ratios and standardization
DESCRIPTION
RATIOS AND STANDARDIZATION. RATIOS. Ratio - the value obtained when one quantity is divided by another: Proportion – a ratio where the numerator is part of the denominator # diseased/(# diseased + # not diseased) Percentage – proportion * 100 - PowerPoint PPT PresentationTRANSCRIPT
RATIOS AND STANDARDIZATION
RATIOSRatio - the value obtained when one quantity is divided by another:
Proportion – a ratio where the numerator is part of the denominator# diseased/(# diseased + # not diseased)
Percentage – proportion * 100[# diseased/(# diseased + # not diseased)]*100
Example: Say we have 15 diseased people and 30 not diseased, then:
proportion diseased: 15/(15+30) = 15/45 = 1/3
percentage diseased: 1/3 * 100 = 33.3%
RATIOS
Another common ratio:Rate - change in one quantity divided by change in another quantity (usually a measure of time)
– Incidence rate: number of new cases in an initially disease free population per person-time
– Mortality rate: number of deaths in a population per person-time
RATIOS
Example: have 5 deaths in a population of 50 person months
Mortality rate: 5/50 = 0.10 or 0.1 death per person month
Which is equivalent to1 deaths per 10 person months
or 10 deaths per 100 person months
RATIOS
Basic measures of disease:Prevalence: a proportion measuring current disease
Point prevalence: (at a point in time)
Period prevalence:
Example: 6 people have the flu in a classroom of 30 in JanuaryPoint prevalence = 6/30 = 0.2020% of the class had the flu in January
RATIOSIncidence: a proportion or rate measuring new disease
Cumulative incidence: Incidence rate:
Example: 10 new infections of herpes among the 80 first year MPH students in 2012 which was a total of 65 person-years
Cumulative incidence: 10/80 = 0.125or 12.5% of the first year MPH students contracted herpes in 2012Incidence rate: 10/65 = 0.154or 154 new cases of herpes per 1,000 person-years
RATIOS
Probability:
Odds:
Example: 5 people out of 20 diedprobability = 5/20 = 0.25odds death: 5:15 or 1:3
EXAMPLE FOR RATIOS
---------------------XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXD-----------------------------------------------------XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------D------------------------------------------------------------------------------------------------------------XXXXXXXD______________________________________________________________________________ JAN FEB MARCH APRIL MAY JUNE JULY AUGUST SEPT
X = HAS DISEASED = DIED--- = NO DISEASEPERSON-MONTHS = NUMBER OF MONTHS EACH PERSON WAS DISEASE FREE IN THE STUDY
Find: proportion with disease?prevalance of disease?cumulative incidence?incidence rate until June
RATIOS OF RATIOS
RR: Relative Risk, aka Risk Ratio
RR =
Example: male child aged 2 -19 only
Obese Not obese TotalBelow poverty 113 423 5363.5 times poverty 128 947 1075
RATIOS OR RATIOS
What is the probability of a male child aged 2 -19 being obese if they reside in a poverty level family?
What if they reside in a family whose income is 3.5 times higher than the poverty level?
What is the Relative Risk?
RATIOS OF RATIOSOR: Odds ratio
OR = Example: Want to compare disease among people with
a copy of a gene variant to those without the gene variant
Disease No Disease Total
One copy gene variant 50 50 100
No copy gene variant 2 98 100
RATIOS OF RATIOS
What are the odds of disease in the presence of the gene variant?
What are the odds of disease in the absence of the gene variant?
What is the Odds Ratio?
WEIGHTED AVERAGES
Subject Grade Credit hoursEnglish C 3Math B+ 3Science B 3History A 3PE D 1
What is the GPA?
WEIGHTED AVERAGES
GPA is a weighted average with the credit hours of each course as the weights
GPA = Σ(numeric score*hours)/ Σ hours
WEIGHTED AVERAGES
Subject Grade HoursWeighted score English C 3 2 x 3 = 6MathB+ 3 3.33 x 3 = 10 Science B 3 3 x 3 = 9History A 3 4 x 3 = 12PE D 1 1 x 1 = 1Total 13 38GPA = 38/13 = 2.9
WEIGHTED AVERAGE
An important and much used weighted average is the technique of STANDARDIZATION
Study population(s) are compared using a ‘standard‘ population which is large and stable in size
STANDARDIZATION OF RATES
Standardization refers to methods of adjustment based on weighted averages
Two ways to control for an important characteristic (age, sex, race, etc)
DirectIndirect
STANDARDIZATION OF RATES
Direct method – uses the structure of a third ‘standard’ population and the rates of the populations to be compared – can compare two different populations
Indirect method – uses the rates from a third ‘standard’ population and the structures of the populations to be compared – cannot compare two different population, can only compare each to a standard population
used to calculate SMRs and SIRs
STANDARDIZATION OF RATES
Example: have two populations:
Group 1 has a crude mortality rate of 0.017 per 1,000 people per year
Group 2 has a crude mortality rate of 0.022 per 1,000 people per year
Group 1 crude mortality rate < Group 2 crude morality rate
STANDARDIZATION OF RATES
Group 1 Group 2
Pop. Deaths Pop. Deaths
<15 370 4 104 1
15-24 200 3 110 2
25-44 320 3 133 1
45-64 150 2 143 3
65+ 20 6 192 8
1060 18 682 15
Crude rate 18/1060 = 0.017 15/682 = 0.022
STANDARDIZATION OF RATESChoose a standard population (usually large and stable):
Age groups
Population Deaths Rates
<15 2,400 20 0.008315-24 1,900 17 0.008925-44 2,100 14 0.006745-64 1,900 21 0.0111
65+ 1,800 35 0.0194
10,100 107
STANDARDIZATION OF RATESDIRECT METHOD
First need to calculate age specific death rates for both groups:rateaj = (# deaths in age group/population in age group)
Group 1 Group 2
Age Pop. DeathsAge
specific Rates
Pop. DeathsAge
specific Rates
<15 370 4 0.0108 104 1 0.0096
15-24 200 3 0.0150 110 2 0.0182
25-44 320 3 0.0094 133 1 0.0075
45-64 150 2 0.0133 143 3 0.0210
65+ 20 6 0.3000 192 8 0.0417
STANDARDIZATION OF RATESDIRECT METHOD
Age-specific mortality rates per 1000 population
Expected
Age Standard Group 1 Group 2 Group 1 Group 2
< 15 2,400 0.0108 0.0096 25.92 23.04
16-24 1,900 0.0150 0.0182 28.50 34.58
25-44 2,100 0.0094 0.0075 19.74 15.75
45-64 1,900 0.0133 0.0210 25.27 39.90
65 + 1,800 0.3000 0.0417 540.00 70.89
10,100 639.43 184.16
Age adjusted rates per 1,000: 63.3 18.6
Group 1 age adjusted rate > Group 2 age adjusted rate
STANDARDIZATION OF RATESINDIRECT METHOD
Use rate from standard population:Group 1 Group 2
Standard rates per age group
Pop.Deaths
Pop.Deaths
O E O E
< 15 0.0083 370 4 3.07 104 1 0.86
16-24 0.0089 200 3 1.78 110 2 0.98
25-44 0.0067 320 3 2.14 133 1 0.89
45-64 0.0111 150 2 1.67 143 3 1.59
65 + 0.0194 20 6 0.39 192 8 3.72
Total 1060 18 9.05 682 15 8.04
STANDARDIZATION OF RATESINDIRECT METHOD
Calculate the SMR:– SMR =
SMRGroup 1 = = 1.99
SMRGroup 2 = = 1.87
Both are higher than the standard population
EXAMPLE STANDARDIZATION Country A Country B Country C
Age groups (in years)
Population Deaths Populations Deaths Population Deaths
0-20 300 6 100 2 800 15
21-50 400 80 150 30 600 56
51-85 300 120 250 100 600 110
Find:Crude death rates for each country
Using C as the standard perform:Direct standardizationIndirect standardization