Picture of Kirsten PatrickKirsten Patrick is Deputy Editor at CMAJ


In the last two weeks I’ve attended three very different scientific conferences on behalf of the CMAJ Group. In fact you couldn’t get more different than the 33rd International Conference on Pharmacoepidemiology and Therapeutic Risk Management (ICPE – all Big Data and massive record linkage aimed at finding out more about the benefits and harms of medicines and devices) and the 5th Canadian Conference on Physician Health (mainly focusing on the major problem of physician burnout and what we should do about it). And yet the same study was mentioned by plenary speakers at both conferences to support the same message: that physicians are overburdened by administrative and data-capture demands. Across four medical specialties, “for every hour physicians provide direct clinical face time to patients, nearly 2 additional hours is spent on EHR and desk work within the clinic day,” found the authors of a study published in Annals of Internal Medicine in December 2016. The electronic health record and email are robbing physicians of the time that they could be spending with patients.

At ICPE, the eminent Robert Califf, latterly commissioner of the FDA, now splitting his time between being Vice Chancellor for Health Data Science at Duke Health and working in an advisory capacity for Google biotech spinoff Verily Life Sciences, said that “We are no longer limited by technology” in delivering on the promise of Big Data for health. “The issues that we face now are all related to cultural and business concerns.”

Since just about all health care interactions end up with something going into a computer there is really no excuse for not progressively learning from the system, in real time and at the level of the family and social circle, said Califf. But he did acknowledge the burden that the capture of high quality electronic health data places on physicians, citing research showing that physicians were increasingly leaving the profession well before retirement age.

I noticed this thread picked up at the Conference on Physician Health by plenary speaker, Dr. Tait Shanafelt, who discussed the problem of physician burnout. Shanafelt was recently appointed as Stanford Medicine’s first Chief Wellness Officer but previously studied physician burnout extensively at the Mayo Clinic. In a longitudinal study of physicians at the Mayo, Shanafelt and colleagues found that high burnout scores and declining satisfaction with clinical work (as measured on staff surveys) were strongly associated with actual reductions in professional work time, that is, scaling back on work hours, over the following 24 months (through linkage to admin and payroll records).

It seems that while many are excited about the promises of Big Data for Health, the burden of capturing those data are weighing heavily on health care personnel and physicians in particular.

1 in 20 Google searches are health related, Google has found. That’s 5% of 3.5 billion – 175 million – health-related web searches a day, according to Califf. The biggest category is searches for a ‘condition’; the second biggest category is searches for ‘treatment’; and the third biggest is ‘home remedies’ (that is, people trying to avoid going to the doctor).  According to Califf, Google team’s mission is to “make health information universally accessible and useful”. No easy task considering that millions of peddlers of snake oil and misinformation are using the Internet quite as easily as purveyors of sound, evidence-based medical information.

Califf talked (too) briefly about Google’s findings regarding mental illness. 300 million people suffer from depression around the world. About half don’t get treatment. If they do treatment is often delayed for up to 7 years yet treatment is effective for 70% of sufferers. Tens of millions of people search for information on depression each week. But many don’t do that. Yet advances in machine learning can allow us to tell changes in people’s mental state by how they use a keyboard, how they take pictures, their behaviour on social media and altered sleep patterns inferred by time of internet use. Creepy? Perhaps. But could it save lives? Lead to earlier identification and treatment of mental illness? Very likely. When Google detects a search on a depression-related term, a box pops up now with questions from the PHQ9. Based on how the individual fills out that questionnaire – if they do – what pops up next is different. According to Califf, this is an experiment and it’s not clear ‘where to next’, but Google will collect anonymized information on what people DO do next.

Perhaps a way to relieve the burden placed on doctors by computer time and mandatory data capture is to innovate to find other reliable and comprehensive ways to capture patients’ history and symptoms information…mechanisms that don’t involve the physician. A mixture of passive data collection and administration of validated questionnaires and information forms that patients fill out before seeing a physician, perhaps. Whatever the future of health care data holds it needs to involve innovation that can relieve the some of the burden of electronic data collection from the clinician.