# Measures of Association --- ## Case study: COVID19 and CFS - COVID19 is known to lead to Chronic Fatigue Syndrome ("CFS") - CFS is marked by extreme fatigue for six months or more - Fatigue not relieved by rest - Sleep does not help - Problem with memory, focus, concentration - Dizziness --- ## Case study: Epidemiologists study the linkage - Select 1000 COVID19 positive people in 2020 ("EXPOSED") - Select another 1000 people who did not have COVID19 ("Non-EXPOSED") - At the beginning of study, NO GROUP have CFS - Follow BOTH Groups for 5 YEARS - Count HOW MANY Develop CFS in BOTH groups in 5 years - Compare the two groups --- ## Question 1: How many person-years? - (X) 1000 person for 5-years = 5000 person-years (or 5000 py) - (Y) 1000 person-years - (Z) 5 person-years --- ## Case study: Findings - After 5 years: - Among the EXPOSED we found 200 people with CFS - Among the Non-EXPOSED, we found 50 people with CFS --- ## Question 2: What is the incidence-rate among the EXPOSED? - (X) 200 per 5000 py - (Y) 200 - (Z) 200 per 1000 --- ## Question 3: What is the incidence-rate among the Non-EXPOSED? - (X) 50 per 5000 py - (Y) 50 - (Z) 50 per 1000 py --- ## EXCESS RISK (Attributable risk, AR) - Also called Risk Difference - if the risk of a disease D among the EXPOSED is R1, and - If the risk of disease D among the non-EXPOSED is R2 - Then Attributable Risk = R1 - R2 - (R1 - R2), or the Risk Difference is ATTRIBUTABLE to Exposure --- ## Question 4: What is the AR for CFS and COVID19? - (X) 50 per 5000 py - (Y) 150 per 5000 py - (Z) 100 per 5000 py --- --- ## Concept of Attributable Proportion (AP) - Also called attributable risk percent - The PROPORTION of risk among the EXPOSED that - is DUE TO THE EXPOSURE - Answers **"How Much of the risk among Exposed is due to Exposure?"** --- ## Measurement of Attributable Proportion (AP) - If R2 is the Risk among the Exposed, AND - If R1 is the Risk among Non-EXPOSED, THEN - (R1 - R2) is the EXCESS RISK (due to Exposure), So - (R1 - R2) / R2 is the attributable proportion --- ## Question 5: What is the Attributable Proportion (AP) of CFS in COVID19 patients? - if 150 per 5000 py is EXCESS RISK, AND - if 200 per 5000 py is RISK in the EXPOSED, THEN which is correct? - (X) AP is 75% - (Y) AP is 15% - (Z) AP is 20% --- ## Explanation of Attributable Proportion in this case - if 150 per 5000 py is EXCESS RISK, AND - if 200 per 5000 py is RISK in the COVID19, THEN - Attributable Proportion = (150 / 200) = **75%** - **75% of the CFS among COVID19 patients is due to COVID19 infection** --- ## Relative Risk (RR) - **How many times** higher is the RISK of disease among EXPOSED - Compared with RISK of disease among Non-EXPOSED - Also called Rate Ratio or Risk Ratio - if risk of disease among exposed is R1, AND - risk of disease among non-EXPOSED is R2, THEN - Relative Risk (RR) = R1 / R2 --- ## Question 6: What is the RR of CFS among COVID19 people? - If 200 per 5000 py is incidence (RISK) among COVID19, AND - If 50 per 5000 py is incidence (RISK) among non-COVID19, THEN - What is the Rate Ratio OR Relative Risk (RR)? - (X) 150 - (Y) 4 - (Z) 1/4 --- ## Explanation of RR of CFS among COVID19 people - If 200 per 5000 py is incidence (RISK) among COVID19, AND - If 50 per 5000 py is incidence (RISK) among non-COVID19, THEN - Rate Ratio = (200 / 50) = **4**, this means - CFS in COVID19 positive people are **FOUR** times more likely --- --- ## Concept of Population Attributable Risk - If E is an exposure and D is the disease, then - What PROPORTION of disease D in the COMMUNITY, is - Attributed to exposure E? - That is, - If E WERE to be ELIMINATED completely, by how much would D be also reduced in the community? --- ## Measurement of PAF - PAF = proportion of diseased that were exposed * Attributable Proportion - Also called population attributable risk percent (PAR%) - Another way: - prevalence of exposure * ((RR - 1) / (1 + prevalence of exposure * (RR - 1))) --- ## In this case ... - Prevalence of COVID19 = 0.01 (or 1%) - Relative Risk (RR) = 4 - Population attributable risk % = (0.01 * 3 / (1 + 0.01 * 3)) = 0.03 / 1.03 = 2.9% - This means **if COVID19 were to be eliminated, it would reduce CFS by 2.9%** in the community --- ## Conclusion - Introduced several measures of Association - Other measures include Odds Ratio (did not discuss here) - Odds Ratio is similar in concept to Risk Ratio - Compares Odds of Exposure - Discuss in the context of study designs