“Who Ordered That? The Economics of Treatment Choices in Medical Care,” by Amitabh Chandra, David Cutler, and Zirui Song is a worthwhile read. Below are my highlights. All are direct quotes.
- [T]here are over 7,000 cardiology guidelines for individual clinical decisions. Only 11 percent are based on randomized controlled trials, and 48 percent are from expert opinions, case studies, or prior standards of care. […] There are over 4,000 infectious disease guidelines.
- None of the theories for which there is a lot of evidence can be shown to explain a major part of cross-individual or cross-area variation in treatments.
- [C]ost sharing affects whether a person gets into the system, but not what happens once a person is in the system. To take an example, cost sharing might influence whether a person with chest pain sees a cardiologist, but not what services the cardiologist performs once care has been initiated.
- When cost sharing increases, people use fewer services, but the services foregone are neither uniformly valuable nor wasteful.
- Higher cost sharing deters recommended preventive and chronic care, which may lead to undesirable “offsets” in greater use and spending on other services, such as hospital care. […] Raising costs for prescription drugs increases hospital costs, and lowering costs for preventive care has only a modest effect on utilization if people need to see their primary care physician before accessing preventive care.
- [R]egional variation is extensive in the Medicare population. People living in areas with the highest quintile of spending use, on average, 50 percent more care than people living in areas with the lowest quintile of spending.
- [V]irtually none of these regional differences is accounted for by differences in area income or poverty rates.
- Health status differences explain about 18 percent of the difference in spending across areas, and the rest is unaccounted for.
- [T]he combined impact of supplemental insurance on regional differences in spending is small. [Zuckerman et al. (2010)] estimate that income and supplementary insurance together explain 1 percent of the higher spending in high-cost areas compared to low-cost areas. Thus, while price and income matter for spending, they are unlikely to explain a large part of why spending differs so much across people or areas.
- Our conjecture is that differences in preferences do not explain a large part of treatment variation—not because preferences do not differ, but because they are frequently not accounted for in actual treatment decisions.
- The literature is clear that providers respond to payments, and that the response can be very large.
- [T]he physician that a patient sees matters far more for treatment than the patients’ preferences.
- Among cardiologists given patient vignettes, whether their colleagues would have ordered a cardiac catheterization in the same situation predicts whether respondents ordered a catheterization.
- [E]vidence from Cesarean sections in Florida shows that while physicians do learn from other physicians, residency programs explain less than 4 percent of the variation in rates of operations (Epstein and Nicholson, 2009). Physicians did not seem to update their prior beliefs, even newly trained physicians, producing a within-area variation that approximately doubled between-area variation.
- An example of stinting is obstetricians avoiding high-risk women, for fear that the baby will be impaired and they will be blamed. Negative defensive medicine would arise in a malpractice environment where adverse events such as injuries receive compensation even though malpractice had not occurred. The combined effect of defensive medicine and stinting has been estimated in several studies (Brennan et al., 2004; Localio et al., 1991; Mello et al., 2010). These estimates are not without difficulty: measuring malpractice pressure in an area is difficult, and finding an exogenous measure of that is harder still. Even so, studies have surmounted this problem using area variation in malpractice premiums or other measures of malpractice pressure such as the size of indemnity payments. The results show a surprisingly small net contribution of malpractice concerns to what physicians do. Mello et al. (2010) estimate that medical malpractice and efforts to manage its risks cost the national health care system more than $55 billion a year, about 2.4 percent of annual health care spending. The study summed various components of the medical liability system, including payments made to malpractice plaintiffs; defensive medicine costs; administrative costs, such as lawyer fees; and the costs of lost clinician work time. Defensive medicine costs were the largest segment of total malpractice spending and amounted to approximately $45 billion a year annually. In other work, Baicker et al. (2007) note that malpractice pressure increases the use of imaging procedures but exerts a small overall effect on total spending, perhaps because of the presence of negative defensive medicine.
- In a systematic review of 35 studies, both providers and patients understood natural frequencies better than probabilities in the presentation of risk (Akl et al., 2011). For example, a 50 percent risk reduction was perceived to be substantially larger than an absolute risk reduction from 2 to 1 percent. Statistics presented as number needed to treat were least persuasive, such as 100 people treated to prevent one case of the disease. Perhaps because of this, there is widespread overly aggressive screening and treatment in prostate cancer (Drazer et al., 2011; Schroder et al., 2009).
The Akl et al. paper comparing persuasiveness of different ways to present risks is in my pile.
