Abstract: Embodiments relate to technological systems and methods for determining return on investment (ROI) of a program (e.g., medical intervention), which may be used to retrospectively or prospectively evaluate the value that the program delivered or will deliver to a patient (e.g., a participant in a program). In some embodiments, a first set of training data associated with a patient population is retrieved. The training data may be classified as belonging to a patient treatment group including patients participating in a program or a patient control group including patients not participating in the program. A plurality of base machine learning models may be trained using a dual-training process to generate both counterfactual values and ROI predictions associated with patients. A global machine learning model may then be trained on the counterfactual values and ROI predictions output by the base machine learning models.

Abstract: Disclosed herein are system, method, and computer program product embodiments for creating and utilizing machine learning models to generate a provider performance index. In some embodiments, a first and second target may be selected. The first and second targets may respectfully include target values. First and second adjusted target values may be determined by combining the first and second target values with first and second external weights. The external weights may be generated by respective first and second machine learning models. The machine learning models may correspond to the respective targets. First and second error values may be determined based on differences between respective target and adjusted target values. The error values may be normalized and combined to generate an index value.

Abstract: Embodiments relate to training and applying a parallel-structured neural network. In some embodiments, a dataset of a plurality of claims is received, where each claim of the plurality of claims includes multiple fields. Features, including numerical features and a plurality of categorical features, are extracted from the plurality of claims and based on the multiple fields. The parallel-structured neural network is trained by (1) selecting, based on a feature selection algorithm and a loss function, one or more numerical features and one or more categorical features, and (2) generating, based on the selected features, a categorical layer, a fully-connected layer, and a concatenation layer of the parallel-structured neural network. A claim is received including numerical data and categorical data, and a probability prediction is generated by inputting the claim into the parallel-structured neural network. Finally, a submission including the first claim is generated based on the probability prediction.

Abstract: Disclosed herein are system, method, and computer program product embodiments for creating and utilizing machine learning models to generate a provider performance index. In some embodiments, a first and second target may be selected. The first and second targets may respectfully include target values. First and second adjusted target values may be determined by combining the first and second target values with first and second external weights. The external weights may be generated by respective first and second machine learning models. The machine learning models may correspond to the respective targets. First and second error values may be determined based on differences between respective target and adjusted target values. The error values may be normalized and combined to generate an index value.