This post is also available in: Deutsch (German)
When you think of organisations applying some sort of Big Data Science to one’s advantage, the first thing that pops to mind is Netflix intelligently recommending shows based on your binge history. Nobody thinks of healthcare first as Big Data analysis use cases.
As a matter of fact, global pandemics and expensive health care systems pose a major challenge to mankind. But little do we realise that Big Data Science can make a difference in how healthcare is practiced (unless you’re in the healthcare industry). The possibilities of using Big Data Science are endless for the healthcare industry. In this article, we’ll touch on some of the ways Big Data Science can benefit healthcare.
Biology of diseases
Genomics, which seeks to understand the heredity of diseases by studying genomes, produces massive amounts of data.
Healthcare expert Dr. Bonnie Feldman said that each human genome has 20,000-25,000 genes comprising 3 million base pairs. And this amounts to 100 gigabytes of data, equivalent to 102,400 photos. That’s an immense data volume for one human genome.
Sequencing more genomes would amount to hundreds of petabytes of data, and then, the data created by analysis of gene interactions multiplies those further. It’s like an explosion of data! This is where Big Data Science takes a big step in.
Scientists at the Munich Leukemia Laboratory (MLL) sequenced a series of hot-spots in panels of around 80 blood cancer genes from patients with leukemia and lymphoma to identify mutations which require therapy. But they have to be careful to, not only find the mutation, but also interpret the mutations correctly.
So, MLL teamed up with tech giant IBM and biotech company Illumina to quickly analyse gene sequences and patient histories to locate changes in only a small number of cells or more subtle variants of mutations. They can then re-analyse genomic data from patients who appear to have the same disease to identify more specific sub-groups of diseases and pinpoint new therapeutic targets.
Medical records
The good thing about Big Data analysis is that it helps medical professionals store massive amounts of medical data for every patient and we can see this in the use of Electronic Health Records (EHRs), the digital version of a collection of patients’ medical records.
Different healthcare facilities can share a patient’s records with each other to coordinate care and prevent repetition of tests and treatments. Such highly-detailed shared records, which include individual genomic data, can also enable the doctor to determine a more precise and personalised treatment plan. And obviously, no need for tedious paperwork.
The Cleveland County Health Department (CCHD) in the USA knew very well that tracking down paper records is stressful and a waste of time and energy. This scenario led CCHD to adopt Patagonia Health EHR as their solution. The benefits reaped from adopting an EHR system went beyond digitising their paperwork. The EHR helped improve departmental efficiencies since all patient data is in one same system and can be accessed anywhere, anytime.
Real-time tracking
Another use of Big Data Science in healthcare is the use of wearable trackers. Wearable trackers automate the monitoring and recording of health conditions, so they save about 15 hours a week by removing the need to key in the healthcare data manually. They can also alert healthcare professionals to emergencies in real time.
One example is the Apple Heart Study, a collaborative effort by Apple and Stanford Medicine, which used data from Apple Watch to evaluate how well the Apple Watch detected irregular heart rhythms and sent notifications to the iPhone if an irregular heart rhythm resembled atrial fibrillation (AFib). Out of the 420,000 participants, about 2,200 participants received notifications about possible AFib, showing that the detection and notification features work, although it’s not clear how much they scored in accuracy.
Reducing Emergency Department Wait Times
The use of the Emergency Department (ED) has increased over the years, resulting in longer wait times and hence, delaying treatment for patients. Luckily, researchers at Stanford University came up with a model to predict how many patients will come to the ED so that the ED can find ways to minimise wait times.
Currently, one way EDs in the USA do this is usually by diverting patients to other healthcare facilities while telling ambulance services to transport the remaining patients to those facilities once the ED is crowded. This particular diversion model, however, relies exclusively on congestion data.
The new Stanford model, on the other hand, uses both predictive data and congestion data to determine when congestion will build in the future, enabling diversion to be done even before the congestion can occur and trigger ongoing delays. Alternatively, EDs can also use the predictive data to assign an appropriate number of emergency room staff to a particular facility.
Pandemic action analysis
In order to contain a pandemic situation, governments and scientists can use Big Data tools and Data Science to analyse situations. This is important to avoid overloading healthcare systems.
One way is to use Data Science techniques to visualise and get insights about the spreading or the recovery of a infection like the Coronavirus. Additionally, countries or regions can introduce lockdowns to limit the spreading of an infection.
Big Data analysis tools enable governments then to control the effectiveness of the measures. For example, the mobile cell tower data can be used to determine to which citizens obey lockdown measures and how strong the mobility is.
The result of such a Big Data analysis can then lead to insights on how to strengthen or loosen the restrictions to control a pandemic like that of the Coronavirus better.
Contact tracing for Viruses
In a pandemic situation, healthcare professionals and politicians face the dilemma to decide between pandemic containments and social-economic factors.
A high technology-driven approach is to use smartphone data to determine contacts of a virus-infected person. The solutions range from precise geotracking of all citizens by capturing the data at a central point over app-based solutions for sharing the geolocations down to anonymous and privacy aware contact tracing solutions.
All in all, the idea is always the same to use contact tracing data to break infection chains faster and to allow a higher quality of life for citizens during a pandemic.
What does this mean for people in healthcare?
We can see that Big Data Science can be useful to the healthcare industry in several ways. It’s changed the way genomes are used; the way a patient’s health condition is monitored, recorded and shared; and the way congestion is managed in the Emergency Department. Moreover, Big Data Science applications tend to promise a higher quality of life during a pandemic.
It has also changed the skill sets the healthcare workforce is required to learn to make the best use of Big Data for their jobs. Medical research laboratories have made space for computer facilities to enable medical researchers to process vast amounts of data, a task which requires knowledge in using algorithms.
A few biology-related undergraduate degrees have adapted to this need for data science skills in the field. Newcastle University introduced a scientific computing module for its biology undergraduate students to learn programming and computational modelling.
New York University’s School of Medicine introduced Health Care by the Numbers, a three-year blended curriculum, to train medical students in using Big Data.
Ultimately, data-driven approaches are the future of all healthcare systems. Equipping the healthcare workforce with the necessary tools and skills is the first step in applying Big Data Science in healthcare.