Discussion 4

Name: ________________________________________________[1]
Case Study 1: Measuring ADHD in School-aged Children
Homework 1
Complete the following questions, which are a continuation of the in-class case study. Information from
the in-class case study will inform answers in the homework. Please type all answers (including calculations)
and show all calculations to receive full credit.
NEXT STEP 6
The five years of data collection is now complete and you prepare a new data table.
ADHD Cases – Raw Data
Year
Newly diagnosed
with ADHD
Individuals still at risk after
the school year ends (June
30)
0 (2009-2010
school year)
–
3535
—-
1 (2010-2011
school year)
55
3480
?
2
50
?
?
3
55
?
?
4
50
?
?
5
55
?
?
Annual Incidence
(per 1,000)
UGPH-GU 30 Epidemiology for Global Health | Fall 2021 | Case Study: Disease Frequency
Name: ________________________________________________
6.
[2]
Calculate the number of individuals still at risk after each school year ends and the annual incidence for
each study year using the information in the table above. Use two decimal points for your incidence
calculation. Fill in the table above where there are question marks, but be sure to show your work
below. (Hint: Think carefully about the denominator and the timing of being at-risk) (3 points)
NEXT STEP 7
Your supervisor recognizes your astute work in correctly calculating annual incidence and has now asked
you to calculate a 5-year period prevalence.
7.
What information do you need to calculate a 5-year period prevalence for 2010-2015? (1 point)
UGPH-GU 30 Epidemiology for Global Health | Fall 2021 | Case Study: Disease Frequency
Name: ________________________________________________
[3]
8.
Calculate the 5-year period prevalence. Use two decimal points for your prevalence calculation. (2
points)
9.
Which measure, annual incidence or period prevalence, is more useful in understanding the burden of
ADHD in the 12 school community? Why? (1 point)
UGPH-GU 30 Epidemiology for Global Health | Fall 2021 | Case Study: Disease Frequency
Name: ________________________________________________
[4]
NEXT STEP 8
Up to this point, we assumed that students remain in the population for the entire school year. This may not
be the case, as some families may move in and out of the community during the school year. Additionally,
some students may develop ADHD mid-year and, therefore, no longer be at risk of developing ADHD. As a
result, the calculation of an incidence proportion may not be appropriate.
Incidence rates implicitly account for time in the denominator and are often used when populations are
dynamic. The purpose is to get an improved assessment of the time a person is at risk of developing ADHD.
To get a more accurate assessment of the risk of developing ADHD in these schools, you revisit the data
sources to determine how much person-time each student contributed during the first year of data
collection.
Below is a schematic representing the amount of person-time each student contributed during the first
school year of data collection. Each group represents a subset of the population as opposed to listing the
person-time for each student at risk during the school year. For example, group 1 represents 3,200 students
who completed the entire school year and did not develop ADHD. Alternatively, group 3 represents the 200
students who left school in October after one month of follow-up and did not develop ADHD. We don’t know
what happened to these students, but they do contribute one month of follow-up time because they were at
risk of developing ADHD during that month. A complete school year is 10 months, which is the maximum
number of months a student can be in school and at risk for developing ADHD.
Diagnosed with ADHD
UGPH-GU 30 Epidemiology for Global Health | Fall 2021 | Case Study: Disease Frequency
Name: ________________________________________________
10.
[5]
Calculate the incidence rate/density of developing ADHD in the 2010-2011 school year using
information from the schematic above. Please report the rate per 1000 student-years (10 months = 1
student year), using two decimal places. Show all of your work. (3 points)
UGPH-GU 30 Epidemiology for Global Health | Fall 2021 | Case Study: Disease Frequency
Cause-specific mortality “rate”
https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812826
Cause-specific mortality in 2020 (including COVID-19)
Murphy SL, Kochanek
KD, Xu JQ, Arias E.
Mortality in the United
States, 2020. NCHS Data
Brief, no 427. Hyattsville,
MD: National Center for
Health Statistics. 2021.
Case fatality “rate”
● How many of those infected or inflicted with
disease will die (percent)?
Population at risk of
dying from the disease
Gordis 2004
Table 13. Number of deaths and case fatality rate (CFR) in 2015*, by pathogen,
Foodborne Diseases Active Surveillance Network (FoodNet), United States
Case fatality
“rates” for
foodborne
diseases, U.S.
Total 2015
Pathogen
# of deaths
Campylobacter
Listeria†
Salmonella
Shigella
STEC§ O157
STEC non-O157
Vibrio
Yersinia
Cryptosporidium
Cyclospora
Total
*Data are preliminary
Total # of cases
11
6,309
15
116
32
7,728
1
2,688
3
463
1
796
5
192
1
139
8
1,612
0
64
77
20,107
CFR
0.17
12.93
0.41
0.04
0.65
0.13
2.60
0.72
0.50
0.00
0.38
†Listeria cases defined as isolation of L. monocytogenes from a normally sterile site
https://www.cdc.gov/foodnet/PDFs/Table13.pdf
or, in the setting of miscarriage or stillbirth, isolation of L. monocytogenes from
placental or fetal tissue
§Shiga toxin-producing Escherichia
coli
Proportionate mortality
● What proportion of the deaths in a population
during a specific time were caused by a
specific disease?
Friis & Sellers 2011
Crude mortality “rate”
● Mortality rate from all causes is known as the crude
or all-cause mortality rate:
multiplier
x 100,000
Friis & Sellers 2011
Caution when using crude “rates”!
● Use with caution when comparing disease
frequencies between populations or over time
● Observed differences may be due to systematic
factors (e.g., sex or age distributions) within the
population rather than true variation in “rates”
● Address by using specific or adjusted “rates”
Friis & Sellers 2011
Example: Age distributions
Source: U.S. Census Bureau International Database
(not really a rate)
Friis & Sellers 2011
Adjusted “rates”
● Summary measures of the “rate” of morbidity and
mortality in a population in which statistical
procedures have been applied to remove the
effect of differences in composition of various
populations
● Makes use of data from a standard population
○
To compare across populations, need to adjust using
same standard population
● Example: age-adjusted mortality “rates”
Friis & Sellers 2011
At-home exercise
● State Cancer Profiles:
https://statecancerprofiles.cancer.gov/index.ht
ml
● Identify the most recent cancer mortality “rate”
for your assigned state available
● How is this rate defined?
● What is the rate and what is the year?
General Population Health
Indicators
Birth Rate
Fertility Rate
Number of live
births declined
from ‘07-’10
The general fertility
rate trend mirrors
that for # of live
births; in 2020 the
GFR = 55.8 births
per 1,000 women
age 15-44 years
Martin JA, Hamilton BE,
Osterman MJK, Driscoll AK.
Births: Final data for 2018.
National Vital Statistics Reports;
vol 68, no 13. Hyattsville, MD:
National Center for Health
Statistics. 2019.
Martin JA, Hamilton BE,
Osterman MJK, Driscoll AK.
Births: Final data for 2018.
National Vital Statistics Reports;
vol 68, no 13. Hyattsville, MD:
National Center for Health
Statistics. 2019.
Martin JA, Hamilton BE,
Osterman MJK, Driscoll AK.
Births: Final data for 2018.
National Vital Statistics Reports;
vol 68, no 13. Hyattsville, MD:
National Center for Health
Statistics. 2019.
Martin JA, Hamilton BE,
Osterman MJK. Births in the
United States, 2014. NCHS data
brief, no 216. Hyattsville, MD:
National Center for Health
Statistics. 2015.
Global infant mortality “rate” comparison, 2019
International infant mortality “rate” comparison
OECD (2022),
Infant mortality
rates (indicator).
doi:
10.1787/83dea506
-en (Accessed on
27 January 2022)
Considerations before we begin – I
● What is a “case”?
● Someone with the outcome of interest
● Operational definition to validly (accurately) and reliably (consistently) identify
who has the health problem
● Based on self-reported symptoms or behaviors, medical examination/clinical
diagnoses (reported various ways)
● What is the scope?
● Geographic area, population, time period
Considerations before we begin – II
Ratios and Proportions
● Ratios – comparison of any two values
○ Numerator and denominator can be unrelated
○ ie, sex ratios
● Proportion
○ Numerator is in denominator
○ ie, fraction of cases in a particular clinical setting
How can we tell how common a health state is
in a population?
Count
● Simplest and most frequently performed quantitative measure in
epidemiology
● Number of cases of a disease or other health phenomenon
● Examples?
● What is a limitation of using counts?
Friis & Sellers 2011
Prevalence
● Number of existing cases of disease in a
population at or during some designated
time
● Expressed as percentage or number of
cases per unit size of population
● Indication of extent of health problem
(burden)
● Point vs. period prevalence
Friis & Sellers 2011
Point Prevalence
Point Prevalence =
Number of cases
Total number in the group
at point in time
Friis & Sellers 2011
Point Prevalence: Example
Example (Iowa Women’s Health Study):
Total number of smokers in the group =
Total number in the group
6,234
41,837
= 149.0 per 1,000
or 14.9%
Friis & Sellers 2011
Point Prevalence: Example II
Period Prevalence
# CASES AT BEGINNING OF
TIME INTERVAL + # NEW
CASES THAT OCCUR
DURING THE INTERVAL
(INCIDENCE AND PREVALENCE)
Period Prevalence =
Number
cases
during
a timeofperiod
Average population
OR CAN USE MIDPOINT
POPULATION
Friis & Sellers 2011
Period Prevalence: Example
Example: Prevalence of HIV in 2019
Persons ever diagnosed with HIV =
Average population
2,293
41,837
= 5.5%
Friis & Sellers 2011
Period Prevalence: Example II
Prevalence of HIV in the US, 1977-2006
Cumulative Lifetime Prevalence
Prevalence of cancer (lifetime)
Persons ever diagnosed with cancer
=
Average population
2,293
41,837
= 5.5%
What does Incidence measure?
• Incidence: measures the occurrence of new cases over a
specified time period
• Cumulative incidence (AKA incidence proportion)
• Probability of developing a disease/health condition during a specified
period of time
• Incidence density or rate
• Measures the number of newly identified
condition per unit of time
cases of a disease or health
(Cumulative) incidence, or risk
● Estimates the risk of the health-related event
● Number of new cases during a specified time interval
● Calculated among all individuals in a population thought to be at risk,
not among everyone in the population
Friis & Sellers 2011
Incidence calculation
Three key pieces of information:
1. Number of new cases (numerator)
2. Size of population at risk
(denominator)
3. Time period of interest
Number of new cases
over a time period
Incidence =
Total population at risk
during the same time period
X 100
USE POPULATION AT RISK AT
BEGINNING OF THE STUDY
PERIOD
Friis & Sellers 2011
Incidence calculation
Number of new cases
over a time period
Three key pieces of information:
1. Number of new cases
Incidence =
Total population at risk
during the same time period
X 100
(numerator)
2. Size of population at risk
(denominator)
Example: # new cases of HIV over 1 year of follow-up = 1,085
Population at risk = 37,105
3. Time period of interest
HIV incidence = 1,085 = 0.0292 x 100 = 2.92%
37,105
Friis & Sellers 2011
Example: high blood pressure among NYC
tax drivers
• Cross-sectional study of NYC taxi drivers
• Contact taxi owners
• Recruit all drivers in NYC, N=20,000
• Self-report high blood pressure diagnosis & conduct high blood
pressure screening
Example: high blood pressure among NYC
tax drivers
Prevalence
Cumulative Incidence
7,600
20,000
=
38% or
1,612
380 per
12,400
=
0.13 or
130 per
1,000 taxi
1,000 taxi
drivers
drivers over
one year
Important Considerations in Calculating
Incidence
• The people at “risk” are potential new cases
• What about people with a history of the disease?
• Are they included in the denominator?
• It depends on the disease, if people can get the disease multiple times they may not
considered a new case each time.
• Also depends on the time period- is it a new case during that year?
• Remember: incidence is the rate of new cases or events over a specified period
for the population at risk for the event.
• If it’s a three week illness- then who got it for the first time over the course of a year
Who is at risk?
● In contrast to prevalence, the numerator for incidence is the
number of new cases of disease that develop during a period
of observation, i.e., incidence focuses on the transition from
non-diseased to diseased among those who are “at risk” of
developing the disease.
● Therefore, for incidence calculations the denominator only
includes people in the source population who were at risk of
developing the outcome of interest at the beginning of the
observation period.
Who is at risk? Examples
Outcome of Interest
Who to Exclude
1st heart attack in adults
Those who already had a heart attack
Any heart attack in adults
None (You need to consider the
outcome of interest to decide who to
exclude.)
Polio in Pakistan
Those who have received polio vaccine
or who have had polio
Cancer of the uterus
Exclude men and any women who have
had a hysterectomy.or who have
already had uterine cancer
Assumptions for incidence
● Accurate when:
● Case distribution across study period is relatively even such that the probability
of disease occurring is equally likely throughout the follow up. E.G. a constant
rate.
● An average rate of incidence is not meaningful if cases occur only certain times
during follow up.
● All individuals are followed for the entire study period
● Assumes ascertainment of disease status for everyone in the
denominator.
What does at risk mean?
● The denominator of an incidence proportion is the number of
persons at the start of the observation period. The
denominator should be limited to the “population at risk” for
developing disease, i.e., persons who have the potential to get
the disease and be included in the numerator.
●
For example, if the numerator represents new cases of cancer
of the ovaries, the denominator should be restricted to women,
because men do not have ovaries. This is easily accomplished
because census data by sex are readily available.
Incidence density (rate)
● Incidence measure used when members of population are under
observation for different lengths of time
● Due to attrition (loss to follow up), death, development of study disease
● Variation in incidence rate calculated using person-time of observation
as denominator
● Person-time = sum of periods of time at risk for each member of
population (e.g., person-years)
Friis & Sellers 2011
Formula for incidence density (rate)
Incidence densityPT = Number of new cases during the time period
Total person-time of observation
If period of observation is measured in years,
formula becomes:
Incidence densityPY = Number of new cases during the time period
Total person-years of observation
Friis & Sellers 2011
Calculating person-time, example
300
Friis & Sellers 2011
Prevalence and incidence are related
• If duration of disease is
short and incidence is
high, prevalence is similar
to incidence
• Short duration: cases recover rapidly/are fatal (eliminate build-up of prevalent
cases), e.g. common cold
• If duration of disease is long and incidence is low, prevalence increases greatly
relative to incidence, e.g., HIV/AIDS prevalence
Friis & Sellers 2011
Resources for review
•
https://sphweb.bumc.bu.edu/otlt/MPH-Modules/PH717-QuantCore/PH717Module3-Frequency-Association/PH717-Module3-FrequencyAssociation4.html