A practical approach to identifying autistic adults within the electronic health record.

This study looked at ways to find autistic adults in hospital computer records. Researchers found that just using autism diagnosis codes isn't enough - some autistic people don't have the right codes, and some people with codes aren't actually autistic. They developed a better method that also looks for autism-related words in medical notes. This approach found most autistic patients (95-99%) but sometimes included people who weren't autistic.

The findings suggest hospitals need better ways to identify autistic patients in their records.

Researchers developed a method to identify autistic adults in electronic health records using both diagnostic codes and natural language processing of medical notes. They studied two groups: 418 patients with autism diagnostic codes (86% confirmed as autistic) and 136 patients with autism-related terms but no codes (35% confirmed as autistic). The approach achieved high sensitivity (95-99%) in detecting autism cases, meaning it successfully identified most autistic patients. However, specificity was moderate (68-81%), indicating some false positives.

The study demonstrates that relying solely on diagnostic codes may miss autistic patients or include incorrect cases, while combining codes with text analysis improves accuracy.

Healthcare systems may benefit from combining diagnostic codes with natural language processing to better identify autistic patients. Current coding practices may miss autistic individuals or incorrectly classify others. Improved identification methods could enhance autism research using electronic health records and support better healthcare delivery for autistic adults.

Single institution study limits generalizability. Chart review methodology and classification rubric not detailed. No information on inter-rater reliability. Sample demographics not provided. Validation in other healthcare systems needed.

The electronic health record (EHR) provides valuable data for understanding physical and mental health conditions in autism. We developed an approach to identify charts of autistic young adults, retrieved from our institution's de-identified EHR database. Clinical notes within two cohorts were identified. Cohort 1 charts had at least one International Classification of Diseases (ICD-CM) autism code.

Cohort 2 charts had only autism key terms without ICD-CM codes, and at least four notes per chart. A natural language processing tool parsed medical charts to identify key terms associated with autism diagnoses and mapped them to Unified Medical Language System Concept Unique Identifiers (CUIs). Average scores were calculated for each set of charts based on captured CUIs. Chart review determined whether patients met criteria for autism using a classification rubric.

In Cohort 1, of 418 patients, 361 were confirmed to have autism by chart review. Sensitivity was 0.99 and specificity was 0.68 with positive predictive value (PPV) of 0.97. Specificity improved to 0.81 (sensitivity was 0.95; PPV was 0.98) when the number of notes was limited to four or more per chart. In Cohort 2, 48 of 136 patients were confirmed to have autism by chart review.

Sensitivity was 0.95, specificity was 0.73, and PPV was 0.70. Our approach, which included using key terms, identified autism charts with high sensitivity, even in the absence of ICD-CM codes. Relying on ICD-CM codes alone may result in inclusion of false positive cases and exclusion of true cases with autism.