Methodological aspects of a randomized within-patient concurrent controlled design for clinical trials in spine surgery

INTRODUCTION

Randomized controlled trials are considered the highest level of evidence, but their feasibility in the surgical field is severely hampered by methodological and practical issues. Concurrent comparison between the experimental and control conditions within the same patient can be an effective strategy to mitigate some of these challenges and improve generalizability, mainly by the elimination of between-patient variability and reduction of the required sample size. This article aims (1) to describe the methodological aspects of a randomized within-patient controlled trial and (2) to quantify the added value of this design, based on a recently completed randomized within-patient controlled trial on bone grafts in instrumented lumbar posterolateral spinal fusion.

METHODS

Boundary conditions for the application of the randomized within-patient controlled trial design were identified. Between-patient variability was quantified by the intraclass correlation coefficient and concordance in the primary fusion outcome. Sample size, study duration and costs were compared with a classic randomized controlled trial design.

RESULTS

Boundary conditions include the concurrent application of the experimental and control conditions to identical but physically separated sites. Moreover, the outcome of interest should be local, uncorrelated and independently assessable. The spinal fusion outcomes within a patient were found to be more similar than between different patients (intraclass correlation coefficient 32% and concordance 64%), demonstrating a clear effect of patient-related factors. The randomized within-patient controlled trial design allowed a reduction of the sample size to one-third of a parallel-group randomized controlled trial, thereby halving the trial duration and costs.

CONCLUSION

When suitable, the randomized within-patient controlled trial is an efficient design that provides a solution to some of the considerable challenges of a classic randomized controlled trial in (spine) surgery. This design holds specific promise for efficacy studies of non-active bone grafts in instrumented posterolateral fusion surgery.

High Inter-Patient Variability in Sepsis Evolution: A Hidden Markov Model Analysis

BACKGROUND

Severe sepsis and septic shock are common in the intensive care unit (ICU) and contribute significantly to cost and mortality. Early treatment is critical but is confounded by the difficulty of real-time diagnosis. This study uses hidden Markov models (HMMs) to examine whether the time evolution of sepsis can add diagnostic accuracy or value using a proven set of bio-signals.

METHODS

Clinical data (N=36 patients; 6071 hours), including an hourly personalised insulin sensitivity metric. A two hidden state HMM is created to discriminate diagnosed cases (Severe Sepsis, Septic Shock) from controls (SIRS, Sepsis) states. Diagnostic performance is measured by ROC curves, likelihood ratios (LHRs), sensitivity/specificity, and diagnostic odds-ratios (DOR), for a best-case resubstitution estimate and a worst-case 80/20% repeated holdout analysis.

RESULTS

The HMM delivered near perfect results (95% Sensitivity; 96% Specificity) for best-case resubstitution estimates, but was comparatively poor (59% Sensitivity; 61% Specificity) for worst-case repeated holdout estimations. Adding the time evolution of sepsis did not add to the accuracy of diagnosis from using the signals alone without time history.

CONCLUSIONS

These potentially surprising results indicate significant inter-patient variability in the time evolution of sepsis, preventing effective diagnosis in the context of the bio-signals, data, and HMM topology used. Efforts for improved real-time, early sepsis diagnosis should concentrate on the robustness and efficacy of the bio-signals and data used, as well as the level of model complexity, to create more effective real-time classifiers.