Home Q&A When the figures don’t add up

When the figures don’t add up

November 2012

Published in the November 2012 issue of Today’s Hospitalist

WHEN THE CENTERS FOR MEDICARE AND MEDICAID SERVICES (CMS) announced that it would not pay to treat the complications of hospital-acquired conditions starting in 2008, the agency had high hopes. Because hospitals would no longer be paid extra for treating preventable conditions, the thinking went, they would be motivated to deliver better care.

But researcher Jennifer A. Meddings, MD, MSc, suspected there would be problems docking hospitals for hospital-acquired catheter-associated urinary tract infections (CAUTIs)

CAUTIs was one of the first conditions that the CMS targeted for nonpayment. But in research that Dr. Meddings conducted before the new policy took effect, she learned that the vast majority of catheter-associated UTIs weren’t being identified because of incorrect documentation and coding. In a chart review of 2006-07 discharges, she found that nurses commonly documented urinary catheter use in bedside flow sheets. But hospital coders would review documentation only from providers “physicians, physician assistants and nurse practitioners “according to CMS guidelines for applying diagnosis codes for describing UTIs.

In the wake of Medicare’s new nonpayment policy, would hospitals start doing a better job describing UTIs in billing data as catheter-associated? Apparently not, according to new research that Dr. Meddings published in the Sept. 4 Annals of Internal Medicine.

That study of 96 Michigan hospitals compared 2007 and 2009 data on hospital stays to gauge the impact of the 2008 policy. Dr. Meddings wasn’t surprised to learn that inaccurate coding was still the rule.

But the extent of that failure was startling. The study found that even after the CMS policy took effect, only 2.6% of hospital-acquired UTIs were described as catheter-associated “a rate that an accompanying editorial called “disturbingly” low. Epidemiological data suggest that between 59% and 86% of hospital-acquired UTIs are catheter-associated.

In the Annals study, only 321 hospital-acquired cases of catheter-associated UTI were listed in claims data. Of those, 296 patients had other diagnoses or comorbidities like heart failure that qualified the hospital for extra payment, even when the hospital wasn’t reimbursed for treating the UTI. So only 25 CAUTI cases that year in Michigan “0.003% of hospitalizations “were subject to the CMS’ financial penalty, hardly what you would expect for such a large pool of hospital-acquired infections.

“The policy is well-intended, but it’s using a data set that doesn’t capture the event it’s trying to measure,” notes Dr. Meddings, assistant professor of internal medicine at the University of Michigan Medical Center in Ann Arbor. “It’s not likely to have its intended effect.”

Dr. Meddings talked to Today’s Hospitalist about her results.

You knew CAUTIs were underreported from your prior work. What did you expect to find in this study?
After the new CMS policy took effect, it was possible that hospitals would know the importance of coding correctly because of the difference in payment. That could be thousands of additional dollars per hospitalization.

We were concerned that results showed a problem with Michigan data. So we looked at other states, including California, with data from AHRQ’s free online HCUPnet system for claims. We found the catheter code was used rarely across the nation, so the problem is not unique to our state.

What’s to blame for the coding problems?
For a UTI to be recognized as a complication for which a hospital doesn’t get extra pay, the coder has to do four things correctly: List the catheter-associated code 996.64, list a UTI code, and list both the catheter code and the UTI code as hospital-acquired. When all aren’t listed correctly, hospitals get paid by default.

To what extent are notes part of the problem?
It comes down to how well the diagnosis is described. Most catheter use is well described in nursing notes, but coders don’t usually see those. Even if coders did recognize catheter use in the chart from nursing or physician notes, they cannot apply the catheter-association code 996.64 unless a provider indicates that the UTI was catheter-associated.

Most doctors have no training in how their documentation is translated into diagnoses for billing and public reporting. Yet, how physicians describe UTIs in their documentation is being interpreted very literally. If you just write “UTI,” it’s often taken by default as present on admission instead of hospital-acquired. Sometimes, physicians don’t even list a UTI among the diagnoses for treatment if it is a lower priority than a patient’s other issues.

Can physicians be trained to do a better job?
Training won’t improve documentation significantly unless the system is changed so physicians can more easily identify and describe complications in the time they have available for documentation. In their daily notes, physicians concentrate on the highest priority items “and CAUTI is not likely to become a higher priority problem for most patients.

Would incentives work?
There is currently no incentive for physicians to describe CAUTIs more accurately in their notes. It doesn’t affect payment or performance reporting for most individual physicians. Some hospitalist groups have quality improvement and patient safety assessments where doctors get feedback on the number of complications that occur on their service or in their patients, and those can be tied to promotion and bonuses. So a doc who documents these things very well may just end up looking like a doc who has a lot of complications compared to his or her peers.

Would computerization help?
Computerized ordering systems can issue smart orders to remind people that a catheter is present and to take it out if it’s no longer needed, or an electronic stop order allowing nurses to remove catheters by a protocol. Such reminders and orders can reduce CAUTI rates by 50%.

But this won’t change physician documentation directly. If anything, some hospitals use electronic orders and note prompts to require physicians to document ” as in check off “conditions that are POA to secure payment for these conditions as comorbidities, rather than as complications.

What would be a better way to collect billing data?
The current process is not adequate for collecting accurate, complete data on the development of hospital-acquired complications, particularly catheter-associated UTI. More accurate data collection for hospital-acquired infections involves surveillance methods where trained personnel routinely evaluate a combination of medical record data, clinician queries, exams and laboratory studies to determine if a patient has certain complications.

What are the implications for public reporting?
These data are simply not valid for assessing hospital rates of CAUTIs, and we suspect similar issues may be occurring for other complications.

Using invalid data to compare hospital complication rates can, unfortunately, undermine the intended principles of the policy. Hospitals with higher rates could be providing lower quality of care, treating patients at higher risk for complications or simply doing a better job documenting complications.

I hope our results will inform revision or expansion of the policy. However, Medicare’s Hospital Compare Web site is still reporting hospital rates of catheter-associated UTIs generated from billing data.

Paula S. Katz is a freelance health care writer based in Vernon Hills, Ill.