News: Study finds problem list inaccurate for risk adjustment calculations
EHR records may not be able to accurately determine case mix, according to a study published January 18 in the American Journal of Managed Care, Revenue Cycle Advisor reported.
An EHR problem list-based comorbidity assessment data can be used to detect CCs/MCCs and Charlson Comorbidity Index (CCI) scores. The study compared EHR problem list-based scores for specific MCCs to scores based on data entered in the EHR free-text field.
An EHR-based assessment of the problem list in contrast to a manual review was, in general, inaccurate and not able to capture MCCs although specificity was above 94% for all documented CCs, the study authors concluded. The EHR problem-list based assessment was able to detect and correctly identify:
- Diabetes in 46% of records
- Chronic obstructive pulmonary disease in 42% of records
- Cerebrovascular disease in 32% of records
- Peripheral vascular disease in 31% of records
- Congestive heart failure in 23% of records
- Myocardial infarction in 8% of records
- Liver disease in 1% of records
In contrast, manual review of free-text field data was able to capture more detailed, accurate CCI scores that were also predictive of long-term other-cause mortality.
As reimbursement continues to move from fee-for-service to quality-based and risk-adjusted methodologies, capturing accurate CC/ MCCs will be increasingly vital, according to Revenue Cycle Advisor. Relying on EHR problem lists to capture risk data may seem like a simple method to save time and resources while improving documentation, but the study authors conclude that case mix data based on EHR problem lists should be validated before it is applied to risk scores.
Editor’s note: This article originally appeared in Revenue Cycle Advisor. To read an article about CDI’s evolution to risk adjustment, click here. To watch the free CDI Week 2016 webinar on the topic of risk adjustment, click here. To read an ACDIS Position Paper on the subject of CDI and EHRs, click here.