IADR Abstract Archives
The Benefits of Using Inverse Probability Weighting in Orofacial Pain Epidemiology
Objectives: Prevalence studies on orofacial pain are important not only for our understanding of the normal patterns of development and maintenance of orofacial pain conditions but also for allocation of resources in dentistry and healthcare. In that context, the emerging possibilities for register-based studies on community samples are an asset, but potential selection bias should be addressed. Therefore, our aim was to use Inverse Probability Weighting (IWP), taking into account the probability of being included in a cohort, to assess the representativeness of an orofacial pain cohort.
Methods: This study was conducted among adults in Västerbotten, Northern Sweden. Self-reported data on orofacial pain was collected at the Public Dental Health services between 2010 and 2017. Self-reported data on sociodemographic factors and general health status were retrieved from questionnaires within a community intervention programme (Västerbotten Intervention Programme). IWP was used to adjust the observed prevalence by weighting individuals in the sample inversely according to the probability of being selected. We also calculated the observed prevalence and the adjusted prevalence, considering missing data.
Results: In total, 52,117 individuals (50.6% women, mean age 52.9 years (SD 11.0)) were included in the analysis. In 2010, the observed prevalence of orofacial pain was 5.3% (95% CI: 4.8 - 5.7), and, albeit not significant, the weighted prevalence was lower 4.9% (95% CI: 4.6 - 5.3). This pattern was consistent during the study period.
Conclusions: The small difference between the observed and the weighted prevalence ensures the representativeness of this cohort. These findings reinforce that using Inverse Probability Weighting in epidemiological studies provides scientific support for generalizability of results.
Methods: This study was conducted among adults in Västerbotten, Northern Sweden. Self-reported data on orofacial pain was collected at the Public Dental Health services between 2010 and 2017. Self-reported data on sociodemographic factors and general health status were retrieved from questionnaires within a community intervention programme (Västerbotten Intervention Programme). IWP was used to adjust the observed prevalence by weighting individuals in the sample inversely according to the probability of being selected. We also calculated the observed prevalence and the adjusted prevalence, considering missing data.
Results: In total, 52,117 individuals (50.6% women, mean age 52.9 years (SD 11.0)) were included in the analysis. In 2010, the observed prevalence of orofacial pain was 5.3% (95% CI: 4.8 - 5.7), and, albeit not significant, the weighted prevalence was lower 4.9% (95% CI: 4.6 - 5.3). This pattern was consistent during the study period.
Conclusions: The small difference between the observed and the weighted prevalence ensures the representativeness of this cohort. These findings reinforce that using Inverse Probability Weighting in epidemiological studies provides scientific support for generalizability of results.