This was the first comprehensive mapping and assessment that determined access to VCT, PMTCT and ART services in a conflict affected Northern Uganda. Given the large population affected by this conflict coupled with high HIV/AIDS prevalence, the results of this study confirms the limited access due to inadequacy and uneven distribution of these services among the districts, counties and camps. Complicating the situation is insecurity, ongoing establishments of new settlements and large number of IDP camps scattered in a vast geographical area in these districts. Among the three districts, access to VCT, PMTCT and ART services was relatively better in Gulu evidenced by both, large number of facilities providing these services and utilizations. The likely reasons might be its comparatively large population size and longer humanitarian crisis that attracted most of relief assistance including provision of HIV related services. Similarly, Gulu as a former headquarters for Acholi sub-region, has a well-developed infrastructure system with relatively more capacity to support humanitarian operations.
The finding that most VCT, PMTCT and ART services were clustered in urban areas is consistent to the literature [13
]. This geographical inequity has left most camps and rural areas lacking these services. Conversely, while VCT services were evenly provided across health facility levels, PMTCT and ART were mostly available at Health center IV and hospitals. This might be due to the fact that VCT has been ongoing in Uganda for decades while PMTCT and ART have been recently introduced. Another reason is the policy issue as in Uganda ART services are required to be provided at HC IV and above, with a medical officer being a pre-requisite [6
] while for PMTCT, the policy requires this service to start at HC III and above [10
Availability of minimum essential components also shows a similar pattern of inadequacy at lower health facility levels. The results of the stock-outs of HIV test kits and drugs at lower health facilities reflect limited capacity of supply chain system probably as a result of insecurity or funding. Similarly, shortage of health staff for ART might be due not only to inadequate number but mal-distribution of this cadre within and among the districts. For instance in Gulu, 79% of all medical officers work in the hospitals within the Municipality and 86% of those in Pader work in one hospital [15
]. As most of these hospitals are located in municipalities, this further indicates urban-rural disparities in the access particularly to PMTCT and ART in northern Uganda.
Based on our data, there is evidence of utilization of VCT, PMTCT and ART services in all the districts. For VCT, the data showed a significant proportion of those counseled received HIV testing, however, our study did not assess the characteristics or reasons for those clients who did not take HIV testing. Our results also show that most PMTCT utilization was observed in Gulu despite an even distribution of PMTCT sites among districts. This is likely due to the presence of large number of hospitals compared to Kitgum and Pader reflecting the possible differences in quality of services. The paradoxical finding of ART utilization in which Kitgum had highest proportion of individuals on ART in the last month despite the limited number of ART sites should be interpreted cautiously. Without data on monthly trends in utilizations, this finding can be misleading. Yet, for Pader district the consistent lowest utilizations of VCT, PMTCT and ART shown by this study is likely driven by the inadequacy of services and resources including stock-outs of medicines as reported by other studies in the past [5
]. However, other factors such as demographic and insecurity differences which have not been elucidated by this study, might also be the likely explanation.
In health system perspective, access defined in terms of physical availability and service utilization, is one of the intermediate outcome measures which has been increasingly used to determine health system performance as it influences both, health status and client satisfaction [9
]. In this context, our findings of limited access to VCT, PMTCT, ART might be a reflection of limited HIV/ART system performance in this region. The findings of other studies on high prevalence and mortality due to HIV/AIDS in Northern Uganda [4
], might likely be proximately related by limited access of these services as depicted by this study. Similarly, the gaps observed by this study, underpins the importance of scaling up of VCT, PMTCT and ART in this sub-region. The main challenge however is that most of the health facilities in rural areas or camps comprise of lower health facility levels which lack appropriate health personnel and medicines to offer services such as PMTCT and ART. Critical to this is therefore a review of current policies including the recruitment and retention of appropriate staff so that services like PMTCT and ART can be rolled out to lower-level health facilities.
Despite the evidence that application of GIS methodology in emergency settings is limited [16
], still there were no studies which determined the feasibility of its use in these settings. In our study, the training on the use of GPS receivers, which were ordered locally, was done by Uganda authorities indicating that the local expertise is available. The analysis of data and production of maps was also accomplished locally. Training and complete data collection in this insecure and wide geographical area took less than 1 month. These indicated that using GIS as a tool in health assessments in conflict settings is feasible and can be locally undertaken.
Our finding of access can be limited by several factors. The use of last month facility attendance for measuring and comparing utilization among the districts can be event driven and might not accurately represents monthly average or variations in utilizations. This type of data can not give users' perspective on utilization and may lack comprehensive information of other determinants of utilization such as demographic variations, acceptability or user-satisfaction. Shortage of staff at lower health facilities might likely impair data collection hence underreporting utilizations. Using district population to estimate and compare utilizations among districts is also likely to be misleading. The use of catchment populations which the facility sub-served would have increased the reliability of our comparisons. Moreover, absence of the comparison district which is conflict-free, is another limitation as it would have determined whether the limited access to and gaps in providing VCT, PMTCT and ART services in northern Uganda was attributed to the longstanding conflicts. It is therefore essential to further examine other factors that affect access particularly the correlates of utilization of these services in this conflict region so as to achieve the goal of universal access by 2010.
In conclusion, the study shows that access to VCT, PMTCT and ART services in northern Uganda is geographically limited due to inadequacy and uneven availability and utilization of these services among districts, health facilities and camps with Pader district mostly affected. Addressing the gaps depicted by this study requires policy review, equitable geographical re-distribution or recruitment of appropriate staff and scaling up plans focusing on essential minimum components of services at lower health facilities. This study has shown that measuring access in emergencies not only provides information for health policy and planning but supplements information related to health system performance and health status of the population. Similarly, application of GIS for health need assessments in conflict settings is feasible and maps can be effective in presenting large set of data into simplistic, visual friendly and easily interpretable information.