Framework for increasing data accuracy in MRA.

* Opposite valence factors, “Lack of abstractor training decreases accuracy of abstracted data,” “An incomplete review of the medical record (e.g., not reading all pages from the required time period) decreases the accuracy of abstracted data,” “Data element definitions that lack suggestions for where in the chart to find data values,” “Data abstracted from a complete medical record are more accurate than those that are abstracted from medical records with omissions,” “Abstractor (human) error is a factor in decreasing the accuracy of abstracted data,” and “Data abstracted from a medical record that is free from error are more accurate than those abstracted from a medical record containing errors,” were omitted from framework.† Combined factors “Misuse of the coding system” and “Misunderstanding the coding system,” and moved to the training category.‡ Original text “Abstractor human error” restated to create an actionable item.§ “Data elements requiring the abstractor to do calculations (e.g., convert units or score questionnaires) are less accurate than those that do not” and “Data elements that are abstracted directly from medical records) are more accurate than those requiring mapping or interpretation” were combined.Framework for increasing data accuracy in MRA.