An Algorithm for Predicting Inadequate Response to Initial PAH Therapy Developed

CT scan PAH
CT scan PAH
Research presented at CHEST 2021 used machine learning to develop an algorithm for predicting poor response to initial PAH therapy, based on claims data.

The following article is a part of conference coverage from CHEST 2021, being held virtually from October 17 to October 20, 2021. The team at Pulmonology Advisor will be reporting on the latest news and research conducted by leading experts in the field. Check back for more from CHEST 2021.


A claims-based algorithm developed by researchers performed well in predicting the time to unsatisfactory response to initial pulmonary arterial hypertension (PAH) therapy, according to study results presented at the CHEST 2021 Annual Meeting, held virtually from October 17 to 20, 2021.

Using machine learning and claims data from 2010 through 2019, the researchers sought to identify patients with PAH with the poorest response to initial therapy and who had the greatest need for early combination therapy for PAH.

Unsatisfactory response was defined as the first instance of new PAH therapy; PAH-related hospitalization or emergency room visit; lung transplant or atrial septostomy; PAH-related death; or initiation of chronic oxygen therapy. The predicted time to outcome was defined as the time at which the estimated survival function reached less than 50%.

The analysis included 4781 patients (median age, 69 years; 59% women).

Over a median follow-up of 14 months, 3169 (66%) participants had an unsatisfactory response to therapy. Across risk quartiles, 15%, 49%, 61%, and 76% of patients had an unsatisfactory response (median time, 15, 11, 7, and 3 months, respectively). The algorithm had 82% sensitivity and 55% specificity at 12 months.

Risk factors with the largest importance included health care resource use, including, PAH-related outpatient visits, pulmonologist visits, cardiologist visits, and all-cause hospitalizations; time since first PAH diagnosis; time since index date; Charlson Comorbidity Index; breathlessness; and age.

“This algorithm could be used to identify patients at higher risk of unsatisfactory response to initial PAH treatment,” the researchers said. “The timely identification of these patients could enable earlier intervention and improve prognosis,” they added.

Disclosure: Multiple authors declared affiliations with the pharmaceutical makers and industry. Please refer to the original article for a full list of disclosures.


Gauthier-Loiselle M, Tsang Y, Lefebvre P, et al. Development and evaluation of a predictive algorithm for unsatisfactory response among patients with pulmonary arterial hypertension using health insurance claims data. Presented at: CHEST 2021; October 17-20, 2021; Orlando, FL/Virtual. Abstract A2267-A2268.