Presentation 7 PDF - Introduction to Epidemiology

1 7. ANALYTICAL & EXPERIMENTAL STUDIES AND BIAS Course Title: Introduction to Epidemiology Course Code: (EPI 213) Program: Public health Department: Public health College: Public health and health informatics Dr. Abdulra...

1 7. ANALYTICAL & EXPERIMENTAL STUDIES AND BIAS Course Title: Introduction to Epidemiology Course Code: (EPI 213) Program: Public health Department: Public health College: Public health and health informatics Dr. Abdulrahman Almujaidel  Epidemiology is the study of the distribution and determinants of health and disease in populations.  Epidemiological studies: descriptive and analytical.  Descriptive epidemiology describes the patterns of disease occurrence according to time, place, and person  Analytical epidemiology measures the associations between exposures and outcomes and investigates the causal relationships.  Experimental epidemiology is a subset of analytical epidemiology that involves manipulating the exposure or intervention in a controlled setting. 2 INTRODUCTION  Analytical epidemiology is the field of epidemiology that measures the relationship between a particular exposure and a disease, using information collected from individuals rather than the general population.  Experimental epidemiology is the phase of epidemiological research that uses an experimental model in studies aimed at confirming the causal relationship suggested by observational studies.  Both analytical and experimental studies can provide evidence for causation, but they also have limitations and challenges, such as bias, confounding, and ethical issues. 3 ANALYTICAL EPIDEMIOLOGY  Analytical epidemiology uses observational study designs that compare groups of individuals with different levels of exposure to a potential risk factor or a protective factor.  The main purpose is to identify associations between exposures and outcomes and to test hypotheses about the causes of disease or health conditions.  The most common types of analytical studies are cohort studies, case-control studies, and cross-sectional studies. 4 TYPES OF ANALYTICAL STUDIES  Cohort studies follow a group of individuals with an exposure or characteristic over time and compare their outcomes with another group of individuals who do not have the exposure or characteristic. Cohort studies can be prospective or retrospective and can measure the incidence and relative risk of disease.  Case-control studies compare a group of individuals who have a specific disease or outcome (cases) with a group of individuals who do not have the disease or outcome (controls) and assess their exposure history. Case-control studies are usually retrospective and can measure the odds ratio of exposure.  Cross-sectional studies measure the prevalence of exposure and outcome in a population at a single point in time and examine their association. Cross-sectional studies are also known as prevalence studies and can measure the prevalence ratio of exposure. 5 EXPERIMENTAL EPIDEMIOLOGY  Experimental epidemiology involves manipulating the exposure or intervention in a controlled setting and measuring the outcome in the experimental group and the control group.  The main purpose is to investigate causal relationships between exposures and outcomes and to evaluate the effectiveness and safety of interventions.  The most common type of experimental study is the randomized controlled trial (RCT), which randomly assigns individuals to the experimental group or the control group. Quasi-experimental studies are similar to RCTs, but they do not use randomization or have a control group. 6 TYPES OF EXPERIMENTAL STUDIES  Randomized controlled trials (RCTs) are the gold standard of experimental studies, as they can minimize the confounding and bias that may affect the results. RCTs can measure the efficacy and effectiveness of interventions, such as drugs, vaccines, or behavioral programs.  Quasi-experimental studies are used when RCTs are not feasible or ethical, such as when the exposure or intervention is naturally occurring or cannot be controlled by the researcher. Quasi-experimental studies can measure the impact or outcome of interventions, such as policies, programs, or environmental changes. 7 BIAS IN EPIDEMIOLOGICAL STUDIES  Bias is any systematic error in an epidemiological study that results in an incorrect estimate of the association between exposure and outcome.  Bias can affect the validity and reliability of the study findings, and lead to false conclusions or misleading recommendations.  Bias can arise from various sources, such as the selection of participants, the measurement of exposure and outcome, the analysis of data, or the interpretation of results.  Bias can be minimized by using appropriate study design, sampling methods, data collection tools, statistical techniques, and reporting standards. 8 SELECTION BIAS  Selection bias is a problem that happens when researchers choose who or what to study in a way that is not random. This can make the sample of people or things they study different from the whole population they want to understand. Examples of selection bias  Self-selection bias: When participants choose to join or leave a study based on their own preferences or characteristics, which may be related to the exposure or outcome of interest. For example, people who are more health-conscious may be more likely to enroll in a health intervention study than those who are not. 9 METHODS TO MINIMIZE SELECTION BIAS  Randomization: Making sure that the groups of people who receive different treatments are chosen by chance, so that they are similar in other ways. This helps to avoid unfair comparisons in studies that test new treatments.  Use of appropriate control groups: choosing a group of people who do not receive the new treatment, but are otherwise similar to the group who does, and who come from the same population. This helps to avoid false associations in studies that observe existing treatments.  Ensuring representative sampling: Choosing a group of people who match the features and diversity of the population that the study wants to apply to, and who are willing to participate and stay in the study. This helps to make the study results more reliable and relevant. 10 INFORMATION BIAS  Recall bias: is when people remember things differently based on whether they have a certain condition or belong to a certain group.  Observer bias: is when researchers see or record things differently based on whether the participants have a certain condition or belong to a certain group.  Measurement bias: When the measurement tools or methods used to assess the exposure or outcome are inaccurate, unreliable, or inconsistent across the study groups. 11 METHODS TO MINIMIZE INFORMATION BIAS  Standardized data collection methods: Using the same protocols, instruments, and procedures to collect and record the data from all the study participants.  Training and calibration of data collectors: Providing adequate training and supervision to the data collectors or the investigators.  Use of validated measurement tools: Using measurement tools or methods that have been tested and proven to be valid, reliable. 12 CONFOUNDING BIAS  Associated with both the exposure and the outcome, and that is not an intermediate step in the causal pathway between them. This can result in an overestimation or an underestimation of the true association between exposure and outcome. Examples of confounding bias  Age confounding: When age is a confounder for the association between an exposure and an outcome, because age is related to both the exposure and the outcome, and it is not a consequence of the exposure.  Socioeconomic status confounding: When socioeconomic status (SES) is a confounder for the association between an exposure and an outcome, because SES is related to both the exposure and the outcome, and it is not a result of the exposure. 13 METHODS TO MINIMIZE CONFOUNDING BIAS  Study design considerations: Incorporating strategies in the study design to control for potential confounders, such as matching, stratification, restriction, or randomization.  Statistical techniques: Applying methods in the data analysis to adjust for the effects of confounders.  Randomization in experimental studies: Using randomization as the primary method to prevent confounding bias in experimental studies. 14 PUBLICATION BIAS Publication bias occurs when scientific studies with positive or significant results are more likely to be published than those with negative or null results. This can create a skewed or incomplete representation of the evidence on a particular topic. Some consequences of publication bias are:  Distortion of the literature: Publication bias can lead to an overestimation or an underestimation of the true effect size or direction of the association between an exposure and an outcome, depending on the direction and magnitude of the bias. 15  Waste of resources: Publication bias can result in a duplication or a repetition of studies that have already been conducted but not published, because the researchers are unaware of the existing evidence or they want to confirm the positive results.  Ethical issues: Publication bias can raise ethical concerns about the integrity and accountability of the researchers, the editors, and the reviewers, who may have conflicts of interest or incentives to publish or suppress certain results. 16 METHODS TO ADDRESS PUBLICATION BIAS  Comprehensive literature search: This involves searching multiple databases, sources, and languages for relevant studies, regardless of their publication status or quality. This can help identify unpublished or hidden studies that may have different results from the published ones.  Inclusion of unpublished studies: This involves contacting authors, researchers, or organizations to obtain unpublished data or reports of studies that are not available in the public domain.  Registration of study protocols: This involves registering the design, methods, and outcomes of a study before it is conducted and making it publicly accessible. This can help prevent selective reporting of results and increase the transparency and accountability of research. 17 RECALL BIAS Recall bias is a systematic error caused by differences in the accuracy or completeness of the recollections retrieved by study participants regarding events or experiences from the past.  Cohort studies: In these studies, a group of people are followed over time and their exposure and outcome status are recorded. If the outcome occurs after a long period of time, the participants may have difficulty remembering or reporting their exposure accurately.  Case-control studies: In these studies, cases (people with a disease or condition) and controls (people without the disease or condition) are asked to recall their exposure to a risk factor in the past. Cases may be more likely to remember or report their exposure than controls. 18 METHODS TO MINIMIZE RECALL BIAS Recall bias can be minimized by using various methods to improve the quality and consistency of the data collected from the participants. Some of these methods are:  Standardized questionnaires: These are structured and pre-tested instruments that use clear and specific questions, multiple-choice answers, and memory aids to elicit information from the participants.  Use of objective measurements: These are methods that use biological, physical, or chemical markers to measure the exposure or outcome status of the participants.  Timely data collection: This involves collecting the data as close as possible to the time of the exposure or outcome occurrence. 19 CONCLUSION  Analytical and experimental epidemiology: These are types of epidemiological studies that aim to identify and quantify the associations between exposures and outcomes in populations.  Importance of minimizing bias in epidemiological studies: Bias is a systematic error that can affect the validity and reliability of the data and the results of epidemiological studies. Minimizing bias can help improve the quality and accuracy of the evidence and the conclusions.  Methods to address publication bias, recall bias, and detection bias: These are some of the common types of bias that can occur in epidemiological studies and some of the methods that can be used to reduce or prevent them. 20 Questions 21