Wednesday, December 24, 2008

The Decreasing Problem of Drug-Drug Interactions Among the Elderly & the Role of PBMs, EHRs & Disease Management. Commentary on JAMA.

The Disease Management Care Blog got a holiday present from JAMA today: an article on the prevalence of major drug-drug interactions among the community dwelling elderly. This was an incredibly detailed and nationally representative study that sent researchers into persons' homes to not only ask what drugs were being taken, but the respondents were asked to go get and show the interviewers the actual drug bottles. In addition, persons were asked about over the counter (OTC) and herbal use.

91% swallowed at least one pill a day. 81% used at least one prescription medication. More than half took more than 5 different pills a day, and about 30% took 5 or more prescription drugs a day.

But what caught the DMCB's eye was the finding that 'one in 25' (or 4%) of study subjects were being exposed to a 'potential' major medication interaction. According to the authors, this corresponds to 2.2 million persons being at risk, which caught the eye of the national media here and here. Sounds like a lot.

The DMCB thinks the real newsworthiness of this report is how low the incidence is. To the DMCB's knowledge, a comparably performed study of outpatients looking specifically at drug-drug interactions doesn't exist. Only half of the drug-drug interactions in this study involved prescription drugs. Contrast that with some representative past studies: drug-drug interactions were more common at 6% in the past among Veteran's Administration outpatients, and among inpatients in Arizona the rate among admissions was 6.4%.

Unfortunately, the authors didn't ask the survey respondents if they received their drugs through an insurance plan, if their prescribing physicians used an eletronic health record (EHRs) or participated in a disease management (DM) program. That's because the data bases of pharmacy benefit managers (PBMs) are being successfully used to identify and prevent interactions. While the DMCB is no fan of EHRs in general, they are good at spotting prescription mishaps. Last but not least, disease management - using registries combined with 'live' person alerts for the prescribing physician - have also been effective in preventing injury.

The DMCB suspects the prevalence of drug-drug interactions nationwide is dropping and it thinks that's because of the market penetration of insurance coverage of medications using PBMs, clinicians' use of EHRs and the activities of DM programs. That's good news.

Post script: This JAMA article also identifies the potential for drug-OTC and drug-herbal problems, which accounted for more than half of the interactions. In the 'real' world of clinical practice, this is very hard to follow because patients (in the opinion of the DMCB) frequently change these agents. While EHRs and the practice of asking patients to tediously list every pill they use at every clinic visit (chewing up precious minutes in a high volume patient 'throughput' setting), a better approach may be covering these agents in pharmacy benefit plans. This is a radical notion, but the coverage doesn't have to be generous. In exchange for the insurance expense, the underlying PBM and DM data bases should be able to spot the other 2% of elderly individuals who are unnecessarily exposing themselves to ills from their pills. While some may be shocked, SHOCKED at the notion of insurance coverage for unproven therapies, the DMCB finds distant public policy parallels here and here.

No comments: