Surveys and quasi-experimental designs

8


Surveys and quasi-experimental designs




Introduction


In the previous chapter, the basic features of the randomized controlled trial were presented. Randomized controlled trials (RCTs) allow the study of causal relationships between variables. Although the RCT is held up by some commentators, such as those using the Cochrane Collaboration framework, as producing the highest level of research evidence, most health sciences research does not use an experimental design. In this chapter we will examine a variety of approaches used for investigation of research questions that do not lend themselves to an experimental approach.


The specific aims of this chapter are to:




Non-experimental research designs


If experimental designs are supposed to provide the tightest possible control of extraneous factors, why should we resort to alternative non-experimental research designs? There are a number of reasons why non-experimental research designs ought to be employed instead of experimental designs:



1. Many variables are not amenable to experimental manipulation, i.e. they are unchangeable. For example, if the research question is concerned with sex differences in responses to heart surgery, then sex cannot be manipulated by the researcher. Similarly, if the researcher is interested in age differences, the ages of the participants cannot be altered by the researcher.


2. Often, it is ethically inappropriate to investigate research questions using an experimental design (see Ch. 6). For example, if a researcher wished to perform a study on the effects of smoking upon health, studying this in an intervention study would require the researcher to allocate participants randomly to the smoking or non-smoking group. Clearly, it is unethical to force some participants to smoke and others not to smoke. In intervention studies using a non-treatment control group, valuable and effective treatment might be withheld from participants. This situation involves serious ethical concerns that might lead to an experiment not being approved by the relevant ethics committee because of potential harm to the participants.


3. Experiments are best used to study simple causal relationships between variables, i.e. does the intervention cause positive therapeutic effects? However, many human diseases and illnesses are not determined by a single cause but rather by a number of causes interacting in a complex fashion. For example, heart disease may be caused by a combination of factors such as smoking, excessive stress, inappropriate diet or genetic factors. To identify such possible causal (or risk) factors, we need to study systems as they function in nature. That is, we should investigate patients in their natural setting, even with the difficulties this entails.


Depending on the nature of the evidence we require, non-experimental designs are appropriate alternatives to experimental designs. In the health sciences, experimental designs are focused on a rather narrow band of research questions that relate to the effectiveness of interventions and the causes of diseases. There are many other interesting research questions that do not relate to intervention effectiveness and, even when the questions do relate to interventions, there are circumstances in which non-experimental methods are more appropriate. Descriptive surveys are non-experimental designs, which provide data essential for progress in the clinical sciences and public health.



Surveys


Surveys are investigations aimed at describing accurately the characteristics of populations for specific variables (see Fig. 8.1). Surveys are commonly used in health research for the following purposes:



The statistics obtained from surveys can present us with an overview of the patterns of states of health, illness and the use of health services in a given community. In this way, we can gain insights into issues such as the prevalent causes of death or the health-related requirements of the population. The outcomes of the surveys can be the bases for hypotheses and theories concerning the causes of illness in a community. The area of health science concerned with such matters is called epidemiology.



Epidemiology


Epidemiology is the field of study that is concerned with the distribution and determinants of health and illnesses in groups of people. Epidemiology has been defined as the study of determinants of health and illness in a given population. Epidemiological studies may be descriptive, where researchers study the incidence and prevalence of health and illness, or analytical, where researchers aim to identify the multiple and interacting factors that determine health and illness in a specific community. Epidemiological investigations are essential to data collection in public health, as this discipline is concerned with the study of health populations. It differs from much clinical research in that it is oriented to the population or group rather than individual level. Epidemiology goes back to the time of Hippocrates who was concerned with the effects of environments upon the health of populations. The work of John Snow in the 19th century concerning the epidemiology of cholera is also considered to be a major landmark in the development of the discipline. Snow mapped the distribution of cholera cases in London and demonstrated that the water supply to different houses supplied by different water companies was involved with the transmission of the disease. The existence of the cholera organism was inferred from these observational data.


Descriptive epidemiology focuses on specific target populations and compares the occurrence of different health or disease states within these populations. Thus the numbers of cases, new cases and the relative risk of having the target condition are key components of the epidemiological approach. Two key terms in epidemiology are incidence and prevalence. The incidence of a disease is the rate at which new cases occur in a specified population during a specified period. The prevalence of a disease is the proportion of a specified population that have the target condition at a specified point in time.


These basic definitions are often used to construct indicators of population health that can be compared across different countries. The World Health Organization (WHO) maintains statistics concerning rates such as the birth rate, the infant mortality rate and perinatal mortality rate. The WHO website provides a wealth of comparative epidemiological data (http://www.who.int/en/).


Analytical epidemiology is concerned with how diseases are transmitted or spread within populations. The classic epidemiological approach conceptualizes disease within the framework of host, agent, vector and environment. The host is the human, the agent is the infection or health problem, the vector is the means by which the disease is carried (e.g. in the case of malaria this may be a mosquito) and the environment is the setting or mechanism that promotes the exposure. Disease can be transmitted directly (from person to person) or indirectly (via water, such as was demonstrated in Snow’s classic study of London water supplies). Although this framework was developed initially within the context of infectious diseases, it is actually much more broadly applicable. Contemporary epidemiologists also study the prevalence and determinants of non-communicable diseases (e.g. Parkinson’s disease) in the population.


The pattern of outbreaks of disease within populations is also an area of interest to epidemiologists. Common patterns of occurrence of diseases include endemic, epidemic and pandemic patterns. A disease is said to be endemic to a particular area if it habitually occurs within the area. An epidemic is the occurrence of a disease that is above and beyond the usually expected rate. We often talk of flu epidemics, meaning an unexpectedly high rate of the occurrence of influenza. A pandemic is a worldwide epidemic. We have become acutely aware of pandemics such as human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) and other feared pandemics such as severe acute respiratory syndrome (SARS), and, more recently, avian flu.


Epidemiology uses both experimental and non-experimental designs for collecting evidence. (Interested readers should consult Bonita et al 2006.)

Related posts:

  1. Critical evaluation of published research
  2. Measurement
  3. Standard scores and normal distributions
  4. Measures of central tendency and variability

Stay updated, free articles. Join our Telegram channel
Tags:
Apr 12, 2017 | Posted by in MEDICAL ASSISSTANT | Comments Off on Surveys and quasi-experimental designs

Full access? Get Clinical Tree
Get Clinical Tree app for offline access