Coding Audits

Select Data spots 9 common coding errors

Example: A patient has a diagnosis of hyperlipidemia in M0240f and anemia in M0240g. Both have a severity rating of two, but anemia, which is a case-mix diagnosis, was put on a diagnosis line that brings no dollars. Given that both have the same severity and that there is adequate documentation to support it, the coding sequence could be changed to improve reimbursement. Generally, many agencies fail to enter diagnoses eligible for case-mix payment within the top six lines even when it would be appropriate to do so because conditions are of equal severity and your documentation supports it, Carmichael notes.

Example: The patient has hypertension but the Select Data team found the patient also has a diagnosis of chronic kidney disease, which must be coded in addition to the hypertension code. The coding guidelines assume arelationship between the two conditions that affect code assignment and sequencing. If the hypertension patient had multiple co-morbidities of equal severity and adequate support documentation, it would be correct to enter the chronic kidney disease code on line M0240g, right below hypertension on line M0240f. That not only maintains sequencing integrity but frees one of the top six lines for a code that may have reimbursement value, Carmichael points out. An agency can sequence this above M0240f, but that would mean using a valuable case-mix line for a non-case mix diagnosis (CKD), says Select Data Coding Manager Elizabeth Clarke.

• Failure to provide the appropriate primary diagnosis

Select Data sees an average 25% error rate in this category, regardless of the agency’s caseload, says Carmichael. While this mistake doesn’t always affect payment, it can mean the coder has missed the focus of care or failed to ask the clinician for additional information, resulting in wrong care priorities. But there can be financial effects, too. Example: A clinician writes “foot ulcer” on the nursing plan for a diabetic patient for whom she is providing wound care, but in her documentation she notes that the referring physician is treating the patient for a diabetic ulcer. The failure to correctly identify the type of wound being treated would have cost the agency under one hundred dollars. By itself, that’s not much. But if this error were repeated in multiple patients, “the loss in revenues would add up quickly,” Carmichael points out. (The revenue loss can be a much greater $380 per episode, if coding changes the “C” score from a C2 to a C3, notes coding consultant Rambusch.)

• Failure to distinguish post-operative complications from routine aftercare codes

In one episode Select Data reviewed, a non-healing amputation site had been coded to routine aftercare. In addition, the patient also had peripheral angiopathy due to diabetes, which the coder also hadn’t spotted. Result: A revenue loss for the agency of $1,789.

• Failure to seek information and drawing conclusions without proper documentation

Example: An agency’s 63-year-old-male patient has both Type II diabetes mellitus and neuropathy, according to the intake referral. The agency then assigned codes for “diabetes mellitus with neurological manifestations” and “polyneuropathy in diabetes mellitus,” assuming that the two conditions must be related. However, the agency shouldn’t have assumed the two conditions were related and instead should have queried the referring provider. The error didn’t affect reimbursement, but by assigning two incorrect codes for which there was no supporting documentation, the agency’s coder put the HHA at risk of a coding compliance deficiency. As noted in CMS’s coding guidelines, only two conditions, osteomyelitis and gangrene, can automatically be linked to diabetes. Similarly, if an “abrasion” is coded to “open wound” without documentation to support the diagnosis, surveyors could see that as upcoding.

• Failure to code to highest level of specificity

Select Data’s audit teams often see 496 COPD, a noncase- mix diagnosis, used when the diagnosis documented in the record is emphysema or chronic obstructive asthma – both of them worth case-mix points. Such unintentional downcoding may cost the agency revenue to which it’s entitled. Conversely, coding COPD as exacerbated when the documentation does not reflect that can appear as to garner higher reimbursement.

• Failure to understand which V codes will pair with case-mix codes in M0246 and which ones won’t

Errors range from using V codes that are not on Table 4 of the CMS pseudocode appendix table to using a V code from table 4 but entering a non-casemix diagnosis in M0246. Example: Using a malignancy code in M0246 for the diagnosis of V10.05 (personal history of malignant neoplasm of the large intestine) would be wrong as none of the V10 codes are on the pseudogrouper table and will not earn casemix points.

• Other V code mistakes

Select Data audit teams regularly see V codes used incorrectly in conjunction with M0246. They also see them used incorrectly in M0190 and M0210. Neither V nor E codes (used to identify the external causes of injury or poisoning) are allowed in M0190 or M0210. Rather, it is the acute condition necessitating treatment or surgery that should be entered in these fields, advises corporate compliance head Carmichael.

• Failure to identify co-morbidities

Example 1: A patient with Alzheimer’s had both dementia and hypertension documented. Neither comorbiditiy was coded by the agency, which lost $402 on that episode. Example 2: Parkinson’s disease coded with no dementia when that condition was due to the Parkinson’s. There also was no documented code for prostate cancer when, in fact, the patient was still on Lupron. The agency would have lost over a thousand dollars in revenue on that episode. – Burt Schorr [bschorr@decisionhealth.com]For more information on Select Data products and services, please email info@selectdata.com or call 1(800) SELECT1.