Hospital quality metrics are becoming more important in healthcare. Some insurers compensate hospitals at more favorable rates if they meet specific quality goals. The federal government requires public reports on various quality measures. And consumers are scanning a number of government and private Internet sites for hospital quality information.
But this information must be approached with caution. Much of it provides an incomplete and possibly inaccurate view of care provided by a particular hospital. The problem is not with the information but in what we expect it to say.
Much of the current information does not detail the actual interventions delivered in the hospital or if the patient got better because of the care in the hospital. As such, the publicly reported indicators do not always give patients the full picture of differences among hospitals.
The data measures for heart attacks, for instance, focus on whether certain medicines were provided in an emergency room, including aspirin, beta blockers and ACE inhibitors, and will soon include measures of how long it takes to get treatment. The measurements look at each medicine individually. So results will show whether a patient received that medicine, which is a good measurement of adherence to accepted medical practices. A good measurement of quality, however, would look at the percentage of patients who received every medicine they should have received, which is not done now.
In addition, current medical literature says the preferred method to treat a heart attack is primary angioplasty, where an obstructed blood vessel is opened by inflating a balloon or placing a stent. But not all hospitals provide primary angioplasty and no public quality measures indicate which ones do.
The data also can be analytically skewed, unintentionally painting a negative picture about hospital quality even when good practices are followed. A hospital not offering primary angioplasty might transfer patients to one that does, an example of a good medical practice. But that risks leaving behind those patients who are too ill to transfer, which could negatively skew the hospital's mortality rate.
To try to compare like populations, the Medicare program has become a source for many quality reports. The advantages of Medicare information include its size -- millions of participants -- and its availability in the public domain. Medicare data also are uniform, unlike data found in private carriers, which allow comparison. However, sourcing from Medicare skews measurements toward ailments in older patients, such as pneumonia and heart attack. With no similar databases to draw from for children, for example, little public quality information exists on pediatric programs in hospitals.
In some cases, quality reports cite measurements that scientific literature may not agree are a benefit to patients. For instance, one measurement looks at the percentage of patients with pneumonia who received an antibiotic within four hours of coming to an emergency room. The scientific literature on pneumonia shows a clear benefit when a pneumonia patient receives an antibiotic within six to eight hours of coming to an emergency room. However, there is no clear consensus that receiving an antibiotic within four hours is necessarily better for a patient.
Overall, what is measured now is too superficial for an accurate look at hospital quality. Hospitals create much more data than are usually used in quality assessments.
Board certification rates of doctors, equipment repair rates, turnover of staff and patient satisfaction often are measured by hospitals but are rarely included in publicly available quality reports. To include such data could help create a better picture of capability, expertise, preparedness and care.
There are many other measurements that should be considered in any quality evaluation. Medical professional organizations are setting standards of quality in their specialties, such as pediatrics and cardiology, and should be part of the national discussion. We need to measure whether medicines such as antibiotics are used inappropriately, such as for a common cold, because of the potential for resistance caused by overuse in the population. Health programs such as smoking cessation, now recommended for cardiac and pneumonia patients, should be recommended for all patients and compliance should be measured. And once all these measurements are set, time and money must be invested to bring best practices and compliance throughout the entire healthcare system.
Information technology also will allow hospitals to gather more information and begin the arduous process of correlating interventions to outcomes. With an in-depth study of the data and public reporting of the outcomes, hospitals will be able to demonstrate a much more vivid picture of what they do so patients can make a more informed determination of quality.
Done properly, the development of quality measures clearly linked to better outcomes can help improve hospital performance and benefit patients. But to identify more accurate measurements of quality, hospitals and those who measure quality must work together. It is more important to do the hard work and truly define quality than simply to rely on easily available statistics regardless of whether they mean anything to patient care.
Steven Corwin, a cardiologist, is executive vice president and COO of New York-Presbyterian Hospital.