STRATEGIES IN EPIDEMIOLOGY: Measurements and Measurement Errors; Assessing Evidence of Disease Causation Flashcards
A number or label assigned to empirical properties of a variable according to rules
MEASUREMENT
numerals that have quantitative meaning
NUMBERS
attributes that have qualitative meaning
LABELS
Categorizing each subject into 2 or more mutually exclusive groupS
CLASSIFICATION (LABELS)
examples of CLASSIFICATION (LABELS)
- NUTRITIONAL STATUS
- SEVERITY OF PAIN
- DISEASE STATUS
The fewer / smaller the errors, the better the measurements
QUALITY OF MEASUREMENTS
errors of QUALITY OF MEASUREMENTS
- MISSCLASSIFICATION
- DEVIATION
Pre-requisite for making measurements
OPERATIONAL DEFINITION
example of OPERATIONAL DEFINITION
- WEIGHT
[OPERATIONAL DEFINITION]
measurement of gravitational force acting on an object
CONTEXTUAL
[OPERATIONAL DEFINITION]
result of an object on a Newton spring scale
OPERATIONAL
give the FOUR ‘SOURCES OF ERROR’
- OBSERVER: examiner, interviewer
- SYSTEM: coding and classifying systems
- SUBJECTS
- INSTRUMENT
[SOURCES OF ERROR]
- Differences or changes in the diagnostic criteria used by most clinicians
- Differences or changes in the application of diagnostic criteria by individual clinicians
- Prior knowledge
OBSERVER: examiner, interviewer
[SOURCES OF ERROR]
- Defects or changes in Classification of diseases / causes of death Coding of diseases / causes of death
SYSTEM: coding and classifying systems
[SOURCES OF ERROR]
- Behavioral
- Interactive responses
- Biologic variability
SUBJECTS
[SOURCES OF ERROR]
‘SUBJECTS’
- Recall problems
- Unwillingness to disclose information
BEHAVIORAL
[SOURCES OF ERROR]
‘SUBJECTS’
- Response modified by the behavior of interviewer
- Response modified by knowledge that one is being observed or studies
INTERACTIVE RESPONSES
[SOURCES OF ERROR]
‘SUBJECTS’
- Random or short-term fluctuations in some biological factors
BIOLOGIC VARIABILITY
[SOURCES OF ERROR]
- Equipment / mechanical instrument
- Single index instrument
- Abstraction form
- Observation Checklist
INSTRUMENT
[SOURCES OF ERROR]
‘INSTRUMENT’
- Analytic or scaling problems of combining information from 2 or more items to form an overall index or indicator of the factor/disease
SINGLE INDEX INSTRUMENT
[SOURCES OF ERROR]
‘INSTRUMENT’
- not properly labeled
- incomplete / unclear
ABSTRACTION FORM
[SOURCES OF ERROR]
‘INSTRUMENT’
- incomplete / unclear
OBSERVATION CHECKLIST
[ABSTRACT and CONCRETE VARIABLES]
- not measured directly
- not easily defined
- measured by combining the results of 2 or more item scores into single index
ABSTRACT
[ABSTRACT and CONCRETE VARIABLES]
- measured directly
- easily defined
- closely related to observed variables
CONCRETE
examples of INDICATORS OF A VARIBALES
- Single
- Composite
- Proxy indicator
[INDICATORS OF A VARIBALES]
- obvious and has one indicator
SINGLE
[INDICATORS OF A VARIBALES]
- not so obvious and can have multiple indicators
COMPOSITE
human errors which involves PROCESSING, DATA ABSTRACTION, TRANSCRIPTION, IMPROPER USE OF SOFTWARE, SOFTWARE VIRUSES/BUGS
DATA PROCESSING
two examples of CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT
- RELIABILITY
- VALIDITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘RELIABILITY’ examples
- INTER-OBSERVER RELIABILITY
- INTRA-OBSERVER RELIABILITY
- INTERNAL CONSISTENCY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘VALIDITY’ examples
- SENSITIVITY
- SPECIFICITY
- PREDICTIVE VALUE (+)
- PREDICTIVE VALUE (-)
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
- the extent to which the measurements obtained are reproducible or repeatable
RELIABILITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘RELIABILTY’
- Across 2 or more observer
INTER-OBSERVER RELIABILITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘RELIABILITY’
- Within the same person looking at the same data, giving same results
INTRA-OBSERVER RELIABILITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘RELIABILITY’
- Similarity among items of a composite measure
INTERNAL CONSISTENCY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
- the extent to which measurements reflect the true values of the theoretical factors that
the observed variable is supposed to measure
VALIDITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘VALIDITY’
- proportion of people labelled positive by the test among those with the disease
SENSITIVITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘VALIDITY’
- proportion of people labelled negative by the test among those without the disease
SPECIFICITY
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘VALIDITY’
- proportion of people who tested positive who have the disease
PREDICTIVE VALUE (+)
[CRITERIA FOR ASSESSING QUALITY OF MEASUREMENT]
‘VALIDITY’
- proportion of people who tested negative among those without the disease
PREDICTIVE VALUE (-)