Information is being added to the cloud at an astonishing pace.
This image available via Intel at http://www.intel.com/content/dam/www/public/us/en/images/illustrations/embedded-infographic-600-logo.jpg
The rate at which this information is being created far exceeds the potential of the human mind to process it. Thus we have to increasingly depend on computers to do this for us. Even when the computer presents us some analysis that we find hard to believe, it is difficult to untangle and trace the algorithmic threads to get down to the primary data. We have a progressively slimmer grasp on the truth as we “knew” it.
How does this apply to medicine? One may think that most of this data is irrelevant to us in medicine. The fact is that Big Data is increasingly impacting health care and will soon be a tsunami that will sweep away the traditional practice of medicine.
Some of the sources of "Big Data" in Healthcare.
- The amount of information available on any topic is growing rapidly - MEDLINE for example adds 2000-4000 citations to its database each day (3)!
- Health care providers are generating huge amount of data in the EHR - according to an informal calculation about 4 MB of text data is added to the EHR for each patient per year (4). This is equal to about 1000 single spaced word processor documents.
- With the advent of wearable computers and sensors; we are entering the age of Internet of Things (5). There will be a steady stream of data like real time blood pressure, pulse, blood sugar feeds that will need to be interpreted to help make the best decisions for each patient.
- As DNA sequencing goes mainstream physicians will have to review DNA sequencing data in addition to getting a history and doing a physical exam. There will be increasing information from research into the clinical impact of a person’s genome. Physicians will have to learn genome-based medicine in addition to evidence-based medicine.
As the amount of medical data increases, we are rapidly approaching a tipping point where the only way we will be able to practice medicine is to rely on computers to interpret the data for us. The question is how do we prepare our trainees for this future.
Traditional Medical Education and Practice
We have traditionally taught our trainees to learn to search for and appraise the evidence before applying it to patient care. We ask them not to just depend on UpToDate and other similar repositories and to learn to search for the primary evidence. We ask them to not take the information in problem lists as gospel but to look up the primary data in the EHR. Thus, when a patient is labelled as having COPD, we ask our students and residents to confirm that the history fits the picture of COPD, to look up the pulmonary function test results or review the chest Xray film. When a specialists starts our patient on an expensive new medication, we look up the trial to calculate the number needed to treat and use this to discuss options with the patients.
With a seemingly impossible amount of information to search and appraise and an equally impossible amount of patient data to sort through, practicing medicine the way we are used to and were trained to, is going to be progressively more challenging. We will progressively need to rely on machines (computers) with artificial intelligence to sort through the morass of patient data and the growing body of evidence in medical literature and present us patient summaries and recommendations so we can make sense of this. Just as we now increasingly rely on UpToDate to find quick answers, we will soon depend on machines with artificial intelligence to guide us in our decision-making.
Just as knowing how to read a map and knowing which way is north is critical for a reality check when using a GPS, what will our students need to know to practice medicine in the future? How do we prepare them for this?
Neil Mehta, MBBS, MS
Web Editor, JGIM
- The GPS a fatally misleading travel companion: NPR
- In Maine, even with a GPS, you cannot get there from here: NPR
- MEDLINE fact sheet: NLM
- The cost of storing patient records: John Halamka
- The Internet of Things: IBMSocial