A scoping Review of tools used to assess patient Complexity in rheumatic disease

Abstract Objective Patients with rheumatic diseases often have multiple comorbidities which may impact well‐being leading to high psychosocial complexity. This scoping review was undertaken to identify complexity measures/tools used in rheumatology that could help in planning and coordinating care. Methods MEDLINE, EMBASE and CINAHL were searched from database inception to 14 December 2019 using keywords and Medical Subject Headings for “care coordination”, “complexity” and selected rheumatic diseases and known complexity measures/tools. Articles describing the development or use of complexity measures/tools in patients with adult rheumatologic diagnoses were included regardless of study design. Included articles were evaluated for risk of bias where applicable. Results The search yielded 407 articles, 37 underwent full‐text review and 2 were identified during a hand search with 9 included articles. Only 2 complexity tools used in populations of adult patients with rheumatic disease were identified: the SLENQ and the INTERMED. The SLENQ is a 97‐item patient needs questionnaire developed for patients with systemic lupus (n = 1 study describing tool development) and applied in 5 cross‐sectional studies. Three studies (a practice article, trial and a cross‐sectional study) applied the INTERMED, a clinical interview to ascertain complexity and support coordinated care, in patients with rheumatologic diagnoses. Conclusions There is limited information on the use of patient complexity measures/tools in rheumatology. Such tools could be applied to coordinate multidisciplinary care and improve patient experience and outcomes. Patient contribution This scoping review will be presented to patient research partners involved in co‐designing a future study on patient complexity in rheumatic disease.


| INTRODUC TI ON
Defining the burden of health issues on individuals and health systems is an evolving and nuanced field. 1,2 Since at least the 1970s, it has been proposed that biological, psychological and social factors influence prevention, causes, presentation, management and outcomes of diseases. 3 This was conceptualized as a biopsychosocial model of medicine by Engel. 4 Despite this long-standing holistic model, for decades in medicine there has been an emphasis on measuring the burden of illness in terms of number and interrelatedness of various health conditions including measuring 'comorbidities' or 'multimorbidity', 1,2 with little attention paid to the interaction between health, illness and the social determinants of health in care delivery at an individual patient level.
An evolving concept is the notion of patient complexity. 2 Patient complexity can be defined according to the vector model 5 as the 'interacting impact of biological, socioeconomic, cultural, environmental and behavioral forces as health determinants'. From a clinicians' perspective, disease factors interact with psychosocial and environmental factors, creating challenges in clinical management, increasing time required in medical appointments and altering health-care resource use. 2 From a patient perspective, a model of cumulative complexity has been proposed by Shippee et al 6 highlighting that patients experience a 'workload' of demands (job, social commitments, disease management) and have a 'capacity' (functional status, socio-economic resources, literacy) to address these demands and that imbalance in workload vs demand contributes to complexity which can impact health outcomes.
Previous strategies to evaluate patient complexity have included examining the number of physician types involved in a patient's care and/or counting the number of a patient's comorbidities. 7 Unfortunately, these strategies still lack integration of the psychosocial and environmental aspects of care needs, or an evaluation of the need for coordination of multiple services for individual patients.
A match between one's case and care complexity is needed for efficient and effective patient-centred health care and can be used to better direct patient support, care coordination and care plan monitoring. 8 A variety of instruments have been developed to measure patient complexity across multiple health-care contexts including ambulatory and acute care settings. These tools have been used in health-care planning and care coordination, and can predict length of stay. 9,10 One of the earliest instruments was the INTERMED.
Developed in the late 1990s, the INTERMED proposed a method for assessing past, present and future health service needs as a means of enhancing communication between health-care professionals for patients with chronic disease to assist coordination of care and services. 11 The information for completing the INTERMED is collected during a 15-minute clinical interview as part of the medical history. It synthesizes data from biological, psychological, social and health-care systems over time with each domain scored from 'no vulnerability or need' to 'high vulnerability or need' (Supplemental Table 1). 11 Beyond its use in research, indicator colours, similar to those of a stop light, have been used to enhance communication to direct the need for health-care team action items in different domains. 12 Rheumatic diseases are typically systemic and therefore have the potential to impact multiple organ systems requiring lifelong frequent encounters with multiple health-care providers.

Complications of inflammatory rheumatic diseases are frequent
and result in additional disease burden. 13 People living with rheumatic diseases may experience challenges with mobility and physical functioning, which affect activities of daily living and can cause role limitation. 14 Rheumatic disease can also cause significant personal and societal economic impacts as well as challenges with employment. [15][16][17][18] People living with rheumatic disease are also vulnerable to psychosocial challenges and concomitant mood disorders, including anxiety and depression. These common conditions contribute to decreased health-related quality of life. [19][20][21][22] Socio-economic status in turn may impact stress, depressive symptoms and disability as has been recently been evaluated in systemic lupus. 23 Addressing a patient's psychosocial distress has been shown to improve coping and self-efficacy, reduce psychological distress and reduce pain in arthritis patients. 24 Patient complexity tools may help elucidate psychosocial challenges in a systematic fashion and can offer the care provider a holistic picture of the patient's experience beyond the physical manifestations of their disease and can be used to target interventions to improve health outcomes. 25 The extent to which complexity tools have been used to better direct care coordination in rheumatology to improve health outcomes is unknown. The purpose of this scoping review was to understand and describe how patient complexity tools/measures have been used in rheumatic diseases to plan a future study aimed at improving care coordination and patient outcomes in rheumatology.

| MATERIAL S AND ME THODS
A scoping review is an appropriate knowledge synthesis strategy for this research topic given the emerging nature of patient complexity as a construct, especially in rheumatic disease. Scoping reviews are used to provide a broad overview of evidence and this strategy is helpful in determining knowledge gaps to plan future research. 26,27 This scoping review was developed and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR). 28,29 A scoping review protocol was developed a priori (available on request). The search strategy was developed in consultation with a medical librarian following a series of iterative steps including 1) a preliminary search of MEDLINE and EMBASE to identify complexity tools and review text words and index terms for the identified articles; 2) a search using all identi- Observational Cohort and Cross-Sectional Studies from the National Heart, Lung, and Blood Institute (NHLBI). 35 Items of the tool were evaluated 'yes', 'no', 'not applicable', 'cannot determine' or 'not reported', which helped guide provision of an overall rating for the quality of each study as 'good', 'fair' or 'poor'.

| RE SULTS
A flow diagram of the search findings is presented in Figure 1.
The search strategy returned 407 articles; 37 were identified as potentially relevant based on title and abstract screen, and 30 were excluded after full-text review. A hand search identified two additional articles for a total of nine included articles. No additional articles were identified after contacting authors of existing complexity tools. Response rate from authors of existing tools was 60%. Table 1 describes the characteristics of the nine included studies: one was in RA, 36 one in general rheumatologic disease 25 and seven in SLE. 12,[37][38][39][40][41][42] In SLE, one study described the development, and five described the application of the Systemic Lupus Erythematosus Needs Questionnaire (SLENQ) in SLE populations in New York, USA; New South Wales (NSW), Australia and the Netherlands. [37][38][39][40][41][42] The INTERMED was studied in three articles. The first was an observational study in RA where cluster analysis was completed based on patients' INTERMED scores to determine clinical and disease associations with raw scores. 36 An RCT used the INTERMED in a general rheumatology population to identify patients that were randomized to a tailored psychiatric intervention. 25 One final article described the application of the INTERMED in a practice case with an SLE patient admitted to a gastroenterology ward to illustrate the utility of the tool but was not a research study. 12  The remaining five articles utilizing the SLENQ were cross-sectional survey studies. Two studies 37,38 highlighted patient characteristics associated with higher needs: patients with chronic symptoms or frequent flares, lower levels of education, those who were unemployed or receiving disability and those on social assistance.

| SLENQ: A tool to measure patient need in SLE
These studies also found that physical appearance changes (eg hair loss) and muscle pain were associated with higher patient-reported needs. Moses et al 39 reported that the most prevalent unmet needs were found in the physical domain with 84% of respondents reporting 'moderate-high' needs. The top 10 highest levels of unmet needs were tiredness (81%), pain (73%), not being able to do things one used to do (72%), fear of exacerbation (72%), sleeping problems (70%), anxiety and stress (69%), feeling down (68%), fearing physical disability (65%), explaining SLE unpredictability to others (64%) and concerns regarding other health problems (63%). 39  proposed to identify patients who could benefit from case management. 45 In this study, 29% of the patients scored at least 21, all of whom were in the complex cluster. 36 Stiefel et al's 25

| Quality assessment of studies
There was a high risk of bias as measured by the Cochrane tool 31 for the RCT by Stiefel et al 25 (Supplemental Table 2). Concern for bias was present in the following domains: blinding for all outcomes; incomplete outcome data; and other possible sources of bias (including absent power calculation, potential selection bias).
The COSMIN checklist [32][33][34] was used to evaluate the study by Moses et al 40 for its measurement properties (Supplemental Tables 3, 4). Unfortunately, we were unable to access the full SLENQ, therefore our own rating of the questionnaire is unavailable, and the overall quality of the SLENQ was rated as indeterminate.
The NHLBI quality assessment tool 35 was used to assess the methodological quality of the five cross-sectional studies and one observational study [36][37][38][39]41,42 (Supplemental Table 5). Outcomes were used to assess the internal validity and risk of bias for each study, and the overall quality was rated. All six studies had an overall rating of fair.

| D ISCUSS I ON
This scoping review of complexity measures used in patients with rheumatic disease highlights a paucity of measures/tools used to evaluate patient complexity in this field and also emphasizes the evolving construct of patient complexity. Two tools were identified that have been investigated in populations with rheumatic disease to date: the INTERMED and the SLENQ.
As the SLENQ was developed specifically for patients with SLE, the study describing its development and psychometric properties was included in our review and evaluated using the COSMIN checklist. We have concluded that the tool was of indeterminate quality, limited by the inability to provide our own rating of the questionnaire as instructed by COSMIN. Nonetheless, the SLENQ is reported to be acceptable, valid and reliable. 40  in 'workload-capacity demands' and therefore understanding complexity from a patient-centred perspective. Further work is clearly needed to better understand and define these concepts and their relationships in rheumatic disease.
As the INTERMED was not developed specifically for patients with rheumatic disease, the articles describing its development 11 and psychometric properties, including its validity, 47 predictive validity 48,49 and reliability, 11 were not evaluated in the present scoping review. Nevertheless, based on the limited findings of our review it appeared to appropriately identify patients with rheumatic disease (primarily RA) who had higher patient complexity that was not explained by their disease activity and there was limited evidence that this tool could be used to direct psychosocial interventions to improve patient outcomes.
Through initial stages of our scoping review, while developing our search strategy, we identified a number of other complexity tools; however, none of these met the inclusion criteria as they had not been used in rheumatic disease populations. In general, these other tools are used to evaluate complexity, identify factors interfering with care, indicate multidisciplinary care and facilitate care coordination, [50][51][52][53][54][55][56][57] with the exception of three that were designed in particular for end-of-life care [58][59][60] (Table 2). There is significant overlap between the domains of these other assessment tools with those of the INTERMED and SLENQ (Table 3). Of note, many of the tools The presence of multiple chronic conditions is a growing reality for many patients with rheumatic diseases. 13

ACK N OWLED G EM ENT
The authors would like to acknowledge Nicole Dunnewold MLIS, a Research and Learning Librarian at the University of Calgary for her assistance with development of a search strategy for this scoping review.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.