doc.ai has seen awesome demand for a new kind of prospective research study called a data trial. Our first data trial was launched this past fall and called “Can Artificial Intelligence Predict Your Risk of Allergies?”
The study gathered data from doc.ai users about their allergy symptoms and triggers. Study participants are asked to document, via the doc.ai app, the severity and time of their symptoms (e.g., high, medium, or low severity).
We’ve analyzed the data from the first 1,441 participants who reported allergy symptoms between September 1 to November 2018, and are learning about how allergy triggers and symptoms vary across the United States, as well as gathering data that will shape future AI efforts.
Study participants were active from the start.
For September 1 through November 2018, here’s an overview of reporting:
How many reports were submitted? There were between 11 and 1674 records submitted on any given day. On average, the number of daily reports submitted was 810.
How many reports per active users per day? There were on average 5.29 records per day per user! On an average day, there were anywhere from 3 to 332 users actively using the app (with at least one allergy report filed).
How many days did participants use the app? Individuals used the app between 1-65 days. An “average” user inputted data on nearly 8 separate days.
Where were participants from? Everywhere! But skewed towards urban areas of the U.S.:
Map of doc.ai allergy data trial participants
Of the ~70k reports submitted to doc.ai, 47% reported the user having an allergy attack (n=33,739 reports). Of these allergy reports, the majority of reports mentioned that the allergy attack came in morning (71% of all reports), followed by evening (45% of all reports) (the reports can report more than one time of the day). Very few came from the night.
Study participants were also asked about what triggered their allergy symptoms for each report. The specific trigger was indicated by app users in over 17,000 reports, and pollen was the most frequent trigger followed by mold and pets.
How do the participants of the doc.ai allergy trial compare to general folks with (and without) allergies across the United States? To learn how doc.ai participants are similar and different to those in the general US population, we compared doc.ai participants to participants from the US Centers for Disease Control and Prevention (CDC) National Health and Nutrition Examination Survey (NHANES) study.
In short, the NHANES is a health survey that is administered by the US CDC whose primary objective is to ascertain prevalence of disease and health of the United States. Participants of NHANES are representative of the United States and give us a glimpse of the variation of health in the US population.
In 2005-2006, NHANES participants were surveyed about their allergies and allergy-related symptoms.
First, allergy sufferers are more likely to be female, both in the doc.ai allergy trial and in the general United States population:
In the doc.ai allergy trial, more women (n=1021, 71%) reported allergies than men (n=420, 29%).
Likewise, in the the US as a whole (in 2005-2006), more women (58%) suffer allergies than men (42%), compared to 50-50 distribution for participants who do not report an allergy.
In the US as a whole (in 2005-2006), women report their allergy symptoms to a doctor much earlier than men (approximately 26 versus 32 years of age).
In the doc.ai allergy trial, the average age of both women and men were just over 40.
doc.ai study participants also reported their symptoms using the app with itching, followed by headache, difficulty breathing, and wheezing being the top symptoms reported.
In summary, we are starting to learn about the triggers and symptoms of allergy sufferers across the United States from this exciting new kind of platform. We will be releasing more in-depth analyses soon as more data become available.
If you’re interested in participating in trials check out doc.ai. If you’d like to learn more about some of the data science techniques we are using to analyze and visualize these data, read more and sign up for data science courses coming online soon!
If you are interested in signing up, please visit this sheet to tell us about you!