Lecture 8 Mortality & Morbidity Statistics Flashcards
What are crude death rates? And why are they not helpful?
Crude death rates are raw data of deaths
Firstly does not take into account population size (annual death rate does).
And is not adjusted for population variation according to age, gender, race (e.g. age-adjusted rates do)
How do we calculate annual death rate?
ADR = (#deaths / total population) * 1000
*1000 to get number of deaths per 1000 population, whole number easy to use
What are specific death rates?
Death rates that are specific to a demographic of the population e.g. age-specific, or specific to a cause of death e.g. cancer deaths
Define confounding factors.
Confounding factors generally distort a measure of the effect of an exposure on an outcome because other factors that are associated to the exposure are influencing the outcome
How are adjusted (standardised) rates calculated?
Use a fictitious population matrix
Using age-specific rates (e.g. 10 deaths in ages 20-30) for comparison population and multiply with age-distribution in fictitious ‘standard’ population matrix (e.g. age 20-30 stand for 20% of population).
Explain the principle of age-adjusted rates.
Rates adjusted to what they would be if the 2 populations being compared had the same age distributions as some arbitrarily selected standard populations, so they can be compared.
What are morbidity statistics used for?
Insurance companies
Governments (planning)
What does direct standardisation use?
Age-specific rates in comparison populations.
Age-distribution in ‘standard’ population
What does indirect standardisation use?
Age-distribution in comparison population.
Age-specific rates in ‘standard’ population
How do screening, hospital records, and long-term registration differ in the volume and quality of their data collection?
Screening: high volume, low quality
Hospital records: medium volume and quality
Long-term registration: low volume, high quality
Explain screening in disease control.
Screening is when you test non-diseased people to detect early and prevent progression.
Only presumptive diagnosis, so require confirmatory tests if positive.
Explain hospital records in disease control.
Full examination. Confirmed diagnoses.
However often don’t have all data for a patient in the same medical record
Explain permanent long-term registration e.g. National Cancer Registry
Best quality data. Costly
How is the sensitivity of the test calculated?
How well you detect disease
True positives / all with disease
How is the specificity of the test calculated?
How well you detect negative people
True negatives / all without disease
