HIMSS14: Going beyond predictive modeling for ICD-10

ORLANDO—With ICD-10 looming, leaders at Rice Memorial Hospital, a 136-bed hospital in Minnesota, suspected clinical documentation and coding problems could hinder future compliance efforts.

Also, the hospital struggled with departments “communicating in silos,” managing the many layers of regulatory initiatives, and improving its quality indicator scores with a larger goal of realizing savings, said Jackie Hinderks, director of revenue and reimbursement, speaking at Health Information and Management Systems Society's annual conference.

The hospital hired a consultant to get a bigger picture of where it stood with documentation, and how to improve. The hospital not only looked at predictive modeling, but underwent a comprehensive gap analysis program of the current environment to avoid the built-in assumption that ICD-9 is done correctly.

As part of this effort, the consultant performed a risk stratification, which involved GEMS analysis to convert DRGs to ICD-10. “The real objective was obtaining a sample selection of charts that are the best choice of records for an audit of red flags,” said Jill Wolf, vice president of compliance, VitaWare, who worked on the project.

The analysis looked at both high-volume DRGs and low-volume DRGs. “High-volume DRGs are not necessarily where the risk is. We didn’t want Rice to miss out on any opportunities,”

Only three of the 10 DRGs most commonly used at Rice were high risk, Wolf noted—thus it made sense to identify opportunities to improve coding in all of them. The analysis zeroed in on several areas: DRG shift, ICD-9 and ICD-10 groupers and principal diagnoses.

In the case of Rice Memorial Hospital, the analysis revealed coding mistakes and insufficient documentation under the current system. Coding errors stemmed from a knowledge gap among junior coders, guideline ambiguity, workload and simple mistakes, Wolf said.

The analysis also identified potential financial gains and losses from the shift from ICD-9 to ICD-10, and found 14 DRGs with net negative reimbursement and 12 DRGs with net positive reimbursement. While they nearly balanced each other out, Rice emphasized that improving coding as a whole, regardless of the financial incentive, leads to improved processes and care.

With the insight gleaned from this analysis, Rice Memorial put a “laser focus” on coding education—which includes side-by-side comparison with coders—and clinical document improvement. “It was helpful to physicians to show how to change documentation practices,” Wolf said.

The analysis enabled Hinderks to obtain buy-in from executives to financially support this training, she said.