Article Text
PDF
Abstract
Objectives Although shared decision making (SDM) is advocated in rheumatoid arthritis (RA) treatment, it is largely unclear when, how and to what extent SDM is applied in routine clinical care of patients with RA. This study aimed to investigate the level of SDM in RA treatment from an observer perspective and to assess associations between the level of SDM and characteristics of the clinician, patient and consultation.
Methods The level of SDM was investigated by scoring audio-recordings of 168 routine consultations with unique patients with the observer patient involvement (OPTION) scale (scale 0–100, higher OPTION scores indicating higher levels of SDM). Associations between the level of SDM and characteristics of the clinician, patient and consultation were assessed using multilevel modelling. Statistical significance was set at p<0.05.
Results The mean OPTION score was 28.3 (SD=15.1). The multilevel model included four characteristics: clinician age, patient age, consultation duration and type of treatment decision. There were significant, positive associations between the level of SDM and the consultation duration (b=0.63, 95% CI 0.16 to 1.11), decision for stopping and/or starting medication (b=14.30, 95% CI 5.62 to 22.98), decision for adjusting medication doses (b=8.36, 95% CI 3.92 to 12.81) and decision for administering single dose glucocorticoids (b=15.03, 95% CI 9.12 to 20.93). Thus, a higher level of SDM was significantly associated with a longer consultation duration and the type of treatment decision. No other significant associations were found.
Conclusions Overall, the level of SDM in RA treatment leaves room for improvement. To foster SDM in routine clinical care, training programmes on patient-centred communication skills may be helpful.
- rheumatoid arthritis
- qualitative research
- DMARDs (synthetic)
- DMARDs (biologic)
Statistics from Altmetric.com
Key messages
What is already known about this subject?
Studies investigating clinicians’ and patients’ perceived level of shared decision making (SDM) in rheumatoid arthritis (RA) treatment suggest that SDM applied only to a limited extent.
Empirical studies investigating the level of SDM in RA treatment from an observer perspective are needed to provide a more accurate and unbiased picture of current practice.
What does this study add?
Using an observer-based measure, a low to moderate level of SDM in RA treatment was found.
The efforts of clinicians to involve patients in decision making varied substantially and a higher level of SDM was significantly associated with a longer consultation duration and the type of treatment decision.
How might this impact on clinical practice or future developments?
Overall, the level of SDM in RA treatment leaves room for improvement. Targeted strategies to support the application of SDM (eg, training programmes on patient-centred communication skills) are warranted in this era of patient-centred care.
Introduction
Disease-modifying antirheumatic drugs (DMARDs) are the cornerstone of rheumatoid arthritis (RA) treatment. A variety of DMARDs are available, each with different risks and benefits to consider. However, a substantial proportion of patients with RA do not adhere to DMARDs, resulting in increased disease activity and radiological progression.1–4 It is increasingly recognised that patient preferences play an important role in adherence. Matching treatment to patient preferences has been shown to improve adherence.5 6 Previous studies have shown that patient preferences for RA treatment vary widely and often differ from those of clinicians.7–11 Therefore, it is important that clinicians and patients make treatment decisions together using the best available evidence and accounting for patient preferences.12 Indeed, international guidelines for RA treatment state that all treatment decisions should be made through shared decision making (SDM) between clinicians and patients.13 14 Although SDM is advocated in RA treatment, it is largely unknown when, how and to what extent SDM is applied in routine clinical care of patients with RA.
Previous studies have investigated clinicians’ and patients’ perceived level of SDM in RA treatment, suggesting that SDM is applied only to a limited extent.15–18 In a study of 157 clinicians treating patients with RA, 27% of clinicians reported to apply SDM.18 Likewise, in a study of 892 patients with RA, psoriatic arthritis and ankylosing spondylitis, 43% of patients reported to perceive no involvement in decision making.17 However, clinician-reported and patient-reported measures may not accurately reflect reality. Clinicians tend to overestimate their efforts to involve patients in decision making, while patients may not fully understand and identify SDM if they have not perceived it previously.19–23 Consequently, they are likely to report how satisfied they were with the consultation in general instead of their perceived level of SDM. Although observer-based measures are considered more valid, research using these measures is lacking in the field of RA. Thus, empirical studies investigating the level of SDM in RA treatment from an observer perspective are needed to provide a more accurate and unbiased picture of current practice.
Finally, insight into factors associated with the level of SDM is of great value for the development of targeted strategies to support the application of SDM. Previously, studies in various clinical settings (eg, primary care, anaesthesiology and vascular surgery) have identified a number of factors associated with the level of SDM.22 24–26 These factors were related to characteristics of the clinician (eg, sex), patient (eg, age) and consultation (eg, duration). However, results were inconsistent between studies and may not be generalised to RA treatment. Therefore, the aim of this study was twofold: first, to investigate the level of SDM in RA treatment from an observer perspective and second, to assess associations between the level of SDM and characteristics of the clinician, patient and consultation.
Methods
Study design
This study had a cross-sectional design and was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology statement.27 This study used data that were collected, but not used, in a large, longitudinal study of van Heuckelum and colleagues.28 In their study, van Heuckelum and colleagues explored the role of implicit and explicit attitudes towards synthetic DMARDs as possible target for improving medication adherence in RA treatment. This study, however, focused on SDM. We used data (eg, audio-recorded consultations) that were not previously analysed and/or published.
Participants
The study of van Heuckelum and colleagues was conducted in the two largest rheumatology centres in the Netherlands: Reade in Amsterdam and the Sint Maartenskliniek in Nijmegen. Recruitment took place between July 2016 and November 2017. First, all clinicians practicing as a rheumatologist or physician assistant in one of the two rheumatology centres were invited for participation. Thereafter, participating clinicians invited a consecutive series of patients after verifying their eligibility. Patients were eligible for participation if they were aged 18 years or older, diagnosed with RA by a rheumatologist and using at least one synthetic DMARD.
Data collection
Characteristics of the clinician and patient
Following recruitment, clinicians completed a short questionnaire on sociodemographic characteristics: rheumatology centre, sex, age, profession and work experience. Patients also completed a short questionnaire on sociodemographic characteristics: sex, age and educational level. Clinical characteristics (disease duration, presence of comorbidities, number of synthetic DMARDs in use, use of biologicals and use of glucocorticoids) were extracted from patients’ medical records. All comorbidities registered in a patient’s medical record were included, regardless of type and severity. Furthermore, patients’ medication beliefs were assessed with the Beliefs about Medicines Questionnaire.29 This 10-item questionnaire comprises two subscales, assessing patients’ perceived need to take medication (necessity subscale) and their concerns about potential adverse consequences (concerns subscale). Scores range from 5 to 25 with higher scores indicating stronger beliefs. The health status of patients was assessed with the Dutch consensus version of the Health Assessment Questionnaire.30 The range of scores is between 0 and 3 where 0 represents no disability and three complete disability.
Characteristics of the consultation
In addition, routine consultations were audio-recorded. Per patient, one audio-recording was made during a single consultation. Audio-recordings were checked for sound quality and completeness. Audio-recordings of poor sound quality and the ones not containing the full consultation were removed. The duration of each consultation was registered. The type of treatment decision that was made during the consultation was also registered. This study only took treatment decisions regarding synthetic DMARDs, biologicals and glucocorticoids into account. Treatment decisions were categorised into four types: stopping and/or starting medication, adjusting medication doses, administering single-dose glucocorticoids and continuing medication. The first two categories refer to changes to a patient’s treatment. The third category refers to the additional administration of single-dose glucocorticoids. This type of medication was usually administered during or shortly after the consultation for short-term control of flares. The fourth category refers to the unchanged continuation of a patient’s treatment. If multiple treatment decisions were made during the consultation, only one type of treatment decision was registered, following the above-mentioned sequence. For example, if it was decided to increase the total weekly dose of methotrexate and the patient also received an intra-articular injection of prednisone, the type of treatment decision was registered as adjusting medication doses.
Level of SDM
The level of SDM was investigated by scoring the audio-recorded consultations with the five-item observing patient involvement (OPTION) scale.31 The OPTION scale is a reliable and valid measure, used across various clinical settings (eg, primary care, anaesthesiology and vascular surgery) to assess the efforts of clinicians to involve patients in decision making from an observer perspective.22 24–26 32 This measure is based on a conceptual framework describing five core dimensions of SDM: justify deliberative work (item 1), justify deliberative work as a team (item 2), inform, describe options and exchange views (item 3), elicit preferences (item 4) and integrate preferences (item 5). Each item was scored on a five-point scale from 0 (no effort) to 4 (exemplary effort). Item scores were summed to obtain an overall score. The overall score, or OPTION score, was rescaled to lie between 0 and 100. Higher OPTION scores indicate higher levels of SDM. To ensure a reliable assessment, two observers (EM and FK) completed an online training provided by the developers of the OPTION scale.33 Guided by a previous study of Stubenrouch and colleagues, the observers independently scored ten random consultations and percentages of absolute agreement and kappa values were calculated for each item to assess inter-rater reliability.22 According to predetermined cut-offs of 80% (percentages of absolute agreement) or 0.6 (kappa values), an acceptable level of agreement was reached.34 35 Therefore, the remaining consultations were divided and scored by only one observer (EM scored 65% of consultations and FK 35%).
Data analysis
Descriptive statistics were computed for characteristics of the clinician, patient and consultation. Continuous variables were reported as means and SD or medians and IQR, depending on the normality of their distribution. Categorical variables were reported as frequencies and percentages. The mean OPTION score, range and mean item scores were reported. In addition, the median OPTION scores per clinician were reported and presented in box plots. The data had a hierarchical structure with patients (level-1) nested in clinicians (level-2). Therefore, multilevel modelling was used. To account for the number of events per variable, we first used univariate regression analyses to assess associations between the level of SDM and characteristics of the clinician (rheumatology centre, sex, age and work experience), patient (sex, age, educational level, disease duration, presence of comorbidities, medication necessity beliefs, medication concerns and health status) and consultation (duration and type of treatment decision). Thereafter, factors with a p value <0.2 were included in a multilevel model with random intercepts. Model fit was evaluated using likelihood ratio tests. Statistical significance was set at p<0.05. As for characteristics of the clinician, work experience was not included in the multilevel model because of multicollinearity with age (r=0.9). All analyses were performed using Stata V.13.
Ethical considerations
Clinicians and patients gave re-consent to use the data for the purposes of this study. Handling of the data complied with the General Data Protection Regulation (GDPR) and the Dutch act on the implementation of the GDPR.
Patient and public involvement
There was no patient and public involvement in this study.
Results
Characteristics of the clinician, patient and consultation
One hundred and seventy-nine routine consultations with unique patients were audio-recorded. Eight audio-recordings were removed because of insufficient sound quality and/or incompleteness. In addition, three patients did not give reconsent to use the data for the purposes of this study. This brought the total number of audio-recordings to 168. Characteristics of the clinician, patient and consultation are shown in table 1.
Characteristics of the clinician, patient and consultation
Level of SDM
The mean OPTION score was 28.3 (SD=15.1) (scale 0–100, higher OPTION scores indicating higher levels of SDM). The range of OPTION scores was between 0 and 75. Disregarding consultations in which the type of treatment decision was registered as continuing medication, the mean OPTION score was 35.9 (SD=14.6). On a scale from 0 to 4, the mean scores for item 1–5 were, respectively 1.0 (SD=1.0), 1.0 (SD=0.7), 0.8 (SD=0.9), 1.2 (SD=0.7) and 1.6 (SD=0.7). Quotes from the consultations that illustrate the efforts of clinicians to involve patients in decision making are shown in Box 1.
Quotes from the consultations
Item 1: justify deliberative work
For the health issue being discussed, the clinician draws attention to or confirms that alternate treatment or management options exist or that the need for a decision exists. If the patient rather than the clinician draws attention to the availability of options, the clinician responds by agreeing that the options need deliberation.At this stage, we have two options. We either wait and see, right, if things get better by themselves. Because you said: overall, I feel quite good, but maybe I have just been doing too much. On the other hand, because there are so many inflamed joints at the moment, it may be a good thing to see if we try to get you moving again with an injection of prednisone, for example. (rheumatologist, male, 39 years old)
→Item score: 2
Item 2: justify deliberative work as a team
The clinician reassures the patient or re-affirms that the clinician will support the patient to become informed or deliberate about the options. If the patient states that they have sought or obtained information prior to the encounter, the clinician supports such a deliberation process.Ultimately, I do think that it is important for you to decide what happens. Also, that you understand what we are suggesting so that you can also decide what is best. (rheumatologist, male, 45 years old)
→Item score: 3
Item 3: inform, describe options, and exchange views
The clinician gives information or checks understanding about the options that are considered reasonable (this can include taking no action), to support the patient in comparing alternatives. If the patient requests clarification, the clinician supports the process.The only other thing we could also try… [ … ] is to start you on a biologic agent after all, to see if we can manage your fatigue problem a little better. However, I do not think that this will provide you with much more benefit. (rheumatologist, female, 52 years old)
→Item score: 1
Item 4: elicit preferences
The clinician makes an effort to elicit the patient's preferences in response to the options that have been described. If the patient declares their preference(s), the clinician is supportive.What is worrying you so much about methotrexate at the moment? (rheumatologist, male, 40 years old)
→Item score: 3
Item 5: integrate preference
The clinician makes an effort to integrate the patient’s elicited preferences as decisions are made. If the patient indicates how best to integrate their preferences as decisions are made, the clinician makes an effort to do so.I will reduce the dose to once every two weeks . I have no problem with that. [ … ] Do you know what I mean? Apart from the fact that you do not look forward to it. Are you sure you understand that this is what I suggest? (rheumatologist, male, 59 years old)
→Item score: 0
The number of consultations per clinician ranged from 1 to 18. The clinician with the lowest OPTION scores (seven consultations) had a median OPTION score of 5 (IQR=0–5). The clinician with the highest OPTION scores (six consultations) had a median OPTION score of 42.5 (IQR=30–55). Figure 1 shows the OPTION scores per clinician.
OPTION scores per clinicianThe numbers on the x-axis refer to the participating clinicians with in parentheses the number of routine consultations with unique patients per clinician. Boxes represent values between the 25th and 75th percentiles, whiskers the upper and lower adjacent values and horizontal lines the medians. Outliers are shown as dots. Clinician 1 (n=7) had five OPTION scores of 20, one OPTION score of 10 and one OPTION score of 60.
Associations between the level of SDM and characteristics of the clinician, patient and consultation
From the univariate regression analyses, four factors met the predefined selection criterion (p<0.2) and were included in the multilevel model: clinician age, patient age, consultation duration and type of treatment decision. There were significant, positive associations between the level of SDM and the consultation duration (b=0.63, 95% CI 0.16 to 1.11, p=0.01), decision for starting and/or stopping medication (b=14.3, 95% CI 5.62 to 22.98, p≤0.01), decision for adjusting medication doses (b=8.36, 95% CI 3.92 to 12.81, p≤0.01) and decision for administering single dose glucocorticoids (b=15.03, 95% CI 9.12 to 20.93, p≤0.01). Thus, a higher level of SDM was significantly associated with a longer consultation duration and the type of treatment decision. More specifically, if the consultation duration increased with 10 min, this led to an increase of 6 points on the OPTION scale. Also, compared with the decision to continue medication, the decisions to stop and/or start medication, adjust medication doses and administer single dose glucocorticoids led to an increase of respectively 14, 8 and 15 points on the OPTION scale. No other significant associations were found (table 2). The intraclass correlation indicated that 22% of the total variability in the level of SDM resided between clinicians.
Multilevel model (patients, level-1, nested in clinicians, level-2) of associations between the level of SDM and characteristics of the clinician, patient and consultation
Discussion
This is the first study that used an observer-based measure to investigate the extent to which SDM is applied in routine clinical care of patients with RA. Using the OPTION scale, we found a low to moderate level of SDM in RA treatment. The efforts of clinicians to involve patients in decision making varied substantially. Furthermore, our results showed that a higher level of SDM was significantly associated with a longer consultation duration and the type of treatment decision.
This study supports the results of qualitative studies and surveys, suggesting that SDM is not current practice in RA treatment.15–18 However, the level of SDM found in this study is higher than the level of SDM reported in most studies. A systematic review of 29 studies using the OPTION scale has shown that only 38% of studies reported a mean OPTION score ≥25.24 The higher mean OPTION score reported in this study (28.3) may be explained by the clinical setting. It has been argued that SDM is particularly suitable for long-term treatment decisions in chronic care, relative to treatment decisions in acute care that are often urgent and irreversible.36 Moreover, medical specialists and non-physicians have shown to involve patients in decision making more frequently than general practitioners.24 In addition, SDM has been making headway in healthcare policy over the last decade. This may be reflected by equally increased levels of SDM found in current research.26 Nevertheless, our results show that there is room for improvement. Indeed, a low to moderate level of SDM in RA treatment does not align with the fact that the vast majority of patients with RA prefer to be involved in decision making.17 37
In this study, the third item of the OPTION scale (inform, describe options and exchange views) scored lowest. This is in contrast to most studies in which the second item of the OPTION scale (justify deliberative work as a team) scored lowest.24 38 Previous studies have shown that patients with RA express a strong need for full disclosure of all available treatment options and their associated risks and benefits.39–41 Moreover, it has been shown that clinicians often underestimate or undervalue patients’ information needs.42 It is important that patients receive the information necessary to enable SDM. This may be achieved by the use of decision aids that help patients become well informed and prepared for decision making. Furthermore, our results showed that there was a substantial variety between the efforts of clinicians to involve patients in decision making. Training programmes on patient-centred communication skills have proven to be effective to help clinicians share treatment decisions with patients.43 However, clinicians may perceive barriers that keep them from applying SDM.44 Future research should focus on identifying barriers to the application of SDM in RA treatment as perceived by clinicians.
We found that a higher level of SDM was significantly associated with a longer consultation duration. Although the application of SDM may add time to the consultation, it is plausible that by applying SDM clinicians lay the groundwork for quicker follow-up care.45 Furthermore, it should be noted that, in this study, adding 10 min to the consultation only led to a modest increase of the level of SDM (an increase of 6 points on the OPTION scale). Our results also showed that the level of SDM was significantly higher in consultations in which it was decided to make changes to a patient’s treatment (stopping and/or starting medication, adjusting medication doses and administering single-dose glucocorticoids). This implies that SDM may be more suitable for some treatment decisions than others. Other researchers have also stated that the level of SDM will probably always differ, depending on the situation.26 Moreover, a higher level of SDM is not necessarily better, for example, if a patient does not prefer to be involved in decision making. Therefore, it is important for clinicians to always be sensitive to the suitability of SDM in a particular situation.
There are several strengths and limitations of this study that need consideration. A strength was the use of an observer-based measure to assess the level of SDM in RA treatment. We had access to a large number of audio-recorded consultations (n=168), strengthening the ecological validity of our results. Audio-recording the consultations may have affected the behaviour of both clinicians and patients. However, at the time of the consultation neither physicians nor patients were aware that the audio-recordings would be used to investigate the level of SDM. As we used audio-recordings instead of video material, non-verbal behaviour could not be taken into account.38 Furthermore, only a single consultation per patient was audio-recorded. Decision making may take place over multiple consultations. Therefore, it is possible that we have not always captured the full process. We were also limited by the availability of the data. Other factors, such as clinician–patient familiarity and patient health literacy, may also impact the level of SDM in RA treatment, but were not included in this study. Finally, the cross-sectional design of this study did not allow for conclusions about causality.
In summary, we found that a higher level of SDM was significantly associated with factors related to characteristics of the consultation, namely a longer duration and the type of treatment decision. No significant associations were found between the level of SDM and factors related to characteristics of the clinician and patient. Overall, the level of SDM in RA treatment leaves room for improvement. Targeted strategies to support the application of SDM are warranted in this era of patient-centred care. For example, training programmes on patient-centred communication skills may be helpful to foster SDM in routine clinical care. Nevertheless, clinicians should always be sensitive to the suitability of SDM in a particular situation.
Acknowledgments
The authors would like to thank Floris Kok for his help with scoring the audio-recorded consultations and Renske Hebing for coordinating this study in Reade.
References
Footnotes
Handling editor Josef S Smolen
Contributors EGEM: study conception and design, collection of data (scoring the audio-recorded consultations), analysis and interpretation of data, drafting and critical revision of the manuscript, and approval of the final version of the manuscript. JEV: study conception and design, analysis and interpretation of data, critical revision of the manuscript, and approval of the final version of the manuscript. CDP: analysis and interpretation of data, critical revision of the manuscript, and approval of the final version of the manuscript. BJFvdB: study conception and design, acquisition of data, analysis and interpretation of data, and critical revision of the manuscript. FK: collection of data (scoring the audio-recorded consultations).
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval The medical research ethical committee of Arnhem-Nijmegen, the Netherlands, waived ethical approval since this study was not subject to the medical research involving human subjects act (file number 2018–4180).
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request.