Istilah dan Kapan Saja Bisa Terjadinya Bias. By Dr.dr. Ardik Lahdimawan, Sp.BS (K).

Summary

This video on epidemiology, presented by Dr. Ardiklah Dimawan, covers the concept of bias in research, explaining its various forms and when they can occur during the stages of a study. The video addresses biases that can appear in stages like reference search, sample selection, experimental intervention, measurement, data analysis, and publication. Dr. Dimawan elaborates on how biases such as selection bias, diagnostic bias, and publication bias can influence research outcomes. The video also emphasizes the importance of understanding and managing biases to ensure the integrity and validity of scientific studies.

Takeaways

• 😀 Bias can occur at multiple stages of research, from reference searching to data analysis and publication.
• 😀 Bias in reference searching includes rhetorical bias, one-sided references, and positive result bias.
• 😀 Selection and sample specification bias can arise from popularity bias, referral bias, and diagnostic bias among others.
• 😀 Intervention bias can happen due to non-randomized selection, diagnostic suspicions, and adherence differences between groups.
• 😀 Exposure and outcome measurement bias can include insensitivity, recall bias, and instrument calibration errors.
• 😀 Data analysis bias may result from skill degradation, rounding errors, and selective reporting of data.
• 😀 Interpretation bias can arise from mistaken identity, cognitive bias, and misinterpretation of statistical significance.
• 😀 Publication bias refers to the tendency to report favorable results while ignoring negative or inconclusive findings.
• 😀 The bias in clinical studies can be exacerbated by small sample sizes or improper matching of control groups.
• 😀 Bias can affect the accuracy of clinical diagnoses, especially if patient behavior or knowledge influences outcomes.
• 😀 Preventing bias requires awareness at every research phase, from design to publication, to ensure validity and reliability of findings.

Q & A

What is the main focus of the video transcript?

The main focus of the video is on the different types of bias that can occur in epidemiological research and how they can affect the validity and reliability of study outcomes.

What is reference search bias?

Reference search bias occurs when the researcher selectively chooses sources or literature that support their hypothesis or argument, excluding contradicting or controversial findings.

What are some examples of bias in sample selection?

Examples of bias in sample selection include popularity bias (where patients are concentrated in specific institutions), diagnostic suspicion bias (where prior knowledge of exposure influences diagnosis), and bias due to sample size (where small samples cannot provide conclusive evidence).

How does bias affect experimental studies?

In experimental studies, bias can occur through therapeutic personality bias (where the researcher's expectations influence results) and dropout bias (where participants who leave the study may affect the results).

What is recall bias, and how does it occur?

Recall bias occurs when participants in a study are asked about their exposure or behavior and may not remember accurately, leading to distorted responses, especially when cases are asked more frequently than controls.

How does data analysis contribute to bias in research?

Bias in data analysis can occur due to skill degradation bias (where rounding or measurement errors obscure differences between groups) or selective omission of data, leading to incomplete or misleading results.

What is significance bias in the interpretation of results?

Significance bias happens when results are deemed statistically significant, but they may not be clinically meaningful or relevant, often leading to misleading conclusions about the effectiveness of an intervention.

What is publication bias, and why is it a problem?

Publication bias refers to the tendency of researchers or journals to publish studies with positive or significant findings while underreporting studies with negative or neutral results, which distorts the overall body of evidence.

How can diagnostic tools contribute to measurement bias?

Diagnostic tools can contribute to measurement bias if they are not calibrated properly or if there are inconsistencies in how measurements are taken across different study subjects, such as with nadi (pulse) or blood pressure measurements.

Why is randomization important in experimental studies to avoid bias?

Randomization is crucial in experimental studies because it ensures that participants are assigned to treatment or control groups in a way that minimizes selection bias and ensures the groups are comparable.