NYTimes: The Evidence Gap. The Pain May Be Real, but the Scan is Deceiving

Gina Kolata has an interesting piece in the NYTimes right now, looking at the gap in evidence between use of scanning technology like MRIs and CAT scans to track down causes of pain, the discovery of a problem, and whether or not fixing the problem found via scan fixes the actual problem.

This hits home for me, because several years ago I went through a long series of scans to see what in the world was possibly causing the severe pain in my right arm. The MRI eventually discovered a disc bulge very high up in my c-spine, at an area that would have significant risks attached to surgery. Thankfully, my primary care physician, anesthesiologist and neurologist held a joint meeting with me and took the time to explain not only how to read MRIs, but the number of people who have something "wrong" with them without that "wrongness" being the cause of their pain.

In my case, terror about the risks associated with surgery high on my spinal column evaporated, and I was left with a different diagnosis and management routine for the disease.

But many people aren't so lucky, and go through unnecessary surgery because they have a pain, a scan shows a deviation, therefore the deviation must be linked to the pain. They have surgery, expecting to feel fine after your typical recovery period, and instead, no dice. (And if someone is unlucky enough to have some kind of pain condition that likes to spread when nerves are injured, things could be made even worse.)

Because of this, it's nice to see people - reporters and medical folks alike - are actually taking a look at the prevalence of "abnormal" scan results. Results that, in the end, are not so abnormal after all. However, it would have been nice to see whether or not doctors who realize that disc bulges or torn meniscus's aren't necessarily related to the pain you're feeling treat the pain, rather than the absence of a fixable issue. Knowing you have arthritis and that's the cause of your pain issues means very little if you're not getting pain relief.

Moving away from a concept of a single norm for the body is good - but hopefully as we make that transition, people recognize that simply because there is not a fixable cause to pain, doesn't mean the pain itself cannot be successfully treated.
-Kelly Hills

Clinical Trials Miss Many Populations

A new report by the Chronic Disease Prevention & Control Research Center at Baylor College of Medicine, in conjunction with the Intercultural Cancer Council, brings the grim but not unexpected news that clinical trials for testing new drugs has routinely excluded or under-represented a broad category of people, including women, minorities, the disabled, elderly, and those who don't live near major research hospitals/urban areas. (And in fact, this very subject is one of the first conversations I remember having with Kathryn Hinsch - the fact that even animal research is done on male models.)

According to the report, there are about 80,000 clinical trials run in this country every year, and only about 1% of the population participates in them (working out to around 2.3 million people). The report didn't limit its critique to the statistical information of those participating in the trial, but also criticised wasted resources and duplicate efforts between government and private funding and the lack of training of IRB members. But the major focus of the report was on the constituencies of trial populations:

The research looked at cancer clinical trials and found that only 25 percent of patients in such trials were over the age of 65. In addition, older people were often excluded from studies focused on Alzheimer's, arthritis and incontinence... As evidence of the problem, [the researchers] honed in on a study of clinical trial composition that found that, between 1995 and 1999, blacks, Asian-Pacific Islanders, Hispanics and Native Americans together made up for less than 10 percent of patients included in new cancer drug trials. Under-representation of this sort, they say, leads to results that do not account for a host of factors -- genetic, cultural, racial, religious, linguistic, as well as variables related to age and gender -- that could have a huge impact on how well new drugs do in the real world.

The researchers also acknowledge that while there has been a lot of discussion about clinical trial populations and their make-up in recent years, very little has been done to redress the issue. To that end, the report actually also offers nine concrete policy suggestions to fix/improve clinical trials in the United States:

• government regulatory changes


• increased collaboration between government and private industry on clinical trial design


• increased community involvement in patient participation


• scientific journal oversight of patient breakdowns


• new, specialized training for review boards


• reallocation of research funding to avoid duplication and address disparities


• increased public education


• increased focus on easing the patient participation process


• guaranteeing insurance coverage for all related costs.

And on a personal note, we'd like to extend congratulations to friend of the blog (and fellow blogger) Daniel Goldberg, one of the researchers on this project and the leading quote in the Washington Post coverage of the story. Daniel's off enjoying the cherry blossoms in DC right now, but perhaps when he's back from his conference, he'll step over here and talk to us a little bit more about his research.
-Kelly Hills

PatientsLikeMe -- a new paradigm?

Interesting story in Sunday's NYT Magazine. It's about a new website, PatientsLikeMe.com, that goes beyond the usual online disease support groups to actual data collection. Currently offering communities for amyotrophic lateral sclerosis (ALS, or Lou Gehrig's disease), multiple sclerosis (MS), HIV, "mood," and Parkinson's, the site gathers data on treatments and side effects and turns it into aggregate data that users can use to see how their health status, symptoms, and treatment plans compare with those of others. The company was named a Top Health IT Innovator last year; CNNmoney.com has named them one of "the next disruptors," the 15 companies that will change the world.

More than being just a "support-plus" site, though, PatientsLikeMe could have some important implications for the way health care works. First, and most obviously, it puts patients in a much more active role--and not just in terms of adhering to their doctor's instructions, but in investigating whether their prescribed regimen could be improved. Second, it could be a parallel track to current ways of doing research: outside academia, outside clinical trials, bodies of data being built up. There's already a DIY trial on lithium for ALS going on among members. One key difference here is that the subjective data patients consider salient could be different from, and not included in, existing study designs. Third, PatientsLikeMe is a for-profit company ... it's seeing pharmaceutical companies as its financial lifeblood, which raises the usual privacy (etc.) concerns.

Overall, the model could challenge our current ideas of who does what in health care; the validity of data from outside the research establishment; how IT and the internet can better serve health needs -- beyond just transferring existing processes from paper to electronic media; and the need to protect personal health information and whether the general public is, or should be, afraid of this kind of radical openness ... lots to think about.