Examination of risk scores to better predict hospital-related harms.

Combining scores or reducing to two-item question-responses did little to change predictive accuracy. This study highlights the limitations of hospital harm predictive scores and emphasises the importance of rigorous testing of predictive scores.