In my email today I received an announcement of a new paper “Price subsidies, diagnostic tests, and targeting of malaria treatment: Evidence from a randomized controlled trial,” by Jessica Cohen, Pascaline Dupas, and (disclosure: my friend) Simone Schaner. They report a field experiment from Kenya, studying how people react to different prices for a new, effective malaria treatment, and how people’s decisions about whether to buy this medicine are influenced by the opportunity to learn if they, in fact, have malaria.
There are many interesting details in the paper, but one that stood out to the authors and to me is that “only 56 percent of those buying such a highly subsidized ACT at retail sector drug shops test positive for malaria.” In other words, quite a lot of people are buying – and presumably often using – medicine to treat diseases that they don’t have.
This is worse than just wasting money. Infectious diseases can develop resistance to overused treatments. Drug resistance is evolution occurring on a small scale: the germs that can survive a treatment will multiply. So, if I inappropriately take medicine to treat what I think is a disease, it makes you and everybody else a little worse off, because – if you actually get the disease – the germs will be just a little bit hardier, and the medicine just a little bit relatively weaker.
So, the authors note, one solution might be to test people: if people know they don’t have malaria, they presumably won’t buy the medicine. Part of the experiment was to randomly offer subsidies for “rapid diagnostic tests” (RDTs) for malaria. Yet, as their figure 9 shows, the subsidized tests did not increase the fraction of medicine-takers who actually had malaria by much, and in some cases not at all. That is, a lot of people still buy the medicine, even when they know (or should know) that they don’t have malaria!
This is puzzling. As the authors suggest, people might not trust the test. Or, they might simply be stocking up on discounted medicine, just in case. It appears that overmedication is not a problem that will be easily solved.
All of this surely has to be terribly bad in India, too. As Jeff Hammer and coauthors’ research has noted, antibiotics are given as the solution to everything, not least because of market pressure to appear to be doing something. We need somebody clever to find an exogenous geographic source of variation in providers’ incentives to overmedicate with antibiotics (is this better in public facilities?), and estimate effects of this unfortunate market on the spread of drug resistance (is there data on this?).