Cohort and Intervetion Studies

Learning Objectives

• Review prevalence, incidence, relative risk
• Concept of a cohort
• Measures used in cohort studies
• Concept of Intervention Studies
• Differences and similarities between cohort and intervention studies

Prevalence

Town T has 1000 residents. Doctors found 100 people with diabetes among the 1000 residents in a survey. What is the prevalence of diabetes in residents of T?

• A 100
• B 10 %
• C 1000
• D I do not know

Answer to Prevalence

• Number of people had diabetes = 100
• Total population of T = 1000
• Prevalence = number with health condition / Total number
• Hence, prevalence of diabetes in T = 100 / 1000 = 10%

Concept of person-time

In Town T, doctors studied 1000 people for 5 years. When the study began, everyone was free of diabetes; in the 5 years of study, no one moved out of town and no new person came in. How many person-years did they study?

• A 1000 person-years
• B 5 person-years
• C 5000 person-years
• D I do not know

Answer to person-time

• 1 person is followed for 1 year = 1 person-year
• 100 persons are followed for 1 year = 100 person-years
• 1000 people were followed for 5 years, then
• Correct answer = 5000 person-years

Concept of incidence

In Town T, doctors studied 1000 people for 5 years.

In the beginning of the study, everyone was free of diabetes, and in the 5 years of study, they found 100 NEW cases of diabetes.

What is the incidence of diabetes in T?

• A 100 per 5000 person-years
• B 100 per 1000 person-years
• C 100
• D I do not know

Answer to incidence

• Incidence = Number of NEW CASES / total number people X Time followed through
• In this case,
• Number fo NEW CASES = 100
• Time followed through = 5 years
• Total number of people = 1000
• Person-time = 1000 X 5 = 5000 person-years
• Incidence = 100 per 5000 person-years

A Study of Association

In Town T, doctors studied 1000 people for 5 years. When study BEGAN, EVERYONE was free of diabetes. 500 people reported they led SEDENTARY life (No exercise, and carbohydrate rich diet); 500 other people reported they led ACTIVE life.

Doctors wanted to study the ASSOCIATION between LIFESTYLE and RISK of DIABETES

Incidence of diabetes among SEDENTARY people

In Town T, doctors studied 1000 people for 5 years. In the beginning of the study, EVERYONE was free of diabetes. 500 people reported they led SEDENTARY lives. After 5 YEARS, 75 people in this group were diagnosed with diabetes; No one out of 500 people died or left or no new people entered T

What is the incidence of diabetes among SEDENTARY people?

• A 75 per 2500 person-years
• B 75 per 500
• C 75 per 1000 person-years
• D I do not know

Answer of incidence among SEDENTARY people

• 75 per 500 X 5 = 2500 person-years

Incidence of diabetes among ACTIVE people

Town T, doctors studied 1000 people for 5 years. In the beginning of the study, EVERYONE was free of diabetes. They found 500 people who led Physically Active lives. After 5 YEARS, 25 people in this GROUP were diagnosed with diabetes; no one out of 500 people died or left or no new people entered T

What is the incidence of diabetes among ACTIVE people?

(500 people followed for 5 years, 25 new cases of diabetes)

• A 25 per 2500 person-years
• B 25 per 500
• C 25 per 1000 person-years
• D I do not know

Answer to the INCIDENCE among ACTIVE people

• 25 per 500 X 5 = 2500 person-years

Incidence is Risk, which group has higher risk of diabetes?

• A SEDENTARY people
• B ACTIVE people
• C Both groups have same risk
• D I do not know

Answer to which group has higher risk

What is the RR among the SEDENTARY people for diabetes?

(RR = Relative Risk or Rate Ratio)

• A 1/3
• B 3
• C I do not know

Answer to the Relative Risk or Rate Ratio or RR

• RR = Relative Risk or Rate Ratio or Risk Ratio
• Risk of diabetes among SEDENTARY people: 75 per 2500 person-years
• Risk of diabetes among ACTIVE people: 25 per 2500 person-years
• Therefore RR = 75 per 2500 person-years / 25 per 2500 person-years = 3

Concept of Cohorts

Cohort = Group

• Each group will have something in common among group members
• Each such GROUP constitutes a COHORT
• If there is nothing common among ALL GROUP members, then it is NOT a cohort

Concept of cohorts with Lifestyle and Diabetes

How many cohorts?

In town T, doctors studied 1000 people for 5 years. 500 people reported they led SEDENTARY life (No exercise, and carbohydrate rich diet); 500 other people reported they led ACTIVE life. Doctors wanted to study the ASSOCIATION between LIFESTYLE and RISK of DIABETES. How many cohorts are there in this study?

• A 1
• B 2
• C I do not know

Answer to how many cohorts

• In one group all members were SEDENTARY, this was COMMON among them
• In another group, all members were ACTIVE, this was COMMON among them
• Hence, number of COHORTS = 2

Exposure and Outcome

• Exposure is what people are “Exposed to”
• Outcome, what happens as a result of that Exposure
• Example: If you want to study whether Air Pollution Causes Asthma, then
• Air Pollution is Exposure
• Asthma is Outcome

Characteristics of Cohorts in a Cohort Study

• Some COHORT will have One/Other type of Exposure
• Cohorts MUST be identified BEFORE OUTCOME can occur
• When Cohorts are identified, everyone in ALL cohorts MUST BE FREE of Outcome
• ALL Cohorts MUST be followed through in time
• For ALL cohorts, INCIDENCE rates are calculated and compared

Is this a cohort study?

In town T, doctors studied 1000 people for 5 years. In the beginning of the study, EVERYONE was free of diabetes. 500 people reported they led SEDENTARY life (No exercise, and carbohydrate rich diet); 500 other people reported they led ACTIVE life. Doctors followed each GROUP for 5 years to study their RISK of DIABETES. Is this a COHORT STUDY?

• A Yes
• B No
• C I do not know

Answer to “Is this a cohort study”

• In the beginning everyone was FREE of DIABETES
• Some people had SEDENTARY, others had ACTIVE LIFESTYLE as Exposure
• ALL Cohorts were followed in time
• Therefore this is a COHORT STUDY
• This type of Cohort study is referred to as Prospective Cohort Study

What is the exposure?

In a factory F, doctors wanted to study if working with a chemical C was a risk factor for Diabetes. They identified 100 factory workers who worked with chemical C and 100 other workers who did not work with chemical C from the registers when HEALTHY employees with No Diabetes first started working in F. What is the Exposure?

• Working with Chemical C
• Work in the factory
• Diabetes
• I do not know

Answer to “What is the exposure?”

• Working with Chemical C is the EXPOSURE
• Diabetes is the outcome
• Although many people might work in the factory,
• Only those WITH and WITHOUT working with Chemical C is important here

Is this a cohort study as well?

Doctors identified 200 factory workers, 100 of whom worked with Chemical C, and 100 of whom DID NOT work with Chemical C initially free of diabetes; **50 in EXPOSED group SUBSEQUENTLY developed diabetes after 5 years; 25 in NONEXPOSED group subsequently developed diabetes after 5 years. Is this a Cohort study?

• A Yes
• B No
• C I do not know

Answer to “Is this a cohort study as well”

• In the beginning of identification of the cohorts, ALL employees
• Were healthy with no signs of diabetes
• Some were exposed to C
• Some were NOT EXPOSED to C
• All were followed up till 5 years later
• All were tested to see if they developed Outcome (Diabetes)
• Hence this IS a cohort study
• This is referred to as Retrospective cohort study

How would you INCLUDE someone in a Cohort study?

• First determine the exposure
• Then, ON THE BASIS of exposure,
• Identify an EXPOSED cohort (group that is exposed)
• Identify a NON-EXPOSED cohort (group that is non-exposed)
• Make sure that ALL participants are FREE of OUTCOME

Question: what determines who can be included a cohort study?

• A Whether they are EXPOSED or NON-EXPOSED
• B Whether they have the OUTCOME or NOT
• C BOTH
• D I do not know

Answer: what determines

• All persons must be free of OUTCOME in the beginning
• Therefore, if a person has an outcome, then that person cannot be included
• All persons must EITHER have or have not EXPOSURE under study
• Therefore, correct answer: BOTH

Concept of Results Table of Cohort Study

• PYEXP = Person-Time of Exposed
• NEXP = Number of Cases among Exposed
• PYNEXP = Person-Time of Non-Exposed
• NNEXP = Number of Cases among Non-Exposed

Results Table of the LIFESTYLE-DIABETES study

• PYEXP = Person-Time of Exposed = 500 X 5 = 2500
• NEXP = Number of Cases among Exposed = 75
• PYNEXP = Person-Time of Non-Exposed = 500 X 5 = 2500
• NNEXP = Number of Cases among Non-Exposed = 25

Chemical-Diabetes study

In a factory F, doctors wanted to study if working with a chemical C was a risk factor for Diabetes.

They identified 100 factory workers who worked with chemical C and 100 other workers who did not work with chemical C from the registers when HEALTHY employees with No Diabetes first started working in F.

Then found 50 in EXPOSED group SUBSEQUENTLY developed diabetes after 5 years;

They found 25 in NONEXPOSED group subsequently developed diabetes after 5 years

Results table of the Chemical-Diabetes Study

• Each COHORT of 100 people were followed for 5 years

What is the correct figure for ??

• A 500
• B 100
• C I do not know

Answers to the results of Chemical-Diabetes study

• 100 people in EACH group
• Followed up for 5 years
• Hence, Person-Time = 500 person-years

Measures of Association studied in Cohort Studies

• Rate Ratio or Relative Risk (RR)
• RR = Risk of outcome among the EXPOSED / Risk of outcome among Non-EXPOSED

Advantage and limitations of Cohort Studies

• Can study MANY OUTCOMES from common exposure
• Strong study design to generate evidence of CAUSE and EFFECT
• Here, investigators can ONLY OBSERVE
• Investigators CANNOT manipulate the EXPOSURE for participants
• TIME CONSUMING and EXPENSIVE

Concept of Observation vs Intervention

• EXPOUSURES in Nature are BEYOND the control of Investigators
• Investigators can only WATCH and RECORD the EXPOSURE status
• Such studies are “OBSERVATIONAL STUDIES”
• If the Investigators can INTERVENE OR ACT, then
• Such study is an INTERVENTION TRIAL

What kind of a study is this?

In town T, doctors studied 1000 people for 5 years. In the beginning of the study, EVERYONE was free of diabetes. 500 people reported they led SEDENTARY life; 500 other people reported they led ACTIVE life. Doctors followed each GROUP for 5 years to study their RISK of DIABETES

• A Observational Study
• B Intervention Trial
• C I do not know

Answer to the kind of study

• Sedentary Life or Active Life was NOT CONTROLLED by the doctors
• Doctors were only going to observe what would happen
• Hence this is an Observational Study

What kind of a study is this again?

In hospital H, doctors studied 1000 DIABETIC patients for 5 years. In the beginning of the study, EVERYONE was free of FOOT ULCER. They SELECTED 500 patients who received special socks to protect their feet, and 500 other patients received usual care. Doctors followed each GROUP for 5 years to study their RISK of FOOT ULCER.

• A Observational Study
• B Intervention trial
• C I do not know

Answer to the kind of study again

• Doctors DECIDED who will receive what treatment (sock or usual treatment)
• Doctor INTERVENED rather than watch the patients develop foot ulcer
• This is why this is an Intervention study

Characteristics of an intervention trial

• In the beginning ALL participants MUST be free of the OUTCOME under study
• ONE group will receive the ACTIVE INTERVENTION
• Other group will NOT receive the intervention
• Investigators will FOLLOW UP BOTH groups to study OUTCOMES
• Investigators will COMPARE OUTCOMES in both groups

Similarities of Intervention trials with Cohort Study

• Participants CANNOT have the OUTCOME of interest in the beginning
• Investigators MUST FOLLOW UP the participants
• In BOTH, investigators MUST assess INCIDENCE Rates
• In BOTH, measures of association are RR

Differences between intervention trials and Cohort studies

• Intervention Trials, EXPOSURE is under the control of the Investigator
• Cohort Studies, EXPOUSRE is NOT under Investigator’s control
• Investigator can ALLOCATE who will receive what EXPOSURE
• Investigator can only OBSERVE in Cohort study

What kind of Measure of Association is appropriate for Intervention trials?

• A Prevalence
• B Rate Ratio or Relative Risk
• C I do not know

Answer to Measure of Association for Intervention trial

• Rate Ratio or Relative Risk (RR) because
• Investigators COMPARE
• incidence of outcome among those who DO receive the intervention and
• Incidence of outcome among those who DO NOT receive the intervention

Summary of key points

• Cohort studies and intervention studies are two different study designs
• Cohort study is an OBSERVATIONAL study design
• Intervention study allows for Investigator to INTERVENE or CONTROL Or ALLOCATE exposures
• Both study designs allow for estimation of relative incidence
• Intervention trials are similar to controlled experiments

Questions and Comments

• Please complete the post-class survey