I found myself disoriented recently as I sat researching a conference on big data and analytics in the healthcare information technology (HIT) sector.
I spent many years in my career developing feasibility studies to justify the return on investment (ROI) for expenditures in business units and procurement departments. But as I listened to Chief Medical Information Officers and Chief Data Officers from leading healthcare enterprises, I found an industry segment where ROI can be viewed much more qualitatively.
Many of the HIT leaders I spoke to were direct about saying, “If we had to quantify a monetary ROI for these deployments, we’d never have a big data strategy.” Many referred to a “positive return on investment” as the key factor in justifying big data and analytics investment.
And while many large healthcare institutions spent more time on data ROI, it was spread across capital expenditures they had budgeted for years. This was especially the case for technologies that support emerging precision health and machine-learning approaches, which process massive amounts of data to validate or enhance human diagnosis and treatments.
So, even with my decades-old predilection to justify the financial benefits of IT investments, it became clear that many HIT organizations get a pass on ROI for data analytics for a quite obvious reason: What is the ROI on saving a life?
How can the ROI for curing an infant of cancer or slowing the effects of Alzheimer’s be quantified with a predicted dollar value? I’m sure there are actuarial scientists whose raison d’etre is to do exactly that. But I sensed that the executives from small to medium-sized hospitals erred on the side of humanitarian ROI, even when statistics on the monetary payback were available.
Despite this softer ROI based on saving lives, HIT data professionals admit they’re still scrutinized when data expenses are more than expected. (The get-out-of-jail-free card only lasts so long in the CFO suite!)
Most of these overruns center on the inability to predict the sheer volume of data growth as a byproduct of the data analytics strategy. As with hundreds of enterprises in other verticals, healthcare organizations tend to believe that big data analytics only work if you gather massive amounts of data.
And in this industry, there can be a massive number of intimate data sets generated from medical labs, patient engagements, billing systems and physician visits. These represent the famous “Big Data 4Vs” of volume, variety, velocity and veracity quite like no other industry. (Storing and processing these Vs come at a cost that is sometimes never factored into an already-soft ROI equation.)
The other challenge that arises is acquiring the talent to make the data findings actionable. Even in huge cognitive computing installations like Watson Health, where findings are theoretically produced in an actionable format, there still is a need for talent to understand and put to work the machine’s advice.
What experiences has your healthcare organization had with justifying the expenses of big data analytics deployments?