Background This study presents the findings of a nationwide study of acute pesticide poisoning (APP) outcomes, including outcome predictors such as physician and hospital volume and associated factors.

Methods This study of data contained in the Taiwan National Health Insurance Research Database analysed 27 046 patients who had been hospitalised for APP from January 1996 to December 2007. Patient characteristics were then compared among quartiles. The primary outcome measures were length of stay (LOS) and hospital treatment cost. Effect size (ES) was compared among three equally divided periods, and multiple regression models were used to identify outcome predictors.

Results The overall prevalence of APP per 100 000 patients decreased from 12.43 in 1996 to 6.87 in 2007. The LOS for APP treatment was negatively associated with physician volume during the study period. Both LOS and hospital treatment cost were lowest in the high hospital volume subgroup. Comparisons of LOS and hospital treatment cost among the three periods showed that high-volume hospitals and high-volume physicians had better ESs compared to low-volume hospitals and low-volume physicians. Age and number of co-morbidities had significant positive associations with LOS, while admission year, male gender, hospital level, hospital volume and physician volume had significant negative associations with LOS (p<0.05). Hospital treatment cost and hospital level correlated positively with admission year, number of co-morbidities and LOS but correlated negatively with hospital volume and physician volume (p<0.05).

Conclusions In APP patients, treatment by a high-volume physician can reduce LOS and treatment cost.