Stakeholders in the workers' compensation (WC) system largely agree that health care reform measures such as those contained in the Patient Protection and Affordable Care Act (PPACA) have the potential to affect the cost and composition of medical care received under WC. So far, however, there has been no broad consensus on the magnitude or even direction of the likely impacts of reform. Moreover, the limited research literature on this topic draws mostly qualitative conclusions and highlights the considerable uncertainty that exists regarding reform's future impacts.
In this study, I used the experience of Massachusetts, which implemented a health care reform package in 2006 containing provisions similar to many of PPACA's key provisions, to develop quantitative, empirical evidence on the effects of reform. My analysis focused in particular on hospital care, drawing data from the State Emergency Department Data (SEDD) and State Inpatient Data (SID) files for Massachusetts. These data included observations on over 9.5 million emergency room (ER) visits and 3 million inpatient hospitalization episodes, and included detailed information on the diagnoses and medical procedures for each patient, patient demographics, billed charges, and source of payment.
Relative to nearby states, Massachusetts has a lower rate of recorded workplace injuries and illnesses, and its WC fee schedules provide a comparatively low rate of reimbursement to providers for medical services. These somewhat unique features of Massachusetts's WC system may affect the ability to generalize from Massachusetts to other contexts. Examining SEDD and SID data on the identity of payers over time, I found that the reform was successful at its primary goal of increasing health insurance coverage and reducing the number of uninsured. The share of patients accessing the ER who had no insurance fell from approximately 15 percent prior to the reform to about 9 percent after the reform, a 40 percent decrease. The increase in coverage can be explained by new enrollments in Medicaid rather than by uptake of private health insurance.
To arrive at empirical estimates of the reform's impact on WC bill and treatment volume and charging patterns, I conducted a projection analysis using 2005 data to predict the number of WC ER bills observed in future quarters, adjusting for the mix of patients observed in the ER. This analysis demonstrated that realized WC ER bill volume fell by 7 percent relative to expected volume in the period after the reform began, and the timing of the decline matched that of the reform's implementation. In addition, the results of a “dose-response” analysis suggest that segments of the population that experienced the largest increases in health insurance coverage during the reform also saw the largest declines in WC bill volume, providing further evidence that the reform reduced WC bills. These estimates suggest that the reform can account for a roughly 5 percent to 10 percent decline in WC ER bill volume.
The implementation of Massachusetts's health care reform partly overlapped the recession that began in December 2007, and a considerable body of evidence suggests that WC claiming activity is linked with business cycle conditions. However, four pieces of evidence suggest that the reform had impacts separate and apart from any changes in WC billing that occurred because of the business cycle: (1) segments of the population whose coverage expanded the most following the reform had the largest changes in WC billing; (2) declines in WC billing occurred in late 2006 and 2007, when the reform was being implemented but before the recession began; (3) parts of the state that saw little change in unemployment rates saw declines in WC billing that began as the reform was implemented; (4) the relationship between health coverage expansions and WC billing declines was strong even after measures of business cycle activity were directly controlled for.
Using projection and dose-response analyses that focused on hospital inpatients (whose average billed charges were $21,000, versus $900 for ER patients) and the 20 percent of ER patients with the most expensive medical conditions, I examined whether the decline in WC bills was observable for seriously injured patients. Both subgroup analyses showed a decline in WC bills, suggesting that reform may affect both lower-cost and high-cost WC patients.
To assess whether the reform affected hospital charges, I analyzed the evolution of billed charges for patients covered by WC and patients covered by other types of insurance, controlling for changes over time in the composition of the patient population. Although charges grew for all types of patients between 2005 and 2008, there is no evidence of differential charge growth among WC patients. Similarly, although available measures of treatment volume are somewhat limited, a differences-in-differences analysis comparing changes in treatment volume across the reform period for patients covered by WC versus patients covered by other types of insurance yielded little evidence of important impacts on treatment volume.
The findings of these analyses present important first evidence suggesting that health care reform may reduce WC billing volume and costs. Important questions do, however, remain. First, how readily these results will generalize to other states that implement health care reform is unclear. Second, only hospital care was considered; whether these results translate to other care modalities, such as outpatient visits or pharmaceutical prescriptions, remains to be seen. Third, only the short-run impacts of reform were assessed. Fourth, additional outcomes likely to be of interest to stakeholders, such as worker health or quality and appropriateness of treatment, were not considered. Despite these limitations, however, the results provide important evidence that health care reform may generate spillover effects on non–health insurance lines such as WC.
