“The medications are the decision-makers…” Making reproductive and medication use decisions among female patients with rheumatoid arthritis: a constructivist grounded theory

Objective To examine how female patients with RA form decisions about having children, pregnancy, and medication use. Methods We employed a constructivist grounded theory design and recruited female participants who are 18 years or older, have a rheumatologist-confirmed RA diagnosis, live in Canada, and are able to communicate in English or French. We collected data through semi-structured individual and focus group interviews using telephone or video conferencing technology. Data collection and analysis were iterative, employed theoretical sampling, reflexive journaling, and peer debriefing, and culminated in a theoretical model. Results We recruited 21 participants with a mean age of 34 years and median 10 years since RA diagnosis. Overall, 33% had never been pregnant, 57% had previously been pregnant, and 10% were pregnant at the time of interview. Of those who had experienced pregnancy, 64% had at least one pregnancy while diagnosed with RA and of those, 56% used DMARD(s) during a pregnancy. We constructed a patient-centred framework depicting the dynamic relationships between 4 decision-making processes—(1) using medications, (2) having children, (3) planning pregnancy, and (4) parenting—and the substantial impact of healthcare providers on patients’ experiences making these decisions. These processes were further influenced by participants’ intersecting identities and contextual factors, particularly attitudes towards health and medications, disease onset and severity, familial support system, and experiences interacting with the healthcare system. Conclusion Our framework provides insight into how patients make reproductive decisions in the context of managing RA and the opportunities for providers to support them at each decision-making process. A patient-centred care approach is suggested to support female patients with RA in making reproductive and medication choices aligning with their individual desires, needs, and values.


Introduction
Managing rheumatoid arthritis (RA) during pregnancy is a therapeutic challenge. Recent evidence suggests a decrease in RA activity in 60% of pregnant patients [1,2] with only 20-40% attaining remission by the third trimester of pregnancy [3,4]. Consequently, the majority of female patients with RA require some form of medication treatment during the perinatal period [5]. Nevertheless, several studies show low utilisation and considerable discontinuation of pregnancy compatible medications for RA perinatally [6][7][8][9]. Moreover, female patients with RA have information needs about medications in pregnancy [10][11][12][13], despite recent evidence-based guidelines supporting the safety of some disease-modifying anti-rheumatic drugs (DMARDs) in pregnancy [14][15][16][17].
Female patients with RA have fewer biological children than desired and as compared to other female individuals [18][19][20]. Reasons for smaller family size may include fertility issues, disease-and medicationrelated decrease in sexual activity, concerns about caring for a child (e.g., due to disease-related physical and functional limitations), uncertainty around medication use (e.g., stopping medications, adverse pregnancy and foetal outcomes), and fears about RA hereditability [19][20][21][22][23][24][25]. Despite recognition that females with RA are not meeting their reproductive goals, there are no studies examining the process of reproductive decisionmaking; specifically, the trade-offs related to managing disease activity and potential pregnancy. We aimed to develop a constructivist grounded theory of patients' reproductive decision-making within the context of living with RA.

Design
Informing this 'MOTHERS' study is a Feminist Intersectional framework for understanding the relational and cumulative nature of independent systems of privilege and oppression that shape individual health and lead to health disparities [26][27][28]. Our approach was guided by constructivist grounded theory, which aims to generate theory to explain human phenomena and acknowledges the researcher and participants as coconstructers of its meaning [29][30][31][32][33]. Our analysis used both inductive and deductive approaches to conceptualise decision-making processes. Our patient research partner (LP) was involved throughout the research process, including grant submission, interview guide development, recruitment, data interpretation, and knowledge translation. This study was approved by the Behavioural Research Ethics Board at the University of British Columbia.

Participants
We recruited participants using posters in rheumatology clinics across Canada and social media posts through investigators' and patient organisations' channels. Female participants were eligible if they were 18 years or older, had a rheumatologist-confirmed RA diagnosis, lived in Canada, and were able to communicate in English or French. Interested participants were provided a study URL to a questionnaire administered by the online survey platform Qualtrics through which they provided written informed consent electronically and responded to demographic information to aid in scheduling interview sessions and adapt the interview guide to their specific pregnancy and rheumatic disease experiences. We purposively sampled participants [34] to ensure diversity with respect to pregnancy intentions and experiences, disease and medication taking history, and province of residence. Moreover, we specifically sought participants who had never been pregnant, who had been pregnant and used medication, and who had been pregnant and not used medication.

Data collection
We collected data through semi-structured synchronous video or telephone one-on-one interviews (~1 h) and online focus groups (~2.5 h) using the iTracks platform. We conducted focus groups with participants with similar pregnancy experiences (e.g., never been pregnant, previously pregnant and used RA medications) when possible. One bilingual author (NR) conducted the English interviews and was present to guide French interviews conducted by a Research Coordinator fluent in French. Interviews were digitally recorded, professionally transcribed, and translated to English where relevant. Data collection and analysis were simultaneous, with analysis informing subsequent data collection, and iterative, continuing until we achieved saturation of themes.

Analysis
Our coding procedures included steps of line-by-line, focused, and theoretical coding. We used line-by-line coding to organise data into concepts and key phrases followed by focused coding to identify and group in vivo codes into categories [33]. Then, we used theoretical coding to identify connections and relationships between categories to provide insight into possible theories [33]. We employed constant comparison of data within and between transcripts to elevate the analysis to a conceptual level [33,35]. Data analysis was conducted by the first author (NR) using NVivo 12. To support this process, NR created memos and conceptual diagrams to practice critical reflexivity, facilitate theoretical integration, and develop the resultant patient-centred grounded theory framework. To mitigate the impact of being an outsider without personal experience of living with a rheumatic disease or making healthcare decisions while considering the effects of those choices on a pregnancy, NR regularly consulted and debriefed with a patient partner, collaborator, and co-author (LP), an insider, throughout the research process.

Results
We recruited 21 participants with a mean age of 34 years and median 10 years since diagnosis (Table 1). Most (86%) were married or co-habited with a romantic partner. All participants had a post-secondary education, most (67%) were employed full time, and the majority reported white (71%) or Asian (24%) ancestry.
Participants had diverse histories of RA, pregnancy, and medication use (see Table 2). Overall, 33% had never been pregnant, 57% had previously been pregnant, and 10% were pregnant during the interview. Of those who were pregnant or had previously experienced pregnancy, 64% had at least one pregnancy while diagnosed with RA and of those, 56% used DMARD(s), 33% used prednisone only, and 22% did not use medications during a pregnancy. Over half intended to have future children.

Decision-making processes
We identified three dynamic, closely connected, and interactive decision-making processes related to participants' reproductive decisions and experiences: (1) having children, (2) planning pregnancy, and (3) parenting. A

Targeted DMARDs
Anti-tumour necrosis factor (anti-TNF) agents 13 (62) Others (e.g. abatacept, rituximab, tocilizumab, tofacitinib) 7 (33) a Rural is defined as less than 400 people per square kilometre. Urban is define as more than 400 people per square kilometre b Cumulative percentage may be greater than 100, as multiple categories may be relevant to each participant c Missing response from one participant d Defined as any medication the participant has been exposed over the course of their RA treatment fourth process, using medications, was present throughout participants' reproductive lives, influencing the aforementioned decision-making processes. Contextual factors influencing these processes are presented in Fig. 1, with representative quotes presented in Table 3.

Having children
Participants described 4 decisional aspects influencing their decisions to prevent unplanned pregnancy and to choose whether to have children.
First, participants considered their individual desires and beliefs about having children, which were influenced by their age, expectations of motherhood, and mindset towards their childrearing options. Some 'always knew [they] wanted kids' while others felt less certain about their reproductive desires. Attitudes towards the environmental impact of biological children, perceived and experienced physical impacts of pregnancy, societal messaging about being a woman and having children, and size and composition of their childhood family influenced how participants evaluated their options (e.g., childbearing, adoption, and fostering). RA further complicated this consideration. Some participants contemplated how adoption would lessen the impact of disease management but remove the 'intimacy' of experiencing pregnancy. Some believed pregnancy was not an option    due to needing to prioritize managing their RA in medication decisions, having insufficient perinatal support from their healthcare team, or receiving discouraging messages about pregnancy from healthcare providers. Regardless of outcome, having the option to become pregnant was important.
Participants also considered the impact of their disease on having children, with their understanding dependent on the timing of disease onset and disease severity. For some, RA 'forced' them to consider their reproductive desires earlier than their peers, including whether having children was 'feasible' . While RA was not the Moreover, participants considered their avenues of support for growing a family. Rheumatologists were influential in participants' decisions to become pregnant and use medications perinatally. Participants with positive provider relationships felt supported, informed, comfortable discussing their reproductive options, and confident taking medications. Others struggled to have 'open conversations' about their reproductive goals with rheumatologists. Several participants believed female providers were more 'sensitive' to their reproductive care needs. Participants also considered their partner's age, health, desire to have children, and capacity for fulfilling additional caregiving responsibilities. Two participants without a partner at the time of their decision, acknowledged that this impacted their decisions to delay pregnancy and not have a child. Overwhelmingly, participants' partners actively supported them through learning about RA, engaging in discussions about reproductive planning, and listening and validating their worries. Additionally, participants considered the availability of childcare assistance from extended family.
Lastly, participants considered their capacity to have children within the context of their life, particularly their personal readiness and reproductive window for becoming pregnant. They envisioned achieving specific personal and professional milestones (e.g., completing their education, securing a permanent job, buying a house, being in a committed relationship) prior to having children. Additionally, the 'timing' of their diagnosis, disease severity, and access to parental leave and insurance coverage for pregnancy compatible medications affected their decisions. Most participants actively prevented unplanned pregnancy, which was a source of anxiety and fear.
Participants on teratogenic medications were vigilant in using 'consistent, reliable' contraceptives. Two participants considering permanent contraception options (i.e., tubal ligation, vasectomy) once they finished having children. Although some participants considered age a limiting factor for timing pregnancy, older participants and those with children wanted to '[keep] some agency' and not be discounted from having children.

Planning pregnancy
Participants who were planning pregnancy, were pregnant, or had previously experienced pregnancy described 4 decisional aspects influencing their conception and pregnancy decisions.
Participants considered their experiences trying to conceive, with their desired pregnancy timelines often at odds with timeframes for medication changes and conception with RA. Medications were the 'first thing' diagnosed participants considered when planning and timing pregnancy. For some, ensuring safe pregnancy removed the 'romance' and 'spontaneity' of trying to conceive with their partner. Often, conception required more time than expected, with some experiencing disease flares during this period. Additionally, some participants felt pressured to conceive quickly by their rheumatologist. Three participants were referred to a fertility specialist and one took medications to aid with conception.
Participants considered their experiences of medication use for disease management pre-conception and during pregnancy. Participants' perinatal medication choices are reported in Table 2. Prior to planned conception, participants followed medication recommendations from their rheumatologists, with some stopping DMARDs compatible with pregnancy and others starting biologic DMARDs (bDMARDs). Five participants used DMARDs during pregnancy. Some felt reassured their medications were safe while others worried about potential risks. Participant 12 experienced an unplanned pregnancy while using a teratogenic medication. Overall, most participants recognised the importance of controlling their disease activity for their own and their baby's health.
Participants who had experienced pregnancy considered their emotional and physical experiences of pregnancy, reporting mixed experiences with disease activity (see Table 2). All three participants who only used prednisone experienced flares during pregnancy, while two of the three participants who used bDMARDs maintained good disease control. Participant 12 experienced active disease while on a bDMARD. Of the two participants who did not take any medications, one went into remission during their first pregnancy but experienced active disease during their second pregnancy, while the other experienced a flare during their first pregnancy and decided to use DMARDs during their second pregnancy. Two participants diagnosed postnatally suspected they experienced untreated RA during pregnancy. All diagnosed participants who became pregnant reported feeling anxious throughout pregnancy planning about 'things that could wrong' related to the uncertainty of perinatal disease management. Many shared needing to 'let go of any expectations' and navigate 'one thing at a time' . Those using medications shared concerns about perinatal medication use, with some deciding to maintain a positive outlook and 'trust in the science' . When making medication decisions, many participants noted 'focusing on the baby', which at times came at a detriment to their own health and disease management.
Finally, participants considered their experiences interacting with healthcare providers. Those with planned pregnancies worked closely with rheumatologists to make medication and pregnancy decisions. Participants using medications were closely monitored by their healthcare team throughout the perinatal period. Although this was reassuring to most, for some it caused additional anxiety. Notably, two participants who stopped all medications preconception felt unsupported by their healthcare team as they were not closely monitored.

Parenting
Participants with an RA diagnosis during pregnancy or post-partum described 2 decisional aspects influencing their parenting decisions.
Participants considered their experiences of disease management post-partum. Six reported experiencing post-partum flares, which influenced their infant feeding decisions (i.e., whether to breastfeed and for how long). Of those who breastfed, some delayed starting medications incompatible with breastfeeding and two used compatible bDMARDs. Another two stopped breastfeeding to start new medications. Some participants described physical challenges of breastfeeding, particularly if RA affected their arms or wrists. Most participants considered breastfeeding an important motherhood experience and opportunity for 'connection' with their baby. Several shared the harm of encountering dogmatic messages such as 'breast is best' .
Relatedly, participants considered their emotional and physical experiences becoming a caregiver to a child, with many feeling 'unprepared' to navigate the impacts of RA on caring for a baby. While most felt relieved postpartum, some shared anxieties about their decisions, including questioning whether they could have 'done something different' and 'prevented' unwanted outcomes (e.g., premature delivery). Some worried about long-term impacts of perinatal medication use on their child, even if 'the research looks great' , and changing their medications for pregnancy putting their health on a 'different track five or ten years down the road' . Many participants experienced physical challenges taking care of newborns which required 'strategic planning' (e.g., related to infant dressing, picking up children, and navigating disease-related fatigue). Participants described prioritising being a 'present' parent by purposefully allocating their limited energy and identifying RA-friendly family activities. Overall, parenting with RA involved accepting that their children would have different experiences than their peers and recognising that growing up more independently taught their children resilience, empathy, and responsibility. All participants with children shared a caregiver role with a romantic partner and described how they divided parenting duties, including having their partners 'take charge' of more physically involved tasks. Some further received help from family and friends, including assistance with carrying their child, childcare, and financial support. Feeling limited in their capacity to perform tasks attributed to motherhood took an emotional toll on participants.

Using medications
Throughout their reproductive years, participants described 3 decisional aspects influencing their decisions to start, use, and stop DMARDs and analgesic medications.
Participants considered their attitudes and experiences using medications broadly. Although some were not accustomed to taking medications regularly prior to diagnosis, all came to realise they needed medications to function. This required time and a grieving process before reaching acceptance. Some depicted not having a choice if they wanted a 'livable' life. Several participants described struggling to find effective medications at some point in their treatment. For some, it was important to minimise medication exposure by taking 'as little medications as possible' and the 'mildest' medication, which was often related to concerns about long-term medication use. Participants discussed the importance of being presented with choices in their care, particularly with medications that affect their reproductive options.
Participants also considered their attitudes and experiences managing their health. Overall, participants perceived their RA management was medication focused. Although participants recognised the importance of medications, they wanted rheumatologists to support other aspects of disease management, including non-medication guidance and referrals for their physical, nutritional, mental, and sexual health. Those with children discussed how motherhood motivated them to '[pay] more attention to [their] body' , including recognising when they needed to prioritize their health to 'be a strong mamma'.
Finally, participants considered their medication options for pregnancy and breastfeeding. Some wanted to prioritize starting 'pregnancy safe medications first' to avoid a potential flare when stopping incompatible medications preconception. Most participants, particularly those who had tried several different therapies, felt their pregnancy options were limited. Although newer bDMARDs provided additional treatment options, some feared unknown long-term effects on an exposed child's health that may be identified in the future. Participants acknowledged that their perception of the safety of medication use was strongly influenced by their rheumatologist's beliefs and attitudes towards medications.

Patient-centred constructivist grounded theory
Our resultant theory (shown in Fig. 2) centres the female patient among contextual factors that form their decision-making environment as they navigate reproductive decision-making processes while living with RA.
The patient's reproductive window is set within the black box encompassing the figure. Encircling the patient Fig. 2 A constructivist, patient-centred, grounded theory of reproductive and medication use decision-making while living with RA are contextual factors related to being a woman (individual identity), a person with a chronic disease (patient identity), and a member of a family (familial identity) that influence their reproductive and medication use decision-making environment. These identities are fluid, overlapping, and influence one another. The emotional effect of interactions between these identities is depicted by the rose-coloured gradient, which is overwhelmingly dominated by fluctuating acute and chronic experiences of anxiety, fear, and guilt -emotional experiences resulting from the degree of confidence the patient experiences within the decision-making processes and their capacity to cognitively and emotionally address their psychological state. These identities and contextual factors are related to their reproductive journey (purple circle) preconception, during pregnancy, and post-partum and associated temporally with the reproductive decisionmaking processes: (1) having children, (2) planning pregnancy, and (3) parenting. The fourth decision-making process, (4) using medications, occurs in a dynamic context between an individual's identities, reproductive journey, and reproductive decision-making processes.

Discussion
Our study is the first to examine reproductive and medication decision-making among female patients with RA. Our patient-centred constructivist grounded theory demonstrates the fundamental role of medication use and disease management in patients' reproductive decision-making-'The medications are pretty much the decision maker' (Participant 2). Moreover, our results reveal that contextual factors-individual desires, fears, beliefs, values, and needs, disease and patient experiences, expectations and experiences of motherhood, and support from partners, family, and friends-influence how patients form decisions about having children, planning pregnancy, and using medication. Our theory's dynamic and multidimensional structure reflects the complex and relational nature of women's lives with implications for supporting a patient-centred approach to delivering reproductive care within rheumatology practice.
Our results have implications for improving perinatal standards of care for patients with RA. Despite research showing female patients with RA stop their medications during pregnancy [6][7][8][9] and have information needs regarding perinatal medication use [10][11][12][13], no studies have examined how they approach these decisions. Given the results from our study examining this phenomenon, we recommend that rheumatologists initiate discussions with patients about their current and future pregnancy prevention (e.g., contraception) and support (e.g., medication management, interdisciplinary referrals) needs early in the therapeutic relationship. Addressing reproductive topics pre-emptively and regularly ensures rheumatologists are supporting patients in preventing unplanned pregnancies on teratogenic medications and establishing a plan for managing RA perinatally. Building patient trust may be particularly important among male rheumatologists, with whom female patients may face additional barriers discussing reproductive care. We further recommend that rheumatologists continue to see patients regularly throughout pregnancy, including those who stop medications as they are more likely to experience pregnancy flares with an associated increased risk of untoward pregnancy outcomes. Post-delivery, rheumatologists are well situated to support breastfeeding decisions and provide referrals to RA-friendly resources for parenting. Finally, given the risks of unintended pregnancy post-partum [36], rheumatologists can support patients starting teratogenic medications or seeking to prevent subsequent pregnancy establish a contraception plan.
Our results reflect others [37][38][39] that show patients' perception of their provider relationship and healthcare experiences are key aspects of being activated in their care and that well-informed patients feel more confident and empowered in healthcare decision-making. Collaborations with rheumatologists and patient partners strengthened our data collection, analysis, and interpretation. Remote interviews facilitated recruiting a geographically diverse sample; however, recruitment was limited to individuals with access to telephone and video conferencing technology. Our data collection was further enriched through facilitating focus groups among participants with similar pregnancy histories. It is important to note that our sample comprised predominantly white and highly educated participants in heterosexual relationships. Further studies should examine the applicability of our theory to patients with historically marginalised racial and sexual identities as well as diverse family structures (e.g., single parent families). As systemic barriers to healthcare access contribute to significant gaps in care, higher disease activity, and poorer pregnancy outcomes among Indigenous, Black, and Hispanic patients with rheumatic diseases [40][41][42][43], our findings suggest an imminent need to address reproductive and medication decision-making needs among racialized patients.

Conclusions
Our study identified how gaps in care impact patients' reproductive decisions and perceived capacity to grow their family. By understanding the practical and emotional aspects of patients' decision-making processes, healthcare providers can identify opportunities for intervention and care adaptation leading to better reproductive choices that align with patients' individual desires, values, and needs.