How Apple is lowering the bar for medical research through software development

[[132519]]

One of the biggest difficulties in medical research is obtaining enough samples for clinical trials. I have some experience with this: when I was studying, the medical school of my university sent me emails every week to recruit research volunteers, and most of them were thrown directly into the trash.

Most people have reservations about being a volunteer for medical research. In addition to the discomfort of being a guinea pig and the distrust of hospitals, being a research volunteer is not an easy commitment. Although sometimes you can get a good salary, the condition is that you must cooperate with the doctor's work throughout the process. Many studies also require long-term recording of physical indicators and regular follow-up visits. Therefore, for medical research institutions, recruiting research volunteers is the biggest problem in history. Kathryn Schmitz, a doctoral student at Penn Medicine, mentioned to Apple that she sent out 60,000 emails to recruit research volunteers and received only 305 responses.

Apple realizes that smartphones, which are in everyone's hands, are a powerful tool for collecting data for clinical research. Users are accustomed to allowing smartphones to collect personal information from fingerprints to locations, and the current smartphone hardware has long supported the development of more complex software. The problem is that medical researchers are not good at making user-oriented software, and they don't have the time and money to independently develop an app and design the UI for each function.

Apple is trying to lower the threshold for recruiting volunteers for medical research by developing software. Last week, Apple officially announced the developer framework of ResearchKit, which was not discussed much at the March press conference. Unlike HealthKit, which is aimed at users to monitor personal health information, ResearchKit targets developers and aims to innovate the sample collection methods for clinical medical research through standardized development processes and usage methods.

ResearchKit's developer framework includes three customizable modules: Participant Consent, Survey, and Active Task. It complies with Mac/iOS programming standards, but is also an open source project on GitHub. The biggest benefit for medical researchers is that it provides a standardized paradigm and a controllable GUI (user interface), which simplifies the development process.

ActiveTask is the basic unit in ResearchKit development and follows a protocol called ORKTask. Apple presets some basic active tasks based on the hardware functions of the iPhone and lists the sensors they call and the data they collect. For example, gait and balance, click speed test, health, spatial memory test, etc. For example, after adding the gait and balance active task, the application can call the acceleration sensor and gyroscope and collect data from the device's pedometer. Developers can also add their own active tasks.

Activities consist of a series of objects arranged in an ordered sequence: "steps". Steps range from simple "yes" answers, such as "Did you read and understand the above?" to actions such as displaying a table, collecting information from audio or memory tests.

Therefore, developers can set up a series of steps for collecting key data like building blocks, and combine them into active tasks to form the main functions of the app. This is much easier than developing ordinary iOS apps. To further simplify the operation, ResearchKit also provides many preset options, such as adding the simplest "yes" answer to the step, selecting answers and values, or entering dates and text.

In addition, there is a view controller (ORKStepViewController) between each step and the activity task to generate interaction. The ArsTechnica article mentioned that, generally speaking, developers need to write the interaction details in Xcode. With the ResearchKit framework, you only need to set the step object (ORKQuestionStep) and add it to the activity task. After the setting is completed, the interaction interface will be automatically generated by the underlying framework. This also further reduces the development threshold.

ResearchKit does not provide features such as background sensor data collection, which means that the app cannot continuously collect data, but can capture data through HealthKit and CoreMotion API. ResearchKit also does not support preset surveys or activity tasks, so apps based on ResearchKit cannot send scheduled reminders to users. In addition, for the purpose of user privacy protection, Apple will not store data on its own servers, so researchers must ensure that the data is securely transmitted back to the research institution.

ResearchKit not only simplifies the first step of research, but also simplifies the steps for users to authorize participation in medical research. Users only need to read the statement and sign on the phone screen to participate in the research. Corey Brideges, CEO of LifeMap Solution, which cooperated with Mount Sinai and Weill Cornell Medical College to develop an asthma application, believes that the most revolutionary part of ResearchKit is that it provides a standard authorization agreement, which greatly reduces the time and economic cost of users participating in medical research, and increases the scope and number of samples.

Bridege said in an interview with ArsTechnica that generally, the number of participants in clinical studies is limited by geographic location, obtaining permission, and the cost of including participants. The cost of clicking on the authorization permission on the app is very small, and users do not need to go to the hospital for follow-up visits. They can provide research data at any time by carrying a smartphone.

Bridge said that in the first week of ResearchKit's release, the app he developed with researchers at Mt. Sinai Hospital had received licenses from 4,500 participants. He said:

“E-consent is groundbreaking because it can help medical researchers break through a major barrier to clinical research. It can reach iPhone-wielding Internet users around the world, increase the number of volunteers participating in studies, and significantly expand sample sizes, all while reducing costs.”

Bridge also believes that "Compared with traditional research, ResearchKit improves the efficiency of the entire data collection and research process and can collect more comprehensive data." Environmental data is a good example. For example, an app that monitors asthma can help establish the connection between the environment and symptoms in medical research through comprehensive indicators such as patient symptoms, geographic location, and air quality. This is accurate information that is difficult to obtain in current clinical data collection.