Identify metrics when piloting patient monitoring apps

Hospitals frequently use information technology services and products from large, blue-chip corporations like Epic Systems Corp. or McKesson Corp. But as they hunt for apps that can help them monitor and engage patients in the months following discharge, they are dealing with a whole new world comprised of many, small startups.

Many of these companies' wares have a limited track record. And the firms themselves sometimes lack stable funding.

And, the gee-whiz techies behind the app usually haven't developed valid metrics for evaluating its capabilities. Measures like readmission rates or cost-per-patient are too broad for determining the value of a post-discharge monitoring tool.

“We need to prove what works” before scaling it across the entire system, said Dr. Zenobia Brown, a medical director at Northwell Care Solutions in Great Neck, N.Y.

Hospitals are looking for post-discharging monitoring technology so they can remain financially viable in the services being thrown into Medicare's new bundled payment programs. Nearly 800 hospitals are now responsible for patients' clinical and financial outcome 90 days after being discharged for a hip or knee replacement. If they fail to hit a target price set by the CMS, they will lose money on the entire episode of care.

In July, the CMS expanded the program to include bundled payments for episodes involving heart attacks, coronary bypasses operations, and hip and femur fractures. The cardiac care bundles alone represent about $11 billion in annual Medicare payments, based on a CMS analysis of historical episode spending between 2012 and 2014.

Many of the startup companies trying to help hospital systems prepare for this brave new world of bundled payments do not have solid financial backing and that can undermine a project. Froedtert Health, for instance, partnered with a company that offered a congestive heart failure application for patients.

The hospital received data from electronic scales, blood pressure cuffs and pulse oximeters in people's homes and used the company's analytic system, which incorporated machine learning, to identify early sign of risks and intervention.

After five months, even though the application performed well, the company lost its venture capital funding and the project folded.

Large advisory firms are very familiar with hospital's workflow. But this is less true for startup companies, who might be eager to do a pilot, free of charge, just to get some more data points on how the application performs in the real world.

“There are a lot of solutions without clear problems,” said Matthew Fenty, director of innovation and strategic partnerships at St. Luke's University Health Network in Bethlehem, Pa., explaining that often vendors have trouble articulating the value of their product and require operational insight from the hospital, which can take extra resources and time to vet.

But the startup vendor and the hospital system need to be very clear what new capabilities the technology brings to the table and if that capability is a “need or a nice to have,” Fenty said. To determine an app's usefulness, key players like physicians, nurses and administrators need to be involved from the onset.

But it is important to consider the less obvious players like the information technology department. Startup companies sometimes expect an unreasonable amount of man-hours from the IT department to deploy their technology, which takes time away from other priorities, said Fenty.

Most importantly, hospitals need to figure out the best metrics for measuring a pilot project's success. These metrics are not always obvious because the field is so new. Until recently many hospitals didn't track what happened to patients once they left the premises.

Reducing readmission rates and preventing possible complications are key reasons for hospitals to engage and monitor patients during the post discharge period. But these can be murky indicators.

“How do you attribute a decline in readmission rates to a specific intervention? It might have had no effect, it may have been the sum total or somewhere in between,” Brown said.

Indeed, many academic researchers are still trying to pinpoint the key variables that cause readmission. Hospitals should look at more quantifiable factors like engagement rates, possibly broken down by age group since there is some concern that older people won't respond well to apps.

Engagement rates can also be broken down by length of time to see if patients interact with the app through the entire episode of care. Or it can be broken down by specific task. Did the patient use the app to send messages, watch videos, complete daily check-ins?

Hospitals are also investigating the possibility of using apps to gather patient-reported outcomes. These surveys are a component of a hospital's overall quality score, which helps determine if they receive a bonus for meeting the quality standards set by Medicare.

And it's not just about Medicare's penalties and rewards. A hospital's reputation ultimately is at stake. “We want to know what percent of our patients are thrilled with the experience. Then how many post public reviews online,” said Jodi Rosen, director of innovation at Northwestern Memorial HealthCare in Chicago.

Other key metrics include: reductions in-bound phone calls to specialty clinics or providers and the ability for nurse navigators to see more patients.

A big part of setting up a pilot is usually defining how it can be expanded into other departments or across a system. This can mean thinking about cost or logistics of expansion before the technology is deployed.

While claiming to handle every condition may be a selling point for would-be vendors, there are obvious situations where different methodologies are needed because they involve different types of patients. For example, an interface designed for a patient undergoing an elective surgery may be very different than one designed for someone who is chronically ill.

For now, even a successful pilot won't offer a clear path to a systemwide technical solution to post-discharge monitoring. “It's too early to pick winners and losers,” Brown said.

Lisa Ward is a freelance writer based in Mendham, N.J.