To improve monitoring of the quality of resuscitation after cardiac arrest staff performance in 47 simulated cardiac arrests was assessed. The exercise revealed many unsuspected deficiencies. Nurses did not apply mouth-to-mouth resuscitation, use the hand-held defibrillator, or give supplementary oxygen while awaiting the arrival of the arrest team; the availability of different models of defibrillators led to confusion and delay in defibrillation; suction equipment did not suck hard enough; lack of special protocols for areas such as paediatrics led to medication errors. Remedial measures included in-service education for nurses, purchase of new defibrillators and portable suction machines, and preparation of detailed protocols for paediatric, haemodialysis, and obstetric units. The use of mock arrests helps in the immediate and definitive identification of deficiencies in the quality of cardiac resuscitation. Furthermore information obtained by the use of mock arrests spurs physicians and hospital administrators to correct the deficiencies promptly.