This blog has not be afraid of a fight when it comes to characterizing poorly designed CER as an excuse for rationing new technologies. On the other hand, those who oppose CER have to do a lot more than just bleat about it's negative consequences. It has to stop spending money on lobbyists and start spending money on researchers who can create a science-based path for the approval and use of diagnostics and clinical data for selecting and evaluating treatments, preventing illness and improving health. And it has to spend money to demonstrate that it can do all of the above.
The current approach to reimbursement will not suffice. Case in point, the preliminary CMS decision not to pay for gene-guided warfarin testing until long term randomized clinical trials determine if such tests actually reduce bleeding and clots over the long term. It is not the right way to go... in a long term study (> 8 - 12 weeks) it is impossible to link initial dose to long term outcomes of bleeds/clots unless one were to let INR fluctuate freely which is clearly unethical. And loading dose is what PgX testing is all about since the major risk is in overdosing in the first month of treatment.
That being the case I was glad to see the approach taken by Sean Tunis and Alan Garber in their NEJM article on how CER could be used to advance personalized medicine. I was particularly interested to see that Tunis and Garber were open to using post hoc subpopulation analysis to generate leads for benefit where one size fits all previously showed none existed We will see if this is merely a sop or the beginning of a shift in approach that is sensitive to advances in science and other considerations.
In addition, AHRQ is holding a symposium on alternative study designs to take into account heterogeneity of treatment effect.
These are well overdue signs that CER will not be manipulated overtly by political masters. However more needs to done. Money needs to be allocated to accelerate the testing of predictive markers for treatment selection and for health outcomes evaluations in prospective, observational settings. And legislation needs to be introduced to insure that the absence of evidence does not become the rationale for absence of reimbursement by insurers, public or private. That is, both should pay for new technologies as along as their is a "meaningful" effort to link use to efforts to measure value and clinical benefit and that such information and the clinical practices associated with the technology are shared.
In any event, I commend Garber and Tunis (Tunis and Garber) for their article and hope that more such efforts are in the offing.
Read more here.
The current approach to reimbursement will not suffice. Case in point, the preliminary CMS decision not to pay for gene-guided warfarin testing until long term randomized clinical trials determine if such tests actually reduce bleeding and clots over the long term. It is not the right way to go... in a long term study (> 8 - 12 weeks) it is impossible to link initial dose to long term outcomes of bleeds/clots unless one were to let INR fluctuate freely which is clearly unethical. And loading dose is what PgX testing is all about since the major risk is in overdosing in the first month of treatment.
That being the case I was glad to see the approach taken by Sean Tunis and Alan Garber in their NEJM article on how CER could be used to advance personalized medicine. I was particularly interested to see that Tunis and Garber were open to using post hoc subpopulation analysis to generate leads for benefit where one size fits all previously showed none existed We will see if this is merely a sop or the beginning of a shift in approach that is sensitive to advances in science and other considerations.
In addition, AHRQ is holding a symposium on alternative study designs to take into account heterogeneity of treatment effect.
These are well overdue signs that CER will not be manipulated overtly by political masters. However more needs to done. Money needs to be allocated to accelerate the testing of predictive markers for treatment selection and for health outcomes evaluations in prospective, observational settings. And legislation needs to be introduced to insure that the absence of evidence does not become the rationale for absence of reimbursement by insurers, public or private. That is, both should pay for new technologies as along as their is a "meaningful" effort to link use to efforts to measure value and clinical benefit and that such information and the clinical practices associated with the technology are shared.
In any event, I commend Garber and Tunis (Tunis and Garber) for their article and hope that more such efforts are in the offing.
Read more here.
