Evidence-Based Medicine

Timing of bias

Type of bias


Before a clinical trial begins

Selection bias

Study subjects improperly chosen

Channeling bias

Prognosis of study subject influences group or cohort assignment

During a clinical trial

Interviewer bias

Variability in manner by which information is gathered or recorded from subject to subject

Chronology bias

Comparison group includes historical controls

Performance bias

Exposure to other factors apart from intervention

Recall bias

Variability in accuracy of subject retrieval of prior experience

Attrition (Transfer) bias

Excess study subjects lost to follow-up

After a clinical trial

Detection bias

Partiality when assessing outcomes

Expectation bias

Expecting a certain result can influence assessment of outcomes

Correlation bias

Correlation improperly implied as causation

Publication (Citation) bias

Favor distribution of studies showing a difference between groups as opposed to studies supporting null hypothesis

[3, 10]

Table 1.2
Bias in observational studies

Type of bias


Self-selection bias

Social, cultural, linguistic, and health values or barriers may promote or hinder patient enrollment in a study, screening program, or measured point of care

Lead-time bias

Influence of a screening program on calculation of stage-specific survival rates: patients living in areas with a screening program for disease X may be diagnosed earlier and therefore concluded to live longer with disease X than age-matched counterparts living in an area without a screening program

Ecological fallacy

Assumption that observed associations can be extrapolated from a population to an individual

Exclusion bias

Evaluating data leaving a particular group or groups out of a sample population

Referral bias

Influence of referral from primary to tertiary and subspecialty centers on makeup of a study population: an observational study consisting of a patient population from a tertiary or subspecialty clinic may have higher percentage of complex or severe cases and adverse outcomes

Spectrum bias

Change in the performance characteristics of a particular test when applied to different patient subgroups

[10, 11]

Levels of Evidence and Grading of Recommendations

When determining what is “best evidence,” the most rigorous study designs are, by nature, those least prone to errors from chance, confounding, and bias. Table 1.3 illustrates how study design directly contributes to a study’s position within the hierarchy of evidence. It should be noted that this tier system for assigning proportionate value to study design is by no means absolute; a flawed randomized control trial, for example (e.g., with inherent bias, poor design or control of external variables), would certainly not hold the same weight as a rigorously designed and executed study. Rather, this is to serve as a rough outline for how practitioners should approach impact of study design, and how committees who review available data establish guidelines and clinical recommendations .

Table 1.3
Levels of evidence

OCEBM level

Therapy/prevention, etiology/harm



Differential diagnosis/symptom prevalence study

Economic and decision analyses


SR (with homogeneity) of RCTs

SR (with homogeneity) of inception cohort studies; CDR″ validated in different populations

SR (with homogeneity) of level 1 diagnostic studies; CDR″ with 1b studies from different clinical centers

SR (with homogeneity) of prospective cohort studies

SR (with homogeneity) of level 1 economic studies


Individual RCT (with narrow confidence interval)

Individual inception cohort study with >80 % follow-up; CDR″ validated in a single population

Validating cohort study with good reference standards; or CDR″ tested within one clinical center

Prospective cohort study with good follow-up

Analysis based on clinically sensible costs or alternatives; systematic review(s) of the evidence; and including multi-way sensitivity analyses


All or none

All or none case series

Absolute SpPins and SnNouts

All or none case series

Absolute better-value or worse-value analyses


SR (with homogeneity) of cohort studies

SR (with homogeneity) of either retrospective cohort studies or untreated control groups in RCTs

SR (with homogeneity) of level >2 diagnostic studies

SR (with homogeneity) of 2b and better studies

SR (with homogeneity) of level >2 economic studies


Individual cohort study (including low-quality RCT; e.g., <80 % follow-up)

Only gold members can continue reading. Log In or Register to continue

Aug 28, 2017 | Posted by in OTOLARYNGOLOGY | Comments Off on Evidence-Based Medicine
Premium Wordpress Themes by UFO Themes