Visual App Development Combats Technical Debt in Healthcare

Humana deploys testing app in a single month that would have taken months under previous agile hand-coding methods.

The treacherous world of application development has swallowed many IT budgets, inside and outside of healthcare. For decades, as modern applications evolved, IT shops were restrained by the abilities of teams of developers hand-coding applications, until a new term emerged—"technical debt"—described as the ever-widening gap between software requirements, and the ability of IT shops to fulfill those requirements.

An application development executive at health insurer Humana thinks the payer has discovered a better way, one that draws down the technical debt in the process.

Using a relatively new form of self-documenting visual application development, Humana has been able to reduce application development time, as well as decrease time needed to update those applications, says Bruce Buttles, digital channels director at Humana.

Technical debt doesn't just refer to legacy applications. It can be compounded by previous agile-based software development methods, which place a value of speed of development, but not necessarily documentation or reuse, Buttles says.

Pictured: Bruce Buttles is the digital channels director at Humana. Photo courtesy of Bruce Buttles.

Buttles joined Humana three and a half years ago, having already been a user of one such visual programming system, provided by OutSystems, for two and a half years.

He had been skeptical about the capabilities of visual application development, compared to traditional hand-coding of applications. And yet, when he compared OutSystems to those methods six years ago, it was able to produce the same level of optimized code as hand coders—if not better.

"When I took this job at Humana, I said, 'If you guys are open to this idea, then I would continue to be a consideration for candidacy,' " he said. " 'If you're not open to this idea, tell me now, because I'm not interested.' That's the course of my career. I will never take another job that isn't completely willing to embrace the idea of automated software engineering like this. I just can't do it. I can't go back."

At Humana, Buttles hit the ground running, "The very first solution we built we literally did over a weekend, and launched it," he says. The application replaced an unruly spreadsheet that Humana call centers were using to refer callers to locations offering tests to Humana members. From there, Buttles and his team were able to quickly build master/detail query screens in the app, allowing Humana call center staff to search for testing locations using basic search on a web browser.

The team kept evolving the application, so that after a month of work, it also included a map interface —similar to Google searches—accessible by the general public, as well as Humana staff. "At the height, we had around 12,000 testing locations being tracked and updated," Buttles says.

Traditional development would have taken months

Buttles estimates that a sharp team of hand-programmers would have taken at least three months to deliver the same application—or, more likely, four to six months.

The visual programming system outputs the standard sort of code found in enterprise applications today—Microsoft C# code for the application tier, running on the .Net Framework, as well as standard SQL code. On the front end, it generates standard HTML, CSS, and JavaScript.

Code reuse is also more feasible with the self-documenting system, "especially if you abide by the OutSystems best practice, which is the four-layer canvas. This four-layer canvas breaks things down into four tiers. That makes it very logical," Buttles says.

Recently, OutSystems conducted a survey about technical debt among 521 IT decision-makers spanning enterprises (companies with more than $2.5 billion in revenue), commercial companies (with $1 billion to $2.5 billion in revenue), and small-to-medium businesses (with $50 million to under $1 billion in revenue), inside and outside of the healthcare industry.

On average, responding businesses devote more than a quarter (28%) of IT budgets to addressing technical debt, compared to about a third (33%) to innovating and building new capabilities, and just shy of 40% to running status quo operations, the OutSystems report found.

Buttles says this method of application development also makes much more predictable estimates of how long applications will take to build.

"As a business and technical executive, the hardest part of my job is trying to predict how long is it going to take, how much is it going to cost, are we going to be on time and on budget," he says. "Even the best agile teams in the world deliver every two weeks. On a good day, they're 80% accurate." With the visual system, developers often work fast enough to get ahead of application requirements creation.

Traditional application development also works off of requirements documents that quickly become out of date, as code gets rapidly built and revised. With a self-documenting system, the requirements document is an integral part of the self-documenting whole, Buttles says.

Also feeding the visual development trend is the increasing availability of application programming interfaces (API) in use throughout enterprises, including healthcare, which products such as OutSystems support.

"Some of the best apps we see our customers build come from the use of a couple of different [enterprise] platforms to create a new kind of experience for an end user," says Robson Grieve, chief marketing officer at OutSystems.

One example in healthcare is applications that onboard new employees while provisioning their addition to many different systems in use by the healthcare organization, Grieve says.

Pediatrix Medical Group, part of MEDNAX, used OutSystems to develop BabySteps Cloud, a cloud-based clinical documentation app for material, newborn, and pediatric services, Grieve says. By modernizing the app, MEDNAX was able to push new features every two weeks, rather than years apart, which led to research and best practices that saved the lives of an estimated 100,000 babies, he adds.