Medicaid Managed Care and Emergency Department Utilization: A North Carolina Analysis
by Temitope Ayokunmi Ojo
Abstract
In July 2021, North Carolina Medicaid switched from a traditional fee-for-service model to a Medicaid managed care (MMC) network. This thesis explores the effect of this policy change on Emergency Department (ED) utilization for Medicaid beneficiaries in North Carolina. A linear difference-in-difference model was used to estimate the change in ED visits between the treatment group, Medicaid beneficiaries, and two control groups, non-Medicaid 19–64-year-olds and 65+ NC residents. The results indicate a statistically significant decline in ED visits, about 11% decline from pre-policy visit rates, for Medicaid beneficiaries after the mandatory switch to managed care. The reduction in visits was most persistent for those related to chronic condition treatment. Furthermore, we find evidence consistent with both medical care disruption and better management of health as drivers of the decline in ED visits. Determining the cause of these patterns should be explored by deeper analyses of trends in other healthcare delivery avenues (i.e. PCP appointments or hospital admissions) post-policy implementation.
Professor M. Kate Bundorf, Faculty Advisor
Professor Grace Kim, Seminar Advisor
JEL Codes: I11, I13, I18
Keywords: Medicaid, Insurance, Emergency Department
The Effect of Community Uninsurance Rates on Access to Health Care among the Insured
by Isabella Antonio
Abstract
While the direct effects of being uninsured have been studied extensively, there is significantly less research on how a high community uninsured rate can impact health care access for insured individuals. Using data from SMART BRFSS, I examine the effect of community uninsured rates on access to health care for insured individuals ages 18 to 64 years old. Controlling for MMSA-level fixed effects and year fixed effects, I estimate the effect of community uninsurance on the likelihood of an insured individual skipping care due to cost, the likelihood of an insured individual having at least one personal doctor, and the likelihood of an insured individual delaying a physical exam, cholesterol check, or pap smear. Results suggest that a 10 percentage point increase in the community uninsured rate decreases the likelihood of an insured individual having at least one personal doctor by 0.304 percentage points and increases the likelihood of delaying a physical exam, cholesterol check, or pap smear by 0.590 to 2.31 percentage points. These findings suggests that policies aimed at reducing the uninsured rate, such as the Affordable Care Act and Medicaid expansion, may produce widespread benefits for all Americans, both the uninsured and the insured.
Professor Jeffrey DeSimone, Faculty Advisor
Professor Michelle Connolly, Faculty Advisor
JEL Codes: I1, I11, I13
Keywords: Health insurance, Health care access
(Email for access to data)
The Impact of Medicare Nonpayment: A Quasi-Experimental Approach
By Audrey Kornkven
In October 2008, a provision of the Deficit Reduction Act of 2005 known as Medicare “Nonpayment” went into effect, eliminating reimbursement for the marginal costs of preventable hospital-acquired conditions in an effort to correct perverse incentives in hospitals and improve patient safety. This paper contributes to the existing debate surrounding Nonpayment’s efficacy by considering varying degrees of fiscal pressure among hospitals; potential impacts on healthcare utilization; and differences between Medicare and non-Medicare patient populations. It combines data on millions of hospital discharges in New York from 2006-2010 with hospital-, hospital referral region-, and county-level data to isolate the policy’s impact. Analysis exploits the quasi-experimental nature of Nonpayment via difference-in-differences with Mahalanobis matching and fuzzy regression discontinuity designs. In line with results from Lee et al. (2012), Schuller et al. (2013), and Vaz et al. (2015), this paper does not find evidence that Nonpayment reduced the likelihood that Medicare patients would develop a hospital-acquired condition, and concludes that the policy is not likely the success claimed by policymakers. Results also suggest that providers may select against unprofitable Medicare patients when possible, and are likely to vary in their responses to financial incentives. Specifically, private non-profit hospitals appear to have been most responsive to the policy. These findings have important implications for pay-for-performance initiatives in American healthcare.
Advisors: Professor Charles Becker, Professor Frank Sloan, Professor Grace Kim| JEL Codes: I1, I13, I18
Assessing the Impacts of an Aging Population on Rising Healthcare and Pharmaceutical Expenditures within the United States
By Rahul Sharma
This paper studies the impact of aging on rising healthcare and pharmaceutical expenditures in the United States with the goal of contextualizing the future burden of public health insurance on the government. Precedent literature has focused on international panels of multiple countries and hasn’t identified significant correlation between age and healthcare expenditures. This paper presents a novel approach of identifying this correlation by using a US sample population to determine if age impacts an individual’s consumption of healthcare services and goods. Results suggest that age has a significant impact on healthcare and pharmaceutical expenditures across private and public insurance.
Advisors: Gilliam D. Saunders-Schmidler and Grace Kim | JEL Codes: H51, H53, I12, I13, I18, I38
The Cost-Effectiveness of Shared Medical Appointments for Type II Diabetes at Duke Family Medicine
By Lauren Nahouraii
With increasing healthcare expenditures above the rate of inflation, new health care delivery models are needed. Since care for chronic health conditions accounts for a majority of spending, more cost-effective ways to manage these conditions are especially necessary and could be the most effective in decreasing health care costs. Shared medical appointments (SMAs) are a promising solution because they increase patient education through group appointments while simultaneously increasing productivity by allowing a provider to see patients in a group but bill for them individually. In this study, 38 patient volunteers participated in an SMA as part of a pilot program at Duke Family Medicine (DFM). As part of this program, patients were randomly assigned to groups that offered varying versions of an SMA curriculum over the course of 3 years. Data collected included HbA1c scores, number and type of medications, type of insurance and payments, number and type of visit (including hospital admissions, emergency room visits, primary care and specialty visits), laboratory tests completed, and home address. Data was collected during, after, and for the six months prior to starting the SMAs. Data points from six months prior to the SMAs serve as a control. HbA1c served as the measure of health outcome while the rest of the data was used in estimating the total healthcare costs of control and treatment periods. Any changes in HbA1c were converted into changes in quality adjusted life years (QALYs) for the cost-effectiveness calculations. The estimated total costs and changes in QALYs were used to calculate the average cost-effectiveness of both the control and treatment periods. Given the small sample size, the SMAs appeared to be more cost-effective for patients that attended a majority of the SMA sessions. The cost-effectiveness comparison for all patients was inconclusive. This study’s calculations should be repeated once more patients complete SMAs in order to increase the power of the tests and provide conclusive results for all patients.
Advisor: Tracy Falba, Ralph Snyderman | JEL Codes: I10, I12, I13, I18
The Professor and the Coal Miner: The effect of socioeconomic and geographical factors on breast cancer diagnosis and survival outcome
By Shelley Chen
Previous studies reported that patients who live farther from cancer centers do not necessarily experience delayed cancer detection and shortened survival. However, the results are biased because of the incomplete observation of patient survival, which cannot be properly accounted for with the multivariable regression model. In this thesis, I isolated the effect of the breast cancer patient’s distance to a comprehensive cancer center on the stage of diagnosis and survival using the Cox Proportional Hazards model. I linked data from the Kentucky Surveillance, Epidemiology, and End Results 18, the Kentucky Life Tables, and the Kentucky Area Health Resource Files and identified 37654 patients diagnosed with breast cancer. I estimated the effect of distance on marginal probability of cancer mortality, controlling for non-cancer related death, socioeconomic status, and demographic factors in patients. After controlling for covariates, travel distance between the patient and the nearest comprehensive cancer center was statistically significantly on the breast cancer mortality probability, but not on the stage of diagnosis. In the Kentucky population, patients who were located farther from comprehensive cancer centers experience an increased marginal probability of mortality (proportional hazard = 1.004; 95% CI: [1.000502 1.007311]). The linkage of SEER 18 and AHRF data provided more comprehensive information on the socioeconomic risk factors of cancer mortality than past study datasets. For the stage of diagnosis, a low physician to population ratio and high county-level Medicaid coverage were associated with more advanced stages of diagnosis. In turn, a more advanced stage of diagnosis, lower physician to population ratio, and identification as African American increased the marginal probabilities of mortality.
Advisor: Charles Becker, Kent Kimbrough | JEL Codes: I1, I13, I14 | Tagged: Breast Cancer, Cancer Mortality, Health Outcomes, Inequality, Socioeconomic, Stage
Incentives and Characteristics that Explain Generic Prescribing Practices
By Rahul Nayak
This study uses the National Ambulatory Medical Care Survey (2006-2010) and Health Tracking Physician Survey (2008) to study the incentives and characteristics that explain physician generic prescribing habits. The findings can be characterized into four main categories: (1) financial/economic, (2) informational, (3) patient- dependent and (4) drug idiosyncratic effects. Physicians in practices owned by HMOs or practices that had at least one managed care contract are significantly more likely to prescribe generic medicines. Furthermore, physicians who have drug industry influence are less likely to prescribe generic medicines. This study also finds consistent evidence that generic prescribing is reduced for patients with pri- vate insurance compared to self-pay patients. Drug-specific characteristics play an important role for whether a drug is prescribed as a generic or brand-name – in- cluding not only market characteristics, such as monopoly duration length, public familiarity with the generic and the quality of the generic, but also non-clinical drug characteristics, such as the length of the generic name compared the length of the brand-name. In particular, the public’s familiarity with the generic has a large effect on the generic prescribing rate for a given drug. There are few differences between the generic prescribing habits of primary care physicians and specialists after controlling for the drugs prescribed.
Advisor: Frank Sloan, Kent Kimbrough | JEL Codes: D82, D83, I11, I13, I18 | Tagged: Drug Market Characteristics, Efficient Prescribing, Electronic Prescribing, Generic Prescribing, HTPS, Industry Influence, NAMCS, Patient Preferences, Physician Incentives, Principle- Agent Problem
Debunking the Cost-Shifting Myth: An Analysis of Dnamic Price Discrimination in California Hospitals
By Omar Nazzal
Cost-shifting, a dynamic form of price discrimination, is a phenomenon in which hospitals shift the burden of decreases in government-sponsored healthcare reimbursement rates to private health insurers. In this paper, I construct a data set spanning 2007 – 2011 that matches financial metrics of California hospitals to hospital- and market-specific characteristics with theoretical implications in price discrimination. The subsequent analysis is split into three stages. In the first and second stages, I use a fixed-effects OLS model to derive a point estimate of the inverse correlation between private revenue and government revenue that is consistent with recent empirical work in cost-shifting, a body of literature almost entirely reliant upon fixed-effects and difference-in-difference OLS. These types of models are encumbered by the inherent causality loop connecting public and private payment sources. I address this endogeneity problem in the third stage by specifying a fixed-effects 2SLS model based on an instrument for government revenue constructed with data from the California Department of Health Care Services and the U.S. Census. This instrument performed well in canonical tests for relevance and validity. I find that an increase in government payments causes an increase in private payments, and that the relationship is statistically-significant at all reasonable levels. In addition, I comment on properties of the data set that suggest that the original inverse correlation was due to inadequate measurements of market power. I conclude with policy implications and suggestions for future research.
Advisor: Frank Sloan | JEL Codes: I11, I13, I18, L11, L80 | Tagged: Health Insurance, Market Structure, Medicaid, Medicare, Price Discrimination
Policy in Competitive Insurance Markets: Incentivizing Risk Sharing and Cost Efficiency
By Ross Green
In the setting of a population with heterogeneous risk of illness, informational asymmetries in a competitive health insurance market can cause the gains from risk sharing to fall short of social optimality in equilibrium. Traditional policies meant to address the under-provision of insurance, like mandating open enrollment or community-rated premiums, can be prohibitively costly or impossible to implement. I consider three policy regimes in the context of a competitive insurance industry in which firms maximize profits by exerting effort to monitor the provision of health care. When multiple risk types are present in the population, I find that a subsidy rule based on the marginal costs of insuring high risks can induce a Pareto-improvement to risk sharing gains, at a cost to the efficiency of health care provision. The novelty of the subsidy rule lies in the way it incentives pooling equilibria.
Advisor: Curtis Taylor | JEL Codes: I0, I13, I18 | Tagged: Health Care Efficiency, Insurance Contracts, Private Information
