Epidemiological methods
Download as PPTX, PDF29 likes51,868 views
The document outlines various epidemiological methods, including Bayesian methods and causal inference, emphasizing their application in research design and analysis. It discusses observational and experimental study designs, types of data analysis, and the importance of methodologies such as meta-analysis and longitudinal data analysis. Additionally, it provides insights into outbreak investigation processes, census vs. surveys, screening versus diagnosis, and various sampling theories used in epidemiology.
Epidemiological methods
- 1. Epidemiological Methods Epidemiology stringently focuses on the application of appropriate study designs and analysis to draw an inference. Bhoj R Singh Division of Epidemiology Indian Veterinary Research Institute, Izatnagar-243122, India
- 2. Types of Methods used Bayesian Methods: A general tool to explicitly incorporate prior knowledge and fit more complicated regression models. Causal Inference: Epidemiologists address causality as a primary target. The counterfactual theory of causation has provided a unified way of conceiving of, implementing and analyzing epidemiologic studies. Latent Variable Modelling: Many of the constructs in our studies are not directly measurable. Latent variables often called confounders are one way to combine multiple, incomplete measures of these constructs into usable variables in our models. Longitudinal Data Analysis: Longitudinal or repeated measures data are ubiquitous in epidemiologic research. Numerous techniques are used to analyze these data, including covariance pattern models, generalized estimating equations, random coefficient/growth curve modelling, survival analysis, and time series modelling, to name a few. Meta Analysis: Observational and interventional studies often provide the best available evidence, however individual studies may lack power to provide either definitive or hypothesis-generating knowledge. Thus, statistical techniques for combining data from several studies with similar hypotheses is becoming increasingly popular in epidemiology as Meta-analytical methods.
- 3. Study methods/designs in Epidemiology • Observational study Designs [Descriptive studies (occurrence and distribution) and analytical studies (testing validity of hypothesis) – Static Studies: Case report, case series and Cross-sectional Studies: Show only static description of a occurrence of the disease, determines point prevalence, some times help is formation of causal hypothesis. – Follow-up studies: To estimate period prevalence, and incidence and for development of causal and interventional hypothesis. Cohort study designs. • Experimental/ Interventional study designs: (to confirm the hypothesis) Have experimental/ intervention and control groups. • Quasi-experimental Study Designs – Case-control – Nested case-control – Case-cohort
- 4. When and where? Which study? • Descriptive: – Little is known about the problem – Rely on pre-existing data pertaining to who/, When? & Where? – To establish potential causal hypothesis. • Analytical: – Used when insight into various aspects of problem are already known. – Rely on generation/ development of new data. – To find the answer of Why? – To evaluate the causal hypothesis.
- 5. Indices used in different Methods Relative risk (RR): used in cohort studies to measure the strength of an association. RR = (incidence in exposed) / (incidence in non-exposed) Attributable risk (AR): Number of cases attributable to the putative risk factor AR = (incidence in exposed) - (incidence in non-exposed) Attributable fraction (AF): Maximum proportion of a disease in a population that can be attributed to a risk factor (or maximum proportion of a disease that would be eliminated in the absence of the risk factor) AF = (prevalence of exposure) x (incidence in exposed - incidence in non- exposed) / (incidence in the overall population) Odds ratio (OR): Used in case-control studies for estimation of relative risk. OR = (cases in exposed group x non-cases in not exposed) / (non-cases in exposed group x cases in not exposed group) Short term fluctuations: Epidemic Curves Long term Fluctuations: Trends
- 6. Observational study design measures of disease, measures of risk, and temporality Study design Measures of disease Measures of risk Temporality Ecological Prevalence (rough estimate) Prevalence ratio Retrospective Proportional mortality Proportional mortality Standardized mortality Proportional mortality ratio Standardized mortality ratio Retrospective Case-crossover None Odds ratio Retrospective Cross-sectional Point prevalence Period prevalence Odds ratio Prevalence odds ratio Prevalence ratio Prevalence difference Retrospective Case-control None Odds ratio Retrospective Retrospective and prospective cohort Point prevalence Period prevalence Incidence Odds ratio Prevalence odds ratio Prevalence ratio Prevalence difference Attributable risk Incidence rate ratio Relative risk Risk ratio Hazard ratio Retrospective only Both retrospective and prospective, Prospective only
- 7. Observational study design strengths and weaknesses Study design Strengths Weaknesses Ecological Very inexpensive Fast Easy to assign exposure levels Inaccuracy of data Inability to control for confounders Difficulty identifying or quantifying denominator No demonstrated temporality Proportional mortality Very inexpensive Fast Outcome (death) well captured Utilize deaths only Inaccuracy of data (death certificates) Inability to control for confounders Case-crossover Reduces some types of bias Good for acute health outcomes with a defined exposure Cases act as their own control Selection of comparison time point difficult Challenging to execute Prone to recall bias No demonstrated temporality Cross-sectional Inexpensive Timely Individualized data Ability to control for multiple confounders Can assess multiple outcomes No temporality Not good for rare diseases Poor for diseases of short duration No demonstrated temporality Case-control Inexpensive Timely Individualized data Ability to control for multiple confounders Good for rare diseases Can assess multiple exposures Cannot calculate prevalence Can only assess one outcome Poor selection of controls can introduce bias May be difficult to identify enough cases Prone to recall bias No demonstrated temporality Retrospective and prospective cohort Temporality demonstrated Individualized data Ability to control for multiple confounders Can assess multiple exposures Can assess multiple outcomes Expensive Time intensive Not good for rare diseases
- 8. Outbreak investigation Descriptive Epidemiology Why to investigate? Identify the source (and eliminate it), Develop strategies to prevent future outbreaks, Evaluate existing prevention strategies, Describe new diseases and learn more about known diseases, Address public concern, It’s your job! 10 Steps in the Process 1. Identify investigation team and resources 2. Establish existence of an outbreak 3. Verify the diagnosis 4. Construct case definition 5. Find cases systematically and develop line listing 6. Perform descriptive epidemiology/develop hypotheses 7. Evaluate hypotheses/perform additional studies as necessary 8. Implement control measures 9. Communicate findings 10. Maintain surveillance
- 9. Census, Survey and Screening • Census: No sampling • Surveys: To map a disease or problem. Sampling required. – Cross Sectional (cross sectional studies) and – Longitudinal (cohort studies, panel studies and trend studies) – Methods • Questionnaires – Advantages: Ideal for asking closed-ended questions; effective for market or consumer research – Disadvantages: Limit the researcher’s understanding of the respondent’s answers; requires budget for reproduction of survey questionnaires. No response problem • Interviews – Advantages: Follow-up questions can be asked; provide better understanding of the answers of the respondents – Disadvantages: Time-consuming; many target respondents have no public-listed phone numbers or no telephones at all. Interviewer effect. • Screening: – Mass screening. No sampling. It is the application of a test to detect a potential disease or unidentified disease or condition or problem in a group, a farm or a population who has no known signs of that disease or condition i.e., apparently healthy. Aim is to size/map the iceberg of the disease. – Targeted screening/ selected or High risk screening: No sampling – Multipurpose: For more than one problem using more than one test – Multi-phasic: Different tests are used as for Brucellosis MRT, Slide test, STAT, ELISA – Opportunistic: case finding to bring the case for treatment, most of the modern hospital do it by free camp etc.
- 10. Concerns in Surveys • Population issues: Is individual identity there? All members of population are equally available or accessible for sampling? Geographical barriers? Co-operation! • Sampling issues: Accessibility for re-sampling, finding of the sample, availability for sampling etc. • Question issues: Formulation of case definitions, questions and way of asking questions! Type of questions, Response and response scale. • Bias issues • Administrative issues • Financial issues • Temporal issues (time available) • Personnel issues • Facilities
- 11. Screening Diagnosis Done on apparently healthy individuals Done on diseased or sick individuals Applied on groups, farms or population Applied on individuals Results are arbitrary and final Diagnosis is never final Based on one criteria and cut-off limit Based on diagnostic test results, laboratory findings & a number of signs and symptoms Based on less accurate, highly sensitive tests with moderate specificity More accurate Less expensive and quick More expensive Not the basis of treatment Used as basis of treatment Initiative comes from investigator, authorities or administrators Initiative comes from patient owner or on recommendation from a clinician Screening versus Diagnosis
- 12. Sample and Sampling Theories • Random sampling: In this case each individual is chosen entirely by chance and each member of the population has an equal chance, or probability, of being selected. • Systematic sampling: Individuals are selected at regular intervals from a list of the whole population. • Stratified sampling: In this method, the population is first divided into sub-groups (or strata) who all share a similar characteristic and then on each strata either random or systematic sampling can be done. • Clustered sampling: In a clustered sample, sub-groups of the population are used as the sampling unit, rather than individuals. • Quota sampling: This method of sampling is often used by market researchers, each researcher is given a defined quota for taking survey. • Sampling of ease or Convenience sampling: Convenience sampling is perhaps the easiest method of sampling, because participants are selected in the most convenient way, and are often allowed to chose or volunteer to take part. • Snowball sampling: This method is commonly used when it is hard to reach target or population groups. Existing samples or individuals in the study are asked to nominate further subjects known to them, so the sample increases in size like a rolling snowball. • Strategic Sampling or Targeted sampling • Sequential sampling • Multistage sampling