No abstract

The Poly Implant Prothèse (PIP) scandal in France prompted a revision of the regulations regarding the marketing of medical devices. The new Medical Device Regulation (MDR [EU]) 2017/745 was developed and entered into force on May 25, 2017. After a transition period of 3 years, the regulations must be implemented in all EU and European Economic Area member states. The implementation of this regulation bears many changes for medical device development and marketing, including medical device software and mobile apps. Medical device development and marketing is a complex process by which manufacturers must keep many regulatory requirements and obligations in mind. The objective of this paper is to provide an introduction and overview of regulatory affairs for manufacturers that are new to the field of medical device software and apps with a specific focus on the new MDR, accompanying harmonized standards, and guidance documents from the European Commission. This work provides a concise overview of the qualification and classification of medical device software and apps, conformity assessment routes, technical documentation, clinical evaluation, the involvement of notified bodies, and the unique device identifier. Compared to the previous Medical Device Directive (MDD) 93/42/EEC, the MDR provides greater detail about the requirements for software qualification and classification. In particular, rule 11 sets specific rules for the classification of medical device software and will be described in this paper. In comparison to the previous MDD, the MDR is more stringent, especially regarding the classification of health apps and software. The implementation of the MDR in May 2020 and its interpretation by the authorities will demonstrate how app and software manufacturers as well as patients will be affected by the regulation.

Background Mobile health (mHealth) is a rapidly emerging market, which has been implemented in a variety of different disease areas. Tuberculosis remains one of the most common causes of death from an infectious disease worldwide, and mHealth apps offer an important contribution to the improvement of tuberculosis treatment. In particular, apps facilitating dose individualization, adherence monitoring, or provision of information and education about the disease can be powerful tools to prevent the development of drug-resistant tuberculosis or disease relapse. Objective The aim of this review was to identify, describe, and categorize mobile and Web-based apps related to tuberculosis that are currently available. Methods PubMed, Google Play Store, Apple Store, Amazon, and Google were searched between February and July 2019 using a combination of 20 keywords. Apps were included in the analysis if they focused on tuberculosis, and were excluded if they were related to other disease areas or if they were games unrelated to tuberculosis. All apps matching the inclusion criteria were classified into the following five categories: adherence monitoring, individualized dosing, eLearning/information, diagnosis, and others. The included apps were then summarized and described based on publicly available information using 12 characteristics. Results Fifty-five mHealth apps met the inclusion criteria and were included in this analysis. Of the 55 apps, 8 (15%) were intended to monitor patients’ adherence, 6 (11%) were designed for dosage adjustment, 29 (53%) were designed for eLearning/information, 3 (6%) were focused on tuberculosis diagnosis, and 9 (16%) were related to other purposes. Conclusions The number of mHealth apps related to tuberculosis has increased during the past 3 years. Although some of the discovered apps seem promising, many were found to contain errors or provided harmful or wrong information. Moreover, the majority of mHealth apps currently on the market are focused on making information about tuberculosis available (29/55, 53%). Thus, this review highlights a need for new, high-quality mHealth apps supporting tuberculosis treatment, especially those supporting individualized optimized treatment through model-informed precision dosing and video observed treatment.

No abstract

Pharmacometrics (PM) and machine learning (ML) are both valuable for drug development to characterize pharmacokinetics (PK) and pharmacodynamics (PD). Pharmacokinetic/pharmacodynamic (PKPD) analysis using PM provides mechanistic insight into biological processes but is time- and labor-intensive. In contrast, ML models are much quicker trained, but offer less mechanistic insights. The opportunity of using ML predictions of drug PK as input for a PKPD model could strongly accelerate analysis efforts. Here exemplified by rifampicin, a widely used antibiotic, we explore the ability of different ML algorithms to predict drug PK. Based on simulated data, we trained linear regressions (LASSO), Gradient Boosting Machines, XGBoost and Random Forest to predict the plasma concentration-time series and rifampicin area under the concentration-versus-time curve from 0–24 h (AUC0–24h) after repeated dosing. XGBoost performed best for prediction of the entire PK series (R2: 0.84, root mean square error (RMSE): 6.9 mg/L, mean absolute error (MAE): 4.0 mg/L) for the scenario with the largest data size. For AUC0–24h prediction, LASSO showed the highest performance (R2: 0.97, RMSE: 29.1 h·mg/L, MAE: 18.8 h·mg/L). Increasing the number of plasma concentrations per patient (0, 2 or 6 concentrations per occasion) improved model performance. For example, for AUC0–24h prediction using LASSO, the R2 was 0.41, 0.69 and 0.97 when using predictors only (no plasma concentrations), 2 or 6 plasma concentrations per occasion as input, respectively. Run times for the ML models ranged from 1.0 s to 8 min, while the run time for the PM model was more than 3 h. Furthermore, building a PM model is more time- and labor-intensive compared with ML. ML predictions of drug PK could thus be used as input into a PKPD model, enabling time-efficient analysis.