A treatment recommender clinical decision support system for personalized medicine: method development and proof-of-concept for drug resistant tuberculosis

Abstract

Background Personalized medicine tailors care based on the patient’s or pathogen’s genotypic and phenotypic characteristics. An automated Clinical Decision Support System (CDSS) could help translate the genotypic and phenotypic characteristics into optimal treatment and thus facilitate implementation of individualized treatment by less experienced physicians. Methods We developed a hybrid knowledge- and data-driven treatment recommender CDSS. Stakeholders and experts first define the knowledge base by identifying and quantifying drug and regimen features for the prototype model input. In an iterative manner, feedback from experts is harvested to generate model training datasets, machine learning methods are applied to identify complex relations and patterns in the data, and model performance is assessed by estimating the precision at one, mean reciprocal rank and mean average precision. Once the model performance no longer iteratively increases, a validation dataset is used to assess model overfitting. Results We applied the novel methodology to develop a treatment recommender CDSS for individualized treatment of drug resistant tuberculosis as a proof of concept. Using input from stakeholders and three rounds of expert feedback on a dataset of 355 patients with 129 unique drug resistance profiles, the model had a 95% precision at 1 indicating that the highest ranked treatment regimen was considered appropriate by the experts in 95% of cases. Use of a validation data set however suggested substantial model overfitting, with a reduction in precision at 1 to 78%. Conclusion Our novel and flexible hybrid knowledge- and data-driven treatment recommender CDSS is a first step towards the automation of individualized treatment for personalized medicine. Further research should assess its value in fields other than drug resistant tuberculosis, develop solid statistical approaches to assess model performance, and evaluate their accuracy in real-life clinical settings.

Sequencing Mycobacteria and Algorithm-determined Resistant Tuberculosis Treatment (SMARTT): a study protocol for a phase IV pragmatic randomized controlled patient management strategy trial

“…Next, the drug resistance profile of a strain has to be translated into an individualized regimen. Using data mining and machine learning methods, computational models can be generated to automatically, accurately and efficiently translate WGS Mtb data into individualized RR-TB treatment regimens 34 . These recommender models are highly flexible and can easily integrate new drugs or new knowledge on genotypic-genotypic associations as they become available.…”

“…The intervention consists of the use of Mtb WGS for determining the drug-resistance profile and automated individualized treatment recommendation of the optimal 4-drug regimen 34 , with communication of the recommended regimen to the health care worker via a mobile phone app (Figure 1).…”

“…Regimens are scored based on the quantitative drug and regimen features including bactericidal and sterilizing activity, propensity to acquire resistance, toxicity, mechanism of action, route of administration, antagonistic and synergistic interactions, and cost. After taking into account clinical data (presence of contraindications, drug toxicity) and health systems data (drug stock out), the highest ranked 4-drug individualized treatment regimen is recommended for six months 34 . When healthcare workers question the WGS-guided treatment recommendation, a central clinical committee (CCC) of international experts in RR-TB treatment and clinical microbiology provides evidence-based guidance to the care provider.…”

Bedaquiline resistance probability to guide treatment decision making for rifampicin-resistant tuberculosis - insights from a qualitative study

“…Physicians found it challenging to adopt the BDQ resistance probability into their clinical decision-making because of their limited experience with DST for BDQ, unfamiliarity with gDST and poor understanding of WGS results. Physicians recommended that the use of BDQ resistance probability information in routine care should be accompanied by a clear interpretation guide and a clinician-friendly decision support system, such as a computerized treatment recommender recently developed for RR-TB (52). Regarding patient characteristics, the clinical and microbiological response to treatment, resistance pro le and BDQ exposure history emerged as factors that strongly in uence the interpretation of BDQ resistance probability value.…”

Bedaquiline resistance probability to guide treatment decision making for rifampicin-resistant tuberculosis: insights from a qualitative study

“…The increasing knowledge of the molecular basis of resistance in the coming years should improve the clinical usefulness of the BDQ resistance probability value. Once ready to be implemented in routine care, the BDQ resistance probability information should be accompanied by a clear interpretation guide and a clinician-friendly decision support system, such as a computerized treatment recommender recently developed for RR-TB [ 55 ]. Regarding patient characteristics, the clinical and microbiological response to treatment, resistance profile and BDQ exposure history emerged as factors that strongly influence the interpretation of BDQ resistance probability value.…”