Phenotyping and Privacy
What is Digital Phenotyping?
Digital Phenotyping is an advancing multidisciplinary field of science. It was described by Jukka-Pekka Onnela, an Associate Professor of Biostatistics at the Harvard T.H. Chan School of Public Health and Director of the Health Data Science Program in Finland, as the “moment-by-moment quantification of the individual-level human phenotype in situ using data from personal digital devices.”
A phenotype is a generic term for the composite observable characteristics or traits of an organism. Digital phenotyping is the study of or capability to determine medical, mental health, or other personal attributes about an individual based on their use of technology. In other words, the way that you use social media or your cell phone can give details about your state of mind, personal habits, and accurately pinpoint your personality.
Data collected can be divided into two categories: active data and passive data. Dynamic data is the type of data that requires a response from the user or input to be generated. Passive data, which includes sensor data and phone usage patterns, requires no user’s active participation.
Due to their widespread availability, smartphones are an excellent source for digital phenotyping. Data can be compiled regarding the extent that the users are engaged with their device. Smartphones can study behavioral patterns, social interactions, physical mobility, gross motor activity, and speech production. Smartphones are a rich source of data that can give personality info, health info, mental health status, GPS locations. The phones collect some of the most personal data about you. It has been noted that passive data collection from smartphones generates foundational information that is relevant to determining psychiatric and other phenotypes. This data type can monitor distances traveled, locations, movement, gross motor activity, calls, and texts sent or posted on social media.
Digital Phenotyping Applications
Here is a listing of some typical applications that use digital phenotyping – some are already in use:
• Facebook – Added an algorithm that can determine from posts if users are exhibiting signs of suicidal thoughts. An alert is then sent to a Facebook review team.
• Mindstrong Health – Is a California-based startup that developed a research app to monitor users’ phone habits, keyboard accuracy, and speed. The data demonstrates the typical usage patterns of the user. It can note changes, including memory changes, in the use patterns to pinpoint depression or anxiety. The company shows biomarkers built into human interactions with smartphones that correlate with cognitive control and reward systems.
Transparency or Violation?
Paul Dagum, CEO of Mindstrong Health, stated that he believes that this data can be used at an aggregate public health level. He can use this data to create a ‘heat map’ of the earth and show areas impacted by high emotional volatility rates. Maps can also be generated to show a cognitive decline, stress levels, contagions, or even toxins.
Jukka-Pekka Onnela explained how a simple text message could give details about a person’s mental state or health. “It’s like a micro-cognitive assessment. You have to have executive function, memory, linguistic function — it’s these little things that turn out to be incredibly informative about a person’s state.”
As individuals, we leave trails of our data on our phones and our computers. The data sold to third parties left open on the internet or exposed through our metadata can give clues about our personal lives. Not knowing who or what 3rd-party data crunchers are doing with your data – does it feel like your life is an open book showing transparency, or do you feel like your civil rights are being violated? It is a question that will continue to face society as the number of uses for our technology and data proliferates.
Onnela continued to explain that, in the future, digital phenotyping will become a great help to medical doctors as they use it to diagnose mental illness, depression, or anxiety. One of the most significant advancements and results for health data from digital phenotyping is diagnosing Parkinson’s Disease. Scientists are now capable of analyzing how people use their phones and predict Parkinson’s with 100 percent accuracy.
Other forms of technology are also showing promising results in predicting and diagnosing disease. The accelerometer data from cell phones and smartwatches, the technology that measures your steps or distance traveled or your movement through space, can give even more detailed information regarding Parkinson’s. By reviewing the data provided by your device and how it moves, doctors can estimate the severity of Parkinson’s tremors. The ability to monitor and measure tremors can help both the patient and doctor understand how the disease progresses or becomes more severe.
The Future of Phenotyping
New uses for phenotyping are being found every day. Digital phenotyping, through the use of algorithms, seeks to utilize the potential of data that is automatically measured by smartphones, wearables, or other connected devices. They can be used to measure human behavior and health problems.
Today’s data stream includes sensor measurements, logs our activities, and reviews our user-generated content. Studies show that future use of digital phenotyping anticipates becoming mainstream in routine clinical practice. It will have the capability of strengthening the clarity and accuracy of clinical diagnosis and treatments. Digital phenotyping may lead to earlier detection of disease, report relapses, and measure treatment responses.
Digital phenotyping will push forward ‘sensing’ technology for medical and mental health. It will be done through the commercialization of various products and the general media. The user must consider the impact of how their data will be used after they receive the results.
The future will allow medical professionals to use digital traces to identify the characteristics of their patients. It will create new fields in medicine involved with ‘sensing’ or projecting results. It will be used to incorporate sensing into genetics, diagnosis, and prognosis.
Having so many benefits is that digital phenotyping does not explain to the average individual that they will be under surveillance. This point is overlooked or ignored. Privacy legislation will be necessary when data collection becomes societal surveillance.
New technology comes at us faster than legislation can keep up. It may be compared to the changes that occurred in the US Federal government after 9/11. Laws provide us with safety and security, so Americans were willing to allow the Patriot Act to strip away some fundamental civil rights. In the name of the positive expected result: safety. Technology that helps collect data for digital phenotyping may result in similar outcomes. Individuals want good health. The benefits of gaining better health and feeling better outweigh the intrusion of surveillance in their lives. Perhaps we should think about how the data will be used after diagnosis. Who will it be shared with or sold to for further use?
Privacy and Future Legislation
We should stop and think about how digital phenotyping could be an invasion of privacy. Health data could be gained without consent or knowledge of the user and then sold to third parties. This could include banks, employers, and insurance companies. Discrimination could be pushed against those whose phone data points to diabetes or depression. Individuals can be denied jobs or mortgages. Knowing who accesses our data and how it is used is becoming vital.
According to Giovanni Stanghelini, a Center for Studies on Phenomenology and Psychiatry Medical Faculty, and Federico Leoni, Professor in the Department of Human Services, Verona University, we will have to learn to stay on top of our health data. “We should learn how to take advantage of the instruments that contemporary technology provides for us, looking at the phenomena they show us and the possibilities of intervention that they open up. On the other hand, we should also learn to look at the instruments themselves, without being dazzled by the phenomena they seem to apply. When we worry that big data involves a privacy issue, it’s already too late—even though we should worry about privacy. The real problem is not that we have to manage certain data about our bodies properly.”
We need to learn to manage data. Effectively and securely keep track of our personal data, who has access, and for what purposes. When it comes to health data, beware – it may come with other costs; once paid, you can’t get a refund.