You will get efficient and thoughtful service from NAILI.
Analysis of the primary outcome: time to healing
The trial analyses will follow a detailed pre-specified statistical analysis plan (SAP), which will be approved and made available on the funder&#;s website before the end of patient recruitment. The primary outcome will be time to healing of the reference ulcer as defined in the &#;Primary outcome&#; section. If, prior to healing of the reference ulcer, participants reach the end of study follow-up, are lost to follow-up, withdraw from follow up and do not consent to continued healthcare records access for healing data or experience competing events (death or amputation of the reference leg), healing times will be censored as detailed in Fig. .
The primary analyses will include both non-inferiority and superiority comparisons. For the non-inferiority comparison (two-layer bandage vs evidence-based compression), two treatment effect estimands will be targeted.1 An unconstrained treatment policy estimand whereby all intercurrent events (e.g. treatment switches/cessations and receipt of surgery) will be accepted as part of completely pragmatic treatment policies (essentially an intention-to-treat analysis) and a constrained estimand whereby certain constraints are placed on the intercurrent events that are accepted as a part of the treatment policies.
For more venus compressorinformation, please contact us. We will provide professional answers.
Intercurrent events (such as treatment switches) will often result in the care received by patients allocated to different treatment groups being more similar, potentially resulting in attenuation of differences between treatment groups. Hence, estimands that do not address the influence of intercurrent events are potentially anti-conservative in the context of investigating non-inferiority. This provides the rationale for the constrained treatment policy estimand; the purpose of this estimand is to remove the influence of intercurrent events that could either be mitigated in practice (e.g. treatment switches that could be avoided by appropriate use of analgesia), or that might not be applicable to all patients (e.g. surgical treatments that some patients cannot/will not receive), whilst accepting the influence of intercurrent events that cannot be easily avoided as an unalterable part of the treatment strategies being compared (e.g. treatment switches due to allergy/adverse events). For the superiority comparisons, just the unconstrained treatment policy estimand will be targeted. Precise description of the unconstrained and constrained treatment policy estimands, and the estimators/methods that will be used to estimate them, will be given in the SAP.
For the non-inferiority comparison, the unconstrained treatment policy estimand will be estimated by modelling reference ulcer healing time using a Cox proportional hazards model with a three-level variable for group allocation (evidence-based compression vs two-layer bandage vs compression wraps), conditioning on the following baseline covariates: reference ulcer area, reference ulcer duration, participant age, participant mobility status and recruitment site (via shared frailties for participants recruited at the same site). For the constrained treatment policy estimand, patients will have their reference ulcer healing times censored at the first of any relevant departures from the constrained treatment policies. These healing times will then be analysed using a similar model as used to estimate the unconstrained treatment policy estimand. For both analyses, the point estimates and 95% confidence intervals (CIs) of the hazard ratios (HRs) for all between-group contrasts will be reported, together with estimated differences in median healing times (with 95% CIs based on non-parametric bootstraps). For each estimand, the estimated upper 95% confidence limits of the HRs for the evidence-based compression vs two-layer bandage contrasts will be compared with the non-inferiority margin of 1.33. If either of these upper limits is greater than 1.33, then the null hypothesis that two-layer bandage is inferior to evidence-based compression will not be rejected. If both estimates are less than 1.33, then the null hypothesis that two-layer bandage is inferior to evidence-based compression will be rejected.
If the null hypothesis that two-layer bandage is inferior to evidence-based compression is rejected, then a further model will be fitted with the two-layer bandage and evidence-based compression groups combined and compared with compression wraps (i.e. a two-level variable for group allocation&#;evidence-based compression&#;+&#;two-layer bandage vs compression wraps), with the same fixed effects and shared frailties as previously detailed. The point estimate and 95% Wald method CI of the HR for the two-layer bandage&#;+&#;evidence-based compression vs compression wraps contrast will be reported, together with the estimated difference in median healing time (with 95% CIs based on non-parametric bootstraps).
Want more information on compressor maintenance kits? Feel free to contact us.
Comments
Please Join Us to post.
0