This cross-sectional study examines associations between health-related social needs and health care quality and utilization outcomes among Medicare Advantage beneficiaries with type 2 diabetes.
Key Points
Question
Which health-related social needs (HRSNs) are associated with health care quality and utilization outcomes in a Medicare Advantage population with type 2 diabetes?
Findings
Of 21 528 beneficiaries in this cross-sectional study, 56.9% had at least 1 HRSN; the most prevalent were financial strain, food insecurity, and poor housing quality. Loneliness and lack of transportation were associated with lower care quality and increased acute care utilization; financial strain, utility insecurity, and housing insecurity with lower care quality; and food insecurity with increased acute care utilization.
Meaning
The findings suggest that some self-reported HRSNs are associated with diabetes-specific health and utilization outcomes; patient-level information may be relevant to diabetes interventions.
Abstract
Importance
Recent research highlights the association of social determinants of health with health outcomes of patients with type 2 diabetes (T2D).
Objective
To examine associations between health-related social needs (HRSNs) and health care quality and utilization outcomes in a Medicare Advantage population with T2D.
Design, Setting, and Participants
This cross-sectional study used medical and pharmacy claims data from 2019. An HRSN survey was given between October 16, 2019, and February 29, 2020, to Medicare Advantage beneficiaries. Inclusion criteria were diagnosis of T2D, age of 20 to 89 years, continuous Medicare Advantage enrollment in 2019, and response to the HRSN survey. Data were analyzed between June 2021 and January 2022.
Exposures
Enrollment in Medicare Advantage, diagnosis of T2D, and completion of a survey on HRSNs.
Main Outcomes and Measures
Quality outcomes included diabetes medication adherence, statin adherence, completion of a glycated hemoglobin (HbA1c) laboratory test in the past 12 months, and controlled HbA1c. Utilization outcomes included all-cause hospitalization, potentially avoidable hospitalization, emergency department discharge, and readmission.
Results
Of the 21 528 Medicare Advantage beneficiaries with T2D included in the study (mean [SD] age, 71.0 [8.3] years; 55.4% women), most (56.9%) had at least 1 HRSN. Among the population with T2D reporting HRSNs, the most prevalent were financial strain (73.6%), food insecurity (47.5%), and poor housing quality (39.1%). In adjusted models, loneliness (odds ratio [OR], 0.85; 95% CI, 0.73-0.99), lack of transportation (OR, 0.80; 95% CI, 0.69-0.92), utility insecurity (OR, 0.86; 95% CI, 0.76-0.98), and housing insecurity (OR, 0.78; 95% CI, 0.67-0.91) were each associated with lower diabetes medication adherence. Loneliness and lack of transportation were associated with increased emergency visits (marginal effects of 173.0 [95% CI, 74.2-271.9] and 244.6 [95% CI, 150.4-338.9] emergency visits per 1000 beneficiaries for loneliness and transportation, respectively). Food insecurity was the HRSN most consistently associated with higher acute care utilization (marginal effects of 84.6 [95% CI, 19.8-149.4] emergency visits, 30.4 [95% CI, 9.5-51.3] inpatient encounters, and 17.1 [95% CI, 4.7-29.5] avoidable hospitalizations per 1000 beneficiaries).
Conclusions and Relevance
In this cross-sectional study of Medicare Advantage beneficiaries with T2D, some HRSNs were associated with care quality and utilization. The results of the study may be used to direct interventions to the social needs most associated with T2D health outcomes and inform policy decisions at the insurance plan and community level.
Introduction
The Centers for Medicare & Medicaid Services (CMS) has emphasized the importance of understanding and addressing adverse social determinants of health (SDOHs), such as food, housing, and transportation insecurity, as a means to reduce health inequities and improve population health.1 This emphasis is based on research demonstrating the association of social factors with health outcomes and health inequities; cumulative social adversity is associated with poorer outcomes.2,3 Prior research has established significant associations between health care utilization and social adversity in type 2 diabetes (T2D),4,5,6 which affects more than 26% of people aged 65 years and older7 and costs more than $200 billion each year in the US.8
Due to the substantial consequences of T2D for the health care system, identifying ways to improve the health of patients with T2D is a high priority. Given the associations between social adversity and T2D, efforts to improve diabetes-related outcomes are increasingly focused on SDOHs.6,9,10 Much of the existing work at the intersection of health care and SDOHs relies on area-level SDOH measures because patient-level data are not consistently available. However, it is not clear that area-level measures are adequate; in a national community health center sample, Cottrell et al11 found that area-level SDOH measures would miscategorize 52% of patients who self-reported social risk factors. With the growing availability of self-reported data, more studies can instead explore whether patient-reported health-related social needs (HRSNs) can better explain health outcomes.12,13,14
To facilitate collection of HRSN data in a large, national demonstration project, the CMS developed the Accountable Health Communities HRSN screening tool for use with Medicare and Medicaid beneficiaries.15 The survey tool includes items assessing food insecurity, transportation, utilities, housing quality, and housing instability as well as supplemental questions on financial strain, loneliness, and social support. In previous studies,14,16 about half of a national sample of Medicare Advantage beneficiaries reported having at least 1 HRSN when surveyed with the Accountable Health Communities screening tool, and HRSNs were associated with acute care utilization. In this study, we extended the prior work examining HRSNs among Medicare Advantage beneficiaries to specifically look at self-reported HRSNs among beneficiaries with T2D and to assess associations between HRSNs and T2D-relevant health care quality and acute care utilization.
Methods
Setting and Participants
This cross-sectional study relied on a survey conducted with a cohort of Medicare Advantage beneficiaries with T2D to examine associations between self-reported HRSNs and health care quality and utilization outcomes. A national sample of Medicare Advantage beneficiaries enrolled in plans offered by Humana Inc, a large, national provider of Medicare Advantage plans, was surveyed about HRSNs from October 16, 2019, through February 29, 2020. The survey included 7 items assessing the following HRSNs: (1) financial strain, (2) food insecurity, (3) loneliness, (4) housing insecurity, (5) poor housing quality, (6) utility insecurity, and (7) unreliable transportation. Participants included in the analysis had a diagnosis of T2D in 2019 (International Statistical Classification of Diseases, Tenth Revision, Clinical Modification code of E11 from 2 different outpatient claims or 1 inpatient claim) and were not contractually excluded from research. Respondents not meeting the age requirement of 20 to 89 years and without continuous enrollment for 2019 were excluded. For the study cohort, survey responses were combined with medical and pharmacy claims to identify associations with diabetes quality of care and utilization outcomes. The Humana Healthcare Research Human Subject Protection Office reviewed this retrospective, cross-sectional study, which used a limited data set, and determined that it did not meet the criteria of human participants research; thus, the study was deemed exempt from institutional review board review and informed consent. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Measures
Health-Related Social Needs
We used the HRSN survey to assess the following HRSNs: (1) financial strain, (2) food insecurity, (3) loneliness, (4) housing insecurity, (5) poor housing quality, (6) utility insecurity, and (7) unreliable transportation. Binary indicators for each HRSN were created and scored as positive (1) or negative (0). We then collapsed HRSN items into binary indicators of the presence or absence of HRSNs for the descriptive statistics. The HRSN-positive group had at least 1 HRSN; the HRSN-negative group reported no HRSNs. In each regression model, we included the 7 individual HRSNs, each coded as binary to indicate the presence or absence of that particular need. Further details of the survey, such as the response rate and responses considered positive, were published previously16 and are available in the eMethods in Supplement 1. Previous analysis demonstrated among the broad sample that survey responders were not meaningfully different from nonresponders on observed characteristics.14
Covariates
Independent variables in the models extracted from the administrative data were age, race and ethnicity, sex, Social Security disability status, Medicare Part D low-income subsidy status, Medicare-Medicaid dual eligible status, geographic region, population density, and risk arrangement of their attributed primary care organization. Covariates were selected based on availability in the record and association with quality measures and/or health care utilization.3,14,17 Demographic characteristics were assessed as of December 31, 2019. Race and ethnicity categories were based on the CMS beneficiary race code, which reflects data self-reported to the Social Security Administration, and were reported as Black, White, other (including Asian, Hispanic, North American Native, and other race or ethnicity), and unknown (owing to inaccuracies in classification for beneficiaries of a race and ethnicity other than Black or White). The Elixhauser Comorbidity Index score18,19,20 and Diabetes Complications Severity Index (DCSI) score were calculated using medical claims from 2019. The DCSI uses 7 categories of complications (cardiovascular disease, nephropathy, retinopathy, peripheral vascular disease, stroke, neuropathy, and metabolic) to calculate a score to predict adverse outcomes, including hospitalization and mortality, based on the number and severity of complications associated with diabetes.17
Health Care Quality and Utilization
We used medical and pharmacy claims data to assess health care quality and utilization metrics in 2019. We assessed health care quality based on the Healthcare Effectiveness Data and Information Set (HEDIS) for comprehensive diabetes care and medication adherence.21 We evaluated the following HEDIS measures: diabetes medication adherence, statin adherence, completion of a glycated hemoglobin (HbA1c) laboratory test in the past 12 months, controlled HbA1c, annual eye examination, and annual kidney function laboratory test. To assess diabetes medication adherence, we identified prescription fills for oral and injectable glucose-lowering medication (excluding insulin) and statins. The HbA1c laboratory results were extracted for the subset of beneficiaries with laboratory results available. We defined medication adherence as a proportion of days covered (PDC) greater than 80% and controlled HbA1c as levels less than 7.0% of total hemoglobin (to convert percentage of total hemoglobin to proportion of total hemoglobin, multiply by 0.01). Kidney and eye examinations were measured but not used as model outputs due to the high number of patients having been administered the examinations. Each quality outcome was scored as binary for meeting the quality indicator (1) or not (0).
We measured acute care utilization as all-cause hospitalization, length of stay for all-cause hospitalization, potentially avoidable hospitalization, emergency department discharge, and readmission. Avoidable hospitalizations were defined using the overall composite and the diabetes composite from the Agency for Healthcare Research and Quality’s Prevention Quality Indicators definition.22 Emergency discharge was defined as a visit to the emergency department (ED) that did not result in inpatient admission. Inpatient encounters included inpatient level of care and skilled nursing facility admissions. Readmission was defined as an inpatient hospital stay within 30 days of the previous inpatient admission discharge date. Utilization measures were summed to create total events per beneficiary during 2019.
Statistical Analysis
Analyses were conducted between June 2021 and January 2022 using SAS Enterprise Guide, version 8.2 (SAS Institute Inc). Descriptive statistics included means and SDs of continuous variables and counts and percentages of categorical variables. Standardized mean differences (SMDs) were calculated to compare demographic and clinical characteristics of the study groups with and without HRSNs. An SMD greater than 0.25 was considered meaningful.23 Unadjusted utilization and quality measures were compared between groups with independent-samples t tests or χ2 tests. For adjusted analyses, we used regression models among complete HRSN cases. Logistic regression models were used to examine the association between HRSNs and binary care quality outcomes. Negative binomial regression models were used to examine the association between HRSNs and the utilization outcomes. Incidence rate ratios and marginal effects per 1000 beneficiaries were calculated for ease of interpretation of the negative binomial regression models. Two-sided P < .05 was considered significant.
Results
Of the total sample of 447 270 Medicare Advantage beneficiaries, 95 091 were contacted and responded to at least 1 of the HRSN survey questions (response rate, 21.3%). Among survey respondents, 26 986 patients (28.4%) had a diagnosis of T2D in 2019 and were not contractually excluded from research. After excluding respondents not meeting the age requirement (n = 359) and without continuous enrollment for 2019 (n = 5099), the final study cohort consisted of 21 528 Medicare Advantage beneficiaries with an HRSN response and a diagnosis of T2D. Mean (SD) age was 71.0 (8.3) years; 55.4% of beneficiaries were female, 44.6% were male, 23.4% were Black, 70.6% were White, 4.9% reported other race and ethnicity, and 1.1% had unknown race and ethnicity. The response rate for each HRSN was as follows: housing insecurity (99.4%), poor housing quality (91.9%), food insecurity (91.9%), unreliable transportation (90.0%), utility insecurity (89.4%), financial strain (88.3%), and loneliness (87.9%). The response rate to the HRSN measures was consistent with the order of the questions, such that the first question had the highest response rate, which gradually decreased over the subsequent questions. All 7 HRSN measures were completed by 82.7% of the cohort. We found no difference in observable characteristics or HRSN prevalence rates between those who completed all HRSN items and those who did not (eTable 1 in Supplement 1). Financial strain and food insecurity were moderately correlated (φ = 0.52), with all other HRSN combinations weakly correlated (φ = 0.10-0.29) (eTable 2 in Supplement 1).
Among the study population, 56.9% reported having at least 1 HRSN. The demographic and clinical characteristics of Medicare Advantage beneficiaries with T2D are reported in Table 1 and grouped according to HRSN-positive or HRSN-negative status. The mean (SD) age of the HRSN-positive group was slightly younger than the HRSN-negative group (69.5 [8.7] vs 72.9 [7.2] years; SMD = 0.42). The HRSN-positive group had higher mean (SD) Elixhauser Comorbidity Index scores (4.97 [2.68] vs 4.35 [2.42]; SMD = 0.25), though there was not a meaningful difference in the mean DCSI score. More than twice as many respondents in the HRSN-positive group qualified for disability (46.6% vs 22.6%; SMD = 0.52) or low-income subsidy (39.6% vs 16.7%; SMD = 0.53) or were Medicare-Medicaid dual eligible (29.7% vs 11.9%; SMD = 0.45) compared with the HRSN-negative population. In the HRSN-positive group, the most frequently reported HRSNs among respondents were financial strain (73.6%), food insecurity (47.5%), poor housing quality (39.1%), and utility insecurity (19.1%).
Table 1. Demographic and Clinical Characteristics by Presence of HRSNs in Beneficiaries With Type 2 Diabetes Enrolled in Medicare Advantage.
| Characteristic | Beneficiariesa | SMDb | ||
|---|---|---|---|---|
| Overall (n = 21 528) | ≥1 HRSN (n = 12 252) | No HRSN (n = 9276) | ||
| Age, mean (SD), y | 71.0 (8.3) | 69.5 (8.7) | 72.9 (7.2) | 0.42 |
| Age category, y | ||||
| 20-49 | 293 (1.4) | 235 (1.9) | 58 (0.6) | 0.12 |
| 50-64 | 3655 (17.0) | 2831 (23.1) | 824 (8.9) | 0.40 |
| 65-89 | 17 580 (81.7) | 9186 (75.0) | 8394 (90.5) | 0.42 |
| Sex | ||||
| Female | 11 927 (55.4) | 7188 (58.7) | 4739 (51.1) | 0.15 |
| Male | 9601 (44.6) | 5064 (41.3) | 4537 (48.9) | 0.15 |
| Race and ethnicityc | ||||
| Black | 5032 (23.4) | 3371 (27.5) | 1661 (17.9) | 0.23 |
| White | 15 201 (70.6) | 8114 (66.2) | 7087 (76.4) | 0.23 |
| Other | 1064 (4.9) | 662 (5.4) | 402 (4.3) | 0.05 |
| Unknown | 231 (1.1) | 105 (0.9) | 126 (1.4) | 0.05 |
| Geographic region | ||||
| Northeast | 656 (3.1) | 377 (3.1) | 279 (3.0) | 0.00 |
| Midwest | 4571 (21.2) | 2460 (20.1) | 2111 (22.8) | 0.07 |
| South | 13 950 (64.8) | 8152 (66.5) | 5798 (62.5) | 0.08 |
| West | 2351 (10.9) | 1263 (10.3) | 1088 (11.7) | 0.05 |
| Population density | ||||
| Urban | 13 594 (63.2) | 7575 (61.8) | 5761 (62.1) | 0.01 |
| Suburban | 5374 (25.0) | 3028 (24.7) | 2346 (25.3) | 0.01 |
| Rural | 2400 (11.2) | 1440 (11.8) | 960 (10.3) | 0.05 |
| Unknown | 160 (0.7) | 209 (1.7) | 209 (2.3) | 0.04 |
| Medicare-Medicaid dual eligible | 4740 (22.0) | 3638 (29.7) | 1102 (11.9) | 0.45 |
| Low-income subsidy eligible | 6394 (29.7) | 4848 (39.6) | 1546 (16.7) | 0.53 |
| Social Security disability eligible | 7811 (36.3) | 5715 (46.6) | 2096 (22.6) | 0.52 |
| Elixhauser Comorbidity Index score, mean (SD) | 4.70 (2.59) | 4.97 (2.68) | 4.35 (2.42) | 0.25 |
| Diabetes Complications Severity Index score, mean (SD) | 1.75 (1.71) | 1.84 (1.76) | 1.62 (1.65) | 0.13 |
| Individual health-related social needsd | ||||
| Financial strain | NA | 8417 (73.6) | NA | NA |
| Food insecurity | NA | 5588 (47.5) | NA | NA |
| Poor housing quality | NA | 4641 (39.1) | NA | NA |
| Utility insecurity | NA | 2205 (19.1) | NA | NA |
| Unreliable transportation | NA | 1919 (16.6) | NA | NA |
| Housing insecurity | NA | 1663 (13.7) | NA | NA |
| Loneliness | NA | 1493 (13.3) | NA | NA |
Abbreviations: HRSN, health-related social need; NA, not applicable; SMD, standardized mean difference.
Data are presented as the number (percentage) of beneficiaries unless otherwise indicated.
Standardized mean differences greater than 0.25 were considered to be meaningful.23
Race and ethnicity were assessed according to the Centers for Medicare & Medicaid Services beneficiary race code, which reflects data self-reported to the Social Security Administration. The “other” category includes Asian, Hispanic, North American Native, and other race or ethnicity.
Among respondents to each HRSN.
Unadjusted results for quality measures and utilization are reported in Table 2. Hypoglycemic and statin medication adherence and HbA1c control were better in the HRSN-negative group. In the HRSN-positive and HRSN-negative groups, 95.1% and 95.7%, respectively, were receiving an annual HbA1c test; 97.4% and 97.5%, respectively, were receiving annual kidney laboratory tests; and 99.6% and 99.6%, respectively, were receiving annual eye examinations. Rates of all utilization measures were higher in the HRSN-positive group. Compared with the HRSN-negative group, beneficiaries with at least 1 HRSN had a higher rate of hospitalization (17.9% vs 13.7%; P < .001), hospital readmission (2.0% vs 1.3%; P < .001), all-cause avoidable hospitalization (5.2% vs 2.8%; P < .001), and ED visits (36.5% vs 26.3%, P < .001).
Table 2. Quality Measures and Utilization by Presence of HRSNs Among Beneficiaries With Type 2 Diabetes Enrolled in Medicare Advantage.
| Measure | Beneficiariesa | P valueb | ||
|---|---|---|---|---|
| Overall (n = 21 528) | ≥1 HRSN (n = 12 252) | No HRSN (n = 9276) | ||
| Quality measures | ||||
| Hypoglycemic medication adherence | ||||
| Prescription fill | 15 480 (71.9) | 8872 (72.4) | 6608 (71.2) | .06 |
| Adherent to medication | 11 997 (77.5) | 6742 (76.0) | 5255 (79.5) | <.001 |
| Hypoglycemics PDC, mean (SD) | 0.86 (0.23) | 0.85 (0.23) | 0.87 (0.22) | <.001 |
| Statin medication adherence | ||||
| Prescription fill | 16 879 (78.4) | 9702 (79.2) | 7177 (77.4) | .001 |
| Adherent to medication | 12 510 (74.1) | 6999 (72.1) | 5511 (76.8) | <.001 |
| Statins PDC, mean (SD) | 0.84 (0.23) | 0.83 (0.23) | 0.86 (0.22) | <.001 |
| Annual HbA1c test completed | 20 537 (95.4) | 11 656 (95.1) | 8881 (95.7) | .04 |
| HbA1c test result available | 14 830 (68.9) | 8360 (68.2) | 6470 (69.7) | .02 |
| HbA1c result, mean (SD), % | 7.06 (3.35) | 7.17 (4.03) | 6.91 (2.15) | <.001 |
| HbA1c controlledc | 8903 (60.0) | 4834 (57.8) | 4069 (62.9) | <.001 |
| Annual kidney laboratory test complete | 20 987 (97.5) | 11 939 (97.4) | 9048 (97.5) | .65 |
| Annual eye examination complete | 21441 (99.6) | 12 200 (99.6) | 9241 (99.6) | .59 |
| Utilization measures | ||||
| Any hospitalization | 3467 (16.1) | 2198 (17.9) | 1269 (13.7) | <.001 |
| Hospitalizations, mean (SD), No. | 0.22 (0.60) | 0.25 (0.66) | 0.18 (0.51) | <.001 |
| Hospital length of stay, mean (SD), d | 1.40 (5.71) | 1.68 (6.70) | 1.03 (4.01) | <.001 |
| Hospital readmission in 30 d | 360 (1.7) | 243 (2.0) | 117 (1.3) | <.001 |
| All-cause avoidable hospitalization | 890 (4.1) | 634 (5.2) | 256 (2.8) | <.001 |
| Diabetes avoidable hospitalization | 212 (1.0) | 156 (1.3) | 56 (0.6) | <.001 |
| Any emergency visit with discharge | 6913 (32.1) | 4478 (36.5) | 2435 (26.3) | <.001 |
| Emergency visits with discharge, mean (SD), No. | 0.81 (2.04) | 0.97 (2.30) | 0.60 (1.62) | <.001 |
Abbreviations: HbA1c, glycated hemoglobin; HRSN, health-related social need; PDC, proportion of days covered.
SI conversion factor: To convert percentage of total hemoglobin to proportion of total hemoglobin, multiply by 0.01.
Data are presented as the number (percentage) of beneficiaries unless otherwise indicated.
P values reflect results of independent-samples t tests for continuous variables and χ2 tests for categorical measures.
Controlled HbA1c was defined as an HbA1c less than 7% of total hemoglobin.
The results of the adjusted models examining the associations between individual HRSNs and quality measures are shown in Table 3. Financial strain was associated with lower odds of statin adherence (odds ratio [OR], 0.91; 95% CI, 0.83-1.00) and having controlled HbA1c (OR, 0.83; 95% CI, 0.76-0.91). Loneliness and unreliable transportation were associated with both lower diabetes medication adherence (loneliness: OR, 0.85; 95% CI, 0.73-0.99; transportation: OR, 0.80; 95% CI, 0.69-0.92) and statin medication adherence (loneliness: OR, 0.79; 95% CI, 0.69-0.92; transportation: OR, 0.80; 95% CI, 0.70-0.91). Additionally, utility insecurity (OR, 0.86; 95% CI, 0.76-0.98) and housing insecurity (OR, 0.78; 95% CI, 0.67-0.91) were associated with lower odds of diabetes medication adherence.
Table 3. Logistic Regression Results of Quality Outcomes Among Beneficiaries With Type 2 Diabetes Enrolled in Medicare Advantage.
| Measure | Odds ratio (95% CI) | ||
|---|---|---|---|
| Diabetes PDC >0.8 | Statin PDC >0.8 | Controlled HbA1c | |
| Age | 1.01 (1.00-1.01)a | 1.01 (1.01-1.02)a | 1.02 (1.02-1.03)a |
| Elixhauser Comorbidity Index score | 1.01 (1.00-1.01)a | 0.97 (0.95-0.99)a | 1.08 (1.06-1.10)a |
| DCSI score | 1.04 (1.00-1.07)a | 1.03 (1.00-1.06) | 0.85 (0.83-0.87)a |
| Sex | |||
| Female | 0.96 (0.88-1.05) | 0.93 (0.85-1.00) | 1.09 (1.01-1.17)a |
| Male | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Race and ethnicityb | |||
| Black | 0.71 (0.64-0.79)a | 0.71 (0.65-0.78)a | 1.07 (0.98-1.18) |
| White | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Other | 0.84 (0.69-1.01) | 0.76 (0.64-0.91)a | 0.96 (0.81-1.14) |
| Social Security disability eligible | |||
| Yes | 0.97 (0.86-1.08) | 1.01 (0.91-1.12) | 0.97 (0.88-1.07) |
| No disability | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Low-income subsidy eligible | |||
| Yes | 1.21 (1.03-1.42)a | 1.08 (0.94-1.25) | 0.87 (0.76-1.00) |
| No | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Medicare-Medicaid dual eligible | |||
| Yes | 1.11 (0.93-1.32) | 1.08 (0.92-1.26) | 1.06 (1.00-1.36) |
| No | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Region | |||
| Northeast | 0.96 (0.75-1.24) | 1.14 (0.89-1.44) | 1.03 (0.79-1.35) |
| Midwest | 1.04 (0.94-1.16) | 1.01 (0.92-1.12) | 0.78 (0.70-0.86)a |
| South | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| West | 0.90 (0.79-1.04) | 0.90 (0.80-1.03) | 0.87 (0.77-0.99)a |
| Population density | |||
| Suburban | 1.06 (0.95-1.18) | 1.02 (0.93-1.12) | 0.92 (0.84-1.01) |
| Rural | 1.10 (0.96-1.28) | 1.17 (1.02-1.34)a | 0.90 (0.78-1.03) |
| Urban | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Practitioner risk arrangement | |||
| Part risk | 1.10 (0.99-1.22) | 1.07 (0.97-1.18) | 1.04 (0.94-1.15) |
| Full risk | 1.08 (0.96-1.21) | 1.06 (0.95-1.18) | 1.10 (1.00-1.22) |
| No risk | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| HRSN | |||
| Food insecurity | 0.93 (0.83-1.05) | 1.02 (0.92-1.13) | 1.02 (0.92-1.13) |
| Financial strain | 0.91 (0.82-1.01) | 0.91 (0.83-1.00)a | 0.83 (0.76-0.91)a |
| Loneliness | 0.85 (0.73-0.99)a | 0.79 (0.69-0.92)a | 0.96 (0.83-1.11) |
| Unreliable transportation | 0.80 (0.69-0.92)a | 0.80 (0.70-0.91)a | 0.94 (0.83-1.08) |
| Utility insecurity | 0.86 (0.76-0.98)a | 0.97 (0.86-1.09) | 0.94 (0.83-1.05) |
| Housing insecurity | 0.78 (0.67-0.91)a | 0.96 (0.83-1.11) | 0.92 (0.80-1.07) |
| Poor housing quality | 1.00 (0.90-1.11) | 0.92 (0.84-1.01) | 1.03 (0.94-1.13) |
| No HRSN | 1 [Reference] | 1 [Reference] | 1 [Reference] |
Abbreviations: DCSI, Diabetes Complications Severity Index; HbA1c, glycated hemoglobin; HRSN, health-related social need; PDC, proportion of days covered.
Statistical significance at P < .05.
Race and ethnicity were assessed according to the Centers for Medicare & Medicaid Services beneficiary race code, which reflects self-reported data to the Social Security Administration. The “other” category includes Asian, Hispanic, North American Native, and other race or ethnicity.
Adjusted models examining associations between individual HRSNs and utilization outcomes, reported in Table 4 with full results in eTable 3 in Supplement 1, showed that food insecurity was associated with significantly more ED visits (84.6 per 1000 beneficiaries; 95% CI, 19.8-149.4 per 1000 beneficiaries), inpatient encounters (30.4 per 1000 beneficiaries; 95% CI, 9.5-51.3 per 1000 beneficiaries), and avoidable hospitalizations (17.1 per 1000 beneficiaries; 95% CI, 4.7-29.5 per 1000 beneficiaries). Loneliness was associated with an additional 173.0 ED visits per 1000 beneficiaries (95% CI, 74.2-271.9 per 1000 beneficiaries); unreliable transportation was associated both with more ED visits (244.6 per 1000 beneficiaries; 95% CI, 150.4-338.9 per 1000 beneficiaries) and with more inpatient encounters (41.8 per 1000 beneficiaries; 95% CI, 13.5-68.3 per 1000 beneficiaries). In contrast, utility insecurity, financial strain, poor housing quality, and housing insecurity were not significantly associated with any of the utilization measures.
Table 4. Association Between HRSNs and Utilization Among Beneficiaries With Type 2 Diabetes Enrolled in Medicare Advantage.
| HRSN | Marginal effects of HRSNs on rates of utilization, per 1000 beneficiaries (95% CI)a | |||
|---|---|---|---|---|
| Avoidable hospitalization | ED visit | Inpatient encounter | 30-d Inpatient readmission | |
| Food insecurity | 17.1 (4.7 to 29.5)b | 84.6 (19.8 to 149.4)b | 30.4 (9.5 to 51.3)b | 8.2 (−1.2 to 17.5) |
| Financial strain | 4.6 (−6.4 to 15.6) | 40.0 (−16.9 to 96.9) | 6.8 (−12.0 to 25.5) | 2.3 (−5.8 to 10.4) |
| Loneliness | 3.9 (−11.8 to 19.7) | 173.0 (74.2 to 271.9)b | −6.3 (−32.1 to 19.5) | −4.5 (−14.5 to 5.4) |
| Unreliable transportation | 13.5 (−1.6 to 28.6) | 244.6 (150.4 to 338.9)b | 41.8 (13.5 to 68.3)b | 1.0 (−9.3 to 11.3) |
| Utility insecurity | 0.00 (−13.2 to 13.2) | −15.7 (−88.5 to 57.2) | −11.5 (−33.9 to 11.0) | −6.3 (−15.1 to 2.6) |
| Housing insecurity | 6.3 (−10.0 to 22.5) | 55.4 (−38.0 to 148.8) | 16.1 (−12.4 to 44.5) | 10.2 (−3.6 to 24.0) |
| Poor housing quality | −2.7 (−13.4 to 8.0) | 1.1 (−57.7 to 60.0) | −15.0 (−33.3 to 3.2) | −1.6 (−9.6 to 6.3) |
Abbreviations: ED, emergency department; HRSN, health-related social need.
Calculated from a negative binomial regression model estimating the association between HRSNs and utilization measures. Models were adjusted for age, sex, race and ethnicity, disability, low-income subsidy, dual eligibility, geographic region, population density, Elixhauser Comorbidity Index score, and Diabetes Complications Severity Index score. The reference group for all marginal effects was beneficiaries not reporting that specific HRSN.
Statistical significance at P < .05.
Discussion
Using self-reported HRSN data from a national sample of Medicare Advantage beneficiaries with T2D, we found that the majority (56.9%) of beneficiaries reported at least 1 HRSN. This proportion is higher than the 49.3% prevalence of HRSNs in a broader, non–disease-specific Medicare Advantage population,16 though the previous study did not include beneficiaries younger than 65 years. In the current study, the most prevalent HRSNs in the population with T2D were the same as in the non–disease-specific group: financial strain, food insecurity, and poor housing quality. Loneliness and unreliable transportation were each associated with both lower medication adherence and increased acute care utilization outcomes. Financial strain, utility insecurity, and housing insecurity were each associated with lower medication adherence, while financial strain was the only HRSN associated with controlled HbA1c. Among the utilization measures, emergency visits were associated with the highest number of individual HRSNs (food insecurity, loneliness, and unreliable transportation). Food insecurity was the HRSN most consistently associated with higher utilization (avoidable hospitalization, all-cause hospitalization, and ED visits).
We found a high prevalence of HRSNs among beneficiaries with T2D and significant associations between individual HRSNs and measures of quality of care and health care utilization. While we were unable to determine the temporal association between development of HRSNs and T2D in this cross-sectional study, prior studies4,5,6 have shown that poor diet, inability to access and/or afford medical care, and lack of social support are associated with the development and management of T2D; managing and treating diabetes also may be associated with worse social adversity. For example, healthy diets are an important component of T2D management, and our findings showed that food insecurity, uniquely, was consistently associated with increased acute care utilization. Poor diet or food insecurity is associated with development of T2D, increased likelihood of complications, and increased acute care utilization events.24,25 In parallel, the physical and economic burdens of living with T2D may make it harder to afford, access, and prepare health-promoting foods.26 This bidirectionality underscores the potential impact of clinical, community, and policy interventions that can reduce social adversity for individuals and populations.
Several HRSN screening tools have been developed over the past few years to support clinicians and other health care practitioners to assess patient HRSNs.15,27 Our study findings underscore that using these tools to better understand social needs may be a key component for improving care for patients with T2D. In theory, collected information might then be used by clinicians and health care delivery organizations to improve care.28 For instance, for patients who screen positive for an HRSN, practitioners can adjust care decisions to reduce the impact of the unmet need, help to connect to appropriate social services, or advocate for more resources.28,29,30 Community-based organizations and health systems can coordinate activities and cross-sector communication to expand opportunities for community members to reduce HRSNs and increase opportunities for healthy lifestyles. Health policy changes may include expanding innovative payment models for practitioners of care to geographically and socioeconomically disadvantaged populations,31 incorporating support services professionals such as community health workers into the care team,32 and increasing flexibility for health plans to offer supplemental or value-added benefits to address the HRSNs of beneficiaries with T2D.
Limitations
Several limitations should be considered when interpreting these findings. The HRSN measures are derived from a self-report survey and subject to biases of self-report and survey nonresponse; though evidence of response bias was not found on observable characteristics among the broader survey population, it may impact HRSN prevalence.16 The survey was offered only in English and Spanish, so it may underreport racial and ethnic minority groups, and race and ethnicity data were limited. No causal inference can be ascertained from this study, as it was an observational study using retrospective claims data. Moreover, the associations we found may be affected by unmeasured HRSNs or other explanatory factors. Limitations common to studies using administrative claims data apply to this study. For instance, we were not able to include data about potentially relevant biomarkers or health behaviors (eg, weight and health behavior information) and we could not correct for claims coding errors. Additionally, the HRSN survey was completed from October 2019 to February 2020; however, we linked the survey data to utilization data from 2019, prior to survey completion, because of the consequences the COVID-19 pandemic had for health care utilization in 2020. Given that HRSNs are typically not static, it is possible that HRSN status captured in the surveys would have been different over the utilization period used in the analysis. As this study used data from 1 Medicare Advantage health plan, though large and national, the results may not be generalizable to the broader population.
Conclusions
In this cross-sectional study of a large HRSN survey data set of Medicare Advantage beneficiaries with a diagnosis of T2D, most beneficiaries reported having at least 1 HRSN, and some HRSNs were associated with lower care quality and increased utilization. Organizations developing both federal and state standards are actively developing policies related to patient-level interventions to screen for and address HRSNs.33,34,35 Findings from this study may inform future policies and decisions about which HRSNs to screen for and to intervene on for patients living with T2D, maximizing the likelihood that evolving programs address the needs most associated with T2D health outcomes and utilization.
eMethods. Health Related Social Needs (HRSNs) Screening Instrument
eTable 1. Demographic and Clinical Characteristics by Completion of HRSN Survey Items
eTable 2. Correlation Between Individual Health-Related Social Needs
eTable 3. Negative Binomial Regression Results of Utilization Outcomes
Data Sharing Statement
References
- 1.Centers for Medicare & Medicaid Services . Opportunities in Medicaid and CHIP to address social determinants of health (SDOH). Accessed February 25, 2021. https://www.medicaid.gov/federal-policy-guidance/downloads/sho21001.pdf
- 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(suppl 2):19-31. doi: 10.1177/00333549141291S206 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hill-Briggs F, Adler NE, Berkowitz SA, et al. Social determinants of health and diabetes: a scientific review. Diabetes Care. 2020;44(1):258-279. doi: 10.2337/dci20-0053 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Walker RJ, Smalls BL, Campbell JA, Strom Williams JL, Egede LE. Impact of social determinants of health on outcomes for type 2 diabetes: a systematic review. Endocrine. 2014;47(1):29-48. doi: 10.1007/s12020-014-0195-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Berkowitz SA, Meigs JB, DeWalt D, et al. Material need insecurities, control of diabetes mellitus, and use of health care resources: results of the Measuring Economic Insecurity in Diabetes study. JAMA Intern Med. 2015;175(2):257-265. doi: 10.1001/jamainternmed.2014.6888 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Jiang DH, O’Connor PJ, Huguet N, Golden SH, McCoy RG. Modernizing diabetes care quality measures. Health Aff (Millwood). 2022;41(7):955-962. doi: 10.1377/hlthaff.2022.00233 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Centers for Disease Control and Prevention . National Diabetes Statistics Report 2020. Accessed July 14, 2021. https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf
- 8.American Diabetes Association . Economic costs of diabetes in the US in 2017. Diabetes Care. 2018;41(5):917-928. doi: 10.2337/dci18-0007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gold R, Kaufmann J, Gottlieb LM, et al. Cross-sectional associations: social risks and diabetes care quality, outcomes. Am J Prev Med. 2022;63(3):392-402. doi: 10.1016/j.amepre.2022.03.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Uppal TS, Chehal PK, Fernandes G, et al. Trends and variations in emergency department use associated with diabetes in the US by sociodemographic factors, 2008-2017. JAMA Netw Open. 2022;5(5):e2213867. doi: 10.1001/jamanetworkopen.2022.13867 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cottrell EK, Hendricks M, Dambrun K, et al. Comparison of community-level and patient-level social risk data in a network of community health centers. JAMA Netw Open. 2020;3(10):e2016852. doi: 10.1001/jamanetworkopen.2020.16852 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chen M, Tan X, Padman R. Social determinants of health in electronic health records and their impact on analysis and risk prediction: a systematic review. J Am Med Inform Assoc. 2020;27(11):1764-1773. doi: 10.1093/jamia/ocaa143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zulman DM, Maciejewski ML, Grubber JM, et al. Patient-reported social and behavioral determinants of health and estimated risk of hospitalization in high-risk Veterans Affairs patients. JAMA Netw Open. 2020;3(10):e2021457. doi: 10.1001/jamanetworkopen.2020.21457 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Canterberry M, Figueroa JF, Long CL, et al. Association between self-reported health-related social needs and acute care utilization among older adults enrolled in Medicare Advantage. JAMA Health Forum. 2022;3(7):e221874. doi: 10.1001/jamahealthforum.2022.1874 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Billioux A, Verlander K, Anthony S, Alley D. Standardized screening for health-related social needs in clinical settings: the Accountable Health Communities screening tool. Discussion paper. NAM Perspect. Published online May 30, 2017. doi: 10.31478/201705b [DOI] [Google Scholar]
- 16.Long CL, Franklin SM, Hagan AS, et al. Health-related social needs among older adults enrolled in Medicare Advantage. Health Aff (Millwood). 2022;41(4):557-562. doi: 10.1377/hlthaff.2021.01547 [DOI] [PubMed] [Google Scholar]
- 17.Young BA, Lin E, Von Korff M, et al. Diabetes complications severity index and risk of mortality, hospitalization, and healthcare utilization. Am J Manag Care. 2008;14(1):15-23. [PMC free article] [PubMed] [Google Scholar]
- 18.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. doi: 10.1097/00005650-199801000-00004 [DOI] [PubMed] [Google Scholar]
- 19.Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53(12):1258-1267. doi: 10.1016/S0895-4356(00)00256-0 [DOI] [PubMed] [Google Scholar]
- 20.Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-1139. doi: 10.1097/01.mlr.0000182534.19832.83 [DOI] [PubMed] [Google Scholar]
- 21.National Committee for Quality Assurance . Comprehensive diabetes care (CDC). Accessed August 30, 2022. https://www.ncqa.org/hedis/measures/comprehensive-diabetes-care/
- 22.Agency for Healthcare Research and Quality . Guide to Prevention Quality Indicators: Hospital Admission for Ambulatory Care Sensitive Conditions. US Department of Health and Human Services; 2001. [Google Scholar]
- 23.Stuart EA. Matching methods for causal inference: a review and a look forward. Stat Sci. 2010;25(1):1-21. doi: 10.1214/09-STS313 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.López A, Seligman HK. Clinical management of food-insecure individuals with diabetes. Diabetes Spectr. 2012;25(1):14-18. doi: 10.2337/diaspect.25.1.14 [DOI] [Google Scholar]
- 25.Gucciardi E, Vahabi M, Norris N, Del Monte JP, Farnum C. The intersection between food insecurity and diabetes: a review. Curr Nutr Rep. 2014;3(4):324-332. doi: 10.1007/s13668-014-0104-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Johnson KT, Palakshappa D, Basu S, Seligman H, Berkowitz SA. Examining the bidirectional relationship between food insecurity and healthcare spending. Health Serv Res. 2021;56(5):864-873. doi: 10.1111/1475-6773.13641 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.LaForge K, Gold R, Cottrell E, et al. How 6 organizations developed tools and processes for social determinants of health screening in primary care: an overview. J Ambul Care Manage. 2018;41(1):2-14. doi: 10.1097/JAC.0000000000000221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.National Academies of Sciences, Engineering, and Medicine. Integrating Social Care Into the Delivery of Health Care: Moving Upstream to Improve the Nation’s Health. National Academies Press; 2019. [PubMed]
- 29.Senteio C, Adler-Milstein J, Richardson C, Veinot T. Psychosocial information use for clinical decisions in diabetes care. J Am Med Inform Assoc. 2019;26(8-9):813-824. doi: 10.1093/jamia/ocz053 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hessler D, Bowyer V, Gold R, Shields-Zeeman L, Cottrell E, Gottlieb LM. Bringing social context into diabetes care: intervening on social risks versus providing contextualized care. Curr Diab Rep. 2019;19(6):30. doi: 10.1007/s11892-019-1149-y [DOI] [PubMed] [Google Scholar]
- 31.Centers for Medicare & Medicaid Services . ACO REACH. Accessed August 16, 2022. https://innovation.cms.gov/innovation-models/aco-reach [PubMed]
- 32.Trump LJ, Mendenhall TJ. Community health workers in diabetes care: a systematic review of randomized controlled trials. Fam Syst Health. 2017;35(3):320-340. doi: 10.1037/fsh0000283 [DOI] [PubMed] [Google Scholar]
- 33.Reynolds A. Social need: new HEDIS measure uses electronic data to look at screening, intervention. National Committee for Quality Assurance . November 2, 2022. Accessed Feb 27, 2023. https://www.ncqa.org/blog/social-need-new-hedis-measure-uses-electronic-data-to-look-at-screening-intervention/
- 34.Notices. Fed Regist. 2022;87(208):65207. [Google Scholar]
- 35.Joint Commission . New requirements to reduce health care disparities. R3 Report. June 20, 2022. Accessed February 27, 2023. https://www.jointcommission.org/-/media/tjc/documents/standards/r3-reports/r3_disparities_july2022-6-20-2022.pdf
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods. Health Related Social Needs (HRSNs) Screening Instrument
eTable 1. Demographic and Clinical Characteristics by Completion of HRSN Survey Items
eTable 2. Correlation Between Individual Health-Related Social Needs
eTable 3. Negative Binomial Regression Results of Utilization Outcomes
Data Sharing Statement