A Recipe for Success: Workflow Automation, Business Rules, and Artificial Intelligence

Pat Stricker, RN, MEd
Senior Vice President
TCS Healthcare Technologies

Technology today makes the healthcare industry entirely different from what it was when I started my nursing career. There were no computers to track and monitor the numerous processes in the hospital; no cell phones; no automatic devices for IV drips, respiratory machines, or telemetry; no automatic thermometers or BP devices; no electronic medical records; and no electronic pharmacy ordering/delivery system. Everything was done manually. Surgery schedules, staff schedules, admitting logs, and all other informational documents were typed or handwritten anew each day. Consequently, it took a long time to get anything accomplished, yet we were able to provide excellent care considering what we had to work with. 

Compare that to today when we are used to having everything electronic and at our fingertips.  Processes in the hospital are electronic and automated — computerized records, all types of electronic monitoring devices, order entry for all ancillary services, remote monitoring, computer programs for all aspects of patient care, cell phones for instant communication with others and access to internet information, and computerized logs, reports, and any other type of data that is available instantly in real time. Consequently, things get accomplished very quickly today, and most of these things are even done automatically without us having to intervene.     

The banking and retail industries have surpassed the healthcare industry in automating processes — teller machines (ATMs), automated banking on our computers and cell phones, self-checkout at the grocery store, cellphone-controlled thermostats and home alarm systems, self-driving vehicles, and a variety of other automated processes.The healthcare industry, on the other hand, lags behind these industries, because it is capital and hardware intensive and constrained by numerous safety regulations. This makes automating processes more difficult and leads to slower progress. However, healthcare reforms and increased competition have created the need to focus on increasing efficiency and reducing costs. These are the two main financial priorities cited by three out of four hospital and health system CEOs. In order to accomplish these goals, automating manual tasks and processes has become a key strategy to improve performance and create more time for the staff to devote to higher-level cognitive functions that require human intervention. 

Workplace Automation and Workflows

Let’s start by defining some of the terms used in the automation process:

• Automation:  1) The technique of making an apparatus, process, or system operate automatically by using mechanical or electronic devices that take the place of human labor; 2) a technology, method or system of operating or controlling a process by highly automatic means, as by electronic devices; 3) decreasing human intervention to a minimum; 4) a mechanical device that functions automatically without continuous input from an operator. 
• Workflow:  A defined process involving a series of tasks that must be done by a specific individual(s) in a specific sequence in order to obtain pre-defined results. 

Workplace automation in the past was typically associated with manufacturing. One of the first introductions of workplace automation was done on the assembly line at the Ford Motor Company in 1913 to improve the work process and reduce costs. With this innovation, Ford achieved a dramatic reduction in the time to produce a car, from 12 hours to 1.5 hours! In addition, the number of cars produced was increased with the same number of workers and the workers were happy because they no longer had to perform repetitive, boring tasks. So it was a win for the company and the workers.

Other workplace automation involved the use of robots that took the place of humans. This created a negative perception of workplace automation. This was a concern when automation and robotics were first introduced in the pharmacy. Some feared robots would replace the pharmacists; however, they found automation allowed the them to eliminate counting medications that required little cognitive value and let them focus on more clinically relevant work that was more productive and rewarding.  

Today’s workplace automation is not focused on replacing humans, but rather empowering humans by complementing or augmenting their abilities, in order to allow them to reduce repetitive tasks with little cognitive value and spend more time on meaningful, relevant, higher-level decision-making functions.    

Automation will be even more essential as more of the baby boomer nurses retire. A study in 2009 found that 260,000 registered nurses are projected to retire by 2025 — six years from now. That will be twice as large as any nursing shortage experienced in this country since the mid-1960s. This shortage of RNs will make it imperative that we achieve optimum work efficiency by eliminating redundant, repetitive work and manual tasks. 

Another reason to work toward optimizing automated workflows is to be able to manage population health management (PHM) programs. There are not enough providers to manage these large patient populations, so this makes automation a “must have” rather than a “nice to have.” A report by the Institute for Health Technology Transformation says, “Automation makes population health management feasible, scalable and sustainable.”

Automation has also been shown to reduce repetitive tasks and increase more meaningful tasks. A time-in-motion study was conducted by a Florida hospital to measure nurse work behavior before and a year after the implementation of an electronic health records (EHR) system with clinical documentation. The administration wanted to determine if the move toward automation actually increased the time at the bedside, decreased the time spent on documentation, and decreased time spent on administrative tasks. Results showed a significant increase in the time nurses spent on both direct care and in EHR documentation. However, they found that the increased time spent in both of those areas came from a 12 percent decrease in the time spent on administrative tasks after implementing the automated documentation system. This validated the value of automating tasks. Nurses were able to spend more time on direct care and thoughtful documentation, rather than repetitive administrative tasks.

Automation can be a set of tools within a business software program that performs repetitive, easy-to-replicate tasks without the need for human interaction. Anything that adds value that is done more than once should be considered for automation. In a hospital setting, that may be a routine procedure that is done frequently on a large population, such as joint replacements. Once these repetitive, routine tasks become automated, the process will become easier to adapt the concepts to higher-level tasks.

Business Rules, Business Engines, and Workflow Engines

In order to automate tasks, business rules need to be defined. A business rule is: 1) a policy or procedure that guides conduct or action; 2) a definition or constraint of some aspect of business which always resolves to either true or false; 3) a description of operations and constraints that apply to an organization; 4) a process that provides business structure or controls/influences the behavior of a business; 5) a criteria for decision-making.   

Business rules set expectations and provide guidelines for daily business activities. They also help organizations stay in compliance with local, state, and federal regulations. 

Business rules contain an IF/THEN statement – IF a certain condition exists, THEN a certain action should take place. Examples include: 1) IF a new case is created, THEN send a Welcome Letter; or 2) IF a certain quality issue or risk is identified, THEN send an alert to a certain person or department.     

A Business Rules Engine is a software system that executes one or more business rules (from regulations, company policies, or other sources). The system allows non-programmers to add and change business logic without the intervention of the vendor or IT department. This is a key component, since the department or organization needs to be able to customize the rules to meet their unique workflow needs. The rules can also be applied to data for analysis resulting in process improvement and improved outcomes. 

A Workflow Engine is a key software component that manages and monitors business processes and workflow activities (processing, approving, and determining new activities to transition to, based on defined workflows). It facilitates the flow of information, tasks, and events, allocating tasks to different users while communicating data to other participants. It can execute a number of arbitrary steps and sequences. Examples include: assigning a new case to a clerical group or user based on the type of program the patient was enrolled in vs. assigning a new case with the same type of enrollment to a clinical group or user based on a risk factor that was identified for the patient.     

Workflow engines typically have three main functions:

• Verifying to see if a process is valid depending on the current status.
• Determining if the user is permitted/has the authority to execute the task.
• Executing a task, after verifying the above two conditions are met. If not met, an error report is created and the task (change) is rolled back.

Healthcare software applications have business rules and workflow engines, but some work better and are easier to use than others. If purchasing a software application be sure to ask pointed questions about how the business rules and workflow engine works. Provide the vendor with a case study of one of your most difficult workflow processes and ask them to automate it for you. Also ask them to show you how a non-programmer would create this automated workflow.  

Automation and Artificial Intelligence (AI)

Automation leads to the next big trend in healthcare today — artificial intelligence (AI). There are numerous definitions of AI, but these define its essence: 1) a branch of computer science dealing with simulation of intelligent human behavior in computers, 2) the capability of a computer system to imitate human intelligence (learning, reasoning, and self-correction), 3) a collection of multiple technologies that enable machines to perform administrative and clinical functions, 4) computer systems able to perform tasks normally requiring human intelligence (visual perception, speech recognition, decision-making, and language translation).   

Some examples of AI include: machine learning, natural language processing (machine translation, question answering, and text generation), image recognition, speech to text or text to speech, and robotics. 

AI does not rely on technology that uses algorithms and/or tools to complement a human being. AI attempts to truly augment human activity by imitating and surpassing the abilities of a human. Today, the basic goal of AI is to use human reasoning as a model, not as an end goal of creating a perfect replica of the human mind. We should not be afraid of AI replacing humans, but rather embrace it as a powerful tool that empowers humans to focus on their highest potential. 

Systems using AI are taught to recognize patterns in unstructured data and turn it into structured data that enables automation. AI innovations in electronic health records (EHR), revenue cycle, and operations will continue to increase exponentially over the next few years. AI will be integrated into clinical workflows, empowering providers with real-time data at the point of care.

AI has the power to make improvements in cost, quality and access. Therefore, it is experiencing explosive growth. According to Accenture analysis the health AI market in 2014 was at $600 million, but that is expected to increase to $6.6 billion by 2021 – an amazing eleven-fold growth in just seven years and a compound annual growth rate of 40 percent. However it is expected to increase more than 10 times over the next 5 years and they predict that key clinical healthcare AI applications can potentially save U.S. healthcare $150 billion by 2026.

Automation and AI is a growing trend in all area of business, especially as technology becomes more sophisticated. Previously implementation and new projects required large amounts of time and cost to get completed, thereby limiting the number of projects that could be done and cancelling a lot of innovative new programs. Now with automation and AI, projects will be able to be completed in 25-50 percent less time and with less cost, thereby allowing those resources to be allocated to other needed projects and programs. This increased efficiency, productivity and lower costs will result in better profitability, so this is definitely going to be an key imperative for healthcare organizations.

The Benefits of Automation

• Saves Time and Improves Productivity
• Streamlines Processes
  • Reviewing workflows helps identify and eliminate unneeded or unnecessary steps
•  Improves Efficiency and Throughput
  • Allows system to be scalable and staff to handle more patients

• Improves Reliability and Accuracy
  • Eliminates human error due to heavy workloads and fatigue in the workplace

• Reduces Costs and Improves Profitability
• Improves Quality and Consistency
  • Consistent basis for care activities, medical records, order entry and decision support leads to reduced deaths, non-compliance and costs.

• Increases Predictability of Outcomes
  • Standardized care plans, supported by automation, make it more likely for a patient to follow the plan. Automation can also detect when the patient has deviated from the plan and alert the care team so they can intervene.
• Allows More Time for Human Creativity and Higher Cognitive Functioning
• Provides Ability to Analysis Large Amounts of Data to Support Decision-Making
• Increases Patient Experience
• Improves Project Implementation (Less Tasks to Teach the Users)
• Performance and Program Improvement/Optimization
  • Data from Automation Provides Continuous Feedback That Can Be Used to Increase Performance Over Time

Problems or Issues with Automation

• Unrealistic Expectations
• Poor Design and Testing
• Lack of Time and Attention Required to Maintain Automated Rules
• Technology Problems
• Organizational Problems

Steps to Take to Implement Automation

• Review workflows, policies, and procedures closely to determine what steps should be added, deleted, or changed.
• Look for processes that are routinely done that do not have a defined workflow.
• Look for processes performed on large populations on a routine basis, e.g., joint replacement.
• Think “outside the box” when defining a workflow process. Don’t include unnecessary steps.
• Challenge yourself on every workflow — Do all these steps need to be taken? What can be eliminated? What can be automated?
• Review documentation standards. Look for ways to automate repetitive documentation by providing appropriate options that can be selected in a dropdown. Streamline documentation, if possible, while still making sure it is appropriate.
• Track and identify problems with patients with home devices. Contact immediately to resolve issues and analyze progress over time.
• Look for ways to automate quality and identify risk factors.
• Look for ways to automate scripting, scheduling, reporting, and analysis.
• Make sure any software applications you purchase have business rules and automation tools. They should each department is able to make changes to meet their own needs, not just the vendor or IT department. Learn how to customize workflows within the system.
• Develop a Change Process to document all changes, why they were made, who made them and when. Assure the Change Process has management oversight and approval.
• Develop IF/THEN statements for automation. If “this occurs.” THEN auto-generate “this action.”
• Make automation “part of the culture” within the organization.

Examples of Automated Processes

• Letters: welcome, follow-up, appointment, closure, instructions, education. Trigger a specific letter based on a documentation field (new enrollment, type of enrollment, non-adherence with care plan, need for additional education, etc.). Can also set up a text field that can be pulled into a letter for more personalization. Letters can be set to be delivered by email, text, or mail.
• Identify and enroll candidates for programs based on a diagnosis, number or types of admissions, prescribed medications, treatments, risk factors, etc.
• Provide “Gold-card” service for providers. Allow certain providers automatic approval of procedure (UM) requests.
• Send automated reminder to re-schedule for patients who miss appointments.
• Set up auto-data loads on a routine basis for patient’s labs, pharmacy scripts, or imaging results.
• Identify risk factors and create automated action (call, letter, enrollment, etc.)
• Create admission and discharge automation that may include processes for registration, billing, insurance, providers, CMs, ancillary team, pharmacist, therapists, etc.
• Develop process to find:
  • Patients with risk factors for new programs being offered
  • Candidates for marketing programs
  • Certain providers and analyze patients’ progress
  • Quality issues – lack of compliance, risk factors, falls, etc.
  • Medication non-adherence

It seems inevitable that we are headed towards a future with more automation and AI and that they have the potential to transform the economy at large. However, one thing seems certain — if something can be automated, it will be.

Pat Stricker, RN, MEd, is senior vice president of Clinical Services at TCS Healthcare Technologies. She can be reached at pstricker@tcshealthcare.com.