One of the great things about a list of predictions is that seldom does anyone come back and see how accurate you were!
In spite of that, these are based on my best sense of the top priorities facing information technologists in the arenas of advanced health technology, mobile health, analytics, clinical research and precision medicine.
Here they are:
1. Natural Language Processing: The importance of the majority of the content in most electronic medical records – unstructured text - will continue to increase. The needs of clinical decision support, predictive analytics, cohort identification, clinical trial recruiting, and comparative effectiveness efforts will drive clinicians and researchers to demand the reliable mining of this promising resource. And the ability of computer algorithms, ontologies and the informaticists that build and maintain them to derive real value from the vastly diverse mixture of unstructured text will continue to fall short. But millions will be spent in pursuit. Because we have to keep trying.
2. Predictive analytics: Real value will be delivered by this technology. In fact the question will shift from “Can we predict?” to “Who do we inform about a prediction?” and “How soon is too soon to predict a future event?” The value of prediction is only realized if those who can intervene to prevent a predicted bad outcome or ensure a good outcome are informed at the right moment and in the appropriate workflow to easily intervene. False positives and false negatives must be avoided at all costs lest clinicians lose faith and ignore the predictions.
3. Wearables: All three of my sons are now wearing FitBits and competing with each other! My Apple Watch and a great app – Heart Watch – provide me almost 24 hour monitoring of my heart rate. The benefits of the devices and associated apps directly to the user are obvious and catching on rapidly. Enabling the patient and their provider to share this data in real time will spread more slowly. But it will spread in 2016. Specific chronic conditions and clinical use cases (post–surgery mobility, wound management, infection detection, hypertension management, glucose monitoring, sleep studies, etc.) will be tracked or measured on your smartphone. Your clinical team will use the data to help you heal, stay healthy or remotely predict when to remove your surgical drain. And your employer may reduce your insurance premiums if you meaningfully use a device to boot!
4. Mobile health: I pioneered real-time mobile health on a Palm VII in 2000. Really. So why am I not a multi-millionaire venture capitalist now? Well it was 10 years before the iPhone for one. Secondly, I ran out of patience and money waiting for more than 5 percent of the clinicians to adopt the fragile, battery challenged, connectivity starved technology. Clearly we have passed that point! In fact almost every nurse, doctor, case manager, and staff member is expecting to use a smartphone to complete some portion of their daily work tasks in the same way they use a mobile device to complete more and more of their non-work tasks (messaging, banking, social media, shopping, video watching, gaming, etc.). This does not mean we can remove all of the desktops and WOWs from the buildings. But mobile health is here to stay and each task performed by our mobile health workers needs to be assessed to determine if it can be more efficiently performed on a smartphone. And many of us will need to buy or build this technology independent of what the mainstream vendors (Epic, Cerner, etc.) provide. Health workers will demand the same cool look, feel and efficiency they currently enjoy with their personal apps and the vendors will be challenged to keep pace.
5. Clinical research: Clinical researchers have grown tired of having to maintain separate research databases containing discrete data manually abstracted from the electronic medical record. They don’t want to use clinical trial management systems that function in a parallel research universe separate from the clinical enterprise. Industry sponsors of research at academic medical centers want to mine the electronic medical record content before the trial starts to ensure inclusion and exclusion criteria are realistic and trial recruitment goals are achievable. All of these trends will put pressure on EMR vendors to enable their products to integrate more tightly with the processes and data associated with clinical research. Academic medical centers, clinical research organizations, pharmaceutical companies and medical device companies will develop ways to work more closely to shorten the drug and device development timelines and reduce development costs.
6. Precision Medicine: It has been stated that there is a lot of precision but not a lot of medicine in the world of precision medicine. This sentiment stems from several challenges: bringing the genetic result data into the EMR in a way that clinicians can act on it has yet to be developed by EMR vendors; the science of genetics is changing rapidly and determining the best course of clinical action based on clinically actionable genetic variants (germ line or somatic) is a moving target and, consequently, the software that manages the genetic pipeline processing and result generation continues to evolve at a rapid pace; external genetic testing vendors are reluctant to share their discrete data results and have not built the infrastructure to allow for two way exchange of test orders and test results with EMRs; and, lastly, standards for electronic exchange of discrete genetic results are insufficient and continue to evolve as the science evolves.
These challenges will not be resolved in 2016 but leading clinical research enterprises will make progress. Another mini-prediction: Most healthcare organizations will choose to develop their own solutions in-house.
Brian Wells is associate vice president of health technology and academic computing at Penn Medicine.