Smartphone App PACOM to Provide Advice Regarding Self-Triage for the Acute Primary Care Needs of Children: Accuracy of Algorithms.

Abstract

We developed a phone app, PACOM (Parents Application Conseils et Orientation Médicale), to provide medical advice to caregivers based on several algorithms and a series of binary questions related to children's symptoms. We compared the recommendations of the PACOM algorithms and clinicians for children visiting the emergency department (ED). Between January and February 2022, we prospectively recruited French-speaking parents of children without any chronic disease who presented to the pediatric ED with any complaint except for mental health problems or trauma. Isolated head trauma was included. They completed questionnaires and the various PACOM algorithms. The first algorithm, called "Quick Look," was developed to identify children with life-threatening emergencies. The standard reference was the advice of the ED clinicians who were blinded to the parental responses to the PACOM algorithm questions. The recommendations included "call urgent medical support," "visit the ED," "visit your general practitioner within 24 hours," and "visit your general practitioner in the next days." The study included 269 parents. The response rate was 75%. The median age of the children was 3 years with interquartile range: 1 to 7 years. In total, 268 children completed the "Quick Look," 141 "fever," 83 "abdominal pain," 72 "cough," 70 "vomiting," 130 questionnaires relative to other proposed symptoms, and 70 "other symptom" questionnaires. The PACOM recommendations were "call urgent medical assistant" for 98 children, "ED visit" for 131, "visit general practitioner within 24 hours" for 13, and "visit general practitioner during the next days" for 24. The sensitivity and specificity of the PACOM recommendation to "call urgent medical support or visit the ED" were 98.1% (95% confidence interval, 95.5-100.00) and 22.1% (95% confidence interval, 15.3-28.8), respectively. The PACOM algorithms has high sensitivity but low specificity for reducing ED visits and calls for urgent medical support.