Use of machine learning to identify patients at risk of sub-optimal adherence: study based on real-world data from 10,929 children using a connected auto-injector device

Amalia Spataru,Paula van Dommelen,Lilian Arnaud,Quentin Le Masne,Silvia Quarteroni,Ekaterina Koledova
DOI: https://doi.org/10.1186/s12911-022-01918-2
2022-07-06
Abstract:Background: Our aim was to develop a machine learning model, using real-world data captured from a connected auto-injector device and from early indicators from the first 3 months of treatment, to predict sub-optimal adherence to recombinant human growth hormone (r-hGH) in patients with growth disorders. Methods: Adherence to r-hGH treatment was assessed in children (aged < 18 years) who started using a connected auto-injector device (easypod™), and transmitted injection data for ≥ 12 months. Adherence in the following 3, 6, or 9 months after treatment start was categorized as optimal (≥ 85%) versus sub-optimal (< 85%). Logistic regression and tree-based models were applied. Results: Data from 10,929 children showed that a random forest model with mean and standard deviation of adherence over the first 3 months, infrequent transmission of data, not changing certain comfort settings, and starting treatment at an older age was important in predicting the risk of sub-optimal adherence in the following 3, 6, or 9 months. Sensitivities ranged between 0.72 and 0.77, and specificities between 0.80 and 0.81. Conclusions: To the authors' knowledge, this is the first attempt to integrate a machine learning model into a digital health ecosystem to help healthcare providers to identify patients at risk of sub-optimal adherence to r-hGH in the following 3, 6, or 9 months. This information, together with patient-specific indicators of sub-optimal adherence, can be used to provide support to at-risk patients and their caregivers to achieve optimal adherence and, subsequently, improve clinical outcomes.
What problem does this paper attempt to address?