Saturday, May 21, 2005

Motality & hospital report cards

Two years ago I was concerned about pneumonia cases. A large insurance company released an updated version of its report card on hospitals, and for the second year in a row CGH statistics were good in all areas – except pneumonia. According to the insurer, the death rate from pneumonia at CGH was more than it should be when compared with the Central New York region.

What was the cause? Were CGH patient records properly coded?[1] An outside review of our records showed that we were accurately coding them.

Was CGH care being provided properly? Yes, it was. We reviewed every single case.

Were pneumonia patients at CGH somehow different from those at other hospitals? We looked at the numbers of patients transferred to CGH from nursing homes. We looked at the average age of pneumonia patients. CGH pneumonia patients were somewhat older than those in other hospitals. On average the CGH patients were 75 years old. The average age of pneumonia patients at other hospitals ranged from 50 to 69 years. Still, age alone did not explain it. We were puzzled.

The quality committee that was reviewing our pneumonia care focused its attention on advance directives.[2] One form of advance directive is a DNR order (“do no resuscitate” order). Patients with DNR orders request not to have cardiopulmonary resuscitation (CPR) under certain circumstances.

When we examined the records of patients with pneumonia who died at CGH, we found that that 90% of them had DNR orders. In other words, most deaths were by patients who did not want aggressive efforts to resuscitate them if their conditions seriously deteriorated. Of the hundreds of non-DNR pneumonia patients cared for at CGH, the death rate was less than one percent.

This distinction was apparently not reflected in the analysis used by the insurance company for its hospital report card. This point was clearly demonstrated last year when Dr. Dan Carlson, Vice President – Medical Affairs, and I attended a meeting on hospital report cards with about a dozen upstate hospitals. The large insurance company showed pneumonia mortality rates that it prepared for hospitals attending the session – and, sure enough, CGH appeared to have an above average mortality rate. Next to present was a national company that also produced quality comparison data. In this case, however, CGH was shown to have a lower mortality rate than the other hospitals when adjusted for more patient risk factors. Both presentations used the same data to generate different hospital rankings, based on their different statistical methods.

I was reminded of our pneumonia investigation when I read “Informed Choices For Better Care,” a recent story in the Post-Standard. The story said that for many years patients and families were “kept in the dark” about hospital comparison information, but report cards are now letting “consumers shop for health care much the same way they do for cars.”[3] The article also quoted the publisher of one hospital report card as saying that “hospital special interests” may no longer claim that “those poor dumb yokels (the general public) couldn’t possibly decipher how to use hospital data.”[4]

That report card publisher apparently sees hospitals as an adversary, but CGH has supported the development of report cards for many years (and we never disparaged the public as “dumb yokels”). In the 1990’s CGH was the first and only upstate hospital to publish its own comparative quality report card.[5] In doing so, CGH consulted the literature on quality reporting, used available data from national studies, and conferred with state government, local health maintenance organizations, and national organizations on the design and data used in the CGH report. This was before insurance companies and the federal government developed their hospital comparison websites.

Hospitals are not perfect, and neither are report cards, as the above pneumonia example suggests. I would argue, however, that website report cards educate both the public and providers, and they encourage competitive hospitals to get better. These are important tools in helping the health care industry make continuous improvements in quality management.

The pneumonia example demonstrates how seriously CGH investigates itself. I believe all hospitals take variations in quality seriously, and as the tools for measurement get better, hospital results will improve too.



[1] The government and payers look at the disease codes assigned to each patient record – this forms the basis for analysis.
[2] Advance directives are the written instructions patients give to doctors and hospitals specifying the type of care the patients want to receive (or don’t want to deceive) in the event they cannot make future medical decisions for themselves.
[3] “Informed Choices for Better Care,” Post-Standard, May 9, 2005.

[4] Post-Standard, May 9, 2005
[5] CGH’s self-published Quality Report Card started in 1996. It was posted on the CGH website and was issued annually until 2000, when the internally-prepared documents ceased when CGH joined the Health Alliance of CNY. In 1999 CGH’s report card cited with approval the Report of the President’s Advisory Commission on Consumer Protection and Quality in the Health Care Industry (1998): “A widespread and ongoing consumer education strategy should be developed to deliver accurate and reliable information on quality when choosing health plans, providers, and treatments.”