By Kay Lazar
“A rampant prescription, a hidden peril” aimed to pull back the curtain on a long-running but shrouded practice in many nursing homes of using antipsychotic drugs to sedate residents, particularly those with dementia who often have challenging behaviors.
Research published in early 2010 found that patients newly admitted to nursing homes with some of the highest rates for prescribing antipsychotics were significantly more likely to receive the drugs than patients entering homes with the lowest prescribing rates – regardless of whether they had conditions that warranted use of the drugs.
That made us wonder: How did Massachusetts nursing homes stack up?
Not well, I discovered. Using publicly available data on the Centers for Medicare & Medicaid (CMS) website, I found that Massachusetts nursing home residents were far more likely to be given powerful antipsychotic drugs that were not intended or recommended for their medical condition, compared to most other states.
That March 2010 story prompted an avalanche of e-mails from readers who wanted to know about the prescribing practices at specific nursing homes – information we could not give them because federal regulators would not release it.
So my colleague, Matt Carroll, and I submitted a Freedom of Information Act request to CMS for data on each of the nation’s 15,000-plus nursing homes. We had no hint of the 19-month odyssey that would follow, trying to pry the information out of the agency.
Our aim was to identify factors such as staffing levels or facility size that are linked to a nursing home’s inappropriate use of these powerful sedatives, which can be lethal to elderly people with dementia. And we wanted to provide readers with an interactive database that would show antipsychotic use in facilities in their communities.
But CMS took the position that facility-specific information regarding antipsychotic use should not be public because, agency officials said, it violated patients’ privacy. Never mind that CMS publishes on its Nursing Home Compare website detailed information about the percentage of residents in each home who lose control of their bowels or bladder, are physically restrained, have pressure sores, etc.
The FOIA we submitted requested information about antipsychotic use at each of the nation’s nursing homes from 2005 through 2009 (which was updated to include 2010 data because of the long wait) and also detailed information about each facility’s staffing levels, total number of residents, type of ownership (for profit, not for profit), etc.
During the 19 months that we hounded the agency for the data, several patients’ rights groups were also lobbying CMS to post the information on its Nursing Home Compare website. We were variously told by CMS that we would have to wait our turn – that there were hundreds of requests submitted before ours – or that the agency was still debating whether and what to make public and even that they had simply lost our request.
Then, in late September 2011, CMS administrator Don Berwick came to the Globe for an editorial board meeting. After that meeting, I made a personal plea for the long-stalled FOIA data, suggesting that the “transparency” that Berwick had championed during his conversation with editors about reforming America’s health care system should also extend to requests for information about that system.
Berwick seemed genuinely surprised to hear about our FOI impasse.
We received our data a few weeks later – 90,000 records in seven databases – and we felt a bit like the dog who chased the fire engine; thrilled to have it but not entirely sure how to proceed.
Carroll spoke with statistics experts to ensure that the way he was planning to analyze the data was sound. He merged the material into one database and sorted the nursing homes by the percentage of patients who received antipsychotic drugs contrary to CMS recommendations. The homes were broken into quartiles and a median was calculated for each quartile.
I talked to experts who specialize in nursing home research to pinpoint the characteristics we should focus on in our analysis, ultimately choosing staffing levels, method of payment (Medicaid vs. Medicare), and percentage of residents reported by staff to have behavior issues.
We had some false starts.
After the first round of analysis, we realized CMS inadvertently sent 2005 data twice, but labeled one of the years 2007. They sent corrected data, and Carroll re-did the analysis.
Something still didn’t look right. Turns out the agency sent us raw staffing data, but told us it was already calculated in a pre-set and rather complex formula. (CMS counts the number of hours each staffer works over two weeks, then divides that by 14, then divides that by the total residents of a home to compare facilities’ nursing hours per resident, per day.) They apologized, sent the formula, and Carroll re-did the analysis again.
The results showed wide disparities, with heavy antipsychotic use in many homes, and virtually none in others. It also showed that higher drug use typically equated with lower staffing, more residents covered by Medicaid, and more residents reported by staff to have “behavior” problems.
But finding families willing to speak publicly about their experiences in nursing homes, especially in facilities with high antipsychotic rates, proved challenging. Some, contacted through an advocacy organization, agreed then backed out, fearful of offending nursing staff, and potential retribution. Local ombudsmen referred our calls to the state regulatory agency, which was not helpful. Ultimately, we found some families willing to be interviewed through elder law attorneys.
We told the story in two parts to keep each concise. Day one focused on our findings and used one family’s experience woven through the story to humanize often complex issues. Day two was devoted to alternative methods to antipsychotics for calming agitated residents.
The day part two was published, we received a letter to the editor from the state’s nursing home trade association saying that it was joining its national counterpart in a goal of reducing antipsychotic use by 15 percent by the end of this year. It was heartening to hear that our series may help improve the quality of life for thousands of nursing home residents, but we also know that the industry has much work to do to hit this goal.
CMS has also announced that it will post facility-specific antipsychotic information on its Nursing Home Compare website in July. [Don’t miss a special webinar on July 18 for AHCJ members about the data.]
On June 8 we received a second batch of documents – requested in an April FOIA to CMS – that we hope will help us take a closer look at nursing homes cited for unnecessary medication use, antipsychotics and other drugs, during annual inspections. As was the case with the first go-round, there are problems with the data – it is missing some years. We have fingers crossed that this wait for CMS data will not be another 19-month odyssey.
