Study Designs PDF PUBH151 Spring 2025

Study Designs PUBH151, Spring 2025 1  attendance CDEM 3741 CMED 2  Learning Objectives Understand and differentiate the concepts of study design and research methods. Classify main types of research designs applied in clinical resea...

Study Designs PUBH151, Spring 2025 1  attendance CDEM 3741 CMED 2  Learning Objectives Understand and differentiate the concepts of study design and research methods. Classify main types of research designs applied in clinical research. Differentiate between descriptive and analytical studies Understand the aspects and types of descriptive studies (case reports, correlational surveys, and ecological studies) and recognize their advantages and disadvantages. Differentiate between observational and experimental studies. Understand aspects and types of analytical studies (cross- sectional, case-control, cohort, RCT, Quasi experimental studies) and recognize their advantages and disadvantages. 3 Study design vs Research methods Study design – A plan to ensure that the evidence obtained enables us to answer the initial question as unambiguously as possible - overall structure of the research study. 4  Study design vs Research methods Research methods – processes, procedures, and tools used to collect and analyze data. – Sampling mode, method of data collection (e.g. questionnaire, observation, measurements), design of questions, statistical analysis plans always confirm with study design. 5  Study Design vs Research Methods Research question/problem Determines the most appropriate Type of study planned? Study design Kind of results expected? Can be implemented by following specific Specific and most suitable methods to Research methods collect and analyze needed evidence 6 Study Designs Descriptive surveys, Ecological studies 7 Descriptive vs Analytic Studies – Descriptive Hypothesis generating, no comparator – Analytical Hypothesis testing with comparator group(s) 8 1-Descriptive Studies 9  Descriptive Studies Describe characteristics of disease (outcome) or exposure (risk factor) with regards to: – Populations (i.e. demographics, socio-economic status SES, lifestyle) – Geographic distribution/place – Frequency variations over time (i.e. seasonal variations observed in time-trend studies) Research question format (PO question). Generate hypothesis that can be later tested in analytic studies. 10  1- Case reports and case series (a group of case reports): Detailed descriptions of: One or a few patients Unusual disease, unexpected recovery/complication Exposure to something that, it is speculated, may have caused the disease *New syndrome described in gay men with similar histories and PCP pneumonia 11  “Time Trend” or “Secular Trend” studies describe a disease by measuring its occurrence over time using a series of measurements. Assess fluctuations in rates of disease with time, & if such fluctuations correlate with other changes in the community. 12 2- Descriptive cross-sectional studies (surveys) Examine the prevalence of a disease by measuring variables from the sample drawn from a population at one point in time. 13 3- Cross-sectional Ecologic studies (population correlational studies) In ecological studies the unit of observation is the population or community. Disease rates and exposures are measured in each of a series of populations. Uses data collected on population-level instead of individual-level. Example: Is there a relationship between the average income in different countries and CHD rates? 14  Ecological Study: example Researchers designed and conducted a study of CHD in several countries of the world relating CHD rate to average income by analysing country-level data. Across the countries studied, CHD rate was higher in the richer ones than in the poorer ones. Concluded that being rich increases your risk of heart disease… 15  Ecological Study: example However, when researchers repeated the study using individual-level data within countries, they found opposite results: In UK, USA and Sweden, poor people had higher CHD rates than rich ones. 16  Ecological fallacy CHD rate CHD rate 17 Income per country Income per individual  Ecological fallacy An erroneous assumption (inference) that an association observed at group level applies at individual level. Usually interpreted as a major weakness of ecological analyses. Ecological analyses, however, inform us about forces which act on whole populations. 18 Advantages and Disadvantages of Descriptive studies Advantages: Disadvantages: Quick and inexpensive. Cannot establish causality, conclusions are Use databases which are sometimes misleading already available. (due to lack of Efficient allocation of temporality, bias or resources, planning of confounding). education and health May be subject to promotion programs. ecological fallacy. Answer non-clinical research questions 19  2-Analytical studies: A-Observational 20  Analytical Studies: Observational or experimental Primarily hypothesis testing using comparator groups. Goal is to determine if intervention (or exposure) affects (or is associated with) outcome. Different analytical designs can answer different types of clinical questions 21 Advantages and disadvantages of Observational studies Advantages: Researcher simply records intervention and outcome, no intervention or manipulation. Useful for naturally occurring exposures or whenever it is unethical or impossible to randomize participants into treatment and control. (e.g. effect of high sugar intake on dental caries in children, effect of dust storms on asthma exacerbations) Can be time-saving and may apply efficient sampling (e.g. Study of rare diseases using case-control design) Disadvantages: More prone to bias and confounding than experimental designs. 22 Axes of observational analytical study designs Directionality Sample selection – Exposure – Outcome – Random/all Type of outcome – Incident – prevalent 23  Directionality It is important to distinguish between when measurements are actually made in the course of a study and direction of analysis by the study investigators: Exposure first  Cohort Outcome first  Case-control Both together  Cross-sectional 24  Sample selection Sample selection will follow directionality – Exposure selection -> Cohort – Outcome selection -> Case-control – Random/all -> Cross-sectional 25  Outcome type Incident -> Cohort Prevalent -> Case-control or cross-sectional Incident outcome: New cases of disease that occurred after the time of study initiation. Prevalent outcome: Disease cases identified during the study period (old +new) 26 2-Analytical A- Observational Analytical Cross-sectional study 27  Analytical Cross-sectional Study Cross-sectional (Are we related?) Examines the prevalence of a disease by measuring the variables from the sample drawn from a population at one point in time (a snapshot). Essentially a correlational study of individuals Explores correlations between exposure and disease (no directionality, cannot infer causation). 28  Steps in an Analytical Cross-sectional Study 1. Select a sample from the population. 2. Measure predictor (exposure) and outcome (disease) variables at a single time point (no specific direction, no follow up). 3. Estimate prevalence of outcome. 4. Assess correlation between exposure and outcome (using Odds Ratios OR). 29 30  Examples Research Question: In patients with Polycystic Ovarian Syndrome (PCOS), is neuro-endocrine dysfunction associated with irregular cycles? Identify PICO elements? P – PCOS patients E – Neuroendocrine dysfunction C – No neuroendocrine dysfunction O – Irregular cycles Study design? Cross-sectional analytical study 31  Examples Research Question: Is Positron Emission Tomography (PET) a sensitive and specific test in diagnosing coronary artery disease as compared to coronary angiography? P –Patients with CHD I – PET C – Coronary angiography O – Sensitivity and specificity Study design: Cross-sectional analytical study 32 Analytical A- Observational Case-control study 33  Analytic Studies - Observational Case-control – why me?? Purpose is to examine the people with or without disease (usually a rare disease) and find the differences in predictor variables that may explain why the cases got the disease and the controls did not. Typically retrospective: – Looks back in time to examine how certain exposures or risk factors may have influenced the occurrence of a specific outcome (e.g., a disease or condition). 34  Steps in a Case-control Study 1. Start by selecting cases and controls. Classify individuals based on presence/absence of disease: a. Select a sample from a population of people with the disease (cases). b. Select a sample from a population at risk that is free of the disease (controls) c. Each case is matched (ideally) with a similar control. 2. Compare exposure history between cases and controls 35 Case-control Study Design YOU ARE HERE 36  Study Base If cases are selected from a certain study base (a defined study population), controls should be selected from the same study base, except that they are not cases. For example, if cases are selected exclusively from hospitalized patients, controls must also be selected from hospitalized patients. This is important to minimize bias. 37  Example: Research question: In patients with Reye’s syndrome (a rare but serious childhood illness), is there an association between the use of aspirin and the development of Reye’s syndrome? 30 patients with Reye’s syndrome were accessible to the investigator for study. 60 patients were drawn from the much larger population of accessible patients who have had minor viral illnesses without Reye’s syndrome. Subjects in both groups asked about history of aspirin use. Subjects who had minor viral illness and took aspirin had 9 times the odds of developing Reye’s syndrome. 38 2 X 2 layout in a case-control study OR= ad/bc 39  Odds ratio Odds in cases Odds in controls Measure of strength of association Can be greater or less than 1 (zero to infinity) Dimensionless OR >1 positive association OR=1 no association OR