[Intercensal reconstruction of population and descriptive epidemiological measures in Italy: what is the impact on the cancer incidence rates?].

Objectives: to test the effect on cancer incidence rates when using precensal computation (computed population) or intercensal reconstruction of population (reconstructed population).

Design: comparison between computed and reconstructed population by area and period in 2002-2011; evaluation of the effect on cancer rates using Italian cancer registries data. Setting e participants: population data from the Italian National Institute for Statistics, cancer data from Italian cancer registries; specific analysis involves data from Parma (Emilia-Romagna Region, Northern Italy) and Taranto (Apulia Region, Southern Italy) cancer registries.

Main Outcome Measures: ratio between computed and reconstructed population by area, gender, age, and period; ratio between corresponding age-standardized incidence rates.

Results: Italian population estimates by precensal computation for years 2002-2011 was generally higher than that obtained by intercensal reconstruction especially in 2011, when this has been found in more than 86% of Italian Municipalities. In the same year a smaller proportion of Municipalities (11%) showed an inverse population ratio. Among the most populated Municipalities, the City of Milan showed the higher precensal to intercensal population ratio (1.076), while the City of Taranto showed the lower precensal to intercensal population ratio (0.956). The ratios between age standardized rates obtained with precensal population to those obtained with intercensal population show similar differences; in particular, for all cancer in males and females they were, respectively, 0.985 and 0.982 in the Province of Parma, 0.974 and 0.968 in the City of Parma, 1.023 and 1.013 in the Province of Taranto, and 1.08 and 1.051 in the City of Taranto.

Conclusion: using precensal population as denominator for the year 2002- 2011 produces a remarkable distortion of both temporal trend and geographical comparisons. It is, therefore, necessary that researchers take into account this possible distortion when reporting descriptive measures in the years between the last two censuses in Italy.