Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Dec 1.
Published in final edited form as: Arthritis Care Res (Hoboken). 2023 Aug 7;75(12):2529–2536. doi: 10.1002/acr.25174

Social Determinants of Health Documentation Among Individuals with Rheumatic and Musculoskeletal Conditions in an Integrated Care Management Program

Sciaska N Ulysse 1, Mia T Chandler 1,2, Leah Santacroce 1, Tianrun Cai 1, Katherine P Liao 1, Candace H Feldman 1
PMCID: PMC10725994  NIHMSID: NIHMS1913685  PMID: 37331999

Abstract

Objective

Social determinants of health (SDoH), such as poverty, are associated with increased burden and severity of rheumatic and musculoskeletal diseases. We studied the prevalence and documentation of SDoH needs in electronic health records (EHRs) of individuals with these conditions.

Methods

We randomly selected individuals with ≥1 ICD-9/10 code for a rheumatic/musculoskeletal condition enrolled in a multihospital integrated care management program that coordinates care for medically and/or psychosocially complex individuals. We assessed SDoH documentation using terms for financial needs, food insecurity, housing instability, transportation, and medication access from EHR note review and ICD-10 “Z” SDoH billing codes. We used multivariable logistic regression to examine associations between demographic factors (age, gender, race, ethnicity, insurance) and ≥1 (vs. 0) SDoH needs (Odds Ratio [OR] 95% CI).

Results

Among 558 individuals with rheumatic/musculoskeletal conditions, 249 (45%) had documentation in EHR notes by social workers, care coordinators, nurses, and physicians of ≥1 SDoH need. 171 individuals (31%) had financial insecurity, 105 (19%) transportation needs, 94 (17%) food insecurity; 5% had ≥1 Z billing code. In the multivariable model, the odds of having ≥1 SDoH need was 2.45 times higher (95% CI 1.17-5.11) for Black vs. White individuals and significantly higher for Medicaid or Medicare beneficiaries vs. Commercially insured individuals.

Conclusion

Nearly half of this sample of complex care management patients with rheumatic/musculoskeletal conditions had SDoH documented within EHR notes; financial insecurity was the most prevalent. Only 5% of patients had representative billing codes suggesting that systematic strategies to extract SDoH from notes are needed.

Introduction

Social determinants of health (SDoH), the conditions in which people work, live, and grow, contribute significantly to health behaviors and to inequities in healthcare access and outcomes.1,2 These nonmedical factors exist across medical specialties including rheumatology. In a national sample of patients with rheumatoid arthritis (RA), researchers demonstrated that faster declines in function over time and overall poorer functional status were observed in patients living in zip codes with more deprivation (measured using the Area Deprivation Index).3 In a study which identified adults with RA from the National Health and Nutrition Examination Survey (NHANES), >30% had food insecurity, which was also associated with higher odds of depression.4 A study of patients with systemic lupus erythematous (SLE) showed that moving out of poverty led to lower mean scores of newly accumulated disease damage, similar to scores of participants who were never in poverty.5 Across rheumatic conditions, living in areas of high heat or social vulnerability has been associated with higher odds of recurrent hospitalizations.6

To date, certain SDoH have been examined among individuals with rheumatic diseases through research studies and qualitative interviews. However, we do not have a clear understanding of how SDoH-related needs are documented in the electronic health records (EHR) of individuals with rheumatic diseases and the burden among individuals who may be at highest risk for adverse health outcomes related to these needs. We therefore aimed to systematically assess the documentation of SDoH-related needs in unstructured notes in the EHRs of patients with rheumatic conditions. We also aimed to determine whether structured billing International Classification of Diseases, Tenth Revision (ICD-10) Z codes, a set of standardized codes that are used to report the determinants that affect health-related outcomes7, were being utilized and the degree to which they overlapped with unstructured documentation of SDoH. Extraction of SDoH using structured billing codes would be a significantly easier way to understand population-level needs compared to manual chart review. However, we hypothesized that these Z codes would be underutilized and thus would underestimate the extent of SDoH-related needs. An understanding of both the burden of documented SDoH-related needs and the way in which they are documented will inform the development of algorithms that, if indicated, could combine natural language processing of unstructured notes with structured data (e.g., billing codes) to extract SDoH to guide clinical care and future research studies. In addition, by understanding the distribution and prevalence of SDoH-related needs by rheumatic condition, resources can be better allocated to help rheumatology clinics develop infrastructure to better meet the needs of their patients, in turn reducing disparities in access and outcomes.

Methods

Patient Population

In our multi-institution academic hospital system, Mass General Brigham (MGB), a subset of medically and psychosocially complex individuals who receive their primary care through MGB-affiliated hospitals are enrolled in an integrated care management program (iCMP). 8 The iCMP includes a multidisciplinary team of nurses, social workers, community health workers, community resource specialists, and pharmacists and aims to coordinate and improve care, and reduce costs. Individuals are identified for iCMP enrollment either through referral by their primary care physician or by a claims-based algorithm,9 which includes combinations of healthcare utilization patterns (e.g., recurrent emergency department visits), presence of complex and/or multiple medical issues, and/or a history of psychosocial needs (e.g., mental health diagnoses). During the time frame of this study, the algorithm did not include granular SDoH-related needs. A qualitative study of 20 providers demonstrated that disease characteristics including complexity, the diagnoses themselves and disease control, the patients’ environment (notably availability of social support), and patient literacy/ability to navigate the healthcare system, were considered when referring patients for the iCMP.10

Each patient enrolled in iCMP is assigned a specific care management lead (e.g., a nurse for patients where multiple medical issues drive complexity, or a social worker for psychologically complex patients). This care management lead conducts an initial assessment, creates a care plan, and manages the patient with the assistance of other members of the team depending on needs they uncover. The program, established in 2006, was initially supported by the Medicare Care Management for High Cost Beneficiaries Demonstration Program. 8 In 2012, iCMP was extended and expanded without significant changes to the structure through MGB’s participation in the Pioneer Accountable Care Organization (ACO) contract and is now seen as the main driver of the multihospital system’s positive performance on ACO risk contracts. 8 The shared savings earned from those risk contracts provide the main funding mechanism for iCMP; enrollment is restricted to aligned beneficiaries.8

SDoH are not systematically collected or documented as part of routine rheumatic disease care. However, screening for SDoH-related needs is part of the iCMP care manager’s initial assessment for enrolled patients and was repeated as indicated, and more recently, is encouraged at least once yearly. Care managers’ initial high risk assessments include both free text documentation of reasons that render the patient “high risk” including living situation, functional status and financial concerns, and a series of multiple choice questions (Supplemental Material 1). This information, however, is not consistently available or complete for all historically enrolled patients, is not in a location that is readily accessible by other providers, does not include coded fields to facilitate data extraction, and at times lacks the granular detail included in more descriptive notes. SDoH-related details and needs uncovered during subsequent conversations are often documented in free text notes. To understand variations in documentation including but not limited to information collected in the baseline assessment, and the prevalence of SDoH among individuals with the highest likelihood of being screened in detail for SDoH needs (compared to the general population), we included adults ≥18 years of age with ≥1 ICD-9 or 10 code for a systemic rheumatic condition, crystalline arthritis, or osteoarthritis enrolled in iCMP across MGB between 1/1/12 (the year iCMP was expanded) -10/18/21. To qualify for iCMP, individuals were required to have primary care providers within the MGB system, which ensured that they also had notes in the EHR.

Literature Review

The study team (SU, MC) conducted literature reviews through PubMed and leveraged the Unified Medical Language System® (UMLS), a collection of medical terms which includes some SDoH, to expand terms from the literature review to develop a dictionary of SDoH terms to guide in-depth EHR reviews. SDoH were defined using the categories of financial insecurity, food insecurity, housing instability, access to transportation, education, childcare, and access to medications. Together, the reviewers (SU, MC, CF) developed a detailed standard operating procedure for the EHR reviews defining the date range of notes to review, types of notes, and search terms informed by the literature review (Supplemental Material 2), a prior pilot study of SDoH in a small subset of individuals with SLE and EHR data linked to billing claims,12 and UMLS.

SDoH Data Extraction: Manual EHR Review

Notes eligible for review included text written by physicians, nurses, social workers, psychiatrists, psychologists, dieticians/nutritionists, pharmacists, physical therapists, rheumatologists, and other health care professionals. Study team members (SU, MC, CF) first reviewed the same 5 charts and adjudicated discrepancies. A second set of 5 charts were then reviewed and the adjudication process was repeated. After discrepancies were adjudicated and definitions were agreed upon, the review team refined the systematic method for data extraction and divided the remaining charts with weekly meetings to review together charts that raised questions for the primary reviewer. For each SDoH examined in this study, reviewers reported whether there were definite needs documented (“yes”), documentation of no need (“no”), the possibility of a need (“possible”), or no documentation (“not mentioned”). SDoH variables included financial insecurity, food insecurity, housing instability, transportation, education, childcare, medication access, and medication adherence. If an individual ever had a description of a need, they were categorized as having that need (“yes”), regardless of whether another note at a different time countered that narrative. Chart reviews were conducted during the dates of iCMP enrollment to focus on those notes most likely to have existing needs documented. A SDoH-related need was categorized as “possible” if there was a suggestion of a need but after comprehensive chart review by the team, a definite conclusion could not be reached. For each chart reviewed, demographic information was also extracted (age, gender, race, ethnicity, primary language, primary and secondary insurance), and the rheumatic condition.

ICD-10 Z Code Identification

Z codes, introduced in the end of 2015, with studies of uptake beginning in 2016, are a subset of ICD-10-CM (International Classification of Diseases, Tenth Revision, Clinical Modification) codes used to report social determinants of health.7 The MGB Research Patient Data Registry (RPDR) was used to identify relevant ICD-10 Z codes for SDoH during the dates of iCMP enrollment. RPDR is a clinical data warehouse that integrates clinical information across the MGB healthcare system for research purposes.13 The Z-codes provided by RPDR were from claims submitted by providers within our system. While a range of SDoH Z codes exist, we focused on those most relevant to the SDoH we were studying (e.g., problems related to education, employment, housing and economic circumstances, and problems related to medical facilities and other healthcare),14 and stratified by the number of Z codes per year beginning in 2016.

Statistical analyses

We used descriptive analyses to examine the overall prevalence of SDoH in this population from chart review and by Z code, and then examined SDoH identified from chart review by rheumatic condition. We used multivariable logistic regression including age, gender, race, ethnicity, insurance status, and rheumatic condition, to examine associations between demographic factors ≥1 (vs. 0) SDOH-related needs (Odds Ratio [OR] with 95% Confidence Intervals (CIs). We conducted an additional analysis removing transportation from the outcome recognizing that this isolated need may be distinct from other SDoH measured. Analyses were conducted using SAS version 9.4 (SAS Institute) and R version 4.2.2 (R Core Team). P-values were 2-sided and statistical significance was set at an alpha level of 0.05. Additionally, we determined whether ICD-10 Z Codes for SDoH were documented and the distribution of these codes over time. This study was approved by the MGB Institutional Review Board.

Results

Among 20,395 individuals (≥18 years) with rheumatic conditions enrolled in MGB iCMP, we randomly selected 600 individuals. We excluded individuals without iCMP documentation (N=35), or a clear rheumatic or musculoskeletal disease diagnosis (N=7). Among the 558 remaining individuals, the mean (SD) age was 73.7 (13.2) years, 62% were female, 80% were White, 9% were Black, and 82% were non-Hispanic (Table 1). The mean (SD) period of iCMP enrollment was 3.3 (2.4) years; there was no statistically significant difference in presence vs. absence of SDoH need documentation by mean enrollment time. There were 148 patients (27%) with a systemic rheumatic disease, 120 (22%) with crystalline arthritis, and 290 (52%) with osteoarthritis without systemic or crystalline disease (categories are not mutually exclusive). Systemic rheumatic conditions included rheumatoid arthritis, palindromic rheumatism, systemic lupus erythematosus, systemic sclerosis, juvenile idiopathic arthritis, ankylosing spondylitis/sacroiliitis, Sjogren’s/sicca syndrome, psoriatic arthritis, mixed/undifferentiated connective tissue disease, vasculitis, sarcoidosis, inflammatory myositis, polymyalgia rheumatica. Crystalline disease included gout and pseudogout.

Table 1.

Baseline Characteristics (N=558)

Age- mean (SD) 73.7 (13.2)
Gender – N (%) Male 210 (38)
Female 348 (62)
Race – N (%) Black 50 (9)
White 449 (80)
Other/Not disclosed 60 (11)
Ethnicity – N (%) Hispanic 49 (9)
Non-Hispanic 459 (82)
Other 50 (9)
Primary Language – N (%) English 516 (92)
Spanish 28 (5)
Other 14 (3)
Primary Insurance – N (%) Medicaid 32 (6)
Medicare 462 (83)
Commercial 39 (7)
Other 25 (4)
Rheumatic Condition – N (%) Systemic Rheumatic Condition* 148 (27)
Crystalline Disease# 120 (22)
Osteoarthritis 290 (52)
*

Includes rheumatoid arthritis, palindromic rheumatism, systemic lupus erythematosus, systemic sclerosis, juvenile idiopathic arthritis, ankylosing spondylitis/sacroiliitis, Sjogren’s/sicca syndrome, psoriatic arthritis, mixed/undifferentiated connective tissue disease, vasculitis, sarcoidosis, inflammatory myositis, polymyalgia rheumatica

#

Includes gout and pseudogout

Of the 558 charts reviewed, 249 (45%) had documentation of at least one definite (“yes”) SDoH-related need. Overall, 171 individuals (31%) had definite evidence of financial needs, 105 (19%) transportation needs, 94 (17%) food insecurity, and 30 (5%) housing instability. Inclusive of charts that were marked possible, 90% (490) contained documentation indicating at least one yes or one possible SDoH-related need. There were 126 individuals (23%) with possible evidence of financial needs, 176 (32%) with possible transportation needs, 92 (16%) with possible food insecurity, and 40 (7%) with possible housing instability. We also stratified documentation by age <65 vs. ≥65 years and found that among those with definite SDoH needs, 88 (35%) individuals were <65 and 161 (65%) were ≥65 years. In addition to these SDoH-related needs, we also assessed for documentation of education (N=10) and childcare-related (N=8) concerns. As this was an older population, we found that these needs were infrequently described and therefore they were not included in the final models.

Prevalence of documented SDoH needs varied by rheumatic condition (Figure 1). Among individuals with a systemic rheumatic disease, 39 individuals (26%) had evidence of financial insecurity, 32 (22%) transportation needs, and 22 (15%) food insecurity. For individuals with osteoarthritis, 98 individuals (34%) had evidence of financial insecurity, 47 (16%) transportation needs, and 51 (18%) food insecurity. Furthermore, among individuals with a crystalline disease, 34 individuals (28%) had evidence of financial insecurity, 26 (22%) transportation needs, and 21 (18%) food insecurity.

Figure 1:

Figure 1:

Percentage of individuals with SDoH-related needs by rheumatic condition

We found significant heterogeneity in the descriptions and terms used for each SDoH need in the notes. Descriptions of financial insecurity included terms like “can’t afford”, “financial assistance”, “limited income”, and “struggling financially.” Food insecurity was described with terms such as “food stamps” and housing instability was often implied with discussions of “subsidized housing.” For our patient population, transportation needs were indicated by “PT-1”, or “The Ride”, a Massachusetts-based public transportation service for patients with temporary or permanent disabilities.15 We also noted terms and descriptions that required more context for interpretation. For example, “home delivered meals” was often used to describe Meals on Wheels, a service provided to seniors who experience physical declines or financial hardship,16 however this is not a universal term for this service. While the terms “poor” and “poverty” are often used to describe individuals with insufficient funds, they were more frequently used in notes by physicians to describe “poor functional status” or “poverty of speech” rather than financial insecurity, emphasizing the importance of context alongside commonly used terms, when delineating SDoH-related needs.

Descriptions and terms were found in various structured and unstructured note types such as telephone encounters, patient care coordination notes, smart form templates, discharge summaries, progress notes, and consults. Notes were written by physicians in various fields such as rheumatology, primary care, psychiatry, and physical rehabilitation services and by iCMP nurses, physical therapists, occupational therapists, pharmacists, and medical assistants. Among the notes reviewed by the study team, 245 notes had evidence of financial insecurity. Ninety-five (39%) of the notes for financial insecurity were recorded by an iCMP nurse, 53 (22%) notes were documented by a social worker, and only one (0.4%) note was written by a rheumatologist. Most mentions of housing instability and food insecurity were similarly included in notes written by the iCMP nurses or social workers. For housing instability, study team members extracted a total of 42 notes with a positive mention; 13 (31%) of the notes were written by a social worker and 11 (26%) written by an iCMP nurse. For food insecurity, study team members extracted a total of 119 with a positive mention; 51 (43%) of the notes were written by an iCMP nurse and 20 (17%) by a social worker. There were no notes written by rheumatologists that indicated housing or food needs among those patients with clear documentation of these needs elsewhere in their charts.

In the multivariable model, the odds of having ≥1 SDoH need was 2.45 times higher (95% CI 1.17-5.11) for Black vs. White individuals, 6.72 times higher (95% CI 2.79-16.21) for Medicaid vs. Commercial insurance beneficiaries, 3.04 times higher (95% CI 1.32-6.97) for Medicare vs. Commercial insurance beneficiaries, and 4.12 times higher (95% CI 1.30-13.04) for individuals without insurance vs. Commercial insurance beneficiaries (Table 2). We did not observe statistically significant differences by age, rheumatic condition, gender, or ethnicity. The multivariable model without transportation in the outcome resulted in similar findings (Supplemental Material 3).

Table 2.

Multivariable logistic regression model examining the odds of >1 SDoH-related needs vs. no need (N=558)

Descriptive Categories Odds Ratio 95% Confidence Interval
Age (years) 0.97 0.95-0.98
Rheumatic Conditions (Ref=Osteoarthritis)
  Systemic Rheumatic Disease 0.79 0.51-1.22
  Crystalline Arthritis 1.15 0.71-1.84
Gender (Ref=Female)
  Male 0.70 0.47-1.05
Race (Ref=White)
  Black 2.45 1.17-5.11
  Other 0.99 0.48-2.04
Ethnicity (Ref=Not Hispanic)
  Hispanic 1.67 0.77-3.65
  Other 1.41 0.75-2.66
Insurance (Ref=Commercial)
  Medicaid 6.72 2.79-16.21
  Medicare 3.04 1.32-6.97
  No insurance 4.12 1.30-13.04

Bolded values indicate statistical significance

We also examined the overall prevalence of SDoH by Z code (Table 3). Among our sample of 558 charts, we found that 26 (5%) individuals were assigned at least 1 SDoH Z code. The most frequently used SDoH Z code was Z59.9, defined as “Problem related to housing and economic circumstances, unspecified.” SDoH Z codes were also examined per year starting in 2016. After excluding individuals who died prior to 2016 or did not have a diagnosis or encounter code after 2016, uptake remained under 5% each year and was highest in 2019 (3.4%) (Supplemental Material 4). One individual did not have a SDoH need in the reviewed categories but was labeled with a billing Z code of Z75.8 (“Other problems related to medical facilitates and other health care”). A secondary analysis yielded similar results when this individual, who had no SDoH identified by manual chart review, was reclassified as having ≥1 SDoH need in the multivariable model (Supplemental Material 5).

Table 3.

Number of Individuals with SDoH-related Billing “Z” Code (N=26)

ICD-10 Codes to Identify Social Determinants of Health Number of individuals with ICD-10 codes*
Z55 – Problems related to education and literacy
   Z55.0 Illiteracy and low-level literacy 1
   Z55.9 Problems related to education and literacy, unspecified 5
Z56 – Problems related to employment and unemployment
   Z56.0 Unemployment, unspecified 6
Z59 – Problems related to housing and economic circumstances
   Z59.0 Homelessness 9
   Z59.1 Inadequate housing 2
   Z59.4 Lack of adequate food and safe drinking water 7
   Z59.48 Other specified lack of adequate food 2
   Z59.6 Low income 1
   Z59.7 Insufficient social insurance and welfare support 4
   Z59.8 Other problems related to housing and economic circumstances 4
   Z59.9 Problem related to housing and economic circumstances, unspecified 10
Z75 - Problems related to medical facilities and other health care
   Z75.8 Other problems related to medical facilities and other health care 1
*

532 individuals were missing a relevant ICD-10 code

Discussion

Social determinants of health play a central role in disparities in care and outcomes in rheumatic and musculoskeletal conditions. Detailed chart review in this population of medically and/or psychosocially complex patients with rheumatic conditions receiving care at a multihospital academic medical center uncovered documentation of SDoH-related needs in nearly half of the charts, and when expanded to include those with possible needs, documentation increased to 90% of the charts. The large discrepancy between the positive and possible mentions demonstrates the difficulty of accurately capturing information through unstructured data. Notably, we found that SDoH documentation was often recorded by iCMP nurses and social workers rather than by rheumatologists with only one rheumatologist’s note indicating a financial, food or housing-related need, considerably less frequent than in primary care physicians’ notes. Similarly, while 45-90% of the charts reviewed in this population had a definite or possible indication of a SDoH-related need, only 5% of charts had an SDoH billing Z code of interest without any notable increases in uptake over time. Prior studies in the general population have similarly demonstrated low uptake of Z codes to date,1719 although utilization overall has increased since Z codes were implemented.7 While the utilization of Z codes remains a promising option to understand population-level SDoH needs, rheumatologists at our multihospital institution are not trained or incentivized to use these codes or to screen for these needs. Patients seen in rheumatology who have primary care physicians within our system and who qualified for enrollment in iCMP, were screened by iCMP care managers and there were community resource specialists available to address the needs that were uncovered. However, both structured SDoH needs assessments and resources to meet these needs were not available as part of the rheumatology clinic infrastructure, which may in part explain the absence of described needs in rheumatology notes.

We also observed a higher prevalence of financial insecurity compared to other SDoH-related needs. Higher SDoH-related burden was noted among Black individuals and among individuals without insurance, as well as Medicaid and Medicare beneficiaries. The stark differences between racial and socioeconomic groups support findings from prior studies2023 and contribute to the underlying differences in health and health outcomes, even more so among a population with complex care needs.24 We found an overall higher prevalence of SDoH needs, and financial insecurity in particular, among individuals with osteoarthritis compared to other rheumatic conditions however this difference was not statistically significant in adjusted analyses.

Several studies to date demonstrate the relevance of financial needs among individuals with rheumatic conditions. A 2016 study found that Medicaid patients were less likely to receive care from a rheumatologist and more likely to have delays in care in receiving medications, highlighting the role of socioeconomic status (SES) in healthcare access.25 In another study by Callahan et al., researchers found an association between lower individual- and community-level SES and poor physical health outcomes, highlighting the of role SES in disease outcomes for individuals with arthritis.26 Additionally, for individuals with RA, low SES has been associated with worse clinical outcomes and delays in treatment.27 Fewer studies to date examine the burden of financial needs, food insecurity, and transportation needs among individuals with rheumatic conditions and more research is needed to demonstrate associations with medication and healthcare use and outcomes.

Strengths of our study include an understanding of SDoH documentation within EHR notes of medically and/or psychosocially complex individuals with rheumatic and musculoskeletal conditions. Through intensive chart reviews, we have identified SDoH-related needs among individuals with rheumatic conditions, the way this information is being documented, and the providers who are recording this information. We have also examined uptake of ICD-10 Z codes among patients with rheumatic conditions and identified a gap for further educational efforts to promote greater utilization, and possibly another avenue to understand patient needs and complexity at a population level for individuals with rheumatic conditions. Prior studies that examine SDoH among patients with rheumatic conditions use methods including literature reviews,28 patient questionnaires,29,30 observational studies,31 and scoping reviews.32 However, we aimed to examine the prevalence of various SDoH-related needs using EHR, which allows us to approximate real-world data. By understanding SDoH documentation in EHR notes, we can inform both the allocation of resources to meet these needs and future systematic data extraction strategies.

Limitations of this study include lack of data outside of our population cohort, as we have only analyzed individuals enrolled in the iCMP program who may be more likely to both have higher prevalence and documentation of SDoH-related needs compared to a less complex population. Compared to a prior study of individuals receiving rheumatic disease care within our multihospital system,6 this population, on average, was about 10 years older, had a similar gender and race distribution, and had a higher percentage of Hispanic individuals and Medicare beneficiaries. As iCMP only requires the primary care physician to be within the multihospital system and not the subspecialists, there may be misclassification of rheumatic/musculoskeletal conditions. Furthermore, our population was older due to Medicare insurance-related eligibility for the original iCMP and as such, our findings may not represent the prevalence or distribution of SDoH in younger populations. With the absence of systematic screening at our institution in rheumatology clinics during the timeframe of this study, understanding need in this enriched population allowed for the identification of strategies and infrastructure for future efforts. Additionally, the Z codes examined were limited to the billing data submitted by providers within our multihospital healthcare system and do not capture information from providers outside of our system. We do not expect that we are missing a significant number however, as the primary care setting would be the most likely for use of these codes and all patients in this cohort had their primary care team within our system. Efforts to increase awareness of Z codes across providers at our institution may increase their use and allow for these factors to be better accounted for when understanding medical complexity, resource allocation, and care utilization patterns. Another limitation to note is that reviewing charts and labeling them is difficult as SDoH are, by definition, dynamic, making it challenging to classify individuals as having versus not having needs overall rather than at specific time points. Z codes are important for population-level data, but our chart reviews demonstrated the importance of clinical context from narrative notes to truly understand the extent of SDoH-related-needs. Further, SDoH documentation was infrequently structured or standardized in notes, and differences may, in part, reflect ascertainment bias within this academic-based complex care population. It is plausible that providers, including rheumatologists, may ask about SDoH but not routinely document them in their clinical notes or do not ask as they do not have the necessary tools to address them and as such, needs may be even higher than what was uncovered in this study. Lastly, we did not examine neighborhood-level factors as our focus was on note-based documentation of SDoH however future analyses are planned to link these SDoH needs to area deprivation indices and neighborhood environmental exposures in this population.

This study illustrates the high burden of SDoH-related needs among individuals with rheumatic and musculoskeletal conditions and the importance of infrastructure to document and address these needs. Despite the high prevalence of SDoH in this population, in our chart review, we found only one note by a rheumatologist documenting financial, food or housing-related needs, suggesting that heightened awareness is needed for rheumatologists, and infrastructure is required in rheumatology clinics to meet the uncovered needs. Future studies should develop processes that effectively incorporate SDoH screening and EHR documentation into routine rheumatology care and that efficiently extract these data, and the actions taken in response. A strategy implemented in primary care at Boston Medical Center screened patients for SDoH and their responses were linked to their EHR and incorporated into the structured data in their charts using ICD-10 Z codes. Then, if patients requested assistance, referrals (in the patients’ primary languages) were provided to guide them to necessary resources.33 This strategy has the potential to be replicated in other institutions including our own. We encourage rheumatologists to understand their patient needs both biomedically and psychosocially and to advocate for resources and referral systems within their institutions to address these needs. In a clinical and public health sphere, SDoH information will allow for better care access and quality, more equitable enrollment in clinical trials, and more comprehensive research studies that appropriately account for the key contributions of social determinants to care utilization, medication adherence and outcomes. Understanding SDoH allow healthcare providers to provide integrated care, and by connecting patients to appropriate services to address these needs, disparities in healthcare access and outcomes can be reduced.34

Supplementary Material

Supinfo2

Supplemental Material 2: Search Terms for Standard Operating Procedure

Supinfo1

Supplemental Material 1: High Risk Baseline Assessment

Supinfo3

Supplemental Material 3: Multivariable logistic regression model examining the odds of ≥1 SDoH-related needs vs. no need (N=558) excluding transportation

Supinfo4

Supplemental Material 4: Number of individuals with Z codes in each year

Supinfo5

Supplemental Material 5: Multivariable logistic regression model examining the odds of ≥1 SDoH-related needs vs. no need (N=558) inclusive of positive Z code

Significance and Innovations.

  • Through intensive chart reviews, we found a significant burden of social determinant of health (SDoH)-related needs, especially financial insecurity, and transportation challenges, in a subset of patients with rheumatic and musculoskeletal diseases enrolled in a multihospital complex care management program.

  • While nearly 50% of individuals with rheumatic conditions enrolled in a complex care management program had at least one documented SDoH-related need, only 5% had an ICD-10 SDoH-related billing (“Z”) code. Use of Z codes did not increase over time, suggesting that structured claims data do not capture the burden of need.

  • There was nearly no documentation of SDoH-related needs by rheumatologists suggesting that further education of rheumatologists, and rheumatology-based infrastructure to screen for and address these needs, are essential.

Acknowledgements:

We would like to acknowledge Christine Vogeli, PhD and the Population Health team at Mass General Brigham for their assistance with data from individuals enrolled in the integrated care management program.

Funding:

This study was funded by the NIH NIAMS P30AR072577 and K23 AR071500. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

Financial Disclosures:

Dr. Feldman receives research support to her institution from the NIH, the Rheumatology Research Foundation, the Arthritis Foundation, Pfizer Pharmaceuticals and the BMS Foundation. She has served as a consultant on grants to the University of Alabama, the American College of Rheumatology, the Lupus Foundation of America, received conference travel reimbursement from RILITE Foundation, and consults for OM1, Inc. She previously served on the American College of Rheumatology Board of Directors, and currently serves on the DEI Task Force of the Arthritis Foundation and the Research Governing Board of CCHERS, Inc.

References

  • 1.Social Determinants of Health - Healthy People 2030 | health.gov. Accessed December 13, 2022. https://health.gov/healthypeople/priority-areas/social-determinants-health
  • 2.Braveman P, Gottlieb L. The Social Determinants of Health: It’s Time to Consider the Causes of the Causes. Public Health Rep. 2014;129(1_suppl2):19–31. doi: 10.1177/00333549141291S206 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Izadi Z, Li J, Evans M, et al. Socioeconomic Disparities in Functional Status in a National Sample of Patients With Rheumatoid Arthritis. JAMA Netw Open. 2021;4(8):e2119400. doi: 10.1001/jamanetworkopen.2021.19400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Cai Q, Pesa J, Wang R, Fu AZ. Depression and food insecurity among patients with rheumatoid arthritis in NHANES. BMC Rheumatol. 2022;6(1):6. doi: 10.1186/s41927-021-00236-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Yelin E, Trupin L, Yazdany J. A Prospective Study of the Impact of Current Poverty, History of Poverty, and Exiting Poverty on Accumulation of Disease Damage in Systemic Lupus Erythematosus. Arthritis Rheumatol. 2017;69(8):1612–1622. doi: 10.1002/art.40134 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Santacroce L, Dellaripa PF, Costenbader KH, Collins J, Feldman CH. Association of Area-Level Heat and Social Vulnerability With Recurrent Hospitalizations Among Individuals With Rheumatic Conditions. Arthritis Care Res. n/a(n/a). doi: 10.1002/acr.25015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Utilization of Z Codes for Social Determinants of Health among Medicare Fee-for-Service Beneficiaries, 2019. Published online 2021.
  • 8.Hsu J, Price M, Vogeli C, et al. Bending The Spending Curve By Altering Care Delivery Patterns: The Role Of Care Management Within A Pioneer ACO. Health Aff Proj Hope. 2017;36(5):876–884. doi: 10.1377/hlthaff.2016.0922 [DOI] [PubMed] [Google Scholar]
  • 9.Vogeli C, Spirt J, Brand R, et al. Implementing a hybrid approach to select patients for care management: variations across practices. Am J Manag Care. 2016;22(5):358–365. [PubMed] [Google Scholar]
  • 10.Haime V, Hong C, Mandel L, et al. Clinician considerations when selecting high-risk patients for care management. Am J Manag Care. 2015;21(10):e576–582. [PubMed] [Google Scholar]
  • 11.A Team Approach to Care. Massachusetts General Hospital. Accessed March 29, 2023. https://www.massgeneral.org/primary-care/integrated-care-management/our-team
  • 12.Taber KA, Williams JN, Huang W, et al. Use of an Integrated Care Management Program to Uncover and Address Social Determinants of Health for Individuals With Lupus. ACR Open Rheumatol. 2021;3(5):305–311. doi: 10.1002/acr2.11236 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Identify Subjects / Request Data | Mass General Brigham RISC. Accessed January 23, 2023. https://rc.partners.org/research-apps-services/identify-subjects-request-data [Google Scholar]
  • 14.PRUP135_ICD10-km.pdf. Accessed December 22, 2022. https://www.hopkinsmedicine.org/johns_hopkins_healthcare/providers_physicians/resources_guidelines/provider_communications/2021/PRUP135_ICD10-km.pdf
  • 15.The RIDE | Accessibility on the MBTA | MBTA. Accessed December 13, 2022. https://www.mbta.com/accessibility/the-ride [Google Scholar]
  • 16.Meals on Wheels America. Accessed December 13, 2022. https://www.mealsonwheelsamerica.org/
  • 17.Weeks WB, Cao SY, Lester CM, Weinstein JN, Morden NE. Use of Z-Codes to Record Social Determinants of Health Among Fee-for-service Medicare Beneficiaries in 2017. J Gen Intern Med. 2020;35(3):952–955. doi: 10.1007/s11606-019-05199-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Liss DT, Cherupally M, Kang RH, Aikman C, Cooper AJ, O’Brien MJ. Social Needs Identified by Diagnostic Codes in Privately Insured U.S. Adults. Am J Prev Med. 2022;63(6):1007–1016. doi: 10.1016/j.amepre.2022.07.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Torres JM, Lawlor J, Colvin JD, et al. ICD Social Codes: An Underutilized Resource for Tracking Social Needs. Med Care. 2017;55(9):810–816. doi: 10.1097/MLR.0000000000000764 [DOI] [PubMed] [Google Scholar]
  • 20.Reduce the proportion of people living in poverty — SDOH-01 - Healthy People 2030 | health.gov. Accessed December 16, 2022. https://health.gov/healthypeople/objectives-and-data/browse-objectives/economic-stability/reduce-proportion-people-living-poverty-sdoh-01
  • 21.Probst JC, Ajmal F. Social Determinants of Health among the Rural African American Population. Published online 2019.
  • 22.Noonan AS, Velasco-Mondragon HE, Wagner FA. Improving the health of African Americans in the USA: an overdue opportunity for social justice. Public Health Rev. 2016;37(1):12. doi: 10.1186/s40985-016-0025-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Williams DR, Mohammed SA, Leavell J, Collins C. Race, socioeconomic status, and health: Complexities, ongoing challenges, and research opportunities. Ann N Y Acad Sci. 2010;1186(1):69–101. doi: 10.1111/j.1749-6632.2009.05339.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.The social determinants of health: why they matter to improving health outcomes | Maximus. Accessed December 16, 2022. https://maximus.com/article/social-determinants-health-why-they-matter-improving-health-outcomes [Google Scholar]
  • 25.Cifaldi M, Renaud J, Ganguli A, Halpern MT. Disparities in care by insurance status for individuals with rheumatoid arthritis: analysis of the medical expenditure panel survey, 2006–2009. Curr Med Res Opin. 2016;32(12):2029–2037. doi: 10.1080/03007995.2016.1227775 [DOI] [PubMed] [Google Scholar]
  • 26.Callahan LF, Martin KR, Shreffler J, et al. Independent and combined influence of homeownership, occupation, education, income, and community poverty on physical health in persons with arthritis. Arthritis Care Res. 2011;63(5):643–653. doi: 10.1002/acr.20428 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Molina E, Del Rincon I, Restrepo JF, Battafarano DF, Escalante A. Association of socioeconomic status with treatment delays, disease activity, joint damage, and disability in rheumatoid arthritis. Arthritis Care Res. 2015;67(7):940–946. doi: 10.1002/acr.22542 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Jackson LE, Danila MI. Healthcare disparities in telemedicine for rheumatology care. Curr Opin Rheumatol. 2022;34(3):171–178. doi: 10.1097/BOR.0000000000000869 [DOI] [PubMed] [Google Scholar]
  • 29.Looper KJ, Mustafa SS, Zelkowitz P, Purden M, Baron M. Work instability and financial loss in early inflammatory arthritis. Int J Rheum Dis. 2012;15(6):546–553. doi: 10.1111/1756-185X.12009 [DOI] [PubMed] [Google Scholar]
  • 30.Callhoff J, Ramos AL, Zink A, Hoffmann F, Albrecht K. The Association of Low Income with Functional Status and Disease Burden in German Patients with Rheumatoid Arthritis: Results of a Cross-sectional Questionnaire Survey Based on Claims Data. J Rheumatol. 2017;44(6):766–772. doi: 10.3899/jrheum.160966 [DOI] [PubMed] [Google Scholar]
  • 31.Barton J, Trupin L, Schillinger D, et al. Racial and Ethnic Disparities in Disease Activity and Function among Persons with Rheumatoid Arthritis from University-Affiliated Clinics. Arthritis Care Res. 2011;63(9):1238–1246. doi: 10.1002/acr.20525 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Thomas M, Harrison M, De Vera MA. Reporting of Determinants of Health Inequities in Rheumatoid Arthritis Randomized Controlled Trials in Canada: A Scoping Review. Arthritis Care Res. Published online July 6, 2022. doi: 10.1002/acr.24978 [DOI] [PubMed] [Google Scholar]
  • 33.Buitron de la Vega P, Losi S, Sprague Martinez L, et al. Implementing an EHR-based Screening and Referral System to Address Social Determinants of Health in Primary Care. Med Care. 2019;57(Suppl 2):S133–S139. doi: 10.1097/MLR.0000000000001029 [DOI] [PubMed] [Google Scholar]
  • 34.HealthITAnalytics. Costs Fell by 11% When Payer Addressed Social Determinants of Health. HealthITAnalytics. Published June 5, 2018. Accessed December 13, 2022. https://healthitanalytics.com/news/costs-fell-by-11-when-payer-addressed-social-determinants-of-health [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo2

Supplemental Material 2: Search Terms for Standard Operating Procedure

Supinfo1

Supplemental Material 1: High Risk Baseline Assessment

Supinfo3

Supplemental Material 3: Multivariable logistic regression model examining the odds of ≥1 SDoH-related needs vs. no need (N=558) excluding transportation

Supinfo4

Supplemental Material 4: Number of individuals with Z codes in each year

Supinfo5

Supplemental Material 5: Multivariable logistic regression model examining the odds of ≥1 SDoH-related needs vs. no need (N=558) inclusive of positive Z code

RESOURCES