Limiting Private Insurance Reimbursement to Medicare Rates Would Reduce Health Spending by About $350 Billion in 2021

Methods

To calculate the amount that would be spent on health care for people with private insurance at Medicare rates, we used ratios of 2018 Medicare reimbursement to private insurance allowed amounts calculated by FAIR Health. FAIR Health used their private health care claims dataset. FAIR Health received approximately 2 billion medical and dental claims for the 2018 year. FAIR Health calculated Medicare reimbursement rates by applying the rules of the inpatient prospective payment system (IPPS), outpatient prospective payment system (OPPS), Medicare Physician Fee Schedule (MPFS), Medicare home health prospective payment system (PPS), Medicare Inpatient Psychiatric Facility PPS, Medicare Inpatient Rehabilitation Facility PPS, Medicare Skilled Nursing Facility (SNF) PPS, Medicare Ambulance Fee Schedule, Anesthesia Reimbursement Fee Schedule, Medicare Durable Medical Equipment, Prosthetics/Orthotics & Supplies Fee Schedule (DMEPOS), Medicare ASP Drug Pricing Files and the Clinical Laboratory fee schedule (CLAB) made available from CMS to services on claims submitted for those with private insurance. As discussed in the limitations section, FAIR Health did not include adjustments to Medicare reimbursements for indirect payments such as DSH, uncompensated care and IME. The FAIR Health ratios of Medicare reimbursement to commercial allowed amounts were calculated by metropolitan statistical area, service category, and by age group. The service categories that were used are: inpatient, outpatient, physician office, skilled nursing facility, laboratory, emergency department and urgent care center. Both facility and physician fees are included. The age groups that were used in our analysis are: 0 to 3, 4 to 18, 19 to 35, 36 to 44, 45 to 54 and 55 to 64.

As discussed in the limitations section of the report, the private-to-Medicare ratios that were used in our analysis are lower than those found in many other studies. Private insurance payment rates in those earlier studies varied based on the characteristics of the markets examined by each study and the author’s methods. Higher private payment rates were often seen in studies looking at more consolidated health care markets where providers have stronger negotiating leverage over insurers. Other factors influencing the results of the studies KFF examined were the representativeness of hospitals, physicians, and insurers used in the analysis and the data collection period, as well as differences in the services included across studies. Additionally, the FAIR Health ratios compare spending rather than prices. If people with private insurance use services more often that have relatively higher Medicare rates, then the ratio of private to Medicare spending for a set of services will be lower than the ratio of individual prices. Since our analysis focuses on spending, using ratios that compare spending rather than prices should give us a more accurate result.

The ratios from FAIR Health were then applied to 2018 spending data from IBM Health Analytics MarketScan Commercial Claims and Encounters Database (IBM Corporation). We indexed 2018 spending to the current year (2021) by applying the projected growth in spending and enrollment for those with private health insurance from the National Health Expenditures data to our estimates of 2018 spending. MarketScan is a convenience sample of health care claims provided primarily by large employers and health plans. Our analysis used claims for almost 18 million people representing about 22% of the 82 million people in the large group market in 2018. The advantage of using claims information to analyze spending is that we can look beyond plan provisions and focus on actual payment liabilities incurred by enrollees. MarketScan includes information on cost-sharing, as well as enrollees that do not have any health spending.

These data reflect out-of-pocket spending incurred under the benefit plan, but do not include balance-billing payments made to health care providers for out-of-network services or out-of-pocket payments for non-covered services. Costs for retail prescription drugs were excluded from our analysis. Our approach would not have been appropriate for prescription drug spending because there is not one Medicare price for retail prescription drugs, since those drugs are covered through the Medicare Part D program, which relies on private Part D plans that negotiate with drug manufacturers.

We reweighted the MarketScan data to represent the distribution of the entire population with private insurance. Weights were applied to match counts in the 2018 American Community Survey (ACS) for non-elderly enrollees with either employer or non-group coverage by sex, age and state. Weights were trimmed at eight times the interquartile range. In total, our analytic sample reflects a universe count of 173 million people. The ACS asks respondents about their health insurance coverage at the time of the survey. Respondents may report having more than one type of coverage; however, individuals are sorted into only one category of insurance coverage. Hospital costs were trimmed to exclude the lowest 1.5% and highest 0.5% of hospital costs within DRG. We also excluded MarketScan observations with negative spending. Due to data constraints, our analysis assumes the distribution of spending for people with private non-group insurance matches the distribution of spending for people with large group employer insurance.

To calculate what spending by private insurance would have been if Medicare rates were used instead of private insurance allowed amounts, we applied the FAIR Health ratios to total spending calculated using MarketScan within age and service category. For example, the ratio for spending on outpatient services for ages 0-3 was applied to total outpatient spending using MarketScan data for people with private insurance who are ages 0-3. We then summed spending across all age and spending categories. This approach assumes the distribution of services in MarketScan is similar to those in the FAIR Health data set on which the commercial to Medicare ratios were estimated.

We break out estimates of the reduction of spending for those that directly purchase health insurance, out-of-pocket spending for enrollees in employer plans, and the reduction in payments to providers made on behalf of enrollees in employer insurance (financed by employee and employer premium contributions). To estimate the reduction in spending attributed to services for those that directly purchase health insurance, we calculated the share of the population with private insurance in each age-sex category that was enrolled in a direct purchase plan. We multiplied the estimates of spending under current commercial rates, Medicare rates, and the reduction in savings by these shares and summed across all age and sex categories. We did not breakout out-of-pocket spending and payer spending for the direct purchase market. For those with employer insurance, we calculated out-of-pocket spending by applying the out-of-pocket share of total spending observed in the MarketScan data for each age group to the estimated spending under Medicare rates and aggregated across all groups. To estimate the share of payer spending financed by employer and employee premiums, respectively, we used MEPS data to calculate the average share of premiums covered by employer contributions (73%) and employees (27%).¹

We made several simplifying assumptions to allocate the distribution of the decrease in spending among enrollee premiums and out-of-pocket spending, and employer contributions. Notably, our analysis did not account for changes in benefit design, including enrollee cost-sharing, supply of health care, or enrollment patterns if payment rates for all private insurers were set at Medicare rates. Instead, we assumed that the actuarial value of the average employer plan would stay the same if Medicare rates were used, and we used that percentage to allocate the reduction in spending across out-of-pocket costs and premiums. We also assumed that employers would keep their share of premiums constant and that the use of health care would stay the same if prices dropped.