Wired / The prognosis for data-led medicine is healthy

This article appears in the June 2014 issue of Wired magazine.

Today’s medicine is more practice than science, albeit intelligent practice. Too much is left to qualitative judgments based on tradition, experience and dated, often un-tested, beliefs. But we are about to make a quantum leap in the capabilities of medicine driven by digital health technologies, sensors and data science. It’s only a matter of time before this leap in technology empowers individuals to become the CEOs of their own health.

There are a number of nascent efforts, Lumiata and IBM’s Watson among them, that capture medical literature and use natural language-processing and curation by human doctors to make this information available to medics and consumers. There are also efforts to validate today’s accepted medical practices using patient records to assess the performance of the approximately 15,000 therapies that we currently employ. As the body of data and data-science insight grows, this knowledge will become increasingly personalised to the condition of each individual and be made available to patients. This will allow patients to understand their medical choices and empower them to make more of the treatment decisions they otherwise are forced to accept today for lack of knowledge.

These systems will be complemented by wearable sensors, which will collectively generate swathes of data. Initially, this data will represent traditional medical variables such as blood pressure, respiratory rate, blood-glucose correlations or cardiac output. Over time, patient activities, in addition to thousands of biomarkers from blood, urine and breath, will be monitored by smartphone apps.

Mobile-phone-based ECGs, blood-pressure and heart-rate monitors, skin-hydration sensors, blood-glucose monitors, contextual data trackers, stress monitors and many more sources of data will feed into increasingly sophisticated machine-learning systems. Smartphone apps will serve as constant monitors, help with behaviour modification and even replace drugs. In the US, the Food and Drug Administration has already approved the AliveCor ECG device as an over-the-counter purchase so any cardiac patient can take their ECG at any time. It has also approved WellDoc, a type-2 diabetes management app. And the smartphone app can monitor a patient’s mental-health status more accurately than any human.

The best doctors will monitor and train these systems and curate their input and output, and eventually the systems will exceed the capabilities of their creators. Taken together, the human/machine combination will exceed the performance that either could possible achieve alone.

Technology also will take into account factors such as comorbidity, genetics, lifestyle and many others that tend to be neglected by doctors. The volume of data and insights about patients’ complex conditions is more than humans can easily interpret, but computers can manage far more inputs and recognise patterns and anomalies. When human doctors are not present or when only those with lesser medical training are available, these systems will be able to personalise care. Patient options today are sometimes limited by the bias of individual doctors; the balance of power in such decisions will shift to the patient.

The new systems will not eliminate the need for the human element of care, but the kinds of people providing care will change. Over time, the need for full medical degrees to perform most functions will decline, and medical professionals with lesser training will be able to do a lot more. And when an MD is involved, they will have more information, knowledge and guidance available to them.

Data will fundamentally change healthcare, and individuals will be empowered to manage their care as technology grants them unprecedented tools.