2. Introduction
• In medicine, improving prognosis is the ultimate goal of all diagnostic
and therapeutic decisions.
• Prediction models have been developed to predict the outcome of
patients in many clinical fields.
• Clinical prediction models are also referred to as clinical prediction
rules, prognostic models, or nomograms or prognostic model, or risk
score.
3. Introduction
• A clinical prediction model is developed to calculate estimates of the
probability of the presence/occurrence or future course of a
particular patient outcome from multiple clinical or non-clinical
predictors in order to help individualize diagnostic and therapeutic
decision-making in healthcare practice.
• Framingham risk score: Prediction model used to estimate the
probability of the occurrence of a cardiovascular disease in an
individual within 10 years.
4. Development of Prediction Model
• Adequate sample size
• Statistical models for prediction :
• regression, classification, and neural networks.
5. Calibration
• Calibration is related to goodness-of-fit, which reflects the agreement
between observed outcomes and predictions.
6. Discrimination
• Discrimination refers to the ability of a prediction model to
differentiate between two outcome classes.
• Receiver operating characteristic (ROC) curves.
0 5 0 1 0 0 1 5 0
0
5 0
1 0 0
1 5 0
R O C c u rv e : R O C o f R O C c u rv e d a ta
1 0 0 % - S p e c ific ity %
S
e
n
s
itiv
it
y
%
7.
8. Prediction Models for Public Health
Targeting of Preventive Interventions : Framingham risk
functions
Prediction models for Clinical Practice
Surgical Decision-Making
9. Schematic representation of surgical decision-making on short-term vs. long-term risk
in replacement of a risky BScc heart valve. Square indicates a decision, circle a chance node.
Predictions (‘p’) are needed for four probabilities: surgical mortality, long-term survival, fracture,
and fatality of fracture