SHEA: infection rate disclosure laws for all states are inevitable
This is going to happen; we just need to own it’
The era of noscomial infection rates protected as privileged and confidential information rapidly is coming to a close, warned several attendees recently in Philadelphia at the annual meeting of the Society for Healthcare Epidemiology of America (SHEA).
Driven by scathing media reports and consumer groups carrying the banner of patient safety, the demand for infection data has increased with such speed that epidemiologists already are looking at state laws as a foregone conclusion.
"The disclosure train is leaving the station," said Patrick Brennan, MD, an epidemiologist at the University of Pennsylvania in Philadelphia. "Providers must get on board."
Laws requiring nosocomial infection rate disclosure have been passed in Illinois and Pennsylvania. Similar laws are under discussion in a number of states, including Virginia, Florida, Missouri, and California. It is becoming increasingly critical for the infection control community to develop a national response on the issue, which gathered a lot of momentum before it began appearing on the radar.
"We need to get a hold of this," said David Henderson, MD, medical epidemiologist at the National Institutes of Health Clinical Center in Bethesda, MD. "It is essential that we get invested and find processes that make sense to get this done. I am not against doing it; I just want to do it right."
The Centers for Disease Control and Prevention (CDC) is working on a guidance document on rate infection disclosure issues, which is expected to be finalized by the end of the year. The Association for Professionals in Infection Control and Epidemiology (APIC) issued a position paper on the subject in 1998, which stated, "Crude or overall infection rates, not adjusted for risk, are inappropriate for comparison and potentially misleading." As laws are framed in various states, local ICPs and epidemiologists are scrambling to get involved to ensure the resulting reporting systems produce meaningful data.
"This is going to happen; we just need to own it," Michael Edmond, MD, epidemiologist at the Medical College of Virginia in Richmond told SHEA attendees. "We should probably be driving it and setting the methodology. Otherwise, we will be forced to use methodology that might not be very good."
Edmond became involved when a bill was introduced in Virginia that would require hospitals to report raw infection data without standardized definitions and other statistical controls.
"We’re very fortunate that our office is right across from the state capitol," he said. "We were able to have a nurse or physician sit through subcommittee hearings; and when the bill was introduced, we were able to speak to it. We argued against the bill on the basis of the fact that the state had not asked for any standardized methodology whatsoever. They are just asking all hospitals to report all data on all nosocomial infections."
At the suggestion of Edmond and colleagues, legislators agreed to table the bill until a survey could be conducted to determine what type of definitions and other aspects of infection surveillance currently are being used in Virginia hospitals. "[We are trying] to identify areas that might be of common interest to hospitals or might be relatively easy to start our surveillance efforts," he said. "I think all of us in Virginia realize, as of next year when the General Assembly meets, there will be a [disclosure] bill."
In getting involved in the process, Edmond met with members of the Consumers Union, publishers of Consumer Reports, which has set up a web site — www.stophospitalinfections.org.
Citing CDC estimates of 90,000 deaths and costs of $5 billion each year, the group emphasizes that disclosing infection rates will give hospitals the "strongest possible incentive" to improve.
Edmond said he used the opportunity to educate the group, but the example conversation he cited highlighted the chasm between consumer and clinical knowledge of infection control issues. "They asked me questions like, What is the difference between surveillance and treatment?’" Edmond told SHEA attendees.
In general, epidemiologists fear such laws may give short shrift to critical denominator data and risk-adjustment factors such as patient severity of illness.
"Better reporters, those who do better surveillance, may appear worse in the outcome of these reports," Brennan said.
"The lessons from Illinois and Pennsylvania are that this is a highly politically charged issue. Health care providers cannot assume that they own the high ground on this issue. There is significant public interest in having access to these data," he explained. "Attempts to communicate the epidemiology have been viewed in both instances as attempts at stonewalling the issue and hiding the information. Communicating the correct epidemiology to a lay audience in this environment has been very difficult."
The big chill
Some fear such laws — however well intentioned — will create disincentives to aggressively pursue and report all infections. Hospital Infection Control asked Brennan whether such a chilling effect might be the result of disclosure laws.
"I am not aware of any provision in the Pennsylvania legislation that provides for sanctions against those that would withhold the information," he said.
"But the process has been so rapid in Pennsylvania that at least the initial data that go in are very likely to be significantly flawed. . . . Every member of the [state advisory] panel anticipates the need for additional staffing to identity all of the device-related denominators. I’m not sure that every hospital in the state is going to achieve that. I have significant misgivings at this point as to whether we can reduce hospital infection rates by this method," Brennan pointed out.
If infection control is to be conducted more transparently in a new age of patient safety, resources to meet such mandates must be provided. For example, consumers are not expected to have much understanding of resource-conscious approaches such as targeted surveillance, instead expecting that hospitals track and report every infection that occurs.
"I think that public accountability really holds our feet to the fire," Brennan noted. "I’m not sure that the information we put out there is going to be the correct information; but it certainly is going to make us pay greater attention to it, and I think that is a good thing. [The state], to its credit, is trying to force hospitals to apply greater resources to infection control."
The issue may end up in the nation’s courts, particularly if hospitals that are doing an honest and thorough job of tracking and reporting infections are compared unfavorably to facilities that look better due to slack surveillance efforts.
"It seems to me if somebody really starts getting killed because they are doing very good surveillance — and they lose a lot of patients and are threatened financially — the only recourse would be court," said C. Glen Mayhall, MD, epidemiologist at the University of Texas Medical Branch in Galveston.
"It might be a good place to have these things play out with respect to lawyers talking and bringing expert witnesses in and making points as to why we [look bad] when we are doing very good surveillance and the other folks are not," he added.
The battle of Pennsylvania
Indeed, Brennan said he originally got involved in the situation in Pennsylvania because of a previous incident involving reporting of mortality data following vascular surgery. A state report suggested his hospital was a "mortality outlier," indicating increased risk for patients undergoing surgery there, he explained.
"Our vascular surgeons, within a month of the report, lost volume," Brennan said. "They got calls from the referral base wanting to know what was happening with their patients, and the referrals dropped off. When we got underneath it, we found out that when looking at procedure codes they actually had lower than expected mortality for procedures, but many of these deaths were related to [cancer] patients who had interventional radiology. They died of their cancer, not at the hands of an interventional radiologist or a surgeon."
The state was receptive to the explanatory information and changed the report accordingly, he said. Thus, Brennan was hoping for cooler heads to again prevail when the rate disclosure debate began heating up in Pennsylvania.
The situation there began late last year when the Pennsylvania Health Care Cost Containment Council (PHC) adopted new reporting requirements for nosocomial infections. The Hospital & Healthsystem Association of Pennsylvania (HAP) stepped into the fray, arguing that requirements be limited to nosocomial infections that represent the highest risk to patients and result in the highest costs.
HAP urged the council to work with clinicians to define the data collection methodology and develop clinically credible public reports that will be useful to the public, purchasers, and clinicians.
Brennan said he found himself between the vying groups, not necessarily opposed to reporting the data but trying to determine the best way to do so.
"I realized they were two warring parties, and I had to tell them I wasn’t going to be a double agent for either one," he said at SHEA. "I was going to talk to everybody, so they should assume that everything they said to me I would say to the other party and vice versa. They seemed to accept that."
Understanding the perspective of the other parties involved — including payers, clinicians, consumers, and worker unions — is essential to finding common ground, Brennan added.
"Physicians and ICP participation in the process shaped the final results in both [Pennsylvania and Illinois]," he said. "There has been both a health care perspective and a purchaser perspective, and there has been tremendous discord between those two."
The law passed in Pennsylvania requires acute care hospitals submit hospitalwide data that eventually will include some surgical procedures, device-related infection rates, ventilator-associated pneumonias, central venous catheter-related bloodstream infections, and device-related urinary tract infections.
"It is anticipated that by Jan. 1, 2006, all 13 categories of infection and body sites will be required," Brennan said. "They essentially expect us to go back to housewide [surveillance]. That really is their expectation."
Whistle-blowers protected in Illinois
The Illinois Hospital Report Card Act states that consumers have a right to access information about the quality of care in hospitals in order to make better decisions about health care providers.
The requirements in that state address staffing issues, calling for reports of nursing hours per patient day and on nursing vacancy and turnover rates. The hospital infections that must be reported include class I surgical-site infections, ventilator-associated pneumonias, and central line-related bloodstream infections.
There also is protection provided for whistle-blowers who report violators, Brennan said. A problem in both Pennsylvania and Illinois is the state health departments never had significant involvement in the legislation, he said.
"Both of these [laws] were initiated by the legislature and independent state agencies, but at the impetus of significant media coverage about this issue," Brennan said.
"But there was no state health department involved in either state, resulting in a significant disconnect between the regulatory and epidemiology departments in state government. . . . There has been a lack of understanding about epidemiological methods, and the communication of the infection control perspective was difficult in both states," he explained.
Indeed, depending upon what quality indicator is measured, hospitals can have widely divergent results even within their own communities in various reporting and report card systems, Brennan noted.
"For example, depending on where your browser takes you, Johns Hopkins either has the third best neurosurergy program in the United States or the 14th best neurosurgery program in Baltimore," he said. "Two different report cards, two different methods: One that is very heavily complication-oriented in its assessment, which downgrades their performance, and the other is not so heavily weighted toward complications. So the devil is certainly in the details, and this could conceivably end up in the courts."