A platform‐based approach to assisting rheumatoid arthritis management

Abstract

Rheumatoid arthritis (RA) is one of the major health issues affecting around 14 million people worldwide. It is characterized by an attack on healthy T-cells and B-cells by the immune system and by periods of low disease activity and flares.1 The underlying causes of this disease have not been fully determined so far. This disease costs the healthcare system and indirect expenses of 19.3 billion dollars annually in the USA.2 It is therefore a major concern in terms of both quality of care for the patient and healthcare cost reduction, an issue that could be addressed by a platform-based approach. Currently, the management of RA faces numerous challenges related to the control of disease activity. Diagnosis is often conducted at a late stage, while it conditions the evolution of the disease and the development of comorbidities.3 It is difficult to follow the disease activity, especially outside of clinical settings, and to rely on accurate data to adjust the treatment plan. This is further enhanced by the fact that flares are also hard to predict and are responsible for irreversible joint damage.4 Supportive treatments have demonstrated their increasing importance over the last few years but continue to be insufficiently integrated into the care pathway. Asia Pacific League of Associations for Rheumatology (APLAR) has committed to the development of digital medicine and telemedicine. The treat-to-target (T2T) strategy has been advocated for the past 2 decades. However, this strategy remains limited due to low patient adherence and inappropriate treatments. These challenges result essentially from the lack of an integrative approach and thus a better understanding of the disease. A platform-based approach leverages different steps of the care pathway while allowing the caregiver to focus on value-added tasks. Governments and companies have understood the importance of launching many joint projects, such as the PICASO Project.5 First, centralization and crossing of data improve the detection and assessment of the patient's disease stages. The integration of all the patient results associated with machine learning solutions within the same platform allows the improvement of evaluation accuracy of the patient's health status. The platform allows a large amount of data to be aggregated, such as clinical, imaging, genomic, administrative, and population data. For example, the classification of RA is associated with machine learning solutions and therefore could be performed with more precision.6 This requires both harvesting infrastructures and analysis tools. Their development should accelerate the future identification of new biomarkers, such as metabolomics biomarkers and at-home testing, such as dried blood spot.7 Teleconsultation has shown its impact during COVID-19 for RA patients maintaining continuity of care at home.8 Moreover, it is appropriate for a population that is sensitive to mobility loss, by allowing patient follow-up and the renewal of prescriptions. This approach is likely to be extended and deepened within a hybrid virtual and face-to-face system, fostering stakeholder collaboration throughout the patient's care pathway. The challenge for the caregiver is to access data in a context where routine evaluation tests would be increasingly performed at home. A recent study has shown that the performance of a self-assessment of a swelled joint in remote monitoring was comparable to the assessment with healthcare professionals.9 Second, a platform also allows collection of large amounts of data, including passive data from wearables and smartphone apps to track non-invasive digital biomarkers. It enables continuous monitoring outside the clinical setting at a low cost of use. A consensus is emerging on the key monitoring parameters, especially activity and pain.10 Additional biomarkers that are more specific to the disease are being developed. For example, a study showed the ability of a solution to detect joint swelling with a specificity of 75% and a sensitivity of 88% using real-world images.11 These data gathered along electronic patient-reported outcomes (ePROs) into the electronic health record would help clinicians to have a more accurate picture of the patient's health status. The process is furthermore systematic and standardized while avoiding the bias of self-reported data. By identifying patterns, it contributes to the anticipation and prevention of flares. Activity monitoring associated with machine learning could, for example, predict the occurrence of flares with a sensitivity and specificity >95%.12 Third, adaptation of the treatment plan is enhanced with remote monitoring and resources for decision support. A study has shown the relevance of connected monitoring following the initiation of disease-modifying antirheumatic drug therapy with an impact on quality of life while reducing the number of consultations.13 Platforms aim to optimize the treatment using reliable data for evidence-based decisions. The impact of each decision could be assessed using metrics such as disease activity and achieving remission. Faced with a change in disease activity, decisions are made more rapidly, resulting in accurate prediction of outcomes and the risk of error is reduced. Treatment plans could be even more personalized by incorporating data related to patients' characteristics, medication records and lifestyles. It could also enable the establishment of population subgroups and promote access to therapeutic innovation by targeting patients. Such data would allow a better understanding of the etiology of the disease and to move toward precision medicine. This is also made possible using a large volume of data regarding a highly complex disease. Through rethinking the interactions between the patient and the caregiver, it allows the patient to be informed about their own condition and to optimize their care pathway. It contributes to their empowerment in the management of their disease, especially outside the hospital setting. This has an impact on the maintenance of positive behaviors and the risk of relapse. A study has shown the impact of asynchronous monitoring to improve adherence and accelerate remission.14 The platform can also be used to support the development of Health Apps which can reach specifically RA patients on topics like medical information, detection of depression, pain management and behavioral education. A single-blinded randomized controlled trial was conducted on a self-management app focusing on the function of the hand and suggested superior outcomes compared to conventional care.15 In addition, by integrating a platform, time can be released for the caregiver to focus on high-value tasks. This includes prioritizing and targeting patients according to their health and well-being needs. The platform allows continuous monitoring compared to a single consultation that alone does not appear to address the different phases of the disease. It also saves time16 and improves quality of life for the patient. However, many obstacles remain in the implementation and adoption of a platform-based approach. Concerns exist about the collection of accurate and quality data from various sources prior to analysis and generating insights. Differences in performance have been observed between wearables for activity tracking. The use of machine learning solutions on collected data may also not be entirely evidence-based and may present biases. Data protection and privacy is a major stake in complying with regulations and responding to patient anxiety. Patients may be resistant to the use of wearables, which they may consider intrusive due to the perception of being over-supervised. Failure to achieve specific goals may have an impact on their motivation and discourage them from continuing their physical activity.17 It is thus necessary to consider how the patient experiences a platform, including their perception, self-esteem, and feelings. To increase adoption, it is key to involve patients at the center and include them in the development process in a co-creative way. The frequency of using eHealth solutions in the population specifically for RA is still relatively low. A reason might be differences in health literacy and digital literacy. Patients may also already be satisfied with the current monitoring and decision process, thus not needing additional capacity.18 This may be exacerbated by a lack of quality apps that are thus responsible for decreased patient adherence after adoption.19 Human-centered design and subsequent interventions assessment are thus crucial to address needs and increase patient willingness to use an app.20 A platform-based approach is tailored to the evolution of RA and is expected to rapidly expand to address the healthcare burden. It enables us to reach a T2T strategy and overcome the lack of adherence to patient empowerment. It also allows integration of all the dimensions of the disease by removing the boundary between healthcare and well-being but also to nudge positive and long-lasting behavior. To some extent, this approach could help understand the etiology of RA. However, further clinical studies are required to confirm long-term outcomes. PW contributed to the conception and writing of the original draft of this paper. CKK contributed to reviewing and editing of this paper. CWC contributed to the conception, writing, review, and editing of this paper. All authors were involved in the study design, and writing of the report, and had final responsibility for the decision to submit for publication. The authors thank colleagues and staff for the support and help provided in this study. The authors declare they have no conflicts of interest. The authors confirm that the data supporting the findings of this study are available within the article.

Topics & Concepts