charles C, Gafni A, Whelan T. Shared decision-making in the medical encounter: what does it mean? (or it takes at least two to tango). Soc Sci Med. 1997;44:681–692. MEDLINE |
CrossRef
[19]. [19]Charles C, Gafni A, Whelan T. Decision-making in the physician–patient encounter: revisiting the shared treatment decision-making model. Soc Sci Med. 1999;49:651–661. MEDLINE |
CrossRef
[20]. [20]Murray E, Charles C, Gafni A. Shared decision-making in primary care: tailoring the Charles et al. model to fit the context of general practice. Pat Educ Counsel. 2006;62:205–211. [21]. [21]Elwyn G, Edwards A, Kinnersley P, Grol R. Shared decision making and the concept of equipoise: the competences of involving patients in healthcare choices. Br J Gen Pract. 2000;50:892–897. MEDLINE [22]. [22]Elwyn G, Edwards A, Wensing M, Hood K, Atwell C, Grol R. Shared decision making: developing the OPTION scale for measuring patient involvement. Qual Saf Health Care. 2003;12:93–99. MEDLINE |
CrossRef
[23]. [23]Elwyn G, Hutching H, Edwards A, Rapport F, Wensing M, Cheung WY, et al. The OPTION scale: measuring the extent that clinicians involve patients in decision-making tasks. Health Expectations. 2005;8:34–42. MEDLINE |
CrossRef
[24]. [24]Elwyn G, Edwards A, Hood K, Robling M, Atwell C, Russell I, et al. Achieving involvement: process outcomes from a cluster randomized trial of shared decision making skill development and use of risk communication aids in general practice. Fam Pract. 2004;21:337–346. MEDLINE |
CrossRef
[25]. [25]Braddock CH, Fihn DS, Levinson W, Jonsen AR, Pearlman RA. How doctors and patients discuss routine clinical decisions—informed decision making in the outpatient setting. J Gen Intern Med. 1997;12:339–345. MEDLINE |
CrossRef
[26]. [26]Braddock CH, Edwards KA, Hasenberg NM, Laidley TL, Levinson W. Informed decision making in outpatient practice—time to get back to basics. J Am Med Assoc. 1999;282:2313–2320. [27]. [27]Winefield H, Murrell T, Clifford J, Farmer E. The search for reliable and valid measures of patient-centredness. Psychol Health. 1996;11:811–824. [28]. [28]Mead N, Bower P. Measuring patient-centredness: a comparison of three observation-based instruments. Pat Educ Counsel. 2000;39:71–80. [29]. [29]McKinstry B, Walker J, Blaney D, Heaney D, Begg D. Do patients and expert doctors agree on the assessment of consultation skills? A comparison of two patient consultation assessment scales with the video component of the MRCGP. Fam Pract. 2004;21:75–80. MEDLINE |
CrossRef
[30]. [30]Roter D, Larson S. The Roter interaction analysis system (RIAS): utility and flexibility for analysis of medical interactions. Pat Educ Counsel. 2002;46:243–251. [31]. [31] http://www.nhssurveys.org/categories.asp?parent=1212007;. [32]. [32]Strull WM, Lo B, Charles G. Do patients want to participate in medical decision making?. J Am Med Assoc. 1984;252:2990–2994. [33]. [33]Horne R, Weinman J, Hankins M. The beliefs about medicines questionnaire: the development and evaluation of a new method for assessing the cognitive representation of medication. Psychol Health. 1999;14:1–24. [34]. [34]Marteau TM, Bekker H. The development of a six-item short-form of the state scale of the Spielberger State-Trait Anxiety Inventory (STAI). Br J Clin Psychol. 1992;31:301–306. [35]. [35]Bentsen BG, Natvig B, Winnem M. Questions you didn’t ask? COOP/WONCA Charts in clinical work and research. Fam Pract. 1999;16:190–195. MEDLINE |
CrossRef
[36]. [36]Elwyn G, Edwards A, Wensing M, Grol R. Option—Observing Patient Involvement Manual, Version 2. Cardiff University; 2002;. [37]. [37]Bland JM, Altman DG. Statistical methods for assessing agreement between measurement. Biochim Clin. 1987;11:399–404. [38]. [38]Britten N, Weiss MC. What is concordance?. In: Bond C editors. Concordance 2004. The Pharmaceutical Press: London; 2004;. [39]. [39]Epstein RM, Franks P, Fiscella K, Shields CG, Meldrum SC, Kravitz RL, et al. Measuring patient-centered communication in Patient-Physician consultations: theoretical and practical issues. Soc Sci Med. 2005;61:1516–1528. MEDLINE |
CrossRef
a Department of Pharmacy & Pharmacology, University of Bath, Bath BA2 7AY, United Kingdom b Academic Unit of Primary Health Care, Department of Community Based Medicine, University of Bristol, United Kingdom  Corresponding author. Tel.: +44 1225 386787; fax: +44 1225 386114.
PII: S0738-3991(07)00341-2 doi:10.1016/j.pec.2007.09.001 © 2007 Elsevier Ireland Ltd. All rights reserved. | |
|
FULL TEXT
