Estimated Impact Of Medically Tailored Meals On Health Care Use And Expenditures In 50 US States

Poor diet is a leading determinant of disease burdens and health inequities.¹ Medically tailored meals (MTMs) are a “Food Is Medicine” intervention that can improve diet-related health outcomes, reduce financial strain and improve associated well-being, address disparities, and reduce health care spending.^2–4 MTMs are prepared, home-delivered meals, typically provided to people with complex health conditions and high acuity of care based on a referral from a medical professional or health plan. Registered dietitian nutritionists design MTMs on the basis of patients’ nutritional assessments and diagnoses. Existing MTM programs typically provide ten weekly meals for an average of eight months annually.⁵ Recipients often have comorbidities, limitations in instrumental activities of daily living (IADLs), nutritional risk, and high prior health care use.⁶

Research suggests that nationwide adoption of MTMs for patients with diet-sensitive conditions could reduce patients’ health care use and spending by public and private payers.⁶ However, MTM coverage remains limited, although several states are beginning to cover Food Is Medicine interventions through Medicaid demonstration projects.^1,7 These states are important incubators for health care innovation.

US states vary considerably in demographics, disease prevalence, and health care use and spending, which may influence MTMs’ impact and cost-effectiveness. The impact of MTM coverage on health care use and spending by US state remains unknown. This study investigated the one-year and five-year effects of implementing MTMs in US states on hospitalizations, health care expenditures, and net costs among patients with diet-related diseases and ADL limitations under Medicaid, Medicare, and private insurance. By providing state-specific estimates, this analysis offers policy makers and health plans insights to guide MTM implementation decisions at the state level.

Study Data And Methods

Design, Setting, And Population

Adopting a previously published population-level, open-cohort policy simulation model,⁶ we estimated changes in annual hospitalizations and health care expenditures after the implementation of MTMs versus the status quo. The study sample combined 2010–19 data from the Medical Expenditure Panel Survey (MEPS).⁸ We excluded 2020 because of pandemic-related data collection changes.⁹ State-level restricted data were obtained from the Agency for Healthcare Research and Quality Data Center.¹⁰ Because MEPS state-specific data were only available for the twenty-nine most populous states, we estimated the size of the state-specific eligible populations using data from the 2019 Behavioral Risk Factor Surveillance System (see supplement table 1 in the online appendix).¹¹ We could not jointly stratify results by both state and payer because of the diminishing size of the data on the eligible populations within these subgroups.

Patient Eligibility

The analytic sample included US adults older than age eighteen with Medicare, Medicaid, or private coverage who had one or more diet-sensitive conditions and IADL limitations. Diet-sensitive conditions include diabetes, coronary heart disease, myocardial infarction, other heart disease, emphysema, stroke, nonmelanoma cancer, chronic kidney disease, and HIV. We selected conditions based on previous MTM research.^3,12–16 IADL limitations include receiving help using the telephone, paying bills, taking medications, preparing meals, doing laundry, or going shopping.¹⁷

MTM Intervention And Effect Sizes

We modeled the impact of providing ten weekly MTMs for a mean of eight months,⁶ following Food is Medicine Coalition standards.¹⁸ Effect sizes were estimated via an inverse-variance weighted meta-analysis of eight published interventional studies on MTM provision,^{3,12–14,19–22} all of which employed quasi-experimental controls. The analysis suggested that MTM provision was associated with a 47 percent reduction in annual hospitalizations (95% confidence interval: 36.1, 57.9) and a 19.7 percent reduction in annual health care expenditures (95% CI: 7.0, 32.4).

Intervention And Health Care Costs

Annual MTM program costs included clinical screening and program expenses. Screening expenses were based on 2024 Medicare reimbursement rates for a fifteen-minute medical nutritional therapy session with a registered dietitian, ranging from $33.50 (Mississippi) to $46.45 (Alaska).²³ A 2024 survey of MTM organizations within the Food is Medicine Coalition (six responses) estimated that the mean program expense per meal was $11.15 (standard deviation: 1.71) (Alyssa Wassung, Food is Medicine Coalition, personal communication, May 22, 2024). The MTM meal costs represent fully burdened MTM-related expenditures, including labor; overhead; contracting; and direct expenses such as materials (food and packaging) and delivery.

We estimated state-specific health care expenditures among the eligible population using the 2010–19 MEPS-restricted state-level data set, adjusted to 2024 US dollars. State-specific data were available for the twenty-nine most-populous states, representing 88 percent of the US population. For the remaining twenty-one states, we extrapolated annual hospitalizations and expenditures using regional averages (Northeast, South, Midwest, and West), excluding data from the twenty-nine most populous states.

Simulation Model

Input data for the simulation model, programmed in R, included state-level eligible populations and demographics, baseline annual hospitalizations and expenditures, MTM intervention effects, and MTM program costs (supplement table 7).¹¹ Model outputs included changes in annual hospitalizations and expenditures, program costs, and net policy costs evaluated from a health system perspective at one year (2024) and five years (2024–28). Five-year outcomes were simulated annually, incorporating trends in eligible populations and health care expenditures, using 2010–19 MEPS data. These trends were estimated using log-linear regression, stratified by region. We assumed that eligible people received meals for eight months annually, as long as they remained qualified, with a summing of outcomes across years and applying 3 percent annual discounting of health care and intervention costs. This assumption was based on studies used to derive the effect size, which had an average duration of treatment around eight months per twelve months of observation.^3,12–14,22

Probabilistic Sensitivity Analyses

By randomly drawing from the input parameters’ range in a probabilistic sensitivity analysis, 1,000 Monte Carlo simulations accounted for uncertainty. The central estimate was the mean of 1,000 simulations, with the 95% uncertainty interval from the 2.5th and 97.5th percentiles, assuming a normal distribution.

Secondary Sensitivity Analyses

We performed secondary sensitivity analyses to assess the results’ robustness. First, we restricted the receipt of MTMs to people with diabetes or congestive heart failure. Second, we modeled 50 percent of eligible patients receiving MTMs, instead of 100 percent. Third, we varied the intervention effect size across the 2.5th 10th, 25th, 50th, 75th, 90th, and 97.5th percentiles of the effect size. Fourth, we conducted two threshold analyses to estimate minimum changes in the per meal cost and health care expenditures needed for cost-neutrality. Fifth, we evaluated annual discounting rates of 0 percent and 5 percent. Finally, we estimated sustained health benefits from MTMs, assuming that 15 percent of MTM recipients each year would not require MTMs the following year while maintaining reduced hospitalizations and expenditures.

Strengths And Limitations

Strengths of our study include the use of state-specific data on eligible patients, health care use, and expenditures, allowing us to generate findings relevant to local decision-makers. Effects of MTM treatment were derived from pooled interventional studies and program costs from contemporary insurance contracts between MTM providers and health care systems. Patient eligibility criteria were consistent with prior research and existing MTM programs. Our model incorporated an open cohort and probabilistic sensitivity analyses to jointly incorporate uncertainty and report a range of outcomes. One-way sensitivity analyses tested the influence of specific assumptions. We estimated one-year and five-year outcomes, providing a range of clinically relevant and policy-relevant time horizons.

However, our study also had limitations. Modeling MTMs for patients with severe comorbidities and IADL limitations might not generalize to healthier populations or less intensive nutritional interventions. Also, MTM effects may vary across disease states. However, the included conditions were selected on the basis of prior MTM research, which informed our effect size estimates. These estimates represent a reasonable averaged effect across the selected disease states.

The main analysis assumed full coverage of all eligible people to provide a best-case scenario. In practice, scaling MTM services to reach all eligible people would take time, making a 50 percent coverage sensitivity analysis an informative comparison. Although scaling could enhance program efficiency and reduce costs, it could also affect nutritional quality, necessitating empirical research for evaluation.

Although state-specific data were available for the twenty-nine most populous states, estimates for the remaining states were based on regional means. Smaller eligible populations in smaller states contributed to wider uncertainty intervals, influenced by sample size and variance. However, for policy makers, overall trends and expected benefits often can offer actionable insights despite some uncertainty.

Finally, although we could not jointly stratify results by state and payer because of limited sample sizes, MTMs were cost-effective in nearly all states across Medicare, Medicaid, and private payers, suggesting that substantial divergence by payer type in individual states is unlikely.

Study Results

Size And Characteristics Of The MTM-Eligible Population

An estimated 7.05 million US adults qualify for MTMs. Among the twenty-nine most populous states, the number of eligible patients was generally proportional to the state population, with the largest number in California ($n=1,221,035$) and the smallest in Oregon ($n=170,319$) (supplement table 2).¹¹ Nationally, the mean age was 67.8 years (SD: 16.5); 62.5 percent of the study population were female; and 54.1 percent were non-Hispanic White adults, 23.4 percent were non-Hispanic Black adults, 11.3 percent were Hispanic adults, and 8.0 percent were non-Hispanic Asian adults (exhibit 1). The median household income-to-poverty ratio was 1.8 (interquartile range: 1.0–3.3; supplement table 8),¹¹ and 89.7 percent of patients had Medicare or Medicaid (exhibit 1). The most common eligibility diagnoses were cardiovascular diseases (61.6 percent), diabetes (35.2 percent) and cancer (29.4 percent). Among the sample population, mean annual health care expenditures per person were $30,892 (SD: 45,403), and the annual number of hospitalizations per person was 0.53 (SD: 1.02), highlighting the high severity of illness and health care use.

Projecting One-Year Outcomes

Assuming 100 percent uptake among eligible people, MTMs were projected to generate mean net health care cost savings across all states except Alabama (where MTMs were cost-neutral) (exhibit 2; supplement table 2 and supplement figure 2).¹¹ Nationally, after the intervention costs were accounted for, MTMs were estimated to save $32.1 billion (95% UI: −19.8, 87.9) (supplement table 2).¹¹ States with the largest annual net cost savings were Connecticut ($6,299 per patient; 95% UI: 15, 13,974), followed by Pennsylvania ($4,450; 95% UI: −845, 10,824) and Massachusetts ($4,331; 95% UI: −756, 10,148).

SOURCE Authors’ analysis of data from the Medical Expenditure Panel Survey (MEPS), pooling data from the period 2010–19. NOTES Each eligible person was assumed to have received MTMs for 8 months annually. The simulation model ran 1,000 Monte Carlo simulations, using inputs and their uncertainties from the pooled MEPS data; relative risks of annual hospitalizations; and annual percent changes in health care expenditures associated with MTM receipt, screening costs, and meal costs. The results present the mean of 1,000 simulations, ranked by net annual cost savings per person.

MTMs would avert a meaningful number of hospitalizations across all states, despite variations (exhibit 3; supplement table 2).¹¹ The number of treated patients needed to avert one hospitalization annually was lowest in Maryland, at 2.3, followed by Massachusetts and Pennsylvania, each at 3.0. Nationally, MTMs were estimated to avert 3,542,500 hospitalizations (95% UI: 2,266,800, 4,786,800) annually (supplement table 2).¹¹

SOURCE Authors’ analysis of data from the Medical Expenditure Panel Survey (MEPS), pooling data from the period 2010–19. NOTES Each eligible person was assumed to have received MTMs for 8 months annually. The simulation model ran 1,000 Monte Carlo simulations, using inputs and their uncertainties from the pooled MEPS data, as described in the exhibit 2 notes. The results present the mean of 1,000 simulations, ranked by the number of treated patients needed to prevent 1 hospitalization annually.

The annual net per person health care cost savings for each state and averted hospitalizations were modestly correlated ($r=0.46$), with meaningful variation (exhibit 4). Maryland had lower cost savings per MTM-treated patient ($3,248; 95% UI: −1,086, 8,352) while requiring the smallest number of treated patients needed to avert one hospitalization (2.3 patients). Conversely, Connecticut reported large cost savings per MTM-treated patient ($6,299) but required a modest number of treated patients to avert one hospitalization (4.1 patients).

SOURCE Authors’ analysis of data from the Medical Expenditure Panel Survey (MEPS), pooling data from the period 2010–19. NOTES Each eligible person was assumed to have received MTMs for 8 months annually. The simulation model ran 1,000 Monte Carlo simulations, as described in the exhibit 2 notes. The results present the mean of 1,000 simulations. Averted hospitalization is represented as the mean number of hospitalizations averted annually per person. Each dot represents a state; Maryland and Connecticut are labeled for ease of interpreting the exhibit findings as described in the text.

States in the South and Midwest regions demonstrated a higher correlation between per person cost savings and averted hospitalizations, whereas states in the Northeast and West showed more variability (exhibit 4). In the five most populous states, between 195,600 and 265,400 hospitalizations would be averted, with total cost savings of $1.5–$3.7 billion (supplement figure 1).¹¹

Sensitivity Analyses With One-Year Projection

For MTMs serving people with diabetes in our one-year projections, the eligible population would decrease to 5.35 million nationally. MTMs would be net cost saving across all states (supplement table 3),¹¹ with an average savings of $3,403 (95% UI: −1,008, 8,149) per person, representing nearly 150 percent greater savings than in our base case. Per MTM-treated patient with diabetes, states with the largest net cost savings were Connecticut ($11,098; 95% UI: 1,262, 24,164), Washington ($8,702; 95% UI: −463, 19,757), and Pennsylvania ($8,620; 95% UI: 462, 18,982).

When the one-year analysis was restricted to patients with heart failure, 2,896,686 people were MTM-eligible nationally. MTMs for those patients would be net cost saving across all states, with a mean savings of $3,989 (95% UI: −834, 9,109) per person nationally—nearly 174 percent greater than our base case (supplement table 4).¹¹ States with the largest net cost savings per MTM-treated patient were Connecticut ($9,279; 95% UI: 1,262, 18,845), Pennsylvania ($6,876; 95% UI: 21, 14,710), and Massachusetts ($6,686; 95% UI: 129, 13,901).

Assuming 50 percent uptake, MTMs were still estimated to avert 1,771,000 hospitalizations (95% UI: 1,131,200, 2,440,500) and reduce net costs by $16.0 billion (95% UI: −9.9, 44.0) nationally. For a one-year MTM intervention to be cost-neutral, the per meal cost would need to increase from $11.15 to $18.30; the effect size on health care expenditure reduction would need to decrease by around one-third, from 19.7 percent to 12.5 percent.

Projecting Five-Year Outcomes

Based on observed trends among MTM-eligible patients, the eligible population was estimated to increase annually by 2.4 percent (Northeast), 1.7 percent (Midwest), 2.4 percent (South), and 0.3 percent (West) over a five-year period. Inflation-adjusted health care expenditures would increase annually in our sample by 1.5 percent (Northeast), 5.6 percent (Midwest), 1.3 percent (South), and 1.4 percent (West) during these five years. In 2024 dollars with 3 percent discounting, five years of MTM intervention could prevent 14,668,000 (95% UI: 9,919,000, 19,481,000) hospitalizations (supplement table 6)¹¹ and gain $151.0 billion (95% UI: −59.4, 382.3) in net savings nationally. Total projected state-level findings for five years generally mirrored those observed in the one-year projection (supplement figure 1).¹¹

Sensitivity Analyses With Five-Year Projection

For 5.0 percent discounting and 0 percent discounting, five-year net cost savings were $137.2 billion (95% UI: −53.9, 347.3) and $175.1 billion (95% UI: −68.8, 443.2) (supplement table 6).¹¹ Assuming that MTM benefits were sustained into the second year for 15 percent of people (3 percent discounting), MTMs could generate net cost savings across all states, totaling $227.5 billion (95% UI: 19.4, 456.4; data not shown) nationally during the five-year projection period.

Discussion

Our simulation model estimated that state-level MTM coverage among Medicare, Medicaid, and privately insured patients with diet-sensitive conditions and IADL limitations would be associated with reductions in annual hospitalizations and health care expenditures for nearly all states. The overall net cost savings ranged from $6,299 per patient (Connecticut) to cost-neutrality (Alabama). The number of treated patients needed to avert one hospitalization in a year ranged from 2.3 (Maryland) to 6.9 (Colorado).

The variation in MTM outcomes by state can be attributed to eligible population sizes, baseline annual health care expenditures, annual hospitalizations, and state-specific health care management strategies. States with the largest net cost savings, such as Connecticut ($6,299), Pennsylvania ($4,450), and Massachusetts ($4,331), reported high baseline health care costs per capita.²⁴ Maryland was estimated to require the smallest number of treated patients to avert one hospitalization (2.3 patients) but had relatively lower net cost savings per MTM-treated patient ($3,248) than other similar states with higher costs, such as Massachusetts (3.0 patients and $4,331) and Pennsylvania (3.0 patients and $4,450). These findings were likely driven by population health and state-level factors in health care delivery. For example, Maryland hospitals operate under a unique global budget program, which sets fixed annual budgets for hospital operations.^25–27 Although this program discourages unnecessary admissions, it also limits the direct financial benefits that hospitals can realize from each averted hospitalization, in part because of Maryland’s higher baseline number of hospitalizations.

Overall, the modestly correlated state-level net cost saving and averted hospitalizations ($r=0.46$) align with previous findings, suggesting potential inefficiency in health care practices in some states.^28,29 Although meal costs may vary by state, this variation is unlikely to have influenced the overall conclusion, as the cost-neutrality threshold ($18.30) is much higher than the highest reported per meal cost among the major MTM organizations we surveyed. In addition, using BRFSS data to estimate eligibility led to a higher number of eligible people (14.0 million) and greater estimated net cost savings ($32.5 billion) nationally (supplement table 2),¹¹ compared with a previous national-level analysis using MEPS data (6.3 million people, $13.6 billion).⁶ This difference is primarily due to a higher proportion of adults reporting limitations in IADLs in the BRFSS (7.2 percent) compared with MEPS (3.2 percent).

Most long-term MTM programs use eligibility criteria beyond specific diseases, often considering other factors linked to higher health care use. These decisions are typically made by clinical providers or social workers case by case. Among the eight MTM studies informing our effect sizes, one explicitly included disability as a criterion, reporting a greater effect size than found in other studies.¹² Other programs included criteria such as recent hospital discharge with home health or hospice care, indicating likely functional limitations. Because national data lack details on clinical acuity, we used IADL limitations as a proxy to identify higher-need patients. Discussions with major MTM providers confirmed this as a reasonable approach to identifying such patients.

Although the availability of MTM services remains uneven, interest in improving access to such programs has grown through a national network of nonprofit providers that provide training and technical assistance to new programs.³⁰ There has also been a proliferation of for-profit MTM providers.³¹ We included published interventional studies that used nonprofit providers managing meal delivery through health systems, although some studies also involved commercial providers.³² Similar evaluations of for-profit providers are critical to ensuring that nutritional standards and health benefits are not diminished. For example, an analysis by STAT News identified the low nutritional quality of MTMs provided by a for-profit provider to several Medicaid programs.³³ In general, we would expect that the provision of less healthy meals would likely be less effective than our findings, given that the effects of MTMs on hospitalizations and health care use stem in part from improved diet quality.⁶

Most interventional evaluations have been quasi-experimental. There remains a need for large randomized controlled trials (RCTs). In a pilot RCT among 161 patients hospitalized with heart failure, MTMs reduced hospital readmissions by 46 percent at five months and 41 percent at nine months compared with controls.²¹ Conversely, a larger trial among nearly 2,000 patients who had recently been discharged with heart failure, diabetes, or chronic kidney disease¹⁹ found reductions in hospitalizations only in patients with heart failure. However, the intervention was short (mean of 6.9 weeks), and longer periods may be necessary to affect hospitalizations. A recent RCT enrolling patients with HIV found an 89 percent reduction in hospitalizations after six months of receiving MTMs, underscoring the potential for greater impact with longer intervention periods.³⁴

Although cost analyses are relevant to payers’ coverage decisions, the primary goal of MTM programs is to provide high-quality medical care for patients with diet-sensitive chronic illnesses. These programs should not be viewed as cost containment strategies—a high bar that is unrealistic for most therapies, preventive services, and diagnostic testing used in health care today.³⁵ In addition, the estimated effects on hospitalizations and health care expenditures do not incorporate potential additional benefits of MTMs for quality of life, disease progression, caregivers’ well-being, and health equity.

Our findings support the implementation and evaluation of MTM programs in public and private health systems at the state level.

Our findings support the implementation and evaluation of MTM programs in public and private health systems at the state level. As of January 2025, sixteen states had approved or proposed Medicaid Section 1115 waivers, enabling MTM treatment coverage and representing a major pathway to standardizing MTM treatment (Kathryn Garfield, Harvard University, personal communication, January 18, 2025). However, Section 1115 waivers often employ restrictive eligibility criteria and require cost-neutrality, and they require frequent renewal by the Centers for Medicare and Medicaid Services. In our study, 89.7 percent of patients were covered by Medicare or Medicaid, highlighting the importance of these programs in facilitating MTM access. Many Medicare Advantage plans are also starting to include MTMs to help improve chronic conditions and patient outcomes.³⁶

Conclusion

Our state-level simulation model estimated that coverage for MTMs in Medicare, Medicaid, and private insurance for patients with diet-sensitive conditions and IADL limitations could prevent hospitalizations in fifty states and be net cost saving in forty-nine states. These findings, including variations across states, can help inform state-level policy makers and health plans that are considering implementing MTMs in clinical care through state-specific strategies.

ACKNOWLEDGMENTS

This work was supported by the National Institutes of Health (NIH) under Grant Nos. 5R01 DK134452-02, 2R01 HL115189, and R01 MD019094. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the NIH or any affiliated institutions. This article complies with the NIH Public Access Policy and is freely available through PubMed Central (PMC). This research was presented at NUTRITION 2024 in Chicago, Illinois, June 29–July 2, 2024. The authors are grateful to Alissa Wassung and the members of the Food is Medicine Coalition for sharing information on the current costs of medically tailored meal programs. This is an open access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to distribute this work provided the original work is properly cited, not altered, and not used for commercial purposes. See https://creativecommons.org/licenses/by-nc-nd/4.0/. To access the authors’ disclosures, click on the Details tab of the article online.

NOTES

• 1 Mozaffarian D, Aspry KE, Garfield K, Kris-Etherton P, Seligman H, Velarde GPet al. “Food Is Medicine” strategies for nutrition security and cardiometabolic health equity: JACC State-of-the-Art Review. J Am Coll Cardiol. 2024;83(8):843–64. Crossref, Medline, Google Scholar
• 2 Downer S, Berkowitz SA, Harlan TS, Olstad DL, Mozaffarian D. Food is medicine: actions to integrate food and nutrition into healthcare. BMJ. 2020;369:m2482. Crossref, Medline, Google Scholar
• 3 Berkowitz SA, Delahanty LM, Terranova J, Steiner B, Ruazol MP, Singh Ret al. Medically tailored meal delivery for diabetes patients with food insecurity: a randomized cross-over trial. J Gen Intern Med. 2019;34(3):396–404. Crossref, Medline, Google Scholar
• 4 Chen AMH, Draime JA, Berman S, Gardner J, Krauss Z, Martinez J. Food as medicine? Exploring the impact of providing healthy foods on adherence and clinical and economic outcomes. Explor Res Clin Soc Pharm. 2022;5:100129. Crossref, Medline, Google Scholar
• 5 Food is Medicine Coalition. Medically tailored meals (MTMs) frequently asked questions [Internet]. [location unknown]: FIMC; [cited 2025 Jan 29]. Available from: https://static1.squarespace.com/static/580a7cb9e3df2806e84bb687/t/6387be7ea25c7f48e08f2eb1/1669840510456/FIMC-MTM-InfoSheet-20221123.pdf Google Scholar
• 6 Hager K, Cudhea FP, Wong JB, Berkowitz SA, Downer S, Lauren BNet al. Association of national expansion of insurance coverage of medically tailored meals with estimated hospitalizations and health care expenditures in the US. JAMA Netw Open. 2022;5(10):e2236898. Crossref, Medline, Google Scholar
• 7 Hanson E, Albert-Rozenberg D, Garfield KM, Leib EB, Ridberg RA, Hager Ket al. The evolution and scope of Medicaid Section 1115 demonstrations to address nutrition: a US survey. Health Aff Sch. 2024;2(2):qxae013. Medline, Google Scholar
• 8 Agency for Healthcare Research and Quality. Medical Expenditure Panel Survey, data file—search results [Internet]. Rockville (MD): AHRQ; [cited 2024 Jan 29]. Available from: https://meps.ahrq.gov/mepsweb/data_stats/download_data_files_results.jsp?cboDataYear=All&cboDataTypeY=1%2CHousehold+Full+Year+File&buttonYearandData Type=Search&cboPufNumber=All&SearchTitle=Consolidated+Data Google Scholar
• 9 Agency for Healthcare Research and Quality. MEPS HC-224 2020 full year consolidated data file [Internet]. Rockville (MD): AHRQ; 2022 Aug [cited 2025 Jan 29]. Available from: https://meps.ahrq.gov/data_stats/download_data/pufs/h224/h224doc.pdf Google Scholar
• 10 Agency for Healthcare Research and Quality. Medical Expenditure Panel Survey, Restricted data files available through the AHRQ Data Center [Internet]. Rockville (MD): AHRQ; [last revised 2023 Dec 1; cited 2025 Jan 29]. Available from: https://meps.ahrq.gov/data_stats/onsite_datacenter.jsp Google Scholar
• 11 To access the appendix, click on the Details tab of the article online.
• 12 Berkowitz SA, Terranova J, Hill C, Ajayi T, Linsky T, Tishler LWet al. Meal delivery programs reduce the use of costly health care in dually eligible Medicare and Medicaid beneficiaries. Health Aff (Millwood). 2018;37(4):535–42. Go to the article, Google Scholar
• 13 Gurvey J, Rand K, Daugherty S, Dinger C, Schmeling J, Laverty N. Examining health care costs among MANNA clients and a comparison group. J Prim Care Community Health. 2013;4(4):311–7. Crossref, Medline, Google Scholar
• 14 Hummel SL, Karmally W, Gillespie BW, Helmke S, Teruya S, Wells Jet al. Home-delivered meals postdischarge from heart failure hospitalization: the GOURMET-HF pilot study. Circ Heart Fail. 2018;11(8):e004886. Medline, Google Scholar
• 15 Tapper EB, Baki J, Nikirk S, Hummel S, Asrani SK, Lok ASet al. Medically tailored meals for the management of symptomatic ascites: the SALTYFOOD pilot randomized clinical trial. Gastroenterol Rep (Oxf). 2020;8(6):453–6. Crossref, Medline, Google Scholar
• 16 Food is Medicine Coalition. Existing research demonstrates that food is medicine [Internet]. [location unknown]: FIMC; [cited 2025 Jan 29]. Available from: https://fimcoalition.org/research/existing-research/ Google Scholar
• 17 Conwell LJ, Cohen JW. Characteristics of persons with high medical expenditures in the U.S. civilian noninstitutionalized population, 2002 [Internet]. Rockville (MD): Agency for Healthcare Research and Quality; 2005 Mar [cited 2025 Mar 18]. (MEPS Statistical Brief No. 73). Available from: https://meps.ahrq.gov/data_files/publications/st73/draft/stat73.shtml Google Scholar
• 18 Food is Medicine Coalition. Nutrition standards [Internet]. [location unknown]: FIMC; [cited 2025 Jan 29]. Available from: https://static1.squarespace.com/static/580a7cb9e3df2806e84bb687/t/60db83a01157011c9c6527bf/1624998816205/Nutriton+Standards_2_04_2021+docx.pdf Google Scholar
• 19 Go AS, Tan TC, Horiuchi KM, Laws D, Ambrosy AP, Lee KKet al. Effect of medically tailored meals on clinical outcomes in recently hospitalized high-risk adults. Med Care. 2022;60(10):750–8. Medline, Google Scholar
• 20 Nguyen HQ, Duan L, Lee JS, Winn TG, Arakelian A, Akiyama-Ciganek Jet al. Association of a Medicare Advantage posthospitalization home meal delivery benefit with rehospitalization and death. JAMA Health Forum. 2023;4(6):e231678. Crossref, Medline, Google Scholar
• 21 Pedroza-Tobias A. Medically supportive food and nutrition education improves health and reduces re-admission for safety-net patients hospitalized with heart failure exacerbation: a pilot randomized controlled trial. Paper presented at: American Public Health Association Annual Meeting and Expo; 2022 Nov 9; Boston, MA. Google Scholar
• 22 Horton R, Paulson K, Aaron D, Raab R, Hardy RY, Pulling E. Pairing medically tailored meals with diabetes self-management education programs [Internet]. Denver (CO): Project Angel Heart; [cited 2025 Jan 29]. Available from: https://www.projectangelheart.org/wp-content/uploads/2019/07/Telligen_Project-Angel-Heart_Pilot-Project_FINAL-002.pdf Google Scholar
• 23 Centers for Medicare and Medicaid Services. Search the Physician Fee Schedule [Internet]. Baltimore (MD): CMS; [last updated 2025 Jan 1; cited 2025 Jan 29]. Available from: https://www.cms.gov/medicare/physician-fee-schedule/search?Y=0&T=0&HT=1&CT=3&H1=97802&M=5 Google Scholar
• 24 Cooper W. Price of US health: which US states are spending the most on healthcare? William Russell Blog [blog on the Internet]. [cited 2025 Jan 29]. Available from: https://www.william-russell.com/blog/price-of-usa-healthcare/ Google Scholar
• 25 Definitive Healthcare. Examining hospital discharge data by state [Internet]. Framingham (MA): Definitive Healthcare; 2023 Jun 6 [cited 2025 Jan 29]. Available from: https://www.definitivehc.com/resources/healthcare-insights/average-hospital-discharges-by-state Google Scholar
• 26 Dávila Castrodad IM, Mohamed NS, Wilkie WA, Remily EA, Pollak AN, Mont MAet al. Maryland’s Global Budget Revenue model associated with lower inpatient costs and 30-day readmissions in patients undergoing total hip arthroplasty. Arthroplast Today. 2020;6(1):88–93. Crossref, Medline, Google Scholar
• 27 Rosenberg J. Report finds wide variation in average cost of healthcare across the US. American Journal of Managed Care [serial on the Internet]. 2018 Nov 15 [cited 2025 Jan 29]. Available from: https://www.ajmc.com/view/report-finds-wide-variation-in-average-cost-of-healthcare-across-the-us Google Scholar
• 28 Burke LA, Ryan AM. The complex relationship between cost and quality in US health care. Virtual Mentor. 2014;16(2):124–30. Crossref, Medline, Google Scholar
• 29 Do LA, Koethe BC, Daly AT, Chambers JD, Ollendorf DA, Wong JBet al. State-level variation in low-value care for commercially insured and Medicare Advantage populations. Health Aff (Millwood). 2022;41(9):1281–90. Go to the article, Google Scholar
• 30 Food is Medicine Coalition. Accelerator program [Internet]. [location unknown]: FIMC; [cited 2025 Jan 29]. Available from: https://fimcoalition.org/programs/fimc-accelerator/ Google Scholar
• 31 Season Health [home page on the Internet]. Keller (TX): Season Health; [cited 2025 Jan 29]. Available from: https://www.seasonhealth.com/ Google Scholar
• 32 Yu L, Madura A, Gil C, Hepfer P, Palar K. Assessing the health outcomes of the Food Access Pilot Project: an evaluation of a medically supportive food support program for people living with HIV in rural California counties. AIDS Behav. 2022;26(8):2613–22. Crossref, Medline, Google Scholar
• 33 Florko N. Medicaid is paying millions for salty, fat-laden “medically tailored” cheeseburgers and sandwiches. STAT [serial on the Internet]. 2024 Jul 11 [cited 2025 Jan 29]. Available from: https://www.statnews.com/2024/07/11/medicaid-food-is-medicine-medically-tailored-meals-quality-questioned/ Google Scholar
• 34 Palar K, Sheira LA, Frongillo EA, O’Donnell AA, Nápoles TM, Ryle Met al. Food Is Medicine for HIV: improved health and hospitalizations in the Changing Health through Food Support (CHEFS-HIV) pragmatic randomized trial. J Infect Dis. 2024 May 2. [Epub ahead of print]. Google Scholar
• 35 Cohen JT, Neumann PJ, Weinstein MC. Does preventive care save money? Health economics and the presidential candidates. N Engl J Med. 2008;358(7):661–3. Crossref, Medline, Google Scholar
• 36 Sherrell Z. Does Medicare cover meal delivery? Medical News Today [serial on the Internet]. 2024 Jul 11 [cited 2025 Jan 29]. Available from: https://www.medicalnewstoday.com/articles/medicare-meal-delivery Google Scholar