Short Takes On News & Events

Study: CMS Penalties Don’t Change Hospital-Acquired Infection Rates

By Julie Appleby

October 10th, 2012, 5:19 PM

Updated at 8:00 p.m. to add comment from CMS.

A Medicare payment policy designed to push hospitals to cut their infection rates has had no effect in reducing two types of preventable infections among patients in intensive care units, researchers say in a study out Wednesday in the New England Journal of Medicine.

In 2008, the Centers for Medicare and Medicaid Services began denying additional payments to hospitals whose patients became sicker as a result of bloodstream infections and urinary tract infections associated with the use of central lines or catheters.

Researchers looked to see whether denying additional payments would spur hospitals to cut their infection rates, comparing those infections with a type of pneumonia not targeted by the payment policy.

“The financial penalty did not further reduce infection rates, which were already going down because of multitude of (infection control) campaigns and interventions that were already ongoing,” said the study’s lead author Grace Lee, associate professor, Harvard Pilgrim Health Care Institute and Harvard Medical School.

Infections picked up by hospitalized patients are an area of growing concern. It is estimated that about one in 20 hospitalized patients get an infection, resulting in up to $33 billion in additional costs each year. Efforts to reduce the rate of infections include public reporting requirements and the payment policy in Medicare, which is now being expanded into state Medicaid programs.

Other studies have found the payment policy resulted in increased attention by hospital leaders – sometimes at the expense of other infections not targeted by the policy.

As policy efforts expand, the researchers say “careful evaluation is needed to determine when these programs work … and when they have unintended consequences.” They did give did give some caveats about their findings: The study looked only at patients in the ICU, for example, so it can’t say if infection rates in other parts of hospitals changed. Researchers tracked data reported to the National Healthcare Safety Network by 398 hospitals from January 2006 to March 2011.

Many of those 398 hospitals were voluntarily reporting that data even before the payment policy was implemented, which means they may have been further ahead in their infection control efforts than others, thus resulting in no additional slowdown after the penalties, says Lisa McGiffert of Consumers Union, publisher of Consumer Reports.

She and others say efforts to affect hospital payments and to publish infection rates do work.

“When you start affecting payments and publishing results, it does get the attention of leadership in a hospital,” said McGiffert. “This study shows it’s enormously difficult to connect a singular policy with progress.”

Today’s study is consistent with data from other studies, including one recently completed by the American Hospital Association, said Nancy Foster, vice president for quality and patient safety policy at the association. Those studies have shown an ongoing decline in infections that pre-dated the payment policy.

Still, Foster said a key piece of any effort to reduce hospital infections is not just to count the number of patients who have problems, but to have a specific, detailed prevention strategy. Preliminary results from the association study found a 40 percent reduction in central line-associated bloodstream infections in intensive care units when they used a team-based approach.

“With infections, you need a series of things you’re doing right each and every time. Not just washing hands, although that is important. Not just the right antiseptic, not just the right antibiotic. But all of those things together,” said Foster. “That’s the shift in perspective hospitals have gained.”

CMS spokesman Brian Cook did not address the study results directly, but said the 2010 federal health care law strengthens policies to reduce infections in hospitals.  “We’re confident that these polices will improve health care quality and reduce costs,” Cook said.

2 Responses to “Study: CMS Penalties Don’t Change Hospital-Acquired Infection Rates”

  1. Deb Dwyer says:

    Expectations regarding future policies regarding reimbursements will impact behavior. So the fact the rates were declining prior to the legislation does not mean the declines were not, at least in part, due to the legislation. The model does not likely capture that. Survey data would be required.

  2. “Data validation: An elephant in the room”

    If the majority of this study’s central-line infection rate’s numerator (number of infections) and denominator (patient or central line days) – these infection data, which were provided by the almost 400 participating hospitals, were the basis for this study’s conclusions – were not independently audited and validated for accuracy and completeness, might not these data be unreliable, inaccurate, and incomplete?

    Might not, therefore, the authors having found “no evidence that the 2008 Centers for Medicare and Medicaid Services policy to reduce payments for central catheter–associated bloodstream infections and catheter-associated urinary tract infections had any measurable effect on infection rates in U.S. hospitals” be a conclusion that is in question?

    Could not the “actual” infection rate (which is different from and, due to many types of well-known biases, may be greater than the “reported” infection rate), or even the transmission rate of infectious agents (which, incidentally, is crucial although is not ordinarily studied and can be masked by antibiotics and not be detected), have ironically risen by the study’s end in 2011? (The study by Lee et al. does not exclude this possibility and, in fact, appears to subtlety support it.)

    Are not the conclusions of any retrospective or prospective cohort (non-randomized, not controlled) study, like Lee et al.’s, that analyzes trends in infection rates only as sound as their associated infection data are demonstrated to be reliable and valid?

    The important topic of data validation is not discussed in Lee et al.’s study. Nevertheless, it can be argued that failure to place emphasis on and to validate the infection data renders of limited value any conclusions about the impact of an intervention or federal policy that are associated with or based on these potentially faulty infection data. (See my peer-reviewed article on this topic at:

    (Disclosure: I have no financial association, direct or indirect, with the study of central line infections in ICUs.)