The information ecosystem is growing at an unprecedented rate and technologies with advanced capabilities to track and evaluate that information are multiplying. Smartphones. Usable. Connected medical devices. All of these innovations harness the power to transform health outcomes; all require constant data collection and presentation to do so.
Breakdown of Big Data in Healthcare
In a report on macrodata in health care from Healthbox, experts shared their ideas on how to break down the noise in a health care ecosystem with more data than ever before. The report noted that the term ‘big data’ was originally coined in the 1990s to describe data sets that are too large or complex for traditional databases to handle.
According to James Gaston, senior director of maturity models at HIMSS, “[Our cultural definition] is moving away from a brick-and-mortar-centered event to a broader, patient-centered continuum that encompasses lifestyle, geography, social determinants of health, and fitness data in addition to traditional episodic health care data. The industry is about to learn how powerful macrodata is in healthcare, he said.
“The sheer volume, speed, and variety of data being collected pose challenges for leveraging and ensuring its validity to benefit both population-level macroeconomic health and evidence-based precision micromedicine,” the report states. In other words, finding meaning within mountains of data is an enormous task for any individual in any health system role.
This is where the power of innovations such as artificial intelligence (AI) comes in. Lily Peng, MD, PhD, product manager of Google Brain’s Artificial Intelligence Research Group, explained that while human intelligence is best suited to integrate a small number of very large effect factors, artificial intelligence is particularly adept at analyzing and identifying patterns in a large number of small effect or obscure factors.
How Big Data and AI Can Collaborate to Improve Decision Making
In a world flooded with data, people can be confident that while artificial intelligence and big data in healthcare have immense potential, there are still limitations that prevent them from being a substitute for universal decision-making. There should not be a single innovation as a single solution.
Combining the power of machine learning and human intelligence to obtain valuable information from large data sets will require focusing on four different areas, as shared in the report.
- Eliminating biases in data collection
“The lens that each researcher brings to the macrodata creates inherent biases,” the report said. This can include everything from the categorization of data assessed, how the data were collected, and so on. “The power of high-dimensional data is supposed to lie in the absence of hidden confounding factors that remain undisclosed in the data. Unfortunately, this assumption is far from being a forgotten conclusion and represents a threat to the validity of conclusions derived from macrodata through AI techniques”.
- Recognize the inherent conflict between anonymity and specificity
“Due caution should be taken to structure analyses so that reverse engineering of patient identities does not occur; however, it is worth noting that the benefit of open data sharing outweighs the adverse potential for individual reidentification.
“Society will have to weigh the ethical trade-offs between the benefits of open data sharing and the finite but real possibility of re-identifying individuals by reverse engineering segmented data. Human intelligence, not artificial intelligence, will be required to deal with these questions.
- Significant validation and measurable impact of collected data
The use of macrodata in health care can pave the way for providing patients with more detailed and understandable guidance on how to manage chronic diseases and other important health conditions. But will increased access to this information lead directly to better outcomes, satisfaction, and the overall consumer experience?
“Data integration, AI-derived knowledge, and informed clinical decisions must be adopted and closely woven into clinical processes and workflow to generate potential benefits in patient care. Properly structured clinical trials are needed to demonstrate that the incremental benefits of an evidence-based process of care justify the costs (and complications) incurred by these decisions.
- Understanding underlying cause and effect relationships
Healthbox emphasized that in data analysis, it is important to keep in mind the old rule that correlation does not imply causality. It is also important to “make sure that the data under analysis do not suffer from the omission of confounding factors that may be causally related to the measured results.
“The experience of the domain and human intuition will always be necessary to work together with the artificial intelligence to confirm the absence of hidden confounding factors . The use of machines can help humans to reveal these undiscovered or unexpected variables.
With this knowledge in mind, it is clear that through a collaborative approach, we can design a better strategy for success with big data in health care, which will take us further on our way to harnessing the full powers of innovation in health care. The continued advent of artificial intelligence technologies will amplify the value of big data, paving the way for a more collaborative and human-centered approach that is aided, not hindered, by new technologies in health and care.
Read more: https://www.himss.org/resources/big-data-healthcare-breaking-through-noise
