XClose

UCL Institute of Health Informatics

Home
Menu

Principles of Health Data Analytics

Analysing data derived from healthcare forms the cornerstone of evidence-based practice. Understanding the processes in which health data are generated and acknowledging what these data can (and cannot!) tell us is a vital skill for a health informatician. In Principles of Health Data Analytics, you will explore the key principles and concepts of statistics, operational research and machine learning as applied to healthcare. The module will guide you through the variety of mathematical and statistical techniques commonly used in research to improve the efficiency, productivity and quality of healthcare processes and systems. Additionally, the module will introduce you to techniques for analysing and evaluating the performance of organisations, including predicting demand, planning capacity and monitoring patient flow through a healthcare system (e.g. by minimizing waiting times).

Module code

CHME0001

UCL credits

15

Course Length

9 Weeks

Module Start Date

The CHME0001: Principles of Health Data Analytics 2020/21 module will become visible to students in the VLE on 22nd February 2021. This is to allow time before the formal start of the module for you to verify that you are ready and able to begin the module on the start date (e.g. can access the VLE without problems). It is important that you access the VLE within 24 hours of the start of the module so that you can read the online instructions for the activities of the first topic.

SYNCH Days

Wk 5: (Wed-Fri) 24 – 26 March 2021

Assessment Dates

04 May 2021

Module organiser

Prof Martin Utley Please direct queries to courses-IHI@ucl.ac.uk

Content

Content summary
ThemeDescriptionWeek
Introduction to the moduleThis week will explore the different roles and scope of analysis in health care, the types of decisions involved in
the design and delivery of health services and the types of
questions that arise in quality assurance and quality
improvement. You will gain exposure to what good (and
bad) health data analytics looks like.
1
The challenges of analysing
systems that have variability
In this week you will understand the difficulty in
interpreting outcome data through peer-to-peer discussion
and through finding and critically reviewing news stories
that report on findings from health data analytics.
You will be introduced to the concept of synthetic data
and to a simulated data set that will be used later in the
course.
2
Data visualisation: the good,
the bad and the ugly.
Here, you will:
  • Learn how visualisation is a key step for the analyst to understand features of data they are working with.
  • Be able to use visualisation as communication tool, through an appreciation of different types of graphs/visualisations.
  • Discuss the differences between design process for journalistic visualisation vs. analytic visualisation.
3
Introduction to Machine
Learning for analysis of
health care data
You will learn what we mean by ‘machine learning’ and be
introduced to some initial algorithms.
Using the simulated data set introduced in week 2, you will
begin to analyse these data using machine learning
approaches.
4
Applied Health Data
Analytics
The face-to-face week will be used to consolidate the
teachings from weeks 1-4 and form the foundations for the
topics covered in weeks 6-8.
This will include, but is not limited to:
  • Guest lectures showcasing great health data
  • analytics
  • Capacity planning workshop
  • Assignment briefing and tool for the assignment.
5 (face to face)
From decision trees to
Algogeddon
You will learn the uses (and abuses) of different machine
learning algorithms (i.e. what machine learning can and
cannot do).
We will cover best practices of how to develop a clinical
prediction models (and what should be avoided). Through
this, you will learn how to critically review published
prediction models to assess their reliability/ usability in
practice.
6
Demand, Capacity and FlowHere, you will begin to:
  • Appreciate the role of uncertainty in linking demand for services to the capacity required to deliver a service at a particular service standard.
  • Understand how the simple concept of traffic intensity can allow one to identify features of connected systems that can lead to congestion.
  • Consider to what extent historical data are the outputs of the current system, and the implications of this on the use of data in modelling.
  • Think about how service providers and commissioners could alter incentives to promote flow.
7
Barriers and Challenges in
Using Models and
Simulations
This week will unpick the non-analytical skills essential to
effective analysis of health data. For example, how do we
translate the results from statistical models, operational
research or machine learning to wider audiences (e.g.
policy makers).
8
Preparing for assignmentStudents will complete their assignments in this week.9

Teaching and learning methods

This 15 credit module lasts for 9 weeks and comprises roughly 150 hours of learning time, but with a break over the holiday period.. The module comprises in sequence:

  • A 4-week introductory phase that includes a short introduction giving the background and learning objectives of the module and a period of individual study, with structured learning activities accessed through the VLE.
  • A 2-week period including a period of 3 consecutive days of intensely interactive face-to-face learning that take place at UCL. The face-to-face sessions will provide an opportunity to engage further in the topics introduced in the first phase as well as introducing new material which will be developed further in weeks 5-6.
  • A further 3-week period of study, development and consolidation using resources accessed through the VLE. The main focus of this period will be undertaking activities related to the assessment and writing up your final report.
  • Final assessment.

For timetabling reasons, the face-to-face week is scheduled differently in different modules. In this module face-to-face teaching takes place in week 5.

Assessment

This module has a single summative assessment component, consisting of an individual written report. The assessment and marking criteria are outlined below.