Correction: BMC Med 19, 235 (2021)

https://doi.org/10.1186/s12916-021-02108-z

In our original article [1] the multivariate models outlining the association of cohabitation with breast cancer recurrence and mortality were over-adjusted, as models also included marital status, which is used to define cohabitation. This has now been corrected.

  1. 1

    In Fig. 2, the 5 years and overall adjusted IRRs and CIs of recurrence were incorrect and stated as 0.89 (0.61—1.29) and 0.85 (0.60—1.19) in the published manuscript, respectively. The corrected estimates are 1.21 (0.90—1.63) and 1.09 (0.83—1.43).

  2. 2

    In Fig. 3, the 5 years and overall adjusted IRRs and CIs of mortality were incorrectly stated as 1.03 (0.67—1.60) and 1.02 (0.70—1.47), respectively, instead of 1.55 (1.10—2.20) and 1.41 (1.05—1.88).

The correct figures are given ahead.

Fig. 2
figure 1

Incidence rates and incidence rate ratios of breast cancer recurrence by socioeconomic position. Plots illustrate 5-year adjusted IRRs and error bars 95% CIs.a Marital status was adjusted for age; cohabitation for age; income for age, comorbidities, marital status, cohabitation, education, and employment; education for age; employment for age, comorbidities, and education. Abbreviations: CI, Confidence Interval; IR, Incidence Rate; IRR, Incidence Rate ratio; N, Numbers; PY, Person-years

Fig. 3
figure 2

Incidence rates and incidence rate ratios of mortality by socioeconomic position. Plots illustrate 5-year adjusted IRRs and error bars 95% CIs.a Marital status was adjusted for age; cohabitation for age; income for age, comorbidities, marital status, cohabitation, education, and employment; education for age; employment for age, comorbidities, and education. Abbreviations: CI, Confidence Interval; IR, Incidence Rate; IRR, Incidence Rate ratio; N, Numbers; PY, Person-years

We regret the error.