The Centers for Medicare & Medicaid Services Could Use Comprehensive Error Rate Testing Data To Identify High-Risk Home Health Agencies

CMS could use Comprehensive Error Rate Testing (CERT) data to identify high-risk home health agencies (HHAs) as a part of a multifaceted approach that includes targeted probe-and-educate reviews as well as aspects of its Fraud Prevention System to further reduce improper payments and the error rate for claims paid to HHAs. Using nationally reported CERT program data for fiscal years (FYs) 2014 through 2017, we identified 87 high-risk HHAs, which in the CERT sample had an improper payment rate of about 78 percent and approximately $1 million in actual improper payments. Using Medicare program data during this same period, we determined that Medicare paid these 87 HHAs more than $4 billion for services.

Given the amount of Medicare dollars paid to these providers and the high error rate observed in the CERT sample, focusing oversight on high-risk HHAs and the prevalent types of errors could significantly improve the effectiveness of CMS's efforts to reduce both HHA improper payments and the CERT error rate. In addition to this data brief focused on HHAs, our office will issue a report related to CMS's overall efforts to use CERT data to identify high-risk providers of all types in the Medicare program. That report will contain appropriate recommendations for action to further reduce overall CERT improper payments and CERT error rates.

In written comments on our data brief, CMS recognized that HHA claims are a major source of improper payments and described past and future corrective actions that address this issue. These corrective actions include identifying high-risk HHAs by using CMS's Fraud Prevention System, probe-and-educate reviews and targeted probe-and-educate reviews, and a new Home Health Review Choice Demonstration in Illinois. CMS stated that these sustained efforts have decreased the home health improper payment rate from 59 percent in FY 2015 to 17.6 percent in FY 2018 and are reliable and effective.

CMS stated that it does not believe that our methodology for identifying high-risk HHAs is valid, as the CERT data are not designed to be precise at the provider level. CMS stated that providers with a higher number of claims billed are much more likely to be selected by the CERT review. Additionally, CMS stated that in the past it has attempted to use CERT data to identify high-risk providers but found these data to be misleading and ineffective. Therefore, CMS discontinued the practice of using CERT data to identify high-risk providers. CMS also stated that its payment contractors have more accurate data at the provider level.

We recognize that CMS's efforts over the past several years have contributed to the reduction of improper HHA payments. However, the HHA improper payment rate remains higher than the national error rate of 8.1 percent in 2018 for all provider types. Moreover, none of CMS's current work precludes the use of CERT data to improve oversight efforts. Although we agree there are limitations to the CERT data, we maintain that our analysis for identifying high-risk HHA providers is valid and can be used to identify areas of potential risk, including specific HHAs that are high-risk providers needing CMS's attention from either the enforcement or provider education perspective.

We disagree with CMS that our method for identifying high-risk HHA providers is misleading and ineffective. Although we don't dispute that CMS and its payment contractors may have additional information employed in its program integrity efforts, our methodology for identifying high-risk HHA providers is valid. We used valid CERT-sampled data to identify a valid dataset of HHA providers with high error rates. We continue to maintain that CMS can use CERT data as an additional tool in its targeted enforcement and education efforts to effectively reduce improper payments and CERT error rates.