Investigating the relation between placement of Quit antismoking advertisements and number of telephone calls to Quitline: a semiparametric modelling approach.

STUDY OBJECTIVES: Quitline-an antismoking advertising and a telephone helpline service-is an effective public health intervention strategy for tobacco control. The objective of this short report is to model the relation between placement of antismoking advertisements and calls to Quitline on a given day. METHODS/DESIGN: Data on daily Quitline antismoking advertisements, television target audience rating points (TARPS), and calls to Quitline Victoria were studied for the period 1 August 2000 and 31 July 2001. The outcome-calls to Quitline-is a count and thus assumed to follow a Poisson distribution. Generalised partial linear models were used to model the logarithm of mean daily calls as a non-parametric function of time and a linear parametric function of the day of week, number of advertisements, and TARPS. MAIN RESULTS: Peak calls to Quitline Victoria occurred during Monday to Wednesday with around three times as many calls compared with Sunday. Both placement of Quitline advertisements (p<0.001) and an increase in TARPS (p<0.001) on a given day significantly increased the number of calls made to Quitline Victoria. The model adequately captured fluctuations in call volume and diagnostics showed no model inadequacy. CONCLUSIONS: In this short report the emphasis is on modelling the parametric components-day of week, placement of advertisements, and TARPS on call volume. The dynamics of the underlying time trend in call volume is captured in a non-parametric component. Future analysis of hourly data would provide additional information to assess different media buying strategies that might increase call volume.