Documentation/coding can create problems

Florida PEPP focuses on dehydration, pneumonia

The Florida Payment Error Prevention Program (PEPP) determined that two of the major coding issues experienced by hospitals involved coding problems with pneumonia patients and unnecessary cases of dehydration.

These two diagnoses could result in an excessive payment by Medicare of more than $2,000 per case. On the other hand, a hospital could lose thousands of dollars through inaccurate coding and incomplete documentation, according to Mark Michelman, MD, MBA, physician coding advisor for Morton Plant Mease Healthcare of Clearwater, FL. Michelman also is the medical director of quality utilization management for Morton Plant Mease and clinical coordinator with the Florida peer review organization Florida Medical Quality Assurance Inc. of Tampa, FL.

The Florida PEPP focused on those two diagnoses because they were areas in which some of the more common coding and documentation mistakes were made, Michelman says.

In the case of pneumonia, hospitals were sometimes billing Medicare for complex pneumonia (DRG 79) cases, such as gram-negative pneumonia or aspiration pneumonia, when those cases proved through chart reviews to have a more accurate diagnosis of simple pneumonia, such as community-acquired pneumonia (DRG 89). The community-acquired pneumonia cases were typically those involving patients who were generally healthy before coming into the hospital with fever, cough, and congestion. The X-ray would indicate a pneumonia diagnosis, but the patient could be treated with a short hospital stay, he explains.

In the more complex pneumonia cases, the patients are frequently immunocompromised or are frail elderly people living in nursing homes, and the treatment usually is more intense, using different antibiotics.

"We had a predictive formula that predicted the billings of DRG 79 for complicated pneumonia cases based on prior billings," Michelman says. "For example, if a hospital had billed 20% of pneumonias with that DRG last year, we could predict based on hospital demographics what the billing should be for the coming year, and we found that some hospitals had billed more than was predicted by the model."

PEPP reviewers then substantiated the data by reviewing medical records from each of the hospitals that had more DRG 79 diagnoses than predicted. Some of the DRG 79 cases would have been more accurately coded as DRG 89 cases, which is the coding for simple pneumonia, a less costly diagnosis.

Based on the chart review information, reviewers identified 20 hospitals in Florida that had the greatest discrepancy between predicted DRG 79 cases and actual billed DRG 79 cases, Michelman adds.

"We asked these hospitals to submit to our organization a quality improvement plan and, invariably, in many of these cases a major issue was inadequate or illegible physician documentation to support the higher-paying DRG," Michelman explains. "In some cases, hospitals would upcode to generate more income for their hospital."

Since Florida PEPP drew attention to the pneumonia coding problem, the 20 hospitals have decreased their billings for DRG 79, Michelman says.

Likewise, Florida PEPP found that hospitals were incorrectly coding dehydration cases, many of which were situations in which a patient either did not meet the criteria for a diagnosis of dehydration or cases in which a patient with dehydration was inappropriately admitted into the hospital or stayed in the hospital for longer than necessary, Michelman says.

"Many times the admission might have been appropriate, but the physician documentation wasn’t there to support the admission," Michelman adds.

Here is how the hospitals improved their physician documentation and coding accuracy during the Florida PEPP projects:

Hospitals identified root causes.

In the case of the pneumonia project, hospitals mainly identified these root causes: lack of information, lack of knowledge, and the need for more coding education and improved documentation by physicians.

With the dehydration project, hospitals decided the root cause was primarily an issue of physician documentation.

Patients with dehydration typically have excessive fluid losses due to vomiting, diarrhea, and sweating, and their clinical signs include rapid heart rate, low blood pressure, dry mucus membranes, poor skin turgor, and abnormal lab values, Michelman says.

Treatment typically is intravenous fluid replacement. But the problem was that physicians often were diagnosing patients with dehydration (DRG 296/297) when there was no documentation that the patient had any of the clinical signs or manifestations of dehydration, he says.

For instance, a physician might say a patient is 85 years old, weak, and came into the hospital after passing out. While that patient could have dehydration, the description cited does not by itself justify that diagnosis, Michelman explains.

"Many times they have no history, no physical findings, and no lab abnormalities consistent with dehydration," Michelman says.

Also, there often was a problem with the treatment given to these patients. The IV fluid sometimes was administered at an inadequate rate to treat dehydration if the patient actually had the problem. In other cases, there were patients treated for dehydration for three or more days in the hospital when their condition could have as easily been treated on an outpatient basis or observation status.

"We found that two-thirds of the patients did not meet the criteria for the diagnosis or to be in the hospital, and that’s a very high number," Michelman says.

"A coder can code a diagnosis if the doctor says the patient has that diagnosis," Michelman adds. "But the doctor has to document why he’s saying that, based on history, physical findings, and therapy, and the coder needs to make sure it’s documented."

Hospitals determined the process for improvement.

There were five basic areas identified for improvement in making the correct pneumonia diagnosis. These were as follows:

— improving physician education;

— improving education and re-education for coders;

— starting an internal and external audit process for making certain documentation in coding is appropriate;

— assigning a physician coding advisor who would interface with coders and physicians to answer coders’ questions about whether clinical documentation is adequate to support a diagnosis;

— using physician query forms as a communication tool, to be sent to doctors via fax, e-mail, or hard copy, in which the coder would ask the physician to clarify a diagnosis. The physician would have to respond to the query in the body of the medical record, whether it was in the history and physical, progress notes, discharge summary, or an addendum.

"Some hospitals also wanted doctors to be reviewed by their peers in an educational format," he adds.

Letting physicians counsel physicians

In this process improvement, physicians would receive general education about properly diagnosing dehydration or pneumonia. Then, in situations in which a physician is found to have made an improper diagnosis and the improper diagnosis was coded and billed before it was discovered that the patient didn’t meet criteria for the diagnosis, other physicians would counsel the doctor who made the diagnosis to prevent future errors and improve physician documentation, Michelman says.

Demonstrate improvement.

The Florida PEPP re-sampled the 20 hospitals’ charts six months after they had submitted their process improvement suggestions for more accurate pneumonia diagnoses, and reviewers found a decrease in the number of DRG 79 codings that were billed compared to the baseline chart review. In the hospitals that were not targeted for process improvement, there was no change in the DRG 79 codings, Michelman says.

PEPP reviewers found that with the dehydration diagnosis, there were two peaks for length of stay (LOS). The first was at three days and the second was at seven days, he says.

Either of these LOS figures would often be inappropriate, because a patient treated for dehydration typically should be treated and discharged before three days have passed, Michelman says.

"Three days is the length of stay that a patient needs to be in the hospital before being eligible to go into a nursing home, so we felt that the patient was kept in the hospital to justify going into a nursing home," Michelman says.

Those patients who were treated for dehydration for seven days were frequently patients who should have received another diagnosis that would have been more appropriate, Michelman says. They might have had malignancies, renal failure, or some other cause of dehydration; if they had been properly diagnosed and coded, then the hospital would have received a higher reimbursement than it would for dehydration, Michelman adds.

[Editor’s note: For more information about Florida PEPP and its coding and documentation projects, contact Mark Michelman, MD, MBA, at Morton Plant Mease Healthcare, 232 Jeffords St., Clearwater, FL 33756-0210. Telephone: (727) 461-8016. E-mail:]