Sex-specific intergenerational trends in morbidity burden and multimorbidity status in Hong Kong community: an age-period-cohort analysis of repeated population surveys.

Objectives Prevalence of multimorbidity has been increasing worldwide. While population ageing undoubtedly contributes, secular trends have seldom been decomposed into age, period and cohort effects to investigate intergenerational differences. This study examines the birth cohort effect on morbidity burden and multimorbidity in Hong Kong community. Design Sex-specific age-period-cohort analysis with repeated cross-sectional surveys. Setting A territory-wide population survey database. Participants 69 636 adults aged 35 or above who participated in the surveys in 1999, 2001, 2005 or 2008. Main outcome measures Morbidity burden was operationalised as number of chronic conditions from a list of 14, while multimorbidity was defined as a dichotomous status of whether participants had two or more conditions. Results For both sexes, there was an upward inflection (positive change) of risk of increased morbidity burden starting from cohort 1955-1959. For men born after 1945-1954, there was a trend of lower risk (relative risk= 0.63, 95% CI 0.50 to 0.80 for 1950-1954 vs 1935-1939) which continued through subsequent cohorts but with no further declines. In women, there had been a gradual increase of risk, although only significant for cohort 1970-1974 (relative risk= 1.90, 95% CI 1.08 to 1.34 vs 1935-1939). Similar results were found for dichotomous multimorbidity status. Conclusions The trend of lower risk starting from men born in 1945-1954 may be due to a persistent decline in smoking rates since the 1980s. On the other hand, the childhood obesity epidemic starting from the late 1950s coincided with the observed upward inflection of risk for both sexes, that is, notably more drastic increase of risk in women and the levelling-off of the decline of risk in men. These findings highlight that the cohort effects on morbidity burden and multimorbidity may be sex-specific and contextual. By examining such effects in different world populations, localised sex-specific and generationspecific risk factors can be identified to inform policymaking.