Authors:
M Mazloum, HD Bailey, T Heiden, BK Armstrong, N de Klerk, E Milne
Authors notes:
Paediatric and Perinatal Epidemiology. doi: 10.1111/j.1365-3016.2011.01253.x
Keywords:
participation, population-based case-control studies, does the observed decline vary by socio-economic status, random digit dialling, bias, socio-economic status, control recruitment, time trend
Abstract
An Australian study of childhood leukaemia (Aus-ALL) previously reported that Control participation was positively associated with socio-economic status (SES).
A similar study of childhood brain tumours (Aus-CBT) was done four years later, and this paper compares control participation and its relationship with SES in the two studies.
To assess the representativeness of controls in terms of SES, the addresses of controls were linked to Australian Bureau of Statistics Census 2006 Collection Districts (CDs), and hence to area-based indices of SES.
Independent sample T-tests and chi-square tests were used to compare the SES indices of CDs where Aus-CBT controls lived with those where Aus-ALL controls lived and with those of all CDs where Australian families lived.
The overall percentage of eligible families who agreed to participate was lower in Aus-CBT, (53.9%) than in Aus-ALL (70.3%). Control families in both studies were of higher SES than the general population, while the distribution of SES among recruited controls was similar in both studies.
These findings provide some reassurance that the observed decline in research participation over time may not be associated with an increasingly unrepresentative participant population.