A more accurate method to report quality of care by hospitals and/or healthcare providers has been developed by Cardinal Health.

The new methodology is based on clinical data rather than administrative data, according to Cardinal Health.

Cardinal Health said its study across six major diseases showed that using clinical data more accurately accounts for the severity of an illness and leads to more clinically valid quality measurements than traditional methods.

Pay-for-performance (P4P) and public accountability reporting have seen increased interest across healthcare providers, government agencies and insurance companies as incentive-based initiatives to help control rising healthcare costs and improve quality, the company said.

Risk-adjusted mortality rates are widely accepted quality measures for the healthcare industry. Thesequality measures are made by comparing expected mortality rates for specific disease groups to the actual mortality rates that occur.

Traditionally, the models used to predict mortality are based on the analysis of billing or administrative data from the healthcare provider, according to Cardinal Health.

Physicians, healthcare administrators and researchers all point to fundamental flaws in using administrative data because it does not appropriately account for severity of illness, which can disadvantage those hospitals that take in the most critically ill patients, Cardinal Health said.

To address the challenge of developing a clinically valid and cost-effective reporting system, researchers from Cardinal Health, the Center for Outcomes Research at the Children’s Hospital of Philadelphia and the University of Pennsylvania School of Medicine analyzed data from more than one million admissions to create models that are clinically sound and adequately adjusted for severity of illness. The new approach maximizes the use of automated laboratory data, but is more cost-effective than relying on chart abstraction.

The researchers found that laboratory data contributed more to the prediction of mortality compared to any other data source across most diseases except for stroke, where altered mental status is more important than lab results, Cardinal Health said.

Laboratory data proved to be up to 67 times more important in predicting mortality than administrative data alone, the company said.

"These findings are significant in light of increased interest in pay-for-performance and public reporting initiatives and highlight the urgent need for valid, clinically-based risk-adjusted approaches," said Dwight Winstead, group president of clinical technologies and services for Cardinal Health.

Cardinal Health published the final findings in the August issue of Medical Care, the official journal of the medical care section of the American Public Health Association.

All six models were found to be clinically plausible and can be cost-effective to implement for P4P and public accountability reporting.

Address: Cardinal Health, 7000 Cardinal Place, Dublin, OH 43017; (614) 757-5000, .