Behavioral science Flashcards
Cross-sectional study
It collects data from a group of people to assess frequency of disease (and related risk factors) at a particular point in time. It asks “what is happening?” It can measure disease prevalence and can show risk factor association with disease, but does not establish causality.
Case control study
It is retrospective. It compares a group of people with a disease to a group without the disease. It looks for prior exposure or risk factor. It asks “what happened?” It can measure odds ration. Eg patients with COPD had higher odds of a history of smoking than those without COPD.
Cohort study
It can either be prospective or retrospective. It compares a group with a given exposure or risk factors to a group without such exposure. It looks to see if exposure increases the likelihood of disease. The prospective study asks “who will develop disease?” and the retrospective study asks “who developed the disease (exposed vs non-exposed)?
Twin concordance study
It compares the frequency with which both monozygotic twins or both dizygotic twins develop the same disease. It measures heritability and influence of environmental factors (nature vs nurture).
Adoption study
It compares siblings raised by biological vs adoptive parents. It measures heritability and influence of environmental factors.
Clinical trial
Experimental study involving humans. It compares therapeutic benefits of 2 or more treatments, or of treatment and placebo. Study quality improves when the study is randomized, controlled and double-blinded. Triple blind refers to the additional blinding of the researchers analyzing the data.
Phase I clinical trial
It has a small number of healthy volunteers. The purpose is to see if it is safe, to asses safety, toxicity, and pharmacokinetics, and pharmacodynamics.
Phase II clinical trial
It has a small number of patients with disease of interest. The purpose is to see does it work, to asses treatment efficacy, optimal dosing, and adverse effects.
Phase III clinical trial
It has a large number of patients randomly assigned either to the treatment under investigation or to the best available treatment (or placebo). The purpose is to see is it as good or better, compares the new treatment to the current standard of care.
Phase IV clinical trial
It is postmarketing surveillance of patients after treatment is approved. The purpose is to see can it stay, detects rare or long-term adverse effects. It can result in treatment being withdrawn from market.
Evaluation of diagnostic tests (2x2 table)
A 2x2 table compares test results with actual presence of disease. Row is disease (+/-) column is test (+/-). Sensitivity and specificity are fixed properties of a test, while PPV and NPV vary depends on disease prevalence.
Sensitivity
It measures the true positive rate, the proportion of all people with disease who test positive, or the probability that a test detects a disease when the disease is present. A value approaching 100% is desirable for ruling out a disease and indicates a low false-negative rate. A high sensitivity is used for screening disease with a low prevalence. Sensitivity=TP/(TP+FN)= 1-false negative rate. SN-N-OUT=highly SeNsitive test, when Negative, rules OUT disease. If sensitivity is 100%, TP/(TP+FP)=1, FP=0, and all negatives must be TNs.
Specificity
It measures the true-negative rate, proportion of all people without disease who test negative, or the probability that a test indicates no disease when disease is absent. A value approaching 100% is desirable for ruling in a disease and indicates a low false-positive rate. A high specificity test is used for confirmation after a positive screening test. Specificity= TN/(TN+FP)=1-false positive rate. SP-P-IN= highly SPecific test, when Positive, rules IN disease. If specificity is 100%, TN/(TN+FP)=1, FP=0, and all positives must be TPs.
Positive predictive value (PPV)
It measures the proportion of positive test results that are true positive, the probability that a person actually has the disease given a positive test result. Positive predictive value=TP(TP+FP). PPV varies directly with prevalence or pretest probability leads to a high PPV.
Negative predictive value (NPV)
It measures a proportion of negative test results that are true negative, the probability that a person actually is disease free given a negative test result. NPV= TN/(TN+FN). NPV varies inversely with prevalence or pretest probability: a high pretest probability leads to a low NPV.
Incidence rate
Incidence rate= # of new cases/ # of people at risk. During a time period. Incidence looks at new cases (incidents).
Prevalence
Prevalence= # of existing cases/ # of people at risk. At point in time. Prevalence looks at all current cases. Prevalence approximates incidence for a short duration disease (eg common cold). Prevalence also approximates the pretest probability.
Odds ratio (OR)
In case control studies, subjects are chosen by outcome, not exposure. The odds ratio gives the likelihood of the subject developing the adverse outcome as compared to the placebo. The odds ratio approximates relative risk only when prevalence is low because disease with exposure+ no disease with exposure is approximately the same as no disease without exposure. OR=(disease with exposure/ disease without exposure)/ (no disease with exposure/ no disease without exposure)= (disease with exposure x no disease without exposure)/ (no disease with exposure x disease without exposure).
Relative risk (RR)
It is typically used in cohort studies. It measure the risk of developing disease in the exposed group divided by risk in the unexposed group (eg if 20% of smokers develop lung cancer vs 1% of nonsmokers, RR= .2/.01=20). If prevalence is low, OR approximates RR. RR= (disease with exposure/(disease with exposure + no disease with exposure))/(disease without exposure/(disease without exposure + no disease without exposure)).
Attributable risk (AR)
Attributable risk is the risk of an outcome attributable to a given exposure and is expressed mathematically as: AR = Incidence in exposed group (%) – Incidence in unexposed group (%). Example: A study examines the association of strokes with smoking cigarettes. In the exposed group (i.e. smokers), there is a 40% rate of strokes. In the non-exposed group (i.e. non-smokers), there is a 20% risk of strokes. The attributable risk is 20%.
Relative risk reduction (RRR)
It measures the proportion of risk reduction attributable to the intervention as compared to a control (eg if 2% of patients who recieve a flu shot develop the flue, while 8% of unvaccinated patients develop the flu, then RR= 2/8=0.25, and RRR=0.75). RRR= 1-relative risk
Absolute risk reduction
It measures the difference in risk (not the proportion) attributable to the intervention as compared to a control (eg if 8% of people who receive a placebo vaccine develop the flu vs 2% of people who receive a flu vaccine, then ARR=8%-2%=6%=.06) ARR=(disease without risk factor/(disease without risk factor + no disease with risk factor))/(disease with risk factor/(disease with risk factor/(disease with risk factor + no disease with risk factor))
Number needed to treat (NNT)
The “number needed to treat” is the number of patients who need to be treated for therapeutic benefit to be observed in one member of the study population. The measurement of number needed to treat allows comparison of efficacy between different treatments. It is often compared to placebo or no treatment groups. NNT = 1/(absolute risk reduction). Example: If the risk of developing lung cancer in Cincinatti is 1.5% and the risk in St. Louis is 0.5%, the number of people who would have to move from Cinncinatti to St. Louis for one person to avoid contracting lung cancer due to the move would be 100 (NNT = 1/ARR = 1/(0.015-0.005)).
Number needed to harm (NNH)
The “number needed to harm” is the number of subjects who must be exposed to a given risk factor for one person to be harmed. NNH = 1/(attributable risk). Example: Say that exposure to a certain level of benzene is shown to increase the risk of leukemia by 0.5%. The number of people who would have to be exposed to benzene for one additional case of leukemia to be demonstrated relative to controls would be 200 (NNH = 1/(attributable risk) = 1/(0.005)).
Precision
The consistency and reproducibility of a test (reliability). The absence of random variation in a test. Random error decreases the precision in a test. An increase in precision leads to a decrease in standard deviation and an increase in statistical power (1-beta, where beta is the probability of making a Type II error)
Accuracy
The trueness of test measurements (validity). The absence of systemic error or bias in a test. Systemic errors lead to decreased accuracy in a test.
Selection bias
An error in assigning subjects to a study group resulting in an unrepresentative sample. Most commonly a sampling bias. Examples include berkson bias, healthy worker effect, non-response bias. Strategy to reduce bias includes randomization and to ensure the choice of the right comparison/reference group.
Berkson bias
A type of selection bias. It occurs when a study population selected from a hospital is less healthy than the general population.
Healthy worker effect
A type of selection bias. It occurs when a study population is healthier than the general population.
Non-response bias
A type of selection bias. It occurs when participating subjects differ from nonrespondents in meaningful ways.
Recall bias
It occurs when an awareness of a disorder alters recall by subjects. It commonly occurs in retrospective studies. eg patients with a disease recall an exposure after learning of similar cases. A strategy to reduce bias includes a decrease in time from exposure to follow up.
Measurement bias
It occurs when information is gathered in a way that distorts it. For example, a miscalibrated scale consistently overstates weights of subjects. It can be reduced with the use of standardized method of data collection.
