How soon might comparative effectiveness offer significant change? And which entities will guide these changes when they begin? Inside the nation's capitol, CER has become something of a mantra among those hoping to drive down health care costs. But beyond the Beltway, CER appears not to be nearly as potent a concept, for now anyway.
To gauge the extent to which CER is perceived, the National Pharmaceutical Council, a policy and research organization supported by pharma, surveyed 111 people from federal agencies, consumer and trade groups, insurers and academics, among others. And NPC found nearly 60 percent are "very familiar” with CER, but only 30 percent believe CER will lead to moderate improvements in health care decision-making in the next year.
One reason for the lack of near-term optimism may be that the Patient-Centered Outcomes Research Institute, the non profit created by health care reform to oversee CER, is simply too new to have made an impression on many radar screens, according to NPC president Dan Leonard. "It is still very early in the implementation," he says in a statement, but insists "there is a clear recognition that the work isn’t done on CER."
"What’s unclear is whether the 70 percent of respondents who believe there will be only slight or no improvement over the next year would offer a different answer if they were asked about changes over the next three to five years or beyond," according to the NPC summary. "We believe it is a fair assumption to say that respondents are taking the long view when it comes to CER, and not immediate results."
Here are some findings: about 55 percent of survey respondents were “very familiar” with CER; 68 percent believe it will take less than three years to establish research standards for CER Studies; 85 percent felt that CER led to little or no improvement for health care decision-making in the past year; 11 percent thought it led to moderate improvements, and 30 percent believed that CER would lead to moderate improvements in health care decision-making in the next year.
Also, 78 percent believe the Agency for Healthcare Research and Quality will establish research standards and 64 percent look to the National Institutes of Health to do so, but only 13 percent think the pharmaceutical industry will have this role (we should note that respondents were permitted to cite more than one entity in answering this question).
Similarly, most believe the AHRQ and NIH will establish research priorities, although 63 percent look to PCORI to fill this function and 31 percent believe the pharmaceutical industry will do so, or at least play a larger role. When it comes to conducting research, 85 percent see academia doing most of the work, followed by 62 percent looking to the pharmaceutical industry (you can read more here).
pic thx to purpleslog on flickr
9 Comments
Comparative effectiveness findings aren't going to come from agencies looking to evaluate different items. It is going to come from data mining of electronic records from companies like Aetna, Medco, and Wellpoint. In fact it already is/has.
http://www.pharmalot.com/2010/06/wellpoint-comparative-effectiveness-and-boniva/ http://www.fiercepharma.com/story/plavix-vs-effient-shows-payer-power/2009-11-20
It will be up to agencies and pharma to confirm / debunk the comparative effectiveness findings. I hesitate to call data mining research. It is, but not with the connotation that many here would likely consider as research.
Zz
Agreed. As for academia, do I hear any donors out there willing to cough up $10-30 million dollars for a SINGLE, adequately statistically powered large scale Phase IV study, say for example comparing two or more oncology drugs, or two or more drugs for Class IV CHF?
Um, I didn't think so.
industry insider-Yes, this seems pretty farfetched. Perhaps the government would fund this research with the rationale that the findings could be used to cut costs in the long run. Would have been easier to construct this scenario before last November than it is now.
How about the gvt mandating the CER studies be done by industry to agency specifications? What the government wants the government gets. As long as the FDA has the power of approval and CMS power of the purse they can request whatever they want and pharma will have to comply.
as long as you don't compel folks to work in pharma old gal.
So long as the established agencies conduct CER using the old style, traditional clinical trial approach (drug A v's drug B) then CER will fail to deliver any meaningful results. This is because the true benefits of CER come from the ability to assess what happens in real patients, and here the complexity of the range of alternative conditions is so great that no traditional trial could be designed and run that would be both fundable and likely to produce meaningful results in a practical time.
Here Zz is absolutely correct. The use of advanced analytics in conjunction with large scale EMR data is not only capable of producing meaningful results for specific CER questions, but is also capable of producing meaningful results across multiple patient types with multiple co-morbidities. Funding is much less of an issue as the costs are less than 10% of a phase IV study, analyses are totally replicable across patient types and can be run on a near-continuous basis. If industry adopts this approach then CER will be making substantial contributions within a couple of years. Stick with the old approach and CER will never make a real contribution.
I am still waiting for the day when a generic drug has to prove comparable efficacy vs it's branded comparator.
Frank, that day will never come because generic companies, except for a few big ones do not have the financial resources to conduct efficacy studies. For the FDA to to compel them do so would require rewriting the entire Waxman-Hatch Act of 1984, and that, I'm afraid ain't about to happen anytime soon.
It occurs to me that this information may be available as a by-product of an existing initiative. The SENTINEL program was designed for safety studies but the vast well of data generated could be used to generate CER as well. It would not be easy but it would be possible.
Reference: http://www.fda.gov/Safety/FDAsSentinelInitiative/default.htm