October 6, 2010

Over and Under-Coding: Auditing for Length of Stay

By

jwissler100Auditing your inpatient Medicare charts for compliance - in other words, seeking out over-coding in so-called "optimized" or "maximized" cases, as well as under-coding in so-called "missed opportunities" cases - is important for providers. The plan should be getting all the reimbursement you are entitled to (without going over) and mitigating your RAC risk at the same time.

 

The ultimate goal in MS-DRG assignment is to select the most appropriate MS-DRG that best reflects the patient's clinical picture and resource utilization as documented in the record.

 

Over-Coding

 

First, let's discuss looking for instances of over-coding:  start up the data-mining engine. You are looking for cases of one-, two-, or three-day length of stay (LOS). You also want your selection criteria to include a display of the discharge disposition. You can eliminate discharge by transfer to another acute-care facility or patient expiration, as these would be common scenarios not worthy of inclusion in this type of review.

 

A pre-billing technique is to have the coder review the actual LOS versus the GMLOS (geometric mean length of stay) of the assigned MS-DRG. If the actual LOS is less than GMLOS, the first thing to review is the D/D (discharge disposition). If this is a transfer MS-DRG and the D/D includes post-acute care (home health, SNF, acute rehab facility, etc.), then the case probably is coded and grouped correctly and needs no further review, but if the discharge status does not include post acute-care, the case may be over-coded. Remember, you are wearing your RAC data-mining hat at this point. You are looking only for the hot spots in your data, the ones that can be seen from afar. This technique will find the lowest-hanging fruit.

 

The easiest area to review, now that you have narrowed your focus and chart selection, is the procedure coding. Surgical MS-DRGs reimburse very well, so validate that the procedure is coded correctly. If it is, then look at the CC or MCC (major or minor complications/comorbidities). If there clearly are none, the case probably is coded/grouped correctly. However, if there is one, verify that the CC/MCC condition actually was treated or addressed during the stay and that it definitively meets reporting criteria for secondary diagnoses (i.e. it required extended LOS, required nursing care, was investigated or treated, or consumed resources).

 

If conditions/diagnoses are noted by the physician but do not appear to meet reporting criteria, then a query is in order. This is an alert that documentation is an issue for your facility. At least one RAC is denying secondary codes if clinical criteria clearly are not met in the ancillary documentation, even if the diagnosis was listed by the physician. Lastly, review the physician's orders for observation status versus admission to inpatient care.

 

On the flip side, if during this type of review you are tempted to remove a MCC/CC (resulting in a lower-weighted MS-DRG), take the time to triple-check the significant procedures and the PDX (principal diagnosis) so as to arrive at a third choice of MS-DRG that may reimburse between the high and low amounts.

 

If an audit reveals a pattern of over-coding, the issues could lie in your CDI (clinical documentation improvement) team being overly aggressive, your physicians not understanding coding terminology, or a lack of coder knowledge of disease processes and/or secondary code assignment guidelines.

 

Under-Coding

 

Now, let's shift gears on the data-mining machine and look for cases of missed revenue opportunities that result from under-coding. This time, you are going to have to locate Medicare cases (regardless of discharge disposition) in which the actual LOS exceed GMLOS for the assigned MS-DRG by more than three days or approximately 33 percent (this is where the dollars statistically are relevant to the facility). In these cases, the D/D is not an issue when it comes to reimbursement. Of course, you always should bill the correct discharge disposition, but in these cases, no dollars would be at stake (at least that part of reimbursement based solely on the D/D assignment).

 

Perform an MS-DRG validation audit only. This means looking at the PDX and any CC/MCC already coded to see if they are valid. If there is a CC but no MCC, look for an opportunity to shift the CC to an MCC status. If there is no CC or MCC, look to add either. Here are your query opportunities. See if there are clinical criteria for a condition with diagnostic terminology that is not clearly documented, but able to be coded. Determine if your issues lie in the physician not documenting a condition that he or she clearly is treating (a condition that your CDI staff also are not looking into), or if it gets as far as the coder failing to take the opportunity to query (either due to lack of knowledge or because he/she is being pushed to "get the bills out the door.")  If you are applying a medical MS-DRG, look for any missed procedures, including a review of the daily nursing notes for bedside procedures, and re-review the emergency room documentation for procedures performed there.  Also, look closely at any coded procedures to ensure the accuracy of code assignments. Do not overlook this area. Many ICD-9 procedure codes that end in .x9 do not affect MS-DRG, but when the 9 is revised to a more specific digit, it can move into the OR procedure category and change the MS-DRG.

 



 

Educate your coders to know when to query, and allow them to decide when it is worth holding the bill a little longer. Have the coder review the actual LOS versus the GMLOS of the assigned MS-DRG. If the GMLOS is exceeded by 33 percent (again, the point at which there is monetary significance to the review), a second review of the case should be performed, looking at PDX, procedures and missed CC/MCC query opportunities. Question whether the CC/MCC that originally was not coded was missed simply because the doctor didn't document well enough or because the coder did not take the time to query (or didn't know to query). Finding multiple cases of under-coding/under-billing points to issues in clinical documentation and/or lack of knowledge within the staff of the CDI department (and/or lack of knowledge or effort within the inpatient coding arena.)

 

Wrap-up

 

I should mention that we just as easily could have talked about MS-DRG relative weights and CMI (case mix index) throughout this discussion, but actual LOS versus GMLOS is just simple to understand.

 

The object of the auditing exercise is to arrive at a final MS-DRG that has a similar actual LOS and GMLOS. The MS-DRG system was built on historical data that dictates how the GMLOS are assigned. What the RAC data-mining software is doing is simply comparing GMLOS with actual LOS in the common working file to select cases for complex review that may have been over-coded.

 

Keep in mind that during the demonstration project, the three RACs had a 33 percent "hit" rate. This means that for every 100 cases selected for complex review (using the techniques described above), 33 cases (initially) were denied for over-coding, causing facilities to have to enter the appeal process. That is a lot of paper and a lot of paperwork.

 

Conversely, though, for your own facility's financial well-being, you must be reimbursed at the highest weighted MS-DRG possible based on the documentation in the chart and supporting resources utilized, so always triple-check the cases that appear to be under-coded.

 

Remember, no one from CMS is going to come knocking on your door with a Publisher's Clearinghouse-sized refund check just because you under-billed - despite what the RAC scope of services implies!

 

About the Author

 

Janelle I. Wissler, RHIA, CCS, CMT, CCDS, is a Manager of Client Audits Precyse Solutions. Janelle was previously the Data Quality Manager for a 1,400-bed hospital system in Florida during the RAC demonstration project and has over 25 years in the HIM profession.

 

Contact the Author

 

jwissler@precysesolutions.com


This email address is being protected from spambots. You need JavaScript enabled to view it.