1. Chapter 2 Epidemiological Tools for Health Promotion
© John Hubley & June Copeman 2013

2. Epidemiology is the study of the distribution and determinant of health-related states or events in specified populations, and the application of this study to control health problems. (Last et al. 2000)

3. Descriptive epidemiology Analytic
Describing the occurrence of disease according to people, place and time
Determining the causes and risk factors for health and disease
Figure 2.1 The role of epidemiology
Epidemiology is split into two main approaches- descriptive and analytic (see Figure 2.1).
The first approach is to ask questions to establish the health problem and the three key elements people, place and time:
What are the health promotion needs?
What are the health problems?
How many people are affected (people)?
Who is affected (people)?
Which communities are affected (place)?
What are the trends? (changes over time)
The second approach is mainly centred on answering the question Why did it happen? This involves making comparisons between groups with and without a disease and between groups exposed and not exposed to a possible causal factor.

4. Community-determined needs
Felt and expressed needs of community, e.g. ‘wants’
Health service-determined needs
Determined from health data
– prevalence, incidence,
mortality, morbidity
Figure 2.2 Health Service- and community-determined needs
Health service-determined needs. These are needs as identified by health and other services. Typically these focus on the main concerns of health services, which are how to allocate effort and resources and are based on distribution of indicators of health and disease. They include -
Normative criteria. What is the situation here compared with the average for the city, region or country?
Comparative criteria. What is the situation in this location compared to other communities?
Time-based trends .Is the problem increasing?
Community-determined needs or ‘Wants’
expressed needs covering issues that individuals and communities have already brought to the attention of the authorities-through their contacts with health services, politicians, letters to newspapers etc.
felt needs include individual and community concerns including those which have not have been drawn to the attention of authorities.
Figure 2.2 shows the relationship between these two groups of needs. There may be areas of overlap but also perceptions of needs that differ between the health services and communities.
Overlap indicates needs identified by both health services and community

5. Assessing priorities
The classic epidemiological approach to setting priorities involves asking the following questions:
What is the size of the health problem?
Is the problem increasing, i.e. what are the trends?
How serious is the health problem?
How feasible would it be to prevent and control the problem?
Epidemiological questions are:
Are causes known?
Can it be prevented?
Will early detection improve treatment outcome?
Social factors to be considered are:
Is there public support for action?
Is the topic something that the community feels strongly about and would they become actively involved in preventive actions?
Are resources – staff, finance and time – available for tackling the problem?
Assessing priorities
The classic epidemiological approach to setting priorities involves asking the following questions.
What is the size of the health problem? The prevalence of the problem is the number of people affected.
Is the problem increasing i.e. what are the trends? When the incidence is increasing we use the term epidemic
Prevalence. Total number of persons in a population with a disease at a given date (point prevalence) or in a given period (period prevalence).
Incidence. Number of new cases of disease in a given time period (per year, month, week etc).
Epidemic. An increase in the number of cases over past experience for a given population.
How serious is the health problem? What impact is it having on death rates, sickness and quality of life? Can it be cured? Does it result in permanent disability?
How feasible would it be to prevent and control the problem? How feasible is it to tackle the problem? Answering this question involves considering both epidemiological and social factors.
Epidemiological questions are:
Are causes known?
Can it be prevented?
Will early detection improve treatment outcome?
Social factors to be considered are:
Is there public support for action?
Is the topic something that the community feel strongly about and would become actively involved in preventive actions?
Are resources –staff, finance and time - available for tackling the problem?

6. Establishing cause, risk factor or confounding factor
Cause: Any factor that can directly lead to disease, e.g. chemicals, radiation, micro-organisms, environment, lifestyle/behaviour.
Risk factor: Something which can increase the likelihood of disease but on its own is insufficient to cause disease (e.g. age, sex, family history, low income).
Confounding factors: These are factors which appear to be associated with a disease but in fact are not causal.
Cause. Any factor e.g. that can directly lead to disease, e.g. chemical, radiation, micro-organisms, environment, lifestyle/behaviour.
Risk factor. Something which can increase the likelihood of disease but on its own is insufficient to cause disease e.g. age, sex, family history, low income
One of the main challenges in epidemiology is to distinguish true causes from factors that might be associated but not causal – the so-called ‘confounding factors’

7. Analytical-epidemiology study designs: measuring associations
Ecological studies
Cross-sectional studies
Case control studies
The branch of epidemiology that is involved with establishing causal influences on health is called analytical epidemiology. The study designs fall into two broad groups.
Group one analytical epidemiology study designs
These study designs are useful for suggesting possible causes but do not have the power to prove causal relationships because they only measure associations.
Ecological studies. These compare disease in one community with another – for examples studies comparing the dental health of people in some communities and not others provided the first clues that fluoride in water might protect from dental decay.
Cross-sectional studies. These are surveys of samples of the population at a single point in time. Information from such surveys might show that people with high levels of disease also share characteristics – for example diet, housing, exposure to environmental hazards.
Case control studies. These compare a sample of people with the disease or condition under study with a control sample of people without the disease. The probability that the people with the disease are more likely to have a particular characteristic e.g. a life style factor, or exposure is compared with the control and called the odds ratio. An odds ratio greater than 1 suggests that the characteristic is more likely to be found in the people with the disease. Case control studies are particularly useful for studying rare diseases or when information is needed very quickly such as when a new health problem has arisen. The quality of case control studies depends heavily on the selection of controls and people’s ability to recall events which may have happened a long time in the past.

8. Case-control study summary
ADVANTAGES…
good for studying rare diseases
can use smaller sample sizes
cost/time effective when using previously collected (RETROSPECTIVE) exposures
DISADVANTAGES…
subject to bias (selection and recall)
can’t calculate incidence
selecting appropriate controls can be challenging

9. Measure exposure to condition under study and development of disease
Cohort study
Develop disease Do not develop disease Develop disease Do not develop disease
Exposed
Not exposed
Define sample
Follow up over time
Measure exposure to condition under study and development of disease

10. Advantages and disadvantages of cohort studies
Allows assessment of multiple risk factors
More control over group similarity than prevalence study
Data on pre-disease onset (prospective studies) avoids problem of recall bias allowing better determination of cause and effect
Samples tend to be more representative of the general population than say, case control studies, and results may be more generalizable
Disadvantages
Can never rule out presence of unidentified confounders (but can be adjusted if carefully planned)
Hard to measure exposure
Large number of subjects required
Potentially long duration for follow-up (problem of “drop-out” bias
Expensive
Not suitable for rare diseases

11. Relative and attributable risk
Relative risk provides a measure of the strength of a causal relationship.
Attributable risk - how much disease can be prevented and how many lives saved if we were to remove that causal factor?
Relative risk provides a measure of the strength of a causal relationship. The higher the relative risk, the more likely it is that there is a causal relationship. Attributable risk answers the question that politicians and policy makers often ask when faced with public health decisions - how much disease can be prevented and lives saved if we were to remove that causal factor?

12. Criteria for causality
Temporal sequence
Strength
Dose response relationship
Replication of findings
Biological plausibility
Consideration of alternative explanations
Experiment
Coherence with established facts
Box 2.7 Criteria for causality – adapted from Bradford Hill (1965)
Temporal sequence . Did exposure to the factor take place before the occurrence of disease? For example you would need to establish that a worker was exposed to asbestos before he or she developed mesothelioma. If exposure was only after the disease developed you would have to discount asbestos as a possible cause.
Strength. The stronger the association between the disease and the factor (i.e. the relative risk or odds ratio), the more likely it is that the relationship is causal.
Dose response relationship. If the likelihood of disease increases with greater exposure to the factor, this provides support for a causal relationship.
Replication of findings. Is the association also found when others repeat the study in different settings?
Biological plausibility. Does the cause-effect relationship fit in with what we already know about biology and disease. For example studies on the effect of radon gas on cancer are supported by laboratory demonstrations that radiation can damage DNA.
Consideration of alternative explanations. Have all the possible alternative causes been considered and is there evidence to rule them out?
Experiment. What evidence is there from intervention trials that the removal of the causal factor e.g. reduction in smoking, high salt diet, high sugar consumption can lead to improved health?
Coherence with established facts. Does the association should fit in with existing theory and knowledge. While in most cases this applies, it is important to bear in mind that occasionally a discovery might be made that is so radical and revolutionary that it overturns conventional wisdom.

13. Epidemiology and health promotion debates
Can emphasize the prevention of disease, reinforcing the medical model.
Is a study of populations.
Can involve complex research designs and elaborate statistical calculations.
One of the challenges is to develop further the social epidemiology of health and subjective wellbeing.