Week 3 - Health Measures
● A large part of epidemiology involves the use of measures to answer questions about
distribution and determinants of disease → e.g. prevalence, incidence
● First step to understand these = understand proportions + rates
Proportion
● Fraction with numerator included in the denominator
● Often expressed as a percentage (*100)
● Dimensionless (no unit of measurement)
● Values range from 0 to 1 (0 to 100%), e.g. A/(A+B)
Prevalence
● # people with a disease / # people in population at risk (diseased + non-diseased)
○ Can only count ‘disease’ that is known
○ Why is prevalence important?
Rate
● Instantaneous potential for change in one quantity per unit change in another quantity
(often time), e.g. 20 deaths/year, 60 mph
● Not dimensionless (“per unit something”), no upper bound (can approach infinity with
catastrophic event)
Incidence
● # new cases of disease in population in period of time / # people at risk in same
population and time period
How many people in Canada got the flu in 2004?
● Canadian population ~ 30,000,000
● 0.4 incident cases of fly per 1,000 in 2004
● 0.4 per 1000 = 0.4 x 30,000, = 12,000 people
Attack rate
● 400 residents in a long term care facility
● During an outbreak 150 get flu
● What is the attack rate for the residents in the facility?
● (150/400) x 100 = 37.5%
Prevalence and incidence
● P = I x D, which works when:
○ The incidence rate is in a steady state: no epidemic exists
○ The mean duration of disease (time from diagnosis to recovery) is stable
Example: Australian Cancer Registry data (2008)
, ● Incidence of cervical cancer was 4.0 per 100,000 persons
● The duration of cervical cancer was 36 months
What was the prevalence in 2008?
1. Use decimals in formula: 4.0 per 100,000 = 0.00004
2. Convert duration to years since question asks about year 2008
● 36 months = 3 yrs
3. Plug into formula (Prevalence = Incidence x Duration)
● Prevalence = 0.00004 x 3 = 0.00012
● Multiply by 100 to get 0.012%
● Multiply by 100,000 to get 12 per 100,000
Surveillance
● Important part of public health epidemiology
● Definition → ongoing, systematic collection, analysis + interpretation of health data
○ Timely dissemination to personas who need to know
● Example → National Diabetes Surveillance System
○ Uses healthcare administrative data (physician billings, hospital discharge
information) to facilitate the long-term monitoring of type ½ diabetes
○ Within these data, epidemiologists may compute incidence, prevalence, mortality,
etc.
What can we do with this data?
● Determine whether the disease is spreading in some populations (‘high risk’ and not
others
● Assess the seriousness of disease outcomes in all populations
● Develop prevention programs targeted at specific populations
● Compare to other diseases to determine resource allocation
Mortality - Influenza
● Influenza mortality rate in Canada in 2004 was 0.7 per 100,000 people per year (disease
specific mortality)
● Total mortality rate in Canada in 2004 was 572 per 100,000 people per year
What was the proportionate mortality from influenza?
● Influenza mortality rate = 0.7 per 100,000 people per year
● Total mortality rate = 572 per 100,000 people per year
● Proportionate mortality from influenza = 0.7/ 572 x 100 = 0.12%
● 0.12% of deaths in 2004 were due to influenza
Proportionate Mortality (PM)
● Proportionate mortality from avian-borne illness = 0.001%
● A large part of epidemiology involves the use of measures to answer questions about
distribution and determinants of disease → e.g. prevalence, incidence
● First step to understand these = understand proportions + rates
Proportion
● Fraction with numerator included in the denominator
● Often expressed as a percentage (*100)
● Dimensionless (no unit of measurement)
● Values range from 0 to 1 (0 to 100%), e.g. A/(A+B)
Prevalence
● # people with a disease / # people in population at risk (diseased + non-diseased)
○ Can only count ‘disease’ that is known
○ Why is prevalence important?
Rate
● Instantaneous potential for change in one quantity per unit change in another quantity
(often time), e.g. 20 deaths/year, 60 mph
● Not dimensionless (“per unit something”), no upper bound (can approach infinity with
catastrophic event)
Incidence
● # new cases of disease in population in period of time / # people at risk in same
population and time period
How many people in Canada got the flu in 2004?
● Canadian population ~ 30,000,000
● 0.4 incident cases of fly per 1,000 in 2004
● 0.4 per 1000 = 0.4 x 30,000, = 12,000 people
Attack rate
● 400 residents in a long term care facility
● During an outbreak 150 get flu
● What is the attack rate for the residents in the facility?
● (150/400) x 100 = 37.5%
Prevalence and incidence
● P = I x D, which works when:
○ The incidence rate is in a steady state: no epidemic exists
○ The mean duration of disease (time from diagnosis to recovery) is stable
Example: Australian Cancer Registry data (2008)
, ● Incidence of cervical cancer was 4.0 per 100,000 persons
● The duration of cervical cancer was 36 months
What was the prevalence in 2008?
1. Use decimals in formula: 4.0 per 100,000 = 0.00004
2. Convert duration to years since question asks about year 2008
● 36 months = 3 yrs
3. Plug into formula (Prevalence = Incidence x Duration)
● Prevalence = 0.00004 x 3 = 0.00012
● Multiply by 100 to get 0.012%
● Multiply by 100,000 to get 12 per 100,000
Surveillance
● Important part of public health epidemiology
● Definition → ongoing, systematic collection, analysis + interpretation of health data
○ Timely dissemination to personas who need to know
● Example → National Diabetes Surveillance System
○ Uses healthcare administrative data (physician billings, hospital discharge
information) to facilitate the long-term monitoring of type ½ diabetes
○ Within these data, epidemiologists may compute incidence, prevalence, mortality,
etc.
What can we do with this data?
● Determine whether the disease is spreading in some populations (‘high risk’ and not
others
● Assess the seriousness of disease outcomes in all populations
● Develop prevention programs targeted at specific populations
● Compare to other diseases to determine resource allocation
Mortality - Influenza
● Influenza mortality rate in Canada in 2004 was 0.7 per 100,000 people per year (disease
specific mortality)
● Total mortality rate in Canada in 2004 was 572 per 100,000 people per year
What was the proportionate mortality from influenza?
● Influenza mortality rate = 0.7 per 100,000 people per year
● Total mortality rate = 572 per 100,000 people per year
● Proportionate mortality from influenza = 0.7/ 572 x 100 = 0.12%
● 0.12% of deaths in 2004 were due to influenza
Proportionate Mortality (PM)
● Proportionate mortality from avian-borne illness = 0.001%
