Statistics Flashcards
Calculate: crude mortality rate
Calculated by dividing the number of the deaths by the total population size.
Calculate: Cause-specific mortality rate
Calculated by dividing the number of deaths from a particular disease by the total popluation size
Calculate: case fatality rate
calculated by dividing the number of deaths from a specific disease by the number of people affected by the disease.
Calculate: Standardized Mortality Ratio
Calcuated by dividing the observed number of deaths by the expect number of deaths. This measure is used sometimes in occupational epidemiology. SMR of 2.0 indicates that the observed mortality in a particular group is twice as high as that in the general population.
Calculate: Attack rate
an incidence measure typically used in infectious disease epidemiology. It is calculated by dividing the number of patients with diease by the total population at risk. For example, attack rate can be calculated for gastroenteritis among people who ate contaminated food.
Calculate: Maternal mortality rate
Calculated by dividing the number of maternal deaths by the number of live births
Calculate: Crude birth rate
Defined as the number of live births divided by the total population
What are the two basic measures of disease occurence?
Incidence and prevalence
Incidence rate
of new cases over a certain period
Prevalence
of cases over # number of total persons
Odds ratio
Probability of disease among the exposed/probability of disease among the non exposed (3/4)((a/c)/(b/d)). Described as the odds of disease is # among the exposed when compared to the unexposed. Associated with case control
○ odds of having disease in expose group / odds of having disease in unexposed group
= ad/bc
Relative Risk
Culmative incidence of disease among the exposed / culmative incidence of disease among the unexposed. (1/2)(a/a+b)/(c/c+d) Described as exposed are # times more likely to develop X when compared to the unexposed.
Associated with cohort studies.
○ probablity of getting disease in exposed group / probability of getting disease in unexposed group
= [a/(a+b)] / [c/(c+d
Attributable Risk
Decribed ast the number of cases which can be attributed to exposure. Calcuted by Culmative incidence of disease in the exposed minus the culmative indcidence of disease in the unexposed. (1/2)(a/a+b)-(c/c+d)
○ risk in exposed group - risk in unexposed group
= a/(a+b) - c/(c+d)
PAR
PAR%
PAR is # of cases of C among total population attributed to exposure (5-2) and PAR% is the number of cases
Lead time bias
E early dectection confused with increased survival
Late look or Length bias
Information gathered at inapproiate time eq survery to study a fatal disease(only those patients still alive will be able to answer
Type 1 error (alpha)
“find something false” FP, detecting a difference when there is no difference. Reject the null hypothesis when the null hypothesis is true. eq test say man is preggo when he is not. Or FP error. also known as described as alpha which goes up with more power and increase sample size, expected effect size, and precision in measurement
alpha equals
a=1-B
Type 2 error (beta)
“miss something” FN, missing a true difference. Failing to reject a false null hypothesis. eq stating that a woman is not pregnant when she actually is. Described as beta. which goes down or decreases with more power and increase sample size, expected effect size, and precision in measurement
Attributable Risk Reduction%
Done after intervention (AR among control)-(AR among intervention group)
NNT
NNT=1/ARR
Relative risk reduction RRR
RRR=ARR/RR among non exposed
Confounding variable
Just because umbrella come out during the rain doesnt meant that the rain cause umbrella
Dose-reponse
> increased level of exposure shows an increased relative risk of developing/odds ratio of having a disease
> can be used in OR or RR to support causality
Sensitivity (SN)
○ % with disease who test positive
= a/(a+c) = TP/(TP+FN)
Specificity (SP)
○ % without disease who test negative
= d/(b+d) = TN/(FP+TN)
Positive predictive value (PPV)
○ % positive test results that are true positives
= a/(a+b) = TP/(TP+FP)
Negative predictive value (NPV)
○ % negative test results that are true negatives
= d/(c+d) = TN/(FN+TN)
Sensitivity and specificity are intrinsic to the diagnostic test
○ do not change with prevalence
PPV and NPV do change with prevalence
Receiver operating characteristic (ROC) curves are a graphical depiction of a test’s performance
○ Y axis: sensitivity
○ X axis: 1-specificity
○ The higher the curve, the better the test
This is quantified by the AUC (area under the curve); an AUC of 0.5 states that the test performs no better than chance (bad test!), whereas an AUC of 0.9 suggests a better-performing test
Positive Skew
Asymmetrical with tail trailing off toright
Mean > median > model
Negative Skew
○Asymmetrical with tail trailing off toleft
Mean < median < mode
○ Mean very ???to skew
○ Median somewhat ????to skew
Mode very ???to skew
○ Mean very sensitive to skew
○ Median somewhat resistant to skew
Mode very resistant to skew
• A certain percentage of all observations will always fall within +/- certain standard deviations of the mean
○ +/- 1 Standard deviation = ???%
○ +/- 2 Standard deviations = ???%
+/- 3 Standard deviations = ???%
• A certain percentage of all observations will always fall within +/- certain standard deviations of the mean
○ +/- 1 Standard deviation = 68%
○ +/- 2 Standard deviations = 95%
+/- 3 Standard deviations = 99.7%
Capacity vs competence
○ capacity is a ???term
competence is a ????term
medical
legal
Always do this
• Avoid going to court
• Use trained medical interpreters when possible
• Committed mentally ill patients retain their rights
• Never abandon a patient
○ If a treatment (such as abortion, birth control, etc) is against a physician’s personal beliefs - that physician does not have to provide that treatment; however, they are responsible for referring their patient to a provider who is willing and able to provide such care
• Disclose all errors, regardless of harm
-If suspected abuse is occurring, physicians are mandated reporters and MUST report to Child Protective Services or Adult Protective Services
○ 17-year-old girl whose parents cannot be contacted
§ ???
○ 17-year-old girl living on her own
§ ???
○ 17-year-old girl who is requests birth control
§ ??
○ 16-year-old girl refuses but mother consents
§ ???
○ 16-year-old girl consents but mother refuses
?????
○ 17-year-old girl whose parents cannot be contacted
§ physician may treat a threat to health underin locum parentis
○ 17-year-old girl living on her own
§ patient can choose whether or not to give consent
○ 17-year-old girl who is requests birth control
§ provide access even in absence of parental consent
○ 16-year-old girl refuses but mother consents
§ treat
○ 16-year-old girl consents but mother refuses
do not treat
Case control
Observational
Compares the odds of being exposed between patients with disease and patients without the disease
eg Patients with cirrhosis are more likely to have been exposed to heavy alcohol use
OR = (a/b) / (c/d) = ad/bc
Relative Risk(RR) is measure of disease association
Observational
○ asks, “How much more likely are you to get cirrhosis if you drink alcohol?”
RR = [a/(a+b)] / [c/(c+d)]
Cross-Sectional Study
Observational
• Determines disease status and exposure/risk factor status at the same point in time
○ can show an association but not causality
measuresdisease prevalence
Factorial design study
Randomizes patients into different interventions with 2 or more variables being studied in each intervention
Crossover Study
• Participants alternate receiving intervention and placebo
• Participants act as own controls
○ improves power and precision of the study
all patients receive an intervention
Meta-Analysis
• Combines datafrom multiple studies
○ better precision than individual studies
○ improves the generalizability of study findings
○ considered to be the highest level of clinical evidence
○ limited by:
§ quality of individual studies
bias in study selection
Case series
• Report on observations of patients with known exposure or disease
Does not have comparison group so cannot perform hypothesis testing
1° (Primary prevention) ○ ??? § e.g.,??? 2° (Secondary prevention) ○ e??? § e.g., ??? • 3° (Tertiary prevention) ○ re????t e.g., ????
1° (Primary prevention) ○ preventdisease occurrence § e.g.,vaccination 2° (Secondary prevention) ○ early detectionof disease to either prevent or decrease morbidity from disease before onset of symptoms § e.g., colonoscopy • 3° (Tertiary prevention) ○ reduce morbidityfrom disease after symptom onset e.g., medication
Confounding
– A third factor is either positively or negatively associated with both the ???and ???
–Confounders arenot in the ????? pathway
— if not ????? for can distort true ???
either towards or away from the null hypothesis
Confounding
– A third factor is either positively or negatively associated with both the exposure and outcome
–Confounders arenot in the causal pathway
— if not adjusted for can distort true association
either towards or away from the null hypothesis
• Selection bias
Nonrandom ????of study participants leads to ????conclusions
• Selection bias
Nonrandom selectionof study participants leads to erroneous conclusions
• Measurement bias
○ Information is gathered in a way that distorts the information
§ ?????Effect
subjects ?????their ???when they ?????they are being ?????
• Measurement bias
○ Information is gathered in a way that distorts the information
§ Hawthorne Effect
subjects alter their behavior when they know they are being studied
• Procedure bias
Different ???not treated the ???
• Procedure bias
Different groups not treated the same
Pygmalion effect
nvestigator inadvertently conveys his high expectations to subjects, who then produce the expected result
A “self-fulfilling prophecy”
Golem Effect
is the opposite pygmalion: study subjects decrease their performance to meet low expectations of investigator.
Design bias??
The control group is inappropriately non-comparable to the intervention group
Observer bias???
Investigator’s evaluation is impacted by knowledge of exposure statu
Ways to Reduce Bias
- Randomization
- Use placebo as control
- Blind studies
- Crossover studies
Randomizationassociation to ITT
intention-to-treat analysis is used in order to preserve randomization
Examples of Effects that are Not Bias??
- Effect modification
- Latent period
- Generalizability
Crossover studies
> Subject acts as own control
>Limits confounding
Level 1 level of evidence
Randomized Controlled Trial (and meta analysis
Level 2 level of evidence
Prospective Cohort Study
Level 3 level of evidence
Retrospective Cohort Study
Case-Control Study
Level 4 level of evidence
Case Series
Level 5 level of evidence
- Case Report(a report of a single case)
- Expert Opinion
- Personal Observation
- Review
• SEM =???
○ SEM =??
○ SEM < σ
SEM ??as n ??
• SEM = standard error of the mean
○ SEM = σ/√n
○ SEM < σ
SEM decreases as n increases
• T-test
○ compares the means of ??
• ANOVA (analysis of variance)
○ compares the means of????
• χ2("chi-squared")
○ tests whether ??? variables are associated
○ used with 2x2 tables
e.g.,effect of treatment on disease• T-test
○ compares the means of 2 groups on a continuous variable
• ANOVA (analysis of variance)
○ compares the means of 3 or more groups on a continuous variable
• χ2(“chi-squared”)
○ tests whether 2 nominal variables are associated
○ used with 2x2 tables
e.g.,effect of treatment on disease
AR percent (ARP) is the attributable risk divided byincidence in the exposed (Ie)
ARP = (RR-1)/RR
○ ARP = 100* (Ie-Iu)/Ie =100*[a/(a+b) - c/(c+d)]/[a/(a+b)]
note thatrelative risk (RR) = Ie/Iu = a/(a+b) DIVIDED BY c/(c+d)
AR is incidence in the exposed (Ie) - incidence in the unexposed (Iu) = Ie - Iu
○ Ie = a/(a+b)
○ Iu = c/(c+d)
AR = a/(a+b) - c/(c+d)
RCA??
○ a problem solving method thatfocuses on finding the cause (the root)of a medical error in order to identify preventative measureswhich are subsequently implemented
• RCA attempts to answer
○ what happened?
○ why did it occur?
○ how can we prevent this from occuring again?
Failure Mode and Effects Analysis (FMEA)???
Definition: a technique used toevaluate risksin order to identify and eliminate known failure, potential failures, systemic, design, process, and service issues.
This technique isapplied toprevent medical errors from occuring in the first place
What should you look for when suspecting selection bias?
Major discrepancy in sample size limiting generalizability. Such as observation group being 105 in one cohort and 35 in the other.
Berkson bias
disease studies using only hospial based patient not applicabe to total populaition- selection bias
Prevalence bias - Neyman bias
exposure happend long before disease assessmennt which resulted missed diesese in patients that die early or recover. - selection bias
Attrition bias
significant loss in of study parciptatnts may cause bias if those lost to follow up differ significantly from remaining subjects.-selection bias
Surveillance bias
Risk factor itself causes increased monitoring in exposed group relative to the unexposed group, which increased the probability of identifying disease - observational bias.
Observational biases
Inaccurate measurement or classification of disease, exposure, or other variable
Selection bias
Inappropiate selection or poor retention of study subjections
Hazard Ratio
formula, the hazard ratio, which can be defined as the relative risk of an event happening at time t, is: λ(t) / λ0. A hazard ratio of 3 means that three times the number of events are seen in the treatment group at any point in time.
Absolute risk reduction ???????? Relative Risk Reduction ??????? Relative Risk ??? Number needed to treat ?????
Absolute risk reduction ARR: control rate-treatment rate Relative Risk Reduction RRR= ARR/control rate Relative Risk RR= treatment rate/control rate Number needed to treat NNT=1/ARR
Positive likely hood ratio
LR(+)=sensitivity/1-specificity higher the better
Negative Likely hood ratio
LR(-)=1-sensitivity/specificity less than 0.1 and lower is better
Confounder effects»_space;»»>and no associated with gradient..confounding can be reduced by ??????
Effect modifier only?????????? and is associated with a ??????, something ????usually revealed by stratification
Confounder effects both exposure and outcome and no associated with gradient..confounding can be reduced by randomization
Effect modifier only effects outcome and is associated with a gradient, something internal usually revealed by stratification
