Making good health care decisions has always been hard. As informatics teases out patients’ varied health needs, it’s poised to get even more complex. And doctor-patient visits are still only 10 minutes long.
It’s an open secret that half of people taking certain long-term, high-cost drugs won’t benefit from them.
As patients are being asked to shoulder more of the costs, and decisions, for their health care, they face tougher choices about whether the dollars for their daily doses are well spent.
Researchers are starting to figure out which treatments work for which people, but those answers will come as statistics, not certainties.
They promise to cut expensive, unnecessary care, and prevent more-expensive, more-harmful medical injuries.
They will do neither, however, unless doctors are given more training, support tools and time with patients.
Getting the right treatment to the right patient to the right time can save lots of money.
Patient advocacy groups and companies hoping to sell tools to should start lobbying now to get resources to doctors so that they can actually use these tools. More time with patients seems an obvious place to start.
At a conference on personalized medicine this week in San Francisco, Brad Margus, CEO of Perlegen Sciences, described the kind of data that will soon be facing drug regulators and prescribers.
In one example, an effective drug with very unpleasant side effects. (Unfortunately, Margus could not disclose the actual drug or side effects.)
Perlegen found about 20 genetic variants that predict how patients respond to the drug.
Imagine yourself, shivering in your underwear, as your doctor offers this explanation to help you decide whether to take a drug:
If doctors do not prescribe the drug to people with 16 or more of these risky genetic variants, or predisposing alleles, they would exclude 91 percent of patients prone to side effects, but also exclude 40 percent of those likely to benefit without suffering.
Give the drug to everyone with fewer than 21 predisposing alleles, and 97 percent of the patients poised to benefit will get it.
So, too, will 60 percent of the patients bound to suffer. You can call your insurance provider later to figure out how much the treatment will cost.
Granted, many diagnostics will have more clear-cut answers than the above scenario.
Still, both doctors and patients are going to have to deal with more information and ambiguity than they’re used to. Genetic analyses are arguably the furthest along, and few consider effects of age, weight, disease severity and the like.
Of course, patients can educate themselves. One study from University College Londonindicated that an interactive Web site help patients manage chronic disease.
But not all Web sites are out to help. Already, hucksters are angling.
Last week, the Electronic Privacy Information Center submitted comments to the FDAwarning that direct-to-consumer marketers were collecting information to target mature, “impulsive” individuals with particular ailments. (Not all of this is collected on-line.)
Personalized marketing of medicines leaves patients at bigger risk of scams, and more in need of longer chats with their doctors.
But doctors themselves will need help weighing all the available information. One potential, partial solution is clinical decision support systems. Ideally, these incorporate both the latest information and myriad patient characteristics. At least one company is working on a “virtual patient” system, that will run in silico simulations for patients under varying conditions and treatment regimens. But these systems still need to be improved, and doctors need to know when to trust the system, and when to go with their guts. (One small study found that younger doctors relied on such systems too much, older doctors, too little.)
Even so, systems must supplement, not supplant, quality time with patients. I read once how a conversation with a doctor eased a woman who’d suffered months of back pain.
It turned out that the patient owned no furniture and slept on the floor. She had only one pair of shoes, and they were heels.
A trip to social services for a bed and sneakers turned out to be much cheaper, and more effective, than the assistance she’d been receiving.
It took time and compassion, not RAM and ROM, to figure that one out.
M.L. Baker is health IT and biotechnology editor for Ziff Davis Internet’s Enterprise Edit group. She can be reached at [email protected]