Prof Cramer focused on the importance of getting our entry criteria and stratification right in trials – to reduce variance and to provide us with the greatest potential to find an effect when one actually exists. The importance of getting the right patients into trials was emphasised, one of our challenges in the field as heterogeneous as stroke is understanding who those people are. Stratification helps us to breakdown larger groups into important subclusters, this can increase power and reduce the number of patients required in the trial. The other two areas covered by prof Cramer included Endpoint training (training outcome assessors to be able to consistently and accurately measure and outcome) and concommitent experience in trials. Endpoint training can also help reduce variance, save time and money.
He provided an example of how training on the Fugl Meyer reduced number of patients needed in a trial from 137- 88! This is an important development area for our field. The importance of ensuring that training is included in trials of recovery enhancing drugs, or other interventions was also emphasised. To believe that correct rewiring can happen without training is too simplistic.
Prof Eng used her ongoing DOSE trial as an exemplar of the importance of carefully designing the intervention and tracking intervention fidelity. In this 3 arm trial Prof Eng is testing whether significantly higher intensity and amount of walking training can help recovery. She highlighted the recommendations from the first international Stroke Recovery and Rehabilitation Roundtable working group on intervention development and monitoring (available free online from the International Journal fo Stroke and via the NNR website). The importance of understanding exactly what patients receive as intervention in trials was emphasised and while the DOSE trial is ongoing, Prof Eng was able to show that she can have confidence that patients in her trial are receiving significantly different doses of training in each arm of her trial. In the last part of her talk, Prof Eng discussed the potential value of using a Stepped Wedge Design for interventions without know harms that are aiming to be implemented in real world settings. She outlined the fundamental nature of these designs.
Prof Kwakkel presented the case for increasing our understanding of now just who recovers and who does not and by how much (part of our understanding of proportional or spontaneous recovery) but also the pattern of recovery. Again the heterogeneity of the stroke population was emphasised and the urgent need to reduce variation in our trials and stratify our patients better into trials was strongly emphasised. Using a number of examples from his lab, Prof Kwakkel showed how understanding recovery patterns may help in the future stratification – if we can predict up front these patterns. This is still work in progress. Improving the way we develop the dose of interventions (dosing – described by Prof Kwakkel as often “homeopathic”), when we study our patients (timing – early rather than later), what outcomes we select, and what theories underpin our choices will all improve trial quality. Prof Kwakkel outlined the Stroke Recovery and Rehabilitation Roundtable working group on developing core outcomes for stroke recovery trials (again available online free). He also showed an intriguing “brain bus” which allowed a portable EEG to come to the patients rather than the patient needing to come to the clinic.
Prof Bernhardt gave an overview of adaptive trial designs which are commonly used in cardiovascular disease and cancer but rare in rehabilitation. The value of these designs was outlined, they are argued to improve flexibility, precisions and to be more ethical. But they are not simple to run or design, need significant data prior to the study to model the approach and they are not well understood from a regulatory standpoint. Adaptive trials (unlike standard randomised controlled trials which apply a “closed” approach), allow future randomisation into specific treatment arms to be informed by patient outcomes once an early period of recruitment has been completed. This means that multiple treatments can be tested in the one trial, with the optimal treatment compared to control at trial end. Prof Bernhardt gave an example of how she is using this in AVERT-DOSE, the final installment of the AVERT early mobility based intervention trial.