For the “Quantity of Care” investigation, The Seattle Times used several databases. Two were particularly important:
One was the National Plan and Provider Enumeration System (NPPES), which supplies a unique 10-digit number called the National Provider Identifier (NPI) for all healthcare providers. From this, we created a subset of doctors whose primary specialty was neurological surgery or orthopedic surgery of the spine. General orthopedic surgeons were not included, even though some may do spinal procedures.
Another data set was the Comprehensive Hospital Abstract Reporting System (CHARS) data, which is provided by the Washington State Department of Health. CHARS provides de-identified discharge data for patients admitted to hospitals in the state, including name of hospital, attending and operating physician NPIs, total charges and up to 25 diagnosis and 25 procedure codes for each patient. Looking at data from 2010 to 2015, we matched the NPI numbers from the NPPES database to isolate all patients whose physician was a neuro or spine specialist.
A SEATTLE TIMES SPECIAL REPORT
- Swedish CEO Tony Armada resigns
- Top Swedish neurosurgeon resigns
- U.S. Attorney’s Office launches investigation
- Providence CEO Rod Hochman apologizes
When calculating the number of cases for a specific surgeon, we only counted a case if the neuro or spine doctor was listed as the attending physician. The charges that are shown for individual surgeons reflect the total amount that was billed for the patient’s entire hospital stay and not just the surgeon’s portion.
All of the diagnosis and procedure codes were based on the International Classification of Diseases, 9th Revision (ICD-9), except for discharge data from October to December 2015. For those months, CHARS switched to ICD-10 codes. Where possible, we used equivalence tables to make sure the ICD-9 and ICD-10 coding was comparable. When analyzing the External Cause of Injury Codes – or E-codes – to quantify complications during elective admissions involving a brain or spine specialist, we only looked at the ICD-9 E-codes up through September 2015.
Using ICD-9 and ICD-10 coding, we analyzed discharge data from 2010 to 2015 for all patients who were admitted to the hospital with brain aneurysms. We then looked for how many patients underwent a coiling procedure versus a clipping procedure. To get data for the University of California, Irvine we gave the California Office of Statewide Health Planning and Development the same ICD-9 and ICD-10 codes to query their state’s patient discharge data for us.