Could algorithms keep you healthy?

If an apple a day keeps the doctor away, an Apple (or Microsoft or Google) algorithm may help keep vexing healthcare problems at bay.

Take the issue of sepsis, an infection notoriously difficult to diagnose in its early stages and with a 40% mortality rate. In the US alone, severe cases strike more than 1m people annually, with cost of treatment reaching at least $30bn. Over the past two years, however, predictive analytics have helped anticipate and reduce its occurrence. Using a patient’s health data, vital signs and other real-time information, algorithms can now estimate the risk that a particular condition is related to sepsis. If it is, doctors are notified and can treat patients in time. This save lives … and money: Reducing occurrences of sepsis can help a 300-bed hospital recoup up to $2m annually, according to Amara Health Analytics.

Big data, in fact, might become an important treatment for hospitals’ dire financial health. Rising healthcare costs, in particular, have increased the financial impacts of problems of payment collection. Costs can add up: US hospitals have given more than $459bn in uncompensated care since 2000, for example. Understanding cash-flow risk is thus key, notably when hospitals are expected to maintain quality of care while simultaneously cutting costs. Predictive analytics can help. GE just announced the launch of DenialsIQ, a software solution using statistical algorithms to help administrators identify the claims facing the highest risk of payment rejection by insurers. This enables administrators to proactively deal with these claims before they become a cash-flow problem.

Big data and predictive analytics could also help evaluate the quality of healthcare more holistically, facilitating the move away from a fee-for-service model towards a more value-based one. Overall, the market for healthcare analytics could reach $21bn by 2020, according to consulting firm IQ4I.

To get there, however, a few hurdles will have to be overcome, including compatibility between different data sources and the need to protect privacy. The former issue is likely to be addressed via a combination of standardisation and competition between different healthcare platforms.

“We are in a transitional time—one where we need to work with a number of colleagues across the industry to create standardised ways to interoperate,” says Jon Zimmerman, general manager at GE Healthcare. “At the same time,” he notes, “we cannot hold back great innovation to build out these standards. We need a blended world of standards and innovation progressing side by side.”

As for privacy, a proactive approach will be required. “Organisations cannot comprehensively identify gaps in security detection and response by solely focusing on breach prevention,” says Neil Jordan, general manager of Health Worldwide for Microsoft.

With only 15% of healthcare providers actively using predictive modelling, the market is still young. And in a realm where data are key to competitive advantage, the temptation may arise to keep the information to oneself. “We must not lose sight that, at the end of the day, the data is the patients’ data,” says Mr Zimmerman. “If there is a ‘data war’, then everybody loses.” Yet another reason for open platforms.

Author: Holly Hickman writes for GE Lookahead.