For decades, experts and policy wonks have argued that Health care is a uniquely inefficient industry, insulated from conventional market forces that operate in the rest of the economy. Patients are believed to be uninformed about Hospital quality, insurance reduces the incentives to shop for better deals, and government programs aren’t sufficiently responsive to quality. This argument for “health care exceptionalism” supports the view that hospitals with superior outcomes aren’t rewarded in the form of landing more patients, which stymies innovation and quality-improvement efforts. Poorly performing hospitals do not feel pressure from patients to improve quality because standard market forces do not apply to health care.
What is problematic with this assessment is that there has been little work done to prove it. There are strong reasons to think that the conventional wisdom is right but also many reasons to believe that it is untrue. Markets are good at aggregating information and disseminating it to consumers who will use it sensibly — in this case, by picking hospitals with high quality. A hospital’s reputation may reflect valuable information about its performance and may spread widely without quality reporting or prodding. Embedded in this idea is the parallel notion that consumers aren’t stupid, that they’re able to process information and not get systematically duped by bad information or advertising. This is the idea of “market learning” that economist and Nobel laureate Frederick von Hayek was celebrated for.
Innovating for Value in Health CareSponsored by MedtronicExploring cutting edge ways to lower costs and improve quality.
Our research group explored whether health care is uniquely different and found evidence that market forces operate in health care with surprising consistency and impact. We assessed two separate but related components of market forces in health care. In our first study, we considered whether higher-quality hospitals had more market share at a point in time and whether they experienced more growth over time when compared to their lower-performing counterparts. Both should be true if market forces operate in health care and not true if patients or their physicians haphazardly select hospitals or have no idea which hospitals are better or worse.
We found that higher-performing hospitals do, in fact, control greater market share at a point in time and do experience higher growth in that market share over time than do their lower-performing competitors. We established this by looking at five high-volume conditions in the Medicare system where patients are free to go wherever they like: acute myocardial infarction (heart attacks), congestive heart failure, pneumonia, and hip and knee replacements. We defined hospital performance using outcome-based measures such as risk-adjusted survival and risk-adjusted readmission, process of care measures (whether specific, well-defined, evidence-based practices were being used for each condition), and patient-reported satisfaction (as measured through surveys). We found that for the conditions studied, higher-performing hospitals, as measured by outcome-based measures and process-of-care measures (but excluding patient-reported satisfaction), do, in fact, have greater market share and their market shares also tended to grow.
This correlation between performance and market share supports the idea that patients are not simply passive pawns when receiving medical care. Rather, our findings support the idea that patients do exercise the ability to place demand-side pressure on care providers. This fact is reiterated by our finding that even for emergency conditions, Medicare patients are willing to travel to higher-performing hospitals to receive better care.
For heart attacks, patients studied were willing to travel 1.8 miles more to receive treatment at a hospital with a risk-adjusted survival rate that was 1 percent higher. Similarly, we found that the relationship between hospital performance and market share was more robust for patients who had greater choice in which hospital they chose for treatment — for example, patients who did not come to the hospital in an emergency setting, reinforcing the idea that many patients are constrained in their choices.
The relationship between hospital performance and market share has also positively contributed to the health of patients. We found that reallocation of patients to higher-performing hospitals has contributed to patient-survival gains: Twenty percent of the improvement in heart-attack survival for Medicare patients in the period 1996-2008 can be explained by reallocation. In-hospital technology and treatment improvements alone were not enough to explain this improvement in survival.
In our second study, we assessed whether productivity dispersion in the health care sector is similar in magnitude to that of other manufacturing sectors. Productivity dispersion in an industry measures the variation in productivity among the players within a market. If it’s larger in health care — as proponent’s health care exceptionalism have argued — then one is likely to find high- and low-productivity hospitals in the same city. In health care, productivity dispersion is large: We found that Medicare patients treated for acute myocardial infarction (AMI) at a 90th percentile hospital had a survival rate 1.55 times higher than those treated at a 10th percentile hospital.
But if this dispersion is similar to that of manufacturing industries, it would lend support to the argument against health care exceptionalism. One of us (Chad Syverson) has shown that dispersion responds to greater competition. This means if market forces are at work in health care, they should reduce variation in productivity by removing low-quality providers or creating incentives for them to improve. To put it another way, the more competition there is in health care or the easier it is for consumers (patients) to switch suppliers (hospitals) relative to another industry, the tighter the distribution of performance in health care relative to the other industry should be.
To assess this possibility, we measured whether productivity dispersion across hospitals (measured as survival accounting for patient risks and the intensity of treatment) for the treatment of AMI was similar to that of companies producing ready-mixed concrete (measured as cubic yards of concrete adjusted for such inputs as labor, capital and other intermediates).
The comparison of health care to concrete may initially strike some as nonsensical, but as we discovered, the contrast is more informative than meets the eye. Prices for ready-mix concrete aren’t set by government reimbursement systems; concrete is also homogeneous unlike hospital care. But just like health care, it is consumed locally: A general contractor in Boston will not order ready-mix concrete from Providence, Rhode Island, in much the same way that a heart-attack patient in Boston will not travel to Providence. Surprisingly, we found that productivity dispersion within the two markets were relatively similar.
What our studies suggest is contrary to the conventional wisdom that the health care sector is insulated from consumer pressures the opposite largely is true. It means that efforts that strengthen these forces may be beneficial to patients. We have not pinpointed the precise mechanism by which patients or their physicians learn about hospital care. But in a recent New York Times piece Austin Frakt speculates that patients utilize their social network to make these decisions — in other words, that families, physicians, and friends influence where we get treated. It could also be that ambulance drivers and primary care physicians play an important role in selecting hospitals.
Separately, while we did find that patients or their agents seem to pick hospitals with higher patient-survival rates, we also found that patients do not penalize hospitals for using more tests and procedures when producing better outcomes. In other words, we found that patients were attracted to hospitals that used more inputs over hospitals that were just as good but used fewer inputs. This is not a good thing because society is paying for those inputs. This divergence between what a private citizen might desire versus what society may want is a ripe topic for efforts to increase value in health care.
Finally, our finding that demand is a key force in influencing outcomes in the health care sector means it is important to ensure that patients have the ability to make express choices. It suggests that anti-trust efforts should focus on increasing competition and that governments should not try to stall the closing of low-performing hospitals to protect jobs. Otherwise, lower-performing hospitals are more likely to be able to stay in the market, stunting innovation and worsening patient outcomes.