What can shopping transactional data reveal about relative prevalence of menstrual pain and period poverty in England?

Abstract

91% of those who menstruate reported to experiencing associated pain. Despite the ubiquity of this phenomenon, the prevalence, extent and sociodemographic variation of menstrual pain remains understudied at national levels. In this paper, we assess the extent and variation of menstrual pain at a national level. We develop a novel proxy measure for menstrual pain, utilising behavioural data extracted from supermarket shopping logs. Baskets where pain and menstrual items co-occur are investigated, and normalized against baskets in which pain or menstrual items occur in isolation. Propensity of menstrual pain purchases are aggregated temporally and geographically across England, prior to linkage with sociodemographic indicators regionally. Findings indicate high prevalence of menstrual pain across England, with 26.7% of customers in our dataset buying pain relief together with menstrual products. People who menstruate are observed to be four times more likely to purchase pain relief with menstrual items than without. Average regional income provides the strongest predictor of menstrual pain co-purchases, with lower income regions exhibiting a 32% lower menstrual-pain purchase than higher income regions. The robust presence of a consistent 28-day cycle in menstrual-pain purchases provides empirical evidence for the use of behavioural proxies for menstrual pain alongside traditional measures. Significant regional differences observed in the prevalence of menstrual-pain transactions across England brings into light existing disparities. Future research into improved understanding of sociodemographic factors associated with menstrual pain could inform strategies to predict and prevent menstrual pain and its adverse impacts.