Is it possible to give sensitivity training to data? Should we if we could? On-demand reports, data dashboards and faster access to decision-enabling information help health care organizations offer better care and streamline operations. But they also raise new questions about how data informs conversations with patients. As technologists, what can we do? What is the responsible thing to do?
Business intelligence (BI) has already been a great boon to health care. The value of mathematical data mining and pattern evolution cannot be denied. Statistical analyses have been instrumental in achieving standardized patient care quality reporting, revenue cycle management, and health care information/records exchange.
The next step in the transformation of care delivery will be to use similar techniques on predictive disease analysis, proactive diagnosis, treatment and outcomes. A CIO/CXO survey conducted by SearchHealthIT on the importance of health care analytics shows that these objectives are top-of-mind. Almost 70 percent of survey respondents would like to use health care analytics on predictive analysis of diseases; more than 85 percent emphasized it as a tool to improve diagnosis, treatment and outcome.
Moving from BI to true analytics will require a much more holistic view. Like BI, health care analytics start with using super-statistical techniques to assess large data sets and identify historical patterns and trends. The ultimate goal is to mine that historical pattern creation to devise and predict future behavior. But in my humble opinion, health care analytics must be more than just statistical analysis. It must be combined with behavioral models, cultural analysis, and demographic research. These clinical and care-centered objectives are best achieved by the confluence of logic (data), emotion (behavior), and context (culture/demographic).
Patients being treated for the same disease are still very different, and every ailment that a given patient is being treated for has its own idiosyncrasies. The same depth of variance can be applied to the health care provider as well. With all the complex play of emotions that abounds in a health care delivery ‘transaction’ pure mathematical analysis, or, even predictive modeling, only goes so far.
A recent NYTimes.com article highlights the power of analytics, as Target was able to not only identify but also predict consumers’ shopping patterns. This is a radical milestone in the path of optimal consumer intimacy and one that suggests powerful possibilities in the health care environment.
However, health care can also place people in raw emotional situations and difficult psychological circumstances. In my own experiences, I have seen how important it is that a patient asking about the probability of recovery gets a response that is precise and founded on data, but is also crafted and conveyed in a manner that is palatable to his behavioral and cultural mindset.
I believe that as technologists we play a critical role in shaping the future of health care analytics and predictive health care analytics. We need to create a seamless, systemic flow of mined information that can be analyzed to drive tangible decisions/projections. We also need to think about context and think about systems that we build. Do they take the patient’s physical health and mental well-being into account? As often as possible, we need to think about not only what information is presented, but also how it is presented.