Image via WikipediaI am not a lawyer, and while I have some opinions and some knowledge of FDA guidelines, I want to use the current pharma story du jour to highlight something that we need to do.
I am sure anyone who listened to NPR today heard about Wyeth‘s anti-nausea drug, Phenergan. To cut a long story short, the drug was administered to Diana Levine, a musician from Vermont, via the IV-push method, which has a small risk of inducing gangrene (if pushed into the artery). At issue, whether Wyeth should have changed the label having known about the issue. I won’t go into what Wyeth should have done, or not. That case is now before the US Supreme Court.
I want to talk about the importance of pharmacovigilance, about the need to essentially never end clinical trials. We often talk about the future (and in some cases present) of oceanography, seismology, telemedicine etc, in the context of a vast array of sensors, collecting information. The financial services industry is into using streaming data and trying to analyze it in real time. Why don’t we do something similar with health information. I have heard many talk about better ways of signal detection for adverse event reporting and for detecting infectious disease outbreaks. We’ve also talked about the importance of electronic medical records. What we need is to realize how importance these are. A system that allows constant monitoring of health related data, adverse events of all kinds, symptoms that could not be diagnosed, environmental information collected in the case of a particular disease, phenotype-genotype relationships, and have intelligent agents process these data, the kind of processing that allows us to modify labels on drugs to meet the available information, to help us develop better drugs. I know there are many reasons this is not being done today, but not to do this would be a disservice to both the pharma industry and to all of us.
It’s also important to consider the kinds of technologies we use here. Yesterday, I talked about context. That’s even more important here, which is one reason I believe that data expressed as triples, and perpetual analytics are the kinds of technologies/data-driven approaches we need to consider. An adverse event in one context is not the same as a similar adverse event in another context. We need to remember that and not rush to conclusions. I’d love to hear what Jeff Jonas might have to say about the challenges of post-marketing intelligence and pharmacovigilance.
I suppose I lied earlier. I do have one opinion to share. A label should never be set in stone. The lack of pharmacovigilance just makes it easier to take that stance.
Another reason pharmacovigilance is required
I am sure anyone who listened to NPR today heard about Wyeth‘s anti-nausea drug, Phenergan. To cut a long story short, the drug was administered to Diana Levine, a musician from Vermont, via the IV-push method, which has a small risk of inducing gangrene (if pushed into the artery). At issue, whether Wyeth should have changed the label having known about the issue. I won’t go into what Wyeth should have done, or not. That case is now before the US Supreme Court.
I want to talk about the importance of pharmacovigilance, about the need to essentially never end clinical trials. We often talk about the future (and in some cases present) of oceanography, seismology, telemedicine etc, in the context of a vast array of sensors, collecting information. The financial services industry is into using streaming data and trying to analyze it in real time. Why don’t we do something similar with health information. I have heard many talk about better ways of signal detection for adverse event reporting and for detecting infectious disease outbreaks. We’ve also talked about the importance of electronic medical records. What we need is to realize how importance these are. A system that allows constant monitoring of health related data, adverse events of all kinds, symptoms that could not be diagnosed, environmental information collected in the case of a particular disease, phenotype-genotype relationships, and have intelligent agents process these data, the kind of processing that allows us to modify labels on drugs to meet the available information, to help us develop better drugs. I know there are many reasons this is not being done today, but not to do this would be a disservice to both the pharma industry and to all of us.
It’s also important to consider the kinds of technologies we use here. Yesterday, I talked about context. That’s even more important here, which is one reason I believe that data expressed as triples, and perpetual analytics are the kinds of technologies/data-driven approaches we need to consider. An adverse event in one context is not the same as a similar adverse event in another context. We need to remember that and not rush to conclusions. I’d love to hear what Jeff Jonas might have to say about the challenges of post-marketing intelligence and pharmacovigilance.
I suppose I lied earlier. I do have one opinion to share. A label should never be set in stone. The lack of pharmacovigilance just makes it easier to take that stance.
Related articles by Zemanta