Studies indicate there are a variety of contributing factors affecting physician test ordering behaviour. Identifying these behaviours allows development of behaviour-based interventions.Methods Through a pilot study, the list of contributing factors in laboratory tests ordering, and the most ordered tests, were identi ed, and given to 50 medical students, interns, residents and paediatricians in questionnaire form. The results showed routine tests and peer or supervisor pressure as the most in uential factors affecting physician ordering behaviour. An audit and feedback mechanism was selected as an appropriate intervention to improve physician ordering behaviour. The intervention was carried out at two intervals over a three-month period.Findings There was a large reduction in the number of laboratory tests ordered; from 908 before intervention to 389 and 361 after rst and second intervention, respectively. There was a signi cant relationship between audit and feedback and the meaningful reduction of 7 out of 15 laboratory tests including complete blood count (p = 0.002), erythrocyte sedimentation rate (p = 0.01), C-reactive protein (p = 0.01), venous blood gas (p = 0.016), urine analysis (p = 0.005), blood culture (p = 0.045) and stool examination (p = 0.001).Conclusion The audit and feedback intervention, even in short duration, affects physician ordering behaviour. It should be designed in terms of behaviour-based intervention and diagnosis of the contributing factors in physicians' behaviour. Further studies are required to substantiate the effectiveness of such behaviour-based intervention strategies in changing physician behaviour.Keywords: audit and feedback, child, diagnostic tests, laboratory medicine, medical students, physicians, teaching rounds, test ordering, test request, utilisation review Declaration of interests: No confl ict of interests declaredAbstract Introduction