Slow cooking works best when installing new health technologies

In March 2005, a Washington Post report called Los Angeles based Cedars-Sinai Health System's  implementation of electronic medical records a “spectacular failure.”

The new software had moved physicians from jotting notes on a clipboard to logging onto a computer to type in their treatment and medication orders.

“The technology, created in-house, was clunky and slow,” the article said. “Only a fraction of the 2,000 doctors with privileges at the hospital were involved in developing the system, even though they faced a dramatic change in the way they practiced medicine.”

The rollout was flawed, the report warned, because hospital administrators opted for a "'big bang' implementation rather than switching one ward at a time.”

Cedars-Sinai was a very early adopter of electronic health records. Four years after the debacle, only 10% of the hospitals in the U.S. had implemented the EHR systems, according to a study published in the New England Journal of Medicine. They were not the only system to make this mistake. England's National Health Service also had a large centralized rollout, which failed.

“Cedars-Sinai lost about $100 million, while England's National Health Service lost about $10 billion,” said Dr. Robert M. Wachter, a professor and chair of the medicine department at the University of California at San Francisco and author of The Digital Doctor: Hope, Hype and Harm at the Dawn of Medicine's Computer Age.

But their missteps helped the healthcare industry develop a more nuanced understanding of the implementation process. In the decade that has followed, these lessons have been applied not only to EHRs, but to other computer-based technologies.

“The biggest error in healthcare IT was believing it was about technological change only,” Wachter said, adding, this misguided assumption led Cedars-Sinai and other early adopters to underestimate the extent to which new technologies could disrupt workflow and require ongoing support.

“There is a productivity paradox,” he said, explaining departments often have to live with the new technology before they can see increases in productivity.

Cedars-Sinai has learned from its fiasco. Today it is considered one of the most-wired hospitals in the country, according to a survey by Hospitals & Health Networks, a journal of the American Hospital Association. But its approach to innovation changed significantly because of that experience.

Cedars-Sinai now consciously collaborates with people in-house to try to anticipate a new technology's effect on workflow and then make modifications when unexpected hurdles are encountered. It also closely collaborates with other health systems to amplify its return on investment. Technological rollouts are incremental and structured, so lessons learned in one department or in a different system can be broadly applied.

“We thought our needs were unique and special,” said Darren Dworkin, chief information officer at Cedars-Sinai Medical Center. In the aftermath of the fiasco, the health system found itself “caught between two goal posts,” not wanting to build its own software, but not wanting to wait for vendors to fill in the white space.

Now Cedars-Sinai works closely with other healthcare systems and tech companies to fill the gaps when there is a need but no market solution. The approach, which is increasingly common, allows it to share risk and bring small companies with new ideas into the fold. Perhaps most importantly, it allows the system to identify new business models that could be potentially disruptive.

Cedars-Sinai identifies startups with potential in part by investing in early-stage companies through its Tech Stars Healthcare Accelerator, which houses startup firms. Cedars-Sinai and Fountain Valley, Calif.-based Memorial Care Health System also run an $80 million Summation Health Ventures, a venture capital arm for investing in early to mid-staged companies.

One of the major lessons learned at Cedars-Sinai was the need to bring doctors and other key staff members into the decision-making process immediately, since any new technology was bound to affect their daily work routine.

For instance, Cedars-Sinai wanted to end noisy overhead paging and frustrating bouts of telephone tag among doctors, nurses and other personnel. Initially it supplied clinicians with cell phones, but the cellular service was expensive to maintain.

About five years ago the system began using smartphones for its clinical teams and turned to Voalte, an app that could be used to make internet calls and send secure texts. Still, the expectation was that VoIP (Voice over internet Protocol) would be the most popular feature.

It was not. “Shortly after going live we learned what most 12-year-olds already know – texting is so much easier than a call,” Dworkin said, adding that within a short time the number of texts far outpaced calls.

While it identified a solution, Cedars-Sinai held off on implementation. It liked Voalte's app, but it wasn't always reliable.

They passed after an initial test, giving feedback to the developers. The company's engineers then improved the product. Cedars-Sinai retested it and moved ahead with the deployment.

Voalte is now widely used throughout the health system.

The Voalte experience taught Cedars-Sinai to favor incremental rollouts, which allowed staff to test and modify products. Its ongoing rollout of HealthLoop is an example of that strategy.

HealthLoop's software helps health organizations monitor and engage patients. It was initially developed for orthopedic surgery and has been used for other scheduled surgeries.

Patients receive a daily text or email from their doctor, asking them to “check in” and linking to HealthLoop's website via home computer or mobile device. Patients answer several questions designed to assess their health and instruct them on what to do in the days leading up to surgery. It can also communicate and reinforce discharge instructions.

But HealthLoop's potential for unplanned hospitalizations with less predictable trajectories was unclear. The app's appealing features, including the ability to help collect patient-reported outcomes, led system officials to immediately think of large-scale deployment. But with the lessons from the EHR rollout in mind, they decided to slowly test it in different departments.

“We couldn't put all diabetics on a loop,” Dworkin said, adding that it could be useful for a certain subset of patients: new diabetics using insulin for the first time. “There might be situations where patients could come in and off the loop.”

Rather than flipping a switch, as was the case with the medical records, Dworkin now realizes that “go-live is a not a single point in time.”

Part 2 of Scaling Up looks at how Canada's St. Elizabeth, a not-for-profit delivering in-home healthcare, developed technology to provide desperately needed guidance for family caregivers.

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