Automation’s Promise: The 3rd and 4th Industrial Revolutions
The global healthcare automation market valuation is projected to reach $58.98B by the end of 2025, with North America being the dominant regional market. According to the Council for Affordable Quality Healthcare’s 2020 Index (CAQH), the United States healthcare system has saved $122B via automation. It is poised to save another $16.3B from workflow automation alone.1
Whether workflow, data, or decision automation, this growth trajectory is fueled by the momentum of realized value, a regulatory environment2 that is promoting interoperability, reduced information blocking, and technology’s evolution. These factors propel healthcare to transform not simply based on operational efficiencies but more significantly, based on availability, purpose, and effective use of massive volumes of real-time data.
Automation’s Current State
As we healthcare professionals march into this Fourth Industrial Revolution,3 we still grapple with the Third Revolution: fully leveraging computing power for core operational processes. Currently popular, robotic process automation (RPA) is accessible and affordable process automation technology that is often misapplied and sees high failure rates.
It comes on the heels of other graphical forms of automation (e.g., data automation platforms), as well as configured and custom software platforms and rules engines. These are often large, costly, and unwieldy initiatives. Data-intensive automation that relies on artificial intelligence (AI) and machine learning (ML) lies ahead with the possibility of human-like interaction and high purpose. These can be fraught with bias, privacy, security, and safety challenges.
New process automation projects have an initial failure rate of 30-50%.Ernst & Young
To beat the odds, we need to apply the learnings from previous experiences to each new project. Here are the six most common pitfalls to avoid.
Misaligned, Isolated, or Low-Value Improvements
Organizations can create long lists of candidates for automation. However, many do not clearly evaluate and prioritize candidates based on value. Alternatively, they focus solely on cost reduction without considering other business-critical drivers: cycle time (speed and throughput), quality and accuracy (and associated rework), and capacity.
We recommend aligning business and technical stakeholders around the rationale for the automation investment. Creating explicit problem statements, including size and frequency of the situation to be automated, why it matters, and what quantifiable benefit the business will achieve.
“If I had an hour to solve a problem, I’d spend 55 minutes thinking about the problem and five minutes thinking about solutions.”Albert Einstein
Underlying Process to be Automated is Functioning Poorly
In the excitement of a new automation effort, teams may fail to verify that the underlying process is correct and yields the intended result. Complicating matters, personnel continue to accomplish work, despite a poor process, via workarounds, rework, and other inefficient strategies. Automation that replicates this churn is unnecessarily costly to create and maintain. Furthermore, it misses “hidden” workarounds, yielding unacceptable outcomes and risking abandonment. Broken processes should be fixed prior to automation, which can be more challenging than the automation itself.
“The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.”Bill Gates
We recommend observing the process as performed and looking for the following clues:
- Extensive workarounds and personal processes
- Process details aren’t document but in peoples’ heads
- Decision-making during the process is “invisible”
- Lack of process understanding – entrance/exit criteria, tasks, validation points
Underlying Process has Unrecognized Complexity and Interdependencies
Multiple dimensions of complexity may complicate the project: process branching, decisions, interfaces, and handoffs. Additionally, upstream/downstream systems and data interdependencies are notoriously difficult to resolve. When this complexity is not well understood, and accounted for somehow, the overall solution will break. The only question is how soon and how significant? When complexity is understood and managed well, automation’s benefits can increase significantly.
Healthcare example: Prior Authorization is an example of a process with downstream implications on claims denials.
Errors in the prior authorization phase of the revenue cycle may account for nearly 24% of all claim denials.HFMA
We recommend mapping the system and data flows with an emphasis on finding clarity on boundaries, key process steps, any handoffs, and interfaces. In some scenarios, conducting a Failure Modes, Effects and Criticality Analysis will be useful to prioritize breakpoints and determine the appropriate solution.
Impact of Future Business Growth is Not Considered
Change in today’s business world is guaranteed. And while improved accessibility and affordability lessen the burden of adapting an automation solution, it is undesirable to repair or replace a solution frequently. It is costly and introduces an opportunity for unintentional errors.
We recommend a better way to design a more robust and resilient solution. Join forces between business and technical stakeholders to ensure the automation team understands expected business changes. Consider using scenario planning, and create a roadmap for managing your solution’s evolution.
Healthcare example: Adding or expanding government lines of business expands the scope and complexity of enrollment and eligibility rules engines.
Automating with Inappropriate Tool or Technology
Many teams start with a pre-ordained automation tool or technology; however, that can be premature. Form should follow function. Custom development has its place. From business rules engines to configured platforms and tools (including workflow automation, data automation, and business process automation such as RPA) to advanced technologies (e.g., artificial intelligence, cognitive computing, machine learning) each has a potential role to play.
“50% of RPA implementations will fail to deliver a sustainable return on investment without being combined with other solutions.”Gartner
We recommend selecting technologies that will accomplish the immediate purpose and reasonably grow with the business. Choose the best-suited technology by considering factors like process scope and span, process complexity, operational dynamics, anticipated change, and systems interfaces.
Automating without an Adoption Plan
Automation is oft-perceived as a cost- and workforce-reduction initiative, even in scenarios where an alternative imperative is clear, such as freeing capacity for growth or reducing backlogs and stress. Furthermore, automation may be siloed as “technology projects.” Teams often bypass cultural and workforce needs, such as communications about the value and “what’s in it for me” and (re)training.
We recommend that automation teams invest in understanding user needs and the impact of their solutions on individuals. Teams should ensure that iterative discovery, design, and feedback loops are working. Consider engaging users with the solution in the spirit of proven quality principles4. Include organization change management experts as part of the project plan to ensure that all cultural circumstances are addressed.
“Automation is good, so long as you know exactly where to put the machine.”Eliyahu Godratt
Aligning Automation Value, Tool, Process and People
Success in healthcare operational efficiency is about more than choosing the most accessible automation tool. Achieving your desired outcomes, and scaling successfully, requires a thoughtful design and approach. We thrive on removing friction from healthcare, simplifying and de-risking operational processes, systems, and data that transform the organization while satisfying consumers and engaging employees in the meaningful work of healthcare.
1 For the specific subset of processes that CAQH evaluates. CAQH estimates the potential savings per patient encounter to be $43.39, across payer and provider
2Regulations include 21st Century Cures Act; HHS Office of the National Coordinator for Health Information Technology (ONC) and CMS data interoperability rules for APIs and FHIR standards
3Principles of Industry 4.0 include interconnectedness (healthcare translation: internet of medical things + internet of people); transparency (healthcare translation: interoperability); tech-supported decisioning & cognitive computing (healthcare translation: AI/ML assists); decentralized decisioning (by machines; which needs to be diligently tempered with safety in healthcare applications)
4Deming quality circles and Lean Gemba walks are both proven practices of engaging personnel in improvement endeavors.