Were ICPs heard? Report: Don't compare hospitals

Report cites key caveats, grim infection toll

Though recently released with much fanfare in Pennsylvania, the first state infection rate report with hospital-specific data contains a key caveat: it should not be used to compare one hospital against another. The Pennsylvania Health Care Cost Containment Council (PHC4) report detailing infection rates at Pennsylvania's 168 individual hospitals states:

"This report should be used to measure individual hospital performance over time, rather than to compare hospitals to each other. It should be used as a tool to ask hospital representatives informed questions, especially about their infection control and prevention program. It is not intended to be the sole source of information in making decisions about hospital care, nor should it be used to generalize about the overall quality of care provided by hospitals."

The overall impact of hospital infections reflected in the report is as bad or worse then previous state and national assessments, with infections resulting in roughly four- to sixfold increases in such critical areas as patient mortality, length of stay and costs. Given the severity of the situation, ICPs are increasingly on board with the push to disclose infection rates as a quality improvement incentive, but they want a system that is valid and allows fair comparisons. In that regard, the caveats cited in the PHC4 report can be seen as acknowledgement of ICP concerns, particularly the disclaimer about not using the document to compare hospitals.

Though comparing hospitals — as opposed to tracking the rates of one facility over time — is the presumed ultimate goal of the infection rate reporting laws that have swept the nation, the PHC4's acknowledgement of its lack of hospital comparative power could be seen as a concession to epidemiological reality. Infection control professionals have repeatedly warned that such rate reporting efforts could be undermined by a host of variables, including patient mix, risk adjustment methods and the intensity of surveillance. The PHC4 — an independent state agency charged with collecting, analyzing and reporting cost and quality health care information — acknowledged as much.

"Hospitals differ in terms of the volume and types of care provided, and the completeness of infection reporting across hospitals may vary," the report states. "For example, a low number of infections reported by a hospital in this report could mean that they are doing an excellent job in reducing their infection rate and ensuring patient safety. On the other hand, it could indicate that they are underreporting their infection numbers to PHC4. Conversely, a hospital with a high number of infections might appear to be less effective at patient safety. Yet, in reality, they may be doing a very good job of identifying and reporting infections — a positive contribution to patient safety. Hospitals using electronic surveillance approaches may report higher numbers for this very reason, and these hospitals are noted in the report."

'An evolution in the process'

"It is an evolution in the process, where they have reached the point of putting out hospital-specific data," says Patrick Brennan, MD, director of infection control at the University of Pennsylvania in Philadelphia. "It is not useful for interhospital comparisons, but it can generate useful discussions between patients and their providers about why the rates are what they are. Whatever the number is, why is it what it is and what are you doing to make it better?"

That said, the reports tendency to group hospitals into "peer groups" may invite the very comparisons it specifically advises against. "It says it should not be used for comparisons between hospitals," Brennan notes. "But the establishment of peer groups sort of begs that comparison. When you look at the peer groups there are vast differences among the hospitals in peer group 1 — which is where our hospital is located — in terms of numbers of admissions and complexity of services. We get transfers from every other one of the hospitals in eastern Pennsylvania. We have the highest case mix index of any of the hospitals in eastern Pennsylvania, which is a crude measure of complexity and utilization. [This report] furthers the discourse, but it doesn't provide any more insight about the actual performance of the hospital than in previous versions."

Brennan and other epidemiologists and ICPs have previously questioned the validity of the PHC4's previous reports because of the dependence on administrative and billing data to count infections. The new PHC4 report details its data collection process and emphasizes that "billing" or administrative data were not used.

"What they have told us is that these data are the data reported by the infection control professionals — that this is not the administrative data," says Brennan, who also is chairman of the Center for Disease Control and Prevention's Healthcare Infection Control Practices Advisory Committee. "I haven't seen the entire report, so I haven't been able to do the math myself. In previous reports, they had 10 times as many administrative reports as ICP-generated reports."

Regardless of the lingering questions and caveats, consumer advocates seized on the report's groundbreaking aspects. "This report is a first — we are no longer looking at statistics based on estimates or extrapolated data," says Lisa McGiffert, director of the Consumers Union's "Stop Hospital Infections" campaign. "These are real people who suffered from real infections. The personal and financial costs of hospital infections are staggering."

While previous reports have focused on the aggregate quality of care and financial consequences of hospital-acquired infections, the new report, Hospital-Acquired Infections in Pennsylvania, establishes a baseline against which an individual hospital's future performance can be measured, the PHC4 report emphasizes. It includes information on approximately 1.6 million patients treated in the state's 168 general acute care hospitals during 2005. In addition to the number of cases and infection rate per 1,000 cases, information on mortality, average length of stay, and average charges for cases with and without hospital-acquired infections are presented for each hospital.

While the report includes actual numbers, rather than risk-adjusted numbers, PHC4 took several steps to help readers interpret the data. Patients that were hospitalized for an organ transplant, complications of an organ transplant and/or burn treatment were not included in the report because they may be at a greater risk of contracting a hospital-acquired infection. In addition, PHC4 grouped hospitals according to the complexity of services offered, the number of patients treated, and the percent of surgical procedures performed.

In Pennsylvania, the hospital-acquired infection reporting requirements were phased in over a two-year period. Beginning Jan. 1, 2004, hospitals were required to start submitting data on the following types of hospital-acquired infections to PHC4: surgical site infections for circulatory, neurological, and orthopedic procedures; indwelling catheter-associated urinary tract infections, ventilator-associated pneumonia and central line- associated bloodstream infections.

For the third and fourth quarters of 2005, the surgical site infection category was expanded to include all surgical procedures, and for the fourth quarter of 2005, the pneumonia, bloodstream, and urinary tract infection categories were expanded to include hospital-acquired infections that were not device-related. While hospitals were required to submit data on all hospital-acquired infections to PHC4 beginning in January 2006, the hospital-specific report includes only the information on cases for which hospitals were required to report during calendar year 2005. Specific findings of the report include:

  • Hospitals reported 19,154 cases in which patients contracted a hospital-acquired infection, a rate of 12.2 per 1,000 cases. The hospitalizations in which these infections occurred amounted to 394,129 hospital days and $3.5 billion in hospital charges.
  • The mortality rate for patients with a hospital-acquired infection was 12.9%; the mortality rate for patients without a hospital-acquired infection was 2.3%.
  • The average length of stay for patients with a hospital-acquired infection was 20.6 days; the average length of stay for patients without a hospital-acquired infection was 4.5 days.
  • The average hospital charge for patients with a hospital-acquired infection was $185,260; the average charge for patients without a hospital-acquired infection was $31,389.
  • When looking at private sector insurance reimbursements (which do not include Medicare and Medicaid), the average payment for a case with a hospital-acquired infection was $53,915, while the average payment for a case without a hospital-acquired infection was $8,311.

[Editor's note: Copies of the report are free at the PHC4 web site at www.phc4.org or by calling PHC4 at (717) 232-6787].