Leveraging Social Risk Factor Screening Infrastructure in Rheumatology Clinics to Respond to Rising Rates of Tuberculosis Among Individuals Experiencing Homelessness Authors: Rebecca Summit, Dylan B. Tierney, Taussia Boadi, Tyler Green, Karli Retzel, Virginia Bills, Nancy A. Shadick, Leah Santacroce, Sara R. Schoenfeld, Candace H. Feldman Published: 2026 ABSTRACT ObjectiveImmunocompromised individuals with rheumatic conditions have heightened risk of tuberculosis (TB) infection, reactivation, and severe disease. Patients with social risks are especially vulnerable. After an alert of increased TB rates among individuals experiencing homelessness in Massachusetts, we leveraged the existing social risk screening infrastructure to identify those at highest risk.MethodsWe identified patients with rheumatic conditions receiving immunosuppressive or immunomodulatory medications and recent housing instability from the electronic health records (EHRs) of a multihospital medical system. We notified these patients’ rheumatologists and provided latent TB and TB symptom screening information. We tracked EHR documentation, orders, and laboratory results following this alert.ResultsThirty‐three patients with rheumatic conditions experienced housing instability. Of these, 23 (70%) patients cared for by 17 rheumatologists were prescribed ≥1 immunosuppressive or immunomodulatory medications. Following the alert, three (13%) patients had EHR documentation of symptom screening, one provider responded that their patient was not homeless, and one ordered an interferon‐gamma release assay (“T‐SPOT”). After the reminder, seven (30%) T‐SPOT tests were ordered, five (22%) patients had EHR documentation of symptom review at their subsequent rheumatology appointment, and four (18%) patients were messaged by their rheumatologists about their TB risk.ConclusionThe existing rheumatology‐based social risk factor screening infrastructure facilitated a rapid response to a relevant public health concern. Rheumatologists’ responses varied, possibly due to increased clinical stressors with limited support. Despite challenges to reaching patients with housing instability, many actions were taken. Efforts to prompt regular TB screening of patients with these vulnerabilities are underway. INTRODUCTION Social determinants of health (SDoH) are the nonmedical factors, such as where a person lives and works, and their access to resources that affect health. 1 Social risk factors (SRFs) such as poverty, housing instability, and living in disadvantaged neighborhoods have been previously shown to influence inequities in care access and adverse outcomes among individuals with rheumatic conditions. 2 These SRFs are infrequently assessed in the subspecialty setting in part because resources (eg, social work support) to address needs that may be uncovered are rarely available outside of primary care. Starting in June 2022, nine rheumatology clinics across Mass General Brigham (MGB), a multihospital health care system, began to screen for SRFs, provide resource sheets, and connect patients who request assistance with their social needs to community‐based resources. 3 SIGNIFICANCE & INNOVATIONS Social risk factor screening infrastructure is rare in the subspecialty setting. This study demonstrates the unique ability to leverage an existing rheumatology‐based screening program to rapidly address a public health concern of rising tuberculosis (TB) rates among individuals with rheumatic conditions experiencing homelessness.There were challenges reaching patients who would benefit from the screening and there were mixed responses from rheumatologists. Other methods of implementing TB screening should be considered, such as involvement of care coordinators or nursing teams.This intervention demonstrates the direct relevance of social risk factor screening and electronic health record documentation of these risk factors to patients with rheumatic conditions. The presence of this infrastructure facilitated the ability to rapidly identify and address care gaps in the setting of an urgent public health concern. SIGNIFICANCE & INNOVATIONS Social risk factor screening infrastructure is rare in the subspecialty setting. This study demonstrates the unique ability to leverage an existing rheumatology‐based screening program to rapidly address a public health concern of rising tuberculosis (TB) rates among individuals with rheumatic conditions experiencing homelessness.There were challenges reaching patients who would benefit from the screening and there were mixed responses from rheumatologists. Other methods of implementing TB screening should be considered, such as involvement of care coordinators or nursing teams.This intervention demonstrates the direct relevance of social risk factor screening and electronic health record documentation of these risk factors to patients with rheumatic conditions. The presence of this infrastructure facilitated the ability to rapidly identify and address care gaps in the setting of an urgent public health concern. Social risk factor screening infrastructure is rare in the subspecialty setting. This study demonstrates the unique ability to leverage an existing rheumatology‐based screening program to rapidly address a public health concern of rising tuberculosis (TB) rates among individuals with rheumatic conditions experiencing homelessness. Social risk factor screening infrastructure is rare in the subspecialty setting. This study demonstrates the unique ability to leverage an existing rheumatology‐based screening program to rapidly address a public health concern of rising tuberculosis (TB) rates among individuals with rheumatic conditions experiencing homelessness. There were challenges reaching patients who would benefit from the screening and there were mixed responses from rheumatologists. Other methods of implementing TB screening should be considered, such as involvement of care coordinators or nursing teams. There were challenges reaching patients who would benefit from the screening and there were mixed responses from rheumatologists. Other methods of implementing TB screening should be considered, such as involvement of care coordinators or nursing teams. This intervention demonstrates the direct relevance of social risk factor screening and electronic health record documentation of these risk factors to patients with rheumatic conditions. The presence of this infrastructure facilitated the ability to rapidly identify and address care gaps in the setting of an urgent public health concern. This intervention demonstrates the direct relevance of social risk factor screening and electronic health record documentation of these risk factors to patients with rheumatic conditions. The presence of this infrastructure facilitated the ability to rapidly identify and address care gaps in the setting of an urgent public health concern. In September 2024, the Massachusetts Department of Public Health (MDPH) and the Boston Public Health Commission (BPHC) issued an alert of an increase in tuberculosis (TB) among persons experiencing homelessness in the greater Boston area. Specifically, spread of infection was seen in homeless shelters. Patients with autoimmune rheumatic diseases, such as rheumatoid arthritis (RA), psoriatic arthritis (PsA), and axial spondyloarthritis (axSpA) have higher TB infection risk compared to the general population 4 and are at increased risk for poorer outcomes. Disease‐modifying antirheumatic drugs (DMARDs) are used to treat rheumatic diseases and hold varying risks for reactivation of latent TB infection (LTBI). 4 Glucocorticoids and methotrexate carry modestly increased risk of TB reactivation as well. 5 Anti–tumor necrosis factor (anti‐TNF)–targeted biologics are highly effective agents used to treat RA, PsA, and axSpA. 4 However, they also increase the risk of reactivation of LTBI by four‐ to eight‐fold. 5 Between the intrinsic immune impairment from RA, PsA, and axSpA, and the regular treatment of patients with these conditions with anti‐TNF–targeted biologics, there is heightened vulnerability for reactivation of LTBI to active TB. Patients taking biologic DMARDs other than anti‐TNF–targeted biologics tend to have a lower risk of reactivation but still have a higher risk than the general population. 5 According to 2022 EULAR recommendations, patients taking biologic DMARDs and targeted synthetic DMARDs should be tested for TB before treatment, and TB testing should be considered for patients prescribed conventional synthetic DMARDs, glucocorticoids, or other immunosuppressants. 6 Homelessness is also a key risk factor for TB within high‐income countries, 7 but rheumatology‐related guidelines for regular LTBI screening in this population are less clear. We aimed to leverage the newly developed infrastructure to screen for and help address SRFs in rheumatology outpatient clinics to develop and implement a potentially scalable, reproducible process to respond in real time to this public health concern of TB spread among high‐risk individuals with rheumatic conditions experiencing homelessness. We assessed rheumatologists’ responses to notifications, uptake of screening recommendations, and feedback about this process. MATERIALS AND METHODS SRF screening Starting in June 2022, nine MGB hospital rheumatology clinics implemented SRF screening. 3 Per the MGB Institutional Review Board, SRF screening and this TB response initiative were determined to be Not Human Subjects Research on March 1, 2022 and on September 20, 2024, respectively. The SRF questionnaire, which included items related to housing instability, food insecurity, difficulty with transportation, medication access, jobs, and education (Supplemental Figure 1), was developed initially for use in primary care clinics and translated into seven languages. Patient responses are automatically integrated into the electronic health record (EHR). These questionnaires are assigned electronically as part of previsit materials to patients with one or more International Classification of Diseases, Tenth Revision (ICD‐10) codes for systemic rheumatic conditions with a focus on inflammatory arthritis (Supplemental Table 1). Patients who indicated SRFs and requested additional information were sent resource sheets and received outreach from a rheumatology‐based community resource specialist (rCRS). Patient population We used the MGB Research Patient Data Repository to identify individuals age ≥18 with ≥1 ICD‐10 code for a systemic rheumatic condition (Supplemental Table 1) and ≥1 visit with an MGB‐affiliated rheumatologist who completed a SRF screening questionnaire and indicated housing instability. We limited our search to individuals with the greatest level of housing instability, defined as a response of “I do not have housing (staying in a hotel, in a shelter, living outside on the street, on a beach, in a car, or in a park)” to the question “What is your housing situation today?”. We reviewed the medical charts of these patients to identify those who were currently receiving a DMARD or other immunosuppressive medication that increased their risk of TB, or hydroxychloroquine, because it may be a precursor to a future DMARD or immunosuppressive medication prescription (Table 1). We also documented the patient's most recent interferon‐gamma release assay (“T‐SPOT”), history of latent or active TB, and documentation of any previous complete or incomplete LTBI or active TB treatments. Table 1Description of patients identified at high risk for TB exposure receiving immunosuppressive or immunomodulatory medications (n = 23)* ValueSex, n (%)Male6 (26)Female17 (74)Race, n (%)White14 (61)Unknown/Not Reported9 (39)Ethnicity, n (%)Hispanic11 (48)Not Hispanic12 (52)Age, mean (SD)56 (16)Medication, n (%)Abatacept2 (9)Anifrolumab1 (4)Anti‐TNFs (etanercept, adalimumab, certolizumab, golimumab, and infliximab)6 (26)Azathioprine1 (4)Hydroxychloroquine7 (30)IL‐6 inhibitors (tocilizumab and sarilumab)1 (4)Anti‐IL‐17s (secukinumab and ixekizumab)3 (13)Leflunomide1 (4)Methotrexate3 (13)Mycophenolate mofetil3 (13)Prednisone or methylprednisolone6 (26)Tacrolimus2 (9)*IL, interleukin; TB, tuberculosis; TNF, tumor necrosis factor. Description of patients identified at high risk for TB exposure receiving immunosuppressive or immunomodulatory medications (n = 23)* IL, interleukin; TB, tuberculosis; TNF, tumor necrosis factor. Protocol development, EHR review, and provider notifications We conducted a series of meetings first with an infectious disease specialist with clinical and public health expertise in this domain (DBT), and with members of Boston Health Care for the Homeless Program, a multidisciplinary organization with >30 clinical sites in homeless shelters and hospitals. 8 Our team, which included three rheumatologists, two infectious disease specialists, a nurse, research assistants, and the rCRS, developed recommendations for rheumatologists and referral pathways for patients with possible TB exposures and symptoms, and for patients who had not had recent LTBI screening (without documentation of previous treatment). We alerted the rheumatologist via secure email of their patients who were identified as having both the highest level of housing instability and current use of a medication that increased their risk of TB infection or LTBI reactivation using a formatted letter (Supplemental Figure 1). The templated message included the patient's identification information, current rheumatic condition, medications, date of last T‐SPOT, and history of LTBI and treatment, if applicable. We also included information about screening for active TB symptoms and for LTBI. If a patient had a rheumatology appointment scheduled within a month of the provider alert, the patient's EHR was reviewed for documentation of rheumatologist action one week following their appointment date. If they did not have a scheduled appointment within one month, the EHR was reviewed two weeks after the first provider alert was sent. On the review date, we checked for documentation of a T‐SPOT order and/or result, symptom review, chest x‐ray, or confirmation that the patient was no longer experiencing homelessness. If no action was taken by the review date, a reminder email was sent to the rheumatologist (Figure 1). A second review for provider action following the second alert (if applicable) was completed in January 2025, and a final review was completed in April 2025. Workflow for alerting rheumatologists of their patients’ increased risk for tuberculosis (TB). It demonstrates provider actions and responses based on the timing of their patients’ appointments. EHR, electronic health record; ID, Infectious Diseases Clinic; T‐SPOT, interferon‐gamma release assay. As a part of our ongoing SRF screening program in rheumatology, all rheumatologists at the MGB rheumatology clinics were sent an annual evaluation survey. We included two questions regarding the TB screening. One question asked the providers how helpful they found this information and the second asked providers what action they took, if any, following the alert. Analyses We assessed the number and percentage of patients who completed additional TB screening procedures based on the recommendations provided, as well as the results of the procedures completed. We examined the distribution of provider responses to the surveys administered. RESULTS Since June 2022, 33 patients receiving rheumatology care at MGB indicated housing instability or homelessness on their most recent SRF questionnaire. Of these patients, 23 had prescriptions for ≥1 medications that met our response criteria (Table 1). The most common rheumatic disease diagnoses were RA (48%), systemic lupus erythematosus (26%), PsA (13%), and inflammatory arthritis (including axSpA) (9%). Of the 23 patients who indicated housing stability or homelessness and were prescribed an immunosuppressive medication, 4 (17%) had documentation of previous TB treatment. Nine (39%) patients had a T‐SPOT within the previous year. We alerted 17 providers of their patients’ TB risk via secure email (Supplemental Figure 2). Of the 23 patients at risk, 7 (30%) had existing appointments within the subsequent month and were reviewed one week following the appointment date, and 16 (70%) patients did not have a rheumatology appointment in the next month and were reviewed two weeks following the initial alert. EHR review demonstrated that five (22%) patients had documentation of physician action following the initial alert. Two (9%) patients were being treated with hydroxychloroquine monotherapy and were not sent a reminder. A reminder was sent to the providers of the 16 (70%) patients who did not have documented physician action. The actions of rheumatologists following the alerts varied. After the first alert, three (13%) patients had documentation of TB symptom screening, all of which were negative for active TB symptoms, one provider responded that their patient was not homeless, and one ordered a T‐SPOT which was negative. Rheumatologists for two (9%) patients prescribed hydroxychloroquine were notified in case there was a therapy escalation plan, but a reminder was not sent due to lower risk. After the second alert, more actions were taken, and multiple actions were taken for some patients. Seven (30%) T‐SPOT tests were ordered, five by rheumatologists and two by other providers. Five (71%) of the T‐SPOT tests were completed, and one (20%) was positive. One patient self‐reported a negative T‐SPOT. One patient was confirmed to have housing by their rheumatologist. Five (22%) patients had EHR documentation of symptom review at their subsequent rheumatology appointment, all of which were negative. Four (18%) patients were messaged by their rheumatologists about their TB risk. Four (18%) patients were messaged on their patient portal by their rheumatologists about their TB risk. Three (75%) patients viewed this message, but none responded to this outreach. No action was taken by rheumatologists for four (18%) patients. Some actions were taken by providers other than rheumatologists, which may or may not have been due to the TB screening alerts sent. Four (18%) patients had symptoms reviewed and seven (32%) patients had chest imaging ordered by non‐rheumatology providers. No action was taken by any provider following the alerts for two (9%) of patients. One patient was later found to be deceased at the time of the TB alert, unbeknownst to their rheumatology or study staff. Eighty‐four rheumatologists were sent a survey to provide feedback on the SRF screening and the TB response program. Forty providers (48%) responded to the administered survey. Eighteen (45%) indicated that they did receive the TB alert, even though we only sent it to 17 providers. This form was anonymous, so we cannot identify the provider who erroneously indicated that they received the TB alert. When asked how helpful they found the information, nine (50%) providers responded “neutral,” four (22%) providers responded “somewhat helpful,” and five (28%) providers responded “very helpful.” No providers indicated that the alert was not helpful. Sixteen providers (89%) indicated their action(s) following the alert. Three (19%) providers said they reached out to the patient regarding symptoms, five (31%) providers responded that they ordered a T‐SPOT, and eight (50%) providers responded that they did none of the above. Two providers indicated “other” and said that they reached out to schedule an overdue appointment with the patient, and that they contacted a different provider caring for their patient to order the appropriate test. DISCUSSION SDoH have been shown to have a greater influence on overall health than either genetic factors 9 or direct clinical care received, 10 and there is increasing evidence that addressing individuals’ social needs is associated with improved health outcomes. 11 Screening and addressing SRFs is limited, and when hospitals do screen, it is predominantly done in the primary care setting. 12 There is limited precedent for subspecialty‐based SRF assessments, often due to lack of dedicated resources to address the needs uncovered. 3 , 13 , 14 The impetus for the SRF screening program at our multihospital institution was to assist patients with needs that were directly relevant to their ability to adhere to medical appointments and medications, and to manage their rheumatic conditions and overall health more broadly. The alert from the MDPH and BPHC prompted a test of our infrastructure to see whether it could be leveraged in real time to address a pressing public health concern, with the thought that this process could be scalable, relevant, and reproducible in the future. Because of the ongoing SRF data collected for over 12,000 rheumatology patients at the time of the alert, we were able to quickly respond by alerting the rheumatologists of the patients we identified with the highest risk of TB. This adds further impetus for SRF screening and its relevance in the public health context. This TB screening and response initiative identified only a small proportion of patients receiving rheumatology care within our multihospital medical system who were unhoused and receiving an immunosuppressive medication. Since starting the SRF screening program in June 2022, we have screened >15,000 patients and identified ~18% with ≥1 SRF. Although only a small number of patients were identified for this proof‐of‐concept intervention, this is a highly vulnerable group and clear TB screening gaps were identified. We anticipate that safety net hospitals that serve larger unhoused populations would derive even greater benefit from the SRF screening infrastructure that can be leveraged for urgent public health concerns such as rising community TB rates. It is important to note that rheumatologists’ responses to the TB alert varied. Multiple rheumatologists did not initially respond to the TB risk alert, demonstrating the need for additional personnel, such as care coordinators, social workers, or nurses, to follow up with providers and to assist with patient outreach (eg, using different modes of communication, offering transportation, and arranging for local laboratory draws at more convenient locations). Further discussion with rheumatologists is needed to determine what information is most valuable to their clinical practice, how it should be conveyed to facilitate action, and what ancillary support would be most helpful for their highest risk patients. Direct patient engagement to better delineate needs and barriers to clinical interventions like this TB screening initiative is also essential. Our nine‐clinic rheumatology‐based SRF screening infrastructure has evolved since inception in 2022. One challenge has been determining who is responsible for certain aspects of care that may be uncovered. For example, should regular TB screening among individuals experiencing homelessness fall within the domain of primary care or rheumatology? Here, we felt it appropriate to direct this to rheumatology due to the concomitant use of immunosuppressives, which increased risk. More resources for responding to patient needs, such as a community resource specialist and social work support, exist within the primary care setting and are less frequently available in the subspecialty setting. However, rheumatologists often see their patients more frequently than primary care providers and may be more likely to uncover needs. Because of these circumstances, it has been difficult to know where the responsibility to address social needs lies. With our SRF screening infrastructure and our study team in place, we were able to quickly screen and intervene following the public health alert even though it was not part of our original goals, and the sustainability of this infrastructure is essential to consider. Moving forward, we are continuing to screen for heightened TB risk among individuals with housing instability. Individuals experiencing homelessness have a greater risk of TB infection and reactivation, 14 and we aim to increase rheumatologists’ awareness both of their patients’ social needs and of these risks that are directly relevant to the care provided. Although TB screening is recommended before initiation of DMARDs and glucocorticoids, there are no specific recommendations for regular screening unless there are other additional risk factors. 15 Although homelessness is not specifically indicated as a risk factor by EULAR, 6 the World Health Organization does consider this a risk factor, 7 and we plan to encourage yearly TB screening for our highest risk patients with our existing infrastructure that facilitates their identification. The process of quickly and easily identifying the highest risk patients was only possible because of the preexisting SRF screening infrastructure with EHR integration. This process demonstrates the importance of investing in this resource‐intensive infrastructure so that it can be leveraged to respond to a public health concern. The process of identifying patients with heightened TB risk was not automated, though, and the gaps we identified in TB screening for this patient population with housing insecurity and potential TB exposures raise the question of whether a more systematic process should be integrated into routine rheumatology care going forward. This SRF screening infrastructure, and the interdisciplinary involvement, documents, and system developed to respond to this public health need can readily be modified to mobilize in the setting of a future public health concern even more rapidly. In addition, it can serve as a template for other institutions that have begun to similarly screen and address SRFs among individuals with rheumatic conditions. AUTHOR CONTRIBUTIONS All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Dr Feldman confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements. ROLE OF THE STUDY SPONSOR The Bristol Myers Squibb Foundation had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by the Bristol Myers Squibb Foundation. Supporting information Disclosure form. Supplementary Figure 1: Supplemental Figure 1 is the social determinants of health questionnaire that was used to identify unhoused patients. Supplementary Figure 2: Supplemental Figure 2 is the letter template that rheumatologists received to alert them of their patients’ increased risk of tuberculosis. It provides rheumatologists with their patient's drug and tuberculosis information and suggestions for tuberculosis screening. Supplementary Table 1: ICD‐10 Codes For SDoH Questionnaire Distribution : Rheumatologic conditions and corresponding ICD‐10 codes used to identify patients that were sent the social determinants of health questionnaire.