Multilevel analysis showing the specific associations between individual and neighborhood characteristics and mammography non-attendance during 2005–2009 among women aged 48–75 years residing in Malmö at the time of invitation (n = 29,901).

a According to a logistic regression modeling non-attendance and including as predictor variables women’s individual number of screening invitations, year and season of screening invitation, age, marital status, number of children, education, income, employment, country of birth, country of citizenship, time lived in Sweden, and history of migration abroad (for categories see Table 2 and/or Appendix?).b Based on proportions of the population in the neighborhoods that had low education, low income, were unemployed, were not born in Sweden and had not resided in the same neighborhood during 2001–2005.In “a” and “b” we show only the ORs for the 5th and 10th decile groups using the 1st decile group as reference.OR = odds ratio; CI = confidence interval; PCV = proportional change of the neighborhood variance; ICC = intraclass correlation at the neighborhood-level; MOR = Median Odds Ratio; AU-ROC = area under the ROC-curve (receiver operating characteristics); BDIC = Bayesian deviance information criterion.Values are given as odds ratios (ORs) and 95% confidence intervals (CIs).