Table 2 Description of the sensitivity analyses that will be made for the primary outcome
From: Detailed statistical analysis plan for the Danish Palliative Care Trial (DanPaCT)
Imputations | Adjusted | Area under the curve (AUC) | Intention-to-treat (ITT) principle | |
|---|---|---|---|---|
Fully adjusted analysis | Yes, as in the primary analysis | Adjusted for covariates (center, WHO performance status, time since the patient was diagnosed with metastases, treatment status, sex, age, diagnosis and education) that are associated with the primary outcome (P <0.10 in a univariate test) and where missing <5% | Yes, as in the primary analysis | Modified ITT analysis as in the primary analysis. |
Complete case analysis | No | Yes, as in the primary analysis | Yes, as in the primary analysis | Violated, no imputations for patients who did not respond to the questionnaire. Analysis may be biased |
Model for repeated measurement analysis | Yes, as in the primary analysis | Yes, as in the primary analysis | No. The three- and eight-week follow-ups are assessed individually | Modified ITT analysis as in the primary analysis |
Per protocol analysis | Yes, as in the primary analysis | Yes, as in the primary analysis | Yes, as in the primary analysis | Violated, excluding patients who did not receive the experimental intervention (defined as at least one contact to the SPC team). The analysis thus only describes the effect of the intervention on those complying with the intervention, which is a subgroup of patients |
Including patients who died | Yes, as in the primary analysis | Yes, as in the primary analysis | Yes, as in the primary analysis | ITT analysis. Multiple imputations for patients who do not respond or die |