How the nursing home data were analyzed
The Boston Globe examined data on more than 15,600 nursing homes across the nation for its investigation of antipsychotic drug overuse. The information was supplied by the US Centers for Medicare & Medicaid Services, 19 months after a Freedom of Information Act request was submitted. There were two sets of data:
■The percentage of long-term residents without a psychosis or related condition who received antipsychotics contrary to federal nursing home regulators’ recommendations.
■The characteristics of each home, such as staffing levels, number of patients on Medicaid, and the number reported by staff to have behavioral problems.
The information, which CMS compiled from data reported by nursing homes, was for the third quarter of each year between 2005 and 2010, the most recent year for which data were available. Long-term residents are those in nursing homes longer than 90 days. Nursing home regulators include these illnesses as a psychosis or related condition: hallucinations, Huntington’s disease, schizophrenia, and Tourette’s.
The information from the two tables was merged to create one database. For some nursing homes, there were multiple entries for a single year, or no entries. That’s because nursing homes might be inspected twice in some years and none in others. In such cases, data from the inspection closest to the third quarter of each year were used.
The nursing homes were sorted by the percentage of patients who received antipsychotic drugs contrary to regulators’ recommendations. The homes were broken into quartiles and a median was calculated for each quartile. Nursing homes with fewer than 50 residents were excluded because a handful of residents on antipsychotic drugs could significantly alter the percentage.
The Globe asked a CMS data analyst to review the findings to ensure the validity of the methods and the accuracy of the data.