AI-Powered Patient Identification to Optimize Care

Share publication:

Introduction: The continued integration of artificial intelligence (AI) into healthcare is anticipated to improve clinician productivity, patient outcomes and the delivery of care. One key role for AI is to identify patterns in large datasets to help to provide clinicians with insight that would be near impossible to detect in normal practice. This study was designed to determine if AI could be utilized in dermatology to capture plaque psoriasis (PsO) patients, demographic information and PsO severity. 

Methods, Results: Utilizing the AI engine DARWEN™, the electronic heath record (EHR) data of 4 large dermatology clinics were queried using a variety of relevant terms. The engine queried the clinical notes generated from regular clinical assessments and examinations for this data analysis.   

The algorithm scanned the EHRs of 10,295 patients in the four dermatologist practices in Canada using the DARWEN™ engine between May 13, 2022, and December 2, 2022. There were 663 patients identified with PsO. Within this group, 269, 135 and 259 patients were identified with mild, moderate or severe PsO based on their PASI scores, respectively. Each dermatologist was provided with a report of these patients, their PASI score, date of last PASI score and date of last appointment.  

Conclusion: AI technology has the potential to transform healthcare practice. This study demonstrated that AI could identify a complete list of PsO patients, their severity and last visit in busy dermatology practices. The results of this program could be used to identify these patients who may require different interventions. This could include revisiting treatment goals, satisfaction with the treatment regimen, adherence, and the need for intensification/modification of current pharmacotherapy.