Biostats Flashcards
High sensitivity test used for ________
screening
High _______ test used for confirmation after a positive screening test
specificity
Probability that a person who has a positive test result actually has the disease
Positive predictive value PPV
Probability that a person who has a negative test result does not has the disease
Negative predictive value NPV
How does raising the Sensitivity of a test affect
Specificity
NPV
PPV
↑Sensitivity
↑NPV
↓specificity
↓PPV
How does raising the Specificity of a test affect
Sensitivity
NPV
PPV
↑ Specificity
↑ PPV
↓sensitivity
↓NPV
How does Lowering the cut off value affect
Sensitivity
FP
Specificity
FN
- ↑ Sensitivity
- ↑ FP
- ↓ Specificity
- ↓ FN
How does Raising the cut off value affect
Specificity
FN
Sensitivity
FP
- ↑ Specificity
- ↑ FN
- ↓ Sensitivity
- ↓ FP
Should Medical errors be disclosed to patients, independent of immediate outcome (harmful or not)?
Yes
always
Even if it is about document hand offs
Occurs at level of frontline operator
(ex: wrong IV pump dose programmed).
Immediate impact
Active Error
Occurs in processes indirect from operator but impacts patient care
(ex: different types of IV pumps used within the same hospital)
Accident waiting to happen
Latent Error
Adverse Event That Is Identifiable, serious, and usually preventable
(ex: scalpel retained in a surgical patient’s abdomen).
An error that never should have happened
Never event
Retrospective approach. Applied after failure event to prevent recurrence
Uses records, interviews, data
Root cause analysis
Forward-looking approach. Applied before process implementation to prevent failure occurrence.
Uses inductive reasoning to foresee what may go wrong
Failure mode and Effects analysis
Patient comfort is prioritized (positive effect) over potential side effects (negative effect).
Principle of double effects
Program is available to patients ≥65 years old,<65 with certain disabilities, and those with end-stage renal disease
Medicare
(Medicare for elderly)
(Medicaid for desitute)
Criteria to establish Decision making capacity.
(Being retarded doesn’t count as an exclusion)
Informed
Mentally/Mood Stable
Age is over 18
Stable decision making history
Values are consistant
Goals are consistent
Expresses a choice
(IM A SaVaGE)
Determined by a doctor for a specific health decision
Capacity
Determined by a Judge for any/all health decisions
Competency
*A competent Judge works with a Capable doctor
Incapacitated patient’s prior oral statements commonly used as guide. If patient was informed, directive was specific, patient made a choice, and decision was repeated over time to multiple people, then the oral directive is more valid.
Oral advance directive
Patient designates anagent to make medical decisions in the event that he/she loses decision-making capacity. Patients may also specify decisions in clinical situations. Can be revoked by the patient if decision-making capacity is intact. More flexible than a living will
Medical power of attoreny
prohibits cardiopulmonary resuscitation(CPR). Other resuscitative measures that may follow (eg,feedingtube) are also typically avoided
DNR order
What is the priority of surrogate decision makers if a patietn loses capacity without an advance directive in place
(5)
- Spouse
- Children (over 18)
- Parents
- Siblings
- relatives
SPicy CHIPS
List 5 exceptions to patient confidentiality
- Suicidal/Homocidal patient
- Abused or Abusive patient (kid, senior, prisoner)
- Victims of potential harm by a patient
- Epileptic patients need to be reported so they can’t drive (other who have impaired driving skills too)
- Diseases like HIV, STIs, hepititis, Ebola, Food poisoning, TB need to be reported.
witholding information harmful to patient or that undermines decision making capacity
therapeutic privalege
List 3 things that increase in old people
Sleep latency (time to fall asleep)
Waking up early
suicide rate
Describe the 4 steps of disease prevention:
- Primary disease prevention
- Secondary disease prevention
- Tertiary disease prevention
- Quaternary disease prevention
- Prevent the disease (Vaccines)
- Screen for the disease (Pap smear/DRE)
- Treat the disease/complications
- Quit giving unnecessary medical treatments that can harm the patient (Imaging, polypharmy)
How to calculate
Case Fatality Rate
(CFR%)
Deaths from disease/ # People with the disease
x100
How to calculate
Number needed to harm
(NNH)
1/Attributable Risk
*Higher number = safer exposure
How to calculate
Number needed to treat
(NNT)
1/ Absolute Risk Reduction
*Lower # = Better treatment
How to calculate
Odds Ratio
aka: x’s more likely
ad/bc
* what are the odds that I was born: after death/before christ
How is Relative Risk calculated?
(RR)
if comparing 2 groups with 2 elements
(Diseased vs Healthy & Exposure vs No Exposure)
a/(a+b) / c/(c+d)
If comparing one group with 2 elements
(Diseased & Exposure vs No Exposure)
Exposure diseased pts/No exposure diseased pts
RR <1
RR =1
RR >1
↓ disease occurence (it’s good for you)
No association between risk/disease
↑ disease occurence (it’s bad for you)
The proportion of reduction of risk because of the intervention as compared to a control is called what and how is it calculated?
Relative Risk Reduction
RRR = 1 - Relative Risk
Relative risk = a/(a+b) / c/(c+d)
or
treated diseased pts/untreated diseased pts
Attributable Risk (AR)
is the difference in risk with exposed and unexposed groups.
How to calculate it?
a/(a+b) – c/(c+d)
How to calculate percent of attributable percent
(AR%)
(RR-1)/RR
x100
What is the difference in risk because of the intervention as compared to a control?
How is it calculated?
Absolute Risk Reduction (ARR)
c/(c+d) – a/(a+b)
Incidence looks at
New cases
Prevalcence looks at
all current cases
How is incidence is calculated
new cases/ people at risk
(per unit time)
How is prevalence is calculated
existing cases/total population
(at a POINT in time)
Prevalance > Incidence
In what case?
Chronic disease
(ex: diabetes → larger # of existing cases)
Prevalence = Incidence
When?
Short course illnesses
(Ex: the Flu)
How does increased survival time affect incidence & prevalence?
Increases prevalence only
How does Increased mortality affect incidence & prevalence?
Decreases prevalence only
How does Faster Recovery time affect incidence & prevalence?
Decreases Prevalence only
How does Extensive Vaccination affect incidence & prevalence?
Decreases
Prevalence & Incidence
How does decreasing risk factors affect incidence & prevalence?
Decreases
Prevalence and Incidence
Prevalence is increased by
increased survival time
Incidence is decreased by only what 2 things?
Vaccination
Less Risk Factors
What’s the difference between precision and accuracy
Precision: the outcomes are all consistently the same (Reliable)
Accuracy: The outcomes are all at or near the target goal (Validity)
What decreases a test accuracy (validity)
Systematic error
Bias decreases what in a test?
Accuracy (Validity)
Non-random sampling or treatment allocation
selection bias
participants selected from hospital only
Berkson Selection Bias
participants who are lost to follow up have worse outcomes
Attrition selection bias
awarness of disorder alters recall
recall bias
information is gathered in a non-systematic way due to faulty procedure or equiptment
Measurement bias
subjects in different groups are not treated the same or using different resources that could be the same
Procedure bias
The researcher’s beliefs or desires changes the outcome or documentation of the study
observer expectancy bias
An extra or hidden factor at play distorts the outcome of a study
Confounding bias
Early detection is confused with increased survival
lead-time bias
screening tests only effective for detecting diseases with a long latency vs those that are symptomatic earlier
have what bias?
Length time Bias
Randomization can reduce what two biases?
Selection bias
Length time bias
Placebos can reduce what three biases
Measurement
Procedure
Observer Expectancy
Blinding can reduce what 2 bias
Procedure
Observer-Expectancy
Objective, standardized, and previosuly tested methods reduce what bias
measurement bias
Multiple repeated studies, cross over studies, matching in groups reduces what bias
confounding bias
Measuring back-end survival
(survival according to severity)
eliminates what bias
lead time bias
Define Mean, Median, Mode
Mean: sum of values/total values (average)
Median: Middle value (middle)
Mode: Most common value (most)
(Standard deviation)2
Variance
Mean = Mode = Median
when
Bell shaped curve
Normal distribution
Mean>Median>Mode
Positive skewed curve
(aka left leaning)
Mean less than Median less than Mode
Negatively skewed curve
(aka Right leaning curve)
Null (H0)
means
no association
Alternative (H1)
means
There is an association
Stating that there is no effect or difference when none exists
null hypothesis not rejected
Stating that there is an effect or difference when one exists
null hypothes is rejected in favor of alternative hypothesis
Stating that there is an effect or difference when there is none
Alpha Type 1 error
Stating that there is NOT an effect or difference when there is one
Beta Type 2 error
If P < 0.05
what does it mean?
Statistically Significant
*results being by chance is less than 5%
If P > 0.05
what does it mean?
Results NOT statistically significant
Checks the differences between the means of 2 groups
T-test
T is meant for 2
Checks the differences between the mean of 3 or more groups
ANOVA
3 words: ANalysis Of VAriance.
Checks the differences between 2 or more percentages or proportions of categorical outcomes (not mean values).
Chi Squared Test
Chi-tegorical.
Checks the differences between 2 percentages or proportions of categorical, nominal outcomes.
Use instead of chi-square test with small populations.
FIsher’s Exact Test
The closer the absolute value of r is to _____, the stronger the linear correlation between the 2 variables.
1
*Correlation does not = causation
- Positive r value =
- Negative r value =
- Co efficient of determination =
- positive correlation (as one variable ↑,the other variable ↑).
- negative correlation (as one variable↑, the other ↓ variable).
- r2 (variance in one variable can be explained by variance in another variable).
LR+> 10 indicates
LR–<0.1 indicates
a highly specific test
a highly sensitive test.
LR+> 10 =
LR–<0.1 =
sensitive/ 1 – specific
TP/FP
(sensitive TiP FiP)
1- sensitive/ specific
FN/TN
(specific FaN TaN )
________ represents a study’s strength to detect a difference (ie, effect size) between treatment groups when one truly exists. It depends on _______ (among other factors).
Statistical power
sample size
** studies with greater sample sizes have greater power than studies with smaller sample sizes.
Cheap fast way to calculate
TP =
FN =
TP = (Sensitivity) × (Number of patients with the confirmed disease)
FN = (1 − Sensitivity) × (Number of patients with the confirmed disease)
Cumulative incidence (CI) is calculated as the total number of new cases of a disease over a specific period divided by the number of __________
people at risk at the beginning of the period
aka
(Total population – current cases)
The _____ is the ratio of the number of people who contract an illness divided by the number of people who were exposed to the illness or at risk.
attack rate
In a normally distributed curve
How do you calculate % of individuals 2 standard deviations away from the mean?
The center top of graph = MEAN
Out of a group of 100 individuals
2.35% are 2 SDs away from the mean.
*So 3 individuals (round up).
A _______ is a descriptive observational study design in which a group of patients with a similar diagnosis or treatment is described at a point in time or followed over a certain period. This study design has no comparison group; therefore, it cannot establish associations between risk factors (eg, treatments) and outcomes (eg, diseases).
case series
The power of a test is the probability of making the correct decision of rejecting a false H0 (ie determining there is a correlation when one truly exists).
If there is a 10% chance of concluding no relationship between 2 variables what is the power the study?
Power = 1 – ß
Power = 1 – .10 = .90
when ____ is assumed to be true (ie not rejected) it is informally interpreted as the probability that the observed results are due to chance.
H0
(not rejected = results due to chance)
A confidence interval (CI) that includes the null value for an RR (ie, RR = 1) is __________
not statistically significant
A Confidence Interval, CI, that excludes the null value
(ie, RR = 1) is _______.
statistically significant
Prevalence = (Incidence) x _______
(Duration of disease)
The odds ratio (OR) is a measure of association used in case-control studies. It quantifies the relationship between an exposure and a disease
(ie: everyone is exposed to same hazard, but let’s compare why some got the disease and some didn’t).
its null hypothesis value is always ____
1
(Odds Ratio = 1).
As disease prevalence increases, the positive predictive value _______, and the negative predictive value ________.
increases
decreases
__ is the maximum probability of making a type I error that a researcher is willing to accept.
α
*The value of α is typically set at 0.05, meaning that researchers are willing to accept up to a 5% chance of making a type I error.
___ is the probability of committing a type II error .Type II error occurs when researchers fail to reject the null hypothesis when it is truly false.
β
*if β is set at 0.2, the power will be (1 – β) = 80%; there will be an 80% chance of rejecting the null hypothesis when it is truly false (BAD).
Case-Control compares
Risk Factor frequency of effects
Cross sectional compares
disease prevelance
Retrospective cohort uses past records to compare
disease incidence
____ studies are organized by selecting a group of individuals, determining their exposure status, and then following them over time for development of the disease of interest.
Prospective cohort
The accuracy of screening or diagnostic tests is quantified by the area under the ROC curve (AUC). The more accurate the test is (ie, higher sensitivity and specificity), the closer the AUC value is to ____.
1.0
To calculate the probability of a series of independent events (ie that do not affect each other in anyway) happening all at once you must.
multiply their individual probabilities together
an independent event for example is a student taking a test. How one student does on the test does NOT affect how they other two students do.
IF the probability of one student a perfect score on the exam is 9% what is the probability that all 3 students will get a perfect score?
.09x.09x.09
Quick way to calculate
True negatives =
False positives =
True negatives = (Specificity) * (Number of patients confirmed without the disease)
False positives = (1 − Specificity) * (Number of patients confirmed without the disease)
Intention-to-treat analysis includes each subject in their initial randomization group even if subjects ______ or shift to a different intervention. This approach tends to provide a conservative but more valid estimate of the intervention effect in real-world scenarios (ie, clinical settings).
stop the intervention
A 2 × 2 table is normally used to record the presence or absence of exposure and disease in research. Rows and columns represent the different levels for each categorical (ie, exposure and disease) variable. The chi-square test for independence is used to evaluate the association between 2 ____ variables.
categorical
(adverse event rate in the control group) – (adverse event rate in the treatment group)
=
Absolute Risk Increase
