Patient participation in the discussions of options in Spanish primary care consultations

Abstract

Objectives

To determine patients' participation in the discussion of options in primary care consultations. Identify the patients' wish to participate and their perceptions of their participation and explore the potential factors that may influence these.

Design

Cross‐sectional study. Setting. Ninety‐seven general practices. Participants. six hundred and fifty‐eight patients who went to their doctors for unselected reasons. Measurements. All the encounters were videoed, patient participation in decision making (DM) was assessed with two tools. After the consultation, GPs completed a questionnaire about biomedical and relational information. Patients' preferences and perception of participation was explored with different type of questions.

Results

Encounters successfully videoed: 638. Of these, only 90 interviews clearly showed patient participation. In 161 other interviews, patient participation was considered possible. Questionnaires collected: 645. In 60% of the situations (390 encounters), patients wished they could have stated their views about the proposed option(s), but they perceived this did not happen. The degree of participation at the consultation did not relate significantly with the physician's ideas about the type of problem, evolution and treatment. Neither did any of the considered variables influence either the patients' wish to participate in the discussion of the suggested option or their perception of this.

Conclusions

GPs ask patients for their opinion and promote discussion about the suggested plan in few encounters. Patients perceive this, including many patients that previously had declared not to be interested in being involved in decisions. These results revealed an important mismatch between what patients wish and what they perceive.

Introduction

The idea that patients should be involved in decisions about their treatment or care has been increasingly advocated for the past few years and is currently widely accepted. Nevertheless, published studies so far reveal that patient levels of desire for involvement in decision making (DM) differ significantly. For instance, in surveys performed amongst the general population, Levinson et al.1 found that 52% of respondents exhibited a greater preference for leaving the final decision to the doctor. In the same way, Coulter and Jenkinson,2 using a population survey, explored what people from eight European countries thought about who should take the lead in making treatment decisions, and patients' expectations of involvement in decisions about their care were also wide.

In this realm, an important point is that variability in patients' desires to be involved in DM is partly due to either their difficulties to interpret the questions on the surveys3 or other shortcomings of these tools for quantifying this phenomenon,4 but also to a wrong differentiation between diagnosis and decision‐making tasks.3 An additional issue of great importance in this debate is that patients and physicians seem to hold different views as to what the concept of participation in DM means. On one hand, even if there is considerable overlap between the constructs with similar connotations,5 there seems to be agreement amongst theoreticians and professionals that sharing decisions with patients requires the doctor to communicate the treatment options effectively and portray the risks and benefits of each option.6, 7, 8 This conceptualization underpins several variant models of patient involvement, including those referred to as ‘Shared decision making’ (SDM) and ‘Informed choice’. However, other authors have highlighted the problems of basing participation in the presentation and discussion of a menu of options and suggested the need for a broader conceptualization that can focus attention on the relational and subjective‐affective aspects of involvement.9, 10 In relation with this perspective, we have emphasized the relevance of a ‘Participatory Decision‐Making’ model 11 where factors such as the clinicians' communicative efforts to understand patients' perspectives and concerns, the rapport developing and the promotion of mutual trust are of crucial importance. Along this line, Epstein introduced the concept of ‘Shared Mind’ 12 to define the way how new ideas and perspectives can emerge from sharing thoughts, feelings, perceptions, meanings and intentions amongst the patient and physician, highlighting the complexity of the preference construction process. These proposals try to conciliate the doctors' perspectives with those many patients seem to have in relation with the concept of involvement in decision making. Indeed, patients seem to appreciate the informed choice of options less than other key attributes of the consultation, such as the efforts resulting in the physician knowing every patient individually,13, 14 getting relevant information about the whole process 13, 15 or perceiving that they contribute significantly to the discussion.13, 15, 16, 17 Actually, patients use these attributes to define what they consider ‘involvement in decisions’.13, 14, 16

An additional very important issue, yet not sufficiently tackled in all its complexity, is that originally, this SDM model was developed to address single event high‐risk problems and uncertainty faced by patients and physicians who do not know each other very well. In these situations, explicit agreement on which decision or plan of care should be taken would usually be considered as necessary. However, this assumption does not apply to many of the patient–physician discussions about options in primary care. In this setting, the bulk of decisions are clearly minor decisions (low‐risk problems) even if one of them is definitely better. Here also, for many other choices, a concise description of the proposed plan could be adequate. In these situations, few patients are likely to be willing to make any decision themselves, and generally, their doctors do not ask them to voice their views or values.18

Often times, organizational factors make patients and doctors in primary care have a long‐term continuous healing relationship across multiple encounters. These relationships enable the judicious use of communication skills tailored to the patients' situation and usually facilitate openness, respect and trust.19

In this context even more than in others, the importance of specific verbal or non‐verbal actions – that is, attentive listening, expressions conveying empathy, paying attention to a specific clue, invitation to make a discrete comment, exploration of emotions, expressions of curiosity – may eventually encourage not only shared information but also shared deliberation11, 20 Therefore, these ‘aspects which facilitate personal communication’ can be important for patients and clinicians to perceive a particular interaction as participative. Research in this field is certainly open. With these perspectives, it should focus on the very specific, simple and unambiguous conducts that arise at particular moments in a consultation as well as on the patients and clinicians' perceptions and experiences of these conducts.

Having these ideas in mind, we designed and conducted a study with the following aims: to determine the patients' participation in the discussion of options in primary care consultations, identify their wish and perceptions of their participation and explore the potential factors that could influence them.

We carried out this study on a sample of patients who went to their general practitioners (GPs) for unselected reasons using different methods for measuring: the external assessment of different types of communicative behaviour at consultations and the doctors and patients' opinions gathered through a variety of questions.

Methods

A cross‐sectional study was designed. We estimated a minimum of 540 patients needed to be recruited to get a precision of ±4%, a 95% level of confidence and 50% of expected proportion (the most unfavourable situation). On the one hand, each of the 96 clinicians was selected by stratified sample at every Clinic and on the other hand, a researcher selected 6–8 patients from each physician's daily list of patients using systematic sample. Patients went to see their doctors for problems related to a chronic condition or for a new health problem. Family doctors were recruited who had more than 5 years of clinical experience and worked in health‐care centres in five different provinces of Spain. All consultations where a treatment decision was making were included in the study and videoed.

Measurements

Assessment of behaviours

Patients' involvement in DM was initially measured through a specific item from a scale that had been designed to measure the extent of patient centeredness.21 Item no. 8 on this scale assesses how much a provider promotes patient involvement in DM by means of a three‐level scale (‘To what extent did the provider give the patient opportunities to participate in the decision‐making process and encourage them to do so?'), 0: conduct absent; 1: probably present; 2: definitively present). The concept of involvement as defined by this item considers that this may exist if the patient responds to a single suggestion by the physician, provided this has opened the discussion and interest in the patient's opinion about it has been shown.22 The intra‐observer reliability of this item from a similar sample of 30 encounters had a Cohen's Kappa value of 0.72. All clinical encounters scoring 1 or 2 for this item plus those from a randomized sample selection scoring zero had behaviour assessment through the CICAA‐D scale (Table 1). There are three domains in this scale and 17 items. Each one of these has an answer range of three points: 0‐behaviour not detected; 1‐possibly present; 2‐definitely present. The third component of the scale (four items) is assessed only if item no. 8 scores 2; this means that involvement has been detected and more than one possible option has arisen. A group of experts selected the behaviours that make up the CICAA‐D scale in a process that was described elsewhere.23 A trained rater assessed all these interviews. The intra‐observer reliability was obtained through a test–retest in 25 interviews that were performed at 3‐month intervals. The Cohen's Kappa values for each one of its items were: <0.4 (items 11, 15), something between >0.4 and <0.7 (items 1,3,4,5,6,7,12,13,14,17) and >0.7 (items 2,8,9,10 and 16), The global intraclass Correlation Ratio was 0,96 (IC 95%: 0.93–0.98). The study followed the scheme as shown in Fig. 1.

Table 1.

CICAA‐D: Frequency of appearance of behaviours as defined in each of the items

Components	Scale Item
	0(%)	1(%)	2(%)
To what extent did the provider…
1st.‐Identify and understand the problems
1. …explore the emotions and/or feelings that the symptom, process or treatment proposal has caused the patient?	62.3	23.5	14.2
2. …explore the patient's expectations for this visit?	55.5	37.7	6.8
2nd.‐ Reach agreements and help to act
3. …try to explain the process or patient's main symptom?	30.7	41	28.3
4. …properly define the problem regarding the decisions that should be taken?	63.1	28	8.9
5. …explain their treatment proposal?	39.1	43.4	17.5
6. …give information adapted to the patient's problems and needs?	62.3	32.3	5.4
7. …give information the patient could understand?	15.9	69.8	14.3
8. …give the patient opportunities to participate in the decision‐making process and encourage them to do so?	60.7	25.2	14.1
9. …allow the patient to express their doubts?	21.8	34.2	44
10. If some discrepancy appeared, to what extent did the provider try to reach an agreement (by discussing and considering the patient's views)?	69.8	22	8.2
11. …make sure the patient understood the information they were given?	33.4	48	18.6
12. …allow for decision making or deferment?	39.6	46	14.4
13. …get explicit commitment from the patient as to the follow‐up management plan?	46.1	35.3	18.6
3rd‐ Decisions with optionsa
14. …display or encourage the display of possible diagnostic/treatment options?	32.2	46.7	21.1
15. …inform about the different options?	65.6	27.8	6.6
16. …allow the patient to ask questions about the options or the decision‐making process?	58.9	31.1	10
17. …explore the patient's preferred level of involvement in decision making?	53.3	28.9	17.8

^aAssessed only if item no. 8 scores 2.

Figure 1.

Study scheme for the analysis of videos.

Patients and physicians opinions

Figure 2 shows the sequence of questions to the patients. An introductory question explored the patient's opinion about who should make decisions in the consultation (P1) – the physician should decide/by joint agreement between physician and patient/the patient should decide. The patient's perception of the DM process was assessed immediately after consultation by a question exploring whether the patient wished the physician had asked for their views about the suggested option/s in the consultation (P2) (In today's consultation, would you have liked your doctor to ask you for your opinion about what you prefer to do?). A negative answer to this question meant the patient preferred no to take part in DM, and a positive answer that they preferred to be involved in DM. When patients answered positively, their perception of involvement was assessed by asking them whether they believed the physician had encouraged them to give their opinion or not (P3). If they answered positively, this meant they preferred to take part and had a positive perception that they had, whilst a negative answer meant they preferred to participate but perceived they had not. The latter was considered as a discrepancy.

Figure 2.

Study scheme for the patient questions.

Physician' opinions about some aspects that, theoretically, might affect the DM process were collected through a questionnaire that mainly consisted of ten‐level visual analog scales. The physician filled in this questionnaire immediately after the consultation. Such aspects included the physician's certainty about the nature of the health problem and its clinical evolution; the usefulness of the suggested diagnostic tests and the effectiveness of the treatment option they had chosen; the existence of other treatment options and the risk both the problem and selected treatment might pose to the patient. Moreover, physicians also provided their opinion about some aspects related to their acquaintance with the person and their relationship with each particular patient (type of relationship and trust) (Table 2).

Table 2.

Results of the questions asked to the physician in the questionnaire after consultation

Explored Domain	Level on visual analogue scale
	Mean	n	Low (0–3) n (%)	Moderate (4–7) n (%)	High (8–10) n (%)
Certainty about the nature of the problem	7.8	630	24 (3.8)	195 (31)	411 (65.2)
Certainty about the problem's evolution	6.9	630	40 (6.4)	302 (48)	288 (45.6)
Potential risk for patient	4.2	624	294 (47.1)	242 (38.8)	88 (14.1)
Usefulness of the Test for the Diagnosis	7.1	192	17 (8.9)	68 (35.4)	107 (55.7)
Possible effectiveness of treatment	7.2	541	19 (3.6)	240 (44.3)	282 (52.1)
Possible risks of treatment	3.3	543	334 (61.5)	194 (35.7)	15 (2.8)
Physician–Patient relationshipa	7.6	619	2 (0.4)	259 (41.8)	358 (57.8)
Patient's trust in doctor	7.5	621	7 (1.1)	258 (41.5)	356 (57.4)
Doctor's knowledge of the patientb	–	640	27 (4.2)	192 (30.1)	405 (63.3)
Possibility of other handling optionsc	–	559	No	Yes, some	Yes, several
			132 (20.6)	320 (50)	107 (16.7)

^aLow: bad/Moderate: average/High: good or very good.

^bLikert scale (Low: new patients).

^cLikert scale.

The statistical analysis consisted of a descriptive study of the measured variables and an inferential analysis. The tests used were Chi‐square, Student's t‐test and anova. Every contrast was bilateral and for a P ≤ 0.05. As different factors or variables might have an effect on the patient's participation in the discussion – on one hand – and on the patient's wish and perception of having participated – on the other hand – that relationship was studied through multivariable analyses of multiple logistic regression and multiple linear regression. The starting point in both analyses was a maximum pattern and in a step by step method, those variables that failed to reach the significance level were crossed‐out (starting criterion: P < 0.10). The goodness of fit of the different models was checked out.

The study was approved by the Ethical Commission of Cordoba Primary Care Administration Centre. All the participants – doctors as well patients – were informed and handed in the informed consent form.

Results

Ninety‐seven GPs took part in the study from 36 urban health centres and 22 rural health centres. Sixty percentage (60%) of them were men. On average, they had 20 years' clinical experience and a list of around 1500 patients.

Sixty‐five (9%) of the 723 patients who met the inclusion criteria and were invited to participate did not do so because they either failed to give their consent – they felt mainly uncomfortable about being videoed – or quit after waiting for a long time to be seen by the interviewer. Finally, 658 patients (91%) took part in the study, 401 (62%) were women. The patients' education level was as follows: 13% no studies, 43% primary school, 28% high school and 16% university. The age range spanned from 18 to 88, the average being 52 years old. Patients came from urban (70%) and rural (30%) areas. The sociocultural level was higher amongst patients from urban areas (9 vs. 31% primary school and 32 vs. 19% high school; P = 0.001). Their health problems were classified according to the International Classification of Primary Care, second edition (ICPC‐2 http://www.who.int/classifications/icd/adaptations/icpc2/en/index.html) and the most frequently attended to were cardiovascular (175), emotional and psychological (124), respiratory (48), digestive (45), neurological (40) and unspecified (38). Consultations lasted an average of 8.3 min (range: 1–27.3).

Analysis of video recordings

We obtained 638 good‐quality videos (seven were discarded owing to their bad quality, and in 13, we get no videorecordings). Figure 1 shows the interviews that were finally analysed with the different tools (only item no. 8, CICAA‐D 13 items and CICAA‐D 17 items) according to the methodological proposed scheme. Absolutely no participation was found in 387 interviews (60.6%) – these scored zero in item no. 8. In this group, 117 interviews were selected at random and analysed through the CICAA‐D. Item no. 8 scored positively in the remaining 251 interviews (39.3%); in 161 of these (64%) involvement was considered ‘possible’ (scored 1 in item no. 8) and in the remaining 90 (36%), the patient's involvement was more clear (scored 2 in item no. 8). The patients in these interviews showing clear or possible participation were younger than those patients whose interviews showed no involvement (50.7 vs. 53.5; P = 0.035). The average score achieved by the CICAA‐D in the group of analysed interviews (368) was 10 points (34 possible points), CI 95%: 9.40–10.52.

Table 1 shows how frequently the different behaviours that make up the CICAA‐D were found in the analysed interviews.

If we just consider the common items (1st and 2nd dimensions of the CICAA‐D: items 1–13, 26 possible points), the average score of interviews where no involvement was found was 6.3 (IC 95% 5.83–6.93). The average score of interviews where possible involvement was found was 9.8 (IC 95% 9.28–10.45), and the average score of interviews where involvement was definitely found (F: 88.5; P < 0.001) was 14.8 (IC 95% 13.38–16.21). Interviews where patient's involvement was present lasted longer (10.2 min) than those where participation was possible (9.1 min) or non‐existent (7.5 min) (F: 20.62; P < 0.001). Involvement was greater in consultations about neurological problems (87.5%), skin diseases (60%) and cardio circulatory problems (45%) than it was in consultations about nose‐/throat‐/ear‐related problems (20%), digestive (20%), emotional/psychological (22.5%), metabolic (31.5%) and breathing problems (31.3%) (Chi‐square: 115.9; P < 0.001). Eight of the 13 types of behaviour included in the scale were significantly more common in consultations with possible or certain participation (item no. 8: 1 or 2). On average, the CICAA‐D score was higher in urban than rural interviews (10.4 vs. 8.8; P = 0.019) just like the extent of participation that was found (score 1 or 2 in item no. 8: 189 (42%) vs. 62 (32%); P = 0.017).

Physicians' and patients' opinions

Table 2 shows what physicians think about different aspects of the consultation and their relationship with their patient. In most consultations, GPs were quite sure about the problem the patient had, its evolution and the effectiveness of treatment. They considered their relationship with their patient good or very good and believed most patients trusted them. More than half the doctors surveyed were acquainted with their patients and thought that more than one treatment option was possible. In 380 visits, GPs did not require a diagnostic test and in 71 did not prescribe or advise any kind of treatment.

Six hundred and forty‐five (645) patients answered the questions; 41% of them thought that the doctor should only make the decisions during the consultation, 57% believed that decisions should be made between both, and ten patients stated that decisions should be made by patients only (Table 3). About 60.6% of the patients' answers to the second question showed that they would have liked the physician to have asked them for their opinion about what could be done, but the doctor did not encourage them to participate. This revealed a discrepancy between the patients' preferences to contribute their views about the suggested option(s) and their perception of their involvement in the consultation. The remaining patients either answered this question negatively (26%) or confirmed that the physician had indeed asked or encouraged them to give their opinion (13.6%).

Table 3.

Patients' initial preference as to who should make decisions in consultations and their wish/perception after consultation is over, that the physician had invited them to express their opinion about the plan to follow

Patients' generic preferences about who should do the decision making	Wish/perception that they had expressed their opinion in consultation
	n (%)a	No	Yes, (and perceives the physician encouraged them to do it)	Yes, (and perceives the physician did not encourage them to do it)
Physician should decide	266 (41.2)	99 (37.2)	37 (13.9)	130 (48.9)
Physician–patient joint consent	369 (57.2)	68 (18.4)	50 (13.6)	251 (68)
Patient should decide	10 (1.5)	0 (0)	1 (10)	9 (90)
n	645 (100)	167 (26)	88 (13.6)	390 (60.4)

Chi‐square: 34.66; P < 0.001.

^aIt represents the percentage of the whole sample. The rest of percentages in brackets are based on the total in each of the groups.

Table 3 records the patients overall opinion as to who should do the DM and their individual wish after consultation that the physician should have invited them to give their opinion about the plan to be followed. Patients who most wished they could have given their opinion were those who favoured the doctor–patient joint agreement and those who preferred that only the patient should decide. However, up to 63% of patients with a ‘paternalistic view’ that the physician should do the DM wished they could have been able to take part in the consultation and nearly half did not perceive the doctor had invited them to do so.

Analysis of factors related to patient participation

In the bivariate analysis, only those consultations where doctors declared to be more certain about the evolution of the problem (7.1 vs. 6.7; P: 0.04) and suitability of the prescribed test to diagnose the problem (7.7 vs. 6.5; P = 0.001) were less participatory. Every question on the clinical questionnaire (Table 2) along with the patient's sociodemographic variables (age, gender, marital status and level of education) and the doctor's variables, plus the patient's opinion prior to consultation as to who should make the decisions, were included as independent variables in a multivariable model that explored its association to the level of participation as observed at consultations, and no statistical association could be seen with any of the considered variables. Neither the patient wish to participate in the consultation nor their later perception of this could be related in a multivariable model to any of the following explanatory variables included in the model: the full list of CICAA items, the scale's total score, the other patient and doctor's sociodemographic variables and the patient's previous views about how DM should be done at consultations.

Discussion

This study reveals that Spanish GPs ask for the patient's opinion or encourage discussion concerning the suggested treatment in few occasions. Patients perceive this, particularly those who wish to share decisions with their doctors. The study also highlights that patients still want to be given a chance to express their opinion as regards options – even if there is realistically only one choice – and points out that this desire is not associated to any of the factors related to the characteristics of the patient, the doctor–patient relationship or the type of health problem. Finally, the findings did not show whether the patient's level of participation might be influenced by the doctor's ideas about the clinical problem, the doctor's relationship with the patient or any other characteristics of the patient.

The ratings of the videotaped physician‐patient interactions found that in almost two‐thirds of the interviews, physicians did not encourage any kind of discussion regarding treatment, and only in 14% of the situations lively discussion occurred and alternative options were given. Involvement – both actual and preferred involvement – in decisions may happen at different levels depending on the nature of the condition, its severity and chronicity.17, 24 In spite of this, and whilst the study was performed on an unselected group of primary care interviews where patients' ‘active participation’ might not have been suitable for many of their problems,18 this work showed that the levels of involvement were low. Results of other studies regarding patient involvement in primary care consultations using external assessment of behaviours also show low involvement.25, 26, 27, 28 Nonetheless, in this study, we considered that involving patients in DM may simply consist in the physician making a proposal and inquiring – either superficially or in depth – so as to find out the patient's opinion concerning such proposal. Some physicians have put this strategy forward as a form of involvement,22 and likewise, many patients talk of having been involved in DM when actually, they had not been given information about a menu of treatment options and had not influenced the selection of a course of action significantly but they contributed significantly in the discussion.13, 14, 15, 17, 29, 30 This behaviour can eventually be a way of showing interest in the patients' views and become the beginning of information exchange and deeper discussion and participation.11

Our findings could be seen as supporting this perspective: looking at it realistically, even when only a better option exists or/and there are not any decisions at risk, as is common in Primary Care settings, three of four of our patients still would want to be given a chance to express their opinion. Besides, interestingly enough, most patients who at first had a ‘paternalistic opinion’ about who should do the DM (63%) later declared that they wished they had also had their say as regards the decision. We do not know what participation in DM meant for these patients, but up to now our categorization of ‘involvement’ as an attempt to elicit patients' perspectives on wether and how well they were involved has often focused on the question ‘who made the decision?’, because we asked respondents to pick one from a menu of descriptions as we did at the beginning of the consultation. Possibly, the message patients were conveying here is that they would have valued the fact that the physician showed interest in their ideas and allowed them to express their opinion as more relevant for them than the physician explaining several options or putting on to them the responsibility to make a final decision.14, 31 This finding also supports the idea that even if many patients in surveys indicate they would prefer it if the health professional took the lead in ‘making’ treatment decisions,32, 33 this should not necessarily be interpreted as evidence that patients do not want to be or feel involved in decisions about their care. It may reflect differences between the patients' understanding and the dominant conceptualization of patient involvement in DM.9 It is also possible that the very patients in our study who had paternalistic ideas about DM could change their views after a particular consultation provided they were specifically invited to participate. This finding may reflect the influence of experience on their preferences. This has been highlighted in other studies.15, 34 Making the patients aware of the possibility that they can participate in a particular decision (by asking them to) or giving them opportunities to participate might be a way of ‘activating’ them, that is to say, make them become more effective at participating in their own care.

Furthermore, – and contrary to the results of other studies 1, 34, 35, 36 – our results show that when considering the wish to participate as the desire to ‘contribute’ and ‘express one's views’ regarding the action to take, neither the patient's level of education nor their age seem to make any difference. The patient's participatory experience is not related to any of the behaviours included in the CICAA‐D, even though other predominant theoretical models 37, 38 have considered many of these conducts important to the SDM process. This lack of relation between the direct observation of conducts and the patient's experience has also been shown through the exploration of the disease experience 39 or partnership itself.40 Moreover, if we consider that patients' apparent perceptions of the consultation outcomes are better than the observation of conducts, these results back up the importance of factors associated with relationship dynamics such as trust, power and other affective factors, in the communicational process in general and the participatory process in particular. Besides our work and the CICAA items were focused on the selection of one option (or more) and its discussion rather than on the patient's possible engagement in decision about the nature of their problem,41 the question of who defines the options set 10 and/or others key relational and subjective‐affective aspects of patient involvement (as demonstrated by several qualitative studies).

The other most interesting finding of this study is probably that in 60% of the encounters, patients wished they had given their views at the time of making a decision but felt the doctor did not give them an opportunity to. Such a mismatch between what patients wished and perceived is similar to other authors' findings in the same clinical setting 35 and shows how unsuccessfully this task is still accomplished in Primary Care consultations.

Lastly, our study also aimed at exploring the influence some consultation‐related factors and the physician–patient relationship might have on the origin of the hidden decisions GPs took regarding participation. Neither the seriousness of the condition nor the risk of treatment for the patient, as perceived by the physician, seemed to have any influence on involvement, yet both these factors have been considered a useful guide to participation in the DM process.18 The doctor's initial diagnostic certainty did not make any difference. Our physicians classified most of their encounters as having either moderate risk and certainty or low risk and high diagnostic and therapeutic certainty. Both these situations make up the bulk of decision‐related situations in primary care. In the first group of situations, decisions should be negotiated; in the second group, the way of sharing DM depend on the specific choice at hand.18 However, in almost every situation, it would have been appropriate for the physician to prompt the patient's opinion about the proposal for action – as defined in item no. 8 level 1 – (even in situations such as when a physician tells a diabetic patient with a generalized malaise, ‘I'd like to see you again in a couple of days’). This is especially true because the majority of patients wished for this and in more than 75% of the consultations different options for treatment existed. As for the physician–patient relationship, which doctors themselves rated as good or pretty good, it did not seem to influence them encourage participation. Judging by these results, perhaps physicians should be aware that it is necessary to look further into their patients' desire to participate.

Limitations

A major drawback of this study which may contribute to the disagreement between conducts directly observed and included in the scale and the subjective experience of participation is that the range of these conducts is limited and incomplete. If – as we pointed out earlier – we consider patient involvement in DM mainly as a process of information and significant discussion for the patient, perhaps other types of conducts should be emphasized; for example, the active exploration by the physician of their patient's ideas, fears and expectations on the problem and its consequences rather than just what action to take or the physician's answer to specific clues given by their patients42, 43 as opposed to the discussion on options. Likewise, the lack of important associations between the variables of the consultation process represented by the observed behaviour and the patient's desire to participate may also be affected by the absence of this type of affective conduct that represents involvement in a better way.11 The tools that are to be designed to encourage patient involvement in DM should take this into account, yet even if they do so, the probability of them capturing the essence of ‘being in relation’44 as represented by the consultation, will still remain low. It should also be made clear that this analysis reduces the number of cases considerably as it only includes those encounters where all conducts had been directly observed. (This happened in 368 interviews only not in all of the 638).

We have highlighted throughout the discussion that we lack tangible information from patients as to what sharing decisions or participating in DM really means for them. Although there are several theoretical approaches to this,6, 7, 8 our study considers that a simple request by the physician to the patient for their opinion regarding any of the proposed action is in fact a form of involvement. This may be insufficient in many situations; yet we insist on the validity of this approach for many of the decisions taking place in Primary Care settings. Given how little the doctor uses this strategy; this approach is relevant from an educational point of view and is more realistic as a first step to foster patient involvement in this clinical context.

We do not know whether the encounters where patients perceived participation were similar to the models put forward by these authors or not. A more thorough analysis of consultations might clarify this point. Finally, although this study has been conducted in a sample of doctors and patients from Spain, we think that these results could be generalized to other western countries.10, 11, 27, 28

Conclusions

Promotion by physicians of discussion on the suggested plan for action through the exploration of the patient's opinion about it is scarce. Patients perceive this, particularly those who think that DM should be shared. Yet, many patients for whom the physician should make the decisions, once the consultation was over would have wished to be able and express their views about the proposed options; this can be considered as a wish to be involved in DM. GPs should bear in mind that this desire is irrespective of the patients' age or education.

Sources of funding

This study received financial support from FIS (Spanish Fund for Health Research), Andalusia Health Regional Ministry and the Spanish Society of Family and Community Medicine (semFYC).

Conflicts of Interest

None.