Long-Term Trends of Psychotropic Drug Use in Nursing Homes

WHY WE DID THIS STUDY

Nursing home residents and their families rely on nursing homes to provide quality care in a safe environment, and nursing homes are statutorily required to protect residents' rights in this regard. OIG work in 2011 raised quality and safety concerns about the high use of one category of psychotropic drug—antipsychotics—by nursing home residents. CMS began monitoring nursing home residents' use of antipsychotics in 2012, and in May 2021 OIG published a report that determined that CMS's existing methods for monitoring antipsychotic use by nursing home residents did not always provide complete information. Additionally, congressional stakeholders continue to raise concerns that nursing home residents may be inappropriately prescribed other types of psychotropic drugs and that potentially inappropriate use of those drugs may be going undetected.

HOW WE DID THIS STUDY

We used Minimum Data Set (MDS) assessment data from calendar year 2011 through 2019 to identify long-stay nursing home residents aged 65 and older and reviewed Medicare Part D psychotropic drug claims data for these residents. From these data, we identified the number of residents who received a prescription for any of these drugs. We then searched for patterns and characteristics in these data correlated with a higher use of psychotropic drugs in nursing homes. Our review did not assess the administration or medical necessity of psychotropic drugs for nursing home residents.

WHAT WE FOUND

OIG found that from 2011 through 2019, about 80 percent of Medicare's long-stay nursing home residents were prescribed a psychotropic drug. While CMS focused its efforts to reduce the use of one category of psychotropic drug—antipsychotics—the use of another category of psychotropic drug—anticonvulsants—increased. This increased use of anticonvulsants contributed to the overall use of psychotropics remaining constant.

In 2019, higher use of psychotropic drugs was associated with nursing homes that have certain characteristics. Nursing homes with lower ratios of registered nurse staff to residents were associated with higher use of psychotropic drugs. Nursing homes with higher percentages of residents with low-income subsidies were also associated with higher use of psychotropic drugs. Additionally, over time the number of unsupported schizophrenia diagnoses increased and in 2019 was concentrated in relatively few nursing homes. Specifically, we found that from 2015 through 2019 both the reporting of residents with schizophrenia in the MDS and the number of residents who lacked a corresponding schizophrenia diagnosis in Medicare claims and encounter data increased by 194 percent. In 2019, the unsupported reporting of schizophrenia was concentrated in 99 nursing homes in which 20 percent or more of the residents had a report of schizophrenia in the MDS that was not found in the Medicare claims history.

CMS's long-stay quality measure that tracks antipsychotic use in nursing homes excludes residents who are reported as having schizophrenia in the MDS. Thus, nursing homes could misreport residents as having schizophrenia in the MDS to falsely impact CMS's quality measure.

By not collecting diagnoses on Medicare Part D claims, CMS is limited in its ability to effectively conduct oversight of psychotropic drugs. First, not having diagnoses on claims limits CMS's ability to detect patient risk and patterns of potentially inappropriate drug use. Second, the lack of diagnoses makes it difficult for CMS to systematically determine whether claims meet the payment requirement that drugs be used for medically accepted purposes.

WHAT WE RECOMMEND

CMS should: (1) evaluate the use of psychotropic drugs among nursing home residents to determine whether additional action is needed to ensure that use among residents is appropriate, (2) use data to identify nursing homes or nursing home characteristics that are associated with a higher use of psychotropic drugs and focus oversight on nursing homes in which trends may signal inappropriate use, and (3) expand the required data elements on Medicare Part D claims to include a diagnosis code. CMS concurred with the first two recommendations and did not concur with the third recommendation in this report.