Hydroxychloroquine (HCQ) is nearly universally recommended for pregnant patients with systemic lupus erythematosus (SLE) to reduce lupus disease activity and adverse outcomes in pregnancy.1-3 Yet despite strong evidence supporting its benefits, HCQ appears underutilized, with several studies suggesting fewer than half of all women with lupus take this medication during pregnancy.4
How accurately these results reflect patient adherence depends on how we measure adherence and what we assume about those not taking HCQ. Our recent work examining HCQ use during pregnancy seeks to address this issue. We used data on prescriptions written and filled, as well as qualitative clinical notes documenting patient treatment plans.5,6
Medication use requires that physicians place the order and patients fill and take the medication. Prior studies of medication use by women with SLE during pregnancy have examined reimbursement claims, pharmacy prescription fills or self-reported use in surveys or clinical notes.4,7,8 Interpreting low HCQ use measured by patient fill data alone assumes that patient nonadherence is the most significant contributing factor. However, physician prescribing practices provide critical context: Patients can’t get a medication without a physician’s prescription. Additionally, just because a prescription is filled does not guarantee the patient takes the medication as recommended, and a patient’s self-report may be inaccurate. Therefore, discussions of adherence and what actions are necessary to improve it should consider not only the patient, but also prescribing practices.
Primary Adherence
Our new study on HCQ use in pregnant patients with SLE published in ACR Open Rheumatology aimed to address primary adherence, or failure to fill a new prescription order.5 We wanted to find out whether adherence to physician orders reflected overall rates of HCQ use by patients as measured by prescription fills. The researchers examined electronic health record data to study prescription orders and fills for a cohort of pregnant patients with lupus. Patients with two or more ICD-coded visits for SLE at least seven days apart and one year prior to their last menstrual period with pregnancies ending in a delivery—still or live born—from 2011–20 were included (n=419).
The study focused primarily on new (physician or pharmacy-initiated) outpatient orders for HCQ. To improve capturing patient exposure to HCQ, the team also noted whether those without new orders had an existing prescription order from the pre-pregnancy period whose supply carried over into their pregnancy.
Frequency of prescription orders and the proportion filled—ever and within 30 days—was calculated. Due to possible changes in physician or patient adherence to guidelines over time, results were also examined separately for the periods 2011–2015, 2016–2019, and 2020. The year 2020 was considered separately due to COVID-19-related HCQ supply issues. For comparison, fills were also examined one year before the pregnancy (i.e., pre-pregnancy period).
For primary adherence, we found that patients were very likely to fill their new prescriptions for HCQ, often within 30 days of the order. Overall fill rates were 89% in the pre-pregnancy period and 93% during the pregnancy period. Consistent with other studies, we reported that 53–68% of pregnancies in women with lupus during the study period had HCQ available for filling (new orders and refills available), meaning that over 30% of pregnant women with lupus had no order on file to be filled.
Using prescription fill data without accounting for whether patients could fill a prescription in the first place may misrepresent the reason for low rates of HCQ use in pregnant women with lupus. The number of HCQ orders covering the pregnancy period determines the denominator of how many patients could have potentially used the medication.
At the same time, it is difficult to assume what the absence of an order means. Whether a provider did not recommend HCQ or a patient declined it during the visit cannot be inferred through claims or prescription data. Understanding physician adherence to guidelines requires qualitative data on the conversations that took place during these patients’ visits.
Shared Decision Making
To explore reasons for non-use of HCQ by women with lupus during pregnancy, we turned to physicians’ treatment notes. In prior interviews, patients with SLE reported concerns about toxicity and preferences to decrease or stop treatment, while still respecting physician judgment.9 Less is known about how pregnancy affects physician and patient attitudes toward HCQ use.
Qualitative studies of overall medication non-adherence among patients with lupus have suggested that hurried/poor communication with rheumatologists and patient perceptions of minimal benefit from treatment with HCQ may contribute to non-adherence.10-12 One qualitative study from 2018 found that rheumatologists cited patients’ preconceived notions/lack of trust in their recommendations, as well as gaps in other clinicians’ prescribing knowledge, as the primary barrier to medication adherence among women with lupus during pregnancy.13
Using a cohort of deliveries by women with lupus at Stanford Hospital, California, from 2008–16, we identified a case series of 38 pregnant women with lupus not taking HCQ at the first prenatal visit. We included prescription orders and physician notes related to SLE starting one year before the first documented prenatal visit through to the delivery date.
Provider specialty, prior history of lupus medication use, changes to lupus management during pregnancy and reasons for deferring HCQ were abstracted from notes through chart review. Reasons for deferral were separated into explicit and implicit themes, depending on whether physicians’ notes directly addressed HCQ use.
Surprisingly, we found patients and physicians cited a wide range of reasons for deferring use. Over a quarter of patients in our case series had self-discontinued HCQ for pregnancy-related reasons, most frequently citing concerns about ability to conceive on the medication. Several also documented having concerns about maternal or fetal side effects.
On the other hand, physicians frequently reported a patient’s clinical stability while not on HCQ as a reason to defer the medication. When HCQ was not explicitly mentioned in clinical notes, themes included patients being on alternate lupus medications, being off SLE treatment completely and barriers to seeing a rheumatologist.
Given that patients with SLE often see many different specialists, potentially at different institutions, our study was constrained by limited access to external notes and prescription data. It is possible that patients in the case series received HCQ during pregnancy elsewhere and this was not captured in our clinical documentation. Clinical notes are not a substitute for in-depth qualitative interviews specifically designed to elicit patient and physician perspectives on medication use or prescribing practices.
Going Forward
Improving adherence to clinical guidelines requires acknowledging that physicians treat people, not populations, and more personalized, patient-centered interventions may be needed. Patients may express reasonable concerns about medication side effects, may not be aware of risks and benefits of different medications, or may experience barriers to acquiring or taking medication regularly.14 Wherever appropriate, open, informed discussions on preconception counseling or about concerns with patients of childbearing age are important to develop the most appropriate treatment plan.
Our work highlights that how we measure adherence is extremely important to understanding which patients do not receive beneficial medication and why. An upstream issue, such as the absence of a medication order, implies different barriers and solutions than a downstream issue, such as patients not filling or taking a prescription. Qualitative information from such sources as physician notes can provide valuable nuance about physicians’ and patients’ beliefs around using HCQ.
Going forward, results from our research require replication using mixed methods in larger cohorts. We are currently leading a study that will interview patients and physicians to explore more detailed perspectives on their filling and prescribing practices during lupus pregnancy, with a primary focus on HCQ use. What is clear is that if patients are prescribed HCQ, they appear very likely to fill it at least once; our questions lie more in who does not receive or refill that prescription, and why.
Antonia Chan is a fourth-year medical student at Stanford University School of Medicine, Calif., and an incoming internal medicine resident physician at Beth Israel Deaconess Medical Center, Boston. She is interested in rheumatology/immunology, shared decision making and health services research.
Julia F. Simard, ScD, is an associate professor in the Department of Epidemiology and Population Health, Division of Immunology & Rheumatology in the Department of Medicine and, by courtesy, in the Department of Obstetrics and Gynecology at Stanford University School of Medicine. She is interested in misclassification bias, pregnancy outcomes in people with rheumatic disease,and epidemiologic methods.
References
- Sperber K, Hom C, Chao CP, et al. Systematic review of hydroxychloroquine use in pregnant patients with autoimmune diseases. Pediatr Rheumatol Online J. 2009 May 13;7:9.
- Clowse MEB, Magder L, Witter F, et al. Hydroxychloroquine in lupus pregnancy. Arthritis Rheum. 2006 Nov;54(11):3640–3647.
- Costedoat-Chalumeau N, Amoura Z, Duhaut P, et al. Safety of hydroxychloroquine in pregnant patients with connective tissue diseases: A study of one hundred thirty-three cases compared with a control group. Arthritis Rheum. 2003 Nov;48(11):3207–3211.
- Liu LH, Fevrier HB, Goldfien R, et al. Understanding nonadherence with hydroxychloroquine therapy in systemic lupus erythematosus. J Rheumatol. 2019 Oct;46(10):1309–1315.
- Simard JF, Liu EF, Chakravarty E, et al. Reconciling between medication orders and medication fills for lupus in pregnancy. ACR Open Rheumatol. 2022 Dec;4(12):1021–1026.
- Chan A, Hirz A, Chaichian Y, et al. Exploring reasons for non-use of hydroxychloroquine in SLE pregnancy. Arthritis Rheumatol. 2022; 74 (suppl 9).
- Jiang M, Chang Y, Wang Y, et al. Highrisk factors for adverse pregnancy outcomes in systemic lupus erythaematosus: A retrospective study of a Chinese population. BMJ Open. 2021 Nov 16;11(11):e049807.
- Bermas BL, Kim SC, Huybrechts K, et al. Trends in use of hydroxychloroquine during pregnancy in systemic lupus erythematosus patients from 2001 to 2015. Lupus. 2018 May;27(6):1012–1017.
- Dollinger J, Brasil C, Wong M, et al. Patient preferences for hydroxychloroquine in systemic lupus (SLE). Arthritis Rheumatol. 2021; 73 (suppl 9).
- Maheswaranathan M, McKenna K, Corneli A, et al. Patient perspective of helpfulness of lupus medications: A qualitative study of medication use within the Type 1 and 2 SLE model. Arthritis Rheumatol. 2020; 72 (suppl 10).
- Barr AC, Clowse M, Maheswaranathan M, et al. Association of hurried communication and low patient self-efficacy with persistent nonadherence to lupus medications. Arthritis Care Res. 2023 Jan;75(1):69–75.
- Leung J, Baker EA, Kim AHJ. Exploring intentional medication non-adherence in patients with systemic lupus erythematosus: The role of physician-patient interactions. Rheumatol Adv Pract. 2021 Jan 24; 5(1):rkaa078.
- Clowse MEB, Eudy AM, Revels J, et al. Provider perceptions on the management of lupus during pregnancy: barriers to improved care. Lupus. 2019 Jan;28(1):86–93.
- Sun K, Corneli AL, Dombeck C, et al. Barriers to taking medications for systemic lupus erythematosus: A qualitative study of racial minority patients, lupus providers, and clinic staff. Arthritis Care Res. 2022 Sep;74(9):1459–1467.