Feasibility of a musculoskeletal ultrasound intervention to improve adherence in juvenile idiopathic arthritis: a proof-of concept trial

Juvenile Idiopathic Arthritis (JIA) is the most common rheumatic condition affecting children worldwide, with an estimated prevalence rate of 132 per 100,000 [1]. The disease spectrum includes classification into seven subtypes based on the number of joints affected and differences in symptom development and prognosis. Disease control is based on management of inflammation and progression of disease while limiting disability and deformity over time. The treatment regimen for JIA is multifaceted and includes oral and/or injectable medications, physical and occupational therapies, and some lifestyle modifications. Studies focused on patients with polyarticular JIA indicate that early and aggressive therapeutic approaches increased the likelihood of attainment and maintenance of disease remission [2, 3]. Nevertheless, even after remission is attained, the current risk of disease recurrence reaches approximately 40%, which in part may be due to waning treatment adherence [4].

Adherence is defined as the extent to which a person’s behavior follows healthcare treatment recommendations [5]. Factors that contribute to poor adherence include delayed benefits in early treatment, need for consistency over a long period of time, and negative side effects [6]. Unfortunately, poor adherence in JIA impacts health-related quality of life [7] and rates of active disease control [8]. For example, prescription refills for adherence to disease modifying anti-rheumatic drugs (DMARD) and injectable anti-tumor necrosis-alpha drugs only occurred 46.9 and 65.7% of the time, respectively [9]. Our prior work studying adherence barriers in the current treatment landscape indicates that 70% of adolescents with JIA and 77% of their caregivers endorse at least one adherence barrier [10]. The most commonly reported barriers include worry about future consequences, pain, and forgetting, which are all amenable to intervention. In fact, recent meta-analyses have demonstrated that healthcare provider-based adherence interventions have greater effects on adherence compared to other multi-component interventions [11]. To date, there are no evidence-based adherence promotion interventions within standard clinical rheumatology practice and this current study sought to establish proof-of-concept for a rheumatologist-provided intervention.

Musculoskeletal ultrasound (MSKUS) is an emerging form of disease surveillance that an increasing number of rheumatology providers are utilizing in clinical practice. Symptoms of chronic arthritis due to synovial hypertrophy and fluid accumulation within the joint space can accurately be identified in real-time during routine clinic visits [12, 13]. Ultrasound studies are efficient in both time and cost expenditure, as compared to the imaging gold standard MRI, and are able to spare the use of radiation or intravenous (IV) contrast burden [12]. Some evidence has suggested that subclinical disease activity including synovitis can be identified by the use of MSKUS surveillance [14–16]. MSKUS provides a longitudinal, objective characterization of disease status that allows an important opportunity for patient inclusion via demonstrative real-time education. Specifically, visual demonstration modalities reduce limitations imposed by healthcare literacy variation and possibly may better equip patients with knowledge regarding their disease [17]. Additionally, MSKUS may foster increased patient-provider interaction through shared disease visualization and discussion. Studies assessing adults with rheumatoid arthritis (RA) revealed that those who underwent ultrasound evaluation had more confidence in their ability to complete treatment recommendations [18, 19].

As the routine use of ultrasound increases in pediatric rheumatology, our aim in this proof-of-concept study is to assess the impact of MSKUS on adherence in adolescents with polyarticular JIA. This pilot study serves to efficiently establish the merit for more rigorous testing of the intervention. Our hypothesis was that MSKUS would be rated as feasible and acceptable to patients and parents and that following receipt of an MSKUS intervention, participants would demonstrate improvement in electronically-monitored and patient-reported adherence measures.

This proof-of-concept study was designed to test a MSKUS education-based adherence intervention amongst adolescents with JIA and their caregivers. Innovative aspects of the current study includes the novel multi-modal method used to demonstrate oral and injectable methotrexate adherence rates in this population, and the application of a rheumatology provider-based educational intervention for adolescents and their caregivers within the clinic setting.

Based on our retention rate of 80% and intervention completion rate of 100%, our study design and intervention methods appear to be feasible. The majority of participants found the ultrasound intervention to be clear, helpful and beneficial. Additionally, the MEMS TrackCap was an acceptable form of adherence measurement with no malfunctioning or missing data noted in the caps that were returned. Further, it was able to capture medication-taking behavior of both formulations of methotrexate, which is a significant strength of the study. To our knowledge, this is the first time MEMS TrackCaps have been used to measure adherence to injectable arthritis medications in pediatric rheumatology. Although adherence rates were capped at 100% for analyses, it is important to note that one family was actually found to have administered more than the prescribed regimen based on the patient living in two households due to divorce. That particular family valued the use of the TrackCaps for added accountability.

While mean adherence did not change significantly from pre to post intervention and correlations cannot be made based on the study size and scope, maintenance and improvements in patient self-reported and electronically-monitored adherence were observed in some patients. Additionally, the percentage of patients with “good” adherence (> 80%) [35] rose from 50 to 75% percent. Given that adherence typically declines over time [36], improvements and/or maintenance when adherence is already high is considered successful. Notably, 50% of our patients were non-adherent at baseline, which is consistent with the larger literature [23, 34], indicating that adherence is a significant issue for these patients.

Our data assessing self-reported and parent-reported adherence is consistent with prior studies suggesting inflation or over reporting. This may reflect response bias or pressure around social desirability. Prior data assessing self-reported adherence for adolescents with JIA (N = 116) on methotrexate found a median adherence of 98% [37]. Additionally, parent perceived adherence using the PARQ over a one year time period reported similarly high rates, varying between 86.1 and 92.0% [38]. In that study, the most highly correlated factor predicting higher self-reported adherence was believing that the therapy was helpful. This highlights the fact that patient beliefs influence the way self-management behaviors are performed and may be further evidence of the importance of standard patient education, with ultrasound being one method [39].

Perhaps the most salient aspect of this study design is the delivery of the intervention by a rheumatologist compared to other healthcare providers. The majority of JIA-focused adherence interventions were prior to biologic therapies and/or utilized psychologists, nurses or peer-mentors to perform the interventions [6, 40–42]. Since the advent of biologics, families of children with JIA have navigated more complex treatment regimens. Our prior work studying adherence barriers in JIA indicated that worry about JIA treatments was common for caregivers and patients [10], which suggests a need for more communication regarding specific concerns. Further, in a study assessing perceptions of adolescents prescribed biologic therapy, most adolescents preferred an active and involved role with dedicated education during their clinic visits [43]. Real-time, objective ultrasound interventions performed by rheumatologists, informs the patients of up-to-date disease status and presents the opportunity for additional questions or concerns regarding care. In some cases, it also justifies the need for therapy in the absence of perceived symptoms as well as the reverse. Our intervention not only included a review of normal joint anatomy and how ultrasounds can be useful in JIA, but patients were encouraged to actively participate in the ultrasound examination by seeing exactly what fluid or synovitis looks like in their own joints. This prompted many patients to have further discussions with their rheumatologists following the examination with a new sense of knowledge that they would have lacked prior. Previous work has established that both patients and providers value the information provided by ultrasounds [44]; however, more research is needed to measure the potentially compound effects that the receipt of an ultrasound can provide.

One unanticipated finding from this study was the benefit of adjunctive MSKUS for clinical care of patients with JIA. Despite the study being short-term, MSKUS aided in the detection of subclinical disease and informed therapeutic change. In our sample, the rate of abnormalities on ultrasound examination was 87.5% with two-thirds of those being discrepant to the rheumatologist’s physical examinations. Our rate of subclinical disease discovery was higher than prior study in pediatrics [15]. There was no recruitment screening done for patients presumed to be in active disease; thus, many of the rheumatology providers were surprised by the ultrasound findings. Outcomes in our sample revealed improvement overall, likely related to attainment of disease control based on optimized medication therapy. This is further supported by the fact that in the subset of patients that were started on a biologic during the intervention visit, the interval improvement in cJADAS score was more clinically meaningful as compared to patients that did not incur a therapeutic change. While unable to be correlated during this study, uncovering subclinical disease is one clear factor that could help to engage patients to adhere more closely to therapies and activate providers to screen for nonadherence more vigorously. This was mentioned by several rheumatologists upon hearing the results of their patient’s ultrasounds in our study.

As this was a proof-of-concept study, there are several important opportunities for improvement. Our small sample size was composed of a heterogeneous collection of participants including differences in therapy regimens, duration of disease and disease severity. Standardizing these parameters, increasing the sample, and testing the intervention via a randomized-controlled trial would increase our ability to detect the impact of our intervention on outcomes. An attention control arm including a verbal-education only group would aid in assessing the true effect of ultrasound imaging as an educational intervention. Also, the inclusion of two participants that received their intervention despite having a baseline adherence rate of 100% may have altered results due to ceiling effects. Any subsequent work in this area should exclude patients who demonstrate high adherence and thus do not need adherence interventions. Adherence interventions in other disease populations have successfully used this approach [32, 45]. Additionally, our short study duration of 12 weeks may have not provided sufficient time to examine behavioral changes over time. A longer study duration would also better demonstrate changes in our secondary outcomes (i.e. HRQOL and cJADAS) as these typically require longer monitoring to display changes over time. Furthermore, the medication regimen changes made during the study in several patients, may have altered the treatment routines of patients, thus potentially adding or subtracting barriers to adherence. However, the ethical need to respond to new information provided by the ultrasound, precludes avoidance of these therapeutic changes. Additional specific improvements that could be made to the ultrasound intervention itself include multi-provider blinded testing as to decrease biases imparted by the study team.

This study emphasizes the continuing importance of the development of adherence interventions for use in JIA. Following more rigorous testing, a multimodal adherence intervention including visual demonstration of disease using ultrasound could become common practice as either primary or secondary adherence promotion. There is still much work to be done in order to understand and encourage our patients to better navigate self-care and participate in their health. As providers of children with chronic illnesses in the midst of advancing technologies, we must continue to seek solutions to better bridge the patient-provider gap and deliver education to improve outcomes.

Our study highlights an additional beneficial and novel use of MSKUS as a patient education tool. In this proof-of-concept trial, patients and families saw the purpose and helpfulness of real-time disease demonstration via ultrasound. We document that quantitative, electronically-measured adherence to methotrexate is, not only less than self-reported measures, but also below acceptable thresholds for therapy. We have also contributed to growing literature that establishes the utility of MSKUS as a useful clinical tool, with the potential to be harnessed for use in patient activation.