World Usability Day 2006
Landmark College, Putney Vermont
Conference notes by Meg Houston Maker
"Usability of Prescription Pharmaceuticals"
John Urquhart, UCSFHow does one get accurate data on patients' use of prescription pharmaceuticals? One can take behavioral approaches, such as looking at patients' diaries, questionnaires, interviews, histories, etc. But only about 7% of the diary entries are done at the time of actually taking the medicine. You can do ad hoc methods, such as counting returned or untaken tablets, monitoring refills, or looking for chemical markers. Or you can monitor the medication events, meaning the entry into and exit from the package, which can be electronically time-stamped. This last is an indirect method, but it's highly reliable in most situations.
One early implementation of monitoring medication events was an eyedropper bottle (the medicine was for glaucoma) from 1976, which had a battery and magnet that would record, by flipping a switch, the removal of the cap and the tipping over of the bottle. Now, that doesn't mean the eye drop went into the eye, but the data from this unit was, at the time, breathtaking to the medical community. The bottle cost about $500 and weighed a lot, but by 1986 the electronics had been redesigned to be easier to use, smaller, lighter, and a bit cheaper at about $400.
The principle was later applied to oral dosage forms, including the Ortho dial pack for oral contraceptives. These were about $300. By 2000 there was an $80 bottle that monitors dosage. The only hitch is that with this, the patient has to bring the package back. Now they're developing units that have cell phones that let the data be sent back to a central system.
All of this needs a robust taxonomy. The process includes prescription, acceptance by the patient, execution of the drug regimen, and discontinuation (for whatever reason). So basically the process has a beginning, middle, and end, and the parameter "persistence" is how much they adhere to the regimen, and the parameter "quality" is how well they perform the regimen.
Persistence, over time, initially drops off precipitously as you lose those patients who don't accept. And then persistence gradually dwindles over time, so that by the time you're out one year, you've lost about 40% of patients, even for medicines that are for chronic conditions. On day 100, about 80% are still engaged in the regimen, but only about 60-70% of patients took the dosage properly. That gap is the gap in execution, and this can lead to other, larger problems, especially drug-resistant strains for drugs that create selection pressure.
These averages give you some idea of the herd, but doctors treat individuals, so you have to look at individual patients in order to perform good medical care. Urquhart shows individual patient data revealing certain patterns—from very steady and persistent, to weekend irregularities, to a "wobble" that may reveal a patient is about to discontinue.
Are variations in adherence a practical problem? We have examples from several major public health issues, from TB to oral contraceptives to HIV/AIDS. TB reached an apex of cases in 1992, and the only thing researchers could think to do was bring the patients in four times a week and watch them take their medicine. This worked, but that's a brute force solution not useful for, e.g., HIV care, because the doses are so toxic. The Norplant implant for oral contraception had a much better success rate than daily pills: 0.05% failure versus 0.1% for perfect use of oral pills or 5.0% failure for "wobbly" usage of oral pills. The same is visible in use of protease inhibitors for HIV.
The strong focus on the dosing process appears to reinforce patients' grasp of the importance of treatment, with longer persistence with treatment. This falls under the rubric of measurement-guided medication management. Showing the patient how they're doing, and intervening to show them what is optimal during an intervention and discussion session, will make adjustments. Doctors have to keep at them, though, to keep encouraging them to take their doses consistently.
In conclusion, usability of prescription pharmaceuticals is disappointingly low, but can be improved by a simple management technique, grounded in objective data on the patient's record of use. The issues are treatment costs, risk, and costs of failed treatment. Ideally the doctors would not administer this MGMM process—it would be managed by the pharmacists, nurse, or paramedicals. This nascent discipline is called pharmionics.
Comments