In-Progress and Working Papers


Many adults are eligible for public benefits programs but are not enrolled. This paper investigates whether geographically distant social networks can facilitate take-up of Medicaid, the U.S.'s low-income health insurance program. I estimate program take-up changes resulting from increased enrollment and salience among one's friend network due to state Medicaid eligibility expansions during 2010--2018. To isolate effects occurring through social channels, I restrict analyses to states that did not expand their Medicaid program and compare between communities with varying degrees of social connection to the expanding states. Potentially eligible adults with stronger social ties to the Medicaid expansions were more likely to enroll in the program following the expansions, even though eligibility was largely unaffected in their own state. The increase was reflected in decreases in uninsurance rates, with no impact on other sources of insurance. In addition to impacting take-up, social network exposure to the Medicaid expansions increased support for the Affordable Care Act and Medicaid among the broader population. The results show program experiences among one's friends can improve their own program participation and highlight how policy changes can have indirect impacts propagated through social networks.


This paper studies the Great Migration---the early-20th century mass migration of Black Americans out of the U.S. South---and its impact on economic racial inequality in the South. To isolate the impacts of migration from potentially correlated impacts of local push factors, we construct a shift-share style ``demand-pull'' instrument, exploiting variation in preexisting Southern out-migration patterns in 1900-10 and labor demand changes outside the South in 1910--1940. We estimate that counties with one percentile higher out-of-South migration during 1910--1940 had 1.3% higher average Black weekly wages in 1940. We find no effects on White wages, resulting in a reduction in the racial wage gap. Reduced Black labor supply and improved human capital accumulation are investigated as potential mechanisms. The results provide novel evidence of how the Great Migration impacted the Southern communities migrants left. 

Hidden Hypertension and the Labor Force Participation of Older Americans


Nearly half of American adults have hypertension, yet many are unaware of their condition and only 1 in 4 have it under control. Hypertension control is essential for preventing future health shocks, which can have adverse consequences for people’s ability to participate in the labor force, especially for older workers. This paper investigates the role of hypertension in older workers’ risk of leaving the labor force before reaching Social Security full retirement age. Using physical blood pressure (BP) measurements from the Health and Retirement Study (HRS) and Cox proportional hazards models, I find a 1 standard deviation higher systolic BP among working Americans ages 50-61 in 2006-2014 was associated with a higher rate of early labor force exits (hazard ratio=1.15). The relationship is robust to the inclusion of controls for a range of confounders and other alternative specifications. Previously diagnosed individuals had slightly lower risk conditional on BP levels but having BP above 140/90 was associated with significantly higher risk for both the diagnosed and undiagnosed. To assess the potential labor market benefits of hypertension screening I exploit an HRS physical measures data collection protocol wherein the nurse was instructed to leave a warning card if the respondent’s BP was above 160/110, which recommended the respondent see a physician as soon as possible. Using a regression discontinuity design among older workers with high BP, I estimate that receiving the warning card increased their remaining time in the labor force by 6 months, with evidence supporting improved BP management and avoided future hospitalizations and work limiting health problems as potential mechanisms. The results highlight the importance of screening and early risk identification for supporting older workers with chronic conditions. 

Peer-Reviewed Publications

Chapel JM, Tysinger B, Goldman DP, Rowe JW, and The Research Network on an Aging Society. The Forgotten Middle: Worsening Health and Economic Trends Extend to Americans With Modest Resources Nearing Retirement. Health Affairs. 2023; 42(9):xxx-xxx.

Media: The New York Times, A Health Podyssey (podcast), Marketplace, U.S. News & World Report.

Chapel JM, Tysinger B, Goldman DP, Rowe JW, and The Research Network on an Aging Society. The Forgotten Middle: Worsening Health and Economic Trends Extend to Americans With Modest Resources Nearing Retirement. Health Affairs. 2023; 42(9):xxx-xxx.


In an aging US society, anticipating the challenges that future seniors will face is essential. This study analyzed the health and economic well-being of five cohorts of Americans in their mid-fifties between 1994 and 2018 using the Future Elderly Model, a dynamic microsimulation based on the Health and Retirement Study. We projected mortality, quality-adjusted life years, health expenditures, and income and benefits. We classified individuals by economic status and focused on the lower middle and upper middle of the economic distribution. Outcome disparities between people in these two groups widened substantially between the 1994 and 2018 cohorts. Quality-adjusted life expectancy increased (5 percent) for the upper-middle economic status group but stagnated for their lower-middle peers. We found that the combined value of the current stock (financial and housing wealth) and the present value of the expected flow of resources (income, health expenditures, and quality-adjusted life-years) after age sixty grew 13 percent for the upper-middle group between cohorts, whereas people in the lower-middle group in 2018 were left scarcely better off (3 percent growth) than their peers two decades earlier. The relatively neglected “forgotten middle” group of near-retirees in the lower-middle group may require stronger supports than are currently available to them. 

Shantharam SS, Mahalingam M, Rasool A, Reynolds JA, Bhuiya AR, Satchell TD, Chapel JM, Hawkins NA, Jones CD, Jacob V, Hopkins DP. Systematic Review of Patient Characteristics, Intervention Features, and Cost of Self-Measured Blood Pressure Monitoring Interventions with Support . American Journal of Preventive Medicine. 2022; 62(2):285-298.


Introduction: Self-measured blood pressure monitoring with support is an evidence-based intervention that helps patients control their blood pressure. This systematic economic review describes how certain intervention aspects contribute to effectiveness, intervention cost, and intervention cost per unit of the effectiveness of self-measured blood pressure monitoring with support.

Methods: Papers published between data inception and March 2021 were identified from a database search and manual searches. Papers were included if they focused on self-measured blood pressure monitoring with support and reported blood pressure change and intervention cost. Papers focused on preeclampsia, kidney disease, or drug efficacy were excluded. Quality of estimates was assessed for effectiveness, cost, and cost per unit of effectiveness. Patient characteristics and intervention features were analyzed in 2021 to determine how they impacted effectiveness, intervention cost, and intervention cost per unit of effectiveness.

Results: A total of 22 studies were included in this review from papers identified in the search. Type of support was not associated with differences in cost and cost per unit of effectiveness. Lower cost and cost per unit of effectiveness were achieved with simple technologies such as interactive phone systems, smartphones, and websites and where providers interacted with patients only as needed.

Discussion: Some of the included studies provided only limited information on key outcomes of interest to this review. However, the strength of this review is the systematic collection and synthesis of evidence that revealed the associations between the characteristics of implemented interventions and their patients and the interventions' effectiveness and cost, a useful contribution to the fields of both research and implementation.

MacLeod KE, Chapel JM, McCurdy M, et al. The Implementation Cost of a Safety-Net Hospital Program Addressing Social Needs in Atlanta. Health Services Research. 2021;56(3):474-485.


Objective: To describe the cost of integrating social needs activities into a health care program that works toward health equity by addressing socioeconomic barriers.

Data sources/study setting: Costs for a heart failure health care program based in a safety-net hospital were reported by program staff for the program year May 2018-April 2019. Additional data sources included hospital records, invoices, and staff survey.

Study design: We conducted a retrospective, cross-sectional, case study of a program that includes health education, outpatient care, financial counseling and free medication; transportation and home services for those most in need; and connections to other social services. Program costs were summarized overall and for mutually exclusive categories: health care program (fixed and variable) and social needs activities.

Data collection: Program cost data were collected using a activity-based, micro-costing approach. In addition, we conducted a survey that was completed by key staff to understand time allocation.

Principal findings: Program costs were approximately $1.33 million, and the annual per patient cost was $1455. Thirty percent of the program costs was for social needs activities: 18% for 30-day supply of medications and addressing socioeconomic barriers to medication adherence, 18% for mobile health services (outpatient home visits), 53% for navigating services through a financial counselor and community health worker, and 12% for transportation to visits and addressing transportation barriers. Most of the program costs were for personnel: 92% of the health care program fixed, 95% of the health care program variable, and 78% of social needs activities.

Discussion: Historically, social and health care services are funded by different systems and have not been integrated. We estimate the cost of implementing social needs activities into a health care program. This work can inform implementation for hospitals attempting to address social determinants of health and social needs in their patient population.

Hawkins NA, Bhuiya A, Shantharam S, Chapel JM, et al. A Replicable Approach to Promoting Effective Practices: Translating Cardiovascular Disease Prevention Research. Journal of Public Health Management and Practice. 2021;27(2):109-116.


Objective: Significant delays in translating health care-related research into public health programs and medical practice mean that people may not get the best care when they need it. Regarding cardiovascular disease, translation delays can mean lives may be unnecessarily lost each year. To facilitate the translation of knowledge to action, we created a Best Practices Guide for Cardiovascular Disease Prevention Programs.

Design: Using the Rapid Synthesis Translation Process and the Best Practices Framework as guiding frameworks, we collected and rated research evidence for hypertension control and cholesterol management strategies. After identifying best practices, we gathered information about programs that were implementing the practices and about resources useful for implementation. Research evidence and supplementary information were consolidated in an informational resource and published online. Web metrics were collected and analyzed to measure use and reach of the guide.

Results: The Best Practices Guide was released in January 2018 and included background information and resources on 8 best practice strategies. It was published as an online resource, publicly accessible from the Centers for Disease Control and Prevention Web site in 2 different formats. Web metrics show that in the first year after publication, there were 25 589 Web page views and 2467 downloads. A query of partner use of the guide indicated that it was often shared in partners' own resources, newsletters, and online material.

Conclusion: In following a systematic approach to creating the Best Practices Guide and documenting the steps taken in its development, we offer a replicable approach for translating research on health care practices into a resource to facilitate implementation. The success of this approach is attributed to 3 key factors: using a prescribed and documented approach to evidence translation, working closely with stakeholders throughout the process, and prioritizing the content design and accessibility of the final product.

Chapel JM & Wang G. Understanding Cost Data Collection Tools to Improve Economic Evaluations of Health Interventions. Stroke and Vascular Neurology. 2019;4(4)214-222.


Micro-costing data collection tools often used in literature include standardized comprehensive templates, targeted questionnaires, activity logs, on-site administrative databases, and direct observation. These tools are not mutually exclusive and are often used in combination. Each tool has unique merits and limitations, and some may be more applicable than others under different circumstances. Proper application of micro-costing tools can produce quality cost estimates and enhance the usefulness of economic evaluations to inform resource allocation decisions. A common method to derive both fixed and variable costs of an intervention involves collecting data from the bottom up for each resource consumed (micro-costing). We scanned economic evaluation literature published in 2008-2018 and identified micro-costing data collection tools used. We categorized the identified tools and discuss their practical applications in an example study of health interventions, including their potential strengths and weaknesses. Sound economic evaluations of health interventions provide valuable information for justifying resource allocation decisions, planning for implementation, and enhancing the sustainability of the interventions. However, the quality of intervention cost estimates is seldom addressed in the literature. Reliable cost data forms the foundation of economic evaluations, and without reliable estimates, evaluation results, such as cost-effectiveness measures, could be misleading. In this project, we identified data collection tools often used to obtain reliable data for estimating costs of interventions that prevent and manage chronic conditions and considered practical applications to promote their use. 

Zhang D, Ritchey M, Park C, Li J, Chapel J, Wang G. Association Between Medicaid Coverage and Income Status on Health Care Use and Costs Among Hypertensive Adults After Enactment of the Affordable Care Act. American Journal of Hypertension. 2019; 32(10):1030-1038.


Background: Hypertension is highly prevalent among the low-income population in the United States. This study assessed the association between Medicaid coverage and health care service use and costs among hypertensive adults following the enactment of the Patient Protection and Affordable Care Act (ACA), by income status level.

Methods: A nationally representative sample of 2,866 nonpregnant hypertensive individuals aged 18-64 years with income up to 138% of the federal poverty level (FPL) were selected from the 2014 and 2015 Medical Expenditure Panel Survey. Regression analyses were performed to examine the association of Medicaid coverage with outpatient (outpatient visits and prescription medication fills), emergency, and acute health care service use and costs among those potentially eligible for Medicaid by income status-the very low-income (FPL ≤ 100%) and the moderately low-income (100% > FPL ≤ 138%).

Results: Among the study population, 70.1% were very low-income and 29.9% were moderately low-income. Full-year Medicaid coverage was higher among the very low-income group (41.0%) compared with those moderately low-income (29.1%). For both income groups, having full-year Medicaid coverage was associated with increased health care service use and higher overall annual medical costs ($13,085 compared with $7,582 without Medicaid); costs were highest among moderately low-income patients ($17,639).

Conclusion: Low-income individuals with hypertension, who were potentially newly eligible for Medicaid under the ACA may benefit from expanded Medicaid coverage by improving their access to outpatient services that can support chronic disease management. However, to realize decreases in medical expenditures, efforts to decrease their use of emergency and acute care services are likely needed.

Yarnoff B, Khavjou O, Elmi J, Lowe-Beasley K, Bradley C, Amoozegar J, Wachtmeister D, Tzeng J, Chapel JM, Teixeira-Poit S, The Paul Coverdell National Acute Stroke Program Team. Estimating the Costs of Implementing Components of Stroke Systems of Care and Data-Driven Quality Improvement in the Paul Coverdell National Acute Stroke Program. Preventing Chronic Disease. 2019;16:190061.


Purpose and Objectives: We evaluated the costs of implementing coordinated systems of stroke care by state health departments from 2012 through 2015 to help policy makers and planners gain a sense of the potential return on investments in establishing a stroke care quality improvement (QI) program.

Intervention Approach: State health departments funded by the Paul Coverdell National Acute Stroke Program (PCNASP) implemented activities to support the start and proficient use of hospital stroke registries statewide and coordinate data-driven QI efforts. These efforts were aimed at improving the treatment and transition of stroke patients from prehospital emergency medical services (EMS) to in-hospital care and postacute care facilities. Health departments provided technical assistance and data to support hospitals, EMS agencies, and posthospital care agencies to carry out small, rapid, incremental QI efforts to produce more effective and efficient stroke care practices.

Evaluation Methods: Six of the 11 PCNASP-funded state health departments in the United States volunteered to collect and report programmatic costs associated with implementing the components of stroke systems of care. Six health departments reported costs paid directly by Centers for Disease Control and Prevention–provided funds, 5 also reported their own in-kind contributions, and 4 compiled data from a sample of their partners’ estimated costs of resources, such as staff time, involved in program implementation. Costs were analyzed separately for PCNASP-funded expenditures and in-kind contributions by the health department by resource category and program activity. In-kind contributions by partners were also analyzed separately.

Results: PCNASP-funded expenditures ranged from $790,123 to $1,298,160 across the 6 health departments for the 3-year funding period. In-kind contributions ranged from $5,805 to $1,394,097. Partner contributions (n = 22) ranged from $3,912 to $362,868.

Implications for Public Health: Our evaluation reports costs for multiple state health departments and their partners for implementing components of stroke systems of care in the United States. Although there are limitations, our findings represent key estimates that can guide future program planning and efforts to achieve sustainability.

Chapel JM, Ritchey MD, Zhang D, Wang G. Prevalence and Medical Costs of Chronic Diseases Among Adult Medicaid Beneficiaries. American Journal of Preventive Medicine. 2017;53(6 Suppl 2):S143-S154.


Introduction: This review summarizes the current literature for the prevalence and medical costs of noncommunicable chronic diseases among adult Medicaid beneficiaries to inform future program design.

Methods: The databases MEDLINE and CINAHL were searched in August 2016 using keywords, including Medicaid, health status, and healthcare cost, to identify original studies that were published during 2000–2016, examined Medicaid as an independent population group, examined prevalence or medical costs of chronic conditions, and included adults within the age group 18–64 years. The review and data extraction was conducted in Fall 2016–Spring 2017. Disease-related costs (costs specifically to treat the disease) and total costs (all-cause medical costs for a patient with the disease) are presented separately.

Results: Among the 29 studies selected, prevalence estimates for enrollees aged 18–64 years were 8.8%–11.8% for heart disease, 17.2%–27.4% for hypertension, 16.8%–23.2% for hyperlipidemia, 7.5%–12.7% for diabetes, 9.5% for cancer, 7.8%–19.3% for asthma, 5.0%–22.3% for depression, and 55.7%–62.1% for one or more chronic conditions. Estimated annual per patient disease-related costs (2015 U.S. dollars) were $3,219–$4,674 for diabetes, $3,968–$6,491 for chronic obstructive pulmonary disease, and $989–$3,069 for asthma. Estimated hypertension-related costs were $687, but total costs per hypertensive beneficiary ranged much higher. Estimated total annual healthcare costs were $29,271–$51,937 per beneficiary with heart failure and $11,446–$20,585 per beneficiary with schizophrenia. Costs among beneficiaries with cancer were $29,384–$46,194 for the 6 months following diagnosis.

Conclusions: These findings could help inform the evaluation of interventions to prevent and manage noncommunicable chronic diseases and their potential to control costs among the vulnerable Medicaid population.