Assessing fidelity to treatment delivery in the ICONS (Identifying Continence OptioNs after Stroke) cluster randomised feasibility trial

Assessing fidelity to treatment delivery in the ICONS (Identifying Continence OptioNs after Stroke) cluster randomised feasibility trial

BMC Medical Research Methodology, August 2015

26293927

Brigit M. Chesworth, Michael J. Leathley, Lois H. Thomas, Christopher J. Sutton, Denise Forshaw, Caroline L. Watkins, The ICONS project team and the ICONS Patient, Public and Carer Involvement Groups

The implementation of strategies to monitor and enhance treatment fidelity is of paramount importance in trials of complex interventions. A recent framework published by the National Institutes of Health Behavior Change Consortium recommends addressing five areas of treatment fidelity, one of which is delivery of treatment. This study aimed to explore fidelity to treatment delivery of the ICONS intervention (a systematic voiding programme [SVP]). This included exploring the feasibility of a method to assess fidelity to treatment delivery and collecting preliminary evidence of the level of fidelity to SVP delivery in order to inform strategies for improving fidelity in a future trial. Delivery of treatment was recorded by nurses through completion of daily clinical logs, which included: a voiding interval, proposed voiding times and actual voiding times. The a priori method for assessment of fidelity - comparing actual voiding times with proposed voiding times - was trialled on a small amount of data. Due to errors in documentation of the voiding intervals and proposed voiding times it was not possible to assess fidelity directly as planned. A new method was devised, which included identification of 'key quality indicators'. This new approach to assessing fidelity used key quality indicators based upon presence of the data needed to make the comparison between proposed and actual voiding times. The proportion of clinical logs with correct documentation of voiding intervals and proposed voiding times was less than 40 %. For clinical logs with correct documentation, an actual voiding time within 30 min of the proposed voiding time was identified on approximately 55 % of occasions. Lessons learnt from this study have implications for the future ICONS definitive trial and for other trials of complex interventions. Implementation of a complex intervention may often deviate from what is intended. While careful consideration should be given to the best method of fidelity assessment, an iterative approach allowing flexibility to adapt pre-planned methods is recommended within feasibility trials. As fidelity to treatment delivery in the ICONS feasibility trial appeared to be relatively low, more attention to implementation strategies will be required in the definitive trial. Identifier: ISRCTN08609907 ; date registered: 07/07/2010.