2.3 Functional Strength Training
2.3.2 Intervention Fidelity
Intervention fidelity, a term which can also be used interchangeably with treatment fidelity refers to the methodological strategies used to enhance the reliability and validity of clinical interventions which are designed to effect a change in behaviour. Section 2.2 identified that people in the chronic stage of
recovery from stroke may not have engaged in exercise for a prolonged period of time. In order for a study which is designed to test the feasibility of an exercise-based intervention such as FST to succeed, it is apparent that this behaviour will have to change.
The concept of intervention fidelity appears to have been studied predominantly within psychological, social and behavioural research (Bellg et al., 2004, Gearing et al., 2011). In an early opinion paper on this topic Moncher and Prinz (1991: pp. 247) referred to treatment fidelity as being “two related, but distinct, issues”. These were:
The degree to which the experimental interventions were implemented as intended and secondly that there was sufficient differentiation between the two experimental conditions such that the findings from the study could be attributed to the independent variable.
It was reported that poor intervention fidelity could lead to unreproducible interventions, a decrease in statistical power and low uptake of potentially effective interventions (Moncher and Prinz, 1991). With this in mind a working party was subsequently created to examine the issues of intervention fidelity and to produce a set of guidelines for improving this in a particular group of studies examining the effectiveness of experimental interventions targeting change in health behaviours. The Behaviour Change Consortium (BCC) published best practice and recommendations for ensuring intervention fidelity (Bellg et al. 2004). In this the authors have addressed areas broader than the original definition of intervention fidelity posited by Moncher and Prinz (1991) and the working definition for intervention fidelity expanded to “the methodological strategies used to monitor and enhance the reliability and validity of behavioural interventions” (Borelli et al., 2005 pp852). However the underlying goal of improving study rigour remained the same.
The BCC described five areas where intervention fidelity could be addressed.
These are:
- Design of study - Training providers
- Delivery of treatment - Receipt of treatment
- Enactment of treatment skills’ (Bellg et al., 2004).
These will be described below along with barriers and facilitators to each of these with examples from studies published in the field of long term conditions.
There appears to be limited work published in this field specific to exercise and stroke and therefore the literature used has been expanded to include studies which have also evaluated the use of exercise in the management of other cardiovascular diseases.
Design of study: Methodological strategies aimed at ensuring intervention fidelity in this category are intended to ensure that the intervention has been appropriately identified (i.e. that there is sufficient theoretical and clinical underpinning), to avoid confounders such as the delivery of different ‘doses’ of each experimental intervention and strict adherence to the research protocol (Bellg et al., 2004).
Defining and delivering an intervention consistently is important in research trials because of the need to be able to communicate it’s content to care providers if the results suggest that the intervention is effective. This can be particularly challenging in the evaluation and subsequent dissemination of complex interventions such as those designed to manage health behaviour change which may have several components and be delivered across different care settings (Spillane et al., 2007). The development of a robust treatment manual with clearly defined interventions may go some way to addressing this.
Monitoring adherence to a research protocol and ensuring effective delivery of interventions that are self-directed such as engagement in an exercise programme may however prove particularly challenging. In their guidelines for best practice Bellg et al (2004) listed several strategies that could be incorporated into the study design which may help address these potential issues. These included checking by
research staff to ensure adherence to the protocol and the inclusion of methods for enabling participants a means of recording their engagement in the exercise programme. A review of intervention fidelity across exercise trials in participants with diabetes found that studies ensured adherence to the intended quantity of therapy by telephone calls, and motivational techniques such as goal setting (Avery et al., 2012).
Bellg et al. (2004) also suggested that studies could prevent implementation setbacks, such as lack of research staff to deliver the intervention by ensuring a pool of trained researchers. This of course carries its own challenges in that training of any extra staff needs to be maintained and monitored to ensure adherence to a protocol. Although this is likely to increase the costs of a study Henggeler et al. (1997) argued that these were minimal when compared to the expense incurred if ineffective treatments are declared effective because of poor intervention fidelity practices.
The following area published in the guidelines specifically addresses training the intervention providers. The underlying goal for this area is also related to ensuring adherence to the research protocol.
Training providers; strategies within this category suggest standardising training, ensuring provider skill acquisition and maintaining provider skills throughout the study.
Consistent training of the intervention providers is one method of attempting to improve adherence to a protocol. Particular barriers to this include lack of resource dedicated to this aspect of a study and poorly defined interventions. Studies may need to consider the costing implications for ensuring that intervention providers are trained and that this training is monitored and updated throughout the study. The quality of the training should be supported by treatment manuals which describe clearly defined interventions. Where this is not the case then intervention providers may misinterpret the instructions and deviate from
the protocol. Such actions would threaten the rigour of the study findings (Bellg et al., 2004).
In a study evaluating the effectiveness of a treadmill training programme with stroke survivors Resnick et al. (2011) identified that demonstrating intervention fidelity within this category was particularly challenging because of the high turnover of intervention providers. This had led to some discrepancies between the protocol and aspects of the intervention delivery. As the study progressed the research team had managed this by ensuring that there were regular reviews of treatment fidelity which enabled them to be able to reinforce prior training and pre-empt further deviations.
Borelli et al. (2005) developed a tool for assessing intervention fidelity and used this to evaluate the quality of treatment fidelity in 342 trials evaluating interventions targeted at health behaviour change. They found that only 22% of the included studies reported on the training strategies for the intervention providers. This review only reflects how well these studies have reported this aspect of intervention fidelity but does seem to indicate that further consideration may need to be given to the challenges researchers face in ensuring that intervention providers are suitably trained.
Delivery of treatment. This category was created to address aspects such as controlling for provider differences, ensuring adherence to treatment protocol and ensuring that control and intervention are sufficiently different and that there is no overlap between these if they are delivered by the same provider (Bellg et al. 2004).
Where one provider is providing both experimental and control interventions then there is the potential for contamination of either intervention, this may be exacerbated if the provider has a particular bias, inadvertent or otherwise (Bellg et al., 2004). In the SPHERE project which evaluated the effectiveness of secondary prevention measures
including exercise in people with heart disease this aspect of intervention fidelity was managed through quality assurance visits by the research nurses, The study team also carried out randomly selected
‘checks’ of the intervention delivery at different time points in the study (Spillane et al., 2007).
Receipt of treatment – Bellg et al. 2004 created this category to monitor whether studies were able to demonstrate that the participant understood the intervention and whether or not they were able to carry it out.
Where studies rely on active participant engagement with cognitive strategies it is important that the research group determine that the participant understands what is involved and is able to carry out the task.
Interventions such as mental imagery may involve the generation of static images, auditory recall, complex tasks or visual patterns and it may not always be possible for individuals to carry out these activities (Pearson et al., 2013). This may be particularly challenging if researchers are seeking to be inclusive in their sampling strategy. For example, following stroke, potential participants with cognitive problems may be unable to take part, thus biasing the sample.
Studies aiming to deliver exercise such as that conducted by Resnick et al. (2011) need to ensure that the study population is capable of carrying out the required level of intensity and will remain motivated to do so.
Resnick et al. (2011) evaluated the effects of treadmill training on cardiovascular fitness in stroke survivors who were within the chronic phase of recovery from the stroke. The threshold level of intensity (40%
to 50% of the maximal heart rate reserve, 20 minutes continuous exercise, three times per week for six months) set by the study authors was arguably quite challenging for the participants who, as has previously been identified, may not have engaged in physical activity for some period of time. Inclusion criteria for this study therefore focussed on a more physically able sample of stroke survivors making the findings
less generalisable to the wider population (Altman, 1990). Strategies that were then used to facilitate ongoing engagement with the exercise were to deliver the intervention in an environment where the participants were socially engaged and able to motivate one another and to use verbal encouragement by the research staff. These strategies appeared to be effective as participants completed the desired number of exercise training sessions.
Enactment of treatment skills – this refers to the ability of the participant to be able to perform the skills/behaviours required by the research study to ‘real-life’ settings. An example of this described by Bellg et al.
(2004) are the appropriate use of a cognitive strategy to prevent cigarette cravings. This category is distinct from adherence to the treatment protocol which would simply record whether or not the participant had engaged in the intervention not whether or not they had used it in the appropriate way.
Interventions aimed at changing behaviours are generally deemed successful only if they can effect a change within ‘real-life’, therefore by addressing ‘enactment of treatment skills’ in the study design, study authors are prompted to assess whether participants use the experimental skills appropriately throughout the study. This is believed to give an indication of whether this will effect a change in their health behaviour once the study has stopped, a factor which can then be more accurately assessed if participants are subsequently followed up after completion of the intervention phase (Bellg et al., 2004). Adherence to exercise programmes declines following withdrawal of an exercise intervention (Jurkiewicz et al., 2011, Karingen et al., 2011). Arguably studies that address the concept of ‘enactment of treatment skills’ may find that long term adherence to the exercise programme following the end of the study may improve.
Intervention fidelity refers to the strategies that have been developed for ensuring that the findings from studies evaluating interventions targeting health
behaviour change are both reliable and robust (Moncher and Prinz, 1991; Bellg et al., 2004). This section has briefly reviewed some of the literature around intervention fidelity and the barriers and facilitators to its implementation.
Arguably many of the aspects included in the guidelines developed by Bellg et al. (2004) could be considered to be best practice in research design and therefore should be inherent within individual study designs and protocols.
However a review of intervention fidelity by Gearing et al. (2011) found that core components of fidelity such as study design, training providers and intervention receipt were still not being addressed sufficiently. Further reflection on the role of intervention fidelity strategies and their place within the design of the feasibility study described in this thesis will be discussed in chapter seven.
2.3 Summary
Therapies, such as Functional Strength Training, targeting late stage recovery of movement after stroke have the potential to increase an individual’s ability to engage in activities of daily living. These may be more effective if they are integrated into motor learning. Within the framework of developing complex interventions this thesis seeks to describe two studies, the results of which will inform the future development of a novel intervention. The following chapter will outline the aims and objectives for this thesis.