The health system has traditionally been seen as data rich and information poor though this adage does not hold the same validity that it did 25 years ago. While we continue to improve the information value chain as it relates to direct patient care, the unsustainable year after year increases in the costs of providing health care in Canada requires business-like approaches that include more robust and actionable business intelligence commonly referred to as health analytics. In recent years, discussion around health analytics usually includes discussion around the term big data.
If you browse around the Internet you will see many definitions for big data and all have the common theme of variety, velocity, volume and the inability to manage or analyze these data sets using traditional relational data base tools and techniques. While storage of these data sets is no longer as cost prohibitive as it was, the computational power and advanced techniques and technologies to extract value from them requires new skills, new technologies and innovative approaches. This also implies additional resources.
In recent years, four software industry leaders spent over $15 billion acquiring management and analytics software firms in an industry worth approximately $100 billion per year and growing 10% annually, approximately twice as fast as the rest of the industry.1 One thing is certain, big data is big business.
What does all of this mean for health care in Canada? The McKinsey Global Institute estimates that in United States health care, effective utilization of big data to drive efficiency and quality could result in more than $300 billion in annual value of which two thirds could be cost reductions equivalent to 8% of national health care expenditure.2 A compelling statistic even if they are half right. Attaining more value from the current health care dollar is something we cannot ignore in this country given the current cost trends. It is time for us to get more serious about big data.
Aggregate data is currently available that could be used (and sometimes already is used) to mine data for knowledge related to unwanted drug interactions and predicting disease occurrences before symptoms occur. Sometimes this data is stored in data warehouses that can be used to identify causal correlations and drive predictive modeling. The increasingly valuable area of genomics research also requires large aggregate data sets to be created, maintained, accessed and analyzed where appropriate.
However, too often we are collecting information for direct patient care that may not be as easily accessible as it should be for driving the additional value that analytics and big data promise. This needs to be addressed as we continue to deploy expensive systems that may be limiting long term value-add capabilities. Incremental short term investments in current initiatives may reap substantial long term benefits and more importantly we need to consider analytics and big data as part of any new system deployment projects.
New techniques and approaches are currently employed that capture large streams of information in real time for purposes of safer and higher quality health care. There is an initiative called Project Artemis that has researchers at Toronto’s Hospital for Sick Children using this technology to aggregate and examine the large streams of vital sign data collected from premature babies who later developed sepsis. The researchers have found patterns in the data that could predict up to 24 hours in advance when sepsis was developing in these most vulnerable of patients. In the past, these kinds of insights derived from the data either were not possible or cost prohibitive. This is good stuff and we need more of it.
Over the last few years, the Conference of Deputy Ministers of Health has taken a leadership role in the national health analytics agenda. That has resulted in a Canadian Institute for Health Information and Canada Health Infoway collaboration, along with participation and meaningful engagement from provinces and territories as well as other key stakeholders across the country. A vision paper is currently under development that hopefully will raise the awareness at senior levels of government as to the potential short term and long term value of investing in analytics and big data tools and technologies. The next logical step in this process should be a national strategy to promote and assist the numerous health systems in this country to avail of this value-add and prepare for future developments. A huge gap that needs to be addressed is the human resource capacity and skills required to operate and extract knowledge from these complex systems and technologies as well the expertise to ask the right questions for the right reasons. Stay tuned.
As a public sector executive managing in fiscally challenging times, I do want us to approach the analytics and big data issue frugally at this time. First of all, there are large data sets currently available that undoubtedly hold valuable information to support the health system and its clients. It is time to leverage this information.
Secondly, we need to collaborate nationally and learn from each other as we move deeper into the analytics and big data picture. One of the compelling characteristics of big data is the sharing of the knowledge and wisdom that comes from complex analysis of these data sets. Thirdly, we must be careful to protect the information assets we have in the public health sector such that any benefits that are derived from the “peoples” information, commercial or otherwise, go toward creating an even stronger health system and healthier population in the future.
Big data is a big deal. Let us take advantage of it in a big way.
1 A special report on managing information: Data, data everywhere | The Economist, February 25th, 2010.
2 Big Data: The Next Frontier for innovation, competition, and productivity. The Mckinsey Global Institute, May 2011