Individual (N-of-1) trials can be combined to give population comparative treatment effect estimates: methodologic considerations

Abstract

Abstract/SummaryObjective-To compare different statistical models for combining N-of-1 trials to estimate a population treatment effect.Study Design and Setting-Data from a published series of N-of-1 trials comparing amitriptyline therapy and combination treatment (amitriptyline + fluoxetine ) were analyzed to compare summary and individual participant data meta-analysis, repeated measures models, Bayesian hierarchical models, single-period, single-pair and averaged outcome crossover models.Results-The best fitting model included a random intercept (response on amitriptyline) and fixed treatment effect (added fluoxetine). Results supported a common, uncorrelated within-patient covariance structure that is equal between-treatments and across patients. Assuming unequal withinpatient variances, a random effects model was favored. Bayesian hierarchical models improved precision and were highly sensitive to within-patient variance priors.Conclusion-Optimal models for combining N-of-1 trials need to consider goals, data sources, and relative within and between patient variances. Without sufficient patients, between-patient variation will be hard to explain with covariates. N-of-1 data with few observations per patients may not support models with heterogeneous within-patient variation. With common variances, models appear robust. Bayesian models may improve parameter estimation but are sensitive to prior assumptions about variance components. With limited resources, improving within-patient precision must be balanced by increased participants to explain population variation.

Translating rare-disease therapies into improved care for patients and families: what are the right outcomes, designs, and engagement approaches in health-systems research?

“…44 In both n-of-1 (single participant) and crossover trials, participants receive two treatments in random order with a washout period in between, such that each participant acts as his or her own control; 43,44,50 variations of these designs may incorporate more than two treatment periods, which can help to mitigate potential bias due to carryover and period effects. 44 n-of-1 trials are explicitly oriented toward supporting individual treatment decisions.…”

“…51 Multiple n-of-1 trials can also be combined to provide population-level effect estimates, although there are some challenges with the approach when the sample is small. 44,50 Although both n-of-1 and crossover designs are susceptible to carryover and period effects, these approaches are ideal designs for investigating heterogeneity in treatment effects across participants ( Table 2). A drawback of n-of-1 and crossover trials is their reliance on short-term outcomes with rapid response to the intervention.…”

“…Methods for systematic reviews of RCTs are well established, 62 and we have mentioned the potential value of combining multiple n-of-1 trials to estimate population treatment effects. 50 Methods for synthesizing the results of studies that have used different study designs are more challenging; heterogeneity in study methods and participants means that meta-analyses, in which results from multiple studies are summarized quantitatively, are difficult to produce. However, if challenges to meta-analysis can be overcome through either standardization of study methods or statistical adjustment for key methodological differences, network meta-analyses may be a promising approach for evaluating multiple interventions for rare diseases.…”

Cystic Fibrosis: The Dawn of a New Therapeutic Era

“…To date, CF studies are increasingly implementing formal group sequential stopping rules to improve the safety and ethics of ongoing trials and decrease the average sample size of trials. N-of-1 study designs provide a unique opportunity to apply patient-centric "precision medicine" to rare CFTR mutations that might otherwise be excluded from larger scale clinical trials (61); such studies can be pooled and leveraged for population-level inference in addition to their application in individual patient care decisions. The participation rate in clinical trials is high among those with CF (62), and patient, family, and physician engagement in therapeutic trials is highly encouraged and valued in the CF community.…”

Optimizing behavioral health interventions with single-case designs: from development to dissemination

“…In addition, advances in effect size estimates has led to several meta-analyses of results from SCDs [48,[58][59][60][61]. Zucker and associates [62] explored Bayesian mixed-model strategy to combining SCDs using, which allowed population-level claims about the merits of different intervention strategies.…”