Biostatistics

Biostatistical design, analysis & reporting for theses and manuscripts

We support clinical, epidemiological, and public-health research from protocol design and power calculation to survival analysis and CONSORT-ready figures/tables.

Clinical & public health Rigorous assumptions Journal-style reporting

What’s included

Study design-outcomes, endpoints, covariates, randomization/stratification.
Power & sample size-for t/ANOVA, proportions, survival, clustered designs.
Data preparation-CRF mapping, coding books, missing-data strategy.
Analysis-classical tests, regression/GLM, mixed models, survival analysis.
Reporting-effect sizes, CI, assumptions & limitations, CONSORT/STROBE tables.

Toolstack

R Stata SPSS SAS* Python REDCap Qualtrics

*SAS support on request add to scope at discovery.

Study types

RCTs Cohort/Case-control Cross-sectional Diagnostics Survival Clustered

Sub-services

Protocol review

Endpoints, estimands, confounders, bias mitigation.

Randomization

Simple/blocked/stratified; allocation concealment notes.

Sample size & power

Superiority, non-inferiority, equivalence; attrition allowance.

Data management

Codebooks, cleaning rules, missing-data plan (MAR/MNAR notes).

Comparative tests

t/ANOVA/ANCOVA, χ²/Fisher, non-parametric alternatives.

Regression & GLM

Linear/logistic/Poisson/neg-bin with diagnostics & effect sizes.

Mixed models

Repeated/clustered data (LMM/GLMM) with random effects.

Survival analysis

Kaplan–Meier, log-rank, Cox PH, diagnostics, time-varying covariates (intro).

Diagnostics & ROC

AUC/ROC, calibration, sensitivity/specificity, decision thresholds.

Reporting

CONSORT/STROBE/PRISMA tables, APA/AMA figures, manuscript methods/results.

Quick reference

AreaExamplesNotes
DesignParallel, crossover, cluster-randomizedICC, carryover effects, stratification
Comparisonst/ANOVA/ANCOVA, χ²/FisherMultiple comparisons, adjusted p-values
RegressionLinear, logistic, Poisson/Neg-binCollinearity, link tests, residual checks
Mixed modelsLMM/GLMMRandom intercept/slope, covariance structure
SurvivalKM, log-rank, Cox PHProportional hazards checks, Schoenfeld
DiagnosticsROC/AUC, calibrationThreshold selection, decision curves (intro)

All analyses documented with assumptions, effect sizes, CI, and limitations for defensible conclusions.

Engagement examples

Power & Design (8–12 hrs)
Starting scope

Design consultation + power/sample size memo + template CRFs.

Custom quote
Request proposal
Chapter-Ready (25–35 hrs)
Most popular

EDA + primary analyses + 3–5 figures/tables + methods/results write-up.

Custom quote
Request proposal
Manuscript Pack (50–70 hrs)
Best for submission

Advanced models + CONSORT/STROBE tables + submission formatting.

Custom quote
Request proposal

Pricing depends on design complexity, data quality, and turnaround. You’ll receive a clear plan after discovery.

Process & timeline

1Discovery

Aims, endpoints, constraints; journal/ethics requirements.

2Design & power

Randomization, estimands; sample size memo.

3Data setup

Templates, codebooks, cleaning rules, missing-data plan.

4Analysis

Primary & sensitivity analyses with diagnostics.

5Reporting

Tables/figures + methods/results write-up.

6Handover

Scripts, outputs, and change log; submission tips.

Typical deliverables

  • Power & sample size memo with assumptions and references
  • Cleaned dataset (as permissible) with data dictionary/codebook
  • Reproducible scripts (R/Stata/SPSS/SAS) or annotated notebooks
  • CONSORT/STROBE-style tables, survival plots, and APA/AMA figures
  • Methods & Results write-up with effect sizes, CI, diagnostics, and limitations

FAQ

Yes we align endpoints with objectives/estimands and assess power implications.

We can generate randomization/stratification lists and document procedures.

We plan for missingness (MAR/MNAR considerations) and may use multiple imputation or sensitivity analyses.

Yes KM curves, log-rank tests, Cox PH with proportional-hazards diagnostics.

Yes LMM/GLMM with appropriate covariance structures and marginal means.

CONSORT, STROBE, PRISMA as applicable; figures/tables formatted per journal.

Absolutely scripts/notebooks and a change log are included for reproducibility.

We use robust/non-parametric alternatives, transformations, or re-specify models and report rationale.

Design/power: days; full analyses/manuscripts: weeks with milestone plan.

By design complexity, data condition, and scope. You’ll receive a transparent quote after discovery.

Start Biostatistics support