Class-1 Measures of Frequency and Association PDF

Measures of Disease Frequency and Association Measuring Disease Frequency What is Epidemiology? Epidemiology is the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to the control of health problems. epi=upon de...

Measures of Disease Frequency and Association Measuring Disease Frequency What is Epidemiology? Epidemiology is the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to the control of health problems. epi=upon demos=people logos=study “the doctrine of what is among or happening to people” "the study of what is upon the people“ Objectives Identify the factors that are causes of disease. ▪ Count the cases of disease, ascertaining how much disease is in a population. ▪ Study the natural history and prognosis of disease. ▪ Guide and evaluate interventions to improve health. ▪ Disease Frequency  As clinicians, you may be required to estimate the frequency of diseases  This skill would require an understanding of:    Prevalence  Point Prevalence  Period Prevalence Incidence  Cumulative Incidence  Incidence Rate Mortality, Morbidity, Attack Rate Prevalence  The number of people with a disease in the population divided by the total number of people in the population # of people with a disease at a specified Prevalence = ----------------------------------------------------total # of people in the population time Prevalence What fraction of the population is affected by the disease at a given time Ranges from 0 to 1 Useful for assessing the burden of disease in a population Valuable for planning health services NOT useful for determining what caused disease Tell us how much but not why or how Prevalence Point Prevalence Prevalence Period Prevalence Point and Period Prevalence   Point prevalence is the proportion of a population that has disease at a specific “point” in time. It is the proportion of persons with a particular disease on a particular date/time.  The “point” can be a specific calendar time, or…  The “point” can be a lifetime “event” (birth, death, entry into the military). Period prevalence is the proportion of a population that has the disease during a given time period of interest. “Past 12 months” is a commonly used period. “Point” Prevalence  Suppose there are 20 students in a class, X represent those who caught flu. Assume those who caught flu recovered in 7 days:  January February March What is the estimated point prevalence for this population on 28th February?  P28 Feb=(# diseased cases ) 28 Feb / total population = 2/20= 10% Period Prevalence  Suppose there are 20 students in a class, X represent those who caught flu. Assume those who caught flu recovered in 7 days: January February March  What is the period prevalence during January?  PJan=(# diseased cases)Jan/total population =3/20= 15% Incidence Incidence  Frequency of new cases during a span of time in people “at risk”.  Focus is on measuring the probability of developing disease during a span of time. # 𝒏𝒆𝒘 𝒄𝒂𝒔𝒆𝒔 𝒅𝒖𝒓𝒊𝒏𝒈 𝒂 𝒈𝒊𝒗𝒆𝒏 𝒕𝒊𝒎𝒆 𝒑𝒆𝒓𝒊𝒐𝒅 𝐈𝐧𝐜𝐢𝐝𝐞𝐧𝐜𝐞 = 𝒕𝒐𝒕𝒂𝒍 𝒑𝒐𝒑𝒖𝒍𝒂𝒕𝒊𝒐𝒏 "𝒂𝒕 𝒓𝒊𝒔𝒌" Population “At Risk” Incidence should be assessed in people who are “at risk” of developing the outcome. The denominator should NOT include:  Those who already have the disease.  Those who can’t get it, such as those who are immune or don’t have the organ. Incidence Cumulative Incidence Risk Incidence Rate  Both focus on # new cases of disease (numerator) during a period of observation.  The difference is the way they handle time. Cumulative Risk aka Incidence Proportion  The number of new cases of a disease that occur during a specified period of time divided by the number of persons at risk of developing the disease at beginning of the time period Incidence = # of new cases of disease over a specific period of time # of persons at risk of disease at the beginning of the study period Cumulative Risk Example  A study in 2020 examined asthma among persons with dementia. The study recruited 200 adults with dementia. Of the 200, 60 had a prior diagnosis of asthma. Over the first year, 7 adults developed asthma.  𝐈𝐧𝐜𝐢𝐝𝐞𝐧𝐜𝐞 = =  # 𝒏𝒆𝒘 𝒄𝒂𝒔𝒆𝒔 𝒅𝒖𝒓𝒊𝒏𝒈 𝒂 𝒈𝒊𝒗𝒆𝒏 𝒕𝒊𝒎𝒆 𝒑𝒆𝒓𝒊𝒐𝒅 𝒕𝒐𝒕𝒂𝒍 𝒑𝒐𝒑𝒖𝒍𝒂𝒕𝒊𝒐𝒏 "𝒂𝒕 𝒓𝒊𝒔𝒌" 7 200-60 = 0.05 The one year incidence of asthma among adults with dementia is 5.0%  Can also be expressed as 50 cases per 1,000 persons with dementia Cumulative Risk Its not valid when large number of people leave your study Why would people not stay in your study?  Loss to follow-up occurs when people drop out of your study for any reason Why is loss to follow-up a problem? The dropout rates are different between study groups; or The patients who drop out are different from those who do not drop out. Risk (incidence proportion)  Is it then valid to use risk as a measure of disease frequency? YES!  When? Risk is more valid in studies where the follow-up period is short and loss to follow-up and competing risk is low.  Eg: Effectiveness of flu vaccine in children Incidence Rate  One solution to dealing with competing risk and loss to follow up problems faced by incidence proportion is to consider incidence rates.  Incidence rate is the rate at which people come down with a disease/illness.  It is the number of new cases of disease divided by the person time at risk, or the total amount of time that people were at risk to get the disease of interest. Incidence Rate  Incidence-density rate  Person-time rate  Hazard rate  Incidence rates are concerned with the number of new cases during a particular person-time of follow-up  Time at risk-the time during which each person is at risk of getting the disease. This means if a person has already contracted the illness/disease, they often are not at risk to get it again. Incidence Rate  Incidence Rate= Number of new cases of disease Total time at risk of persons followed, or amount of person time at risk Conceptually, person-time=[#persons]X[amount of time] In practice though many people will contribute different amounts of person time, due to different times they are enrolled, and different times that they develop disease throughout the study. Assumptions  People are at risk for a disease until: They have the disease event. Then they are no longer at risk for experiencing that event (eg: having a ‘first heart attack’-disease cannot recur) Or they die Or they are lost to follow-up Or the study ends A person-time example #1 Blue means that the person was at risk. : ‘Person Time at Risk’ Green means that the person was NOT at risk Lightning Bolt means that the person had an event (disease event) Understanding person-time (for ex: years or days) T1 P1 P2 People P3 P4 P5 P6 P7 P8 P9 P10 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T20 A person-time example #1 Blue means that the person was at risk. : ‘Person Time at Risk’ Green means that the person was NOT at risk Lightning Bolt means that the person had an event (disease event) T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T20 P1 NOT at risk Person time at risk T1+T2+T3+T4+T5+…….+T13+T14=14 units (this is what we are interested in) Person time T15+T16+T7+T18+T19+T2 0=6 units (this is not of interestdon’t have to calculate it) A person-time example #1 Blue means that the person was at risk. : ‘Person Time at Risk’ Green means that the person was NOT at risk Lightning Bolt means that the person had an event (disease event) Understanding person-time (for ex:years or days) People (person 1, person , etc) T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Calculate time at risk for person 4 and person 5 T16 T17 T18 T19 T20 A person-time example #1 What is the incidence rate of disease from T1 to T20? Understanding person-time (for ex:years or days) People (person 1, person , etc) T1 P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T20 Incidence Rate  Incidence Rate= Number of new cases of disease Total time at risk of persons followed, or amount of person time at risk  Incidence rate= 3 14+20+11+11+20+20+10+20+2+10 =3/138=0.022 or 22 per 1000 person time Incidence: Summary  CR and IR focus on measuring the transition from health to disease  CR  Easy  Less to calculate and understand accurate when its assumptions are not met  More  useful for fixed populations IR  Greater accuracy  Person-time  More denominator more difficult to calculate and understand useful for dynamic populations Prevalence vs. Incidence 2003 2004 2005 2006 X X XX X Prevalence: Existing cases 2007 XX X 2008 2009 X X XX 2010 Incidence: Frequency of new cases during a span of time in people at risk. Prevalence is the probability of having disease at a point in time/ or a given period of time. Incidence is the probability of developing disease during a span of time. Which measure of disease frequency best describes: “The frequency of new cases of tuberculosis in Boston over a calendar year”? A. Prevalence B. Cumulative incidence C. Incidence rate D. None of the above Which measure of disease frequency best describes: “The number of previously healthy women who had a stroke during 40,000 person-years of follow up”? A. Prevalence B. Cumulative incidence C. Incidence rate D. None of the above Relationship between Prevalence and Incidence Prevalence ~ Incidence x Duration What will happen to prevalence of a disease if a new treatment increases the survival among those affected by the disease (all else remains the same)? A. Prevalence will increase B. Prevalence will decrease C. Prevalence will stay the same D. Not enough information How would prevalence of a disease change when a new treatment prevents new cases of the disease from occurring? (Everything else remains the same) A. B. C. D. Stay the same Increase Decrease Fluctuate in either direction Mortality Rate Mortality Rate is the number of deaths divided by the total population during a specified time period. In one year there were 1,807 deaths from TB in the US (population=231,534,000), so the mortality rate for TB was 7.8 per million. 1,807/231,534,000 = 7.8/million per year In the same year there were 1,973,000 deaths from all causes in the US, so the all-cause mortality rate was 852/100,000 population. 1,973,000 /231,534,000 = 852/100,000/year Morbidity Rate Morbidity Rate is the incidence of non-fatal cases of a disease in a population during a specified time period. Example: in 1982 there were 25,250 non-fatal cases of TB in US population. Midyear population was 231,534,000, Morbidity rate of TB =25,520 / 231,534,000 = 11.0 / 100,000 in 1982 Attack Rate The proportion of those exposed that develop the disease. COVID-19 exposure Measures of Association Causation  What is a cause? A cause of a disease is an event, condition, or characteristic that preceded the disease event and without which the disease event would not have occurred at all or would not have occurred until some later time. Individual vs. Population Cause An exposure can be a cause of disease even if every individual who is exposed does not develop disease.  We mean cause in a probabilistic sense. (where everyone exposed has a higher “risk” of developing disease).  …Not a deterministic sense (where everyone exposed to an exposure gets the disease).  This is an important thing to understand about how epidemiologists view “cause”. Association v. Causation Association v. Causation Exposure/risk factor (X) Diseases/outcome (Y) Exposure/risk factor (X) Diseases/outcome (Y) Association: X and Y tend to occur together Causation: The presence of X brings about the presence Y Models of Causation  Miasmas- 19th century  Germ Theory  Koch’s postulates  Webs of causation  Hill’s criteria for causality  Sufficient-Component Cause Model Evidence for Causality: Hill’s criteria for Cause  Temporality  Strength of association  Replication of findings/Consistency  Specificity of association  Biological Gradient/Dose-response  Biological plausibility  Coherence with established facts  Experimental Evidence -- Cessation of exposure  Analogy Temporality  Exposure must precede disease Strength of Association  The magnitude of the association  Stronger the association, the more likely there is a causal relationship  Weaker the association, the more likely it is explained by bias, confounding or random error Dose-Response Relationship  Higher amounts of exposure associated with higher risk of disease Replication of Findings  Relations that are demonstrated in multiple studies are more likely to be causal  Consistency across multiple studies Biologic Plausibility  The proposed mechanism should be biologically (etiologically) plausible in the context of current biological knowledge.  Example: Gambling causes lung cancer?  Found increased risk of lung cancer among gamblers  Biologically plausible? Lung Cancer Gambling Smoking Measures of disease association A measure of association quantifies the relationship between exposure and disease. Measures of Association Absolute vs Relative scale  Relative differences tell us the relative increase or decrease in effect comparing one quantity to another.  Absolute differences tell us the absolute increase or decrease in effect. Measures of Association Relative Scale  Prevalence ratio  Risk ratio  Incidence rate ratio  Odds ratio Measure of Association Prevalence Ratio = 𝑃𝑟𝑒𝑣𝑎𝑙𝑒𝑛𝑐𝑒 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑃𝑟𝑒𝑣𝑎𝑙𝑒𝑛𝑐𝑒 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 Then we calculate prevalence for each of these exposure groups, separately Prevalence ratio is not used commonly in epidemiology, but important to understand it conceptually Measure of Association Risk Ratio (RR)= 𝑅𝑖𝑠𝑘 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑅𝑖𝑠𝑘 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑  It ranges from 0 to +∞, and has no units.  We need to divide the population into two groups: those who are exposed and those who are not exposed.  Then we calculate the risk of disease in each exposure group separately, to get the risk of disease in the exposed and the risk of disease in the unexposed. Measures of Association  The Incidence Rate Ratio (IRR) is the ratio of incidence rates for two groups, the exposed vs the unexposed. IRR = 𝐼𝑅 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝐼𝑅 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 It ranges from 0 to +∞, and has no units. Measures of Association  Odds Ratio (OR) OR = 𝑂𝑑𝑑𝑠 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑂𝑑𝑑𝑠 𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 Calculating the measures of association  2X2 Table  a- # of people with disease who are exposed  b-# of people without the disease who are exposed  c-# of people with the disease who are not exposed  d-# of people without the disease who are not exposed Disease No Disease Total Exposed a b a+b Not Exposed c d c+d Total a+c b+d Calculating the measures of association Risk (exposed) Risk (unexposed) = = Risk Ratio (RR)= 𝑎 𝑎+𝑏 𝑐 𝑐+𝑑 𝑎 𝑎+𝑏 𝑐 𝑐+𝑑 Calculating the measures of association   Odds exposed = 𝑝𝑒𝑥𝑝𝑜𝑠𝑒𝑑 = 1−𝑝𝑒𝑥𝑝𝑜𝑠𝑒𝑑 Odds unexposed = 𝑎 𝑎+𝑏 𝑎 1− 𝑎+𝑏 = 𝑝𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 = 1−𝑝𝑢𝑛𝑒𝑥𝑝𝑜𝑠𝑒𝑑 Odds Ratio (OR)= 𝑎 𝑏 𝑐 𝑑 𝑎 𝑎+𝑏 𝑎+𝑏−𝑎 𝑎+𝑏 = 𝑐 𝑐+𝑑 𝑐 1− 𝑐+𝑑 = 𝑎∗𝑑 𝑐∗𝑏 = 𝑐 𝑐+𝑑 𝑐+𝑑−𝑐 𝑐+𝑑 𝑎 𝑏 = 𝑐 𝑑 this is cross product Calculating the measures of association  We cannot calculate the rate and prevalence measures with this specific data from the 2X2 table.  Why not? No person-time information No information on prevalence only disease incidence data Measures of Association Absolute Scale Risk Difference (RD)= Risk (exposed) −𝑅𝑖𝑠𝑘(unexposed)  It ranges from -1 to +1, and has no units.  This is the additional risk among those exposed when compared to those who are unexposed. Measures of Association  Incidence Rate Difference (IRD)  The difference of incidence rates comparing those who are exposed vs those unexposed. IRD= IR(exposed) −𝐼𝑅(unexposed) It ranges from -∞ to +∞, and has unit of time. Interpreting the measures of association  Null means there is no effect, or no association between the exposure and disease of interest.  If there is no effect on the relative scale then the PR, RR, IRR and OR will be equal to 1 Relative scale the null is equal to 1 Interpreting the measures of association  Null means there is no effect, or no association between the exposure and disease of interest.  If there is no effect on the absolute scale then the RD, IRD will be equal to 0. Absolute scale the null is equal to 0 Interpreting the measures of association  What if the RR or IRR or OR is greater than 1 ?  What if the RD or IRD is greater than 0?  Those who are exposed exhibit a higher (risk/rate/odds) of disease, compared to those who are unexposed.  Positive Association Interpreting the measures of association  What if the RR or IRR or OR is less than 1 ?  What if the RD or IRD is less than 0?  Those who are exposed exhibit a lower (risk/rate/odds) of disease, compared to those who are unexposed.  Negative Association Example-1 Had Incidental Appendectomy? Wound Infection No Wound Infection Total Yes 7 124 131 No 1 78 79 Risk in exposed=7/131=5.34% Risk in unexposed=1/79=1.27% Risk Ratio=5.34/1.27=4.2 Interpretation: In this study patients who underwent incidental appendectomy had 4.2 times the risk of post-operative wound infection compared to patients who did not undergo incidental appendectomy. http://sphweb.bumc.bu.edu/otlt/mph-modules/ep/ep713_association/ep713_association_print.html  Risk Difference=5.34-1.27 = 4.07  Interpretation: Subjects who had an incidental appendectomy had 4 additional cases of wound infection per 100 people compared to subjects who did not have an incidental appendectomy. http://sphweb.bumc.bu.edu/otlt/mph-modules/ep/ep713_association/ep713_association_print.html Example-2 Treatment Myocardial Infarction No Infarction Total Risk Aspirin 139 10,898 11,037 139/11,037 = 0.0126 Placebo 239 10,795 11,034 239/11,034 = 0.0217 Risk in exposed=0.0126 Risk in unexposed=0.0127 Risk Ratio=0.0126/0.0217=0.58 Interpretation: Those who take low dose aspirin regularly have 0.58 times the risk of myocardial infarction compared to those who do not take aspirin. http://sphweb.bumc.bu.edu/otlt/mph-modules/ep/ep713_association/ep713_association_print.html  Risk Difference=1.26-2.17= -0.91  Interpretation: Those who take low dose aspirin had 1 less case of myocardial infarction per 100 people compared to those who took the placebo drug. http://sphweb.bumc.bu.edu/otlt/mph-modules/ep/ep713_association/ep713_association_print.html Measures of Association  Risk Ratio and Risk Difference are different takes on the same information.  Risk Ratio provides information on the strength of the association.  Risk Difference provides a measure of the public health impact of the risk factor, and focuses on the number of cases that could potentially be prevented by eliminating the risk factor. Sugar consumption, Dental Caries and Obesity The Relative Risk (RR) for high sugar consumption and dental caries is 4.5; RR for high sugar consumption and obesity is 1.4. High sugar consumption is a stronger risk factor for? A. Dental Caries B. Obesity C. Both D. None Calculate Prevalence (of autism), Relative Risk & Risk Difference Autism Vaccinated for MMR 25 Not Vaccinated for 20 MMR Total 45 A. 20/100, 0.8, 5 B. 45/200, 1.25, 5 C. 25/200, 1.25, -5 D. 45/200, 0.8, 5 No Autism Total 75 80 100 100 155 200 Coming Up  Study Designs