HEAL 307 Calculation Basics

Prevalence
# 𝑜𝑓 𝑝𝑒𝑜𝑝𝑙𝑒 𝑖𝑛 𝑠𝑎𝑚𝑝𝑙𝑒 𝑊𝐼𝑇𝐻 𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐
𝑃𝑟𝑒𝑣𝑎𝑙𝑒𝑛𝑐𝑒 =
𝑡𝑜𝑡𝑎𝑙 # 𝑜𝑓 𝑝𝑒𝑜𝑝𝑙𝑒 𝑖𝑛 𝑠𝑎𝑚𝑝𝑙𝑒

• Proportion of the population who has a disease or characteristic at a particular time

• Assesses the overall burden of disease in a population
• All people counted in the numerator are also included in the denominator
o Numerator is a subset of the denominator
• Typically expressed as a percentage
Cumulative Incidence (AKA Risk or Incidence Proportion)

# 𝑜𝑓 𝑛𝑒𝑤 𝑒𝑣𝑒𝑛𝑡𝑠 𝑜𝑟 𝑐𝑎𝑠𝑒𝑠 𝑜𝑓 𝑑𝑖𝑠𝑒𝑎𝑠𝑒

𝐶𝐼 =
𝑡𝑜𝑡𝑎𝑙 # 𝑖𝑛 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 𝑟𝑖𝑠𝑘 𝑓𝑜𝑟 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑡𝑖𝑚𝑒 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙

• The probability of developing a disease/outcome over a stated period

• Measures the number of new cases per person in the population over a defined period
• Estimates risk
• Must specify a period
• Used when follow-up time is short, and there is relatively little loss to follow-up
Incidence Rate (AKA Person-Time Rate)

# 𝑜𝑓 𝑛𝑒𝑤 𝑐𝑎𝑠𝑒𝑠 𝑜𝑓 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑜𝑟 𝑖𝑛𝑗𝑢𝑟𝑦 𝑑𝑢𝑟𝑖𝑛𝑔 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑒𝑑 𝑝𝑒𝑟𝑖𝑜𝑑

𝐼𝑅 =
𝑡𝑖𝑚𝑒 𝑒𝑎𝑐ℎ 𝑝𝑒𝑟𝑠𝑜𝑛 𝑤𝑎𝑠 𝑜𝑏𝑠𝑒𝑟𝑣𝑒𝑑, 𝑡𝑜𝑡𝑎𝑙𝑒𝑑 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑝𝑒𝑟𝑠𝑜𝑛𝑠

• Measures the number of new cases (incidence) per unit of time (rate)
• Denominator is expressed as person-time
• Each person is observed from starting time until an endpoint is reached
o Onset of disease
o Death
o Migration out of the study (lost to follow-up)
o End of the study
Risk Ratio (AKA Relative Risk)

𝐶𝐼 𝑖𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝
𝑅𝑖𝑠𝑘 𝑅𝑎𝑡𝑖𝑜 =
𝐶𝐼 𝑖𝑛 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝

• Indicates the strength of association between variables

o 1 (or close to 1)
▪ No/little difference in risk
o >1
▪ Increased risk of outcome in the exposed group
o <1
▪ Reduced risk in the exposed group
• Organize information in a 2x2 table to compute
• The unexposed comparison group must be specified when expressing findings
o Example for a risk ratio < 1:
▪ Individuals who have a BMI in the normal range for height have 0.47 times
the risk of osteoporosis compared to those who are underweight
o Example for a risk ratio > 1:
▪ Mechanically ventilated patients had 3.7 times the risk of nosocomial
pneumonia compared to patients who were not mechanically ventilated
• Can also be expressed as percent relative effect (the percent change in the exposed group)
o Assumes the unexposed group has 100% of the risk and expresses the exposed group
relative to that
o Example when the risk ratio is <1: (1 – 0.47) x 100 = 53
Individuals with a BMI in the normal range for height had a 53%
reduction in risk of osteoporosis compared to those with a BMI
categorized as underweight
𝐼 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒 = (𝑅𝑅 − 1) × 100
o Example when the risk ratio is >1: (3.7 – 1) x 100 = 270
Mechanically ventilated patients had a 270% increase in the risk of
nosocomial pneumonia as compared to patients who were not
mechanically ventilated
% 𝑑𝑒𝑐𝑟𝑒𝑎𝑠𝑒 = (1 − 𝑅𝑅) × 100

• Avoid using “more” or “less” when interpreting findings, as this leads to errors
o Correct expressions of findings
▪ “Those who <name the exposure> had <insert risk ratio> times the risk
compared to those who <describe the group without the exposure>.”
▪ “The risk of <name the disease/outcome> among those who <name the
exposure> was <insert risk ratio> times as high as the risk of <name the
disease/outcome> among those who did not <name the exposure/control
group>”
Rate Ratio
 𝐼𝑅 𝑖𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝
𝑅𝑎𝑡𝑒 𝑅𝑎𝑡𝑖𝑜 =
𝐼𝑅 𝑖𝑛 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝

• The ratio of IR in the exposed and unexposed groups

• Using a 2 x 2 table is recommended to organize the information required to calculate the
rate ratio
• Refer to findings as the ratio of rates (it is not as precise to refer to rate ratio as risk)
o “Those who <describe exposure> had <insert rate ratio> times the rate of
<describe disease/outcome> compared to <describe unexposed group>”
Risk Difference (AKA Attributable Risk)

𝑅𝐷 = 𝐶𝐼 𝑖𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠 − 𝐶𝐼 𝑖𝑛 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠

• Focuses on the absolute effect of the risk factor

• Establishes the excess risk of disease in those who have the factor compared with those
who do not
• Focus on the interpretation of excess risk in the exposed group when stating findings
• Always specify the time interval (how long the subjects were observed)
• If your CI is presented as a percentage, convert the percentage to a fraction so that you
can express it as excess risk
o Example: CI with surgery = 60% (60/100)
CI without surgery = 20% (20/100)
RD = 40/100
Subjects who had surgery had 40 additional cases of disease
compared to subjects who did not have surgery
Rate Difference

𝑅𝑎𝑡𝑒 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 = 𝐼𝑅 𝑖𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠 − 𝐼𝑅 𝑖𝑛 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠

• Calculated by subtracting the IR in the unexposed group from the IR in the group with
the exposure
• Assumes that there is a valid association and cause-effect relationship between the
exposure and the outcome
• Expressed as the number of additional cases in the exposed group per person-time
• Can never be expressed as a percentage due to the retention of person-time in the
denominator
Attributable Proportion (AKA Attributable Fraction, Attributable Risk %)

𝑅𝑖𝑠𝑘 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒: 𝐶𝐼𝑒𝑥𝑝𝑜𝑠𝑒𝑑 − 𝐶𝐼𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑

𝐴𝑡𝑡𝑟𝑖𝑏𝑢𝑡𝑎𝑏𝑙𝑒 𝑃𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 =
𝐶𝐼 𝑖𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝

• The proportion of disease in the exposed group that can be attributed to the exposure
• The proportion of disease in the exposed group that could be prevented by eliminating the
risk factor
• Can be expressed as a proportion or percent
o Proportion
▪ Incidence of <describe disease/outcome> in the <describe exposed group>
was 10 per 100; of this, 8 per 100 could be attributed to <describe
exposure>
o Percentage
▪ Multiply attributable proportion by 100
▪ 80% of the <describe disease/outcome> in the group that <describe
exposed group> could be attributed to <describe exposure>
Preventive Proportion (AKA Preventive Fraction)

𝑅𝑖𝑠𝑘 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒: 𝐶𝐼𝑒𝑥𝑝𝑜𝑠𝑒𝑑 − 𝐶𝐼𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑

𝑃𝐹 =
𝐶𝐼 𝑖𝑛 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝

• Used when exposure reduces the risk of disease rather than increases it
• Can be expressed as a percent by multiplying findings by 100
o State findings as the percentage of disease reduction in the exposed group that can
be attributed to the beneficial exposure
Population Attributable Proportion (AKA Population Attributable Fraction)

# 𝑜𝑓 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 & 𝑤𝑖𝑡ℎ 𝑜𝑢𝑡𝑐𝑜𝑚𝑒/𝑑𝑖𝑠𝑒𝑎𝑠𝑒

𝑃𝐴𝑅 = × 𝐴𝑃 𝑖𝑛 𝑡ℎ𝑒 𝑒𝑥𝑝𝑜𝑠𝑒𝑑
𝑡𝑜𝑡𝑎𝑙 # 𝑤𝑖𝑡ℎ 𝑜𝑢𝑡𝑐𝑜𝑚𝑒/𝑑𝑖𝑠𝑒𝑎𝑠𝑒
• Percentage/proportion of cases in an entire population that can be attributed to the
exposure
• Risk (incidence) in the overall population that can be attributed to the exposure
• Multiply findings by 100 to express as a percentage
• State results as “<Insert PAF as %> of <describe disease/outcome> in the population are
attributable to <describe exposure>
• Alternate formula when risk or rate ratio (indicated by RR in the formula below) is
known:
𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠 𝑖𝑛 𝑒𝑛𝑡𝑖𝑟𝑒 𝑠𝑡𝑢𝑑𝑦 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 × (𝑅𝑅 − 1)
𝑃𝐴𝑅 =
𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑠𝑢𝑏𝑗𝑒𝑐𝑡𝑠 𝑖𝑛 𝑒𝑛𝑡𝑖𝑟𝑒 𝑠𝑡𝑢𝑑𝑦 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 × (𝑅𝑅 + 1)
Odds Ratio (AKA Cross-Product Ratio

# 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 & 𝑤𝑖𝑡ℎ 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 (𝑎) # 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑏𝑢𝑡 𝑤𝑖𝑡ℎ 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 (𝑐) 𝑎𝑑
𝑂𝑅 = ( × )=
# 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑏𝑢𝑡 𝑛𝑜 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 (𝑏) # 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑎𝑛𝑑 𝑛𝑜 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 (𝑑) 𝑏𝑐

• The odds of exposure among cases and non-diseased controls

• Measure of association for a case-control study
o Because case-control studies don’t measure and compare the incidence, only the
odds ratio can be used as an estimate of relative effect size
o Quantifies the relationship between exposure and a disease
• Tells us how much higher the odds of exposure are among cases than among controls
o OR of 1
▪ Odds of exposure among cases and controls are the same or similar
▪ Exposure is not associated with the disease/outcome
o OR > 1
▪ Odds of exposure among cases are greater than odds of exposure among
controls
▪ Exposure may be a risk factor for the disease/outcome
▪ The further away the OR ratio is above 1, the more likely it is that the
relationship between the exposure and the disease/outcome is causal
o OR < 1
▪ Odds of exposure among cases are lower than the odds of exposure among
controls
▪ Exposure may be a protective factor against the disease/outcome
• Both the numerical value and strength of association should be expressed when
interpreted
o Example using an odd ratio of 10:
▪ The odds of <describe exposure> were ten times higher among cases than
controls.
▪ As the OR is greater than 1.0, <describe exposure> may be a risk factor for
<describe outcome/disease>
▪ The magnitude of the OR suggests a strong association between <describe
exposure> and <describe outcome/disease>