1. Course Welcome and Outline

class: center, middle, inverse, title-slide

.title[
# 1. Course Welcome and Outline
]
.subtitle[
## HDS5103<br />Statistical Modelling in Health Data Science
]
.author[
### Neil O’Leary
]
.institute[
### University of Galway
]

---

## Outline

- Objectives

- Course outline

- Course details

- Course reading

---
## Learning objectives
- Learn how to do the following with statistical models in a wide-range of settings
  - **plan** (*choose*)
  - **implement** (*use*) 
  - **interpret**  (*peruse*)
  - **explain**  (*put across*)

- Write statistical reports based on a research question and some data to demonstrate your knowledge

- There will be lots of examples from epidemiology and clinical research studies

---
## Course outline
- **Section 1.** Statistical modelling fundamentals (Week 1-2)

- Principles and uses of regression
    - Linear regression recap
<br>
<br>
- **Section 2.** Generalised linear models (Weeks 3-5)

- Extending the linear model
    - Likelihood and information
    - Logistic regression
    - Poisson regression
    
---
## Course outline
- **Section 3.** Mixed-effects models (Weeks 6-9)

- Hierarchical models and covariance structure
    - Linear mixed-effects models
    - Logistic mixed-effects models  
<br>
<br>
- **Section 4.** Survival or time-to-event data (Weeks 9-12)

- Kaplan-Meier estimation
    - Cox proportional-hazards models
    - Other parametric models

---
## Course schedule

**Lectures** will be in person

- **Monday at 10am** in ADB-G021
- **Wednesday at 11am** in ADB-G021

**Tutorials** will start in Week 2

- **Thursday at 1pm** in ADB-G021

Tutorials will cover the exercise sheets from the previous week, a combination of R practicals or theoretical/interpretation problems

---
## Course plan

---
## What this course will not cover
Some elements of these, but you will encounter these in-depth elsewhere
<br>

- Clinical prediction models (HDS5101) 
- Causal inference (ST4120) 
- Bayesian Inference
- Extensive data cleaning, and processing

**This is an applied course.**

The algorithms of model fitting and the theory of statistical inference will only be covered to the extent that enables you to select, fit, interpret and critique statistical models for a given dataset and research question.

---
## Course outline

- **Lecture material**
    - Lectures: two hours per week in person
    - Slides
    
- **Tutorials**
    - Tutorials: one per week in person
    - Practicals
    
- **Assessment**
    - In-class assessments, written and using R (30%)
    - Exam (70%)

---
## R and Rstudio

.pull-left[
<img src="data:image/png;base64,#images/rdetails.png" width="100%" />
]

.pull-right[
- `lm` and `glm` in base R
- `lme4` and `nlme` packages
- `survival` package
- `tidyverse` for some data-formatting
- `broom`, `modelsummary` and other packages for summarising and plotting model outputs
]

---
## Course reading (free, online)
.pull-left[
<div class="figure">
<img src="data:image/png;base64,#images/ros_book.jpg" alt="https://avehtari.github.io/ROS-Examples/" width="80%" />
<p class="caption">https://avehtari.github.io/ROS-Examples/</p>
</div>
]

.pull-right[
<div class="figure">
<img src="data:image/png;base64,#images/amm_book.jpg" alt="https://ebookcentral.proquest.com/lib/nuig/detail.action?docID=1896285" width="75%" />
<p class="caption">https://ebookcentral.proquest.com/lib/nuig/detail.action?docID=1896285</p>
</div>
]

---
## Summary

- Welcome to the module!

- We'll cover a wide range of statistical models with lots of examples

- R and Rstudio will be used for data analysis

- Tasks for Week 1:

1. Look at worksheet at end of week
    2. Refresh your knowledge of simple and multiple linear regression