ANALYTICAL STUDIES AND KEY STATISTICAL CONCEPTS Flashcards
ROLE AND CHARACTERISTICS OF ANALYTICAL STUDIES?
- AN ANALYTICAL STUDY ALLOWS TESTING OF HYPOTHESES AND ESTIMATION OF THE RISK OF A CERTAIN ILLNESS
- THEY ARE ALWAYS OBSERVATIONAL (I.E. NO INTERVENTION IS IMPLEMENTED LIKE IN RANDOMISED CONTROL TRIALS)
WHY ARE EPIDEMIOLOGICAL STUDIES DONE?
- TO UNDERSTAND WHAT HAPPENED
- TO ESTIMATE THE EXTENT OF THE OUTBREAK
- TO IDENTIFY THE POTENTIAL SOURCE (SO IT CAN BE CONTROLLED AND PREVENTED IN THE FUTURE)
- TO CREATE EVIDENCE AND KNOWLEDGE
ROLE OF EPIDEMIOLOGICAL STUDIES (ANALYTICAL EPIDEMIOLOGY)?
THEY CAN GIVE AN IDEA OF HOW LIKELY IT IS THAT AN EXPOSURE IS ASSOCIATED WITH ILLNESS
(THEY CAN TEST THE HYPOTHESIS GENERATED BY DESCRIPTIVE EPIDEMIOLOGY)
WHAT IS RISK IN EPIDEMIOLOGY?
- THE PROBABILITY AN EVENT WILL OCCUR
- THE PROPORTION OF INITIALLY DISEASE FREE INDIVIDUALS WHO DEVELOP DISEASE OVER A DEFINED PERIOD OF OBSERVATION
- DEFINITION OF RISK HERE DOESN’T INCLUDE PERCEPTIONS OF RISK AS E.G. DANGEROUS SITUATIONS (THOSE ARE JUST SUBJECTIVE JUDGEMENTS)
2 MAIN TYPES OF ANALYTICAL STUDIES?
COHORT STUDY AND CASE CONTROL STUDY
WHAT IS A COHORT IN EPIDEMIOLOGICAL TERMS?
A GROUP OF PEOPLE WITH SOMETHING IN COMMON, USUALLY AN EXPOSURE (SAME EXPERIENCE IN THE SAME PERIOD OF TIME)
WHICH ANALYTICAL STUDY DESIGN IS USUALLY USED WHEN OUTBREAK INVESTIGATION IS DONE FOR A DEFINED POPULATION?
COHORT STUDY
WHAT ARE THE OUTPUTS OF COHORT STUDIES, I.E. WHAT CAN BE CALCULATED?
- ATTACK RATE (RISK)
- RISK RATIO/RELATIVE RISK (RR)
EXAMPLES OF WHEN COHORT STUDIES MIGHT BE USED?
- WEDDINGS, BBQs..
- EVENTS WHERE THERE IS A GUEST LIST (IF TOO BIG A COHORT STUDY WON’T BE POSSIBLE, RESOURCES NEED TO BE CONSIDERED)
- EVENTS WHERE IT CAN USUALLY BE IDENTIFIED WHO BECOME ILL, WHERE THE MENU WAS LIMITED ETC
INTERPRETATION OF RELATIVE RISK/RISK RATIO (RR):
RR>1 RISK FACTOR
RR=1 NO ASSOCIATION
RR<1 PROTECTIVE FACTOR
WHAT IS ATTACK RATE AND HOW IS IT CALCULATED?
- AKA RISK
- THE PROPORTION OF PEOPLE WHO BECAME ILL WITH A DISEASE IN A POPULATION INITIALLY FREE OF THE DISEASE
(NUMBER OF CASES/TOTAL NUMBER OF PARTICIPANTS)x100
HOW IS ATTACK RATE (RISK) USED IN COHORT STUDIES?
FOR EACH EXPOSURE OF INTEREST, WHEN CAL CALCULATE ATTACK RATE IN EXPOSED AND IN UNEXPOSED PARTICIPANTS AND USE THEM WHEN DETERMINING THE RR
WHAT IS RELATIVE RISK/RISK RATIO AND HOW IS IT CALCULATED?
A RATIO OF THE EXPOSED AND UNEXPOSED ATTACK RATES
RISK OF DISEASE IN EXPOSED/RISK OF DISEASE IN UNEXPOSED
WHAT DOES A RISK RATIO DETERMINE?
IT DETERMINES IF THE ATTACK RATE IN THE EXPOSED INDIVIDUALS IS HIGHER THAN IN THE UNEXPOSED AND BY HOW MANY TIMES (HELPS DETERMINE RISK FACTORS AND PROTECTIVE FACTORS)
WHEN ARE CASE CONTROL STUDIES USED IN EPIDEMIOLOGY AND HOW ARE PARTICIPANTS SELECTED?
- WHEN THERE’S NO DEFINED POPULATION; E.G. INCREASE IN CASES IN A PARTICULAR GEOGRAPHICAL AREA
- ALL CASES ARE INCLUDED AND THEN CONTROLS ARE SELECTED (CONTROLS NEED TO BE FROM THE SAME POPULATION); BOTH ARE ASKED THE SAME QUESTIONS
WHAT IS THE OUTPUT OF A CASE CONTROL STUDY?
AN ODDS RATIO
WHAT DOES AN ODDS RATIO DETERMINE?
THE DIFFERENCE OF THE FREQUENCY OF EXPOSURE BETWEEN CASES AND CONTROLS
INTERPRETING ODDS RATIO:
OR>1 RISK FACTOR
OR=1 NO ASSOCIATION
OR<1 PROTECTIVE FACTOR
EXAMPLES OF WHEN CASE CONTROL STUDIES ARE USED?
- NATIONALLY DISPERSED OUTBREAKS ASSOCIATED WITH FOOD
- TAKE-AWAYS
- MOST OUTBREAKS
DEFINITION OF THE P VALUE?
THE PROBABILITY OF OBTAINING RESULTS AT LEAST AS EXTREME AS THE OBSERVED RESULTS, ASSUMING THAT THE NULL HYPOTHESIS IS CORRECT
WHAT IS THE NULL HYPOTHESIS IN EPIDEMIOLOGY?
THE HYPOTHESIS STATING THERE IS NO ASSOCIATION BETWEEN THE CASES AND THE EXPOSURE; STATES THERE IS NO DIFFERENCE BETWEEN OBSERVED GROUPS
RANGE OF VALUES THE P VALUE CAN TAKE?
A NUMBER BETWEEN 0 AND 1
WHEN DOES P VALUE INDICATE STATISTICAL SIGNIFICANCE?
WHEN IT IS LOW; USUALLY <0.05 (5%)
INDICATES THAT THE ASSOCIATION IS LIKELY AND WE CAN REJECT THE NULL HYPOTHESIS
WHEN TESTING A HYPOTHESIS, WHAT DOES THE P VALUE TELL US ABOUT THE RESULTS WE GET?
IT HELPS DETERMINE THE STATISTICAL SIGNIFICANCE OF THE RESULTS
WHEN CAN’T YOU REJECT THE NULL HYPOTHESIS?
WHEN p>0.05
WHAT DO P VALUES DEPEND ON?
- SIZE OF THE EFFECT
- SIZE OF THE SAMPLE
HOW DOES SAMPLE SIZE INFLUENCE THE P VALUE AND HOW IS THIS ADDRESSED?
- IF THE SAMPLE IS LARGE ENOUGH, EVEN VERY SMALL DIFFERENCES WILL BE CONSIDERED ‘STATISTICALLY SIGNIFICANT’
- THAT’S WHY P VALUE IS OFTEN ACCOMPANIED BY CONFIDENCE INTERVALS WHICH ARE MORE INFORMATIVE
WHAT ARE CONFIDENCE INTERVALS (IN ANALYTICAL STUDIES)?
RANGES OF POTENTIAL VALUES FOR THE OR & RR AS THEY ARE BOTH ESTIMATES CALCULATED FROM A SAMPLE; CI INDICATE THE PRECISION WITH WHICH THE SAMPLE ESTIMATE IS LIKELY TO REPRESENT THE POPULATION FROM WHICH THAT SAMPLE WAS DRAWN
WHAT DOES A NARROW CONFIDENCE INTERVAL IMPLY?
A LARGE SAMPLE SIZE OR THAT THE RESULTS WERE VERY SIMILAR/PRECISE
INTERPRETING CONFIDENCE INTERVALS FOR OR AND RR:
WHERE A 95% CONFIDENCE INTERVAL INCLUDES 1 THERE IS NO EVIDENCE AT THE LEVEL OF P=0.05 THAT THERE IS A TRUE DIFFERENCE (THAT A HYPOTHESIS ALTERNATIVE TO NULL HYPOTHESIS IS CORRECT)
WHAT IS A CONFOUNDER IN EPIDEMIOLOGY?
SOMETHING ASSOCIATED WITH BOTH EXPOSURE AND THE OUTCOME
WHAT IS MULTIVARIABLE ANALYSIS IN EPIDEMIOLOGY?
- ANALYSIS LOOKING AT ASSOCIATIONS FOR SEVERAL X VARIABLES (EXPOSURES) SIMULTANEOUSLY, BUT ONLY ONE Y (ILLNESS)
. MANY DIFFERENT STATISTICAL TECHNIQUES ARE USED, REGRESSION BEING THE MOST COMMON
WHAT IS THE ROLE OF ADJUSTED ODDS RATION aOR?
IT PRODUCES THE ODDS RATIO FOR EACH INDIVIDUAL VARIABLE ASSUMING THAT THE OTHER VARIABLES WERE HELD CONSTANT
WHAT ARE THE AIMS OF THE MULTIVARIABLE ANALYSIS?
- UNDERSTAND WHICH VARIABLES ARE ASSOCIATED WITH AN OUTCOME
- ACCOUNT FOR THE EFFECT OF THE OTHER VARIABLES WHEN MEASURING ONE VARIABLE
- ADJUST FOR ‘CONFOUNDERS’
WHAT ARE ZOONOSES?
- INFECTIOUS DISEASES WHICH JUMPED FROM A NON-HUMAN ANIMAL TO HUMANS
- THEY SPREAD TO HUMANS THROUGH DIRECT CONTACT OR THROUGH FOOD, WATER OR THE ENVIRONMENT
- CAUSES CAN BE BACTERIA, VIRUSES, PARASITES AND OTHER UNCONVENTIONAL AGENTS LIKE PRIONS
- SOME DISEASES CAN BEGIN AS ZOONOSIS BUT MUTATE TO BECOME HUMAN ONLY (LIKE HIV)
MEANING OF ONE HEALTH APPROACH WHEN ADDRESSING ZOONOSES?
- CONTROL REQUIRES CO-OPERATION OF HUMAN, ENVIRONMENTAL AND ANIMAL HEALTH PARTNERS
- IMPOSSIBLE FOR JUST ONE PERSON/ORGANISATION/SECTOR ALONE TO ADDRESS THE ANIMAL-HUMAN-ENVIRONMENT INTERFACE ADEQUATELY
