The interventions were conducted in Kitgum Matidi Sub County in Kitgum District, Northern Uganda. Kitgum Matidi (HC) III serves as the highest level health facility providing health care services in Kitgum Matidi Sub County. The facility serves an estimated population of about 15,000 people. The health facility provides a range of preventive and curative health services including: ANC, skilled attended delivery, other outpatient and inpatient services.

The project interventions focused on strengthening both the community and health facility systems in order to increase demand and quality of services, and ultimately utilization of ANC and skilled delivery services at the health facility. The project interventions were implemented between January, 2010 and September, 2011. The interventions mainly targeted pregnant women and their male partners. Details of the program interventions are described below. The key components of the interventions are also summarized in Table 1.

All data which also included the estimated (expected) number of pregnant women in the first trimester were abstracted from the Health Information Management System (HIMS) data collection and reporting tools as well as registers of MOH that are used routinely by the health workers to capture patient care and treatment data. These included: the ANC register, delivery and birth register as well as quarterly and annual HMIS reports. Based on the population demographics, the health facility with technical support from the district health office and MoH set the annual targets on expected pregnancies. These facility level targets contributed to the district and national annual targets [24,25]. The data that was used for this paper was collected over a period of 21 months (from January, 2010 to September, 2011); during implementation of interventions. In order to evaluate the outcomes of the project, data for the same indicators was collected for the year preceding the interventions (January to December, 2009). The data was abstracted from the registers using a standardized tool and entered into an excel sheet for storage and analysis.

Data was analyzed using Excel and Epi Info soft ware. Although data was collected on a monthly basis, during analysis data was collapsed into quarters and years (annually). The variables of interest included: total number of pregnant women attending first ANC visit and this was compared with the estimated annual number of pregnant women in the first trimester who were expected to attend ANC. Other measures included; total number of women attending the fourth ANC, number of pregnant women attending the first ANC visit who were counseled, tested for HIV and received their results (this was compared with the total number of pregnant women attending the first ANC visit rather than the total number of pregnancies in the community), number of male partners of pregnant women attending the first ANC and counseled, tested for HIV and received their results together with their wives (this was compared with number of pregnant women attending the first ANC visit who were counseled, tested for HIV and received their results);first-time ANC pregnant women were expected to be counseled, tested for HIV and given their HIV status results together with their husbands, as couples. The other variable was total number of skilled attended (health facility) deliveries recorded in the facility registers. Pregnant women who attended ANC but did not return to deliver in the health facility were considered to have delivered in the community/at home with the help of unskilled personnel. Using EPINFO StatCal option, we performed a Chi square test for trend and generated Odds Ratios (OR), 95%CI and p-Values to check whether the trends in these indicators were statistically significant. The Confidence Intervals (CI) were computed using the usual Chi-square in EPINFO.

This manuscript was based on routine program data analysis for purposes of program monitoring and evaluation. The data that were analyzed had no identifiers for any of the individuals that were served and was exempt from ethical approval.

Generally the total annual first ANC attendance increased from the year 2009 to 2011. In 2009, 56.0% (425/753) of the expected women attended ANC compared to 69.0% (537/778) in 2010 (OR 1.72; 95% CI 1.39 – 2.12; p < 0.001) and 93.7% (562/600) in 2011 (OR 11.41; 95% CI 7.97 – 16.34; p < 0.001). Refer to Table 2. More specifically, in 2009, there was no significant change in the number of mothers attending both first and fourth ANC visits. The numbers however increased from 114 women in the first quarter of 2010 to 202 in the third quarter of 2011. This trend continued steadily for the new ANC attendees up to the third quarter of 2011. A rise in fourth ANC visit was registered in 2010, from 89 to 107 in the first and second quarters respectively of 2010, with a slower rise from the third quarter of 2010 to third quarter of 2011. In all the quarters from 2009 to 2011 fewer women attended 4th ANC visit compared to those who attended 1st ANC visit (Figure 1).

The number of pregnant women who were counselled, tested and received their results for HIV was relatively stable in 2009 (pre-intervention). During this period, 82 (76.6%) pregnant women in the first quarter, 90 (79.6%) in the second quarter, 83 (79.0%) in the third quarter and 78 (78.0%) in the fourth quarter were counselled and tested. However, a steady rise was registered from 92 (80.7%) in the first quarter to 146 (97.3%) in the fourth quarter of 2010 (OR 3.40; 95% CI 2.31 – 4.99; p < 0.001) and then to 201 (99.5%) pregnant women in the third quarter of 2011 (OR 32.94; 95% CI 13.27 – 81.77; p < 0.001) (Figure 2 and Table 3).

The number of male partners undergoing HIV counseling and testing together with their pregnant wives during ANC was computed. In the first two quarters of 2009, male ANC attendance was at zero. The number of male partner ANC attendance steadily rose from 13 (16.7%) in the fourth quarter of 2009 to 130 (89%) in the fourth quarter of 2010 (OR 31.97; 95% CI 20.64 – 49.51; p < 0.001) and then to 180 (89.6%) in the third quarter of 2011 (OR 92.18; 95% 57.43 – 147.96; p < 0.001) (Figure 3 and Table 4).

From the first quarter of 2009 to the second quarter of 2010 the number of pregnant women delivering in the health facility remained relatively stable. The number of pregnant women delivering in the health facility did not change much before distribution of mama kits, at 67 (58.8%) in the first quarter and74 (55.2%) in the second quarter of 2010. However, there was an increase to 100 (71.9%) and 149 (99.3%) in the third and fourth quarters of 2010 respectively. Mama-kits were provided at the end of second quarter of 2010 and lasted up to the fourth quarter of 2010. In the first and second quarter of 2011 after the mama-kit stock had run out, the number of pregnant women delivering in the health facility dropped to 126 (72.4%) and 127 (68.3%) respectively but remained higher than the period before mama-kits were provided. When mama-kits were re-stocked at the beginning of the third quarter of 2011 the number of pregnant women delivering in the health facility increased again, to 173 (85.6%) in the third quarter of 2011 (Figure 4).

ANC attendance provides an opportunity for assessment of women’s health and planning for delivery and is an important level in the improvement of maternal and child health outcomes [6-8]. The rising first ANC attendance in this evaluation is consistent with those in another study conducted in a rural district in Tanzania, which showed increased ANC booking following community engagement [15]. The community engagements in these program interventions also addressed cultural beliefs and barriers which have been shown to influence ANC attendance [26]. In this evaluation, the fourth ANC attendance was however not impacted much suggesting that more efforts are required for communities to appreciate the need for follow-up visits, and facilities to ensure retention of women after the first visit [27].

The uptake of HIV testing by pregnant women rose during the implementation of the interventions and can be attributed to community mobilisation as well as emphasizing testing at first ANC visit. Deferring HIV testing to subsequent visits, which sometimes happens in ANC programs, is a missed opportunity as several women will not return after the first ANC visit, as demonstrated in this evaluation and other studies [28]. Ensuring availability of supplies like test kits at the facility may have also contributed to the increased uptake of HIV testing by women and their partners [29].

These findings show a rapid increase in male partner involvement in ANC and male partner uptake of HIV testing, an area that is very crucial in the delivery of maternal and child health services. This is a very important finding given the continued challenges that programs have faced with promoting male involvement in various services including family planning, ANC, delivery and PMTCT. For example, in Uganda men have been observed to be particularly important in ensuring that their pregnant women attend skilled deliveries [30]. The male partner access clubs and VHTs recruited men who had been engaged in ANC and whose wives had delivered at health facilities, and these may have acted as 'change agents’ [18]. These kind of interventions may also overcome the cultural barriers that influence ANC attendance within communities [22].

Skilled delivery care is crucial in saving the lives of mothers and their infants [31]. Provision of delivery supplies had a significant and fairly sustained impact in terms of increasing health facility deliveries; whether or not these supplies are delivered as mama-kits or other form, ensuring availability of delivery supplies is an important intervention in increasing facility based deliveries. The additional items in the mama kits for the babies may have also acted as an added incentive for facility based deliveries, an area that may require further scrutiny, since it was implemented alongside a myriad of other interventions. Whereas other interventions such as community mobilization and male partner involvement may have contributed, there was a clear link between the availability of delivery kits and facility based deliveries. As documented in various studies, the availability of mama-kits could have addressed the challenges associated with poor families procuring delivery items as is often required in times of stock outs at facilities, [32-34].

We cannot fully explain the sustained facility deliveries after the mama-kits stocks ran out. However, it is possible that availability of mama-kits attracted pregnant women to deliver in the health facility and once they got to the facilities, the women got exposed to other interventions (e.g. health education) that may have influenced the development of positive attitudes towards delivering in the health facility. The interventions aimed at improving quality of health services (e.g. training of providers and availability of other supplies) may have also instilled confidence and sustained the facility deliveries despite the absence of mama kits.

The project team engaged different stakeholders including other non-governmental organizations and the district local government to appreciate the crucial and significant role that mama-kits and other interventions in this project play in promoting ANC attendance by women and their partners, and skilled attended deliveries. This was done with a view of encouraging other stakeholders to take up these interventions and ensure continuity.

This evaluation has several limitations which we acknowledge. We could not completely rule out the contributions by other players in the attainment of the health outcomes described in this paper. However given that ChildFund Uganda was the main implementing agency of these interventions in that region at the time of the evaluation, the results registered can to a great extent be safely attributable to these interventions.

The absence of information on characteristics of the study population and pregnancy history of the mothers as well as male partner variables limits the analysis for the women-specific factors affecting uptake of these services. The authors cannot guarantee the accuracy of estimated/expected annual numbers of pregnant women in the first trimester, which may compromise the accuracy of the proportions of the women reached. Additionally, the data on pregnancy outcomes was not included in the analysis, and the outcomes are limited to the service uptake indicators. Also, this analysis only tracked uptake of HIV testing and not the other services in the ANC package. The exact cost of mama-kits and other interventions were not computed given a number of indirect costs associated their delivery. Finally our paper does not present the number/percentage of pregnant women who delivered at home (under unskilled personnel). However, the evaluation is still important as it demonstrates increased first ANC attendance, facility based deliveries, and male involvement, which are important avenues for increasing all ANC and facility delivery services.