A survey conducted by Paul Krebs and Dustin Duncan (Krebs & Duncan, 2015), shows that just over half of the mobile phone users (934/1604, 58.23%) had downloaded a health-related mobile app. North Americans are increasingly relying on mobile technology and the internet for health-related information and resources (Fox & Duggan, 2012). The proliferation of smartphone ownership among US adults, particularly among traditionally underserved populations (e.g., low-income, racial/ethnic minorities), has expanded the potential reach of healthy eating, physical activity, and weight loss programs. From 2011 to 2015, the percentage of US adults owning a smartphone increased from 35% to 68% (Anderson, 2015). In addition, low income and racial/ethnic minority populations are more likely to be smartphone-dependent, thus relying primarily on their phones for health information (Smith, 2015). Making these apps “smarter” through AI can enhance their accuracy, effectiveness, and efficiency increasing users' trust while allowing input from medical experts, which in turn increases the utility of these applications.
What it does
CANBeWell organizes information on proactive wellness topics and recommended screening tests by age and gender in an easy to use interface that includes an annotated anatomical image where the user can click on different parts of the body. Users log in as either a provider or a patient and by clicking on different parts of the annotated anatomical body image, text and important links with supporting evidence and supplementary information are displayed based on the context of the use provider or patient. Information is presented to the users visually through the Body tab or by alphabetically sorted topics if they click on the Topic tab. The top left corner displays the current configuration.
How I built it
CANBeWell was built upon a React. Our rule-based content management application leverages declarative rules developed by medical experts with information on the common health challenges that affect various organs of the body for different ages and gender. rules are created and managed by medical experts based on research and findings from the literature. These rules are then converted to a machine-readable JSON file and pushed to a mobile app that is made available to users using very interactive, mobile-friendly user interfaces. Rule-based contents are then used to dynamically display the correct content based on the language of the user, gender and age of the patient, and whether they want information appropriate for a patient or a provider. Since most medical experts are familiar with using spreadsheets, we decided to use an Excel spreadsheet to organize and classify information for these experts. Figure 3 shows a snipped of the Excel spreadsheet that was used with our rule-based application. Each of the columns corresponds to the classification of content. Each row is a rule that says what content to display, depending on the value of gender and the age range. he content from the Excel spreadsheet from the experts is then translated into different languages as needed to support the targeted users. These files are then converted to machine-readable JSON files. The choice of a JSON file is because of its flexibility with various web and mobile application development platforms. Regression tests are run on the app to ensure that it is functioning properly. And finally, the medical expert manually tests the app to be sure the content specified is displaying correctly. To support French and English users, there are four content files for the application.
Accomplishments that I'm proud of
Enumerated below are some of the strong features of the application. ▪ Bilingual support (currently supports English and French).
▪ Optimized for smartphones but runs on any device (laptop, tablet, smart TV, cell phones) with a browser.
▪ Body image interface.
▪ Separately and optimally worded text for both patients and providers.
▪ Easy Filtering and Navigation (body parts, age, gender, health topics, common tests, patient vs provider text)
▪ Use of embedded images and supporting links in the content.
▪ Additionally, the current integration of the app has support for content management by the medical expert.
Log in or sign up for Devpost to join the conversation.