Systemic lupus erythematosus (SLE) is the most common life-threatening autoimmune disease and a leading cause of mortality among young women of Black race and Hispanic ethnicity.1 Treatment of SLE is complex, requiring immune-altering medicines to reduce disease activity and prevent disease-related damage. The decision to commit to long-term treatment with immune-altering medicines is difficult, especially given that such medicines carry the risk, albeit low, of side effects, teratogenicity, and toxicity. In the face of such a difficult decision, which is compounded by individual barriers such as economic insecurities, social desirability, mistrust, or cognitive dysfunction (Figure 1), more than 50% of patients self-discontinue medicines within 9 months, and 33% do not refill medicines.2-4 This strikingly high medication nonadherence accelerates the risk of flares requiring hospitalization and drives early mortality in SLE.3,5-8

Conceptual model of factors influencing the decision to take medications as prescribed. Med: medication; SLE: systemic lupus erythematosus.

Taking medications as prescribed is a behavior driven by many interrelated factors (Figure 1). Patients have noted that they frequently receive incomplete information about medications during their often-hurried SLE visits.9 They may resort to using the internet to learn about their medications, which often confirms their preexisting fears of toxicity, prompting self-discontinuation.9,10 Moreover, patient knowledge of the benefits of medications may be suboptimal. For example, in one study, more than 80% of patients with SLE reported that they did not know that hydroxychloroquine could improve survival and protect organs in SLE.10 Thus, dedicated provider-patient discussions on the benefits vs harms of SLE medicines could support a patient’s decision to continue therapy. Such discussions often work but can be time-consuming and difficult to adopt in busy clinical practices and low-resource settings. To further complicate this issue, without objective methods to monitor adherence, rheumatologists might assume that patients are taking medicines as directed, and patients might not broach the topic of medication concerns due to social desirability bias. Sun and colleagues sought to address some of these challenges in their recent study, published in this issue of The Journal of Rheumatology.11

Sun et al11 recruited patients with SLE who had low medication adherence (N = 25), defined as < 80% average proportion days covered (PDC) for all SLE medicines, from a single clinic to test the effectiveness of a clinician-led intervention to improve medication adherence in SLE. A 2-step intervention was used, in which rheumatologists (n = 6) first reviewed the pharmacy refill data available in the electronic health record (EHR) and then discussed the data with the patient participants to identify and address potential barriers to adherence. Patient-physician discussions during this clinic visit were audio-recorded for content analysis. The authors found that an average time of < 4 minutes was spent on the discussion of adherence; > 80% of these discussions involved positive reinforcement and validation, whereas approximately 50% involved open-ended questions. During postintervention interviews, patients and rheumatologists reported that they found the intervention to be useful and efficient. Several rheumatologists noted that the refill data helped them discuss barriers during the visit and that the discussion helped build trust and rapport. Finally, refill data indicated that, post intervention, 44% of patients had a major improvement (≥ 20% increase in PDC) in their adherence, with 40% considered adherent (PDC ≥ 80%).11

In their study, Sun et al11 used objective pharmacy refill data, a previously time- and labor-intensive source of data often involving a pharmacist or care team member gathering data from individual pharmacies. However, with recent advances in interoperability, pharmacy refill data are readily available in many EHRs. Sun et al11 leveraged this functionality in their EHR to build their intervention. This is a major strength of the intervention, given that medication adherence may be optimistically estimated by physicians.12 This objective measure likely enhanced rheumatologists’ awareness of nonadherence and helped them discuss barriers during a visit. Another strength of this intervention was its simplicity and brevity (time spent < 4 minutes), making it feasible to incorporate during study visits. To further explore end-users’ acceptability of the intervention, the authors performed semistructured interviews, in which patients and rheumatologists overwhelmingly supported the use of the intervention. More importantly, patients reported that dedicated discussions driven by rheumatologists to assess and address barriers were “nonjudgmental” and “extremely helpful,” despite adherence being a complex and sensitive topic to address. This clearly highlights that dedicated, open patient-provider discussions, supported by objective adherence data, could support patients in their decisions to continue or start medicines. Overall, this study delivers a relatively simple and potentially effective intervention that was feasible and acceptable to both the patients and rheumatologists participating in the study.11

However, it is likely that this intervention might require some revisions to be adapted in diverse clinics and populations. First, whereas pharmacy refill data are an incredibly helpful measure of adherence, these data are often available only as an add-on EHR functionality, which might incur costs. In clinics with limited resources or smaller practices, this may be an insurmountable barrier. As noted by Sun et al,11 refill data, although objective, are not perfect measures of adherence; patients may have automated refills and still skip doses, take the wrong dose, or not follow instructions in taking medications, which can reflect problems with health literacy, finances, and/or cognition (Figure 1). Thus, refill data are not always equivalent to medications taken. More objective measures, such as blood levels of medicines, could be used to assess adherence and start an open discussion to understand barriers to adherence.12 Therapeutic drug-level monitoring bypasses clinical variables, including dosing, absorption, and clearance, and could help balance safety and efficacy of medicines at the individual patient level and encourage use by alleviating fears of toxicity.12-15 However, these advantages should be balanced with the potential additional burden on the patient (ie, blood draws) and potential erosion of trust if patients experience such monitoring as punitive rather than informative. Second, Sun et al11 found that only half of encounters involved open-ended questions and that barriers were not addressed in about a third of visits, suggesting the need for more time for these conversations and for generic prompts for the clinicians to start an open, nonjudgmental conversation, at least for some patient visits. Third, the intervention would have to be adapted to address individual, potentially interrelated barriers, which may vary by geography and practice type. Such barriers include, but are not limited to, cognitive issues, which can cause patients to forget medications; health literacy issues, which can cause patients not to understand their condition or how medications can improve it; poor mental and physical health, which can compete with patients’ needs for SLE treatment; financial constraints, which can make it difficult for patients to afford medications; system factors, which can make it difficult for patients to obtain their medications or result in confusion about which medications to take; and trust and belief, which can conflict with the clinician’s treatment plans or make it difficult to build rapport and trust (Figure 1).16-20

The need for patient and rheumatologist support to facilitate these interventions is clear. The results from Sun et al11 and other studies could inform the development and implementation of shared decision-making tools,10,19-21 which could help patients understand the benefits vs harms of specific SLE medicines and support patients in their decisions to continue or stop medications. Moreover, by using such tools, rheumatologists and patients could prioritize individualized treatment goals, which could improve adherence and, ultimately, outcomes in SLE.10,19-21 Interventions could be additionally facilitated by other healthcare team members, such as nurses, medical assistants, pharmacists, or social workers (when available), alleviating the burden on the rheumatologist by guiding patients toward resources that address underlying barriers, such as financial need or lack of transportation.22 Additional EHR support such as dashboards that create visual trends of both adherence and outcomes like disease activity and physical and mental functioning over time, could help patients understand how their trajectory changes over time with better adherence and, thus, encourage patients to continue SLE medicines. Moreover, EHR vendors could make accurate refill data available without additional costs or at reduced costs to support wider adoption across diverse clinics.

Finally, some of the above problems are further compounded in patients with language barriers or limited health literacy (Figure 1). For example, Spanish-speaking US patients often refer to healthcare visits as “batalla” (a battle) and face a higher risk of nonadherence, driving disproportionately poor outcomes.23 Thus, developing tools and interventions in different languages that could be used by patients with limited health literacy and proficiency in the local language to feel supported in their decisions to continue or stop medications.20 Moreover, with the advances in the role of artificial intelligence in health care, it is possible to leverage large language models (LLMs) in adherence initiatives. For example, LLMs could generate prompts and strategies that staff could use to discover and address common barriers to adherence, such as queueing 90-day prescriptions, using mail-order pharmacies, and translating medicine instructions.24

In conclusion, improving medication adherence in SLE is crucial for improving health outcomes. It requires a comprehensive, multistep approach that includes the assessment of adherence by one or more methods; dedicated discussions of adherence and barriers, led by one or more clinicians; shared decision making about medications; and longitudinal follow-up of adherence and barriers (Figure 2). This approach addresses often complex barriers to adherence and emphasizes the importance of patient-physician communication and relationships. Sun et al11 took on the nearly herculean task of improving medication adherence in SLE, and, as a result, they provide a promising model for a simple, feasible, and effective intervention. It is time to evaluate the long-term effects of such multifaceted medication adherence interventions across diverse populations and clinics in SLE.

Proposed comprehensive, multistep approach to improving medication adherence in SLE. APPs include physician assistants and NPs. APP: advanced practice provider; B/w: between; MA: medical assistant; MD: doctor of medicine; NP: nurse practitioner; PharmD: pharmacist; RN: registered nurse; SW: social worker.

In the past 3 years, SG has had unrelated grant funding from the National Institutes of Health National Center for Advancing Translational Sciences through a Clinical and Translational Science Award, the University of Wisconsin-Madison, Institute for Clinical and Translational Research (UW ICTR), the Rheumatology Research Foundation, and the Lupus Foundation of America.

The authors declare no conflicts of interest relevant to this article.

See Clinician-led SLE adherence intervention, page 884