Seen statistically in hospitalized patients' complication and mortality rates, and in the percentages of individuals with chronic diseases who are kept out of the hospital with effective preventative care, these quality "outcomes" are the overlooked elephant in the room for health reform.
We recently analyzed 22 million hospital admissions
across the country, examining two dozen common conditions, including such widespread illnesses as heart disease, diabetes and post-operative infections. What we found were surprisingly large variations between the best- and worst-performing hospitals.
Patients in the worst-performing hospitals -- those in the bottom 10% -- were three times more likely to die and 13 times more likely to experience complications than those in the top-performing hospitals. The probability of dying in the hospital after a heart attack or stroke was more than twice as high in low-performing hospitals as in high-performing ones. And patients in low-performing hospitals were nearly 20 times more likely to experience IV line infections and more than three times more likely to contract post-operative sepsis infections than in high-performing hospitals.
It would not be overly dramatic to characterize some of the outcome differences as alarming, including significant differences among hospitals within the same metropolitan area.
For example, if you had called 911 for a heart attack in Phoenix, you could have been treated at a hospital with a 5% death rate or a facility with a 15% death rate, depending on which of 14 local hospitals cared for you. Among the 26 hospitals in New York City, you could have been taken to a hospital with a 4% death rate or one with a 21% death rate.
Most Americans know that hospitals vary in quality, but do they realize the wrong choice can increase their chance of death so significantly?
The standard answer from many medical professionals when confronted with such variations is that they're often caused by factors beyond their control, such as the patient's health or income. But that's only part of the story, or so we found.
After rigorously risk-adjusting for more than 80 distinct measures in patient health, demographics, socio-economics and health system factors -- including whether the hospitals were in urban or rural areas, the prevalence of smoking in the area, and so forth -- we saw that the variations persisted. Indeed, challenging conventional wisdom, the study found
some poor-performing hospitals serving mainly high-income, largely white populations and some high-performing hospitals serving primarily low-income, minority populations. This was true across the country, between states, within states and within cities. In other words, where you live might determine if you live.
We can't say conclusively why such outcome differences exist and persist even after risk adjustment. But it's clear that what happens inside a hospital matters a lot.
Measuring, reporting and acting on outcomes provides a real opportunity to avert harm, save lives and lower health care costs. Small steps are being taken in this area, such as measuring IV line infection rates and reducing reimbursements to hospitals with high readmission rates. But more needs to be done.
The policy changes being contemplated for next year are rooted, advocates say, in the fundamental American ideals of choice and competition. But a consumer-based, patient-centered health system requires individuals, families and third-party payers, whether government or private insurers, to know what they're buying.
Similarly, doctors and health care professionals can't effectively set goals and make improvements if they don't know where they stand. Our experience indicates that doctors and hospital administrators want to provide the best possible care. They mean well. They are mission driven. But they need clear and objective data on their hospital's performance so they know where to focus their efforts.
Interestingly, to conduct our research we were forced to use 2011 hospital data because the Agency for Healthcare Research and Quality, in 2012 -- ostensibly to "enhance confidentiality
" -- eliminated "state and hospital identifiers" from its National Inpatient Sample (NIS) database. That's like knowing there's been a big pileup on the Beltway, but not knowing where.
Outcomes data from every hospital in the United States, analyzed at both the disease level and procedure level, should be compiled and made available for legitimate research and quality improvement purposes.
We're not looking to generate tabloid headlines. Hospitals should be given ample time to react to the data and improve. After that, the same type of information should be made publicly available -- so that patients, in consultation with their doctors, can select hospitals where they're most likely to get the best care, rather than those that simply accept their insurance.
Our research should send a strong message to policy makers that good health insurance, no matter how affordable or accessible, is not synonymous with good care.
You don't go to an Italian restaurant and expect great sushi. Similarly, just because a hospital is good at knee replacements doesn't mean it's good at brain surgery. Collecting and analyzing data on outcomes, providing transparency, and driving performance improvements should be core elements of health-care reform. The American people deserve nothing less.