Classification of patients with chronic disease by activation level using machine learning methods

Patient Activation Measure (PAM) measures the activation level of patients with chronic conditions and correlates well with patient adherence behavior, health outcomes, and healthcare costs. PAM is increasingly used in practice to identify patients needing more support from the care team. We define PAM levels 1 and 2 as low PAM and investigate the performance of eight machine learning methods (Logistic Regression, Lasso Regression, Ridge Regression, Random Forest, Gradient Boosted Trees, Support Vector Machines, Decision Trees, Neural Networks) to classify patients. Primary data collected from adult patients (n=431) with Diabetes Mellitus (DM) or Hypertension (HT) attending Family Health Centers in Istanbul, Turkey, is used to test the methods. [Formula: see text] of patients in the dataset have a low PAM level. Classification performance with several feature sets was analyzed to understand the relative importance of different types of information and provide insights. The most important features are found as whether the patient performs self-monitoring, smoking and exercise habits, education, and socio-economic status. The best performance was achieved with the Logistic Regression algorithm, with Area Under the Curve (AUC)=0.72 with the best performing feature set. Alternative feature sets with similar prediction performance are also presented. The prediction performance was inferior with an automated feature selection method, supporting the importance of using domain knowledge in machine learning.

Design and development of a smartphone app for hypertension management: An intervention mapping approach

Background

Several research studies have demonstrated the potential of mobile health apps in supporting health management. However, the design and development process of these apps are rarely presented.

Objective

We present the design and development of a smartphone-based lifestyle app integrating a wearable device for hypertension management.

Methods

We used an intervention mapping approach for the development of theory- and evidence-based intervention in hypertension management. This consisted of six fundamental steps: needs assessment, matrices, theoretical methods and practical strategies, program design, adoption and implementation plan, and evaluation plan. To design the contents of the intervention, we performed a literature review to determine the preferences of people with hypertension (Step 1) and necessary objectives toward the promotion of self-management behaviors (Step 2). Based on these findings, we implemented theoretical and practical strategies in consultation with stakeholders and researchers (Steps 3), which was used to identify the functionality and develop an mHealth app (Step 4). The adoption (Step 5) and evaluation (Step 6) of the mHealth app will be conducted in a future study.

Results

Through the needs analysis, we identified that people with hypertension preferred having education, medication or treatment adherence, lifestyle modification, alcohol and smoking cessation and blood pressure monitoring support. We utilized MoSCoW analysis to consider four key elements, i.e., education, medication or treatment adherence, lifestyle modification and blood pressure support based on past experiences, and its potential benefits in hypertension management. Theoretical models such as (i) the information, motivation, and behavior skills model, and (ii) the patient health engagement model was implemented in the intervention development to ensure positive engagement and health behavior. Our app provides health education to people with hypertension related to their condition, while utilizing wearable devices to promote lifestyle modification and blood pressure management. The app also contains a clinician portal with rules and medication lists titrated by the clinician to ensure treatment adherence, with regular push notifications to prompt behavioral change. In addition, the app data can be reviewed by patients and clinicians as needed.

Conclusions

This is the first study describing the design and development of an app that integrates a wearable blood pressure device and provides lifestyle support and hypertension management. Our theory-driven intervention for hypertension management is founded on the critical needs of people with hypertension to ensure treatment adherence and supports medication review and titration by clinicians. The intervention will be clinically evaluated in future studies to determine its effectiveness and usability.