Transcript PPT(A) - cancer-epidemiology.org

EPI 5240:
Introduction to Epidemiology
Disease Classification; Vital Statistics; Measures
of Mortality/morbidity
September 21, 2009
Dr. N. Birkett,
Department of Epidemiology & Community
Medicine,
University of Ottawa
9/2009
1
Session Overview
• Review basis of disease classification
systems
• Overview main measures of mortality and
morbidity.
9/2009
2
Traditional Epidemiology Questions
Review
• Who gets disease ‘X’?
• Why did someone get disease ‘X’?
• What is going to happen to someone who has disease
‘X’?
• What can we do to prevent someone getting disease ‘X’?
• What can we do to help someone with disease ‘X’?
• Why are more (or fewer) people getting disease ‘X’ now
than before?
• Why do people living in ‘Y’ get more (or less) of disease
‘X’ than people living in ‘Z’?
9/2009
3
Classification (1)
• To answer questions like these, need to be
able to group or classify people with
similar conditions.
• Can be hard
– Focus on similarities
– Focus on differences
– Heterogeneity vs. homogeneity
• E.g. Psychiatry vs. cancer
9/2009
4
Classification (2)
• John Graunt (1662)
– Grouped deaths into common causes.
• E.g. old age, consumption, smallpox, plague,
diseases of teeth, worms
• James Farr (1860’s)
– Developed an early disease classification
system.
9/2009
5
Classification (3)
• What features should be used as basis for
classification?
– Site of disease
– Hot vs. cold
– Yin/yan
– Imbalance of the four humors
– Behavioural vs. psychological constructs vs.
biological neural factors
– And so on.
9/2009
6
Classification (4)
• Main disease classification is the
International Classification of Diseases
and Related Health Problems (ICD).
– Developed from Farr’s work with first version
around 1900.
– Up-dated about every 10 years.
– Current: ICD-10
– ICD-9 is widely used in epidemiology
9/2009
7
Classification (5)
• ICD-10
– 21 major Chapters.
• Based mainly on organ systems with no implied etiological
value
• Largely based on traditional diagnostic groupings but
includes ‘external’ and other causes.
• Heterogeneous
• Manifestational vs. experiential classification
– Each chapter is divided into paragraphs and subsections.
– C34.4 – Lung cancer, lower lobe
– I60.4 – Subarachnoid hemorrhage, basilar artery
– V95.4 – Spacecraft accident injuring occupant
9/2009
8
Classification – ICD10 (6)
I.
Infections and Parasitic Diseases (A-B) *
II. Neoplasms (C-D)
III. Diseases of blood &blood-forming organs (D)
IV. Endocrine, nutritional and metabolic Diseases(E)
V. Mental & Behavioural Disorders (F)
VI. Diseases of the Nervous System (G)
VII. Diseases of the eye and adnexa (H)
VIII. Diseases of the Ear and Mastoid (H)
IX. Diseases of the Circulatory System (I)
X. Diseases of the Respiratory System (J)
XI. Diseases of the Digestive System (K)
9/2009
9
Classification – ICD10 (7)
XII. Diseases of the skin (L)
XIII. Diseases of the musculoskeletal system (M)
XIV. Diseases of the Genitourinary system (N)
XV. Pregnancy, childbirth, etc. (O)
XVI. The Perinatal period (P)
XVII. Congenital conditions, etc. (Q)
XVIII. Symptoms, signs, NOS (R)
XIX. Injury, poisoning (S-T)
XX. External causes (V-Y) *
XXI. Factors influencing health status and contact
with health services (Z)
9/2009
10
Classification (8)
• Many other classification systems exist
• Cancer
– ICDO
– Snomed
• DSM
• Impairments and disabilities
• Conditions in Primary Care
9/2009
11
Vital Statistics
• Information on main life events
– Births
– Deaths
– Marriages
• Usually collected by the local (municipal)
level).
– Reports sent to provincial government
– Federal government (Statistics Canada)
collates information into reports.
9/2009
12
9/2009
13
9/2009
14
9/2009
15
9/2009
16
9/2009
17
Mortality & morbidity measures (1)
• Prevalence
– More correctly: point prevalence
– Similar to results from a political poll.
– The probability that a person has a disease or
condition TODAY. There is no time
dimension.
– ‘The prevalence of hepatitis C in intravenous
drug users in Ottawa is 60%.’. This means
that 60% of intravenous drug users in Ottawa
have Hepatitis C.
9/2009
18
Mortality & morbidity measures (2)
• Incidence (general)
– Measures the development of NEW cases of
a disease or condition.
– Requires a time dimension. Often this is
given implicitly as ‘one year’.
– Two types of incidence are recognized
• Cumulative incidence or incidence proportion
• Incidence rate or incidence density
– Sorting this out is tricky and somewhat
advanced. I’ll give you a simplified approach.
9/2009
19
Mortality & morbidity measures (3)
• Cumulative incidence or Incidence proportion
– Measures the probability of developing a NEW case
of a disease or condition in a set period of time.
# new cases
Inicid prop 
# people at risk of becoming cases
– ‘The cumulative incidence of esophageal cancer is
4/100,000 in one year.’
• The probability that an adult will develop esophageal
cancer in the next year is 4/100,000
9/2009
20
Mortality & morbidity measures (4)
• Incidence rate (the hard one)
– Measures the rate at which people develop a NEW case
of a disease or condition.
# new cases
IR 
person - time at risk
– Person-time = # people X time spent at risk
– Allows for ‘loss to follow-up’.
– This is not a probability. It can take on any value from 0 to
+∞. It has units: time-1 or cases/person-year.
– ‘The Incidence Rate of influenza is 4 cases per 100
person-years’.
9/2009
21
Mortality & morbidity measures (5)
• ‘Rate’ is used very loosely in epidemiology.
You need to know from context what is
meant (e.g. prevalence rate).
• You will see three broad types of ‘rates’.
– Crude
– Specific
– Adjusted or standardized
9/2009
22
Mortality & morbidity measures (6)
• Crude mortality rate (crude death rate)
# of deaths in a year
mid - year population
• Mid-year population is usually used for
denominator
• The simplest rate used.
9/2009
23
Mortality & morbidity measures (7)
• Cause-specific mortality rate
# of deaths due to the specific cause in a year
mid - year population
• Can also be specific to ‘sex’, ‘geography’, ‘age’,
etc.
9/2009
24
Mortality & morbidity measures (8)
• Age-standardized mortality rate
• A ‘fictitious’ rate designed for comparing groups
which differ in their age distribution
– the crude mortality is higher in Canada than in Haiti.
Is the risk of dying really higher in Canada or is this
due to the Haitian population being younger?
• Will be discussed more later in the course. For
now, just learn the name and purpose.
• Can be adjusted for factors other than age.
9/2009
25
Some important rates (1)
Crude Birth Rate
# of life births in a year x1000
mid - year population
General Fertility Rate
# of life births in a year
x1000
# woman aged 15- 49 in population
9/2009
26
Some important rates (2)
Total Fertility Rate
• Average number of children who would be born
alive to a woman during her lifetime IF she
were to pass through all her child bearing
years conforming to the age-specific fertility
rates for a given year.
• Complex! Essentially, it estimates the number
of children per women which would have been
expected if the current fertility patterns had
applied through-out her life.
9/2009
27
Some important rates (3)
Completed Fertility
• Average number of life births per woman who
has reached the end of her child bearing
period.
• Similar to Total Fertility Rate but is based on
the actual fertility rates through-out a woman’s
life-time rather than on assuming the rates are
the same today.
• Consider this graph:
9/2009
28
9/2009
29
Some important rates (4)
• Traditionally, epidemiology has lots of rates
related to pregnancy and child development
Infant Mortality Rate
# of children dying before 1st birthday in year
x1000
# children born in the year
• Excellent indicator of public health services.
– High rates indicate unmet health/environmental conditions
• Nutrition; sanitation; education
• Widely used for international comparisons
– Canada (2001): 4.9/1000
– Sierra Leone: >100/1,000
9/2009
30
Some important rates (5)
Maternal Mortality Rate
# of deaths due to abortion, pregancy, delivery in year
x100,000
# live births in the year
• Can be strongly influenced by illegal abortions.
• Increases with maternal age
• In Canada, any maternal death is most likely
due to medical negligence
– Canada:
0.3/100,000
– Sierra Leone: 450/100,000
9/2009
31
Some important rates (6)
Live birth
• Complete expulsion from the mother of a product of
conception which breathes or otherwise show any sign
of life after expulsion. One breath is enough. But, the
umbilical cord must have been cut.
Fetal Death
• A death of the product of conception prior to the
complete expulsion. There must be no sign of life postexpulsion.
These definitions are controversial and not consistently
used (e.g. early miscarriages; therapeutic abortions).
9/2009
32
Some important rates (7)
Stillbirth
• There are at least three definitions used:
– A fetal death occurring after a gestation of at least
20 weeks.
– A fetal death occurring after a gestation of at least
28 weeks (the WHO definition when I last checked)
– A fetal death occurring after a gestation of at least
20 weeks or with a fetus weighing more than 500
grams
• Variation partly due to improvements in
neonatal care/survival.
9/2009
33
9/2009
34
9/2009
35
Some important rates (8)
Life expectancy
• Not really a rate but it fits in here.
• The number of years which a person can
expect to live
– Usually reported by the media as life expectancy at
birth (about 79 for men and 82 for women)
– Can be used at any age.
• Life expectancy at age 50 is the number of additional years
the person can expect to live given they have survived to
age 50.
9/2009
36
Some important rates (9)
Life expectancy (cont)
• Computation is complex and uses life tables.
We won’t get into it in this course
• Is strongly affected by deaths in early
childhood
– This is the main reason why life expectancy was so
low pre-1900 and is still low in developing countries
– After reaching adulthood, there is less discrepancy
between countries
9/2009
37
UK data
9/2009
38
Some important rates (10)
Life expectancy (cont)
• Why the increase over the 1900’s?
– Marked reduction in early childhood mortality
• Nutrition
• Sanitation
• Immunization
– Decrease in infectious disease mortality
•
•
•
•
•
9/2009
Nutrition
Sanitation
Housing
Immunization
antibiotics
39
Other Outcome Measures (1)
• PYLLs
– Standard mortality measures ignore the age
of death
– Our society values a death of a child ‘more’
than the death of an older person
– PYLL’s weight deaths by the age they occur to
show impact in a different way.
– Generally, PYLLs rate acute illnesses and
injuries, to younger people, as more important
than chronic diseases
9/2009
40
Other Outcome Measures (2)
• PYLLs
– For each death, establish the ‘expected’ age
at death
• Life expectancy
• A fixed age (e.g. 75) [The most common approach]
– Determine how many years of life have been
‘lost’ due to the death
– Sum up over all deaths.
– Died at age 30  75-30 = 45 PYLL’s
9/2009
41
Other Outcome Measures (3)
• QALYs
– ‘Quality Adjusted Life Years.
– The number of years of life added by an
intervention adjusted for their ‘quality’
– ‘Normal health’ = 1.0
– ‘Dead’ = 0.0
– Use patient’s utilities to assign value to other
states
9/2009
42
Summary
• ICD-10 is main classification system used
• Three key measures of mortality/morbidity
– Prevalence
– Incidence proportion or cumulative incidence
– Incidence rate
• Other measures can be used
– PYLLs
– QALYs
• Lots of ‘rates’ related to pregnancy
– IMR
– MMR
9/2009
43