IRR Fundamentals
What is: The number of new cases (of a disease), devided by total person-time at risk
Incidence Rate
Exposed: 20 cases / 1,000 person-year
Unexposed: 10 cases / 1,000 person-year
IRR=?
2.0
What is the most common study design where IRR is used?
Cohort Studies
What were the 3 studies examined in our main article?
Dietary Inflammatory Index (DII)
Alternative Healthy Eating Index-2010 (AHEI-2010)
Mediterranean Diet Score (MDS)
An IRR of 0.67 means the exposed group has this relative to the unexposed group
33% lower incidence
What is: Incidence Rate in the exposed group / Incidence rate in unexposed group
Incidence Rate Ratio (IRR)
Exposed: 15 events over 500 person-years.
Unexposed: 10 events over 800 person-years.
The IRR is approximately:
2.4
When is IRR prefered over risk ratios?
when there is varying lengths of follow up (person-time)
What was the IRR for the highest vs lowest quintile of the Dietary Inflammatory Index and type 2 diabetes?
1.49
An IRR of 1.49 means the exposed group has this relative to the unexposed group.
The exposed had a 49% higher incidence rate?
Q: an IRR of 1.0 indicates what in the relationship between exposure and outcome?
That there is no association
In the British Doctors Study, non-smokers had ~10 lung cancer deaths per 100,000 person-years and current smokers had ~83.
What is the approximate IRR?
8.3
A classic study where IRR was used in the assessment of British doctors, looked at what factors?
smoking and lung/lung cancer deaths
What was the IRR for the highest versus lowest quintile of the Healthy Eating Index (AHEI-2010) and incident type 2 diabetes.
0.67
IRR is particularly useful for rare outcomes or long follow-up because it accounts for this.:
Person-time at risk
An IRR >1.0 (greater) indicates what about the exposed group compared to the unexposed group?
The exposed group had a higher incidence rate
Exposed: 40 events / 2,000 person-years.
Unexposed: 20 events / 4,000 person-years.
The IRR is this.
= 4.0
What type of Randomized human trials use IRR
Pharmacoepidemiology
In our article, was diet-diabetes association mediated by body size (BMI/Waist to Hip Ratio)?
Yes
In public health, an elevated IRR for a modifiable exposure helps identify this.
potential targets for prevention
A group of Nutrition students at TMU go to the rooftop garden every week, to buy hot peppers. While another group of TMU Nutrition students don't go to the rooftop garden. The IRR of low assignment scores among those who went is 0.35. What is the difference in rate of low scores among those who go to the rooftop garden, vs those who dont?
Those who went to the rooftop garden and bought hot peppers, had a 65% lower chance of low assignment scores
Unexposed: 98,877 in 38,000 person-year
IRR=
IRR = 1.07
What type of studies among workers at factory use IRR
Occupational studies of workplace exposure, commonly cancer risk
What was two limitations stated by the authors?
Self-reported data from FFQ
Self-reported diabetes diagnosis
Self-reported Physical activity
The mediation analysis in our article suggested that body size explained a substantial portion of the association, implying this practical implication:
That improving diet quality may reduce diabetes risk partly by influencing body composition?