In healthcare, we talk a lot about transformation, but what does that mean? Typically, these conversations come down to ‘better’ or ‘improved’ integration of care and funding models.
We don’t need ‘better.’ We need different.
As healthcare organizations, we agree that transformation needs to occur but where we start or how that looks is a mystery…until now. Big Data is here to stay and it’s time for it to lead the way to transformation.
Poor Data Leads to Poor Insights
Let’s paint a common picture. A patient tries to go to their family doctor but can’t get an appointment for more than a week so they go to a walk-in clinic. They wait two hours to see the doctor who proceeds to ask for all the same information their family doctor has already. Once done collecting information, the doctor proceeds with a preliminary diagnosis, a prescription for a medication, sends the patient to the laboratory for some blood tests and tells them to follow-up with the family doctor next week. But the family doctor, has no idea what the diagnosis was because they have no records from the walk-in visit or the laboratory test results.
Today’s patient is responsible for keeping track of their own records, visits, and medications. A bank can keep track of your purchases from around the world but healthcare providers can’t keep track of what happens to patients within a one kilometre radius.
This is because there is no standardized approach to collecting or sharing data. Data collection is provider-centred rather than patient-centred. As providers ask questions it’s up to the patient or family-member to answer despite this information residing in multiple provider datasets.
Although Ontario is generally a health-data rich province, using this data to inform decision-making and converting data into actionable insights is a gold-mine we’ve only started to tap.
Now we aren’t under the illusion that transforming this system is easy – care is complex, funding comes from multiple sources, science and technology is changing constantly. Data collection has increased drastically but the data remain siloed – hospital medical records are housed in one silo, family doctor records in another, public health in another, and the data are never brought together to create the patient journey, resulting in lack of visibility into continuity between settings (e.g. hospital to community care to home) or over time (e.g. transition from child and youth to adult).
Do these issues seem unique to healthcare? They aren’t. The retail sector, hospitality services, transportation services and so many more have all dealt with similar issues but they understood the value of using data to learn about their consumers to better serve their needs. This is the same business and cultural understanding that we have to embrace: understand the consumer to better serve the needs of your consumer. At the end of the day, patients are consumers of healthcare.
Use flight bookings as a gauge – most of us don’t use travel agents to book flights anymore. We look at a couple of travel sites like Expedia or Google Flights and pick a flight based on the best value. We use these sites because they aggregate data from various flight carriers and allow us to do our own cost-benefit analysis based on options for airline, date, and flight path. In other words, travel sites use aggregated data from multiple sources and organize the data around what’s relevant to the customer instead of having them visit one carrier page at a time. In understanding the needs of the consumer flight bookings were transformed for the better.
Addressing the Problem, not the Symptoms
Siloed data are a major barrier to transformation. The major risk is that siloed data cause us to mistakenly identify symptoms of the problem as the true problem.
To drive transformation, we first need to understand the problems. Symptoms are very important to observe but they are bread-crumbs leading us down the path to the root problem; they are not the destination itself. By ignoring data we will continue throwing resources at fixing the symptoms. Consider hospital capacity issues. Do we need more beds? Or do we need to redirect patients before they get to the hospital? We don’t know because we aren’t looking at the data.
Poor data collection and sharing prevents us from understanding our most devastating crises. Opioids are responsible for hundreds of overdose deaths a year yet the root causes are not well understood because we don’t survey prescriptions that are publicly and privately funded in one system. We can describe prescription patterns but we can’t actually tell what’s driving them.
There is Hope
Big Data methods “allow us to augment what can be understood from massive amounts of data like how a microscope enhances eyesight,” Harlan Krumholz, a professor at Yale University, argued in Health Affairs.
Big Data analytic platforms are necessary for the transformation of healthcare. They allow us to collect and purposely link data from various sources in various formats in order to analyze data faster, making meaningful insights that are actionable.
Initiatives such as Reconnect Community Health Services show the value of functional linkages with disparate datasets. Using their linked data for community health services they dispelled the hypothesis that patients were accessing the same service from multiple community health service providers. In fact, they found more than 80 percent of unique clients received services from one community health service provider.
Their dataset also enables them to show a single client’s journey through separate health service organizations over a longitudinal time period. These seemingly simple insights allow us to observe critically important issues and to dig deeper into the real problem.
It’s time to transform healthcare by answering the known-unknowns (questions we currently have without answers) and start to discover the unknown-unknowns (things we could discover and knew nothing about, but would truly surprise us).
The Big Data wave has arrived – and one of the many forms it comes to us in is a problem-identification-enabler which can lead the transformation we need.
The comments section is closed.
Informative article.
Much of this is already being accomplished through the work on ConnectingGTA and ConnectingOntario. We should be careful that theoretical benefits don’t become a bludgeon to impose poorly designed electronic medical records on already overburdened health providers, as occurred in the US following the adoption of the ACA.
One real benefit of big data would be to use the data collected in our various hospitals for machine learning projects, which Ontario and Quebec would be well placed to do given their proximity to world leading researchers and companies in the area.
Buzzkiller
Everyone hates silos and loves data. That being said, healthcare’s present love of data is reminiscent of its love of computerized records a decade ago. It’s been a painful and expensive transition to the digital world, with relatively little evidence that the results have been worth the costs.
For dissemination of information and communication, technology has been a godsend. But health care is a human (and political) enterprise. Data needs input, which these days is diverting an ever-increasing share of health professionals’ time. Data needs analysis, which demands commitment of talented people and resources. And data analysis needs action, or the entire exercise becomes that very Canadian activity of studying a problem to exhaustion then doing whatever’s politically expedient.
I too was a great enthusiast for electronic records and data sharing, until that data became a bludgeon for the bureaucracy to threaten funding, and a treasure trove of excuses for political inaction.
Data works magnificently for the financial industry because it’s numbers on a screen representing numbers in a bank account. But it can’t capture or reflect the burned-out caregiver spouse of an elderly person languishing in hospital with dementia; or the family doctor that inherited a patient on 10 times the maximum recommended dose of morphine, who’s dealing with the College investigation into her prescribing; or the casual nurse mother who can’t get enough hours to pay the bills; or the careerist administrator who won’t compromise an inch on policy, even in extraordinary circumstances.
Data is information, nothing more, nothing less. It’s who and what we do or don’t with it that matters.
Excellent article.
Good article with examples from other sectors.
Data well-used can help redesign our system.
Talented people like Arif and Morgan are fundamental to improving health and healthcare.
It’s terrific to see a thoughtful piece advocating for better organization and use of healthcare data to drive clinical practice, healthcare system performance, etc. Congratulations on putting this together. The historical investments in legacy systems and infrastructure, particularly in smaller hospitals / centres, are often cited as a major barrier to change in a relatively conservative sector (i.e. healthcare). What advice would you give to those who feel they may be ‘too small’, ‘too vested in legacy systems and infrastructure’, or ‘not sufficiently resourced’ to move to infrastructures that enable advanced analytics?
Thanks for the question Muhammad. Whether your organization seems too small, too vested in legacy systems, or insufficiently resourced, I would encourage us to think big but start small.
Start by addressing one problem at a time. Pick a problem that is aligned to your organizational strategy, feasible with currently available data with an appropriate level of quality, and most of all, something that is actionable.
Creating value through actionable insights is the ultimate goal. Starting small allows us to take that ‘risk’ by not putting much at risk.
Big data presents oppourtunity in many industries. Do you think the healthcare system has the incentives to attract top talent in competition with financial services/ technology/ other industry?
Great question Warren. In healthcare we need to entice talent to join through the motivation to contribute to the health of a community. Other industries can compensate much better than healthcare but I believe it starts with compassion and knowing that this is heart-work. Compensation will come later – once we use big data to identify problems and solve the right ones, the savings will come.