Making Sense of Critical Appraisal

Olajide Ajetunmobi

• Publisher: CRC Press
• Year: 2021
• ISBN:
  9781000423464
• Edition: 1
• Topic: Medicine
• Subtopic: Medical Theory, Practice & Reference

This handy pocket companion provides all the necessary guidance on how to understand medical research publications, read them critically and decide whether the content of those papers is clinically useful in the care of patients. Illustrated throughout with medically relevant examples, the accessible text encompasses all relevant aspects of study design and clinical audit to give a clear framework to support critical reading for the novice and more experienced reader.

Basic stats

1

Introduction

Descriptive statistics

Hypothesis testing and statistical inference

Estimation

Statistical tests

Correlation and regression

Non-essential information

Introduction

Critical appraisal is concerned with the acquisition of necessary skills with which to discern clinical research papers accurately. Although it is only natural that some degree of overlap does exist between the two subjects, critical appraisal should not really be regarded as just another aspect of medical statistics. This misconception often leads to an overestimation of the level of statistical knowledge required for critical appraisal.

In this chapter, only the statistical concepts that are relevant to the purposes of critical appraisal are discussed. These concepts are carefully presented in a gentle step-by-step approach with aid of ‘easily understandable’ vignettes as necessary.

A separate ‘Non-essential information’ section is included for the more adventurous reader who may also want to consult more formal statistical texts for a more in-depth treatment of the various statistical concepts discussed.

Descriptive statistics

Types of data

Data can be described according to a variety of characteristics. Distinguishing between different types of data is a fundamental issue in statistics because it determines the way that data are handled and presented (Figure 1.1).

Figure 1.1 Classification of data.

Categorical data

These are qualitative data that cannot be measured on a scale and have no numerical value. Instead, they are identified, named and counted. The different types of categorical data can be described as follows:

• Nominal data (binary) – These can only be grouped into two categories. Examples include either/or, dead/alive, improved/not improved, healed/unhealed, male/female observations, etc.

• Nominal data (multi-category) – These can be grouped into more than two categories. Examples include racial groups, religious groups, social class, hair colour, television channels or questionnaire responses such as ‘strongly disagree/disagree/don’t know/agree/strongly agree’, etc.

• Ranked data – These also have multiple categories but here, a hierarchy can be seen to exist between the respective categories. For example, following a head injury patients may be grouped into ranks of mild, moderate and severe impairment; chemotherapy treatment response can be ranked as ‘no response/partial response/partial remission/full remission’, etc. Ranked categorical data are also called ‘ordinal data’.

Ranked data can...

Table of contents

• Cover
• Half Title
• Dedication
• Title Page
• Copyright Page
• Table of Contents
• Introduction
• Chapter 1 Basic stats
• Chapter 2 Relevant study design issues
• Chapter 3 Critical appraisal of studies on diagnostic tests
• Chapter 4 Critical appraisal of cross-sectional surveys
• Chapter 5 Critical appraisal of prospective studies
• Chapter 6 Critical appraisal of retrospective studies
• Chapter 7 Critical appraisal of randomized clinical trials
• Chapter 8 Handling survival time data
• Chapter 9 Critical appraisal of systematic reviews
• Chapter 10 Critical appraisal of clinical audit
• Chapter 11 Critical appraisal of economic evaluations
• Chapter 12 Critical appraisal of qualitative studies
• Glossary of terms
• References and suggested further reading
• Index