Priorities to accelerate workflow automation in health care

Abstract

Inefficient workflows affect many health care stakeholders including patients, caregivers, clinicians, and staff. Widespread health information technology adoption and modern computing provide opportunities for more efficient health care workflows through automation. The Office of the National Coordinator for Health Information Technology (ONC) led a multidisciplinary effort with stakeholders across health care and experts in industrial engineering, computer science, and finance to explore opportunities for automation in health care. The effort included semistructured key informant interviews, a review of relevant literature, and a workshop to understand automation lessons across nonhealth care industries that could be applied to health care. In this article, we describe considerations for advancing workflow automation in health care that were identified through these activities. We also discuss a set of six priorities and related strategies developed through the ONC-led effort and highlight the role the informatics and research communities have in advancing each priority and the strategies.

WORKFLOW, AUTOMATION, AND 21ST CENTURY HEALTH CARE DELIVERY

Health care in the 21st century includes a combination of complex tasks¹ and processing an ever-expanding amount of data.^2–5 Health care delivery involves a series of interconnected clinical, administrative, and population-level workflows, or “the sequence of physical and mental tasks performed by various people within and between work environments,”⁶ that involve patients, caregivers, clinicians, and staff. “Digitized” paper-based workflows that simply copy how a paper-based workflow is performed have led to an ecosystem that contributes to burnout^7,8 and impedes the full use of technology to optimize patient care through the use of automation.^9,10 Inefficient workflows are a pervasive problem that affect everyone in health care,^11–14 including clinicians facing burnout due to managing care delivery tasks,¹⁵ and patients and caregivers facing complex care management tasks.¹

The increased adoption and use of health information technology (IT)^16,17 and the availability of modern computational technology^18,19 provide new opportunities for more effective and efficient workflows through automation. In particular, automation, or “the creation and application of technology to monitor and control the delivery of products and services,”²⁰ can improve efficiency in health care delivery in the United States across several health care domains as Figure 1 illustrates. Automation integration into daily workflows in health care has not been the same as in other industries.^21,22 Lessons learned from the use of automation in these nonhealth care industries may offer key insights for health care.²¹

Figure 1.

Health care domains that can be automated.

As the lead health IT policy agency in the United States,²³ the Office of the National Coordinator for Health Information Technology (ONC) recently led a project to identify priorities that could accelerate health care workflow automation through the use of health IT and modern computing.²⁴ A multistep approach was taken including conducting semistructured interviews with key automation experts across multiple industries and analyzing peer-reviewed and gray literature that included sources both within and outside health care.²¹ A report resulting from these activities identified automation opportunities, approaches, and barriers and facilitators and informed a multidisciplinary workshop that brought together experts from health care, including clinicians, patient advocates, health IT developers, and payors; industrial engineering; computer science; and finance.²² This diverse group discussed automation approaches, workflows to automate, and relevant market and policy levers.^22,25 ONC presented a summary of project activities and preliminary observations to the American Medical Informatics Association (AMIA) Public Policy Committee and at the 2021 ONC Annual Meeting,²⁶ and further refined project findings based on audience feedback. Draft priorities and strategies were shared for review with workshop participants. Through these activities, we elucidated considerations for effective design, implementation, and use of automation in health care. This article summarizes the considerations along with the resulting priorities and related strategies identified through this effort.

CONSIDERATIONS FOR ADVANCING HEALTH CARE WORKFLOW AUTOMATION

Barriers and facilitators for workflow automation identified through the key informant interviews, literature review,^21,22 and workshop²⁵ were used to elucidate the types of considerations necessary for automation to be effective in health care. While not meant to be exhaustive, the considerations summarized in Figure 2 can aid health care stakeholders including patients, caregivers, clinicians, staff, technology developers, researchers, provider organizations, payors, policymakers, and accrediting bodies, successfully address or engage in the myriad design, implementation, and use factors posed by automating workflows in health care.

Figure 2.

Considerations for advancing health care workflow automation.

Leverage high-quality data

Access to high-quality data is one of the most critical considerations when seeking to automate solutions for health care workflows.^27–29 The quality of data used to support automation can have a significant effect on user (eg, patient, caregiver, clinician, staff) experience, safety, and outcomes. The data and metadata used to power automation solutions may include a range of administrative, cost, frequency, duration, clinical, and outcome data. It is also important to understand data’s limitations to mitigate potential issues regarding completeness, consistency, reliability, and accuracy. Use of industry consensus standards may improve the quality of the data used to support automation. When paired with evidence-based workflows, such standards can reduce the need for ad hoc mapping, improve the reliability and repeatability of the automated task, and enable reuse of automation approaches.^30,31

Understand relevant workflows

Introducing automation introduces change. Whether it is one task or an entire health care workflow, automation will change the way that work is performed. Developing automation solutions and measuring their effects hinges on how well understood the workflow and its intended goals are. The tasks or steps, technology, organization and context, and individuals involved in the workflow under consideration for automation all need to be evaluated and accounted for in the proposed solution.^32–34

Build trust through effective design and implementation

Securing buy-in and trust among relevant stakeholders is essential to the successful design and implementation of automation solutions.^35,36 For workflow automation solutions to be implemented, adopted, and used, those most affected (eg, patients, caregivers, clinicians, or staff) need to be invested in the change. A robust, systematic process with fully engaged stakeholders, including direct users (eg, patients or clinicians) and others impacted by a given workflow (eg, organizational leadership) is necessary to identify automation needs and the anticipated benefits. Clear communication is vital to establish discrete, commonly understood goals as well as to explain what automation is meant to improve upon, how it is expected to work, and what it will and will not replace. Poor communication and user experiences, including from a lack of well-designed training, can quickly erode trust because the intended improvements from automation are blurred by frustration and diminished enthusiasm for change. This is why automation solutions for health care workflows must be accurate, reliable, and not create new risks, including those to patient safety.

Automate to add value, not burden

Ensuring success when it comes to introducing automation to health care workflows should be driven by a commitment to deliver high-quality, equitable care. When seeking to add automation to health care workflows, an important factor to consider is whether its introduction can add value, increase safety, and/or improve efficiency.³⁷ This includes ensuring automation will serve all patient populations, and not create biases or advantage a particular population. In addition, seeking to introduce automation to a health care workflow often provides opportunities to redesign the workflow as a whole to support a re-envisioned workflow with automation.

Conduct ongoing testing and evaluation

Testing and evaluation will be paramount, prior to, during, and after implementation.³⁸ Automation efforts must contain feedback loops that allow for continuous monitoring of the automation technology itself and its impact. This should facilitate and enhance understanding as to whether the automation is working and performing as intended, measuring its impact, and determining whether it has led to improvements and met automation goals. Implementing automation needs to lead to improvements across one or more dimensions in the health care delivery ecosystem, and automation efforts must be designed to measure improvements.

The aforementioned considerations can help support a path forward for advancing workflow automation in health care along with clearly articulated priorities and relevant strategies.

PRIORITIES AND STRATEGIES TO ADVANCE HEALTH CARE WORKFLOW AUTOMATION

Findings from across the three project activities (1) interviews,²² (2) literature review,^21,22 and (3) workshop²⁵ combined with external feedback obtained through presentations and review were used to develop six priorities and six related strategies that can help accelerate workflow automation in health care through health IT and modern computing.³⁹

The priorities, summarized in Table 1, address an ecosystem of organizations and individual stakeholders—including patients, caregivers, clinicians, and staff—and the technology that must collectively support advancement in health care workflow automation. The priorities inform near-term actions to enable automation and guide long-term planning to reach the desired end-state.

Table 1.

Health care workflow automation priorities, desired end state, and goals

Priorities	Desired end state	Goals
1. Mobilize nationwide scalable automation in near-term “sprints” and long-term “marathons”.	• The health care industry aligns to pursue automation for widespread gains.	Identify smaller, feasible automation opportunities that can be implemented rapidly where numerous stakeholders benefit in the near-term. Plan longer-term efforts that fully automate entire processes. Use technologies that easily can be plugged into an electronic ecosystem, regardless of the size, type, and sophistication of an organization’s technology.
2. Enable discovery of redundant tasks.	• Organizations independently identify automation opportunities.	Create an operational routine and culture of identifying workflows to automate across health care organizations. Promote a workforce with the skills needed to analyze existing workflows and data, design and implement automation, and conduct ongoing monitoring and optimization. Scale automation initiatives across health care organizations in a manner that supports organizations of different sizes, resources, and patient population.
3. Ensure a ready clinician base for workflow automation.	• Clinicians embrace automation to relieve workflow burden.	Reinvigorate the joy of practice by designing automation that supports cognitive tasks and patient interaction. Improve situational awareness for care management by identifying and filling data blind spots to provide person-centered care. Uphold strict and high levels of data security and patient safety. Streamline data collection, review, and translation for patient care and population health.
4. Enable all stakeholders to effectively and efficiently engage in health and health care tasks.	• Individuals’ time shifts from repetitive to cognitive tasks of managing and supporting care.	Simplify redundant data entry at multiple points within and across health care encounters. Eliminate “hunting and gathering” data across multiple platforms by enabling easy access to information for use in decision making. Present complete, accurate data for workflows through tools that streamline care management. Improve feedback and status monitoring across the continuum of care and across settings.
5. Improve patient and caregiver interactions with health and health care.	• Patients and caregivers have a more seamless and integrated experience in health and health care management.	Streamline information and data sharing across patients, caregivers, clinicians, provider and payor staff, and other stakeholders involved in health care workflows. Simplify administrative tasks such as appointments, registration, portals, and applications for health data access. Empower patients to manage health by facilitating faster access to needed information. Improve care access and outcomes by reducing barriers, allowing care to be managed anytime and anywhere.
6. Leverage interoperable health data for automation.	• Data shared across disparate technologies deliver insights and efficiencies.	Streamline health information exchange workflows. Improve situational awareness for care management by finding, parsing, and presenting data across multiple systems. Reduce use of paper forms and processes that require multiple inputs of clinical and administrative information. Integrate clinician and patient workflows by bringing information, communication, and care into harmony.

Priority 1. Mobilize nationwide scalable automation in near-term “sprints” and long-term “marathons”

As with any change, the adoption of workflow automation in health care will benefit from both incremental activities that demonstrate short-term gains and help build buy-in within the community—while creating more transformative, long-term automation efforts. Establishing an environment that leverages and benefits from workflow automation should be approached in iterative phases moving toward the desired end state. The informatics community is already contributing to development of relevant tools and methods (eg, natural language processing, machine learning). Informatics research is needed to rigorously identify “automation-sensitive” workflows and build the evidence base for exploring broader solutions at scale.

Priority 2. Enable discovery of redundant tasks

Making greater use of automation in health care and identifying automation opportunities will require analyzing the infrastructure, processes, and systems that health care organizations use to highlight tasks that may be redundant across staff and operations. Systematic analyses of workflows through established techniques⁴⁰ and technological assessments of existing infrastructure and resources are needed to produce a comprehensive list of tasks and level of readiness and complexity for automation. Informatics approaches and tools that leverage data from health IT systems and can be easily used across health care organizations are needed to help identify redundant workflows. Research and demonstration projects on such approaches and tools are needed to demonstrate how they would work at scale.

Priority 3. Ensure a ready clinician base for workflow automation

Clinicians face new technology, regulatory mandates, and business drivers while trying to provide the best possible care for their patients.^7,8,41,42 Understandably, technology changes can be viewed as adding to existing responsibilities and raise concerns (eg, safety) and, therefore, workflow automation should involve clinicians throughout the process to increase acceptance. Informatics methods and research regarding the design and implementation of health IT^43–45 will be key to advancing workflow automation.

Priority 4. Enable all stakeholders to effectively and efficiently engage in health and health care tasks

Over the years, health care has benefited from technological advances in diagnosis, treatment, and self-management.^46–48 More recently, implementation and use of health IT provided clinicians, staff, patients, and caregivers access to additional health information and care-related electronic functionality.^49,50 Automation solutions should aim to build on lessons learned from health IT and health technology implementation^51,52 and aim to remove redundancy from professionals’, patients’, and caregivers’ work. The informatics research community can help advance development and testing of automation solutions that reduce administrative burden for clinicians, patients, and caregivers.

Priority 5. Improve patient and caregiver interactions with health and health care

Patients and caregivers continue to face a complex web of health care services and interactions that require active care and information management on their part all while juggling care coordination and information sharing with providers, payors, employers, and others. Design and implementation of workflow automation solutions should broadly aim to improve patients’ experiences and care management as well as support their caregivers. Automation solutions should support seamless access to health care services and health data, aim to simplify administrative tasks, and make health and health care management less cumbersome. Consumer health informatics research can advance both development and testing of efficacy of relevant solutions.

Priority 6. Leverage interoperable health data for automation

Access to interoperable data is critical for workflow automation. While recent advances in health IT have increased the interoperable use of health information across the industry,^53–55 workflow automation requires availability of standardized data at scale both within individual organizations (eg, clinics, hospitals, payors) and across health care organizations. Such data are needed to help identify automation opportunities and needs, and to develop and evaluate potential solutions. The informatics community has a significant role to play in continuing to improve interoperability by identifying where access to health data are lacking, outlining health data standards needs, and assessing the availability and suitability of health data both for identifying automation needs and informing their design and evaluation.

This effort also identified six strategies that build on each other, summarized in Table 2, to broadly advance health care workflow automation.³⁹ The strategies are meant to support planned and gradual advancements in workflow automation across priorities. Implementing these strategies will require input and collaboration from across health care stakeholders.

Table 2.

Strategies to advance health care workflow automation priorities

Supporting strategies	Strategy focus
Educate	Increase awareness of the need and opportunity to automate, identify, and address barriers to automation.
Convene	Build national dialog on advancing priorities for workflow automation with multiple stakeholders.
Prioritize	Identify and prioritize workflows for automation and appropriate level of automation.
Demonstrate	Test and evaluate automation approaches.
Incent	Design and implement needed policy and market levels to support workflow automation.
Scale	Disseminate effective automation approaches across health care organizations and workflows.

In particular, patients and caregivers must be included and participate in national conversations regarding workflow automation to ensure their needs are reflected in solution design and implementation. Health care professionals and organizations must also be engaged in prioritizing workflows for automation, design, and implementation. The health IT and technology community needs to support instrumentation and technology development that facilitates workflow automation design and evaluation, in coordination with potential users. Researchers are essential to advancing automation priorities through development and evaluation. Payors, policymakers, and accrediting bodies need to create the environment that incents development, testing, and use of automation solutions that deliver value to patients, caregivers, and health care professionals individually as well as the health care system more broadly.

MAXIMIZING AUTOMATION IN THE 21ST CENTURY

The six priorities and related strategies outlined in this article to advance use of automation in health care are based on input from interviews, a literature review, and a multidisciplinary workshop. They articulate a pathway to collaborate across stakeholders and deploy modern technology to increase efficiency, improve health outcomes, and deliver value for patients, caregivers, clinicians, staff, and other stakeholders who support health and health care through workflow automation. There are near-term workflow opportunities where automation can be implemented rapidly to benefit many stakeholders, as well as other workflows that require long-term planning for design and implementation.

This effort identified several workflows ripe for automation based on their attributes (eg, manual, repetitive or frequent, require data entry) across health care domains as shown in Figure 3.²¹ Administrative workflows that would benefit most from automation in the near-term are those tied to prior authorization, clinical documentation, and reimbursement, which have proven to be burdensome for clinicians and/or patients.^56–58 With respect to treatment and care delivery workflows, automation is needed to better support medication reconciliation, intelligent care transitions and data identification, and laboratory results review and communication as these are all areas that require greater data exchange and place higher cognitive loads on clinicians and burden patients and caregivers as a result. Automation has the potential to assist clinicians in identifying patient data from an external network transaction (eg, from health information networks) and aid with patient communication, by automatically generating a tailored, patient-specific summary or explanation (eg, for laboratory results). With regards to population health management, automation may aid practices in implementing guidelines-based approaches to identify newly diagnosed patients and recommended care management protocols. Finally, reporting-related workflows associated with state and/or federally mandated public health reporting or adverse event reporting might benefit from automation that can identify relevant requirements, extract needed data, and automatically generate and, upon approval, submit reports on behalf of clinicians, practices, or provider organizations to appropriate registries, agencies, or systems.

Figure 3.

Sample workflows to automate.

The priorities and strategies constitute a call to action to begin that work as part of the direction set by the 21st Century Cures Act.⁵⁹ Adding automation requires a mind shift across the entire ecosystem, but will help create opportunities across health care from clinical care to public health reporting. This shift may require new workforce, roles, and training as health care implements more automation and different automation approaches.⁶⁰ It will also require automation of tasks that are focused on ensuring safety, improving patient-provider relationships, providing more time for treatment, and yielding better and faster insights about health and health care.

ONC is creating a health IT ecosystem that will enable automation by advancing the necessary access, exchange, and use of electronic health information through policies and programs.^61,62 Through ongoing investment in data standards development and adoption that will help ensure data quality, ONC is facilitating use of those data across multiple systems and applications.^63,64 ONC projects may serve as models for pilot testing of health care automation approaches.⁶⁵ ONC will continue to collaborate with public and private sector stakeholders while it advances policies and technology-related investments to enable workflow automation.

FUNDING

This work was partially funded through U.S. Department of Health and Human Services Contract Number HHSP233201600030I/75P00119F37001, with Clinovations Government + Health, and by the U.S. National Library of Medicine, National Institutes of Health.

AUTHOR CONTRIBUTIONS

TZ-C and THO led the development of the priorities and strategies through a series of coordinated activities which gathered stakeholder input, and SP provided executive leadership and critical appraisal and input to the priorities and strategies. All authors led drafting of the article. All authors revised the article critically, provided intellectual content, and approved the final version for submission. The order of authors listed in the manuscript has been approved by all authors. TZ-C served as Chief Scientist and THO was senior program analyst at the Office of the National Coordinator for Health Information Technology until February 2021 and November 2021, respectively. TZ-C provided executive leadership, and both TZ-C and THO oversaw the overarching project that examined the use of workflow automation in health care.