Enhancing diagnosis of disorder through artificial intelligence and mobile health technologies without compromising accuracy

A computer system for generating a diagnostic tool for diagnosis of a behavioral disorder of a subject by applying machine learning to a diagnostic instrument for diagnosis of the behavioral disorder, wherein the diagnostic instrument comprises a set of diagnostic questions and corresponding selectable answers, the computer system comprising:

• one or more processors;
• and
• memory to store:
• one or more computer programs, the one or more computer programs comprising instructions for:
• receiving as input diagnostic outcomes and selected answers to the set of diagnostic questions of a plurality of subjects previously evaluated for the behavioral disorder,
• analyzing the diagnostic outcomes and the selected answers of the plurality of previously evaluated subjects with the machine learning to construct a behavioral classifier to distinguish among previously evaluated subjects with different diagnostic outcomes, the behavioral classifier comprising a subset of the set of diagnostic questions and corresponding selectable answers from the diagnostic instrument,
• determining the accuracy of the behavioral classifier comprising the subset of diagnostic questions and corresponding selectable answers, by testing the behavioral classifier against an independent source of clinical data, wherein the behavioral classifier comprising the subset of diagnostic questions and corresponding selectable answers has an accuracy over 90%,
• generating the diagnostic tool for diagnosis of the behavioral disorder, wherein the diagnostic tool comprises the behavioral classifier and the subset of diagnostic questions and corresponding selectable answers having the accuracy greater than 90%, and
• configuring a computing device accessible by a user to display the subset of diagnostic questions and corresponding selectable answers to the user, to collect user selected answers for the subset of diagnostic questions, and to provide the user selected answers as input into the behavioral classifier in order to diagnose the behavioral disorder of the subject with the behavioral classifier.

A non-transitory computer-readable storage medium storing one or more computer programs configured to be executed by one or more processing units at a computer comprising instructions for:

• receiving as input diagnostic outcomes and selected answers to a set of diagnostic questions of a plurality of subjects previously evaluated for a behavioral disorder, wherein the set of diagnostic questions comprise diagnostic questions of a diagnostic instrument for diagnosis of the behavioral disorder, and wherein the diagnostic instrument further comprises corresponding selectable answers for the set of diagnostic questions;
• analyzing the diagnostic outcomes and the selected answers of the plurality of previously evaluated subjects with machine learning to construct a behavioral classifier to distinguish among previously evaluated subjects with different diagnostic outcomes, the behavioral classifier comprising a subset of the set of diagnostic questions and corresponding selectable answers from the diagnostic instrument;
• determining the accuracy of the behavioral classifier comprising the subset of diagnostic questions and corresponding selectable answers, by testing the behavioral classifier against an independent source of clinical data, wherein the behavioral classifier comprising the subset of diagnostic questions and corresponding selectable answers has an accuracy over 90%;
• generating a diagnostic tool for diagnosis of the behavioral disorder, wherein the diagnostic tool comprises the behavioral classifier and the subset of diagnostic questions and corresponding selectable answers having the accuracy greater than 90%;
• and
• configuring a computing device accessible by a user to display the subset of diagnostic questions and corresponding selectable answers to the user, to collect user selected answers for the subset of diagnostic questions, and to provide the user selected answers as input into the behavioral classifier in order to diagnose the behavioral disorder of the subject with the behavioral classifier.