Diagnostics

Question 1

Diagnostic validity

Answer 1

• validly of a diagnosis and corresponding diagnostic tests, assays etc
• relates to sensitivity AND specificity

Question 2

Examples of validity from psychiatry, neurology, clinical psychology

Answer 2

• content validity: refers to symptoms and diagnostic criteria
• concurrent validity: associated correlates or markers, and response to treatment in therapy
• predictive validity: diagnostic stability over time
• discrimination validity: can it discriminate between disorders

Question 3

Diagnostic tests

Answer 3

• predictor variable: test result
• outcome variable: presence or absence of disease
  
  • values can be:
• dichotomous: + or -
• categorical: ++++, +++, ++, + or -
• continuous: milligrams of glucose per litre

Question 4

Observational vs diagnostic studies

Answer 4

• observational studies: association between a predictor and outcome variable
• diagnostic studies: detection/discrimination between + (yes) and - (no)»> more than an association

Question 5

Design

Answer 5

• test phase: what constitutes a positive test result?
  
  • good example of the necessity of pilot studies: crucial in diagnostic studies
• validation phase: assess the tests sensitivity and specificity
  
  • use new sample of subjects

Question 6

Detailed design steps

Answer 6

1. Asses the need for a new diagnostic tool
2. Selection of subjects and samples
3. Identify a reasonable gold standard (for comparison) ie. what will determine a positive result vs a negative result?
4. Design standardized and blinded trials
5. Estimate the sample size (for accuracy)
6. Find the subjects
7. Report accuracy: sensitivity and specificity data

Question 7

Quantitative tools

Answer 7

1. Likelihood ratios: involves the odds
   
   • likelihood that a person with a disease would have a particular test result divided by the likelihood that a person without the disease would have that result
   • very powerful when combined with prior probability of a disease (base rate judgments depend on this): beysian analysis
2. D’, criterion and ROC curves

Question 8

Other important factors

Answer 8

1. Prevalence: how rare is the disease in the population
2. Prior probability: probability that a specific patient has the disease
3. Predictive value or posterior probability of a positive test or negative test

Question 9

Prevalence

Answer 9

• if prevalence is low, the test needs to be specific (need to know that its NOT this disease)
  
  • ex. HIV/AIDS
• if prevalence is high, test needs to be specific (need to know that it IS this disease)
  
  • ex. High blood pressure

Question 10

Prior probability

Answer 10

• base rate
• ex. Prior probability of coronary heart disease in a young, healthy, non-smoker etc: LOW
  
  • prior probability of coronary heart disease in an older, high BMI, smoking etc: HIGH

Question 11

Predictive value (posterior probability)

Answer 11

• predictive value of a positive test (PV+): likelihood of a true positive/(likelihood of a TP + likelihood of FP)
• predictive value of negative test (PV-): likelihood of a TN/ (likelihood of a TN + likelihood of FN)

Question 12

Errors in diagnostic studies

Answer 12

• random error&raquo_space;> chance&raquo_space;> quantify it with CI’s for specificity and sensitivity
• systematic error&raquo_space;> biases
  
  • sampling bias
  • measurement bias
  • reporting bias