Treatment fidelity in clinical trials

Introduction

How many times have you read a research study and either: (1) had no idea what the treatment intervention consisted of; or (2) realized that the “intervention” that was used in the study was nothing like what you would apply in clinical practice? If you’ve encountered these two situations while reading literature, you may have been witness to limitations of treatment fidelity (TF).

Despite its importance, TF is often poorly reported in clinical trials (1-3). This is especially the case in behavioral-based studies that require some degree of clinician interpretation of the patient’s progress and a modification based on that interpretation (4). It may also be because the definition of TF can vary across studies and contexts. Although TF generally refers to the extent to which an intervention is delivered as intended, ensuring consistency and reliability, terms such as “adherence,” “integrity/veracity,” or “implementation fidelity” are commonly used, which may not be anchored to the same underlying concept.

In this viewpoint, we focus on perspectives that have a “clinical context” (with a goal of improving clinician interpretation of TF) and provide a modern definition of TF, by describing two key components of TF (adherence and veracity), discuss examples in the literature in which TF was lacking, and provide methods to improve the implementation of interventions in clinical trials. We hope to show that in addition to commonly measured constructs such as quality, risk of bias, and certainty of evidence, TF should be assessed when interpreting the meaningfulness of a clinical trial.

A modern definition of TF

In the context of clinical trials, TF has historically referred to the extent to which an intervention or treatment adheres to the implementation parameters intended by the researchers who designed the trial (5). Indeed, appropriate implementation is critical as TF is essential in ensuring that the results of the trial accurately reflect the treatment effects of the intended intervention, with no additions or omissions. Adequate TF improves one’s interpretation of the outcome data in research studies, improves the likelihood of reproducibility (if studied again), and is essential for clinical translation (5,6). This demands appropriate reporting of treatment structure used in the trial. Perhaps most importantly, TF is one of the few elements in a clinical trial that equally represents components of internal and external validity (5).

Adherence of TF routinely measures protocol and dosage adherence. Adherence can be considered as “did the researchers do as they indicated they would do?” Protocol and dosage adherence reflect the extent to which the intervention was delivered as planned. It involves an assessment of whether the treatment protocol was followed closely, including the dosage, frequency, and duration of the intervention. Investigators in clinical trials should demonstrate an effort to show that they have optimized dosage capacity by incorporating known parameters of therapeutic effectiveness and an application that is similar to that provided in clinical practice.

Recent consensus-based work has markedly widened the scope of topics that reflect TF (7). In addition to whether the intervention is delivered with a high degree of adherence, TF should include efforts to ensure that the application of the intervention is performed in a way that is known to be therapeutically beneficial (4) (Fig. 1). In other words, was the veracity of the intervention performed and implemented in a manner that should allow someone to improve if performed in a similar clinical situation? To ensure the veracity of TF in a clinical study, one must consider: (a) the quality of delivery and (b) participant adherence.

Quality of delivery assesses both the therapeutic potency of the interventions and the competency of the individuals delivering the intervention. Therapeutic potency reflects whether the clinical parameters such as dosage, time, etc., are performed in a way that allows optimal therapeutic recovery. In a pharmaceutical trial, it would reflect whether the research participant received the appropriate dosages of the medications at appropriate time intervals. Additionally, quality of delivery involves evaluating whether the research administrators have the necessary skills, training, and expertise to deliver the treatment effectively, and ensuring consistency in the provision of interventions between those delivering the treatment.

Participant adherence refers to the extent to which participants engage with and respond to the intervention. It involves monitoring participants’ adherence to the intervention protocol, their understanding of the intervention, and their willingness to participate. The selection of appropriately responsive measures that actually assess patient engagement and change in outcomes within the targeted domain is requisite to ensure these measures have meaning.

Examples and recommendations involving TF in clinical trials

Although critical, it is important to recognize that assessment and implementation of TF procedures in a trial is a challenging process (2). There are numerous studies that have either experienced or highlighted TF concerns. In this section, we outline examples of TF limitations and provide options for improvement in future studies.

Unique challenges of TF for physiotherapy and rehabilitation approaches

Measuring TF in physiotherapy can be challenging compared to other areas of healthcare, such as a pharmacological intervention that uses objective laboratory values to determine a treatment regimen, because the nature of physiotherapy is multifaceted, interventions are often clinician-dependent, and interactions between the clinician and patient are uniquely individual (17). Multiple elements impact the delivery, receipt, and enactment of a prescribed physiotherapy treatment intervention and TF may be impacted by the clinician, the patient, or the actual treatment itself (18). The skill of the physiotherapist, the individual needs of the patient, and the distinct interventions required for each individual widely vary across the physiotherapy field, which can lead to significant difficulties in measuring TF.

Multiple covariates associated with the delivery of physiotherapy or other rehabilitation services, such as the time spent with the patient, the setting, and the therapeutic alliance between the patient and provider, can influence TF (18). Because there is so much variation in physiotherapy, a specific checklist may not allow for enough latitude, leading to an unclear interpretation of how high the TF truly is (19). Adaptability within a research protocol, or “flexible fidelity” (20), allows the adjustment of protocol components in response to individual patient differences, such as tailoring exercises based on an individual’s pain response or strength. In this context, fidelity can be viewed as adherence to the underlying theory outlined in a treatment protocol, rather than to specific activities or behaviors.

Conclusion

In this viewpoint, we outline the components of TF and provide examples in the literature where TF was lacking. We argue that TF is critical to establishing the evidence base of interventions and determining the circumstances under which an intervention is most effective. Interventions need to be delivered with a high degree of TF, which will allow for greater confidence that the outcomes observed are truly driven by the specific intervention. When TF is not adhered to in clinical research, we may rightly be left to wonder what effect minor modifications of the protocol had on patient outcomes. We suggest that there is a risk that minor modifications could potentially erode the true effect of the treatment and influence clinical outcomes, leading to “evidence” that is erroneously adopted into evolving clinical paradigms.

Disclosures

Conflict of interest: The authors report no conflicts of interest.

Financial support: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Authors’ contributions: All authors participated in the study concept, writing, and final manuscript preparation.

Data Availability Statement: There are no data associated with this viewpoint.