Utilization by High-Cost, High-Need Medicaid Patients Receiving Social Worker Care Coordination

ABSTRACT

Objectives: Office-based care coordination programs are increasingly utilizing social work–trained care coordinators to manage high-cost, high-need (HCHN) patients with increased psychosocial risk, including patients covered by Medicaid. Whether utilization of social worker care coordinators (SWCCs) with HCHN Medicaid patients impacts health care utilization is not clear, so we conducted a wait-list randomized controlled trial to test whether social worker care coordination decreases health care utilization and costs in HCHN Medicaid patients.

Study Design: Prospective enrollment with computer randomization 1:1 to an intervention group receiving SWCC office-based care coordination or to a control group.

Methods: HCHN patients (N = 252) were randomly assigned to enrollment in an office-based care coordination program with SWCC engagement or to a wait-list control group. Total numbers of emergency department (ED) visits and hospitalizations and total medical expenditure (TME) within the preceding 12 months were collected at baseline and then again at 1 year and 2 years after study group assignment. Health care utilization and costs were compared using mixed-effects models.

Results: Patients enrolled in an outpatient SWCC program had an increased number of hospitalizations compared with wait-listed control patients. There were no differences in the number of ED visits or in TME between groups.

Conclusions: Enrollment for up to 2 years in an office-based social worker–led care coordination program did not reduce health care utilization or costs in HCHN Medicaid patients.

Am J Manag Care. 2026;32(4):e103-e109. https://doi.org/10.37765/ajmc.2026.89922

Takeaway Points

This randomized controlled trial of 252 adults assessed whether an outpatient social worker–led care coordinator program can reduce health care utilization and costs in high-cost, high-need (HCHN) Medicaid patients.

• Hospitalizations were increased among patients followed by a social worker compared with control patients after 12 months (difference in mean changes, 0.37; SD = 0.18; P = .04) and 24 months (0.41; SD = 0.18; P = .02).
• No differences were observed in emergency department visits or costs.
• Overall, our study found that a social worker–led outpatient care coordination program did not statistically significantly reduce traditional measures of health care utilization or costs in HCHN Medicaid patients.

High-cost, high-need (HCHN) patients have an increased risk of poor health outcomes and preventable costly health care spending.^1,2 Lowering preventable health care utilization and decreasing associated costs have accordingly become important areas of focus among health networks and policy makers, with hospitals and health networks increasingly turning to complex care management and office-based care coordination programs to try to achieve these goals.^3-8

In 2006, Massachusetts General Hospital (MGH) created the Integrated Care Management Program (iCMP) as part of a Medicare demonstration project to address the increasing needs of HCHN patients.^3,9 The iCMP is a primary care–based outpatient care coordination program for HCHN patients designed to assess unmet patient needs, create and implement care plans, educate patients, coordinate patient care, avoid preventable emergency department (ED) visits and hospital admissions, conduct postdischarge follow-up services, and support primary care physicians. The program’s care coordinators are supported by a central management team and ancillary staff including a pharmacist, data support, and community resource specialists. During the initial demonstration project period, registered nurse care coordinators (RNCCs) worked with HCHN Medicare patients with the aim of achieving these programmatic goals. The program was found to effectively reduce preventable health care utilization and health care spending, with details previously published elsewhere.^10-12

Given the initial success of the iCMP Medicare demonstration project—and with the hospital’s increasing realization that HCHN patients also include patients with different risk profiles beyond Medicare patients, such as younger patients and patients with psychosocial risk—the iCMP program expanded to include commercially and Medicaid insured patients along with Medicare patients. With the expansion of the program to additional HCHN patient populations, the program grew to include social worker care coordinators (SWCCs) to work with Medicaid patients, given their different health profiles, risk factors, and sociodemographic resources from Medicare patients; in 2015, the program began using SWCCs to function as primary care coordinators for Medicaid patients with higher psychosocial complexity. Most Medicare patients are 65 years and older and have multiple chronic medical conditions, whereas Medicaid patients tend to be younger, have more psychiatric rather than medical morbidities, and often have social and neighborhood challenges (eg, food insecurity, housing instability, employment- and income-related challenges) driving their poor outcomes. The impact of using SWCCs with Medicaid patients, however, remains unknown, as does whether providing SWCC services to Medicaid patients yields utilization and cost benefits similar to pairing RNCCs with Medicare patients. Prior studies in adult HCHN Medicaid patients that have used mixed staffing models that included social workers in supporting roles or that leveraged community-based partners have shown potential short-term benefits in reducing health care utilization and cost outcomes.^13,14 To empirically assess the benefits of the iCMP for Medicaid patients followed by SWCCs, we thus designed a prospective randomized controlled trial.

This study sought to evaluate the effects of using social workers as the primary iCMP team member providing outpatient care coordination to HCHN Medicaid patients by assessing the impact on traditional HCHN markers of health care utilization and spending.

METHODS

Trial Design

This investigator-conceived and -initiated study was approved as exempt by the hospital’s institutional review board. The study design was a randomized controlled trial to evaluate the impact of using SWCCs as the primary iCMP team member on traditional health care utilization markers—ED visits, hospitalizations, and total medical expenditure (TME)—in HCHN Medicaid patients. The study was conducted at MGH-affiliated primary care practices in greater metropolitan Boston, Massachusetts. Eligible study participants were prospectively enrolled and computer-randomized 1:1 to an intervention or control group. Control participants were wait-listed for enrollment in the program after study conclusion, based on continued clinical need and program capacity. The iCMP staff implemented the study protocol, administered the intervention, and by nature of the intervention were unblinded to study group assignment but were unaware of study results until the study was concluded.

Program Eligibility

This study included patients enrolled in a Medicaid risk contract with the Mass General Brigham (MGB) health care network. Eligible patients were 18 years and older living in Massachusetts, enrolled in either the institution’s Medicaid accountable care organization (ACO) or dually enrolled in the institution’s Medicaid ACO and Medicare ACO plans, followed by an MGH primary care physician, and identified as HCHN by MGH from April 2019 through August 2021. Patients enrolled in a MassHealth Community Partners program or in any other insurance-based care management program were not eligible for enrollment as per MGH ACO policy. Eligible HCHN patients were identified by MGB through the use of a machine learning algorithm that incorporates information on demographics (eg, age, sex), heath status (≥ 20 high-risk complex conditions), health care utilization in the 12 months prior, and social determinants of health information (eg, smoking, area-level poverty).⁹

Participants were excluded if they resided out of state, had no contact information, or were already participating in another insurance-provided care management program or another Medicaid demonstration program.

Intervention

The MGH iCMP provides office-based care coordination to HCHN risk-contract patients in coordination with primary care practices through a team-based approach with specialized staff including nurses, social workers, community health workers, pharmacists, and resources specialists. A primary care coordinator is assigned based on the patient’s primary risk driver (medical, psychiatric, socioenvironmental). For this study, Medicaid patients were assigned an SWCC based on the known high prevalence of psychosocial risk factors in this population.¹⁵ Medicaid patients assigned to an SWCC who have accompanying medical complexity are also able to receive iCMP nurse care coordinator support to assist with comanagement of medical risk factors. Enrolled patients undergo a comprehensive intake assessment that checks for multidomain barriers to care, including assessment of social determinants of health and physical and mental health needs, resulting in an individualized care plan. The SWCC provides longitudinal in-person and telephonic care coordination, supported as needed by additional iCMP team members, that includes routine and ad hoc check-ins at least once every 6 months, coordination with the patient’s mental health and substance use teams, trauma-informed support and coordination, postdischarge assessments, pharmacy support, caregiver needs support, and ad hoc needs assessments.

Data Sources

Information on age, sex, race, ethnicity, primary language, marital status, employment status, and town of residence was collected from patients’ electronic health records. Health care utilization data were collected from MGB’s electronic health database and included data from all hospitals participating in the network. Medical claims data and/or health network payment data were collected from MGB’s enterprise database and used for TME calculations.

Outcomes

We used a combination of clinical and financial data for our primary outcomes analyses. Primary study outcomes were ED visits, hospitalizations, and TME. Total counts over the preceding 12 months for all outcomes were collected separately for each year in both the intervention and control groups at each time point (baseline, 12 months, and 24 months).

Health care utilization. Utilization outcomes included ED visits and hospital admissions at all MGB network hospitals. MGB is the largest health system in Massachusetts and includes 16 institutions consisting of academic medical centers (including MGH), specialty centers, and community hospitals. Hospitalizations were categorized as medical-surgical vs psychiatric, with medical-surgical hospital admissions further categorized using an ambulatory care–sensitive conditions (ACSCs) classification to determine whether hospitalizations were avoidable. We identified avoidable hospitalizations using a systemic approach that included 16 previously validated conditions, as determined by primary admission diagnosis code.^16,17 ACSCs were angina, appendicitis (complicated/ruptured), asthma, cellulitis, chronic obstructive pulmonary disease, congestive heart failure, dehydration, diabetes, gangrene, hypokalemia, hypertension (malignant), pneumonia, pyelonephritis, ulcer (perforated/bleeding), urinary tract infection, and vaccine-preventable conditions.

Medical spending. TME included all services delivered both inside and outside the MGB system within the index year paid by Medicaid, including all inpatient and outpatient costs, ambulance costs, and laboratory costs. Retail pharmacy costs were not included as part of TME.

Mental health. Mental health measures for depression and anxiety were collected using the following validated screening instruments: the 9-Item Patient Health Questionnaire (PHQ-9) score for depression and the 7-Item Generalized Anxiety Disorder questionnaire (GAD-7) score for anxiety. Patients followed at MGH primary care practices are asked to complete the PHQ-9 and GAD-7 mental health screeners prior to their annual health visits. The screeners may also be administered by clinicians during medical visits, ED visits, and hospitalizations based on clinical concerns.

Statistical Analysis

Descriptive statistics are provided for the study population (Table 1). Longitudinal linear mixed-effects models were applied to examine the effect of the intervention on number of ED visits, number of hospitalizations, and TME at specific times. The model included a random intercept and fixed effects of the study group (ie, intervention vs control, discrete time [0, 12, and 24 months], and group*time interaction). Baseline mental health data were added as covariates in the mixed-effects models. For hospitalization, we also ran separate models for medical-surgical and psychiatric admissions. To assess for any differences between study groups over time in frequency of avoidable admissions, we conducted a subanalysis limited to ACSC hospitalizations utilizing a generalized linear model with logit link using generalized estimating equations. To assess for postrandomization confounding given the differences in study group enrollment after randomization, we compared sociodemographic covariates by study group.

Repeated measures of mental health data, including a baseline and at least 1 follow-up value, were missing for most patients, so separate subanalyses of mental health outcome measures were conducted, given the potential of social work support for mental health.

RESULTS

We received a list of 301 HCHN patients enrolled in Medicaid from MGB with a risk score identifying them as iCMP eligible; they were then randomly assigned to a study group. Forty-nine patients randomly assigned to the intervention group were ultimately not enrolled in the program due to the following reasons: unreachable status, patient refusal, medical complexity otherwise not appropriate for SWCC management, serious mental illness, out-of-state relocation, incarceration, and deceased. The final sample for analysis included 252 patients, with 154 in the control group and 98 in the intervention group. All 252 patients assigned to a study group completed the study and were included in data analyses. Baseline demographics were similar among study groups (Table 1). Patients had a mean age of 50 years; were predominantly White, non-Hispanic, and English speaking; and were evenly distributed by sex. Approximately half had at least 1 identified social determinant of health need. Analyses comparing sociodemographic factors by study group to test for postrandomization confounding found no statistically significant differences (data not shown).

Health Care Utilization

Baseline traits. Utilization of ED services (1.9 vs 1.4; P = .2) and hospitalizations (0.9 vs 0.7; P = .2) were similar among control and intervention participants at baseline, as was TME ($21,601 vs $17,962; P = .5) (Table 1).

ED visit rates. Over 2 years of follow-up, HCHN patients assigned to an SWCC had no differences in the number of ED visits after 12 months (P = .5) or 24 months (P = .8) of enrollment in the program compared with controls (Table 2 and Figure). Findings after 12 and 24 months did not change significantly when baseline mental health data were added (data not shown).

Hospitalization rates. Compared with control participants, patients assigned a high-risk SWCC experienced increased hospitalizations at both 12 months (difference in mean changes, 0.37; SD = 0.18; P = .04) and at 24 months (0.41; SD = 0.18; P = .02) of enrollment (Table 2 and Figure). Hospitalizations categorized as medical-surgical increased in intervention patients compared with controls at both 12 months (0.36; SD = 0.17; P = .03) and 24 months (0.35; SD = 0.17; P = .04), but the number of psychiatric admissions remained similar for both groups at 12 months (0.02; SD = 0.06; P = .8) and 24 months (0.06; SD = 0.06; P = .3). Twelve- and 24-month findings did not change significantly when baseline mental health data were added (data not shown). We found no differences in ACSC admissions between study groups at 12 or 24 months (0.078; 95% CI, –1.4 to 1.6; P = .9).

Medical Spending

Over 2 years of follow-up, there were no significant differences in TME between control patients and patients assigned to an SWCC at either 12 months (P = .1) or 24 months (P = .6) (Table 2 and Figure). Findings after 12 and 24 months did not change significantly when baseline mental health data were added (data not shown).

Mental Health

In subanalyses of mental health outcomes analyzing 2 years of follow-up data, we found no difference in the number of patients who screened positive for anxiety (P = .50) or depression (P = .14) based on enrollment status in the high-risk SWCC program (Table 3).

DISCUSSION

In this randomized controlled trial, we found that HCHN Medicaid patients assigned to receive office-based SWCC services had a slight increase in medical hospitalizations and no improvement in ED utilization or cost outcomes after 24 months of enrollment in the program compared with similar-risk control patients who did not receive SWCC services. This equates to less than 1 (0.42) additional hospitalization over 2 years for each patient participating in a social worker care coordination program compared with similar-risk patients not followed in the program. This counterintuitive finding of possible increased health care utilization is surprising because it differs from the results of a prior intervention study showing benefits for Medicare patients receiving outpatient nurse care coordinator services as well as a retrospective cohort study showing benefits for commercially insured patients, which both also examined patients enrolled in the iCMP.^10,18 Of importance, however, this study is the first to prospectively and empirically test the program’s impact of using SWCCs on health care outcomes for HCHN Medicaid patients, likely accounting for the divergent findings.

The unexpected finding of increased hospitalizations could paradoxically be an indication of appropriate care delivery if prior to enrollment in a high-risk program patients were experiencing barriers to care or underutilization of health care services, as has been previously documented in HCHN populations.^10,13 Accordingly, the increased hospitalization rates observed after 1 and 2 years of enrollment in a care coordination program could represent catch-up care rather than avoidable health care utilization. After working with an SWCC, patients previously lost to follow-up or nonadherent to medical recommendations could in turn be receiving more medical care through improved coordination of care with specialists, being better informed about symptoms that require immediate medical attention, and being more willing to undergo hospitalization to receive needed medical care. The noted increase in medical and surgical hospitalizations but not psychiatric admissions in a population of patients with a high burden of psychosocial risk is consistent with this hypothesis. Furthermore, the finding of increased medical and surgical hospital admissions without an increase in avoidable admissions or in ED visits lends further credence to an appropriate increase in utilization, as frequent ED visits are often an indication of preventable utilization in lieu of primary preventive care.

Interestingly, our findings contrast with those from prior studies of the same program, although different populations were examined. Initial studies of the iCMP Medicare demonstration project showed clear benefits to using RNCCs to manage medically complex HCHN Medicare patients.¹⁰ A prior retrospective cohort study further showed the program’s benefits in lowering costs among commercially insured iCMP patients.¹⁸ Importantly, however, neither of these studies assessed programmatic impact on Medicaid patients or isolated the impact on patients followed primarily by SWCCs.

Our findings are also in contrast to studies assessing outcomes of care coordination programs utilizing social workers or involving Medicaid patients, which may be a result of differences in program structure and study design. The Camden Core Model, which was found to decrease hospital readmissions, utilized a mixed staffing model that included nurses, community health workers, and social workers to comanage patients with both medical and social complexity,¹⁹ whereas our study assigned SWCCs as the primary team member focused on caring for patients with predominantly high psychosocial risk. Study outcomes were also different, as our study assessed ED visits, hospitalizations, and TME, whereas Finkelstein and colleagues focused on assessing hospital readmission outcomes.¹⁹ Similarly, a study embedding social workers in rural primary care teams found slight benefits in decreasing ED visits and hospitalizations at affiliated Veterans Health Administration hospitals.²⁰ In contrast to our study, social workers in the veteran study played a supporting role rather than leading care coordination, patients were older, and outcomes were only followed up to 60 days. The CareMore Health (now Carelon Health) study randomly assigned HCHN Medicaid patients to a mixed staffing model that included social workers in a supporting role and showed reductions in TME, hospitalizations, and inpatient bed days; however, community health workers functioned as the primary care coordinators in that model, and outcomes were only measured up to 12 months.¹³

The possibility also exists that the traditional utilization outcomes selected in this study that are focused on health care utilization and cost may not be the optimal outcomes to properly and fully ascertain programmatic benefits in the Medicaid population. Much of the literature on care coordination and health care utilization benefits comes from Medicare populations followed in care coordination programs staffed with nurses.^10,21 Social worker input on psychosocial issues may be too indirect or insufficiently valued relative to other health care roles by patients to impact cost or traditional health care utilization outcomes.²² Medicaid patients may represent a population with a unique health risk profile that requires a different set of assessment tools. Outcomes that assess patient well-being or agency, such as self-reported health, health literacy, self-efficacy in managing one’s health care, or satisfaction with health care delivery, may be better barometers of the impact of delivering office-based social worker–led care coordination to patients with multiple upstream sociodemographic complexities. Similarly, measures of mental health may also be better markers of health status improvement in this population rather than traditional measures of utilization, which are better designed to assess medical complexity. Future studies should consider including outcome measures better suited for assessing health status improvement in patients with primarily sociodemographic and psychosocial complexity.

Although we found that enrollment in a social worker–led care coordination program did not improve health care utilization outcomes in HCHN Medicaid patients, it is important to appreciate that the drivers of utilization are likely upstream factors in this population, including housing instability and social stressors, which typically involve long-term solutions that lie outside the medical domain and are unlikely to be modified in any significant way by a relatively light-touch intervention housed within a hospital-based health care delivery system. Although our study did not find that social worker–led care coordination reduced utilization or cost, it remains plausible that it provided other unmeasured benefits. Health care organizations may therefore and nonetheless still consider providing social worker–led care coordination for psychosocially complex patients who have not benefitted from traditional health care interventions aimed at cost containment.

Fidelity to the SWCC model was monitored monthly using programmatic key performance indicators, reducing the potential for inconsistent program delivery as a cause of the intervention’s lack of effectiveness in reducing traditional health care utilization outcomes. Universal programmatic metrics collected on all patients enrolled in iCMP, including SWCC patients, included rate of care plan completion, frequency of care plan review, average time to completion of high-risk assessment, care coordinator patient panel size, number of patients enrolled and discharged, and number of patients without any contact within the past 3 months.

Limitations

In addition to considerations about outcome measure suitability, this study has several other limitations. Nearly one-third of patients randomly assigned to the intervention arm did not enroll in the program, leading to possible postrandomization confounding. Analyses by sociodemographic categories did not reveal any differences by study group; however, unmeasured residual confounding remains a possibility. This study included only utilization information from institutions within the MGB network and did not capture ED visits and hospitalizations occurring outside the network. Differences in utilization patterns outside the network by study group, such as greater external health care utilization among control patients who were not receiving care coordination services, could account for the observed differences in utilization patterns between study groups. Although control patients were not eligible for enrollment in iCMP during the study period, they may have participated in other care coordination or care management programs, which we did not collect data on and may have accounted for improved utilization measures among controls.

Limited available data in the medical records on mental health outcomes leading to substantial missing data make it difficult to draw definitive conclusions about programmatic impact on mental health. Different medical practices within the health network may have different protocols for when and how to collect mental health screening questionnaire data from patients, likely accounting for the generally low amounts of data availability for the PHQ-9 and GAD-7 scores. Although our hope was to obtain a better understanding of the programmatic impact on mental health in HCHN Medicaid patients, the randomized study design allays concerns about potential confounding by mental health conditions on the main health care utilization study outcomes. The mental health findings were exploratory and underpowered, and future fully powered prospective studies with targeted outcomes are needed to better understand the impact of SWCCs on mental health outcomes.

Not all patients randomly assigned to the intervention arm were ultimately enrolled in the program, so the final intervention group may represent enrollment bias, which may account for slight baseline differences between study groups. Although the groups may have differed at baseline, the randomized study design and the reporting of real-world outcomes demonstrate the true impact of rolling out an office-based social worker care coordination program. This per-protocol study was designed to assess the impact of an existing clinical program over the course of 24 months, a time span that captures the enrollment period of a majority of enrolled patients in the program, to allow for maximum power with outcomes assessment. Because the study included only patients enrolled in the clinical program and did not proactively recruit participants, a priori power calculations were not conducted, and it is possible that the study’s negative findings could be due to inadequate power.This study included patients followed within one urban academic medical center in the Northeastern US, so findings may not be broadly generalizable to all Medicaid populations. Although we analyzed utilization data by ACSCs to assess for the program’s impact on avoidable hospitalizations, data were collected in part during the COVID-19 pandemic, when many symptoms and diagnostic categories previously classified as ambulatory, including viral respiratory infections, may have been less clearly delineated as requiring ambulatory vs inpatient care.

CONCLUSIONS

We found that enrollment of HCHN Medicaid patients in an office-based social worker–led care coordination program did not improve traditional markers of health care utilization or spending. The finding of an increase in overall hospital admissions but not in ACSCs, and without accompanying increases in ED visits or costs, may indicate an appropriate increase in complexity-driven care utilization. Identifying impactful outcome measures to accurately evaluate programmatic impact in Medicaid patients is essential to better assess and quantify the potential benefits of providing outpatient care coordination to this high-risk population.

Author Affiliations: Division of General Internal Medicine, Department of Medicine (NMO, AEE, ZAA), Biostatistics Center (HL), and Division of Gynecologic Oncology, Department of Obstetrics and Gynecology (MGdC), Massachusetts General Hospital, Boston, MA; Harvard Medical School (NMO, HL, MGdC), Boston, MA.

Source of Funding: None.

Author Disclosures: Dr Oreskovic is employed by Massachusetts General Hospital as a physician and the Integrated Care Management Program medical director. Mr Allon is employed by Massachusetts General Hospital and works with the Integrated Care Management Program. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (NMO, AEE, HL, MGdC); acquisition of data (NMO, AEE, ZAA); analysis and interpretation of data (NMO, AEE, ZAA, HL); drafting of the manuscript (NMO, AEE, ZAA); critical revision of the manuscript for important intellectual content (NMO, AEE, ZAA, HL, MGdC); statistical analysis (NMO, AEE, ZAA, HL); provision of patients or study materials (NMO); and supervision (NMO, MGdC).

Address Correspondence to: Nicolas M. Oreskovic, MD, MPH, Massachusetts General Hospital, 125 Nashua St, Ste 8406, Boston, MA 02114. Email: noreskovic@mgh.harvard.edu.

REFERENCES

1. Figueroa JF, Joynt Maddox KE, Beaulieu N, Wild RC, Jha AK. Concentration of potentially preventable spending among high-cost Medicare subpopulations: an observational study. Ann Intern Med. 2017;167(10):706-713. doi:10.7326/M17-0767

2. Figueroa JF, Frakt AB, Lyon ZM, Zhou X, Jha AK. Characteristics and spending patterns of high cost, non-elderly adults in Massachusetts. Healthc (Amst). 2017;5(4):165-170. doi:10.1016/j.hjdsi.2017.05.001

3. Weil E, Kaufman J, Neagle M. Care management demonstration project for high-risk populations: transition, communication and continuity. Collaborative Case Management. 2010:11-14. Accessed March 11, 2026. https://www.scribd.com/document/41915182/2010v8i3a4

4. Wasfy JH, Ferris TG. The business case for population health management. Prim Care. 2019;46(4):623-629. doi:10.1016/j.pop.2019.07.003

5. Blumenthal D, Abrams MK. Tailoring complex care management for high-need, high-cost patients. JAMA. 2016;316(16):1657-1658. doi:10.1001/jama.2016.12388

6. Hong CS, Siegel AL, Ferris TG. Caring for high-need, high-cost patients: what makes for a successful care management program? Issue Brief (Commonw Fund). 2014;19:1-19.

7. Ryan J, Abrams MK, Doty MM, Shah T, Schneider EC. How high-need patients experience health care in the United States: findings from the 2016 Commonwealth Fund survey of high-need patients. Issue Brief (Commonw Fund). 2016;43:1-20.

8. Powers BW, Chaguturu SK, Ferris TG. Optimizing high-risk care management. JAMA. 2015;313(8):795-796. doi:10.1001/jama.2014.18171

9. Vogeli C, Sprit 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.

10. 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 (Millwood). 2017;36(5):876-884. doi:10.1377/hlthaff.2016.0922

11. Bakshi S, Schiavoni KH, Carlson LC, et al. The essential role of population health during and beyond COVID-19. Am J Manag Care. 2021;27(3):123-128. doi:10.37765/ajmc.2021.88511

12. Kodner DL. Managing high-risk patients: the Mass General care management programme. Int J Integr Care. 2015;15:e017. doi:10.5334/ijic.2245

13. Powers BW, Modarai F, Palakodeti S, et al. Impact of complex care management on spending and utilization for high-need, high-cost Medicaid patients. Am J Manag Care. 2020;26(2):e57-e63. doi:10.37765/ajmc.2020.42402

14. Rowe JS, Gulla J, Vienneau M, et al. Intensive care management of a complex Medicaid population: a randomized evaluation. Am J Manag Care. 2022;28(9):430-435. doi:10.37765/ajmc.2022.89219

15. Oreskovic NM, Maniates J, Weilburg J, Choy G. Optimizing the use of electronic health records to identify high-risk psychosocial determinants of health. JMIR Med Inform. 2017;5(3):e25. doi:10.2196/medinform.8240

16. Weissman JS, Gatsonis C, Epstein AM. Rates of avoidable hospitalization by insurance status in Massachusetts and Maryland. JAMA. 1992;268(17):2388-2394. doi:10.1001/jama.1992.03490170060026

17. Guide to Prevention Quality Indicators: Hospital Admission for Ambulatory Care Sensitive Conditions. Agency for Healthcare Research and Quality; 2001. Updated November 24, 2004. Accessed January 1, 2025. https://qualityindicators.ahrq.gov/Downloads/Modules/PQI/V21/pqi_guide_rev4.pdf

18. Benson NM, Price M, Weiss M, et al. Tacking upwind: reducing spending among high-risk commercially insured patients. Am J Manag Care. 2023;29(5):220-226. doi:10.37765/ajmc.2023.89355

19. Finkelstein A, Zhou A, Taubman S, Doyle J. Health care hotspotting — a randomized, controlled trial. N Engl J Med. 2020;382(2):152-162. doi:10.1056/nejmsa1906848

20. Cornell PY, Halladay CW, Ader J, et al. Embedding social workers in Veterans Health Administration primary care teams reduces emergency department visits. Health Aff (Millwood). 2020;39(4):603-612. doi:10.1377/hlthaff.2019.01589

21. Peikes D, Chen A, Schore J, Brown R. Effects of care coordination on hospitalization, quality of care, and health care expenditures among Medicare beneficiaries: 15 randomized trials. JAMA. 2009;301(6):603-618. doi:10.1001/jama.2009.126

22. Heenan D. Hospital social work and discharge planning for older people: challenges of working in a clinical setting. Ageing Soc. 2023;43(6):1333-1350. doi:10.1017/S0144686X21001124