Background: The rarity of skull base pathologies necessitates multicenter data to study them
effectively. One method to achieve this is the use of de-identified nationwide insurance
databases. A major strength of such databases is the large number of otherwise rare
cases that can be identified using diagnosis or procedure codes, which may increase
the relevance of findings and aid in statistical analysis.
Aim: We aim to describe our experience using the Optum Clinformatics Data Mart, an anonymized
nationwide private insurance database.
Methods: Based on our experience in using this database in studying transsphenoidal resection
of pituitary adenomas, with a focus on Cushing’s disease, we describe the benefits
and disadvantages of these data, in addition to lessons learned.
Results: Using both diagnosis and procedure codes to identify transsphenoidal resections of
pituitary adenomas between the years 2001 and 2019, a total of 7,295 records were
found. The quality of available data in this database depended on the documentation
and coding in patients' charts, which ultimately shaped the insurance claims. Familiarity
with different medical coding systems and their use was necessary for effective querying
of these databases. Date-sensitive information, such as date of surgery, was not accurately
reflected in the claims and in some cases may affect analysis. Some records were sparse
in data, likely reflecting patients with dual insurance plans. Some very relevant
medications in risk stratifying patients for skull base surgeries, such as aspirin,
are over the counter and were not captured in insurance claims. Analysis of these
data required taking these caveats into account, and required modification of the
investigation at times in order for analyses to be valid.
Conclusion: Proper planning, expectation setting and analysis is needed when using a nationwide
insurance database for skull base research purposes. These challenges illustrate the
need for a centralized, nationwide, collaborative skull-base registry.
