Worldwide, male attendance of skilled ANC and delivery care remains a challenge to safe motherhood. About 210 million women become pregnant each year with 30 million (15%) developing complications, resulting into over half a million maternal deaths 1 . Developing countries account for more than 99% of all maternal deaths; about a half occurring in sub-Saharan Africa, and a third in South Asia 2 3 . There is slow progress towards achieving the fifth Millennium Development Goal (MDG) in developing countries 4 . Uganda continues to have one of the highest maternal and child mortality worldwide, with an estimated Maternal Mortality Ratio (MMR) at 435/100,000 and child mortality at 137/1,000 live births 5 . The Ugandan government has prioritized reproductive health strategies which focus on accelerated reduction in maternal mortality and severe morbidity related to pregnancy and childbirth.

As early as 2001, the World Health Organization (WHO) established proven safe motherhood interventions that are required at household, community and facility levels to enable every pregnant woman to have a safe pregnancy and childbirth, and to provide couples with the best chance of having healthy infants 6 . The strategies include; providing skilled attendants to prevent, detect and manage the major obstetric complications, together with providing equipment, drugs and other supplies 7 8 . A few studies have found that facility interventions providing routine ANC have minimal beneficial effects to maternal health outcomes, creating a rationale for community-based interventions for pregnant women where educational and support programmes have been successful in increasing/improving skilled delivery, breastfeeding practices and postnatal attendance 9 10 11 12 . This reported success may be mainly attributable to spousal involvement, and not necessarily the community as a whole. Even though it is widely recognized that there is limited research on the role of male involvement during pregnancy 13 14 15 , there are promising links to beneficial effects. Indeed some few prospective studies and literature reviews have successfully demonstrated that prenatal male involvement is associated with beneficial health outcomes such as; higher first trimester ANC visits, abstinence from smoking and alcohol consumption 14 16 , and reduction in low birth-weight infants 14 17 . Evaluations of Prevention of Mother-to-Child Transmission of HIV (PMTCT) which is often implemented within the confines of safe motherhood have similarly shown male involvement to positively influence uptake of HIV testing and preventive interventions for vertical and sexual transmissions of HIV 15 18 19 .

Male attendance of skilled ANC is a fairly new field to research in Uganda. The available estimates depict a low attendance averaging 3% in 2006 20 but are based on health facility information systems that monitor male attendance in the PMTCT programme. Gulu district is located in northern Uganda, a region recovering from a 20-year-old insurgency. The civil war forced majority of its population into Internally Displaced Peoples' (IDP) camps. According to Gulu development plan (2007), the district has a maternal mortality ratio (MMR) of 700/100,000; infant mortality rate of 172/1,000 and ANC HIV sero-prevalence of 8.8% 21 . Gulu district MMR is similar to other districts in Northern Uganda with an MMR averaging 650/100,000 but, higher than neighbouring post-conflict affected Adjumani district in West Nile where the MMR is estimated at 550/100,000 21 . In 2008, the district estimate for male attendance of skilled ANC was 18.4% having risen from about 3% in 2006 22 . Of the total 23,408 pregnant women that were new ANC attendees in Gulu district in 2008, male attendance estimated at 18.4% varied by sub-county; 41.6% (1,107/2,746) in Aswa, 41.5% (2,529/6,264) in Omoro and only 5.0% (660/14,398) in the Municipality 22 . The marked rise in male partner ANC attendance is possibly due to a robust district-wide PMTCT programme. This study assessed the level, perceived benefits and factors associated with male partner attendance of skilled ANC in a peri-urban community in Gulu.

In this study, the level of male attendance of skilled ANC was relatively high at 65.4% (60.3 - 70.5) as compared to 42% for Omoro county in Gulu observed from district health management information system (HMIS) reports in 2008 22 and much higher than the 4% national average 20 . This study reports relatively high levels of male attendance of ANC possibly because the sample was composed of peri-urban males who arguably have better access to health facilities where integrated ANC/PMTCT services are provided than the rural that make up the majority of the population. Studies in Sub-Saharan localities where PMTCT programmes are being implemented as part of the ANC services report male attendance rates at 12.5% in northern Tanzania 26 , about 15 and 16% respectively at different clinics in Nairobi, Kenya 15 27 and 25% in Côte d'Ivoire 19 .

HIV screening was perceived as the main benefit to ANC attendance. Although it was not significantly associated with male partner ANC attendance, it is of public health importance because HIV screening is one of the quality services offered at ANC 2 20 . Furthermore HIV testing is meant to benefit the couple so that assessment of the risk of HIV infection to the unborn child is done during pregnancy and prevented. In case of couples that are HIV discordant, male attendance presents the opportunity for health providers to promote use of barrier contraceptive methods such as condoms. The perceived benefit to male partners of monitoring foetal growth is similarly of public health importance, even though it was not found significantly associated with male attendance of ANC. A study conducted in Taiwan reported anxiety of the male partner about the newborn during the wife's pregnancy to increase their participation in maternal health care 28 . Perceived benefit of foetal monitoring during spousal pregnancy is worth inclusion in subsequent studies to determine whether it is a predictor of male attendance of maternal health care.

The main factors identified by the respondents in the survey as barriers to attendance of ANC were lack of transportation, long waiting time and long walking distance to a health facility. Perceiving the walking distance to the health facility as long (greater than one-hour walk or greater than 5 km) was a prominent barrier and associated with reduced male attendance of ANC (p = 0.034). Long distance as a barrier to male attendance of maternal health care has been reported by other cross-sectional studies in Tanzania, Uganda and Bangladesh 29 30 31 . This finding emphasizes the need for the health sector to further increase access to maternal health services.

On the other hand, males that wanted their spouses to carry another pregnancy were found to be associated with reduced attendance of ANC. This finding could suggest that fertility preferences are modified once a man attends ANC with the spouse during pregnancy. This is possible because health education about family planning which is provided during the pre and postnatal visits encourages child spacing for at least two years 8 . We however also found that male partners with high fertility preferences were less likely to have four or more children, implying that they wanted to have more children because they perceived the number they had as few. Given that total fertility rate in Uganda is 6.7% 5 32 , the latter may be the alternative explanation for this finding.

Interestingly, we found that male attendance of ANC was associated with higher likelihood of spouse having skilled delivery at her last childbirth. To the best of our knowledge, this is a new finding; however, some community-level intervention studies found health education and support programmes to increase skilled attendance at childbirth 9 10 11 12 .

Knowledge about safe motherhood services among the male partners was limited. Only 47.1% respondents knew 3 - 5 services offered at ANC. This finding is consistent with other studies in Taiwan, India and Tanzania where male partners were found to have limited knowledge about safe motherhood and sexuality 28 33 34 35 36 . In Uganda, this finding is not surprising because male attendance of safe motherhood is a new approach that was introduced with the national PMTCT program in 2001 20 . Limited knowledge of male partners about maternal health issues has been reported to be a significant barrier to their attendance 34 37 . In this study, male partners that had reasonable knowledge of services offered at ANC (at least three services) were significantly more likely to attend ANC (p < 0.001). This finding supports the evidence that increasing male partner's knowledge of safe motherhood is positively associated with attendance of ANC 38 . However, the study design does not allow us to conclusively affirm the temporal relation between knowledge and ANC attendance. It is also possible that the more one attends ANC, the more knowledgeable one becomes about ANC services.

Obtaining health information from a facility health worker was positively associated with male partner attendance of ANC with their spouses. The finding suggests that males regard health workers as a credible source of health information, and this is influential on their health behaviour. This finding supports evidence from a systematic review by Sternberg et al (2004) that concluded; health promotion for male involvement in sexual and reproductive health programs has a high promise of success 34 . The alternative explanation to this finding is that male attendance of ANC was positively associated with obtaining health information from facility health workers, which would point towards effectiveness of health education at ANC 38 .

We found that male partner's education was not significantly associated with attendance of ANC in the adjusted analysis. The interpretation of this result is that the variable is collinear with knowledge of ANC services, and we therefore omit it from the final model. A cross-sectional study conducted in Nigeria 39 that interviewed pregnant women, who are considered surrogate respondents in this comparison, showed that low education was associated with reduced utilization of ANC similar to other studies 30 31 .

Recall bias could have been present even after restricting inclusion to males whose spouses last delivered 24 months prior to the survey. Two years is a long period, yet pregnancy may not signify as much importance to the male partner as it does to the female. Additionally reporting bias arising from men wanting to provide socially desirable responses especially regarding the level of ANC attendance could have been a potential pitfall to the study. This could have been overcome by cross validation of the men's responses by women interviews.

In this study, we used multi-stage cluster sampling but analyzed subjects at individual level, which could have potentially reduced the statistical efficiency at analysis 40 . We used a conservative design effect of 1.5 to cater for the sampling error, but this may not have been sufficient 40 .

Our study population was largely peri-urban, therefore our findings may not be entirely generalizable to the rest of Gulu district and other rural settings in Uganda. However, we gain insights into how similar populations can be targeted to improve male ANC attendance.

This study design was cross-sectional which limits us from making any causal inferences in relation to the main outcome and independent variables.

Male attendance of skilled ANC of a peri-urban population in Gulu was high at 65.4% compared to 42% in the rest of the district and an estimated 10% attendance rate in the other regions in Uganda. Willingness to attend subsequent skilled ANC was nearly universal. Male partners had favourable perceptions for attendance of skilled ANC that included foetal growth monitoring, HIV screening and identifying complications during pregnancy.

Factors associated with male attendance of ANC were: knowledge of at least three ANC services, obtaining health information from facility health workers, wife last gave birth at a health facility, distance to nearest health facility less than 1 hour or 5 km and the male partner having no intention for the spouse to carry another pregnancy.

Knowledge about ANC services was the main independent variable associated with male attendance of ANC with their spouses. We recommend that governments and local health authorities should develop strategies to empower men (both in communities and at health facilities) with knowledge about ANC. Such strategies may contribute to increasing male attendance in ANC, improve skilled attendance at delivery and improve maternal health outcomes. Strategies that increase male attendance of skilled ANC are likely to be feasible in post-conflict regions and resource limited settings.

The authors declare that they have no competing interests.

RT: Designed the study, coordinated recruitment of participants, collected and entered data, analyzed data and drafted the manuscript. JNS: Designed the study, participated in data analysis and reviewed all drafts of the manuscript. JKL: Designed the study, participated in data analysis, reviewed all drafts of the manuscript. NTM: Guided data analysis, reviewed all drafts of the manuscript. All authors read and approved the final manuscript.