On Tuesday I discussed a recent upbeat article in the New Old Age blog about a nursing home’s improved approach to the all too widespread problem of medicating the elderly. I also raised questions about why it has taken this long for nursing homes to start addressing this issue when overmedicating the elderly has theoretically been illegal since the Omnibus Budget Reconciliation Act of 1987.

Today I’d like to talk about a report in the American Journal of Geriatric Pharmacotherapy that undermines my hope that things are getting better. (I've never seen this journal before, but it has a good editorial board and has recently formed a relationship with the American Geriatrics Society. That is probably why I'm seeing it now.)

The article, by Dedhiya, Hancock, Craig, Doebbeling, and Thomas (a group of pharmacy researchers at Purdue, Indiana University, and the Regenstrief Center), has the somewhat understated title, "Incident Use and Outcomes Associated with Potentially Inappropriate Medication Use in Older Adults." The authors looked at the entire Indiana Medicaid population living in nursing homes in 2003 (7,594 people) to examine the use of "potentially inappropriate medications" (PIMs), according to the 2003 Beers list of inappropriate drugs for older adults. (See our previous blog on Mark Beers here and the list here.)

Their gross finding was that in the course of a year, 42.1 percent of the people had at least one PIM: 32 percent were prescribed a drug on the list of those inappropriate for older people and 18 percent were prescribed drugs inappropriate for the particular illnesses they had. (Note, some people had more than one PIM.)

The authors used several analytic steps to make their results more meaningful. To focus on new (incident) prescriptions, they excluded residents who had a PIM in the last quarter of the year prior to the study year. So, for example, the person who has been on Valium for 20 years is not included in this analysis. I've always thought that these kinds of cases, while contraindicated by the Beers criteria, deserved some slack under the principle of leaving things be. They also excluded the mere 50 people (!) who had no new drugs in the study year (personal communication). More importantly, they were able to also obtain the records of hospitalization and death for residents after the new PIM prescription, along with a wide variety of demographic and health information.

Controlling for prior hospitalization, number of co-morbid illnesses, age, etc., the authors found that 19 percent of those who had a PIM were hospitalized in the following year versus 12.4 percent of those without a PIM. And 38.5 percent of those prescribed a PIM died within a year versus 31.2 percent of those without. Now, I know that even very good statistical controls in the analysis of an observational study is not the same as conducting a real experiment (i.e., randomly assigning residents to receive PIMs) to really prove a causal relationship between inappropriate medications and hospitalization or death. However, no one is ever going to be able to do that experiment, and given that these data were collected in 2002-2004 we are already very late to be taking action.

What is going on? From prescription through review and on to dispensing, the system seems to be failing these very vulnerable older adults. And by the available evidence they seem to be getting hurt--more likely to be hospitalized and more likely to die. Even though we have a law on the books, it doesn't seem to be working. What should we be doing? Advocating for more enforcement? Encouraging the spread of voluntary cutbacks in medication among nursing homes like the one described in the New Old Age article? I’m not entirely sure, and would welcome suggestions.