Application of an Artificial Intelligence Model to Detect Alpha-1 Antitrypsin Deficiency: Characterizing the Study Population
Rationale: Alpha-1 antitrypsin deficiency (AATD), an under-recognized genetic condition, can cause chronic obstructive pulmonary disease (COPD) and/or liver disease in adults. Symptoms of individuals who have COPD and/or liver disease due to AATD are often indistinguishable from those without deficiency, necessitating additional tools to establish diagnosis. Using claims data from the large United States Komodo Health database, we previously reported an artificial intelligence model to identify symptomatic patients who likely had AATD but were unrecognized. In this current study, we used retrospective deidentified patient data generated from the Cleveland Clinic Electronic Medical Record (EMR) to define and characterize patient cohorts utilized to calibrate and refine the model. Methods: Data from deidentified patient records, sourced from the Cleveland Clinic EMR between January 1, 1998 and August 15, 2024, were analyzed. Patients who had positive AATD diagnoses were grouped into categories depending upon how diagnoses were confirmed (genetic: genotype confirming AATD; severe: severe deficiency in AAT serum levels [<57 mg/dL]; and mild: mild deficiency in AAT serum levels [57-90 mg/dL]). Clinical phenotype and treatment patterns for these subgroups were analyzed based on the presence of relevant medical codes in patients’ records. Results: Of 794 patients with AATD, 290 patients (36.5%) were identified by genotype (genetic), the remaining patients were characterized by serum AAT level in the severe (n=117, 14.7%) and mild (n=387, 48.7%) categories. Overall, 50.4% of patients were male; median age was 63 years. Prevalence of common comorbidities was lower in the mild subgroup versus the total population and other subgroups: COPD (total: 40.1%, genetic: 44.8%, severe: 47.9%, mild: 34.1%), unspecified liver disease (total: 34.9%, genetic: 41.4%, severe: 40.2%, mild: 28.4%), and emphysema (total: 31.7%, genetic: 36.9%, severe: 38.5%, mild: 25.8%). Overall, 12.5% of patients had any AATD augmentation therapy drug in their records. When assessed by subgroup, 22.4% of the genetic group, 17.9% of the severe group, and 3.4% of the mild group had an augmentation therapy drug listed in their records; patients with emphysema had higher rates of receiving augmentation therapy than patients without (Figure 1). Conclusions: Within the Cleveland Clinic EMR, this analysis identified phenotypic and treatment pattern differences among cohorts of patients confirmed to have AATD. Lower frequencies of comorbidities, as well as lower rates of patients who received augmentation therapy for AATD, were observed in patients with mild deficiency of AAT compared with patients with confirmed genetic diagnoses and those with severe deficiency of AAT.
Read more