Study designs
67 likes16,929 views
This document describes different types of epidemiological study designs, including observational studies like cross-sectional, case-control, and cohort studies as well as experimental studies like randomized controlled trials. It provides details on descriptive versus analytical epidemiology and discusses cross-sectional, case-control, and cohort study designs, including their characteristics, uses, strengths, limitations, and how to conduct them. Key points covered include how observational studies observe exposures and outcomes without intervention, while experimental studies test interventions, and how case-control and cohort studies are used to test hypotheses about disease risk factors.
Study designs
- 2. Types of epidemiological studies • Types of studies Alternative name Unit of study • OBSERVATIONAL STUDIES – Descriptive studies – Analytical studies • Ecological Correlation Populations • Cross-sectional Prevalence Individuals • Case-control Case-reference Individuals • Cohort Follow-up Individuals • EXPERIMENTAL STUDIES (Intervention Studies) – Randomized control trials Clinical Trials Patients – Field trials Healthy People – Community trial Community studies Communities
- 3. Difference between Descriptive and Analytical Epidemiology Descriptive Epidemiology Analytical Epidemiology Only one group studies At least two groups are studied for comparison At the start of study there is no explicit hypothesis regarding cause effect relationship At the start of the study there is definite hypothesis regarding an exposure possibly causing an outcome. The study ends in development of possible hypothesis regarding cause and effect relationship but does not confirm or reject such hypothesis At the end of the study it confirms or rejects the hypothesis with which it started
- 5. Reasons for doing a cross-sectional study • To assess the burden of disease in a population and to assess the need for health services. • To compare the prevalence of disease in different populations. • To examine trends in disease prevalence or severity over time.
- 6. Cross-sectional Studies: characteristics • Exposure and disease outcomes are determined simultaneously • Provides a snap shot view and hence measures prevalence, not incidence of disease; Example: community surveys • Know the existence of disease but not the duration • Often used to study conditions that are relatively frequent with long duration of expression (nonfatal, chronic conditions) • Not suitable for studying rare or highly fatal diseases or a disease with short duration of expression • Examining effects on physiologic variables (e.g. Liver enzyme levels, blood pressure, lung function)
- 7. Where can we do a cross sectional study • Any setting or location – Community – Schools – Worksite – Hospitals
- 8. Analytical epidemiology • The subject of interest is the individual within the population. • The object is not to formulate, but to test hypotheses. • Types- Case Control study and Cohort study • We can determine – Whether or not a statistical association exists between a disease and a suspected factor – If one exists, the strength of association
- 10. Definition of a Case-Control Study • A method of sampling a population in which – cases are identified and enrolled, and – a sample of the population (controls) that produced the cases is identified and enrolled. • - Exposures are determined – for individuals in both groups.
- 11. When to conduct a case-control study? • When disease has long latent period - Ex: Cancer, cardiovascular disease • When the disease is rare – Ex: Studying risk factors for birth defects • When exposure data are expensive or difficult to obtain - Ex: Pesticide levels in blood
- 12. Scientific principles for Case Control study • Define exposure beforehand • Define inclusion exclusion criteria for cases • Exclude from controls – with early signs of outcome or contra indications (or no chance at all) of exposure • Minimize recall, detection biases • Record unknown exposure • – as unknown, not absent • Manage participant bias
- 13. Five criteria for case-control studies 1. Both cases and controls - from the same population base 2. Controls -independent of exposure and represent non-diseased in base population 3. Disease - ruled out in controls 4. Exposure determination -same way in cases and controls 5. Sampling - ensures that had “control” been a “case” it would have been sampled.
- 14. Steps of Case Control study • Selection of cases • Selection of controls • Measurement of exposure • Analysis and interpretation
- 15. Steps for case control study Selection of cases :- Identification of suitable group of cases and controls. (a) Case definition : Diagnostic criteria of case specifically defined. (b) Eligibility : Only newly diagnosed cases. Sources of case :- Hospital (single/ multiple), general population in a defined geographical area during specified time period, fairly representative of all cases in community
- 16. Selection of controls :- Controls free from disease and similar to all cases except the disease. Source of controls :- Hospital –relatives of patients or pt.s of other disease and Community - neighbours, general population. Random sampling done. Matching : Important to ensure comparability of cases and controls matched by age, sex, occupation, socio- economic status etc. Measurement of exposure : It is obtained by interviews, by questionnaires or by studying past records. Important to study the association
- 18. Advantages of Case Control Study 1. Inexpensive requires only a few subjects and quick results obtained. 2. Well suited for diseases which have a long latent period. (cancer, AIDS,CVA,) 3. Well studied for rare outcome. 4. Well studied for conditions for which medical care is sought. 5. Helps in examine the risk factors
- 19. Disadvantages: 1. Not a good method for the study of rare exposure. 2. Don’t give any idea of incidence or prevalence only gives odds ratio. 3. Prone to various, selection, information, bias recall bias and observer bias.
- 21. Selection Bias – The cases and controls may not be representative of cases and controls in the population. There may be differences between the characteristics between cases and controls. Berkesonian Bias - The bias may arise because of different rates of admission to hospitals for people with different diseases. 9i.e. Hospital cases and controls). Interviewer’s bias – Bias may occur when the interviewer knows the hypothesis and also knows who the cases are. This prior information may lead him to question the cases more thoroughly than controls regarding a positive history of suspected causal factor.
- 22. Comparison of case control and cohort study Case control Cohort Retrospective (backward in direction) Prospective (forward in direction) From outcome to exposure From exposure to outcome Outcome has already occurred Outcome would have not occurred however in retrospective cohort the outcome may have occurred Neither incidence nor prevalence can be calculated only odd ratio calculated. Calculate the incidence in exposed and non-exposed and Risk Ratio calculated. Results obtained quickly Takes longer time. Potential problem of bias (Recall, Survivorship) Not an issue Loss of follow up not an issue. Potential problem of loss.
- 24. Concept of cohort Cohort is defined a group of people who share a common characteristic or experience within a defined time period (age, occupation, exposure) Birth cohort, exposure cohort Distinguishing features of Cohort study 1. Cohorts are identified prior to the appearance of the disease under investigation. 2. The defined study group are observed for a period of time to observe the frequency among them. 3. The study proceeds forward from cause to effect.
- 25. Types of cohort study : • Prospective (observational): data on exposure status and disease outcome are not available at the outset of the study. • Ambispective: exposure status has been collected in the past and can be obtained from records, but, disease outcome is unknown or incompletely known. • Historical (non concurrent): data on exposure status and disease outcomes is available from historical records. • Nested case-control design: cases and controls are drawn from the cohort study population (eliminates the recall bias inherent in classical case-control studies AND reduces costs).
- 26. Elements of cohort study 1) Selection of study subjects 2) Obtaining data on exposure 3) Selection of comparison group 4) Follow up 5) Analysis
- 27. Selection of study subjects a) General population b) Special groups i) Select groups ii) Exposure groups Obtaining data on exposure 1. Information obtained by personal interview or mail questionnaires 2. Review of records 3. Medical examination or special test 4. Environmental survey
- 28. Selection of comparison groups i) Internal comparison – Single cohort enters into the study and on the basis of information obtained, the members may be classified into several comparasion groups. ii) External comparison – When the information on degree of exposure is not available, it is necessary to put up an external control. Eg. Smokers and non smokers, cohort of radiologists and ophthalmologists. They should be similar in demographic and possibly important variables under study. iii) Comparison with general population rates – If none is available, comparison between general population rates of same geographic area.
- 29. Follow up : i) Periodic Medical examination of cohort members ii) Reviewing physicians and hospital records iii) Routine surveillance of death records iv) Mailed questionnaires or telephone call or periodic home visit.
- 30. Analysis i) Incidence rate of outcome among exposed and non- exposed. ii) Estimation of risk A) Relative risk (RR) – Incidence of disease rate among exposed X 100 Incidence rate among exposed a) Attributable risk (AR) Incidence of disease rate among exposed - incidence of disease rate among non-exposed Incidence rate among exposed X 100
- 31. Advantages of cohort study : 1. Scientifically stronger design compared to case control or cross sectional studies. 2. Temporal association more convincingly demonstrated. 3. No recall bias. 4. Can study many outcome of given exposure. 5. Provide direct incidence of outcome in exposed and non-exposed RR calculated. 6. Results are not biased due to survivorship. 7. Good for study of rare exposure. 8. Any change of exposure at the start of study can be recorded.
- 32. Disadvantage of cohort study : 1) Quite expensive 2) Needs large no. of subjects. 3) Results may take very long time. 4) Ascertainment bias due to differential assessment of exposed and non-exposed. 5) Some subjects may cross over. Some exposed may leave exposure and vice-versa during study. 6) Inefficient for study of rare outcome.
- 33. EXPERIMENTAL EPIDEMIOLOGY 1. Experimental studies are similar in approach to cohort studies excepting that the study condition are under direct control of the investigator. 2. There is some action, intervention or manipulation which are deliberate. AIM OF EXPERIMENTAL STUDY 1. To provide scientific proof of aetiological factor. 2. To provide a method of measuring the effectiveness and efficiency of health services for the prevention control and treatment of diseases. TYPES OF EXPERIMENTAL STUDIES 1. Animal study 2. Human study
- 34. Animal studies : 1. Contributed to our knowledge in Anatomy, Physiology, Pathology, Immunology, Genetics etc. 2. Experimental reproduction of human diseases in animals to confirm aetiological hypothesis and to study the pathogenesis. 3. Testing the efficiency of preventive and therapeutic measures like vaccine and drugs. 4. Completing the natural history of diseases. Advantages of Animal studies. 1. Experimental animals can be bred in laboratories and manipulated easily as per the wish of the investigator. 2. The Animals multiply rapidly.
- 35. Limitations : 1. All Human diseases can not be reproduced in animals. 2. All the conclusions drawn from animal experiments can not be strictly applicable to human beings. Human Experiment 1. Required to investigate disease aetiology. 2. To evaluate preventive and therapeutic measures. 3. More important for diseases that can not be reproduced in animals.
- 36. Important considerations: 1. Unquestionably most inclusive approach to scientific problem. 2. Ethical and logistic consideration often prevent its application to study the disease in human beings. 3. The volunteers are to be made fully aware of all possible consequences of the experiment. 4. It is equally on ethical if a drug or procedure is brought to general use without establishing the effectiveness.
- 40. Randomized controlled trials (RCT) are scientific investigations that examine and evaluate the safety / efficacy of new drugs or therapeutic/ preventive proceduresusinghuman subjects RCT can be used to: § evaluatethe safety of a new drug in healthyhumanvolunteers § assess treatment benefitsin patientswitha specificdisease § compare a new drug against existing drugs or against dummy medications(placebo) The result of the studies are considered to be the most valued data in the era of evidence-basedmedicine. 40
- 41. Randomized – Eliminate bias in selection / allocation – Balances all confounders: known or unknown Controlled – Intervention compared to a control – Control: active or placebo – Both groups identical except for intervention / exposure – Investigator has control over the process Trial – Experimental intervention – Effects unknown to investigator 41
- 42. Population hierarchy for intervention study Reference population Experimental population Exclusion criteria Informed consent Excluded Refused Study population Intervention group Control group Outcome Losses to follow-up Losses to follow-up Random allocation
- 44. Steps of Randomised controlled trials 1. Drawing of protocol 2. Selecting reference and experimental population. 3. Randomisation 4. Manipulation or intervention. 5. Follow up 6. Assessment of outcome
- 45. DESIGN OF A RANDOMISED CONTROL TRIAL SELECT SUITABLE POPULATION (Reference or Target Population) SELECT SUITABLE SAMPLE (Experimental or Study Population RANDOMISE EXPERIMENTAL GROUP CONTROL GROUP MANIPULATION AND FOLLOW - UP ASSESMENT 45
- 46. 1. Drawing of Protocol : (a) It specifies the aims and objectives. (b) Questions to be answered (c) Criteria for selection of study and control group. (d) Size of sample (e) Procedure for allocation of subject into study and control group. (f) Treatment to be applied when and where and how to what kind of patients. (g) Standardisation of working procedures and schedules. (h) Responsibility of the parties involved in the trial. (i) If needed a pilot study to study the feasibility, acceptability, efficiency of certain procedures.
- 47. Selecting reference and experimental population : • Reference population is that population to which the findings of the trial if successful are to be applied. • This population may be the whole mankind or be limited to certain population in specific age, sex, occupation and socio-economic group. Experimental or study population • It is derived from reference population. • Ideally randomly chosen. • The experimental population is defined and its members are invited to participate in the study.
- 48. Better to choose a stable population. Co-operation which assured to avoid losses to follow up. Criteria of the participating volunteers. 1. Give informed consent to participate in the study and informed about the purpose, procedure and possible dangers of the trial. 2. Should be representative of reference population. 3. Should be eligible for the trial.
- 52. 6. Assessment : Assessment is done in terms of the outcome. Positive result – Reduce the incidence of severity of disease and cost to health services. Negative result – Frequency of site effects and complication more. The difference of result in both groups tested vigorously for any statistical significance. Biases encountered 1. Subject variation 2. Observer variation 3. Bias in evaluation
- 53. To overcome bias the technique of blinding is adopted. 1. Single blind trial : The participants are unaware to which group they belong. (study/ control) 2. Double blind trial : The investigator or the participant do not know the group allocation. 3. Triple blind trial : The participants, investigators and the analysers do not know about the allocation of groups. This method is the best.
- 54. Types of Randomized Controlled Trials 1. Clinical Trials 2. Preventive Trial 3. Risk Factor Trials 4. Cessation Experiment 5. Trial of etiological agents 6. Evaluation of health services 54
- 55. Types of Randomized Controlled Trials 1. Clinical Trials: Concerned with evaluating therapeutic agent, drugs/interventions a) Evaluation of beta blockers in reducing cardiovascular mortality in patient’s surviving the acute phase of myocardial infarction. b) Trials of aspirin on cardiovascular mortality and beta carotene on cancer incidence. c) Randomized controlled trial of coronary bypass surgery for the prevention of myocardial infarction. 55
- 56. 2. Preventive Trial a) Vaccines b) Chemo-prophylactic drugs Analysis of preventive trials must result in clear statement about benefits to community, risk involved and cost to health
- 57. 3. Risk Factor Trials Investigator intervenes to interrupt the usual sequence in the development of disease for those individuals who have risk factor for developing the disease - Primary prevention of CHD using clofibrate to lower serum cholesterol 4. Cessation Experiment An attempt is made to evaluate the termination of a habit which is considered to be causally related to disease - Cigarette smoking and lung cancer 57
- 58. 5. Trial of etiological agents To confirm or refute an etiological hypothesis - Retrolental fibroplasia and administration of oxygen to premature babies. 6. Evaluation of health services effectiveness Domiciliary Vs Hospital Treatment in tuberculosis. - Domiciliary treatment of PTB was as effective as more costlier hospital or sanatorium treatment 58
- 59. Phases of clinical trials Phase – I Trial done on group of healthy individuals to know safety, efficacy and the side effects Phase – II carried in diseased group to know safety and efficacy done in multiple centers. Phase – III carried in thousands of people to study safety, efficacy and whether fit for manufacturing. Determine the effectiveness (overall benefit/risk-cost assessment) of new therapies relative to standard therapy Phase – IV :Post-marketing study as the drug has already been granted regulatory approval/license. • continuous on going process to know the long term effects, additional safety information, rare side effects. Often non randomized
- 60. Types of study design 1. Concurrent parallel study design 2. Cross over study design Non-Randomized trials 1. Uncontrolled trials – no comparison group 2. Natural experiment – Smokers and nonsmokers 3. Before and after comparison studies – with or without control
- 61. DEFINED POPULATION RANDOMIZED NEW TREATMENT CURRENT TREATNENT IMPROVED NOT IMPROVED IMPROVED NOT IMPROVED DESIGN OF A RANDOMISED TRIALPARALLEL STUDY DESIGN Patients are randomized into two groups: the new treatment or to the standard treatment and followed-up, to determine the effect of each treatment in parallel groups 61
- 62. RANDOMIZED GROUP 1 GROUP 1 GROUP 2 GROUP 2 GROUP 2 GROUP 2 GROUP 1 GROUP 1 OBSERVED OBSERVED THERAPY A THERAPY B CROSSOVER STUDY DESIGN randomize patients to different sequences of treatments, but all patients eventually get all treatments in varying order. The patient act as, his/her own control Time for elimination / carry over effect 62
- 63. SUBJECTS RANDOMIZED RAMIPRIL PLACEBO VITAMIN E PLACEBO VITAMIN E PLACEBO FACTORIAL DESIGN USED IN STUDY OF RAMIPRIL & VITAMIN E The HOPE Study: Ramipril vs Placebo; Vitamin E vs Placebo
- 65. End Points (Outcome Measures) • Clinical outcome (e.g, death, stroke, BP ) • Symptom/ sign • Lab abnormality 2 types, depending on nature : • Hard end points • Surrogate end points 65
- 66. Hard End Points SurrogateEnd Points A biomarker intended to substitute for a clinical end point • HbA1c • Serum cholesterol levels 66 • Generally clinical end-points • Death due to any cause • Death due to cardiovascular / diabetic causes • Non-fatal myocardial infarction/ strokes • End stage renal disease • Time to clinical event
- 67. SUMMARY
- 68. Different studies have different uses Objective Type of study Prevalence Cross-sectional Incidence Cohort Causal association Cohort Case-control Prognosis Cohort Natural history of disease Cohort Treatment effect Randomised Controlled Trial