With funding from a cooperative agreement from the National Institute on Disability and Rehabilitation Research, we have worked for the past 5 years on the development of a rehabilitation treatment taxonomy, a system of classifying all treatments delivered by all rehabilitation disciplines for all diagnostic groups of patients, whatever the setting in which these services are delivered. To date, we have focused on developing a conceptual framework for such a taxonomy, specifying the scope of the taxonomy and the basis for classifying treatments. A recent supplement of the Archives of Physical Medicine and Rehabilitation contained a series of articles setting forth the background for the project, our approach to conceptual issues, and the need to classify treatments based on a theory of how active ingredients bring about change in a clinical target of treatment (some aspect of patient functioning) through a specific mechanism of action.1 It also contains papers on how therapists view classification and lessons learned during a previous effort at classifying learning interventions, as well as commentaries by various scholars who had an opportunity to review these papers. Here we summarize the key points.
The Need for a Rehabilitation Treatment Taxonomy
The field of rehabilitation has made substantial advances in defining and measuring the functional outcomes of the rehabilitation process and the patient characteristics that are associated with those outcomes. However, we lack a rigorous and shared approach to defining, classifying, and measuring the rehabilitation treatments that are hypothesized to moderate the relationships between patient factors and outcomes. At present, rehabilitation treatments may be described as number of hours or days of service by a particular discipline, without regard to what that service actually entails, or simply by referring to the problems they are intended to address (eg, “gait training,” “memory remediation”), without specifying the content or process used to address them.
Without a way to define and measure the active ingredients of treatments—that is, the specific elements of treatments that are hypothesized to carry their effects—we are limited in our ability to discuss the nature of treatment within and across disciplines, to conduct efficacy and comparative effectiveness research, to evaluate clinical practice variations, and to disseminate treatments that are supported by research. Research, training and education, clinical documentation, and cross-disciplinary communication will be advanced by a classification system, or taxonomy, that organizes rehabilitation treatments by their known or hypothesized active ingredients.
Proposed Conceptual Framework for the Rehabilitation Treatment Taxonomy
We propose a conceptual framework to support the development of a rehabilitation treatment taxonomy that can be used by all disciplines. The key features of this conceptual framework include:
A distinction between 2 broad classes of theory that play critical but separate and reciprocal roles in rehabilitation research and practice. Treatment theories articulate how particular ingredients directly alter specific aspects of functioning, but they cannot explain effects on more distal aspects of functioning. In contrast, enablement theories hypothesize how changes in various levels of functioning are interrelated with one another but are silent as to how the initial change is produced.
A model for the tripartite structure of all treatment theories. Treatment theories specify how the ingredients of treatment, operationalized as clinician behaviors or physical entities (eg, energy, chemical compounds) delivered to the patient, affect the target of treatment. The target is some specific and measurable aspect of functioning that is intended to be directly changed by the treatment through a known or hypothesized mechanism of action (Fig. 1).
A proposal to organize a rehabilitation treatment taxonomy around ingredients, as they are linked to targets by their respective mechanisms of action. Categorizing treatments in this fashion allows efficacy research to build treatment theory, rather than simply supporting or refuting the utility of individual treatments one at a time without advancing knowledge about effective mechanisms.
A delineation of broad classes of mutually exclusive treatments that differ from each other in the kinds of ingredients they contain, the mechanisms of action they put into play, and the types of functional targets they can change directly. Defining mutually exclusive categories of treatment ingredients and mechanisms offers a method for decomposing complex interventions and treatment packages into a finite number of components, each of which may be manipulated separately to determine its relative effects. Perhaps more importantly, it requires us to hypothesize mechanisms of action which are distinct from one another and invites further research to clarify potential ambiguities.
A focus on what happens in patient-clinician interactions, as opposed to other aspects of the health care system (eg, structures, coordinating processes), which ultimately affect the patient through what clinicians do with, for, and to the patient (Fig. 2).
Causal and temporal aspects of the tripartite structure of treatment theory. The causal chain runs from left to right: clinician-provided ingredients, through a mechanism of action, bring about changes in a desired target. In clinical reasoning, the sequence is reversed, with the clinician determining an aspect of the patient's functioning that needs changing, deciding on a mechanism of action that might bring about the change, and then selecting the ingredients that are expected to engage that mechanism. Often, the “mechanism of action” step is not explicit in the latter process. Reprinted, with permission from Elsevier, from: Hart T, Tsaousides T, Zanca JM, et al. Toward a theory-driven classification of rehabilitation treatments. Arch Phys Med Rehabil. 2014;95(1 suppl 1):S33–S44.
The focus of the proposed taxonomic scheme: individual treatments and how they fit into the larger rehabilitation program. The focus of the proposed conceptual scheme is on treatments (Rx 1, 2, and 3) and their respective targets (Tgt 1, 2, and 3)—those measureable aspects of functioning that are changed directly by treatment. Changes in those targets, in turn, may contribute to downstream changes in outcomes referred to as “aims.” Selection and ongoing modification of nature, strength, and dosage of treatments is directed by formal and informal assessments of the patient in a feedback loop. Reprinted, with permission from Elsevier, from: Dijkers MP. Rehabilitation treatment taxonomy: establishing common ground. Arch Phys Med Rehabil. 2014;95(1 suppl 1):S1–S5.
Immediate and Long-term Benefits of the Proposed Framework
In the near term, this conceptual framework is intended to support improved treatment specification and clinical reasoning. Any treatment can be specified by its hypothesized active ingredients and their predicted effects on a specific target of treatment, some aspect of functioning that the clinician (and the patient, or both) thinks needs to be changed. This specification encourages clinicians to predict the direct consequences of the ingredients they select for treatment sessions and to articulate separately the enablement model that predicts additional downstream effects. Furthermore, the specification of active ingredients is valuable for education and consistency of treatment delivery across therapists.
A similar specification should be used in research reports of new therapies: the framework is intended to facilitate better communication of what the investigators did (ingredients), how it was hypothesized to work (mechanism of action), and what were the direct effects on specified, measurable aspects of functioning (targets). Communication of these elements of the tripartite structure will improve reproducibility of treatments for both clinical and research purposes. In the long run, this framework can guide organization of an ingredients-based taxonomy of well-specified treatments.
The key concepts presented in this document (and the more complete description of the conceptual framework in the Archives of Physical Medicine and Rehabilitation supplement) require broad discussion by all disciplines to test their utility and applicability across the range of rehabilitation settings and interventions.
Footnotes
Dr Dijkers, Dr Hart, Mr Packel, Dr Whyte, and Dr Zanca provided concept/idea/project design and writing. Dr Ferraro and Dr Zanca provided data collection and analysis. Dr Whyte and Dr Zanca provided fund procurement. Dr Whyte provided consultation (including review of manuscript before submission).
- Received November 11, 2013.
- Accepted November 12, 2013.
- © 2014 American Physical Therapy Association
Issue highlights
