RESEARCH SUPPORT
Clinical Trial Statistical Analysis Plan
Estimands Defined Before Unblinding
Clinical trial SAP development for investigators and doctoral researchers planning a trial: the estimand stated, the primary analysis and its population fixed, missing data and later events specified, and robustness analyses planned before the database is locked.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
What Does a Trial Analysis Plan Contain?
A clinical trial statistical analysis plan (SAP) fixes, before outcome data are examined, how the objectives of a trial will be analyzed. It states the estimand, the analysis populations, the model for the primary endpoint, the treatment of missing data and of events after randomization, and the supportive analyses planned. Clinical trial SAP development ties these choices to the protocol and the sample size so that all three agree. ICH E9(R1) estimands are stated through five attributes: the population, the treatment, the variable, the handling of intercurrent events and the summary measure. OnlineSPSS is a specialist statistical consultancy for academic, clinical and professional research whose in-house PhD-level statisticians draft the plan with the trial team, while every clinical decision remains the investigators’.
WHEN TO ENGAGE
Where a Trial Stands When the Plan Is Needed
A plan can be drafted from the protocol alone, though the right moment depends on how much is already fixed and on whether any outcome data have been seen.
Protocol Near Final The protocol is approved or close to it, and the objectives must be translated into endpoints, analysis populations and a primary model that agree with the sample size calculation.
Recruiting With Data Blinded Enrollment is under way and outcome data remain blinded, so the plan can still be finalized and dated, with any blinded review of the data limited to quantities such as missingness.
Investigator-Led or Pilot Trial A small trial or a doctoral study has no sponsor statistics team, and the plan must be proportionate yet specific enough that the primary analysis cannot be chosen after the results.
Plan Needs an Amendment A protocol amendment, an interim analysis or an unexpected pattern of discontinuation calls for a revised version, with each change dated and justified.
STRUCTURE OF THE PLAN
Estimands, Populations and the Primary Analysis
The estimand is set first, and the model, the populations and the missing-data approach are chosen to answer it, not the other way round.
Estimand Attributes
Each primary objective receives an estimand in the terms of ICH E9(R1), drawn up as part of Research Planning & Study Design.
Population Definitions
The full analysis set and any per-protocol set are defined by rule before unblinding, so that exclusions cannot depend on the outcomes.
Primary Model
The model and covariates for the primary endpoint are fixed in advance, with the testing order for several endpoints, and for time-to-event data the plan follows Survival Analysis.
Missing Data Strategy
The main approach to missing outcomes is chosen to be consistent with the estimand, and what it assumes about the missing values is stated for review.
Planned Robustness Analyses
Each supportive analysis targets a named assumption, such as the missing-data mechanism or the handling of intercurrent events, and is labeled in advance as supportive.
Shells for Every Planned Table
Table and figure shells accompany the plan, within the discipline of Clinical Trials and the wider Clinical, Survival & Epidemiological methods.
WEAK POINTS IN TRIAL PLANS
Defects That Weaken a Trial Analysis Plan
A plan fails when it leaves a decisive choice open, and the open choices are usually the same few: populations, missing outcomes and events after randomization.
Estimand Left Implicit The plan names a test but not the question it answers, so the handling of treatment discontinuation or rescue therapy is decided after the data show who stopped.
Sensitivity Analysis Aimed Elsewhere A per-protocol or complete-case analysis is labeled a sensitivity analysis of a treatment-policy primary, though it targets a different estimand, so agreement or disagreement says little about the robustness of the primary result.
Missing Outcomes Assumed Harmless Complete cases are analyzed without stating what is assumed about the missing outcomes, though dropout often depends on the treatment or the course of the illness.
Analysis Visit Windows Undefined The plan does not say which measurement counts when a visit falls early or late, so the value used for the primary endpoint can be chosen after unblinding.
Primary Endpoint Left Movable More than one candidate primary endpoint or time point stays in the plan, so the final choice can follow the results.
Changes After Unblinding Unlogged The plan is revised after the data are seen without a dated record, so readers cannot tell which analyses were planned and which came later.
DELIVERABLES
Contents of the Plan Package
The documents fix the analysis in a form that can be dated, versioned and shown to a monitoring board, an ethics committee or a journal.
Versioned Plan Document The plan itself, with sections that follow the trial objectives, a version history and the date of finalization relative to unblinding.
Intercurrent Event Strategy Table Each foreseeable intercurrent event, such as discontinuation, rescue medication or death, listed against the strategy chosen for it and the analysis that strategy implies.
Table and Figure Shells Empty templates for every table, listing and figure planned, so that programming and the trial report follow a fixed layout.
Sample Size Rationale A statement of the assumptions behind the sample size and how they relate to the primary estimand, for the protocol or the statistical section of a grant.
Assumption Register A list of each assumption behind the primary analysis, the check that probes it and the result that would change the interpretation.
Amendment Log Template A dated log for changes after finalization, recording what changed, why, and whether outcome data had been seen at the time.
ILLUSTRATIVE EXAMPLE
Deaths Before the Mobility Assessment
An illustration of how a plan decides, before unblinding, what a mobility endpoint means for participants who die before it can be measured.
ILLUSTRATIVE EXAMPLE · HIP FRACTURE REHABILITATION TRIAL · UNITED STATES
A geriatrician leading a randomized trial of an early home rehabilitation program after hip fracture needed a plan for the primary endpoint, a mobility score at sixteen weeks, in a population where some participants die during follow-up.
The plan was drafted from the protocol while recruitment continued. The complication was death before the sixteen-week visit: no mobility score exists for these participants, and treating them as missing would describe a population in which nobody dies. The team weighed the options and made a composite strategy primary, ranking death below every observed score and comparing the arms with a rank-based method, so that the endpoint reflected both survival and mobility. A while-alive summary was planned as supportive, missed visits among survivors were handled under a stated assumption, and the plan was dated and versioned before unblinding.
Outcome: The investigator received the dated plan, the intercurrent event strategy table, the table and figure shells, the sample size rationale and an amendment log template, with the while-alive analysis labeled as supportive.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
SOFTWARE
Coding the Prespecified Primary Model
Repeated-measures endpoints can be fitted in several packages, so the choice follows the programming standard of the sponsor or institution.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCH RESPONSIBILITY
The Trial Remains the Investigators'
The protocol, the clinical decisions and any submission to an ethics committee or a regulator rest with the investigators and the sponsor, and the plan describes how the data will be analyzed.
Agreed Statistical Support OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, defined in a scope that is confirmed before any work begins.
No Intellectual Contribution OnlineSPSS does not claim any intellectual contribution to your research. Authorship of the protocol, the trial report and any article remains yours.
Your Research Decisions The trial design, the clinical definitions, the interpretation of the results and the conclusions remain yours, with overall responsibility.
Confidential Handling Research data and project details are handled confidentially, project information is not published, and an NDA is available where required.
FREQUENTLY ASKED QUESTIONS
Timing, Intercurrent Events and Smaller Trials
When the plan should be finalized, what an intercurrent event is and whether a small trial needs a full plan.
What does a SAP for randomized trials include?
The plan states the estimand, the analysis populations, the primary model, the handling of missing data and intercurrent events, the supportive analyses and the table shells, and is finalized before outcome data are unblinded. Its structure follows the trial objectives, with one section per estimand and the supportive analyses listed beneath the primary analysis they test.
When should the analysis plan be finalized?
For a blinded trial, before the database is locked and the allocation is revealed. For an open-label trial, before outcome data are examined. Blinded review of quantities such as missingness is compatible with this, and later changes are recorded as dated versions, so the sequence of decisions can be shown to a reviewer.
What is an intercurrent event in a trial?
An event after randomization that affects the existence or interpretation of the endpoint, such as stopping treatment, taking rescue medication or death. The plan names a strategy for each: treatment policy, hypothetical, composite, while on treatment or principal stratum.
Does a small investigator-led trial need a full plan?
It needs a plan, though a shorter one. The essentials remain the estimand, the primary analysis, the populations and the handling of missing data, because they are what prevents the analysis from being chosen after the results are known. The document can be briefer than a sponsor template, and the protocol may hold part of it.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the protocol and the sample size calculation, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Estimands Settled While
the Trial Is Still Blinded
After the allocation is revealed, no revision of the plan can show that its primary analysis was chosen without knowing the results.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014