The final frontier in software testing automation

July 25, 2017

In a modern clinical laboratory, it is safe to assume that there is some form of application software to support and enhance daily operations. It could be a laboratory information system (LIS), one or more middleware products, or software associated with a complex lab instrument. That software—regardless of how much testing the manufacturer has performed—still needs to be validated by the lab for use in its unique environment. And this validation needs to occur with every software update, not just when it is first implemented.

Increasing IT workloads without increasing staff can lead to a lack of adequate personnel to complete this testing in a timely manner. How does the modern clinical laboratory cope? One answer is to automate the software testing—much as the modern laboratory has automated so many other aspects of its clinical workload.

Lean laboratories and automation

We have all worked in, visited, read about, or participated in designing a “Lean lab.” Some of the basic tenets of a Lean operation are elimination of wait time, wasted motion, and errors from a laboratory process. Frequently a batch process is replaced with a single-piece, continuous flow system. It’s no surprise that many times when a laboratory section or process undergoes a Lean assessment, the outcome is a decision to add new automation, or to upgrade existing automation. Automation supports Lean workflow by improving accuracy, increasing productivity, and creating consistency.

Most lab managers probably don’t think of information technology (IT) and software testing as an area that lends itself to Lean assessments, but consider the following:

  • Meetings, overnights, weekends, vacation, sick time—a multi-day test plan waits during all those periods.
  • A person doing testing is frequently disturbed with phone calls, cubicle visitors, meetings, and other distractions that lead to wasted effort as the tester stops work only to have to remember where he or she left off or what was supposed to be done next.
  • Fatigue, boredom and deadlines are reasons that a step or even entire sections of a test plan may be inadvertently skipped. For example, tight deadlines might mean not testing ALL the orderable procedures, which in turn risks missing an incorrect normal range definition or a billing transaction. Discovering these after the software has been moved to live operation may have downstream patient care or revenue impact.
  • In order to achieve 24 hours of software testing in one standard work-day, the lab would need to have three people on three different PCs performing the testing—in other words, a “batch” of testing.

So, what exactly is automated testing, and how can it help the lab to alleviate or even eliminate these issues?

Automated testing: the robot tester

A robust automated testing solution is a software package that runs on the same PC as the application being testing—for example, the LIS. It may be called automated testing, artificial intelligence or robotic testing. If you’ve never seen automated testing in action, it may be difficult to visualize. Think of it as an invisible robot that sits at the PC and emulates a human being’s actions in the application being tested.

The robot is trained to log into the LIS, then use mouse clicks and keyboard entries to perform a series of actions. One simple example would be to place a lab order, result the pending tests, perform an inquiry, and check that the results meet the expected criteria. Do the values displayed match the ones that were entered? Does the display include the correct color coding or flags that indicate critical, normal or abnormal results? If the result is not what is expected, the robot tester will make a note of that for a report and determine if it can continue to the next testing step or that it needs to stop until a correction is made.

Automated testing software also captures screen shots to document the test findings, immediately notifies appropriate personnel of a system error that impedes the testing process, and provides a final report of the testing outcomes.

Since the robot isn’t required at meetings, doesn’t eat lunch, didn’t watch last night’s late game and doesn’t need to sleep, it can test around the clock—even on weekends! With a robot tester laboratories can:

  • Remove the wait as well as the wasted time associated with stopping and re-starting the testing process.
  • Eliminate fatigue as a factor in testing errors.
  • Turn software testing into a continuous flow process by running tests overnight and on weekends.

What the robot cannot do, however, is make the final decision that the way the software is working is appropriate for the lab’s clinical and business needs. That final review and validation of the testing results, and the “OK” to move the software to live production, remains the responsibility of the trained laboratory staff.

Where to begin?

Back in the days before auto-verification (auto-validation) of lab results, before that automated solution was widely accepted and adopted, many lab leaders felt intimidated thinking about how to get started—so they just didn’t. Also, some lab staff were afraid of being “replaced by a computer.” But once auto-verification was in place, a typical staff reaction was “How did we ever get along without it?”

For lab directors who aren’t accustomed to the software world, the very idea of getting started with automated testing may be similarly overwhelming. Rest assured, any vendor that has a solution available has already made a large investment in developing tools that are compatible with the software solution. The robot comes pre-trained (programmed) with the basic knowledge it needs to work in your particular software environment. For example, it is trained to read the lab’s test definitions and find the normal ranges, then follow the instructions to enter results that are above, below, and within those ranges.

After an initial consultation, the company can set up its software to work in any lab environment and begin testing, usually starting within a matter of days. In addition, the company can train staff to use the software to perform other repetitive, rote functions.

What else can a software automation tool do?

If the only thing an automated software testing product could achieve was to assist with testing the LIS upgrades, lab directors might consider putting their investment in an outside consulting company to perform the testing, or budgeting for temporary help to backfill the bench duties while full-time employees are sequestered in the “testing cave.” However, once an investment is made in a robot tester, it can be trained to perform other tasks as well. Examples include:

  • Accession and result every orderable procedure for a billing or EMR interface test.
  • Accession and result several of the same orders on different days to provide lab results for nurse training on lab inquiries in the EMR.
  • Perform annual rules validation for middleware products.
  • Test the LIS configuration for a newly added facility.

Additional uses for the tool turns it into a year-round employee and makes return on investment even faster.

Why invest in automated testing?

We all know that many departments and care providers in the healthcare environment depend on the laboratory to provide timely, accurate results. Indeed many IT projects require laboratory participation to complete their own software testing. What EMR is going to be successful without lab results in it? Should we try ordering and resulting all our lab tests before we say the revenue cycle management system is complete, or will ten or fifteen tests be adequate? What manager wants to be called out as the reason a major enterprise-wide project was delayed, because his or her section didn’t get their testing done on time?

And what laboratory has excess personnel to send off to a testing cave for several days or weeks to meet those deadlines, maybe multiple times per year?

Lab leaders should consider investing in automated testing for the same reasons they invested in a pre-analytic specimen processing system, an automated plate streaker, gel automation, or a digital imaging system: to offload manual labor, increase throughput, reduce turnaround time, achieve consistency, and ultimately improve patient care.

Automation has reached all sections of the clinical laboratory—core lab, microbiology, blood bank and histopath. Software testing and IT may be the last automation frontier the modern lab manager has left to conquer!

Jennifer Lyle is CEO and Founder of Nevada-based Software Testing Solutions, LLC. She founded the company in 1999 to create innovative solutions which automate and accelerate the in-depth testing of healthcare applications.

Photo 241571148 © BiancoBlue |
Photo 75539817 © Vladimirs Prusakovs |
Dreamstime Xxl 75539817
Image by NatalyaBurova @
Coverbackgroundv1 Forstory
Photo 14015956 © Sebastian Czapnik |
Dreamstime Xxl 14015956