Following publication of this article , it has come to our attention that the total number of observations (21,798) has been mistakenly cited as the number of respondents in some paragraphs within the paper. This number appears within the abstract, in the analysis sub-section, within the opening paragraph of the results section, and in the section on HIV prevalence. The number that should have been cited is 11,268 - the total number of respondents in the dataset. 21,798 refers to the total number of observations over the study period. The percentages estimated out of 21,798 have been recalculated. Of the 11,268 individuals enrolled in this study, 81.2% (9,220) were in monogamous marital unions while 18.2% (2,048) were in polygamous marital unions. Of those in polygamous marital unions (n = 2,048), 52.8% were females while 47.2% were males. Thirty eight per cent of the participants (4,236) reported that they had ever received HCT (i.e. individual or couples’ HCT). Overall HIV prevalence was 11.9% (1,337 of 11,268). However, it is important to note that since serial cross-sectional analyses of each of the 4 study visits were used under consideration, the findings shown in Tables 1 and 2 as well as Fig. 2 (A,B,C) are not affected by this error.
In Table 3, percentages are cited showing the number of individuals who were interviewed for at least 3 times. The percentages are shown against their denominators. In citing the numbers, the total number of observations (10,712) was used for those that were interviewed for at least 3 times, instead of 4338 - which is the total number of individuals interviewed for at least 3 times. The corrected version of the table is provided below.