Investigating the Quality of Mobile Apps for Drug-Drug Interaction (DDI) Management using Mobile App Rating Scale (MARS) and K-Means Clustering : Systematic Search in App Stores Analysis (Preprint)

Abstract

Drug-Drug Interactions (DDIs) are a serious issue that can compromise patient care and increase healthcare costs. Mobile apps offer potential solutions for managing DDIs, yet their quality and effectiveness from the user’s perspective remains unclear.

To evaluate the quality of publicly available mobile apps for DDI management in the US using the Mobile App Rating Scale (MARS) and identify patterns in user preferences.

A review was conducted to identify available mobile apps for DDI management, resulting in the discovery of 19 apps. The apps were evaluated independently by two evaluators using the MARS scale. MARS dimensionality scores were computed, and a correlation study was conducted to understand the interrelation of dimensions. K-Means Clustering was used to classify apps in clusters based on the MARS scores. Scatter plots were created to visualise the distribution of apps across different dimensions by their respective clusters. To validate the clustering model and assess the alignment between MARS evaluations and user satisfaction, a comparison of mean weighted app ratings with mean MARS scores by cluster was conducted. Additionally, further correlation analysis was carried out to examine how MARS dimensions influenced app ratings within each cluster, providing deeper insights into the factors driving user satisfaction.

The mean MARS score was 3.54 out of 5, with the information dimension scoring the highest and engagement the lowest. Positive correlations across all dimensions suggest they are interrelated, highlighting the importance of developing well-rounded apps. K-Means clustering identified three distinct app clusters, with Cluster 3 having the highest average MARS scores and Cluster 1 the lowest. Scatter plot analysis emphasized that engagement, functionality, and aesthetics are the primary drivers of user perceptions, while information plays a lesser role in differentiating apps. The strong positive correlation between mean weighted app ratings and mean MARS scores across clusters validated our K-means model, demonstrating its effectiveness in distinguishing apps. However, statistical testing revealed no significant difference between MARS scores and weighted user ratings in Clusters 1 and 3, while Cluster 2 exhibited a significant difference. Further correlation analysis showed that in Cluster 1, functionality and engagement were the primary drivers of user satisfaction, while in Cluster 2, information quality was more critical, with aesthetics and engagement playing secondary roles. Cluster 3 showed balanced importance across dimensions, favoring well-rounded, informative, and visually appealing apps.

This study assesses the quality of mobile apps for Drug-Drug Interaction (DDI) management by integrating the Mobile App Rating Scale (MARS) with K-Means Clustering, offering a novel approach to app evaluation. Through K-Means Clustering, we achieved a structured classification of apps based on MARS scores, identifying distinct clusters that reflect overall app quality. The study revealed that the influence of MARS dimensions on app ratings varies by cluster, highlighting that the significance of these dimensions shifts according to the specific needs and preferences of different user groups.

N/A