Principal investigator: Bruno Benevit
Original title: The Oregon Health Insurance Experiment: Evidence from the First Year
Authors: Amy Finkelstein, Sarah Taubman, Bill Wright, Mira Bernstein, Jonathan Gruber, Joseph P. Newhouse, Heidi Allen, Katherine Baicker, and Oregon Health Study Group.
Location of the Intervention: United States
Sample Size: 90.000 individuals
Sector: Health Economics
Primary Variable of Interest: Use of health services
Type of Intervention: Eligibility for health insurance
Methodology: OLS, 2SLS
Summary
Public policies for financing health insurance are fundamental for low-income individuals to access medical care. Measuring the impact of health insurance coverage is crucial for improving the effectiveness of health policies. To assess the impact of... Medicaid In the state of Oregon, United States, this study conducted a lottery that provided the opportunity to enroll in the program, providing health service coverage to those insured. One year after the lottery, the results indicated that those selected had a 25 percentage point higher chance of being insured, used more health services, had lower medical expenses and debt, and reported better physical and mental health compared to those who were not selected.
- Policy Problem
Health insurance plays a significant role in providing financial protection and ensuring access to quality medical care for policyholders. The coverage provided by these insurances reduces the cost barrier for medical treatments, allowing policyholders to seek care when needed without fear of incurring exorbitant out-of-pocket expenses. By covering a wide range of medical services, from routine consultations and preventive care to hospitalizations and complex procedures, health insurance helps promote a healthier and more productive population. Furthermore, by mitigating the financial risks associated with unexpected medical emergencies, health insurance provides a sense of security and well-being for individuals and their families (FINKELSTEIN et al., 2012).
The program Medicaid It is a significant example of health insurance targeted at socioeconomically vulnerable populations in the United States. Created to provide medical coverage to low-income individuals and families, the Medicaid It provides an essential support network for those who otherwise could not afford medical care. The importance of this program is amplified among vulnerable populations, as it not only improves access to healthcare but also helps alleviate financial pressure by reducing out-of-pocket medical expenses and associated debt.
In the context of the United States, several states implement regional policies associated with Medicaid, as was the case in the state of Oregon in 2008. The Oregon Health Plan (OHP) was created as a federal exemption program (waivers) of the traditional rules of Medicaid, consisting of two distinct programs: OHP Standard and OHP Plus. OHP Standard is an expansion program of Medicaid Aimed at covering low-income adults who are not categorically eligible for OHP Plus (which is aimed at pregnant women, people with disabilities, and families enrolled in other social programs). The Standard version of the program was implemented through a lottery system, allowing for the measurement of the program's coverage impacts. Medicaid in several aspects.
- Policy Implementation Context
OHP Standard serves adults aged 19 to 64 who reside in Oregon, are U.S. citizens or legal immigrants, have been without health insurance for six months, have income below the federal poverty level (FPL), and possess assets of less than $2.000. The program offers comprehensive benefits with no consumer copay, covering medical services, prescription drugs, major hospital benefits, mental health and substance abuse services, hospice care, and some durable medical equipment. However, it does not cover vision or non-emergency dental services. Care is primarily provided by managed care organizations, with monthly premiums ranging from $0 to $20, depending on income. At its peak in 2002, approximately 110.000 people were enrolled in OHP Standard, but due to budget cuts, the program was closed to new enrollments in 2004.
In 2008, due to reduced enrollment and budgetary constraints, the state decided to reopen the program to an additional 10.000 adults. A public awareness campaign was conducted, and a waiting list was created, allowing enrollment by phone, fax, mail, online, or in person. Over five weeks, 89.824 individuals were added to the list. The state conducted eight random selections from March to September 2008, selecting 35.169 individuals from 29.664 households to apply for the OHP Standard. Approximately 30% of those selected successfully enrolled, while the remainder failed to submit the required documentation or were deemed ineligible due to income. Enrollees could remain in the program indefinitely, provided they confirmed their eligibility every six months.
- Evaluation Details
This study used a combination of administrative data and a mail survey to assess the impacts of the OHP Standard. The survey was sent in seven waves during July and August 2009 to nearly all individuals selected by the draw and an approximately equal number of unselected individuals. The basic survey involved three mail delivery attempts, resulting in a response rate of 36%. To increase the response rate, a more intensive protocol was applied to approximately 30% of non-respondents, including additional tracking efforts, mail deliveries, and telephone contacts. This resulted in an additional 22% response rate for those who received the intensive protocol, culminating in an effective response rate of 50% when weighted by the inverse probability of inclusion in the intensive follow-up subsample.
In addition to the main survey, the data were compared with two nearly identical prior surveys conducted with the same population: an initial survey conducted approximately one year after randomization and a six-month survey conducted midway between the initial and main surveys. The six-month survey was applied to a 20% subsample of the sample used in the other two surveys and obtained response rates of 45% and 42%, respectively. This initial data was primarily used to construct “lottery list” variables that helped examine the pre-randomization demographic balance between the treatment and control groups.
Additionally, administrative records from the state were obtained regarding the entire enrollment history. Medicaid The data from participants on the lottery list from before the draw until September 2009 were collected. This data was used as the primary measure of insurance coverage. Administrative records on the benefit history of the programs were also obtained. Food Stamp e Temporary Assistance to Needy Families (TANF) of participants. Outcomes were measured from the date individuals were notified of their selection until the end of September 2009, representing a mean observation period of 16 months after notification and 14 months after coverage approval for those who successfully enrolled in the OHP Standard. If an individual obtained insurance through the lottery, coverage was applied retroactively to a few days after the state sent the application, usually one month after the notification date and one month before the approval date.
- Method
The study used an Ordinary Least Squares (OLS) method to estimate Intention to Treat (ITT) to assess the impact of winning the OHP Standard lottery. This method compared outcomes between the treatment group (lottery winners) and the control group (unselected). Variables such as household size and survey wave were considered to control for observable characteristics between groups. In addition to the main variables, the study included additional lottery-related covariates to increase the robustness of the analyses, controlling for factors such as lottery list demographics, outcome measures before randomization, and the specific lottery draw.
Subsequently, the study used the Two-Stage Least Squares (2SLS) method to estimate the local mean treatment effect (LATE) of being insured by Medicaid via the OHP Standard program. For the first stage of this regression, the probability of being insured was estimated from the program's lottery result. Linear models were employed for all estimates, even for binary results. The standard errors of the models were grouped by household identifier, since the treatment was performed at the household level.
The analyses of the survey data were weighted to reflect the sample design and ensure that the results were not sensitive to different weighting methods. The outcome variables analyzed included the program's effects on the use of hospital services and medical care, preventive care practices, financial stress resulting from healthcare expenses, and individuals' perception of their health.
- Main results
The results related to the (LATE) effect of access to Medicaid Studies conducted via the OHP Standard program revealed an increase in the use of hospital services and medical care. Individuals in the treated group experienced a 2,1 percentage point (pp) increase in the likelihood of hospital admission, concentrated primarily in non-emergency admissions. There was a proportional increase of approximately 20% in the number of hospital days, 40% in total billing, and 45% in the number of procedures, the latter being statistically significant. The analysis also showed a significant increase in hospital utilization for cardiac diseases, but found no substantial effects on the quality of outpatient or inpatient care. Furthermore, there was no detectable change in the proportion of patients going to public hospitals compared to private hospitals.
Specifically regarding medical care, estimates showed that insurance is associated with significant increases in the use of prescription medications and outpatient care, with a 15% increase in the number of medications and a 55% increase in outpatient visits. There was no significant impact on emergency room use or hospitalizations. Overall, insurance increased utilization by $778 in annual spending, approximately 25% more compared to the control group. Furthermore, insurance increased adherence to recommended preventive care, raising cholesterol and diabetes screenings by approximately 10 percentage points and mammograms and Pap smears by approximately 18 percentage points.
The results show that health insurance is associated with a statistically insignificant increase in overall financial pressure, measured by various forms of financial stress such as bankruptcy, foreclosure, judgment, collections, and payment delays. However, there was a significant 4,8 pp reduction in the likelihood of unpaid bills being sent to collections, especially medical bills, indicating a positive impact of health insurance in this specific aspect. Furthermore, measures of financial stress obtained through survey data reveal a statistically significant decrease in out-of-pocket medical expenses, difficulties paying non-medical bills due to medical expenses, and refusal of medical treatment due to medical debt, reflecting direct financial benefits for policyholders. These results suggest that the Medicaid It provides substantial benefits to beneficiaries, not only in terms of health, but also financially.
The results indicate that health insurance is associated with statistically significant improvements in seven self-reported health measures, including a greater likelihood of the individual reporting their health as good, very good, or excellent, and a reduction in the likelihood of being diagnosed with depression. These improvements correspond to an average increase of 13 percentage points in self-reported positive health, a 25% increase compared to the control group. Furthermore, there is evidence of an increase in perceived access to healthcare and in the perceived quality of care received by approximately 32%. Given the subjective nature of the responses, the authors argued that it is difficult to determine to what extent these results reflect improvements in objective physical health or an overall increase in perceived well-being. Despite initial concerns that increased contact with the healthcare system might lead to a worse perception of health due to the diagnosis of new health problems, the results do not suggest that this outweighs the perceived positive effects of health insurance.
- Lessons in Public Policy
This study investigated the impact of the OHP Standard program on low-income beneficiaries in Oregon who were subsequently enrolled in health insurance. MedicaidThrough an analysis that explored the randomized experimental design of the program eligibility draw, the effects of insurance on a variety of indicators were analyzed, including the use of health services, financial issues, and perceptions of health care among beneficiaries.
The results indicate a significant increase in the likelihood of non-emergency hospital admissions, suggesting a potential cost sensitivity on the part of beneficiaries. Furthermore, insurance is associated with a substantial increase in the use of prescribed medications and outpatient visits, indicating an expansion of access to primary healthcare services. The study also investigated the impacts of health insurance on measures of financial stress, revealing a statistically significant reduction in the likelihood of medical bills being sent for collection and in the average number of medical collections. While no overall decrease in financial stress was observed, the results suggest that insurance can alleviate the financial burdens associated with healthcare, especially in cases of medical billing. Additionally, there were perceived improvements in the perceived quality of medical care and access to care, reflecting a positive response from beneficiaries to insurance.
This study emphasizes the importance of health insurance policies that not only increase access to medical services but also reduce perceived financial barriers, contributing to financial security and quality of healthcare among low-income populations. Such evidence is fundamental for improving the targeting of public policies aimed at increasing access to services for socioeconomically vulnerable populations.
References
FINKELSTEIN, A. et al. The Oregon Health Insurance Experiment: Evidence from the First Year*. The Quarterly Journal of Economics, v. 127, no. 3, p. 1057–1106, 1 Aug. 2012.
