PTSD diagnoses among Iraq and Afghanistan veterans: comparison of administrative data to chart review.

To guide budgetary and policy-level decisions, the U.S. Department of Veterans Affairs (VA) produces quarterly reports that count the number of Iraq and Afghanistan veterans with International Classification of Diseases, 9(th) Revision, Clinical Modification (ICD-9-CM) codes for posttraumatic stress disorder (PTSD; 309.81) in their electronic medical record administrative data. We explored the accuracy of VA administrative data (i.e., diagnostic codes used for billing purposes), by comparing it to chart review evidence of PTSD (i.e., medical progress notes and all other clinical documentation contained in the entire VA medical record). We reviewed VA electronic medical records for a nationwide sample of 1,000 Iraq and Afghanistan veterans with at least one ICD-9-CM code for PTSD in their VA administrative data. Among veterans sampled, 99.9% had 2 or more ICD-9-CM codes for PTSD. Reviewing all VA electronic medical record notes for these 1,000 veterans for the full course of their VA health care history revealed that PTSD was diagnosed by a mental health provider for 89.6%, refuted for 5.6%, and PTSD was never evaluated by a mental health provider for 4.8%. VA treatment notes for the 12 months preceding chart review showed that 661 veterans sampled received a VA PTSD diagnosis during that 12-month timeframe, and of these 555 were diagnosed by a mental health provider (83.9%). Thus, the presence of an ICD-9-CM code for PTSD approximated diagnoses by VA mental health providers across time points (89.6% for entire treatment history and 83.9% for 12 months prior to chart review). Administrative data offer large-scale means to track diagnoses and treatment utilization; however, their limitations are many, including the inability to detect false-negatives.