Stats in surgery Flashcards
Assuming that the results are normally distributed, what percentage of values lie within two standard deviations of the mean ?
95%
Properties of normal districbution
symmetrical i.e. Mean = mode = median
68.3% of values lie within 1 SD of the mean
95.4% of values lie within 2 SD of the mean
99.7% of values lie within 3 SD of the mean
Types of audit
Financial - historical- fairness, accuracy, and reliability of financial data
Operational - future- primary sources of evidence are the operational policies and achievements
Departmental review - A current period analysis of administrative functions, to evaluate the adequacy of controls, safeguarding of assets, efficient use of resources, compliance with related laws, regulations and institutional policy and integrity of financial information.
Standards- Comparison of care or passage of care against set and widely agreed standards or outcomes.
Systems - Evaluation of processes occurring within an institution- If the problem is related to a breakdown in the system or technology infrastructure
Which device would be most suited for resection of a parietal lobe tumour?
CUSA device
How Is data from multiple RCT best displayed
Forest plot
Who to report to if patient in a study experiences an adverse outcom
Ethics committee
Calculating RR
Exp event rate
Control event rate
RR=EER/CER
When to use bipolar over monopolar
Bipolar is most commonly used in operations of the digits/penis (to avoid monopolar current focused over a smaller region), in patients with pacemakers (to avoid electrical involvement with the pacemaker), or in microsurgery
What does a p value represent
Risk of type 1 error
Type 1 vs 2 error
Type 1 Error
the null hypothesis is rejected when it is true - i.e. Showing a difference between two groups when it doesn’t exist, a false positive
Rate of type 1 error is the given the value of α
It usually equals the significance level of a test
Type 2 Error
the null hypothesis is accepted when it is false - i.e. Failing to spot a difference when one really exists, a false negative.
Rate of type 2 errors is given the value of β
It is related to the power of the test
Diathermy used in thyroid gland surgery
Bipolar to minimise risk to RLN
What does a null hypothesis state
A null hypothesis (H0) states that two treatments are equally effective (and is hence negatively phrased). A significance test uses the sample data to assess how likely the null hypothesis is to be correct.
Alternative hypothesis states
The alternative hypothesis (H1) is the opposite of the null hypothesis, i.e. There is a difference between the two treatments
Incidence vs prevalence
The incidence is the number of new cases per population in a given time period.
The prevalence is the total number of cases per population at a particular point in time.
Number needed to treat calculation
1/ARR
ARR= RT- RT2
Case control vs cohort
Cohort - measure RR
Observational and prospective
Case control- Observational and retrospective
OR
Useful for studying rare conditions
Prone to confounding
If a dataset requires multiple testing
Use Bonferroni correction
Tests for not normal distributions
Chi Squared and Mann Whitney U tests
Sample size small- Fischer
What has the biggest impact on PPV
Prevelance
Definition of PPV< NPV, Sens, spec
The positive predictive value (PPV) is the probability that an individual with a positive screening result has the disease. The sensitivity is the probability that an individual with the disease is screened positive and the specificity is the probability that an individual without the disease is screened negative.
Its value depends upon the prevalence of the condition being tested for and the sensitivity of the test used.
It may be calculated by dividing the number of true positives by the number of true positives and the number of false positives. §]
Levels of evidence
I - evidence from meta-analysis of randomised controlled trials
II - evidence from at least one well designed controlled trial
III - evidence from correlation and comparative studies or use of historical controls
IV - evidence from case series
V - Expert opinion or founded on basic principles
Components for power calculation
Size of effect
Significance level
Sample size used to detect the effect
Desired power value
Mean in relation to median with skew
In a left-skewed distribution, the mean is less than the median
Child is unstable requires blood, parents refuse what do you do
Treat in emergencies
Pre- and post- test odds and probability
Pre test prob= prevelencne
Pre test odds= PreTP/1-PreTP
Post test odds= PreTO x likelihood
Post test probs= PostTO/1-PostTO
Montgomery ruling
The doctor is therefore under a duty to take reasonable care to ensure that the patient is aware of any material risks involved in any recommended treatment, and of any reasonable alternative or variant treatments.
Cluster trials
Groups are randomised rather than individuals
Require increased recruitment to achieve same stat power as individual
more prone to unit of analyses errors than individual based trial
Analysis is more complex
Intention to treat analysis
An intention to treat analysis considers data according to the treatment to which the patient was randomised, rather than the treatment which was recieved.
Bolam test
decision made by a doctor is in agreement with the professional standard of medical practise.
Consent forms
1- competent adults, consciousness may be impaired
2- adult on behalf of child
3- not impaired
4- adult lacking capacity
When is a MCA advocate needed
Patient requires intervention
Time pressure
Lacks capacity for themselves
a best interests decision is required because the doctor has assessed the person as not having the capacity to make the decision themselves
the person does not have family or friends with whom it is appropriate to consult about the decision.
What does r represent
Correlation coefficient
Levels of evidence
1a- systematic review of RCTs
2- cohort
3- case- control
4- case series
5- expert opinion
a- systematic review of
b- individual
p value of 0.05
<5% due to chance
Friedman test use
3 or more
Non parametrical- not normal distribution
95% CI meaning
That 95% confidence that the true population value lie within this range
Chi squared test vs t test
T test for continuous data
Chi squared for categorical - they either did or they didnt