BIG Data & Healthcare

Authors: Ozan Kaya, Inti Mendoza Estrada, Ana Paz Hernandez, Rodrigo Saavedra Sotelo

Can we learn from the past to become better in the future?

Health Data

The advancements in medicinal technologies, particularly those that involve the sequencing of genomic information to detect possible health threats as well as monitoring activities regarding health information, have recently led to a surging necessity of being able to store, analyze, and access data through an electronic source that is straightforward and fast to utilize. Therefore,  in healthcare there are three possible sources of information that are integrated into the scope of health data these being: genomic data, clinical data, and behavioral data.

Clinical Data

The most generated type of data is generated already by physicians when a patient is given treatment. In recent years this data had been stored at paper files, however, nowadays this information has moved towards electronic services known as Electronic Health Records (EHR). It is a challenge however, to keep clinical data unified in a single type of data source such that common problems like grammatical errors, misspellings, etc., can be avoided and the rich amount of information can be correctly stored.

In this regard some of the challenges that healthcare systems face in data collection are:

• being able to infer complex knowledge from complex heterogeneous patients
• understanding unstructured clinical notes
• efficiently handling large amounts of medical images
• analysing genomic information
• capturing patient behavioural data through sensors

Overcoming these challenges are the goals of implementing data storage to healthcare so that there can be more personalized and efficient ways of treating patients.

Behavioral Data

Whilst clinical data can be gathered at a physician’s center, it is another story to gather information regarding a patient’s health status once they reach their homes. For this reason several sensors have become available to the public that tackle the uncertainty of not having a doctor around. Some examples of these sensors have been developed by Under Armour , the University of California San Diego and the University of Illinois.

 Under Armour is one of the first manufacturers/retailers who create and sell such technologies. The Under Armor UA HealthBox, for example,  is a device capable of measuring sleep, resting heart rate, and the body’s physical status. It can also track body weight, body fat percentage, and overall progress of weight loss as well as track heart rate. All of this information can then be stored in real time in an online database, allowing for easy access of one’s health status. 

A flexible skin patch developed by the USCD can correctly detect alcohol levels of the user through a sweat inducing property released when the device is stuck to the wearer. The levels of alcohol are then synced to a device that tracks this information.  Such technologies may be implemented to monitor alcohol levels in drivers.

Most surprising however, are the temporary sensor-tattoos, called Biostamps, that were developed by John Rogers and his team from the University of Illinois.  Rogers argues that whilst advancements in tracking devices have become essential for athletes and some patients, the data generated by these devices are not enough to determine whether a patient will become ill or not. With the Biostamp developed by his team, it would be possible to detect slight body changes that could warn from a disease.

With these advancements in technology, however, it becomes necessary to reflect on the possible negative outcomes of utilizing these devices. One possibility could be a psychological inclination to constantly look at the device as well as over-worry over one’s health status. Another negative aspect of these trackers are the possible confidentiality issues. When a patient agrees to have their health data stored in a centralized data base there has to be a consensus about how this information will be used, as well as who will be able to access it. Will it only be themselves? or will this information be shared with other organizations such as targeted advertisements or insurance profiling? Having this information shared with insurance companies could in some way affect if a patient can or will be covered for certain medical expense.

Nevertheless, being able to track health information throughout the day such that major illnesses could be prevented is, a major advantage. Furthermore, these are not the only advancements that have been made in the sensor smart wear. Several other start ups have begun appearing with innovative ideas that will also help collect, and analyze data.

Genomic Data

With the finalization of the human genome project in 2003 and the advancements made since then, the cost of DNA sequencing has dramatically fallen. Being able to sequence one’s genome can provide useful information about not only heritage, but also possible propensities towards certain diseases through genotypes. 

According to the National Human Genome Research Institute there is a comprehensive list of genetic diseases that are currently being studied. These genetic diseases include for example Breast Cancer, Down Syndrome, Duchenne Muscular Dystrophy, etc. Through genomic sequencing a robust set of information can be generated about a particular patient.

Once again the recollection of this type of information could be an issue for ethical involvement when it comes to how this information will be used. A certain degree of discrimination could be possible from health insurance companies for example if they knew about a patient’s propensity.

The Treatment of Patients

Ultimately, the goal of any aspect of public health should be the improvement of patient care, be it palliative or rehabilitative. Although it has already been exposed that the gathering of data concerning a particular patient’s health status is rather complex and laborious, it is true that in the long run, if this data has enough structure and is concise, it can serve to identify trends and common behaviors of disease, associating them with specific patient groups.

Medicine is far from being an exact science. Throughout history, the practice of medicine has acquired knowledge through experience and documentation. The implementation of high capacity data storage could therefore certainly be of use for the sharing of information between peers and institutions.

In today’s world, medical procedures and protocols are thoroughly examined not only by patients but by the general public. Health institutions have to make sure that their practices are updated to the highest global standards, are fool-proof, and are reported to provide satisfactory results. Pharmaceuticals are also highly scrutinized regarding their efficiency in the medical field, as well as the possible side effects of their dosages.

In summary, information is extremely important when studying public health, both as means of scientific advance in the area of medicine, and as a reliable source to judge the providers of health services in a global scale. Patients are, in a way, just clients whose information and behavior must be studied in order to pinpoint and satisfy their needs.

Having said this, is translating a human into possibly discriminative data worth the advantages of data collection that could benefit a wider range of patients?

Sources

http://www.lifesciences.sourcebioscience.com/genomic-services/ Retrieved Sep. 30. 2016

https://www.genome.gov/10001772/all-about-the–human-genome-project-hgp/ Retrieved Sep. 30. 2016

http://ucsdnews.ucsd.edu/pressrelease/flexible_wearable_electronic_skin_patch_offers_new_way_to_monitor_alcohol_l Retrieved Sep. 30. 2016
https://www.underarmour.com/en-us/ua-healthbox/pid1292219-001 Retrieved Sep. 30. 2016

http://spectrum.ieee.org/biomedical/devices/a-temporary-tattoo-that-senses-through-your-skin Retrieved Sep. 30. 2016 

https://www.genome.gov/10001204/specific-genetic-disorders/ Retrieved Sep. 30. 2016