Comparative Effectiveness Research (CER) is being talked about more and more as a fulcrum for controlling healthcare costs. For example:
- The Congressional Budget Office issued a report on CER in December 2007 and has highlighted it in more recent analyses and reports about health reform options
- The ARRA legislation included $1.1 Billion for CER
- ARRA included language for the IOM Committee on Comparative Effectiveness Research Priorities to provide a report by June 30, 2009 about how to spend the $400 million allocated to HHS for CER.
All this discussion has kept me thinking about how CER will be done, how the results from this research will actually be used to improve quality and reduce costs, and what are the scope of healthcare issues that CER is, will, or should be applied to help improving.
While understanding what works best in healthcare is certainly a worthy goal, this is far from a simple task. Some of the factors that complicate research to compare the effectiveness of various treatment options are:
- Gold-standard double-blinded trials for clinical research provide information about efficacy – which is different than effectiveness.
- Observational research can provide information about real world effectiveness, but the information from this type of research can be flawed by problems in the data – including selection biases and other conclusion skewing factors.
- Both these types of research methodologies inherently have a lag between the time the research project starts and the time the data is analyzed and conclusions developed. This time lag is often several years, during which new treatment options will likely have been developed. Thus, CER really is only answering questions about the most effective treatment options when the study began, not when its conclusions are presented.
- There is also considerable controversy about what factors to compare in CER projects. That is, should only clinical outcomes be compared, or should costs be a factor? And if cost is a factor, how are indirect costs, such as diagnostic testing, office visits, patient’s time, etc., included? And how is quality of life valued? (Some CER analyses report results according to Quality Adjusted Life Years).
- And of course people interested in biopharma and medical device innovations are concerned that CER will be used not just to inform clinicians and patients, but to justify coverage and payment decisions which will impact R&D in therapeutic areas where reimbursement for innovative products is denied or limited.
All of these factors point towards larger issues of how to ensure that medical practice is maximizing knowledge to optimize clinical care for the good of the patient and society. In some cases, this is termed Evidence Based Medicine (EBM). In theory CER should support good EBM ASAP. And from what President Obama has said, he wants this done PDQ.
Health System CER & Evidence-Based Interventions
While all these challenges for CER are ongoing, there also seems to be opportunities for applying the principles of CER and EBM to more system wide properties of the US healthcare system to increase value and efficiency. For example – and I hope I’m not beating a too tired horse here – but the surgical checklist (and similar quality improving activities) have been shown to increase quality and reduce costly events, but not all hospitals and clinicians are using them. Therefore, how about research to compare the effectiveness of hospitals (or surgeons, etc.) that use and don’t use such practices? Some people might say that we don’t need this research since the value of these practices is already known, but perhaps focused research highlighting this information will serve as a big push to get the laggards on-board.
Similarly, CER type analyses could be applied to Medical Homes to determine what characteristics and capabilities of Medical Home medical practices make them better at improving the quality of patient care and controlling overall spending. In particular, there might be specific features of Medical Homes that would be most important for diabetics, and others for patients with CHF, etc. And currently NCQA’s 3 tiers of Medical Homes build upon each other, but don’t permit greater granularity nor do they distinguish between potentail patient populations. This research might be complicated, but with initiatives such as Medical Homes being proposed as a way to redesign and reconfigure outpatient care in the United States, more focused research beyond the existing and planned demonstrations and pilot projects might be very worthwhile expenditures.
P4P for Cost Containment
Another big push for cost containment in health reform is pay-for-performance (a.k.a. P4P). While the knowledge gained from CER could certainly be fed into P4P practices, P4P itself has some controversy about how well it does or does not work to change behaviors to improve quality and reduce costs. At a breakout session about P4P at a conference on Friday led by Bob Galvin, MD (GE’s Director of Global Healthcare), I stated that two basic criteria that P4P interventions need to be successful are:
- The group effected needs to be small enough that each individual feels that changing their actions will effect their compensation, i.e., a group of 500 clinicians is too large, but it most likely can be larger than 5.
- The information about how the group or the individual is doing towards any P4P goals is delivered often enough to provide timely feedback, i.e. once a year is not frequent enough, perhaps quarterly is OK, and monthly would be great.
Dr. Galvin pointed out that the size of the P4P incentives also needs to be significant, i.e. it can’t max out at $100 per clinician. And another participant noted that the P4P measures need to be controllable in some way by the clinician. For example, while patients’ seat belt use might be somewhat influenced by clinicians’ reminders and admonishments, clinicians are much more able to see that their diabetic patients are getting regular HbA1C testing, eye exams, and appropriate immunizations.
Coming Full Circle From CER to P4P
18 years ago I coauthored a book chapter about the structure of bonus pools and other P4P-type incentives for physicians in nascent managed care organizations. Unfortunately, in the early 1990s, there weren’t robust information systems to provide data about “performance” for these P4P systems to be effectively implemented. Perhaps now – and in the future – as health IT matures and become well integrated into healthcare delivery, better data will be available and P4P can be productive for clinicians, patients and society.
To help make that potential a more likely reality, perhaps some of the CER efforts could also be directed toward determining how to best structure and implement P4P programs to maximally change clinician (and possibly patient) behaviors to better utilize information about what is already known to work best in medical care. And then these same P4P interventions would be in place and prepared to use the new knowledge that will come out of the expanded CER programs starting this year – and which will hopefully enable us to dramatically improve medical care and the medical system in the future.