Cohort Study 3

COHORT STUDIES Dr. A. K. AVASARALA MBBS, M.D. PROFESSOR & HEAD DEPT OF COMMUNITY MEDICINE & EPIDEMIOLOGY PRATHIMA INSTITUTE OF MEDICAL SCIENCES, KARIMNAGAR, A.P. INDIA: +91505417 [email protected]

LEARNING OBJECTIVES LEARNER SHOULD BE ABLE TO KNOW AT THE END OF THIS LECTURE • WHAT IS A COHORT STUDY? • WHEN TO CONDUCT A COHORT STUDY? • WHAT TYPES OF COHORT STUDIES ARE AVAILABLE? • HOW TO TEST THE HYPOTHESIS BY MEANS OF A COHORT STUDY?

PERFORMANCE OBJECTIVES LEARNER SHOULD BE ABLE TO • CONDUCT THE COHORT STUDY FOR ANY NEW HYPOTHESIS WHICH HE IS THINKING CAUSALLY RELATED • APPLY THIS ANALYTICAL METHOD WHEN SITUATION DEMANDS

COHORT STUDIES • ANALYTICAL OBSERVATIONAL STUDIES • ALMOST SIMILAR TO THE EXPERIMENTAL STUDIES

WHAT IS A COHORT? • COHORT IS A GROUP HAVING A COMMON CHARACTERISTIC • e.g. A SMOKER’S COHORT MEANS ALL ARE SMOKERS IN THAT GROUP

PURPOSE • THEY TRY TO PROVE THE CAUSAL INFLUENCE ON THE DISEASE OUTCOME IN A FORWARD DIRECTION FROM CAUSE TO THE EFFECT AND IN A COMPARATIVE FASHION

INDICATIONS 1. FOR FURTHER CONFIRMATION OF CAUSAL HYPOTHESIS ALREADY SUGGESTED BY CASE CONTROL DESIGN 2. COHORT METHOD IS DIRECTLY CARRIED OUT WHEN INDEX OF CAUSAL SUSPICION OF HYPOTHESIS IS VERY GREAT

COHORT STUDY DESIGN • A COHORT, WHICH IS EXPOSED TO A SUSPECTED FACTOR BUT NOT YET DEVELOPED THE DISEASE, IS OBSERVED AND FOLLOWED OVER TIME • THEN, THE INCIDENCE OF THE DISEASE, THE EXPOSURE RISK, THE DISEASE RISK AND THE POPULATION RISKS ARE MEASURED DIRECTLY

COHORT STUDY DESIGN EXPOSURE / SUSPECTED FACTOR (SMOKING)

DISEASE DEVELOPED (LUNG CANCER PRESENT)

DISEASE TOTAL NOT DEVELOPED (LUNG CANCER ABSENT)

PRESENT

a

b

a+b

ABSENT

c

d

c+d

a+c

b +d

a+b+c+d

ANALYSIS 1. IN A SIMILAR MANNER BOTH FOR THE STUDY COHORT AND COMPARISON COHORT 2. DUE CONSIDERATION TO • THE LOSS ON FOLLOW-UP • THE SECONDARY INFORMATION OBTAINED FROM OTHER SOURCES (FAMILY MEMBERS AND NEIGHBORS) WITH REGARD TO ATTRITION

COMPARISON COHORT THE COMPARISON CAN BE DONE • INTERNALLY, • EXTERNALLY OR • WITH GENERAL POPULATION RATES.

INCIDENCE RATES • INCIDENCE AMONG THE EXPOSED (NEW CASES AMONG THE SMOKERS) = (A/ A+B)

• INCIDENCE AMONG THE NON-EXPOSED (NEW CASES AMONG THE NONSMOKERS) = (C/ C+D )

RISKS • EXCESS RISK (ER)

=

(A /A+B) – (C / C+D)

• RELATIVE RISK OR RISK RATIO (RR) = [A / (A+B)] / [C / C +D)] (incidence in the exposed (smokers) / incidence in the non-exposed (nonsmokers)

• Attributable risk (AR = excess risk/ incidence among the exposed (AETIOLOGICAL FRACTION) • Population attributable risk (PAR) = incidence in the total population minus incidence among the non-exposed.

METHODOLOGY SPECIAL GROUPS

STUDY COHORT (EXPOSED GROUP)

GENERAL POPULATION COMPARISON COHORT (NOT EXPOSED) INTERNAL OR EXTERNAL COMPARISON OR WITH THE GENERAL POPULATION RATES

FOLLOW UP FOR INCIDENCE OF NEW CASES INCIDENCE & RISK MEASUREMENT IN BOTH GROUPS RR, EXCESS RISK, AR AND PAR

PROSPECTIVE OR FORWARD LOOKING STUDY

FORWARD SUSPECTED CAUSE

EFFECT/ DISEASE

FOLLOW UP STUDY • COHORTS ARE FOLLOWED TILL NEW CASES ARE OBSERVED HENCE THE NAME

TYPES 1. CURRENT COHORT OR PRESENT COHORT (USUAL COHORT) 2. HISTORICAL COHORT (STUDY COMMENCED AT A POINT OF TIME IN THE PAST UNTO THE PRESENT) 3. RETROSPECTIVE – PROSPECTIVE COHORT (COMMENCED IN THE PAST AND EXTENDED INTO THE FUTURE)

STEPS IN CONDUCTING COHORT STUDY • • • •

FORMATION OF STUDY AND COMPARISON COHORTS COLLECTING BASELINE INFORMATION FOLLOW-UP ANALYSIS FOR INCIDENCE AND RISK MEASUREMENT IN BOTH THE GROUPS AND COMPARISON

FORMATION OF STUDY COHORT EXCLUDE FROM THE STUDY • DISEASED PERSONS • EXPOSED PERSONS

EXPOSURE CONSIDERATIONS

• HOMOGENEOUS COHORT • HETEROGENEOUS COHORT

SOURCES OF STUDY COHORT 1. GENERAL POPULATION 2. SOME SELECTED GROUPS LIKE DOCTORS, LAWYERS, etc 3. SPECIAL EXPOSURE GROUPS LIKE ASBESTOS WORKERS, MINERS, etc.

INTERNAL COMPARISON DIFFERENT SUBGROUPS OF SAME HETEROGENEOUS COHORT ARE COMPARED

EXTERNAL COMPARISON • STUDY COHORT COMPARED WITH A SEPARATE NONEXPOSED COHORT

COMPARISON WITH THE GENERAL POPULATION RATES

IDEAL COHORT WHILE SELECTING THE COHORT, • • • •

A STABLE COHORT, A COOPERATIVE COHORT, A COMMITTED COHORT AND A WELL-INFORMED COHORT

IS TO BE USUALLY SELECTED.

COHORT SELECTION TECHNIQUES

• SPECIAL GROUPS, • EXPOSURE GROUPS, • HOSPITAL REGISTERS, • MEDICAL RECORDS, AND •DEATH CERTIFICATES, ETC.

COHORT SELECTION TECHNIQUES • SELECT BOTH COHORTS FROM SAME POPULATION • SELECT THE EXPOSED GROUPS • MEASURE THE DEGREE AND DURATION OF EXPOSURE • DO FOLLOW UP FOR INCIDENCE OF OUTCOMES.

COLLECTION OF BASE LINE INFORMATION • COLLECTED IN A SIMILAR FASHION FROM BOTH THE GROUPS • SERVES AS A BENCHMARK FOR FUTURE ANALYSIS • TIME, DURATION AND EXTENT OF COLLECTION OF DATA TO BE CONSIDERED.

DATA COLLECTION TECHNIQUES • THE PERIOD OF RECORDING, QUALITY AND APPLICABILITY OF RECORDS • DOSE-RESPONSE RELATIONSHIP

DATA COLLECTION TECHNIQUES-2 • SOURCES OF GOOD OR HIGH QUALITY • QUESTIONS OF EITHER DIRECT OR INDIRECT, EQUAL TIME AND SAME QUESTIONS (USUALLY OPEN-ENDED) • INTERVIEWER BIAS

ATTRITION REDUCTION • BY TAKING THE INFORMED CONSENT • BY ENLISTING THE COMMITMENT TO CONTINUE AND COOPERATE IN THE STUDY • BY TRACING THE LOST ONES AND INCLUDE THEM • BY CONSIDERING THE INFORMATION OF THE LOST PERSONS AT THE TIME OF ANALYSIS • BY KEEPING THE NON-RESPONSE AT A LOW LEVEL TO IMPROVE THE VALIDITY

FOLLOW UP

• WELLDEFINED AND STANDARDIZED END POINTS FOR PROPER FOLLOW UP • SIMILAR END POINTS FOR BOTH COHORTS • DURATION OF FOLLOW UP

FOLLOW-UP TECHNIQUES 1. PERIODICAL MEDICAL EXAMINATIONS AND MAILED QUESTIONNAIRES 3. DIRECT PERSONAL INTERVIEWS OR EXAMINATIONS 5. VIDEOCONFERENCE, NEIGHBORS, FRIENDS AND RELATIVES, ETC 7. LOST PERSONS CAN BE TRACED THROUGH THE LETTERS, FROM THEIR RELATIVES AND FRIENDS 9. MIGRATED COHORT SUBJECTS CAN ALSO BE TRACED THROUGH TRAVEL AND IMMIGRATION AUTHORITIES • DEAD PERSONS - LOCAL OR REGIONAL MORTALITY REGISTERS OR DEATH CERTIFICATES

PROBLEMS FOR FOLLOW UP • THE FOLLOW-UP OF LARGE GROUP (USUAL IN COHORT STUDY) • RESOURCE CRUNCH • TIME SCARCITY • PAUCITY OF TRAINED PERSONNEL • ATTRITION, LOSS ON FOLLOW UP

EXAMPLES 1.HISTORICAL COHORT STUDY ON ARTIFICIAL MENOPAUSE AND BREAST CANCER IN BOSTON AREA (1940) 3.PROSPECTIVE STUDY ON SMOKING AND MORTALITY IN BRITISH DOCTORS (1951) 5. FRAMINGHAM HEART STUDY (1951)

ARTIFICIAL MENOPAUSE AND BREAST CANCER IN BOSTON AREA (1940) 1ST STEP - FORMATION OF COHORTS • WOMEN AGED 55 YEARS AND

YOUNGER AND TREATED IN TWO BOSTON HOSPITALS BETWEEN 1920 AND 1940 ARE SELECTED FOR THE STUDY. DETAILS WERE COLLECTED FROM THE PATHOLOGICAL RECORDS AND DEATH CERTIFICATES.

ARTIFICIAL MENOPAUSE AND BREAST CANCER IN BOSTON AREA (1940) II STEP - COLLECTING INFORMATION • FROM THE AVAILABLE SURGICAL & PATHOLOGICAL RECORDS OF HIGH QUALITY • INFORMATION IS ALSO AVAILABLE ABOUT THE CO-VARIABLES

PROSPECTIVE STUDY ON SMOKING AND MORTALITY IN BRITISH DOCTORS (1951) 1ST STEP - FORMATION OF COHORTS

40637 BRITISH DOCTORS WERE SELECTED FROM BRITISH MEDICAL REGISTER.

PROSPECTIVE STUDY ON SMOKING AND MORTALITY IN BRITISH DOCTORS (1951) II STEP- COLLECTING INFORMATION • MAILED QUESTIONNAIRES WERE USED FOR OBTAINING INFORMATION. • ADDITIONAL QUESTIONNAIRES WERE SENT TO KNOW MORE ABOUT A) SMOKING, AGE WHEN SMOKING STARTED, NATURE, DURATION, CURRENT STATUS B) IF STOPPED, THE

FRAMINGHAM HEART STUDY (1951) • 1ST STEP - FORMATION OF COHORTS • TOTALLY, 5209 INDIVIDUALS WERE REGISTERED FOR THE STUDY.

FRAMINGHAM HEART STUDY (1951) II STEP - COLLECTING INFORMATION • BY INITIAL EXAMINATION AND DETAILED INTERVIEWS, THE DATA WAS GATHERED. • 82 PERSONS WHO HAD CARDIOVASCULAR EVENT AT THE INITIAL EXAMINATION WERE EXCLUDED OUT OF 5209 SUBJECTS. • THE REMAINING 5127 SUBJECTS WERE CLASSIFIED ACCORDING TO THE RISK FACTORS LIKE HIGH SERUM CHOLESTEROL, SMOKING, HYPERTENSION, BODY MASS INDEX AND THE PRESENCE OF OTHER DISEASES.

HURDLES IN COHORT STUDIES • FOLLOW UP • ATTRITION • ETHICAL PROBLEMS

SUMMARY THOUGH COHORT STUDY IS • TIME CONSUMING • EXPENSIVE • LESS BIASED IT IS BETTER ANALYTICAL METHOD TO PROVE CAUSAL HYPOTHESIS AND DIRECT RISK MEASURE MENT • •

REFERENCES • Brian Mac Mahan - Epidemiology -principles & methods • Roger Detels, James Mc Even -Oxford Text Book of Public Health • Maxcy-Rosenau-Last, Public Health & Preventive medicine