SLEEP EPIDEMIOLOGY Flashcards
REM SLEEP DISCOVERY?
REM sleep was discovered by Aserinsky and Kleitman in 1953
2017 NOBEL PRIZE IN MEDICINE?
Jeffrey C. Hall, Michael Rosbash and Michael W. Young
for their discoveries of molecular mechanisms controlling the circadian rhythm
BMI AND SLEEP DEPRIVATION IN US ADULTS OVER TIME
BMI RISING, AVERAGE SLEEP DURATION GOING DOWN (LESS THAN 7 HRS PER NIGHT)
THESE FACTS ALONE ARE NOT ENOUGH TO PROVE ASSOCIATION/ PROVE THAT THE 2 ARE CAUSED BY ONE ANOTHER –> WOULD LEAD TO ECOLOGICAL FALLACY, DIFFERENT DATA NEEDED (NOT JUST CROSS SECTIONAL STUDIES!)
ECOLOGICAL FALLACY?
An ecological fallacy (also ecological inference fallacy or population fallacy) is a formal fallacy in the interpretation of statistical data that occurs when inferences about the nature of individuals are deduced from inferences about the group to which those individuals belong.
ecological fallacy, in epidemiology, failure in reasoning that arises when an inference is made about an individual based on aggregate data for a group
UK AVERAGE SLEEP DURATION?
7 HRS
SLEEP DURATION AND HYPERTENSION?
- SHORT DURATION OF SLEEP HAS AN EFFECT ON BLOOD PRESURE VERY EARLY IN LIFE
- EFFECT ON INCIDENCE OF HYPERTENSION MAINLY ASSOCIATED WITH A REDUCTION IN SLOW WAVE SLEEP
- SLEEP DURATION IMPROVES BP (REVERSIBLE EFFECT)
SHRT DURATION OF SLEEP AND RISK OF ‘INCIDENT’ CHRONIC DISEASE
- RELATIVE RISK >1 FOR MANY CHRONIC DISEASES
- ESP HIGH FOR OBESITY (FOLLOWED BY CORONARY HEART DISEASE, T2D, HYPERTENSION)
ASSOCIATION VS CAUSALITY
- DON’T MEAN THE SAME
- ASSOCIATION MAY BE PRESENT N THE ABSCENCE OF A TRUE CAUSE-EFFECT RELATIONSHIP (DUE TO EPIDEMIOLOGIAL STUDY DESIGNS, SYSTEMATIC AND RANDOM VARIATION, ALTERNATIVE ASSOCIATIONS..)
- THERE ARE SIGNIFICANT LIMITATIONS OF OBSERVATIONAL RESEARCH IN ESTABLISHING A CAUSAL LINK BETWEEN A BEHAVIOURAL EXPOSURE AND HEALTH OUTCOMES
INTERPRETATION OF CAUSALITY; THE ROLE OF ‘RULING OUT’?
RULE OUT:
CHANCE (‘LUCK’)
- MEASURED BY P VALUE (PROVING STATISTICAL SIGNIFICANCE)
- 95% CIs
- CAN USE ‘ERROR FACTOR’ FOR THE CALCULATION (POISSON DISTRIBUTION)
- CAN USE ‘STANDARD DEVIATION’ FOR THE CALCULATION (GAUSSIAN DISTRIBUTION)
BIAS (‘INCORRECT’)
- SELECTION BIAS (NOT REPRESENTATIVE POPULATION)
- INFORMATION BIAS (DIFFERENTIAL RECALL, OBSERVATION, MEASUREMENT, CLASSIFICATION..)
CONFOUNDING (‘ERRONEOUS’), SOMETHING ELSE THAT MIGHT EXPLAIN APPARENT CAUSALITY
- KNOWN FACTORS (E.G. AGE, SEX, ETHNICITY)
- POSSIBLE FACTORS (E.G. SES, DEPRIVATION)
- UNKNOWN FACTORS (E.G. GENETICS)
BRADFORD HILL?
- SIR AUSTIN BRADFORD HILL, ENGLISH EPIDEMIOLOGIST AND STATISTICIAN
- PIONEERED THE RANDOMISED CONTROL TRIAL
- WITH SIR RICHARD DOLL, DEMONSTRATED THE CONNECTION BETWEEN CIGARETTE SMOKING AND LUNG CANCER
- BRADFORD HILL CRITERIA FOR ‘CAUSALITY’ –> ACKNOWLEDGE THE DIFFERENCE BETWEEN CORRELATION AND CAUSATION –> SET OF CRITERIA THAT ARE MEANT TO EVALUATE THE STRENGTH OF EVIDENCE IN FAVOUR OF A CAUSE-EFFECT RELATIONSHIP
BRADFOR HILL CRITERIA FOR INFERRING CAUSATION?
- STRENGTH
- TEMPORALITY
- GRADIENT (DOSE-RESPONSE)
- REVERSIBILITY
- CONSISTENCY
- PLAUSIBILITY
- SPECIFICITY
- ANALOGY
- COHERENCE
(NO PARTICULAR RULE AS TO HOW MANY NEED TO BE MET, THE MORE CRITERIA YOU MEET THE RELATIONSHIP IS MORE LIKELY TO BE CAUSATIVE)
WHAT ARE SOME REASONS SUPPORTING THAT THE ASSOCIATION BETWEEN SHORT DURATION OF SLEEP AND CARDIO-METABOLIC RISK FACTORS MAY REFLECT CAUSALITY?
- THE EFFECTS ARE STRONG (LARGE RELATIVE RISK)
- EFFECTS ARE CONSISTENT (CONFIRMED IN DIFFERENT POPULATIONS FOR SEVERAL END-POINTS)
- EFFECTS SHOW A TEMPORAL SEQUENCE (SHORT SLEEP PRECEDING END-POINTS)
- DOSE-RESPONSE (CONSISTENT TRESHOLD EFFECTS)
- BIOLOGICAL PLAUSIBILITY ( SEVERAL; POTENTIAL MECHANISMS INVOLVED; GENETIC, MOLECULAR, CELLULAR, PHYSIOLOGICAL..)
- REVERSIBILITY (CONFIRMED WHEN TESTED IN CONTROLLED TRIAL CONDITIONS, AT LEAST SHORT-TERM)
