Objectives: In the growing market of health apps (>259,000 on major stores) finding the right app for a specific need is challenging and can be a barrier to adoption. The objective of our research is to develop a framework for automated extraction from the web of meaningful information on the apps related to Cardiology. Methods. Custom software to automatically browse the iTunes app store webpages, analyse the HTML code, retrieve apps' attributes, and detect the language in the app description was developed and applied to the Medical (M) and Health & Fitness (H&F) categories of the US store (as of Apr 25, 2017). To validate in future automated text analytics methods to characterize the apps' features, 250 apps were randomly selected as a training set and manually analysed to extract the information relevant to: (i) medical specialty, (ii) promoter, and (iii) target users. In addition, we focused on a sample of apps for disease prevention, diagnosis, treatment, and lifestyle management in Cardiology. Results: In total, 42046 M and 79851 H&F apps' webpages were browsed, removing 27% of M apps and 23% of H&F apps as not described in English, resulting in 80490 unique apps: 49917 (62%) in the H&F category, 19383 (24%) in the M category, and 11190 (14%) in both categories. The same distribution was reflected in the selected training set. Based on manual analysis of apps in the training set, 42/250 apps (16.8%) were not relevant to health (e.g., entertainment, learning, business apps). Of the remaining 208 apps, 41.3% were for Fitness and Wellness, 23.6% were not related to any specialty (e.g., lab tests, pharmacy, telemedicine), 4.8% were for Nutrition, 4.3% for Cardiology and for Sensory Systems Healthcare (e.g., Hearing, Vision, Balance), 2.9% for Emergency Medicine, for Diabetes care, and for Gynaecology and Obstetrics, and less than 2% for other specialties, e.g. Dentistry, Dermatology, Gastroenterology, Oncology. A total of 135/250 apps (i.e., 53.6%) related to Cardiology for disease prevention, diagnosis, treatment, and lifestyle management (e.g., physical activity, smoking cessation, or risk factors tracking) were found. About 24% were promoted by individual developers, 21% by Healthcare providers, 20% by companies (device, drug, or software manufacturers), 14% by fitness providers, 12% by scientific or educational institutions, 5% by publishers, 3% by governmental Institutions, and 1% by patients Associations. The target users were: citizens (57%), physicians & Healthcare professionals (21%), patients (21%), and families and significant others (1%). Conclusions: This preliminary study revealed the distribution of specialties, promoters, and target users in a random sample of health apps directly or indirectly related to Cardiology. These results will be the basis for future validation of automated methods to characterize apps as a means to build novel tools to support patients and healthcare professionals in informed, aware adoption of apps.
Development of a framework for the automated characterization of apps for disease prevention, diagnosis, treatment, and lifestyle management in Cardiology from app store webpages
Paglialonga A;
2017
Abstract
Objectives: In the growing market of health apps (>259,000 on major stores) finding the right app for a specific need is challenging and can be a barrier to adoption. The objective of our research is to develop a framework for automated extraction from the web of meaningful information on the apps related to Cardiology. Methods. Custom software to automatically browse the iTunes app store webpages, analyse the HTML code, retrieve apps' attributes, and detect the language in the app description was developed and applied to the Medical (M) and Health & Fitness (H&F) categories of the US store (as of Apr 25, 2017). To validate in future automated text analytics methods to characterize the apps' features, 250 apps were randomly selected as a training set and manually analysed to extract the information relevant to: (i) medical specialty, (ii) promoter, and (iii) target users. In addition, we focused on a sample of apps for disease prevention, diagnosis, treatment, and lifestyle management in Cardiology. Results: In total, 42046 M and 79851 H&F apps' webpages were browsed, removing 27% of M apps and 23% of H&F apps as not described in English, resulting in 80490 unique apps: 49917 (62%) in the H&F category, 19383 (24%) in the M category, and 11190 (14%) in both categories. The same distribution was reflected in the selected training set. Based on manual analysis of apps in the training set, 42/250 apps (16.8%) were not relevant to health (e.g., entertainment, learning, business apps). Of the remaining 208 apps, 41.3% were for Fitness and Wellness, 23.6% were not related to any specialty (e.g., lab tests, pharmacy, telemedicine), 4.8% were for Nutrition, 4.3% for Cardiology and for Sensory Systems Healthcare (e.g., Hearing, Vision, Balance), 2.9% for Emergency Medicine, for Diabetes care, and for Gynaecology and Obstetrics, and less than 2% for other specialties, e.g. Dentistry, Dermatology, Gastroenterology, Oncology. A total of 135/250 apps (i.e., 53.6%) related to Cardiology for disease prevention, diagnosis, treatment, and lifestyle management (e.g., physical activity, smoking cessation, or risk factors tracking) were found. About 24% were promoted by individual developers, 21% by Healthcare providers, 20% by companies (device, drug, or software manufacturers), 14% by fitness providers, 12% by scientific or educational institutions, 5% by publishers, 3% by governmental Institutions, and 1% by patients Associations. The target users were: citizens (57%), physicians & Healthcare professionals (21%), patients (21%), and families and significant others (1%). Conclusions: This preliminary study revealed the distribution of specialties, promoters, and target users in a random sample of health apps directly or indirectly related to Cardiology. These results will be the basis for future validation of automated methods to characterize apps as a means to build novel tools to support patients and healthcare professionals in informed, aware adoption of apps.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
