Workshop on Evaluating and Appraising a Diagnostic Test in the Health Related Sciences with Statistical Software Demonstration


Diagnosis is an essential part of clinical practice and patient management.  Very often we want to improve our methods of doing so; new diagnostic tests are being developed at a fast rate and the technology of existing tests is continuously being improved; a new instrument may have come out in the market; a new kind of test is now made available; a new test has been proposed as a diagnostic criteria; how good are we at classifying a patient correctly according to a blood test result; how good is the test at picking those correctly diagnosed; how precise or reproducible is the test; how to choose the best `cut-off’ level for a diagnostic test; how good is the inter-rater agreement?   These are the issues the workshop hopes to deal with. Besides, it will open the scope for a student in need of a research project. It will be informal with the topics flitting between lectures and practical demonstrations.  

Learning outcome:

At the end of the workshop, the candidate will be able to:

  • Get an overview on the relevance and place of diagnostic tests in clinical management
  • Appreciate the importance of diagnostic tests as an ancillary aid in coming to a diagnosis
  • Be aware of the types of diagnostic tests commonly used in clinical practice and their interpretation
  • Appreciate the limitations of diagnostic tests and when to exercise caution in their interpretation

Who should attend:

Doctors, radiologists, biomedical students, pharmacists, clinical scientists and chemical pathologists or anyone who is involved in evaluating diagnostic tests or research in the area of diagnostic medicine and clinical epidemiology.

Workshop Program:

  • The basics: Measures of central tendency and variation. (Mean and median. Variance, standard deviation, standard error, p value and 95% confidence intervals)

  • Precision and repeatability: Does the test measure precisely what it is supposed to do? (Coefficient of variation/ repeatability/ Intra and inter-class correlation coefficient)

  • The validity of a test: Introducing sensitivity, specificity and positive predictive value

  • How good is test in predicting the disease and where is the best cut-off level. (When do we `diagnose in’ and `diagnose out ‘the disease? / Understanding the receiver operator characteristics (ROC) curve)

  • How good is the concordance between the new and the established test? (If the new test is as good and cheaper why not use it? / Understanding the Bland Altman plot)

  • How good is the inter-rater agreement between two or more clinicians in coming to a diagnosis: the kappa statistic


Professor Kulenthran Arumugam and his expertise:

Professor Kulenthran Arumugam has been a clinician for more than 40 years. He is also a distinguish researcher having the rare distinction of holding two Doctoral thesis, an MD and a PhD, both from the University of Malaya.  He is a trained Epidemiologist from the London School of Hygiene, University of London.  His interest and specialty is in clinical epidemiology.  In addition, he has a Certification in Advanced Epidemiological Analysis from the London School of Hygiene, Practical Statistics for Medical Research from University College London and in a Certification in the Conduct of Clinical Trials, University of Bristol. He has run a number of workshops in Clinical Epidemiology in particular, in the area of evaluating a diagnostic test. He is well versed with the STATA Statistical software and in medical research.