Cath Lab Digest

Login to Download
PDF version

Retrospective database review

  Regarding DaVanzo JE, El-Gamil AM, Dobson A, Sen N. A retrospective comparison of clinical outcomes and Medicare expenditures in skilled nursing facility residents with chronic wounds. Ostomy Wound Manage. 2010;56(9):44–54: Anyone who is involved with skilled nursing facilities clearly understands the variability in studies using MDS data from multiple facilities, where the information about wounds can be inaccurate, inconsistent, and diametrically opposed to useful understanding for research purposes (specifically, backstaging pressure ulcers). Using a retrospective approach to analyzing MDS data related to wounds is fraught with indeterminable inaccuracies. Time to healing, as the authors admitted, could not be determined; therefore, the total cost of care was determined by multiple other factors. For most clinicians in wound care, the wound is healed when clinically it is healed; not by analyzing MDS data tabulated on a frequency based on the Centers for Medicare and Medicaid Services (CMS) guidelines, rules, and regulations. Total wound episode, as defined in the study, is based on claims data, not the actual care of or observation of the wound.

  The authors state that the predicted values for all study group dependent variables are calculated independently based on the comparison group costs so the predicted values do not sum to the total expenditures. Then why bother to tabulate Table 6, which explicitly gives the total costs of care for both groups? If the study group Part A and Part B payments are $26,568.58 and $7,270.91, respectively, it is not clear how the total Medicare payments for Part A and B can be only $21,449.64, as noted in Table 6. Using the data from that table, the total costs of care for both groups come very close.

  Further, it is concerning that the article seems to support, through retrospective analysis of MDS claims data and using sophisticated, incomprehensible statistical analysis, a conclusion that supports the funding organization to promote future studies using organizations like theirs. In addition, if the physician group is being paid by the facility to provide wound care services such as debridement, these costs also should be added to the analysis of the total cost of care. Because the debridement codes 97597 and 97598 are the most frequent codes utilized in the skilled nursing facility on Part A patients, the clinicians would not be reimbursed for their services because these codes, for the Part A patients, are “owned” by the facility, not the clinicians.

Scott Matthew Bolhack, MD, MBA, CMD, CWS, FACP, FAAP
CEO, TLC HealthCare Quality Improvement Organization
Tucson, Arizona

Reply

  The article authors appreciate the difficulties encountered when using claims data to understand clinical processes. However, as experienced users of claims data, the authors also believe that claims data are often the most cost-effective, if not only, source of retrospective information on patient outcomes. It is hoped that this study is one of many that points the way toward informed use of claims data to investigate numerous types of clinical processes. As budgets tighten, claims data may become the only way to monitor our healthcare system.

  In order to risk-adjust the patients within the study and comparison groups to ensure an appropriate analysis, the regression analysis calculated a predicted cost of Medicare Part A, Part B, and inpatient hospitalization care separately. As a result, as noted, the sum of these parts does not equal the total cost of care. Rather, using a prediction model, each Medicare payment component is risk-adjusted so patients are comparable across groups. To the same effect, the predicted total Medicare Part A and B payment is risk-adjusted to ensure patient comparability.