Antipsychotic Drugs and Nursing Home Residents: What Do the Different Numbers Mean?

Last week’s Alert discussed the Centers for Medicare & Medicaid Services’ (CMS’s) National Partnership to Improve Dementia Care and the Government Accountability Office’s (GAO) recent report on antipsychotic drugs.[1]  The CMS Partnership and the GAO reported different numbers of nursing home residents receiving antipsychotic drugs.  Some of the differences appear to reflect the different databases and assumptions underlying CMS’s and the GAO’s work.  This Alert explains these differences and advantages and disadvantages of each approach. 

CMS itself reports different rates of antipsychotic drug use for nursing home residents, based on resident assessment information (minimum data set, or MDS 3.0).  The Partnership reports that the national prevalence of antipsychotic drug use for long-stay residents declined from 23.2% in the second quarter of 2012 to 19.4% in the second quarter 2014.[2]  MDS data reported on a CMS website indicate that the rate of antipsychotic drug use for all nursing facility residents (short-stay as well as long-stay) in the fourth quarter of 2014 was 22.42%.[3]

Regardless of the databases, assumptions, and populations of residents studied, however, two points are indisputable:

1. Too many nursing home residents are given antipsychotic drugs, even though they have dementia and not a mental illness for which antipsychotic drugs are intended, and even though the drugs are harmful and life-threatening to them.[4]

The National Partnership may have helped lead to 50,000-60,000 fewer long-stay residents taking antipsychotic drugs inappropriately between March 2012 and December 2014, but more than 200,000 residents continue to take these drugs today, and for the overwhelming majority of them, the drugs are inappropriate and dangerous. 

2. Antipsychotic drugs are too often used as a substitute for adequate staffing.[5]

   Nursing facilities need sufficient numbers of professional nurses and paraprofessional nursing staff who know the residents individually as people; recognize and understand the significance of changes in residents’ conditions; attempt to determine whether residents are communicating problems through their behavior when they do not have words to say what is wrong; and respond appropriately to these problems, rather than call for antipsychotic drugs.  Considerable research documents that residents with dementia are insufficiently treated for pain.[6]  Antipsychotic drugs should not be used as an alternative to needed pain medication.

The National Partnership

The National Partnership’s reporting uses facility-wide data and resident assessment information (minimum data set, or MDS) data to identify the number of long-stay residents taking antipsychotic drugs. 

Advantages of this Approach:

• It reflects facility-wide information, giving a broad perspective on the extent of the use of antipsychotic drugs for long-stay residents.  Although the primary concern is the inappropriate use of these drugs with residents who have dementia, facility-wide data capture all antipsychotic drug use for long-stay residents, including additional residents for whom these drugs are inappropriate. 

Disadvantages of this Approach:

• It reflects facility-wide information, not the specific subpopulation of primary concern: residents who have dementia and take antipsychotic drugs.  The report of facility-wide data masks the high proportion of all residents with dementia who are given antipsychotic drugs.
• The National Partnership looks only at long-stay residents, not all residents in the facility who are receiving antipsychotic drugs.  Short-stay residents are excluded from the data reported by the Partnership.
• MDS data are self-reported by facilities and not audited by CMS.  They are often inaccurate, as CMS’s recent analysis reported.[7]

CMS MDS 3.0 Frequency Report, 4^th Quarter, 2014

This report shows MDS data by section of the MDS.  It indicates that nationwide, 22.42% of all residents received antipsychotic drugs “during the last 7 days or since admission/entry or reentry if less than 7 days” in the fourth quarter of 2014.[8]  (This figure is derived from the report, which records that 77.58% of all residents did not receive an antipsychotic drug in the prior week.  100% minus 77.58% equals 22.42%.) 

The CMS database also reports, by state, how many days residents received antipsychotic drugs in the prior week, ranging from one to seven.  The MDS 3.0 Frequency Report indicates that 79.27% of the residents receiving antipsychotic drugs received them for seven days in the prior week.

Advantages of this Approach:

• It reflects facility-wide information and includes short-stay as well as long-stay residents.

Disadvantages of this Approach:

• It does not focus on the subpopulation of primary concern – residents with dementia.
• MDS data are self-reported by facilities and not audited by CMS.

The GAO

The GAO used Medicare Part D payments and residents with dementia to identify the number of residents taking antipsychotic drugs, with the drugs paid for by Part D plans.

Advantages of this Approach:

• It focuses on the subgroup of residents who are of primary concern – those who have dementia and no diagnosis of mental illness. 
• Part D payments are more reliable than facilities’ self-reported MDS data.

Disadvantages of this Approach:

• It underestimates total antipsychotic drug use for residents with dementia because it does not count residents in a Medicare Part A stay in a skilled nursing facility (SNF).  For Part A residents, antipsychotic drugs are included in the per day rate paid to the SNF; that is, the drugs are not separately billed to the Medicare program or to a Part D plan.
• It underestimates total antipsychotic drug use for residents with dementia because it looks only at long-stay residents (defined as over 100 cumulative days), not, in addition, at residents whose stay is fewer than 100 days.
• It underestimates total antipsychotic drug use for residents with dementia because it does not include long-stay residents who do not have a Part D plan.
• It underestimates total antipsychotic drug use for residents with dementia because it does not consider antipsychotic drug use for Medicare beneficiaries who are in managed care plans that include drug coverage.

Medicare Part D Compare Website

The Medicare Part D Compare website reports atypical antipsychotic drug payments for residents in nursing facilities through their Part D plans.

Advantages of this Approach:

• It focuses on the subgroup of residents who are of primary concern – those who have dementia and no diagnosis of mental illness. 
• Part D payments are more reliable than facilities’ self-reported MDS data.

Disadvantages of this Approach:

• It underestimates total antipsychotic drug use for residents with dementia because it does not count residents in a Medicare Part A stay in a SNF. 
• It underestimates total antipsychotic drug use for residents with dementia because it looks at long-stay residents (defined as over 100 cumulative days), not, in addition, at residents whose stay is fewer than 100 days.
• It underestimates total antipsychotic drug use for residents with dementia because it does not consider antipsychotic drug use for Medicare beneficiaries who are in managed care plans that include drug coverage.
• It underestimates total antipsychotic drug use for residents with dementia because it does not report use of conventional antipsychotic drugs.  These are the older antipsychotic drugs, such as Haldol.

Conclusion

CMS should report accurate, audited facility-wide data on antipsychotic drug use in order to identify how widely these drugs are used in facilities.  The Partnership should report accurately and focus on all residents with dementia who take antipsychotic drugs in order to address the inappropriate use with residents who are at greatest risk.

Reducing the inappropriate use of antipsychotic drugs is an important policy objective, leading to better care for nursing home residents and cost savings to the Medicare program.

March 2015 – T. Edelman