Patient-reported outcomes (PROs) and their measures have a long and important history for determining the status and treatment of patients with rheumatoid arthritis (RA). This article describes the history and evolution of PROs for RA and the current state of the field, with key examples of accepted and widely used measures, and offers some reflection on the roles of PROs for the study and management of RA.
Key points
- •
Patient-reported outcomes (PROs) and their measures have a long and important history for determining the status and treatment of patients with rheumatoid arthritis (RA).
- •
The most important and commonly studied RA PROs are also core measures: physical function, pain, and patient global assessment.
- •
This article describes the history and evolution of PROs for RA and the current state of the field, with key examples of accepted and widely used measures, and offers some reflection on the roles of PROs for the study and management of RA.
What are patient-reported outcomes and why are they important?
Characteristics of rheumatoid arthritis (RA) can be divided into those signs and symptoms that are assessed by the patient, and those that are assessed by someone or something other than the patient. In the latter case, the assessments are considered objective measures of disease activity or damage, like acute phase reactants, the swelling of joints, or the erosions on a radiograph of a hand. In cases of signs and symptoms reported directly by the patient, without interpretation of a third person, clinicians speak of patient-reported outcomes (PROs).
The degree of disease activity and response to treatment are traditionally determined by the evaluation of the RA core set or indices derived thereof. The core set contains 3 PROs: physical function, pain, and a global assessment of disease activity. These PROs have been found to be at least as important as other physical and biochemical (more objective) measures in assessing baseline disease status, improvement during interventions, or prediction of long-term outcome. However, several important areas, such as fatigue and sleep quality, have only recently been identified as important to patients and thus potentially core areas for measurement.
Patients with RA no longer depend on their physicians to tell them what to do, but increasingly take charge of their care processes and functions as partners in obtaining relevant information. Patients and professionals bring different skills, values, and experiences to research.
This article gives an overview on the growth and current value of PRO research in RA, the application of research findings into daily clinical practice, and the gap between PRO research and practice that needs to be filled in the coming years.
What are patient-reported outcomes and why are they important?
Characteristics of rheumatoid arthritis (RA) can be divided into those signs and symptoms that are assessed by the patient, and those that are assessed by someone or something other than the patient. In the latter case, the assessments are considered objective measures of disease activity or damage, like acute phase reactants, the swelling of joints, or the erosions on a radiograph of a hand. In cases of signs and symptoms reported directly by the patient, without interpretation of a third person, clinicians speak of patient-reported outcomes (PROs).
The degree of disease activity and response to treatment are traditionally determined by the evaluation of the RA core set or indices derived thereof. The core set contains 3 PROs: physical function, pain, and a global assessment of disease activity. These PROs have been found to be at least as important as other physical and biochemical (more objective) measures in assessing baseline disease status, improvement during interventions, or prediction of long-term outcome. However, several important areas, such as fatigue and sleep quality, have only recently been identified as important to patients and thus potentially core areas for measurement.
Patients with RA no longer depend on their physicians to tell them what to do, but increasingly take charge of their care processes and functions as partners in obtaining relevant information. Patients and professionals bring different skills, values, and experiences to research.
This article gives an overview on the growth and current value of PRO research in RA, the application of research findings into daily clinical practice, and the gap between PRO research and practice that needs to be filled in the coming years.
History of patient-reported outcomes in rheumatoid arthritis
RA has a rich history of PRO research and still is at the forefront of PRO measure (PROM) development and patient participation in research activities. One of the most characteristic symptoms of RA, pain, was first described as an important outcome at the beginning of the nineteenth century, when therapy for RA was mostly nonpharmacologic, with the exception of salicylates (what are now known as nonsteroidal antiinflammatory drugs) such as ibuprofen and diclofenac, but with a less favorable safety profile and of limited efficacy.
With the introduction of gold compounds in the twentieth century and the discovery of the so-called disease-modifying antirheumatic drugs (DMARDs) sulfasalazine and hydroxychloroquine, as well as glucocorticoids, the development of outcomes in RA research became more and more important. One of the first initiatives in outcome research was that of Steinbrocker and Blazer, who developed the therapeutic score card for RA. This method included the patient global assessment of disease activity, as well as joint tenderness, pain, and functional status. In 1956, Lansbury developed the Systemic Index, the first numerical method to assess and compare disease activity between patients. This index included a measure of duration of morning stiffness, a measure of fatigue (hours after rising before onset of fatigue), grip strength, and pain, which was measured as the number of aspirins required for pain relief.
Development of the Health Assessment Questionnaire (HAQ) in 1980, followed by development of a shorter version, the HAQ Disability Index (HAQ-DI), created a revolution in measurement of functional status in RA that is still in use. Years later, Paulus and colleagues developed the Paulus Criteria, including morning stiffness, joint pain, and the patient global assessment. To harmonize the use of outcome measures across RA clinical trials, the American College of Rheumatology (ACR) established a committee to develop the first RA core set of outcome measures. With support of the first Outcome Measures in Rheumatology (OMERACT) meeting in 1992, consensus was reached, resulting in the ACR core set that includes 3 PROs: pain, patient global assessment of disease activity, and functional status.
Although already included in the systemic index of Lansbury, it was not until recently that fatigue was identified as one of the most important problems identified by patients with RA, and it has proved highly reliable, sensitive to change, and an independent determinant of disease activity. Although fatigue is not part of the ACR core set, OMERACT endorsed fatigue in their core set in 2006, and the European League Against Rheumatism (EULAR) and ACR recommended reporting fatigue in the domain of disease activity in every RA randomized controlled trial (RCT). Although most clinicians agree that morning stiffness is typical for RA, duration of morning stiffness was excluded from the recent update of both the ACR classification and remission criteria. Although the importance of the symptom was acknowledged, it was thought that the instruments to measure it yielded data of insufficient reliability to include stiffness in the criteria. A recent review of patient-oriented measurement instruments that have been developed to assess stiffness in patients with RA identified studies between 1987 and 2010 with instruments limited to duration and severity of morning stiffness found similar insufficient data.
At present, international consensus exists on the importance of measurement and reporting of patients’ assessment of disease activity, pain, and physical functioning.
Other promising PRO domains that are currently being studied include flare, remission, stiffness, (work) participation, worker productivity, self-management, sleep, and emotional distress.
Use of patient-reported outcomes in different current settings
The scientific interest in PROs in recent years has been boosted by societal pressure; patients demand to be an active contributor to their care process, whereas, in contrast, health insurance companies and governments call for transparency of outcomes in order to quantify quality of care delivery because costs are increasing to unsustainable levels. Under the influence of health economist Michael Porter, health care delivery is more and more patient centered, with value-adding activities as a main driver for good outcomes.
As a consequence, physicians are faced with the request to add value for patients, but often lack information or tools to measure domains that patients find important. Hospitals struggle to identify a single PRO that is important and meaningful for all the patients who are treated within their facilities. Although PRO research activities have been focused toward identification of patient-important domains and optimizing psychometric properties of domain-specific questionnaires, there is a growing interest in the translation of all this evidence into clinical practice in order to start measuring, comparing, and improving outcomes and making good choices about reducing costs. However, it is unclear whether PROs and the instruments to measure them allow for benchmarking across health care providers.
Quality Indicators
In recent years, there have been several quality improvement initiatives within rheumatology, including the development of quality indicators; by measuring, reporting, and comparing outcomes, health care providers are stimulated to innovate and improve. Quality of care is operationalized by Donabedian, who distinguished structure, process, and outcome of care. Structure describes aspects in the setting in which care is delivered; for example, the setting, staff, and medical equipment. Process denotes the actions of the health care professionals; for example, whether the protocol is followed or whether the treatment is adjusted in case of a more severe disease course. The outcome reflects the effect of the given care in terms of mortality, morbidity, and health status. It is thought that more desirable outcomes are obtained if the structure provides the opportunity to deliver qualified care processes. A quality indicator is a measurable element of practice performance for which there is evidence or consensus that it can be used to assess the quality, and hence change the quality of care provided.
There is no single, broadly accepted set of quality indicators for RA, but several groups around the world have made an attempt to develop sets of RA quality indicators. These sets, developed by, among others, the National Health Service in the United Kingdom, the European Musculoskeletal Conditions Surveillance and Information Network, the ACR, and the Australian Rheumatoid Association, together contain 82 indicators, with only 9 outcome measures. Remarkably, not even 1 patient was consulted to check whether the chosen indicators were of any relevance to the health care experience. This omission is potentially a major lack of face validity and, in Porter’s words, might not add value to health care delivery. Of the 9 outcome measures, none were PROs. For example, the measurement of physical functioning is regarded as an indicator, but the value or level of functioning is not included. This omission is odd, because from a patient’s perspective it does not create value to know that your level of physical functioning has been measured; value is created when physical functioning is measured over time and actions (like additional physiotherapy) are taken to improve functioning when it lags behind. This perspective has created a shift from process measurement to outcome measurement. However, recent discussions highlight that a good balance between process and outcome measurement is needed to be able to understand quality improvement efforts.
It is evident that there is still a lot of work to be done in quality indicators and the incorporation of patients’ perspectives on quality of care. A recent study by Radner and colleagues showed that the situation is not very different for cohorts; in a survey among 25 European registers and clinical cohorts involving patients with RA, only 2 PROs, namely HAQ and patient global assessment, were identified as commonly collected. Other PROs, such as fatigue, were less frequently collected.
Patient-reported Outcomes in Remission
Treatment of RA increasingly targets low disease activity, remission, or even sustained remission. The recently developed ACR/EULAR remission criteria contain 1 PRO, the patient global assessment. During the development of the criteria, the remission committee evaluated the added value of all core set outcome measures, including the 3 PROs. From this it was learned that the patient global was equally good, if not better, at discriminating active from control treatment. Nevertheless, there has been criticism regarding the need for the patient global to equal 1 or less before a patient is classified as in remission, because comorbidity may influence the level of the patient global; another criticism is that this threshold is too low, because some respondents never use the extremities of a scale, in this case the 0, 1, 9, or 10 on a 10-point numerical rating scale, even if they feel extremely well/extremely bad.
With treatment increasingly targeting remission, it is important to understand the patient’s perspective on this state. An OMERACT working group in collaboration with EULAR has conducted qualitative research to understand which domains of health significantly contribute to patients’ experiences of low disease activity. This qualitative study, conducted in the Netherlands, Austria, and the United Kingdom, identified 3 major themes of patient-perceived remission:
- 1.
Symptoms such as pain, stiffness, and fatigue are either absent or reduced in intensity
- 2.
The impact of RA on daily life diminishes by having increased independence, the ability to do valued activities, an improved mood, and the ability to cope
- 3.
Leading to a return to normality, including the ability to work, enjoy a family role, and be seen as normal by other people
Patients thought that the concept of remission was most influenced by aging, side effects of medication, comorbidities, accrued damage to joints, and disease duration.
In order to prioritize the many domains in order of importance, patients with RA from the Netherlands, United Kingdom, Austria, Denmark, France, and the United States were asked to complete a survey that contained all 26 domains of remission identified in a qualitative study. Patients were asked to rate domains for importance (not important, important, or essential) to characterize a period of remission, and, if important or essential, whether this domain needs to be less, almost gone, or gone to reflect remission. In addition, respondents were asked to determine their personal top 3 most important/essential domains that characterize remission.
Data from 274 patients with RA were collected. The most mentioned domains in patients’ top 3 were pain (67%), fatigue (33%), and independence (19%). Domains that were most frequently rated as essential to characterize a period of remission were pain (60%), being mobile (52%), physical function (51%), being independent (47%), and fatigue (41%). Pain needed to be less (13%), almost gone (42%), or gone (45%) to reflect remission. Similar patterns were seen for fatigue (23%, 40%, and 37%), independence (16%, 31%, and 53%), mobility (16%, 35%, and 49%), and physical functioning (14%, 29%, and 57%).
At present, an international validation study to identify appropriate measurement instruments of these domains in low disease activity is ongoing. The first results will be presented during a special-interest group at OMERACT in 2016. By studying the quantification of the added value of these PROs in defining remission, this could lead to an update of the remission definition.
Patient-reported Disease Activity Scores
In 2012 the ACR Rheumatoid Arthritis Clinical Disease Activity Measures Working Group conducted a systematic review resulting in 63 RA disease activity measures. Of these, only 6 were recommended for clinical use, with 3 being entirely patient-reported: Patient Activity Scale (PAS), PAS-II, and Routine Assessment of Patient Index Data with 3 measures (RAPID3). All 3 of these indices are similar because they are the mean of 3 PROs from the core set: pain, patient global, and functional disability. Their main difference is the respective functional disability tool used: HAQ, HAQ-II, or Multidimensional Health Assessment Questionnaire (MDHAQ). Recent work has shown that RAPID3 remission, which is 3 or less on a scale of 0 to 30, is as strict as the ACR/EULAR remission criteria in the prediction of long-term outcomes.
Other RA activity scores that are completely patient reported but not recommended by the ACR’s working group because of either being too new (not enough published data) or not having remission criteria include the Rheumatoid Arthritis Disease Activity Index (RADAI) and Rheumatoid Arthritis Index of Disease Activity (RAID). The RADAI is older and includes 6 Likert scales on arthritis activity, pain, stiffness, stiffness duration, global health, joint swelling, and joint tenderness. The RAID is the patient-derived score to capture the impact of RA on daily life. It includes visual analog scales for 7 domains: pain, functional disability assessment, fatigue, sleep, physical well-being, emotional well-being, and coping. This score was developed in close collaboration between patients and professionals, has been validated in different settings, and is freely available in 67 languages and dialects from the developer’s Web site. Although the instrument is not in use as a disease activity score, cutoff state values have been proposed.
Specific patient-reported outcome measures by domains in rheumatoid arthritis
As shown in Fig. 1 , the RA core set measures contain 3 PRO domains: physical function, pain, and a patient global assessment of disease activity. Although several PRO indices using measures in these domains have qualities similar to or superior to other core set measures, each is also important to address individually because there is no clear consensus on which measures to use.
Physical Function
The physical function domain has a myriad of PROMs. The original core set stated that, “Any patient self-assessment instrument which has been validated, has reliability, has been proven in RA trials to be sensitive to change, and which measures physical function in RA patients is acceptable.” They recommended the Arthritis Impact Measurement Scales (AIMS), HAQ, and a few others that have seen much less use in rheumatology during the past 20 years. The AIMS and subsequent AIMS2 had more than 60 items and covered several health domains in addition to physical function, which made them difficult to implement outside of clinical trials. Although the original long HAQ also contained additional nondisability items, its short form, the HAQ Disability Index, became the gold standard for use in clinical trials and clinics for decades. With 20 Likert items and 21 questions about devices, the HAQ was similarly difficult for rheumatologists to use consistently in the clinical workplace as pressure mounted on reducing patient wait times and simplifying scoring.
The first major change to the HAQ was the modified HAQ (MHAQ), which used a single item from each of the 8 dimensions in the HAQ and simplified scoring by taking the average item value. Because of large floor effect issues with the MHAQ, Pincus and colleagues added 2 difficult items, playing sports and walking 2 miles, to the MHAQ to create the 10-item Multidimensional HAQ (MDHAQ). Using Item Response Theory (IRT), a quantitative method that assesses item properties and information content, Wolfe and colleagues developed the HAQ-II, a 10-item measure that used 5 legacy items from the HAQ and 5 new items to efficiently reproduce the original 41-item HAQ. One issue with having so many similar measures was that some confused the HAQ, MHAQ, MDHAQ, HAQ-II, and so forth as being interchangeable, although clinicians now better understand their important differences.
Expanding this use of IRT, the US National Institutes of Health developed the Patient Reported Outcomes Measurement Information System (PROMIS), which has created a series of physical function measures for general use. These measures include physical function 10-item (PF-10) and 20-item (PF-20) measures that have been shown to be effective in patients with RA. Fig. 2 shows the comparison of information content (measured by standard error) and physical function (theta) in the HAQ, PROMIS PF-10 and PF-20, and other measures. Of note is how well the HAQ performs in patients with the worst physical function and how the PROMIS measures perform better in patients with the best physical function. As expectations for patients with RA have improved since the HAQ was first developed to measure disability (or worse physical functioning), there has been growing interest in incorporating these newer PROMIS physical functioning measures into rheumatology studies. In addition, PROMIS has developed a computerized adaptive test (CAT) for physical function that requires an electronic interface but can provide similar results to the PROMIS PF-10 in only 4 to 5 items.
Pain
Pain is one of the most common PROMs and, through general quality improvement needs since the late 1990s, it has mostly become established as the fifth vital sign, although its general utility in the clinic is unclear. Although the primary vital sign measure of pain is a numerical or verbal rating scale with a range of 0 to 10, pain in RA is often measured on a 10-cm visual analog scale (VAS) between anchors of 0 (no pain) and 100 (unbearable pain, severe pain, worst imaginable/possible pain, and so forth). Studies in patients with RA and other patients with chronic pain have shown that differences of 5 points or less (out of 100) are negligible compared with minimal clinically important differences, and some clinicians have resorted to 11-point and 21-point numeric rating scales, especially as forms became machine readable. For example, Wolfe and Michaud found that a generally acceptable level of pain in 12,090 patients with RA was less than or equal to 20 and a minimal clinically important improvement (MCII) of 11. Tubach and colleagues found from 358 international patients with RA that their pain MCII was 15 absolute points or a relative 20% improvement. This finding is similar to reviews of other chronic pain conditions ; depending on the baseline amount of pain, the MCII may similarly change (eg, 20, 25, and 30 for least, average, and worst pain baseline).
Limitations of measuring pain through a single item are not recent, and some clinicians have proposed using additional items in the clinic and in trials. The Rheumatoid Arthritis Pain Scale (RAPS) is the only RA-specific pain PRO measure and has 24 items. The Regional Pain Scale (RPS) examines the extent of bodily pain through its 19-item measure, and helps identify concomitant fibromyalgia. Many investigators report pain values as subscales of larger health-related quality-of-life measures such as the Short-form 36 (SF-36) and AIMS. In addition, PROMIS has 2 pain dimensions, pain intensity and pain interference, with CAT and short-form measures providing reliable results in patients with RA.
Patient Global
Patient global assessment of disease activity, or patient global, is a simple and important measure, but controversial in its interpretation. Physicians have always asked patients how they are doing and have often used the answer to help guide their medical decisions. Confusion arises in determining how much of the patient’s current health problems can be assigned to their RA, and whether comorbidities, treatment side effects, and other health problems that may or may not be RA related should be included so that patient global is similar to the patient global assessment of general health. Even though physicians usually synthesize the patient global measure when determining their own physician global assessment of RA activity, these measures often disagree in meaningful ways. The ACR/EULAR remission guidelines recommended the following wording: “For patient global assessment, ‘Considering all of the ways your arthritis has affected you, how do you feel your arthritis is today?’ (anchors: very well–very poor).” Subsequent studies have shown that patient global is the measure most likely to prevent a patient from achieving a score that indicates remission. Future work is needed to determine how to separate the impact of comorbidities and mood disorders; how to test psychometric changes for patients who avoid selecting the VAS anchors; and how sociolinguistic changes, including time windows, affect this outcome measure.
Additional Patient-reported Outcome Domains
Although much of the RA PRO research literature has focused on the 3 core measures mentioned earlier, there are a variety of others to consider, as Fig. 1 shows. A systematic review of PROs in RA clinical trials found 63 PROMs covering 14 domains of health. Fatigue, sleep, and depression are discussed later, and readers are invited to investigate further the several other PRO domains, including work/productivity, coping, well-being, and general quality of life.
Although the high prevalence of fatigue in RA was established early on, there are a variety of ways of measuring it. In their 2007 systematic review, Hewlett and colleagues found 23 different fatigue scales used by patients with RA, but only 6 provided evidence of reasonable validation. Arguments have been made for a validated multidimensional fatigue scale for RA trials at the same time increasing the use of a fatigue VAS in the rheumatology clinic. Because most studies have used a single VAS to measure fatigue with a wide variety of anchors, when analyzed, the MCII have been found to be smaller than pain at 10 (out of 100). In trying to understand the pathways on how to improve fatigue for patients with RA, studies found that improving pain, function, and depression had the greatest impacts. Outside of DMARD treatment, exercise has been shown to have a significant effect on improving fatigue.
Often connected with fatigue, sleep disturbances are increased in RA and have important associations with disease activity, pain, and overall quality of life. Because patients emphasized the importance of quality sleep, there was a transition from polysomnography studies on small groups to using sleep PROs on larger groups. A 2009 review found 45 sleep questionnaire instruments covering 14 domains of sleep and recommended 4 for testing in future RA clinical trials: Athens Insomnia Scale, Medical Outcome Study (MOS) Sleep Measure, Pittsburgh Sleep Diary, and Women’s Health Initiative Insomnia Rating Scale. However, the most commonly used in trials have been simpler items measuring frequency of sleep disturbances, quality of sleep, and sleep VAS. This trend changed during 2 RCTs for abatacept when the MOS sleep questionnaire was used and validated. Other observational studies have additionally helped validate sleep PROMs and have shown their association with RA activity, and the minimal clinically important difference (MCID) was lower for the sleep VAS at 6 (out of 100).
Although depression is increased in RA compared with healthy controls, it does not seem to be increased compared with chronic diseases with similar pain levels. The prevalence of depression is associated with younger age and the screening tool used, with levels ranging from 15% for the Hospital Anxiety and Depression Scale (HADS) to 39% for the Patient Health Questionnaire (PHQ-9). In RA treatment clinical trials there have been at least 3 PRO tools used to measure depression: Beck Depression Inventory (BDI), HADS, and Mental Health Inventory (MHI) depression subscale. Observational studies have shown a strong association of pain extent, fatigue, ethnicity, and low socioeconomic status with self-reported depression or PHQ-9. Even with the large overlap with other domains, there seems to be an independent temporal impact of depression symptoms with RA activity. Although there are several depression PRO questionnaires validated in patients with RA, it is not clear which, if any, are feasible, practical, and helpful for treating RA in the clinic.
Related posts:
Patient Reported Outcomes in Rheumatic Diseases
Patient-Reported Outcomes in Systemic Lupus Erythematosus
Patient-Reported Outcomes in Psoriatic Arthritis
Patient-Reported Outcomes and Fibromyalgia
Challenges and Opportunities in Using Patient-reported Outcomes in Quality Measurement in Rheumatology
The Promise of Patient-Reported Outcomes Measurement Information System—Turning Theory into Reality
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
