14 ⼀BIOSTATISTICS/BIOETHICS Flashcards
What is [Root Cause Analysis] ?
[5-part QA tool] that analyzes then addresses Root Causes leading to an adverse medical event
DCRSA
What are QALY and DALY?
________________
Describe both
________________
Quality-Adjusted Life Years | Disability-Adjusted Life Years
BOTH measures burden of disease for individuals or populations
________________
[TTO (time trade off)] is used for calculating QALY
________________
[Yeas of Life Lost (from premature Death)] and [Years of Life Lived with Disability] are used for calculating DALY
A patient, recently tested for Huntington’s disease, now declines to learn of the test result
What are the ethics guidelines regarding genetic information? (2)
🔲 Always Remember: PATIENTS HAVE RIGHT TO REFUSE GENETIC INFORMATION
BUT
⊙physicians should also be sensitive to psych impact of predictive screening and be willing to explore reasons for that pt’s anxiety
- “Learning about your test results is a very personal decision; I’d like to understand more about what led you to change your mind.”*
describe Rate Ratio
_________________
what’s the purpose of Rate Ratio?
[event occurred less in treatment group than control] < [RR = 1.0(null)] < [event occurred MORE in treatment group than control]
the difference from 1.0 indicates amount benefited or lost
_________________
measures effect of an intervention on an outcome over time
The ⬜ and ⬜ are common [test for heterogeneity]. What are [test for heterogeneity] useful for? (2)
▶[Q statistic(NO heterogeneity if P > 0.05)] :
a (small P<0.05) suggest there is a difference between the groups studied and that the [null hypothesis H0] should be rejected. A (Large P>0.05) suggest there is NO difference between groups studied and [null hypothesis H0] can NOT be rejected
▶[I2 index(heterogeneity level: [25%=low]/[50%=moderate]/[75%=HIGH])]
_________________
- performing meta analysis (of several trials as it can provide insight about combinability of multiple studies)
- comparing different trials
What does overlapping [Standard Error of Measurement] suggest?
suggest [NO significant statistical difference between 2 data sets]
(two data sets are similar)
[Case Control Study] is a study of the (⬜ past | present | future) that looks at [⬜ and ⬜ groups] to determine ⬜
_________________
What question does [Case Control Study] ask?
past ; ([Diseased] v [NONDiseased controls]) ; [RISK FACTORS that caused the disease]
_________________
After comparing the [Diseasded] and the [NONDiseased control], [what risk factor] increased odds of having the disease ?
observational and retrospective
- ________________*
- ” After review of* [group with new GI illness] and [group without new GI illness], having [history of eating at a particular restaurant] increased odds of having [new GI illness] “
in Medical law, the GINA act stands for ⬜
What is the GINA act? (2)
[GINA act] = Genetic Information Nondiscrimination Act ⼀
- prohibits employers and health insurance from requesting genetic testing
and/or
- requiring genetic testing to determine future employment and health insurance eligibility
[T or F] Sample Size and Power of a Study are interdependent
_________________
Explain?
TRUE
_________________
[power ⇪] and/or [small difference between groups] ➜ [⇪ Sample Size required]
Mortality prediction curves are useful for ⬜. How do you properly interpret these curves?
providing mortality risk while accounting for multiple other variables; each variable must be accounted for to arrive at the accurate predicted mortality
[T or F] Patients have a right to refuse knowledge regarding their diagnosis
_________________
What should Physicians do about this?
TRUE
_________________
nonjudgmental exploration into underlying reasoning ⼀ in order to understand patient’s concerns
state the inferiority designation for each event
A good ⬜ test will have high Sensitivity
How is Sensitivity related to NPV?
SCREENING; [⇪ sN = ⇪ NPV]
_________________
[⇪ sN = ⇪ NPV = ⬇︎FN]
Criteria for giving out Pt medical information? - 3
Pt must… PDA
1st: be Present (or otherwise available prior to disclosure)
2nd: have Decision Making Capacity (CURE)
3rd: Agrees to disclose information
Name and describe the [5 Ethical Principles in Medicine]
BJPAN
Describe how to interpret Standard Deviation chart
[68% of all observations will lie within 1 SD]
[95% | 2 SD]
[99.7% | 3 SD]
In terms of Medical Errors, describe
Near-Miss event
What is Statistical power?
_________________
How do you mitigate low statistical power?
ability to detect an association if that association exist. Based on sample size. Larger sample size helps control all confounders ➜ ⇪ Statistical power
_________________
META ANALYSIS (pools data from several studies to INC statistical power)
T or F
It is NEVER acceptable to allow industry-sponsored programs to influence lecture content
TRUE
Physicians have to retain FULL CONTROL over psntn content
What is the current recommendation regarding informal treatment of friends or family? (3)
for friends/family not technically your patient…
- ONLY IF no other physician is available.. you can give
- ACUTE care ONLY
- LIMITED care ONLY
**writing a prescription automatically establishes a medicolegal physician-patient relationship; making MD liable for any/all outcomes
“I would like to help you, but I am uncomfortable prescribing for someone I am not treating.”
What are the guidelines regarding patients themselves requesting copies of their medical records? (3)
▶ADULTS CAN! Adult pts have the right to release their medical records even to themselves (for their own personal record/viewing).
▶PEDS REQUIRE PARENT CONSENT (unless emancipated or received care that did not require parental consent)
▶Ok to Charge “Records Fee” (for printing/mailing/etc)
which kind of study is used to backtrace a restaurant outbreak?
CASE control
What is the Formula for Positive Likelihood Ratio?
Are Positive/ Negative Likelihood Ratios dependent or indepedent of a disease’s prevalence?
P = N / (1-P)
“Number 1 Nigga, Positivity” ….. “Positivity… Number 1 Plan”.
Negative LR = (1 - seNsitivity) / P
Positive LR = seNsitivity / (1 - sPecificity)
INDEPEDENT
Medical encounters may be the only way trafficked victims can get help
Describe what to do once you suspect a patient is being Human Trafficked
How should physicians approach [Complementary and Alternative medicine] with patients? (2)
- have open, explorative and collaborative discussion
- adequately research risk, benefit, and [proven treatment alternatives] for each CAM
[Per-Protocol Analysis] (3)
- Excludes non-compliant participants = analysis of only data from participants who’ve followed rules Per Protocol
- estimates true effect of an intervention but overestimate effect of intervention in a real world setting vs
- opposite of [Intention-To-Treat analysis (which keeps participants in their allocated groups regardless of dropout or non-compliance = better estimate of real world)]
What is [Intention To Treat Analysis]? (3)
▶after subjects are initially randomized to tx vs control, those subjects are maintained to that intervention group regardless of what they do during study period [switch intervention/drops out]→ crossover/attrition …actually become apart of the subjects outcome data and are attributed to the intervention
▶ will provide conservative (but more valid) estimate of the intervention effect
▶preserves benefits of randomization in superiority trials
define SpeciFicity
_________________
~~~
SpeciFicity formula ?
SPeciFicity= [true negative] = [TN / (TN + FP)] = used for test conFirmation
= “a test’s probability (in the absence of disease) a patient test negative”
What is [MLR (Multiple linear regression)] used for?
evaluates association between [1 quantitative dependent variable] and [≥2 independent variables of interest] while controlling for adjustment variables
What does setting statistical significance at [α = 0.05] actually mean?
statistical significance [α = 0.05] means…
Researchers are willing to accept that there’s a LOE5% chance that the observed differences they see will be considered significant when no difference actually exist
What are the recommendations regarding MD and teenager patients having sex.
Should MD discuss with parents?
MD should maintain confidentiality of sexually active teens and ensure teen access to healthcare and counseling
MD MUST BREAK CONFIDENTIALITY IF THERE’S RISK OF HARM TO SELF/OTHERS OR CHILD ABUSE
What is the Doctrine of Implied Consent? (3)
❗️In pts who LACK DECISION-MAKING CAPACITY
❗️but who require 9-1-1 Tx
❗️can receive 9-1-1 tx per [Doctrine of Implied (NO explicit necessary) consent] (and must be tx reasonable people in similar situation would expect)
What’s unique about Survival Analysis ? (3)
-it doesn’t just account for # of events for both the [control group] and [treatment group] …
-but also the TIMING OF THOSE EVENTS throughout the follow-up period for these 2 groups.
-This produces [TIME-TO-EVENT/DEATH] for the [control group] and [treatment group], which may reveal overall superiority between the 2
In research, what is Lead-time bias?
overestimation of survival only solely because patients were diagnosed earlier
_________________
when a screening test appears to prolong survival when actually it’s just informing the patient of their diagnosis earlier on …(and as a result…) time from diagnosis until death is longer ⼀even though there isn’t actual increase in survival (just earlier detection)
What is Number Needed to Harm?
_________________
formula? (2)
Number of people Needed to be exposed to a tx before a harmful event occurs
_________________
NNH
=[1 / ARI]
= {1 / [(AERTX) - (AERPLACEBO)]}
AER: Adverse Event Rate | ARI: Absolute Risk Increase
What is a Pragmatic study?
_________________
What is a Nested study?
seeks to determine whether intervention works in real life conditions
_________________
retrospective observational study in which subsets of controls are matched to cases and analyzed for variables of interest
What should a Specialist do if their plan does not coincide with PCP?
BEFORE MAKING ANY CHANGES, Specialist should call/discuss plan with PCP since PCP has primary responsibility for patient’s care
How is Likelihood Ratio used to stratify clinical significance?
LRs does not change as disease prevalence changes
and LR can be used to grade clinical significance of various results when >2 different test results are possible
_________________
LR = probability of a given test result occurring in a patient with a disorder compared to the probability of the same result occurring in a patient without the disorder
Elder abuse occurs in patients age ⬜, and if [risk factors ( ⬜4 )] are present, physicians should automatically screen these patients for [major signs of abuse ( ⬜4 )]
> >65[age>80], female, physical impairment, psych impairment(dementia/depression)
- spiral fx
- abrasions in unusual locations
- malnutrition
- pressure ulcers
interview patient alone and alert [Adult Protective Services] if elder abuse suspicion supported
Sensitive diagnostic test have lower false (⬜ negative | positive) results
________________
Specific diagnostic test have lower false (⬜ negative | positive) results
test with:
⇪ seNsitivity = ⬇︎false Negatives
________________
⇪ sPecificity = ⬇︎ false Positives
When are IV medications preferred over oral? (4)
- HDUS
- PO INtolerance
- PO meds failed
- < 2 yo
Main features of a DNR order -4
- ✔︎ [Ok to treat temporary or correctable conditions (i.e. sepsis)]
—-but NO VAP ——
- NO Ventilator dependency
- NO ACLS
- NO Prolongation of terminal illness
Recite the breakdown formula for Standard Deviation Curve (3)
[(mean) +/- (SD)] =
[(external SD)% chance > (mean + SD)] and
[(external SD)% chance < (mean - SD)]
In terms of research, describe the 3 characteristics of
Internal validity
[T or F]
It is paramount to determine guardianship status for all adult patients with intellectual disability
TRUE
_________________
especially as to clarify medical decisions
▨ What is a meta-analysis? (2)
_________________
◮ What is the disadvantage to using meta-analysis?
▨ -pooling data from several single studies in order to INC power of 1 LARGER COLLECTIVE study (as to INC that LARGER COLLECTIVE study’s ability to detect (if one exist) a difference in outcome between groups)
▨ If an outcome is rare (or difference between groups is super small), this will be difficult for single small study to detect ➜ unlikely to reach statistical significance = use meta analysis
_________________
◮ pooling from several studies will also pool their biases and limitations all into 1 analysis
Describe the 4 types of Disease Prevention
PSTQ = Prevent/Screen/Treat/Query
[Primary = Prevent future disease (* _P_revent future MI by reducing HLD*)]
[Secondary = Screen possible disease (_S_creen possible cervical CA with pap smear)]
[Tertiary = Treat current disease PATIENT CURRENTLY HAS to ⬇︎ disability from that current disease (_T_reat current severe CAD with CABG revascularization)]
[Quartenary = Query statistics of disease: to identify pts at risk for unnecessary treatment and prevent redundant or unnecessary therapies (shared EMR limits unnecessary or repeat procedures)
define Sensitivity
~~~
_________________
~~~
Sensitivity formula ?
seNsitivity = [true positive] = [TP / (TP + FN)]
= “a test’s probability (in the presence of disease) a patient test positive”
define
Specificity
sPecificity = [true negative] = Negative In Health
= “a test’s probability (in the ABSENCE of disease) a patient test negative”
What does Confidence Interval describe
⭐Range of values in which a [specified probability (usually 95%)] of the [means of repeated samples] are expected to fall within
⭐CI = mean +/- Z(Standard Error of Mean) (ex: [95% CI, Z=1.96] and [99% CI, Z=2.58]
⭐Statistical difference:
a.[95% CI for (mean difference between 2 variables)] that includes 0 = [NSD = H0 not rejected]
b. [95% CI for (odds|relative risk ratio)] that includes 1 = [NSD = H0 not rejected]
c.
CI of an [Odds Ratio] indicates (with a certain confidence level) whether a given OR is statistically different from [null OR = 1].
a CI that excludes [null OR = 1] is statistically significant
H0 = null hypothesis = [NO statistically significant difference exist]
In terms of research, describe the 3 characteristics of
EXTERNAL validity
if relationship is seen in populations outside study also??
[T or F] Romantic relationships between pediatric physicians and parents of their patient is ethically acceptable with full disclosure
_________________
Why or Why not? (3)
FALSE
- Pediatricians dating their patients’ family members is ethically PROBLEMATIC.
- Physicians are ethically obligated to act in best interest of patient .. so intimate involvement may blur judgment =
- these pediatricians need to transfer patient care
In the Kaplan-Meier survival curve graph, ≥2 study groups can be compared and are associated with p-value
What does p-value delineate?
p ≥ 0.05 means
the study groups are NOT statistically different
In terms of Medical Errors, describe the difference between
[Preventable adverse event] and [Negligent adverse event]
PAE = preventable event 2/2 faulty human Practice
NAE = preventable event 2/2 Not adhering to Standard of Care
Recall the [2 x 2 Test vs. Disease] diagram
[Cross Sectional Study] is a study of the (⬜ past | present | future) that looks at [⬜ and ⬜ groups] to determine ⬜
_________________
What question does [Cross Sectional Study] ask?
present ; ([Group with Risk Factor] v [Group wihout Risk Factor]) ; PREVALENCE of a disease
_________________
How prevalent is this [disease and its risk factors] right now??
________________
Collects data from a group of people to determine [frequency of disease and its risk factors] at a specific point in time
________________
what is the Hawthorne effect
when pts modify their behavior just because they know they’re being studied
What are the ethical guidelines regarding receiving gifts from patients? (4)
Since these influence professional judgment, it is Unethical to accept patient gifts that are ___X___
- given with expectation of preferential tx (vs genuine tokens of appreciation)
- LARGE
- EXPENSIVE
- intimate
A patient’s Employer hands you a signed “release of information” for that patient’s medical information.
How does this affect HIPPA?
If given written authorization, HIPPA allows MDs to give the minimum necessary information to satisfy the employer’s request
What is Hazard Ratio?
HR = the likelihood an event will occur in a treatment group (relative to the control)
[not likely/protective] < HR 1.0 < [LIKELY/DETRIMENTAL]
________________
[EVENT less LIKELY to occur in the treatment group than the control group] < HR 1.0
HR 1.0 < [EVENT MORE LIKELY to occur in the Treatment Group than the control group]
Sampling bias
selection bias that → unrepresentative sample
What is [External Validity] ?
how generalizable the results of a study are to other populations
(i.e. a study for middle aged women has low external validity for elderly men)
A 95% Confidence Interval that does not include the [null value 1.0] corresponds to p-value ⬜
<0.05
_________________
95% | <0.05
Describe the ANOVA test (2)
_________________
ANOVA: Analysis Of Variance
▶compares the mean between ≥3 groups
_________________
▶requires [quantitative dependent variable (outcome)] and [quaLitative independent variable (exposure/risk factor)]
▶[Standardized Mortality Ratio] is ⬜
▶▶Formula?
________________
Explain the overall takeaway
▶adjusted measure of overall mortality
——-▶▶by calculating ——–
[observed # of deaths in population of interest]
➗
[expected # of deaths in population of interest]
________________
SMR of 1.75 = “observed # of deaths (in this population of interest) is 75% higher than expected”
What is Number Needed to Treat?
_________________
formula? (2)
of patients needed to treated in order to [prevent 1 disease] or [cure 1 condition] = measures efficacy of a therapy
_________________
NNT
= [1 / ARR]
= {1 / [(failure ratePLACEBO) - (failure rateTX)]}
ARR: Absolute Risk Reduction
What are Advance directives used for?
AD (living will, DNR/DNI) are invoked ONLY WHEN PATIENT LACKS DECISION-MAKING CAPACITY
If patient still retains Decision-Making Capacity, M.D. should still ask patient if they want to refuse intervention (like Intubation) in the treatment of temporary / reversible illness (like PNA).
Describe the recommendations regarding Physicians asking for financial support from patients?
MD altogether should avoid soliciting financial support from patients (espeicially for personal endeavor)
talking about personal experiences is fine, as long as its relevant to patient
What’s the major takeaway with interpreting a ROC Curve Graph? -2
They use (1-specificity)
so
[⇪ (1-specificity) by 12%] actually = [⬇︎ specificity by 12%]
________________
moving up-and-right on ROC = [⇪ N but ⬇︎P]
What are the 5 primary components of a safe and effective patient handoff?
PCPAF
Patient
Clinical_Status
Plan of Care/Summary with anticipated d/c
Anticipated problems with recommended mitigation
Follow ups/Pending Actions/TO DOs
What is the difference between doing research with Incarcerated individuals vs general population? (2)
- IRB requires ADDITIONAL OVERSIGHT than gen pop (to ensure fairness to this vulnerable population)
- otherwise, same rights to refuse or consent to participate
In terms of disclosure, how should physicians approach medical errors? (4)
- ERRORS DISCUSSED WITH PATIENT AND FAMILY IN TIMELY MANNER = OLBIGATION!
- Empathetic apology
- Explanation
- Event prevention discussed
Your colleague’s patient informs you that your colleague performs inappropriate breast exams, that to which your colleague denies
What should you do?
REPORT SUSPECTED (SEXUAL) MISCONDUCT TO STATE MEDICAL BOARD FOR INVESTIGATION = ETHICAL OBLIGATION
_________________
advising your colleague to terminate their physician-patient relationship is NOT ENOUGH
Relative Risk measures ⬜
_________________
What’s the Formula?
ratio between (probability of event<sup>ET</sup> : probability of event<sup>cp</sup>)]
RR = [IET / Icp]
RR = [Incidence of disease in (Exposed/Treated) group] / [Incidence of disease in (control/placebo)* *group]
T or F
- Patient p/w financial hardship at end of her visit.*
- You tell her that Co-pays can be 100% completely waived or partially adjusted based on standardized proof of need*
[T or F] ?
Explain
TRUE
⼀(True if there is standardized proof of need - but note: waiving copays for ALL patients = insurance fraud due to total charge discrepancy and potential inducement of unnecessary medical services)
In Research, describe what a [Type 1 error] is (3)
⭐detecting a difference between groups when a difference does NOT in fact exist lol
or (AKA)
⭐: Mistakenly rejecting [a null hypothesis] … that is actually true and should NOT be rejected = false positive(since rejecting = positive)
⭐{probability of [Type 1 error] occurring = α } and this reflects significance of a test
On a Central Tendency Graph below Identify
Mean, Median, Mode
cOIN
central tendency graph =
1: mOde
2: medIan
3. meaN
What case control statistical device can be used to measure association (if one exist) between a Disease and a specific [exposure or risk factor] ?
Odds Ratio
[(aD) / (bc)]
[(DISEASEOOE)/ (controlOOE)]
_________________
• OOE = Odds of Exposure
• first draw contingency table | • [(OR >1) indicates there IS an association between a DISEASE and the [exposure or risk factor] in question
What is the difference between [Case fatality rate] and [Mortality rate]?
For (Event J)…
Case fatality rate = [proportion of people with (Event J) who die from (Event J)]
_________________
Mortality rate = [proportion of general population who die from (Event J)]
What is the [Family Medical Leave Act]?
[solid employees > half time ≥1y] of [Large companies > 50 employees] have legal federally protected [≤12 weeks unpaid leave] if :
- postpartum
- serious illnes
- caring for immediate family with serious illness
What is Correlation Coefficient (r)?
[{r < 0} indicates ⬜]
& [{r>0} indicates ⬜]
describes
[Direction {neg inverse} vs {positive direct}]
and
[STRENGTH (-1 0 1) - furthest from 0 = stronger]
- of two quantitative variables to each other*
- _________________*
- [{r<0} = {neg inverse} = {A⇪ ⼀B⬇︎}]* 🆚 [{r>0} = {pos direct} = {A⇪ ⼀B⇪ }]
Cross sectional studies are majorly used to determine ⬜, but may also be used to ⬜
prevalence of disease;
obtain snapshots = examines associations between risk factors and outcomes in a well-defined population at one particular time
=takes snapshot of a population and measures freq of risk factors and outcomes simultaneously
although this leaves out temporal relationship between the two
In terms of Medical Errors, describe
Sentinel event
Conflict arises between multiple children of a terminally ill patient regarding his medical decisions
Management?
Hospital ETHICS Committee
Predictive values varies based on disease prevalence
How does disease Prevalance affect Positive Predictive Value?
and Negative Predictive Value?
⇪ Prevalence = ⇪ PPV
and
⇪ Prevalence = ⬇︎ NPV
highly SPecific test are useful for test [⬜ (screening | confirmation)]
PCD = [ProoF|ConFirmation|Diagnostic]
_________________
sPeci_F_icity= [true negative] = [TN / (TN + FP)] = used for PCD test
= “a test’s probability (in the absence of disease) a patient test negative”
PCD = [ProoF|ConFirmation|Diagnostic]
AUC of a ROC curve is a reflection of ⬜. Explain
diagnostic accuracy;
Greater AUC = Greater Accuracy and discrimination (outperforms)
What is the Formula for Negative Likelihood Ratio?
Are Positive/ Negative Likelihood Ratios dependent or indepedent of a disease’s prevalence?
NLR⊝ = [(1-N) / P]
_________________
INDEPENDENT
“Number 1 Nigga, Positivity”….
Positive LR = seNsitivity / (1 - sPecificity)
Negative LR = (1 - seNsitivity) / P
This type of study (known as ⬜ ) can only provide what type of information?
[ecologic correlational study] ;
population-level information
Define | Formula?
[(LR+) Positive likelihood ratio]
________________
[(LR-) Negative likelihood ratio]
[Cohort Study] is a study of the (⬜ past | present | future) that looks at [⬜ and ⬜ groups] to determine ⬜
_________________
What question does [Retrospective Cohort Study] ask?
What question does [Prospective Cohort Study] ask?
[past (retrospective)] and [future (prospective] ; ([group WITH Risk Factor] v [group WITHOUT Risk Factor]) ; [INCIDENCE (development)] of a disease
_________________
- retrospective = Using past records, between [GWF] and [GWOF] who developed the disease?*
- Prospective = Between [GWF] and [GWOF] Who WILL develop disease?*
“Smokers (exposed/risk factor group) had higher RISK of developing COPD than Nonsmokers (NOT exposed/NO risk factor group) had”
What kind of screening test are good for identifying individuals with NO disease? (2)
HIGH SPECIFICITY
High Negative Predictive Value
What kind of study should be used to investigate an
acute infectious disease OUTBREAK?
Case-Control Study
________________
allows quick localization of outbreak source
Draw a [Specificity/Sensitivity contingency table]
Define
Internal Validity (3)
- IV = causality (did your [independent variable-EFFECT] Cause the ∆ in the [dependent variable-OUTCOME] you observed?
- If Yes = [⇪ Internal Validity] 👍
- If No ➜ [solution = double blind]
Define
EXTERNAL Validity (2)
- EV = GENERALIZABILITY ( is the observed relationship seen in the study GENERALIZEABLE (meaning can we see the SAME observed relationship OUTSIDE THE STUDY-IN THE REAL WORLD?
- If Yes ➜ ⇪ External Validity)
Observational studies are classified as Descriptive or Analytical
Name the Analytical Observational studies? (3)
_________________
What do they do?
[case control], [COHORT R/P], [some cross-sectional]
_________________
- Analytical* studies TEST HYPOTHESIS about association between risk factors and disease
- (i.e. study determining that smoking cigarettes is a statistically significant risk factor for Lung CA)*
characteristics of
Crossover design (4)
- smaller sample size
- more precise estimates of treatment effect than studies with parallel design - (since crossover = each subject gets ALL TREATMENTS = acts as their own control) (versus parallel = subjects receive single treatment thoughout entire study)
- takes longer to complete (since subjects have to move from 1 phase to another multiple times during the study)
- carryover effect (1st treatment effect carrysover into the 2nd treatment phase - can be avoided with lengthy washout)
How do you deduce predictive value from [(r) correlation coefficient]? (3)
- [(r) correlation coefficient] range from {-1 to +1} with [0 = NO CORRELATION AT ALL].
- The closer to +1 = VERY POSITIVELY CORRELATED
- The closer to -1 = VERY NEGATIVELY CORRELATED
Briefly describe each term:
a. Premium
b. copayment
c. coinsurance
Briefly describe each term:
d. deductible
e. Out-Of-Pocket Maximum
d. [deductible = (Out-Of-Pocket minimum)] = min patient has to pay before health plan starts to cover any part of health expenses
e. [Out-Of-Pocket MAX] = MAX patient has to pay before health plan starts to cover ALL (100%) of health expenses
Both RR and OR measure the association between ⬜
_________________
Name the difference
Relative Risk | Odds ratio
exposure and outcome
RR = observational / experimental / follow up studies / cohort
OR = case-control / cross-sectional studies
Name and describe the 8 components of (writing) a Prescription
DP dRIVRS
lead toxicity levels ⬜ are a/w permanent cognitobehavioral problems and levels of ⬜ require Chelation therapy
10-20 ; ≥45
note: Chelation does not improve cognitive complications
Medical encounters may be the only way trafficked victims can get help
If you suspect human trafficking, recite the 8 questions you can use to screen for it
Interview Patient ALONE
Medical encounters may be the only way trafficked victims can get help
Name the common warning signs for Human Trafficking (5)
Interview Patient ALONE
What is the purpose of Palliative Care? -2
★ [interdisciplinary ⬇︎of unnecessary medical interventions] while [⇪ quality of life for terminal/seriously ill patients (and their family)].
★ Can occur concurrently with life-prolonging tx
What is the Formula for Positive Likelihood Ratio?
Are Positive/ Negative Likelihood Ratios dependent or indepedent of a disease’s prevalence?
P = N / (1-P)
“Number 1 Nigga, Positivity” ….. “Positivity… Number 1 Plan”.
Negative LR = (1 - seNsitivity) / P
Positive LR = seNsitivity / (1 - sPecificity)
INDEPENDENT
In research, how does [Double Blinding] (which means ⬜ ) affect a study’s internal validity?
keeping individuals involved unaware of participants’ treatment assignments
________________
[Double Blinding] ⇪ internal validity
In Research, describe what a [Type 2 error] is (4)
_________________
What is the formula for [Type 2 error]
- failure to detect a difference between groups when a difference does in fact exist
- or (AKA): failure to reject a null hypothesis … that is actually false and should be rejected = false negative (since rejecting = positive)
- *
- *
—{probability of [Type 2 error] occurring = β } and this is inversely related to [power (which is how large a study has to go to detect a difference when a difference does exist)**]
- smaller study = ⬇︎power = [⇪ β = INC chances of Type 2 Error occurring]
- _________________*
[T2E = β = (1 - power)]
What is Recall bias?
inaccurate recall of past exposure by participants ➜ MISCLASSIFICATION OF EXPOSURE (applies mostly to case control studies)
_________________
individuals with dz are more likely to search their memories for (possibly inappropriate) association between an exposure and their dz ⼀compared to individuals w/out dz
In terms of [Correlation Coefficient ( r )], the null is ⬜. Why is this?
_________________
How is this related to Confidence Interval? (3)
[null = r = 0] ⼀since this indicates no linear relationship _________________
- a CI for r is statistically significant if it EXcludes 0
- CI of a r indicates, with a certain confidence level, whether r is significantly different from its [null= r =0]
- so… if study states [linear relationship between 2 variables ( r )] is significantly different from 0 (i.e. statistically significant | p<0.050) ➜ further illustrated by [CI for that ( r ) must EXclude 0]
negative likelihood ratio (LR-) represents ⬜ and
positive likelihood ratio (LR+) represents ⬜
________________
What does this mean?
[NLR-] = value of a negative test result = [(1-N) / P]
[PLR+] = value of a positive test result = [N / (1-P)]
________________
LR varies from 0 to infinity. [⇪LR = ⇪ likelihood disease is actually absent/present]
Sensitivity
Formula(2) & meaning
Explain How:
to determine 95% Confidence Interval
for the difference in percentage of patients who achieved primary efficacy endpoint between Drug X and Drug Y? (4)
Confidence interval contain [sample value (∆ between 2 sample percentages) at the center of the interval] = [sample value +/- margin of error]
1st: determine sample value [(Drug X 74.0%) - (Drug Y 85.4%) = (-11.4%)]
2nd: determine if [p < 0.001] –(if yes)-> significantly different from 0 and excludes 0 = CI% will also not include 0
3rd: Find range that does not include 0 and has [sample value -11.4%] its the center =
(-13.9% to -8.9%)
What kind of study should be used to investigate an
acute infectious disease OUTBREAK?
Case-Control Study
________________
allows quick localization of outbreak source
Explain how Confidence Intervals work? (4)
Confidence interval contain the [sample value (∆ between 2 sample percentages) at center of the interval] = [sample value +/- margin of error]
_________________
when comparing ≥2 groups (i.e. treatment 1 vs treatment 2), if their Confidence Intervals DO NOT overlap = There is a Statistically Significant Difference ✔︎ between those 2 groups
_________________
when comparing 1 group to placebo, if that groups’ Confidence Interval DO NOT contain the null value (0) = Overall Statistically Significant
_________________
If 2 groups’ CI overlap = there may or may not be a statistically significant difference between those 2 groups
How do you calculate ARR?
[ARR= (riskPLACEBO - riskTreatment)]
How do you calculate RRR? (2)
RRR=
[ARR/riskCTRL]
–or–
[(riskCTRL - riskTreatment)/riskCTRL)]
_________________
[ARR: Absolute| RRR: Relative _ Risk Reduction]
How do you calculate NNT? (2)
[ARR= (riskCTRL - riskTreatment)]
[RRR = ARR/riskCTRL ]
- [ARR: Absolute| RRR: Relative _ Risk Reduction]*
- *
NNT = [1 / ARR] = {1 / [(failure ratePLACEBO) - (failure rateTX)]}
A good ⬜ test will have high Specificity
How is Specificity related to PPV?
PCD ; [⇪ sP = ⇪ PPV]
_________________
[⇪ sP = ⇪ PPV = ⬇︎FP]
PCD = [Proof/Confirmatory or Diagnostic]
highly Sensitive test are useful for test [⬜ (screening | confirmation)]
screeNing
_________________
seNsitivity= [true Positive] = [TP / (TP + FN)] = used for test screeNing
= “a test’s probability (in the presence of disease) a patient test positive”
Length-time bias
Since [slow progressive benign cases] outlive [rapid progressive malignant cases], Length-time bias occurs when a test only ends up detecting those [slow progressive benign cases (which naturally have better prognosis)] ➜ false appearance that the test actually INC survival benefits
_________________
survival benefits of a screening test are overstated due to detection of a disproportionately higher # of [slowly progressive benign cases]
A good ⬜ test will have high Specificity
How is Specificity related to False Positives?
PCD ; [⇪ sP = ⬇︎False Positives] ⼀which is why good Confirmatory test have HIGH sPecificity
_________________
[⇪ sP = ⇪ PPV = ⬇︎FP]
PCD = [Proof/Confirmatory or Diagnostic]
Define [p-value] and its relation to Null hypothesis
[p-value] = Chance that study results happened randomly
[(At CI 95%) [p-value] < 0.05] means you can Reject Null hypothesis since it means there’s lil chance the results happened randomly
What does overlapping [Standard Error of Measurement bars] indicate?
Non-Statistically significant difference between the two data sets being compared
What kind of test are best to rule out a diagnosis? (2)
SnNOut
Sensitivity HIGH + Negative test result = rules Out a disease
Define__X__ (2)| What’s the Formula?
[(NPV) Negative Predictive Value]
NPV= [TN / (TN + FN)] = probability (in the presence of a Negative test result) a patient DOES NOT actually have the disease = ability of test to accurately predict Negative test result_
How much can we trust this test?
Asymmetry in a funnel plot suggest ⬜ bias
publication
Standardized Mortality Ratio
Explain the overall takeaway
SMR of 1.75 = “observed # of deaths (in this population of interest) is 75% higher than expected”
How do you describe a survival analysis study, in which although [80% control group died] and [80% of treatment group died] also, investigators still reported treatment was more effective than the control?
investigators must have analyzed [TIME-TO-EVENT/DEATH] data instead and saw that, although both groups had 80% die ultimately, the treatment group must have had longer [TIME-TO-DEATH] = Treatment group lived longer = Treatment overall more effective
[TIME-TO-EVENT/DEATH] often used in survival analysis
Define__X__ | What’s the Formula?
[(PPV) Positive Predictive Value]
PPV = [TP/ (TP+FP)] = probability (in the presence of a Positive test result) the patient actually has the disease = ability of test to accurately predict Positive test results_
________________
PPV = ability of test to accurately predict Positive test results
How much can we trust this test?
Odds Ratio
formula
[(aD) / (bc)]
[(DISEASEOOE)/ (controlOOE)]
_________________
• OOE = Odds of Exposure
• first draw contingency table | • [(OR >1) indicates there IS an association between a DISEASE and the [exposure or risk factor] in question
What type of Nonnormal distribution is this?
_________________
How does it affect Mean, Median and Mode?
POSITIVE skew
( “it’s skewing away from the POSITIVE direction = POSITIVE SKEW)
mea(N) > med(i)an > m(O)de
