Wednesday, March 18, 2015

Study shows switch will create chaos in measuring patient safety

According to a study published in the January 2015 issue of the Journal of the American Medical Informatics Association, the switch to ICD-10-CM will result in ...substantial hospital safety reporting errors... In other words, the metrics hospitals use to report their performance on patient safety will be so distorted by the switch, valid pre- and post-switch comparisons will result in either (1) underreporting safety, leading patients to unnecessarily doubt the safety of a given hospital or (2) overreporting safety, leading patients to place false confidence in their hospitals.

The reason is that the national system for reporting safety, the so-called Patient Safety Indicators (PSIs) created by the Agency for Healthcare Research and Quality, rely on diagnosis codes for their calculation.  The switch to ICD-10-CM changes the meaning of the calculations, resulting in different PSI values based solely on whether ICD-9-CM or ICD-10-CM is used to compute them.

The study found that of 23 PSIs calculated using ICD-9-CM, 5 PSIs had no mapping to ICD-10-CM, 15 PSIs had 'convoluted' mappings to ICD-10-CM, and only 3 had straightforward mappings to ICD-10-CM.  A 'convoluted' mapping means that the ICD-9-CM code has no equivalent in ICD-10-CM, nor is it a more specific or more general code than the ICD-10-CM code(s) to which it maps.  That is, there is no ICD-10-CM that stands in a parent child relationship with it, or vice versa.

The authors estimated the impact of under- and over-reporting of the 15 convoluted PSIs.  They found a range of 100% underreporting of patient safety to 18% overreporting.

The switch to ICD-10-CM will create chaos for patients trying understand how safe an environment a given hospital offers.  We should not switch.

1 comment:

Unknown said...

Dr. Vergil Slee, pioneer in medical informatics (he developed the first hospital discharge abstract system, and led the team which created ICD-9-CM), has written persuasively on the innate defects of "category coding" as exemplified by ICD-10. See ICD-10-CM: A Close Look andThe Endangered Medical Record: Ensuring Its Integrity in the Age of Informatics (available free here in pdf format or at Amazon).