BP201T · Core · Semester 2
Applied Biostatistics & Data Analytics for Pharmaceutical Sciences
Apply foundational biostatistics and Python-based analysis to clearly fictional clinical, pharmaceutical, QC, PK, and pharmacovigilance teaching datasets—always with transparent assumptions and careful interpretation.
Course objectives
- Explain fundamental statistical concepts relevant to pharmaceutical sciences.
- Interpret clinical and experimental data scientifically.
- Apply statistical methods to fictional pharmaceutical and biomedical datasets.
- Use Python-based tools for transparent data analysis and handling.
- Recognise the role and limits of statistical analysis in evidence-based decision-making and healthcare machine-learning applications.
Unit 1: Descriptive Statistics
Classify fictional pharmaceutical data as nominal, ordinal, interval, or ratio; then compare centre, spread, and distribution shape. NumPy and Pandas support transparent calculations, but interpretation still depends on the data source and design.
Compare mean, median, range, and standard deviation after adding an outlying simulated QC result, then state what skewness could change.
Unit 2: Probability & Statistical Distributions in Healthcare
Use probability rules, conditional probability, Bayes’ theorem, random-variable concepts, and normal, binomial, or Poisson distributions to reason about fictional healthcare data. A probability illustration does not make a clinical decision.
Choose a distribution for a fictional clinical-trial outcome or rare ADR-style count and name the assumption that makes the choice plausible.
Unit 3: Sampling & Statistical Inference
Separate population from sample, name a sampling error or bias risk, use the Central Limit Theorem conceptually, and interpret confidence intervals, hypotheses, Type I/II errors, p-values, and significance with pharmaceutical context.
Compare fictional dissolution samples, then report the p-value together with the confidence interval, an assumption, and one design limitation.
Unit 4: Basics of Correlation & Regression
Interpret Pearson correlation, positive and negative association, scatter plots, simple linear-regression coefficients, and odds-ratio vocabulary in fictional pharmaceutical datasets. Association alone does not establish cause or clinical risk.
Explain why a visible dose-response association may still reflect study design, confounding, or uncertainty rather than a causal mechanism.
Unit 5: Statistical Analysis Using Python — Case based learning
Combine descriptive statistics, correlation, and linear regression in a transparent fictional case using SciPy, Statsmodels, or scikit-learn vocabulary. Read output summaries and p-values carefully, then prepare a bounded statistical report.
Draft a short statistical report for a simulated dissolution comparison, including question, data check, result, p-value, limitation, and no-decision boundary.