The importance of public health for sustainable, safe and healthy societies is becoming increasingly recognized around the world. The public health achievements of nineteenth-century Europe were overshadowed for much of the twentieth century by advances in personal care, especially hospital care.
List of fi gures
List of tables
Preface
The structure of the book has been changed to aid the flow of learning. The number of activities has been increased and we have moved the comments to the end of the chapters for this new edition to help readers test their understanding of the material more effectively.
Acknowledgements
Whether your interest is in clinical or public health medicine, the study of epidemiology is key to improving health. The goal of epidemiology is to use these methods and the resulting data to improve health and survival.
Overview of the book
It provides an overview of the main concepts and methods of epidemiology as a basis for further study. You may find the index at the back of the book useful for finding terms you are unsure of or want to review.
Key principles of epidemiology
The study of epidemiology provides the evidence base for appropriate public health policy, planning and practice. Epidemiology is the study of the distribution and determinants of health conditions or events in specified populations, and the application of this study to control health problems (adapted from Porta and International Epidemiological Association, 2008).
Principles of 1
1 Describe the distribution of cholera deaths in relation to the position of the water pumps in Figure 1.4. The most likely explanation for the sudden increase in cholera deaths would be exposure of the population to a causative agent from a common source.
Measuring the frequency of outcomes
Risk = number of new cases in a given time period Total number of individuals at risk in the population per. 1 What is the total number of person-months at risk of HIV infection observed in this study.
Measures of association 3
Second, you will be introduced to measures of impact: attributable risk, attributable fraction, population attributable risk, and population attributable fraction. If the relative risk is greater than one, then exposed individuals are at greater risk.
The incidence of the outcome in the total sample population can then be calculated as (a+c)/(b+d). The incidence in the exposed group is a/b and the incidence in the unexposed group is c/d.
Interpreting associations
For example, an apparent association between occupation and lung cancer may be the result of the occupational group being more likely to smoke (or smokers more likely to choose a particular occupation) and therefore at increased risk. of lung cancer. However, this means that the results of the study cannot then be extrapolated to women. In each of the examples below, there is an alternative explanation for the association between exposure and outcome.
6 Reversibility: relates to whether an intervention to remove or reduce the exposure results in the elimination or reduction of the outcome.
Epidemiological research studies
An epidemiological investigation begins with the identification of the association of interest and the selection of an appropriate study design. In this chapter you will learn about choosing an appropriate study design and protocol development and data management issues that are common to different epidemiologic study designs. Descriptive data provide information on the frequency of an outcome or level of exposure, but do not analyze a relationship between the outcome and exposure.
Routinely collected data include population censuses and population health surveys (further described in Chapter 6), vital registration systems, outcome registries and health facility data (further described in Chapter 12).
Study design and 5
Analytical studies measure the association between an exposure and outcome, with the goal of inferring causality. Descriptive data may come from routinely collected data sources or epidemiological studies and are used to identify health issues for further study. Descriptive epidemiological studies can be undertaken to assess whether the burden of an outcome is of public health importance, or as part of an analysis of the subsequent epidemiological situation.
5 A survey of the number of children with existing Ascaris lumbricoides (an intestinal worm) infection and the household's access to running water. In this chapter, you have reviewed the characteristics of each of the major epidemiologic study designs. 2 This is a cohort study which will prospectively measure the incidence of heart disease in relation to different exposures recorded at the start of the study.
1 This is a cross-sectional study of the relationship between the prevalence of malaria and the possession of mosquito nets.
Ecological studies 6
The age-specific rate is the incidence rate (number of cases divided by the person-time at risk) which is calculated separately for each age group. 1 Calculate the crude HIV-related death rate and the age-specific HIV-related death rate for country Y in 2010. 3 The age-specific HIV-related death rates appear to be almost identical in both populations.
Age group (years) HIV-related deaths Population mid-year Age-specific HIV-related deaths per 100,000.
Cross-sectional studies 7
This is the ratio of the prevalence of the result in those exposed divided by the prevalence in those not exposed. 3 What is the prevalence ratio for anemia of the effect of having an ITN compared to not having one. 4 What is the prevalence ratio for malaria parasites from the effect of not having a net, either ITN or untreated.
3 To calculate the incidence ratio for anemia of the effect of ITNs compared to no net ownership, your 2×2 table should look like Table 7.4, with appropriate labeling.
Cohort studies
Study participants must be free of the result of interest at the start of the study. For example, if the result is to have cancer, it is important to ensure that all participants are cancer-free at the start of the study. Selection of the study population usually depends on whether the exposure of interest is common or rare.
A study of the effect of HIV infection on the incidence of pulmonary tuberculosis (TB) was undertaken in 2000.
Case-control studies
The sequencing of the human genome has led to the search for disease-susceptible genes, and the case-control design allows examination of several genetic markers (and possibly the entire genome) in the same study. However, common cases may differ from incident cases in ways that may reduce the validity of the study. Case-control studies begin with an outcome assessment, so the collection of information on exposures is almost always retrospective.
Therefore, they cannot directly estimate the prevalence or incidence of the outcome or the frequency of the exposure in the general population.
Intervention studies
4 Investigators may be able to conceal the allocation of the intervention from study subjects and those measuring the outcome, reducing information bias. Selection bias can also be introduced during the course of the study if there is a difference in the follow-up between the intervention arms. It is good practice for a statistician independent of the study team to keep the codes at the intervention allocation.
Given that the study is open-label, the systematic allocation of the intervention is unlikely to introduce further bias.
Epidemiology in public health
In previous chapters, you learned how to measure epidemiologic associations between exposure and outcome and how to estimate the impact of a risk factor. In this chapter, we consider how to use such evidence in different approaches to preventive intervention. You will be introduced to the concepts of primary, secondary and tertiary prevention and how the shape of the exposure-outcome relationship affects whether to use a population or high-risk approach for targeted prevention.
Screening is a key tool for prevention programs and you will learn about the measures of sensitivity and specificity used to assess the validity of a screening method and predictive values used to assess its usefulness at an individual level.
Prevention strategies 11
Consider that Table 11.1 has the same format as Table 3.1 (see Chapter 3) and think of the 'true state' as the 'result' and the 'screening method result'. Note that the sensitivity and specificity of a screening method is independent of the prevalence of the condition in the population being screened. Create a new 2×2 table using this new cut-off and calculate: a) the positive predictive value of FBS in the study population;.
If a large proportion of the outcome is due to those with higher exposure, then it would be appropriate to target preventive interventions at those at high risk.
Surveillance, monitoring 12
In this chapter, you will be introduced to public health surveillance systems and methods for monitoring and evaluating public health programs. Many of the research methods you have learned can be adapted for use in routine settings so that public health can continue to be evidence-based even after the research has ended. A country's health information system (HIS) aims to integrate the collection, analysis and reporting of data on health outcomes, to provide evidence to improve health services and ultimately the health of the population.
Repeated cross-sectional studies for specific diseases can also be performed as part of monitoring the health of a population.
For example, a sudden increase in the incidence of measles in a population can be explored along with possible explanations (for example, a decrease in vaccination coverage). This data can be used for health service planning, and can be compared across populations and tracked over time to identify problems from Figure 12.2 Total AIDS Cases Reported to the Centers for Disease Control in the U.S. from 1982 and 2002. If we have a good estimate of the number of infants in the community, we can calculate coverage.
Health facility data can provide information on the number of women attending for ANC, and demographic data can be used to estimate the number of pregnant women in the population.
Glossary
Frequency A measure of the number of occurrences of an outcome per population (see prevalence) or per unit of time (see occurrence). Intervention effectiveness A measure of the proportion of occurrence of an outcome that can be prevented by an intervention. Protective efficacy A measure of the proportion of occurrence of an outcome that can be prevented by a protective factor.
Relative risk An estimate of the magnitude of the association between exposure and occurrence of an outcome.
Index
