Endocrine Abstracts

Introduction: Pituitary disorders are associated with increased mortality and morbidity. Data on the prevalence of pituitary disorders is scarce. Formal routine coding of diagnoses in outpatient endocrine practice lags behind medical coding of inpatients. Standardised coding could improve our understanding of disease burden and highlight areas of increasing need within our services.

Objective: Automatically extract and assign SNOMED codes for endocrine diagnoses from structured outpatient correspondence. The coded information was used to ascertain the prevalence of pituitary disorders in patients attending endocrine outpatient clinics at Aberdeen Royal Infirmary, Aberdeen, Scotland.

Method: Retrospective study conducted in a tertiary outpatient endocrine clinic. Patients from postcodes within 2 regional areas of NHS Grampian were included (Aberdeen City and Aberdeenshire), who attended clinics between 1st January 2018 to 31st December 2019. Based on the mid population estimates from the National Records of Scotland, the total study population was 489,880 inhabitants. After each clinical consultation, which may be face to face or remote, a structured letter was created, containing a detailed problem list. Structured correspondence was introduced to provide more uniform output from clinic appointments and ensure key information was captured (despite clinics being conducted by several different healthcare professionals). An automated script was developed to extract each problem statement and update a database. Unique problem list entries were manually coded using the ‘disorder’ concepts from SNOMED CT (UK edition).

Results: A total of 1870 patients attended the outpatient services (1251 of them were female with a male to female ratio of 1:2); age ranged from 16-96 years. 464 (21.2%) had pituitary disorders. Pituitary disorders have a prevalence of 94.7 per 100,000 inhabitants, with a mean age of 51yrs (SD ± 18) and male to female ratio at 1: 1.5. The most common diagnosis was pituitary adenoma with a prevalence rate of 65.3 per 100,000 inhabitants. The prevalence per 100,000 for the subcategories were functionless pituitary adenoma (22), acromegaly (7.1) and pituitary-dependent Cushing’s disease (3.1). The prevalence is consistent with previous studies.

Conclusion: Automated text parsing of structured endocrine correspondence allowed the creation of a database of endocrine problems. SNOMED coding enabled us to create a common endocrine reference set. Standardised coding helped assess the prevalence of pituitary disorders in this population. A similar automated approach to allow medical coding of routine medical correspondence could help improve understanding of the epidemiology of conditions managed predominately in outpatient settings, as well as facilitate quality improvement projects.