From algorithm to evigraph: A knowledge translation tool

As a strategy to aid in knowledge translation, the American Occupational Therapy Association (AOTA) began using clinical algorithms in Practice Guidelines in 2022. This article provides an overview of the function of algorithms in clinical care, the unique challenges to developing traditional clinical algorithms to inform occupational therapy practice, and an overview of AOTA’s development of a new type of knowledge translation tool called an evigraph (evidence graphic).

Clinical algorithms are flowcharts that guide practitioners through the logical process of decision making via sets of clinical decision points (Sox & Stewart, 2015). Based on evidence and/or expert opinion, algorithms serve as tools for decision making in clinical and educational environments. Some health care systems have incorporated artificial intelligence (AI) algorithms into electronic health records (EHRs) to guide practice and improve health care quality; however, issues of bias have been identified in some existing AI algorithms (Obermeyer et al., 2019). For the purposes of this article, the discussion of algorithms refers only to decision-making flowcharts, not to AI components. Awareness of the biases in existing health care algorithms did, however, inform the development process for AOTA’s algorithms, with careful attention to issues of equity and occupational justice.

Development and Evolution of AOTA’s Algorithms

In 2019, members of AOTA’s Quality and Evidence-Based Practice (EBP) teams began work to understand how algorithms might be developed to support evidence-based occupational therapy practice. Unlike algorithms used by other health care disciplines to determine medical treatment or medication regimens based on clinical presentation factors and pathways clearly defined in the research, we anticipated that OT algorithms might require a different approach. The breadth and interplay of factors within the domain of occupational therapy, as well as the variability of treatment modalities by setting and the available evidence presented unique challenges. As stated in the fourth edition of the Occupational Therapy Practice Framework: Domain and Process (OTPF-4; AOTA, 2020), “All aspects of the domain have a dynamic interrelatedness …occupational therapists are skilled in evaluating all aspects of the domain, the interrelationships among the aspects, and the client within context” (p. 6).

As an initial trial, the EBP team developed several prototypes of an algorithm based on clinical recommendations for fall prevention from the Occupational Therapy Practice Guideline for Productive Aging for Community-Dwelling Older Adults (Smallfield et al., 2019). The process for development of the initial algorithm (see Figure 1) involved examining the clinical recommendations to identify potential decision points and areas of the occupational therapy domain, as well as considerations that would affect intervention selection. The algorithm prototypes were drafted and reviewed within the EBP and Quality teams, and they chose a final version for review by falls prevention clinical content experts. Reviewers were asked to provide feedback on the flow of the algorithm and its consistency with practice, as well as the potential to inform clinical practice. The prototype was met with positive feedback regarding the applicability to clinical practice and the potential for clinicians to use the algorithms to quickly understand and utilize evidence, as well as recommendations for visual representation of the information. With support from the content experts and staff across AOTA, the EBP team started incorporating algorithms into clinical practice guideline development in 2021.

The first guidelines to include algorithms were the Occupational Therapy Practice Guidelines for Adults With Chronic Conditions (Fields & Smallfield, 2022), Occupational Therapy Practice Guidelines for People With Parkinson’s Disease (Wood et al., 2022), and the Occupational Therapy Practice Guidelines for Adults With Multiple Sclerosis (Cunningham et al., 2022). With collaboration from members of the EBP team, the Practice Guideline authors drafted algorithms illustrating the clinical decision-making process in individual case studies (see Figure 2). Focusing on interventions within case studies allowed the authors to demonstrate the decision-making processes as they might occur in clinical practice.

To expand the usability of the algorithms in the Practice Guidelines beyond demonstrating evidence application in the case studies, the EBP team determined that the next published guidelines should include algorithms with a broader representation of the clinical recommendations. Multiple iterations were drafted in the established algorithm format, but due to challenges in fully representing the breadth of evidence in relation to potential clinical scenarios and the interplay of factors within the domain of occupational therapy, it became clear that a new strategy or format was necessary.

Updates to two of AOTA’s Practice Guidelines are currently in progress. At the outset of these projects, the EBP team trialed new formats beyond the algorithms to best represent the evidence and make the clinical recommendations and related considerations clear and easy to access for busy practitioners. Through multiple trials and feedback from Practice Guideline authors, the team developed a hybrid of an infographic and algorithm, with presentations of the clinical recommendations, citations for related studies, client factors and areas for intervention (i.e., areas of occupational difficulty or goals), and considerations for selecting interventions.

The EBP team reviewed the literature to determine whether similar types of knowledge translation tools were in use, to accurately identify the type of resource. No comparable tools were identified. To distinguish the new type of tool from traditional algorithms, the team developed the term evigraph (evidence graphic) to represent the hybrid infographic/algorithm. The first evigraphs are scheduled for publication in the next year, as part of AOTA’s Practice Guideline series. Figure 3 provides a pre-publication example.

Application of Evigraphs and Next Steps

The evigraphs will allow practitioners to use the combination of the individual client factors and preferences identified in the occupational profile in conjunction with client-centered goals to determine the range of evidence-based intervention options that are most likely to help clients meet their goals. When using the evigraphs, practitioners should refer to the clinical recommendations table and identified studies, along with incorporating their clinical judgment and the client’s context and preferences, to fully inform decisions.

Although the components of the evigraphs will remain consistent, users should note that the format in individual Practice Guidelines may differ depending on the design of the systematic reviews that informed each guideline. This flexibility ensures that the evigraphs provide useful translation information that is consistent with the clinical recommendations, in order to remain evidence based. Stepping away from the algorithm model also allows a greater breadth of application. Multiple evigraphs can be used in conjunction to develop the plan of care, and they support shared decision making by providing a resource that can be used directly with clients to inform them about research findings and effective intervention options.

One of AOTA’s primary purposes is to provide occupational therapy practitioners with reliable, applicable, evidence-based practice resources. After publication of the new evigraph format in the 2023 Practice Guidelines, AOTA staff members will assess and incorporate feedback in a continuous quality improvement process. Translating research into practice is essential for best practice in client care, and it is our intention that developing a tool based on the unique needs of the occupational therapy process will provide another avenue for knowledge translation.

References

American Occupational Therapy Association. (2020). Occupational therapy practice framework: Domain and process (4th ed.). American Journal of Occupational Therapy, 74(Suppl. 2), 7412410010. https://doi.org/10.5014/ajot.2020.74S2001

Cunningham, R., Uyeshiro Simon, A., & Preissner, K. (2022). Occupational therapy practice guidelines for adults with multiple sclerosis. American Journal of Occupational Therapy, 76, 7605397010. https://doi.org/10.5014/ajot.2022.050088

Fields, B., & Smallfield, S. (2022). Occupational therapy practice guidelines for adults with chronic conditions. American Journal of Occupational Therapy, 76, 7602397010. https://doi.org/10.5014/ajot.2022/762001
Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366, 447–453. https://doi.org/10.1126/science.aax2342

Smallfield, S., Elliot, S., & Leland, N. E. (2019). Occupational therapy practice guidelines for productive aging for community-dwelling older adults. AOTA Press.

Sox, H., & Stewart, W. (2015). Algorithms, clinical practice guidelines, and standardized clinical assessment and management plans: Evidence-based patient management standards in evolution. Academic Medicine, 90(2), 129–132. https://doi.org/10.1097/ACM.0000000000000509

Wood, J., Henderson, W., & Foster, E. R. (2022). Occupational therapy practice guidelines for people with Parkinson’s disease. American Journal of Occupational Therapy, 76, 7603397010. https://doi.org/10.5014/ajot.2022.763001

Hillary Richardson, MOT, OTR/L, DipACLM, is Practice Manager, Knowledge Translation, Evidence-based Practice, and Practice Improvement at AOTA.