Epidemiology and Stats

Question 1

Define the following
1. Fetal death (still birth)
2. Perinatal death
3. Early neonatal death
4. Late neonatal death
5. Post-neonatal death
6. Infant death

Answer 1

1. Fetal death - death from 20/40 till birth or >400g if gest unknown
2. Death from 20/40 until 7 days old
3. Death from 0-7 days onld
4. Death from 7-27 days old
5. Death from 28days - 1 year
6. Death from birth until 1 yr

Question 2

Define the following
1. Preterm birth
2. Extreme preterm
3. Very preterm
4. Moderate to late preterm
5. Perinatal mortality rate
6. Low birth weight

Answer 2

1. Preterm - Born before 37/40
2. Extreme preterm - Born before 28/40 - 2%
3. Very preterm - Born 28-32 weeks - 2%
4. Moderate to late preterm - Born 32-37 weeks - 7%
5. Perinatal mortality rate - number of fetal and early neonatal deaths from 20/40 until 7 days per 1000 live births
6. Any baby weighing less that 2499g regardless of gestation

Question 3

Define the following
1. Maternal morbidity
2. Maternal mortality
3. Maternal mortality ratio
4. What are the main causes of MMR in NZ (4)

Answer 3

1. Maternal morbidity - Any health condition that is attributed to or aggrevated by pregnancy
2. Death of a woman while pregnant to 42 days PN from any cause related to or aggrevated by pregnancy. Doesn’t include accidents or incidental causes.
3. The number of maternal deaths per 100 000 live births over a specific time period.
4. Main causes of maternal mortality in NZ
   -Suicide - 50% Maori women 26% (direct cause)
   -AFE 12% (direct cause)
   -Neurological 10% (indirect cause)
   -Cardiac - 10% (indirect cause)

Question 4

Discuss sensitivity
-Calculation
-Interpretation

Answer 4

1. Sensitivity calculation
   -True positive / true positive + false negative A/A+C
   -The ability of a test to detect disease, the chance of a test being positive if you have the disease
   -High sensitivity means that there are few false negatives
   -Good for ruling things out. If test negative then unlikely to have disease if test has high sensitivity

Question 5

Discuss specificity
-Calculation
-Interpretation

Answer 5

-Specificity = true negative / true negative + false positive (D/B+D)
-The ability of a test to identify those without disease, the chance of a test being negative if you don’t have a disease
-High specificity means low false negatives
-Good for ruling things in. If test very specific and is positive then likely to have disease

Question 6

Discuss predicitive values
-Definition of PPV and NPV
-Calculations
-Conditions

Answer 6

1. Defintiion
   -Ability of a test result to predict the prescence of abscence of disease
   -PPV - likelihood that someone has a disease if the test is positive
   -NPV-likelihood that someone doesn’t have the disease if the test is negative
2. Calculations
   -PPV = A/A+B
   -NPV = D/C+D
3. Conditions
   -Highly dependent on prevalence in a population

Question 7

Discuss likelihood ratios
-Defintion
-Calculations
-Conditions

Answer 7

1. Defintion
   -A measure of the odds of having a disease relative to the prior probability of the disease
2. Calculations
   Positive likelihood ratio - True positive rate / false positive rate (sensitivity / 1-sensitivity)
   Negative likelihood ratio - False negative rate / true negative rate (1- specificity / specificity)
   Pre-test odds x LR = post test odds
3. Conditions
   -independent of disease prevalance

Question 8

Define the following:
1. null hypothesis
2. type 1 error
3. type two error
4. power

Answer 8

1. Null hypothesis = there is no association between two vaiabiles
   -Research can either accept or reject the null hypothesis
2. Type 1 error
   -Incorrectly reject the null hypothesis i.e. find an association when there isn’t one
   - alpha
   -Same as p-value 0.05
3. Type 2 error
   -Incorrectly accept the null hypothesis. i.e. don’t find an assoication when there is one
   -Beta
   -Reflects an under powered study
4. Power
   -Power = 1 - beta
   -Reflects the ability of a study to find the true assoication

Question 9

Discuss levels of evidence
-Numerical level system
-Alphabetical level system

Answer 9

1. Numerical level system
   -Level I - Evidence obtained from at least one well designed RCT
   -Level II-1 - Evidence obtained from well designed controlled trials but no randomisation
   -Level II-2 - Evidence has been obtained from cohort or case control studies.
   -Level II -3. Evidence has come from other types of non analytical studies
   -Level III - expert opinion
2. Alphabetical level system
   -Level A - Evidence is based on consistent findings from RCT, cohort studies and has validity in different pops
   -Level B - Evidence from retrospective cohorts or case control studies
   -Level C - Evidence from case series
   -Level D - Evidence from expert opion