Advertisement

DEVELOPMENT AND EVALUATION OF A PREDICTIVE ALGORITHM FOR UNSATISFACTORY RESPONSE AMONG PATIENTS WITH PULMONARY ARTERIAL HYPERTENSION USING HEALTH INSURANCE CLAIMS DATA

      TOPIC: Pulmonary Vascular Disease
      TYPE: Original Investigations
      PURPOSE: Pulmonary arterial hypertension (PAH) is a rapidly progressing and life-threatening disease. A delay in identifying patients at risk of progression can lead to inadequate treatment and poor outcomes. Predictive algorithms could help identify patients with the greatest need for early combination therapy. Current algorithms require access to clinical values that may not be readily available. This study aimed to use machine learning to develop a predictive algorithm for unsatisfactory response to initial therapy using health insurance claims.
      METHODS: Adult patients with PAH were retrieved from Optum's de-identified Clinformatics® Data Mart (1/1/2010-12/31/2019). Patients initiated on a first PAH therapy (index date) were included in the study if they had continuous healthcare plan enrollment in the prior 12 months. The algorithm was developed using patient-month data for each month between the index date and end of follow-up (i.e., earliest of end of continuous enrollment, death, or unsatisfactory response). At each time point, risk factors were assessed in the prior 12 months to allow for variations over the follow-up period. Unsatisfactory response (outcome) was defined as the first instance of: (1) new PAH therapy (i.e., switch or add-on), (2) PAH-related hospitalization or emergency room visit, (3) lung transplant or atrial septostomy, (4) PAH-related death, or (5) chronic oxygen therapy initiation. The algorithm was developed using random survival forests and predicted the survival function (i.e., risk of not having the outcome) at each time point. The time at which the estimated survival function reached <50% was defined as the predicted time to outcome. Algorithm performance was evaluated based on discrimination (c-statistic), sensitivity, and specificity. Risk groups were created based on the distribution of predicted time to outcome.
      RESULTS: In total, 4,781 patients were included (median age=69 years; 59% female). Over a median follow-up of 14 months, 3,169 (66%) had unsatisfactory response. Risk factors with the largest importance included healthcare resource use (i.e., PAH-related outpatient visits, pulmonologist visits, cardiologist visits, all-cause hospitalizations), time since first PAH diagnosis, time since index date, Charlson Comorbidity Index, breathlessness, and age. The c-statistic was 0.73; the algorithm had 82% sensitivity and 55% specificity at 12 months. Across risk quartiles, 15%, 49%, 61%, and 76% had unsatisfactory response (median time: 15, 11, 7, and 3 months, respectively).
      CONCLUSIONS: We developed a claims-based algorithm that performed well in predicting the time to unsatisfactory response to initial PAH therapy.
      CLINICAL IMPLICATIONS: This algorithm could be used to identify patients at higher risk of unsatisfactory response to initial PAH treatment. The timely identification of these patients could enable earlier intervention and improve prognosis.
      DISCLOSURES: Employee relationship with Janssen Pharmaceuticals Please note: June 16, 2017 to now Added 04/26/2021 by Peter Agron, source=Web Response, value=Salary
      Employee relationship with Analysis Group Please note: May 01, 2020 Added 04/22/2021 by Lucas Bennett, source=Web Response, value=Consulting fee
      Removed 04/22/2021 by Lucas Bennett, source=Web Response
      Employee relationship with Analysis Group Please note: 2019-2021 Added 04/22/2021 by Lucas Bennett, source=Web Response, value=Salary
      Employee of Analysis Group, Inc. (AGI) relationship with Actelion Pharmaceuticals, Inc. Please note: 2018-2021 Added 04/22/2021 by Lucas Bennett, source=Web Response, value=AGI received consultancy fees
      Employee relationship with Analysis Group, Inc. Please note: 2012-2021 Added 04/22/2021 by Marjolaine Gauthier-Loiselle, source=Web Response, value=Salary
      Employee of Analysis Group, Inc. (AGI) relationship with Actelion Pharmaceuticals, Inc. Please note: 2018-2021 Added 04/22/2021 by Marjolaine Gauthier-Loiselle, source=Web Response, value=AGI received consultancy fees
      Employee relationship with Analysis Group, Inc. Please note: 2001-2021 Added 04/23/2021 by Patrick Lefebvre, source=Web Response, value=Salary
      Employee of Analysis Group, Inc. (AGI) relationship with Actelion Pharmaceuticals, Inc. Please note: 2018-2021 Added 04/23/2021 by Patrick Lefebvre, source=Web Response, value=AGI received consultancy fees
      Employee relationship with Janssen Scientific Affairs, LLC Please note: Nov 2013 to present Added 04/27/2021 by Karimah Lynum, source=Web Response, value=Salary
      Employee relationship with Janssen Scientific Affairs, LLC Please note: Nov 2013 to present Added 04/27/2021 by Karimah Lynum, source=Web Response, value=Stock
      Employee relationship with Janssen Please note: May 2020 to present Added 04/21/2021 by Sumeet Panjabi, source=Web Response, value=Salary
      Employee relationship with Janssen Please note: May 2020 to present Added 04/21/2021 by Sumeet Panjabi, source=Web Response, value=Ownership interest
      Employee relationship with Analysis Group, Inc. Please note: 2000-2021 Added 04/23/2021 by Jimmy Royer, source=Web Response, value=Salary
      Employee of Analysis Group, Inc. (AGI) relationship with Actelion Pharmaceuticals, Inc. Please note: 2018-2021 Added 04/23/2021 by Jimmy Royer, source=Web Response, value=AGI received consultancy fees
      Employee relationship with Analysis Group, Inc. Please note: 2020-2021 Added 04/22/2021 by Emma Russon, source=Web Response, value=Salary
      Employee of Analysis Group, Inc. (AGI) relationship with Actelion Pharmaceuticals, Inc. Please note: 2021-2021 Added 04/22/2021 by Emma Russon, source=Web Response, value=AGI received consultancy fees
      Employee relationship with Janssen Scientific Affairs, LLC Please note: 2015 to present Added 04/20/2021 by Yuen Tsang, source=Web Response, value=Salary
      Employee relationship with Janssen Scientific Affairs, LLC Please note: 2015 to present Added 04/20/2021 by Yuen Tsang, source=Web Response, value=Stock