Patient-reported outcomes in RA care improve patient communication, decision-making, satisfaction and confidence: qualitative results

Abstract

Objective

To evaluate the impact of integrating patient-reported outcomes (PROs) into routine clinics, from the perspective of patients with RA, clinicians and other staff.

Methods

We conducted a prospective cohort study using a mixed methods sequential explanatory design at an academic arthritis clinic. RA patients completed selected Patient-Reported Outcomes Measurement Information System measures on tablets in the waiting room. Results were immediately available to discuss during the visit. Post-visit surveys with patients and physicians evaluated topics discussed and their impact on decision making; patients rated confidence in treatment. Focus groups or interviews with patients, treating rheumatologists and clinic staff were conducted to understand perspectives and experiences.

Results

Some 196 patients and 20 rheumatologists completed post-visit surveys at 816 and 806 visits, respectively. Focus groups were conducted with 24 patients, 10 rheumatologists and 4 research/clinic staff. PROs influenced medical decision-making and RA treatment changes (38 and 18% of visits, respectively). Patients reported very high satisfaction and treatment confidence. Impact on clinical workflow was minimal after a period of initial adjustment. PROs were valued by patients and physicians, and provided new insight into how patients felt and functioned over time. Reviewing results together improved communication, and facilitated patient-centred care, shared decision making, and the identification of new symptoms and contributing psychosocial/behavioural factors.

Conclusion

PRO use at RA visits was feasible, increased understanding of how disease affects how patients feel and function, facilitated shared decision-making, and was associated with high patient satisfaction and treatment confidence.

Rheumatology key messages

• Reviewing patient-reported outcomes at RA visits is feasible and well-received by patients and clinicians.

• Discussing results facilitates a shared understanding of the patient’s RA symptom experience and everyday life.

• Use of patient-reported outcomes can improve RA decision-making, communication, satisfaction and treatment confidence.

Introduction

Patient-centred care begins with understanding the experiences of people living with their disease(s) [1, 2]. Patient-reported outcomes (PROs) quantify health reported directly by the patient, without interpretation or input from others. Some view PROs as having a central role in transforming healthcare by helping clinicians and patients make better decisions, and allowing better evaluation of health care services [3, 4]. Concerns about trade-offs between the value of PROs to inform health decisions vs the burden of collection have limited their adoption in routine care [5–10].

RA greatly affects physical, emotional and social health. Physicians monitor RA patients primarily by examining joints, reviewing labs, and asking patients about their pain, physical function and disease control [11–13]. RA patients have also identified fatigue, sleep disturbance, emotional distress and social participation as important outcomes to them [14–16].

The Patient-Reported Outcome Measurement Information System™ (PROMIS™) measures physical, mental and social health across diseases. Scores are standardized as T-scores (mean of 50, s.d. 10; higher value represents more of what is measured) to facilitate interpretation of individual results relative to the US general population [17]. We have previously shown PROMIS can reliably and precisely measure RA symptoms that are most important to patients [18].

The study was designed to evaluate the feasibility and acceptability of collecting and reviewing a comprehensive set of PROs as part of routine rheumatology visits, and the impact on care. We hypothesized that PROs would be feasible to collect, focus conversations on a broader range of symptoms that were important to patients, facilitate shared decision-making and be associated with high levels of patient satisfaction and treatment confidence.

Methods

Design

Using a mixed methods sequential explanatory design [19], first quantitative data were collected and analysed (reported elsewhere [18]), and questions generated by the quantitative analysis were further explored through the collection and analysis of the qualitative data contained in this report.

Participants

Patients who were English-speaking, enrolled in our RA registries and receiving care at the Johns Hopkins Arthritis Centre were consecutively invited to participate between September 2012 and August 2015 [18]. We excluded individuals who were not fluent in English and those with current medical or psychiatric comorbidities that we believed would limit their ability to participate. Providers included all attending rheumatologists (n = 8) and clinical fellows (n = 12; 4 per year) at the clinic. Research staff included five coordinators, two assistants and two post-doctoral fellows. Clinic staff included three medical assistants and two office coordinators. The study was approved by the Johns Hopkins Institutional Review Board (NA0071923, NA00066663, IRB00059765). All participants provided written informed consent.

Process

Patients

Study visits occurred in the context of usual care visits. Participants were asked to arrive 30 min early to register for the visit and complete questionnaires on a tablet. At every visit, patients completed 10 PROMIS computer adaptive tests (CATs) (Pain Interference, Physical Function, Fatigue, Participation in Social Roles and Activities, Satisfaction with Social Roles, Sleep Disturbance, Sleep Interference, Anxiety, Depression, Anger). RA patients have identified these domains as important to their health-related quality of life [15].

Clinicians

At the start of the study, we reviewed how to interpret PROMIS results with the rheumatologists; new fellows were subsequently trained by attending rheumatologists. Colour-coded reports (Fig. 1) were printed and given to physicians when entering the examination room. Prior results were available in the chart. Physicians were encouraged (but not required) to review results with patients. We conducted surveys and interviews at the midpoint and end of the study to assess the impact of integrating PRO reviews at visits.

Fig. 1.

Sample PROMIS report

Reprinted with permission from www.healthmeasures.net/PROMIS. PROMIS: Patient-Reported Outcome Measurement Information System.

Questionnaires

Comfort using tablets

At baseline, patients rated their familiarity and comfort using tablets/computers, whether tablets were easy to use and their preference for answering questions (tablet vs paper).

Use of PRO results during visit

Immediately after the visit, patients and physicians each independently completed post-visit surveys. We asked patients to indicate which symptoms were discussed (yes/no) during the visit; about the time needed to review results (about right/more time needed); and if their treatment had been changed during the visit. Physicians were asked if they received a PROMIS report, reviewed it and discussed with patient (all yes/no); about the time spent discussing PROs (about right/too long); whether new symptoms or comorbidities were identified; and whether the discussions influenced a decision to change the patient’s treatment.

Confidence in Decision scale

Patients also completed the COMRADE (combined outcome measure for risk communication and treatment decision making effectiveness) Confidence in Decision scale after the visit [20]. The 10 items reflect core domains of patient-centred care and shared decision-making. COMRADE has been validated in RA and is strongly related to physician trust [21].

Patient interviews

Midway through the study, experienced qualitative researchers (E.D.L., K.C.S.) conducted interviews with nine patients (∼10% of the sample at that time) within 48 h of a visit. Patients were purposively sampled to have maximal variation in potential moderating characteristics including computer literacy and experience, treatment expectations, RA duration and disability, and sociodemographic factors (age, sex, race, education). We asked about: ease of using tablets; quantity, content and relevance of PRO questions; use and impact of having PRO results during visit; and suggestions for continued use in clinical practice. Patients were also asked to if they felt the PROs were relevant to them, impacted discussions and shared decision making with the doctor, and their overall impressions of using PROs as part of RA care. Patient interviews were recorded and transcribed. Analysis loosely followed a simplified framework approach [22], with responses being applied to a data matrix addressing predefined topics to facilitate comparison across participants.

Clinicians and staff

Midway through the study, we also interviewed three attending rheumatologists. Detailed notes were taken during the interviews. At 9–12 months, separate focus groups were held with three clinic staff members and six rheumatology fellows; interviews were held with two research staff members. Discussions were organized around experiences and perceptions of PRO collection, challenges in implementation and perceived value. Rheumatologists were also asked about the value of discussing PRO results during clinical encounters, their perceptions of patients’ reactions to reviewing PROs, implementation challenges, and impact on communication and shared decision-making. Interviews and focus groups with fellows and staff were audiotaped and transcribed.

After the conclusion of the study, we interviewed the nine patients queried midway through the study and an additional six participants (n = 15; eight via phone, seven in person), four research staff who interacted with patients and three rheumatologists. The interviews, which probed questions that arose from analysis of survey data, were recorded and summarized according to themes identified in the initial interviews.

Data analysis

Descriptive statistics were calculated to summarize participant characteristics and survey results. Linear mixed models with random effects were used to assess relationships between topics discussed, time discussing and patient confidence in treatment decisions.

Thematic analysis of focus groups and interviews was aided by qualitative analysis software (ATLAS.ti). Our approach was largely exploratory and inductive, but guided by a priori areas of interest that were being examined in the survey component of the work (i.e. feasibility and utility of PROs in the clinical setting and cohesion between PRO results and participant experiences.) Thematic analysis focussed on feasibility, adequacy of symptoms queried, and the impact of PRO reviews on patient–provider communication, visit satisfaction and shared decision making. Within each category, we used interview data to expand and refine categories and examine relationships between categories to identify common and contrasting experiences and opinions from the perspectives of patients, physicians, clinic and research staff. Our patient research partner (A.L.) provided additional input on data analysis and interpretation.

Results

Participants

The 196 adults were mostly women (81%) with a mean (s.d.) age of 55 (13) years, with 12 (10) years disease duration (Table 1). Most (68%) had RA that was in remission or low disease activity. Eight rheumatologists and 12 clinical fellows provided care for patients; interviews/focus groups were conducted with 4 rheumatologists and 6 fellows. Of the five research coordinators, two research assistants and two research fellows that assisted with this study, we interviewed the four with the most patient contact. Among clinic staff, of the three medical assistants and two office coordinators, three were selected at random and interviewed.

Table 1.

Patient characteristics (n = 196).

	Value
Age (years)	54.8 (13.4)
Women	158 (81)
RA duration (years)	11 (10)
RF or anti-CCP antibody	152 (78)
Race
White	162 (83)
Black	25 (13)
Other	9 (4)
CDAI	6.4 (1.8, 11.8)
Remission	62 (32)
Low disease activity	70 (36)
Moderate to high disease activity	64 (32)
Tender joints (0–28)	0 (0, 2)
Swollen joints (0–28)	1 (0, 4)
Physician global assessment (0–100)	10 (3, 20)
Patient global assessment (0–100)	20 (5, 50)
ESR (mm/h; n = 180)	14 (5, 29)
CRP (mg/dL; n = 184)	0.3 (0.1, 0.7)

Data are presented as mean (s.d.), median (interquartile range) or n (%). CDAI: Clinical Disease Activity Index.

Comfort completing questionnaires on tablets

Early in the study, several patients suggested we add foam tubing to the tablet stylus to facilitate use. Most (86%) patients said they were ‘somewhat’ (24%) or ‘very comfortable’ (62%) using computers. Most (83%) used computers at work or home and were comfortable using tablets; only 7% were ‘somewhat’ or ‘very uncomfortable’ using computers. Almost all (93%) ‘agreed’ or ‘strongly agreed’ that it was easy to answers questions on tablets; 89% preferred tablets to paper questionnaires. Participants reported that answering questions on the tablet become easier over time; research staff confirmed that patients became more efficient and required less assistance over time.

Process of completing PROs

Patients completed PROs at 821 visits (mean 4 visits per patient; range 1–8). On average, 5 items were administered for each PROMIS CAT, with 10 CATs completed in 11 min. Post-visit surveys were available for 806 patients and 818 physicians.

Patients

Interviews were conducted midway and at the end of the study. At the midpoint, no new codes in the categories had occurred by the ninth participant; six additional participants were added to end of study interviews (total 15) to support saturation and enhance the diversity of sample. The emergent themes from the patient interviews provided context and specificity for the survey responses. Almost all said that the number of questions and time spent answering them was appropriate.

‘It’s not taxing. It’s just a few extra minutes, and it can be monumental in its benefits’ (woman, age 43 years).

Physicians

Rheumatologists noted that when patients arrived late, they did not always have time to complete PROs.

Staff

Staff noted that age and tablet experience initially affected completion time. When clinic flow disruptions occurred, it was almost always due to patients arriving late. While some patients initially asked for help using tablets, very few requested help subsequently.

Impact of PROs on clinical interactions

Patients

In post-visit surveys, 92% said PROs prompted them to think more broadly about how they felt and how RA affected day-to-day life. Almost all (97%) felt that ‘adequate time’ was spent discussing results. Frequently discussed topics included impacts on life, pain and overall health (supplementary Table S1, available at Rheumatology online).

In interviews, 80% of patients felt that PROs improved communication with their physician.

‘It focuses your discussion … so you can prioritize the issues you want to address’ (woman, age 50 years).

‘It helps communication with the doctor … He sees the results, and we go through them, and almost every time it brings up some issue or question or clarification about how I’m feeling’ (man, age 64 years)

Many valued the ability to track their symptoms over time, often framed as their ‘progress’ or their ‘new normal’.

‘Sometimes [the doctor] won’t see you for a couple of months … and then they look at [PROs] and say … “Last time you were here, you answered that this was poor and now this is good.” So it helps us see the progress’ (woman, age 25 years)

‘I love seeing data and graphs. It really helps me’ (man, age 64 years)

PROs helped patients discuss what was important to them regarding how well their disease was controlled and facilitated shared decision-making and discussion about the need for additional tests or monitoring.

‘It was extraordinarily helpful when you sit down with your doctor and he says, “This is what I see …” and I say, “Well this is what I see …”. And we talk about how we can improve those areas that I would like to see improved, or maybe manage better those areas that may not be able to be improved’ (woman, age 53 years)

‘It’s probably as important as any of my medications. It opened up a whole new way to talk about what was going on, how I was feeling, and made me remember things that I wanted to discuss’ (woman, age 60 years)

Physicians

Physicians reviewed (≥83%) and discussed (≥76%) PRO results during most visits (average three symptoms discussed; supplementary Table S2, available at Rheumatology online). Across visits, 82% rated the time spent discussing the reports ‘about right’; discussing results took ‘too much time’ at <2% of visits. Impact on life, pain and overall health were the topics that were discussed most frequently.

Physicians tailored discussions to patient needs and visit context. Reviewing PRO results together helped initiate and focus discussions and identify new or worsening symptoms.

‘It helped me understand and pay more attention to some things that aren’t usually brought up’

‘These data help us treat the whole patient by addressing issues that patients have identified as important, improve satisfaction with their overall care, and help patients take charge of their own health care’

Reviewing PRO results provided insight into how well RA symptoms were currently controlled.

‘Some patients are very stoic, but seeing the results helped me show them what they are really going through’

Some physicians noted they were less comfortable discussing emotional results (i.e. sadness, anger and anxiety) with patients. One provider described how a high depression score led to a discussion about tragic events in the patient’s life, and how that was negatively affecting self-management. In another case, an elevated depression score helped both the patient and physician to identify the recurrence of a major depressive episode.

Staff

Staff indicated that patients mentioned to them that having PRO results available improved communication and decision-making with their doctor.

‘Those with RA for a long time mentioned it helped them consider the impact of their disease … and brought up a lot of things they had not thought would be related to RA but actually were’

Staff recalled rare examples where patients disagreed with results (e.g. depression, anger scores). A few expressed concern that their scores may cause their doctor to view them less favourably.

Added value of PRO information

Patients

Overall, patients judged the PROs as highly relevant to their care and helping them see how their disease affected everyday life.

‘They’re extremely relevant … Pain and fatigue, those have always been general markers, but the ones … about being able to participate in life the way I want to, and the sleep questions have improved my treatment’ (woman, age 58 years)

‘It made me think more about what actually was happening. I knew there was sleep disruption; it made me analyze it more … helped me identify problems and think more deeply about them’ (man, age 72 years)

Many patients mentioned that when rating symptoms such as fatigue, it was not clear if they should rate their RA-related fatigue, or general fatigue.

‘It gets a little difficult … is it all related to the RA, or are some problems related to something else?’ (man, age 53 years)

Some noted that monitoring PROs between visits could offer more time for reflection and more accurate assessment of their health.

‘I think that [competing PROs between visits] would help spot trends that might come up or missed symptoms that emerge’ (woman, age 50 years)

Many patients found it was helpful to compare their results against population norms. One noted that he had underestimated the severity of his pain and fatigue.

‘You have this disease long enough and all of the sudden your normal changes … I have felt like this for so long now I almost think it’s normal. But if I look at the graphs, I can see that it isn’t’ (man, age 61 years)

Physicians

In post-visit surveys, physicians indicated that reviewing PRO results influenced decisions to change or adjust RA treatment in 20% of encounters (Table 2). Often (38% of visits), results helped identify new symptoms. The most frequent treatment changes included adding new medications, non-pharmacological treatments (e.g. physical/occupational therapy) and recommending health behaviour changes.

Table 2.

Influence of patient reported outcomes on treatment decisions

	Visit
	1	2	3	4	5–8
	(n = 196)	(n = 181)	(n = 158)	(n = 137)	(n = 146)
Change in arthritis treatment and new symptom identification, n (%)
Change or adjust arthritis treatment	34 (18)	28 (16)	27 (18)	24 (19)	20 (14)
Identify new symptom or comorbidity	73 (38)	62 (35)	49 (33)	33 (27)	38 (26)
Specific changes prompted by discussing PRO resultsa, n (%)
Decide to NOT change treatment	116 (59)	121 (67)	99 (63)	89 (65)	91 (62)
Increase medication dose/frequency	8 (4)	10 (6)	13 (8)	6 (4)	6 (4)
Decrease medication dose/frequency	8 (4)	8 (4)	7 (4)	2 (1)	2 (1)
Add new RA medication(s)	21 (11)	14 (8)	11 (7)	13 (9)	9 (6)
Add new non-RA medication(s)	7 (4)	7 (4)	2 (1)	2 (1)	4 (3)
Add nonpharmacologic intervention	21 (11)	10 (6)	6 (4)	6 (4)	7 (5)
Stop ≥1 RA medicine	0 (0)	0 (0)	2 (3)	1 (1)	1 (1)
Stop ≥1 non-RA medicine	1 (5)	0 (0)	0 (0)	0 (0)	0 (0)
Evaluate potential comorbidity	10 (5)	4 (2)	1 (1)	5 (4)	2 (1)
Refer to another specialist	7 (4)	2 (1)	1 (1)	1 (1)	5 (3)

^aAs reported by 19 attending rheumatologists or rheumatology fellows; could select more than one response. PRO: patient-reported outcome.

‘PROMIS results helped us make better clinical decisions on everything together. I didn’t change [a decision] based on just the scores, but [they] helped me ask questions that may not have been asked otherwise’

Physicians found it helpful to interpret results in relation to population norms.

‘I think my favourite thing was the normalization. There’s so much fatigue and anxiety. Many people were worried … but scoring right at the 50th percentile. I think [we] got some reassurance from that sometimes’

Several physicians commented on how results facilitated monitoring symptom changes over time.

‘I like the PROMIS report because you can show the patient, “Look, this is how you were 3 months ago; this is how you are now. You are better. That’s great.” Or, “You're worse. What happened?”’

PRO results helped uncover psychosocial factors, lifestyle and behaviours including treatment adherence.

‘It’s hard to tease out, but that’s when I start asking “Well, your pain was good here … What's been happening?” And then they tell you, “Oh, well I felt so great that I decreased my dose!”’

‘It has changed my approach to patients. When a patient says, ’I don’t think I’m doing well,’ but I look at their swollen and tender joints or their labs and … think they are doing well, I am much more likely now to ask about fatigue, mood, pain, sleep. I try to understand that disconnect and how can we fix it’

Collective impacts associated with PRO use are summarized in Fig. 2.

Fig. 2.

Potential impact on patients and providers of integrating PRO data collection and use into chronic disease visits

PRO: patient-reported outcome.

Patient confidence in treatment decisions

Mean COMRADE Confidence in Decision scores were very high, averaging 48/50 across visits, suggesting that patients were highly satisfied with how decisions were made (Fig. 3, supplementary Table S3, available at Rheumatology online). Patients strongly agreed that they were aware of treatment choices and felt able to make informed choices that reflected what was best for them.

Fig. 3.

Mean COMRADE Confidence in Decision scores at each visit

In linear mixed models, the average COMRADE score increased by 0.2 for each topic discussed (P = 0.001). However, among the minority of patients who felt ‘more time was needed’ to discuss PROs, median scores were about 6 points lower (P = 0.0007) compared with those who rated the time discussing PRO results ‘about right’.

Discussion

This is the first study to evaluate the impact of collecting and using PROs as part of routine RA clinic visits from multiple perspectives. When questionnaires were completed in the waiting room, the results were immediately available and almost always discussed during the visit. Both patients and physicians felt that reviewing PRO results offered insight into how patients were feeling and their day-to-day function, and facilitated shared decision making. In some instances, new symptoms were uncovered suggesting that a patient’s RA was not well controlled. Patients said that reviewing PROs made visits feel more patient-centred, which translated into very high levels of satisfaction and treatment confidence. In a study of RA patients on DMARD decision making [21], mean COMRADE treatment confidence scores averaged 4.2, whereas our scores averaged 4.8 across visits. High satisfaction and confidence in treatment reflect strong communication and trust in the physician [20]—factors that are highly correlated with better treatment adherence [23].

While PROs are increasingly collected as part of quality initiatives and value-based care, there are limited data on how they influence care. For example, although RA affects multiple aspects of health-related quality of life [14–16, 18], discussions initiated by rheumatologists are often limited to asking patients about their pain, disability and how active their RA is [14–16]. Patients said that PROs prompted them to identify new or worsening symptoms and to consider ongoing physical and social limitations. Physicians noted that in most cases, PROs affirmed clinical findings and increased confidence about whether the current treatment was working as intended. Notably, in nearly two of five patients, new symptoms were uncovered that led to referrals or new interventions; in one of five patients, adjustments were made to RA therapies.

Collecting PROs immediately prior to visits was feasible because we asked patients to arrive earlier for their appointment. Most patients preferred using tablets over paper questionnaires, even when they had limited experience with computers. While physicians said discussing results initially increased the visit length, this was an acceptable trade-off to improve communication and shared decision making. All said that reviewing changes in PROs helped quantify improvement and worsening in symptoms in relation to treatment changes or life events.

Evidence supporting an added value of integrating PRO reviews into care visits is important and timely given that PROMIS and other measures are available in many electronic medical record systems. Electronic medical records can be programmed to contact and invite patients to complete PROs online prior to the visit. For example, RAND analysts recently demonstrated that collecting the PROMIS-29 profile via the web, phone or mail was feasible and valid in older adults with multiple chronic conditions enrolled in Kaiser Permanente Colorado [24]. PROMIS measures have robust psychometric properties across chronic diseases [25, 26], and CATs reduce respondent burden by providing precise estimates with fewer questions [27]. The T-score metric facilitates interpretation, including in reference to the US general population. Both physicians and patients easily understood PROMIS reports, although many asked about what was ‘normal’ for people with RA. RA norms, minimally clinically important differences and cut points linked to clinical actions are needed to enhance interpretability.

Strengths of our study include the use of mixed methods to evaluate the experiences of patients, clinicians and staff. The sample included diverse patients with varying levels of disease severity, comorbidities, and personal and RA experiences. There are also limitations. We conducted this study at an academic centre with an integrated clinical and research operation. We did not assess patient satisfaction and confidence in treatment decisions prior to the intervention and do not know how this was specifically impacted by the intervention. However, a recent prospective study in orthopaedic visits suggested that adding PROs does improve the patient experience [28], and we note that patients in our study reported higher general health-related quality of life than in other RA studies [21]. The extent to which PROs influenced discussions and decision making is unclear; given the skills and experience of clinicians and patient needs, similar discussions may have occurred in the context of usual care. It remains unclear whether changing the process and experience of care improves long-term RA outcomes.

Conclusion

We found that reviewing PROs at routine RA visits was feasible and well received, and encouraged a shared understanding of the patient’s symptom experience and everyday life. Patients felt PROs made the visit more patient-centred and improved communication and shared decision-making, resulting in high satisfaction and treatment confidence. PROs helped clinicians to identify new symptoms and adjust treatment as needed. As PROs are increasingly collected to fulfil population monitoring, quality and policy needs, our results suggest that when results are reviewed by patients and providers together, they can also improve communication and guide patient care.