Pathology/Lab Coding Alert

The federal government continues to scrutinize pneumonia coding for cases of overpayment or fraud. Included in the Office of Inspector General's (OIG) fiscal-year 2001 work plan is the ongoing investigation of pneumonia diagnosis related group (DRG) upcoding for hospital inpatients. However, if the proper laboratory tests are carried out, and the treating physician documents the link between the test results and the diagnosis in the medical record, hospitals should not worry about passing OIG scrutiny, says William Haik, MD, director of DRG review, a hospital coding consulting firm in Ft. Walton Beach, Fla.

The pneumonia-upcoding project seeks to identify hospitals that incorrectly assign pneumonia diagnosis codes so cases are reimbursed under the higher rate of DRG 79 (respiratory infections and inflammations with complications) rather than DRG 89 (simple pneumonia and pleurisy with complications). The OIG and U.S. Department of Justice (DOJ) identify possible overpayment or fraud when a hospital reports a disproportionate number of pneumonia cases with ICD9 Codes assigned to DRG 79, which pays an average of $2,500 more than DRG 89. Upon OIG audit or DOJ investigation, hospitals have been cited for upcoding when the documentation in the medical record does not support the diagnosis used for the DRG assignment. Over 22 hospitals have settled their liability and paid more than $23.6 million, with many more under investigation.

Under particular scrutiny by the OIG are 482.83 (pneumonia due to other gram-negative bacteria), 482.89 (pneumonia due to other specified bacteria) and 507.0 (pneumonitis due to inhalation of food or vomitus), all of which assign to DRG 79. The first two codes are of particular interest to laboratories because those diagnoses are made largely because of cultures and other laboratory tests. Because these two codes do not specify the causative organism, coders have sometimes erroneously assigned pneumonia cases to these categories without adequate documentation. The assignment may be due to a lack of understanding of microbiology and ICD-9 coding principles or, in some cases, intentional upcoding. 

"The key to avoiding or surviving OIG scrutiny for pneumonia coding is to ensure adequate documentation to justify the assignment of the ICD-9 code," Haik says. This would mean having the results of appropriate lab tests, such as some combination of smears and cultures from sputum, bronchial washings or brushings or bronchial alveolar lavage (BAL). "But no test is confirmatory without the physician stating the connection between the lab results and the diagnosis," Haik says. "For example, if a sputum culture indicates the presence of Staphylococcus aureus, the medical record must have the physician's statement to the effect of, 'pneumonia secondary to Staphylococcus aureus' (482.41)."

The coder cannot use the lab results alone to assign the diagnosis; it must be assigned by the physician on the basis of both clinical and laboratory data. In the last example, the presence of  Staphylococcus aureus may have been due to contamination of the sputum specimen, and based on clinical findings, the physician may not diagnose 482.41. "In fact, Staphylococcus or other organisms can be a contaminant in sputum cultures," Haik says. "So for example, in assigning a diagnosis for a patient with a positive Candida culture, the physician takes into account that this would not be an uncommon contaminant, but would be rare as the cause of pneumonia."

Coders can avoid improper code assignments by never assigning a diagnosis on the basis of lab results alone. They must base the code assignment on the physician's final diagnosis, which must be documented in the medical record, along with all supporting test results. "If coders do not find adequate documentation, they should contact the attending physician for written confirmation of the diagnosis," Haik says.

Another common coding error involves a misunderstanding of ICD-9 coding principles. Specifically, coders sometimes assign codes for "other specified" when they should assign a code for "unspecified." ICD-9 has two conventions for "other specified": in Volume 2 (alphabetic index) these codes are listed as NEC (not elsewhere classified); in Volume 1 (tabular list), these codes may be listed as NOS (not otherwise specified). For pneumonia coding, both 482.83 (pneumonia due to gram-negative bacteria NOS) and 482.89 (pneumonia due to other specified bacteria) are NEC codes. This means that a specific causative organism (gram-negative 482.8, or gram-positive 482.89) has been identified, but the organism is not specifically listed by name in any of the other ICD-9 pneumonia codes.

This is in sharp contrast to pneumonia cases that should be reported as 482.9. If the causative organism is not identified beyond "bacterial," the proper code is 482.9 (bacterial pneumonia unspecified). Haik says that without a specific organism identified by lab tests and the physician's statement of a diagnosis linked to that organism, reporting one of the NEC codes, e.g., 482.83 or 482.89, is wrong.

When hospitals report an unusually large number of pneumonia cases using 482.83 and 482.89, it may raise a red flag for OIG scrutiny. Both of these codes represent cases involving a causative organism that is identified but not specifically listed in the diagnosis codes. "ICD-9 has a specific listing for nearly every strain of pneumonia-causing bacteria, so identifying a specific causative organism that is not listed would be uncommon," Haik says.

For example, pneumonia due to other gram-negative bacteria (482.83) includes cases of infection with an organism such as Proteus, which is a gram-negative rod enterobacterium. This bacterium is a normal inhabitant of the bowel and is not commonly involved in cases of pneumonia. "However, pneumonia that fits the description of diagnosis code 482.83 would be more common in immunosuppressed patients who are institutionalized," Haik explains.

"True instances of pneumonia due to other specified bacteria (482.89) would be even more rare, probably less than 1 percent of cases," Haik says. Because "other specified gram-negative bacteria" has its own code (482.83), only gram-positive organisms not listed elsewhere in the ICD-9 pneumonia codes could be reported with 482.89. "For example, a rare form of pneumonia that may occur in immunodeficient (e.g., AIDS) patients is caused by a Corynebacterium, called Rhodococcus equi, and would be reported as 482.89 because these are gram-positive organisms not specified elsewhere in the ICD-9 pneumonia codes."

According to Jodi Garrett, MT (ASCP) (SM), manager of the microbiology division of Nebraska Health Systems laboratory, many lab procedures might be instrumental in the physician's pneumonia diagnosis. "When a patient presents with signs and symptoms of pneumonia, the physician often orders a smear and culture from sputum, BAL or bronchial washings or brushings," Garrett says. The sputum or BAL smear may be stained directly to detect potential disease-causing organisms. This may include a Gram stain (87205) and/or fluorescent or acid fast bacteria (AFB) stain (87206). For specimens with positive smears for mycobacterium, physicians may also request probe techniques on the sputum or BAL specimen such as 87556 (mycobacteria tuberculosis, amplified probe technique). The physician may also request a Legionella fluorescent stain on the sputum or BAL, 87278 (infectious agent antigen detection by immunofluorescent technique; Legionella pneumophila). "These tests are used to give an immediate indicator of the presence and type of possible infective organism," Garrett says. Typically, the sputum or BAL is decontaminated and concentrated for AFB stain and AFB culture, and the concentration technique is reported as a separate service, 87015 (concentration [any type], for infectious agents).

"In addition to a direct smear and stain, cultures are usually ordered and carried out concurrently on the sputum or BAL specimen," Garrett says. Based on clinical findings and the smear results, the physician may order any of three cultures: 87070 (culture, bacterial; any other source except urine, blood or stool, with isolation and presumptive identification of isolates), 87102 (culture, fungi [mold or yeast] isolation, with presumptive identification of isolates; other source [except blood]) or 87116 (culture, tubercle or other acid-fast bacilli [e.g., TB, AFB, mycobacteria] any source, with isolation and presumptive identification of isolates). 

"If the isolate from any of these cultures requires definitive identification, further testing would be used," Garrett says. For example, aerobic bacterial isolates may be identified to the genus or species level using tests such as biochemical panels, 87077. Definitive identification of a fungus would be reported with 87106 (culture, fungi, definitive identification, each organism; yeast) or 87107 (... mold). Mycobacterial identification tests include 87118 (culture, mycobacterial, definitive identification, each isolate) and 87149 (culture, typing; identification by nucleic acid probe).

Although the physician must order lab tests and interpret the results to indicate the appropriate pneumonia diagnosis, laboratories and coders can ensure adequate documentation to support the diagnosis. Laboratory personnel should educate physicians regarding which tests to order for a suspected diagnosis, or which analyses are needed to identify an isolated organism. Coders should properly apply ICD-9 coding principles and ensure that appropriate test results and the physician's statement linking these results to a diagnosis are in the medical record to support the ICD-9 code.