3. Assessment of outcomes.
Objectives: The objective of preventing the spread of HIV among high risk groups, and within the general population, was and continues to be highly relevant to Bangladesh. While 1998-99 sero-surveillance data indicated that prevalence rates were still low among high risk groups, risk factors for rapid spread were present. Globally, the Millenium Development Goals established in September 2000 had established halting and reversing the spread of HIV prevalence as an explicit goal for all countries; and the Global Fund to fight AIDS, TB and Malaria (GFATM) had begun to support large scale prevention and treatment programs across the globe.
Design: The design of the program, with a strong focus on TIs among high risk groups, was in keeping with global best practice. The additional components of advocacy and communication, blood safety and capacity building were appropriate for the HIV program; and they were in keeping with similar projects being implemented in the region and globally.
Implementation: The implementation of the TIs through NGOs was the appropriate strategy, since NGOs had a strategic advantage in dealing with the marginalized communities at high risk of HIV infection. Implementation arrangements for the project were initially envisaged as part of the sector-wide HPSP; but when this was shown to be ineffective, it was changed at the MTR into more of a project mode, to facilitate greater focus and ownership. Partners were also brought on board for the implementation of key components, to strengthen the hands of the NASP.
3.2. Achievement of Project Development Objectives (including brief discussion of causal linkages between outputs and outcomes, with details on outputs in Annex 4):
PDO and Intermediate outcome indicators that were tracked over the project were: HIV prevalence among high risk groups (end of project target: < 5%); proportion of high risk groups reached by NGOs contracted under the project (end of project target: 100 percent); proportion of IDUs that did not share needles in the last week (end of project target: 95 percent); and proportion of FSWs that reported using condom during last sex act with new client (end of project target: 50 percent). The assessment of achievement of development objectives is based largely on these indicators using data from serological and behavioural surveillance surveys. Baseline values for 2001 are available but since project supported targeted interventions were initiated in August 2004 the analysis will focus mainly on changes during this period.
Here it is important to note that targeted interventions are also implemented through other financing streams. HIV preventive services for high risk populations in Bangladesh have been underway for more than a decade. In 1997/98, CARE Bangladesh, with DfID financing, initiated interventions in Dhaka among brothel and street based workers and harm reduction programs for IDUs, expanding to other cities and high risk groups before ending in March 2005 with the withdrawal of DfID financing. In the subsequent period, needle exchange programs and interventions among brothel based FSWs were entirely supported by the project. Preventive efforts with other sub-groups of female sex workers were also implemented by Family Health International (FHI) with USAID support, although HAPP was the larger financier for these groups. Other high risk groups such as MSM, MSW and transgender sex workers were largely financed by FHI. HAPP support for these populations was initiated late and covered a short period of barely seven months. Targeted interventions supported by the Global Fund have been initiated only recently at the end of 2007.
HIV prevalence among high risk groups: The seventh round of serological surveillance undertaken in 2006 confirms that HIV prevalence among all high risk groups with the exception of injecting drug users in Dhaka has remained below 1 percent (FSW <1%; MSM 0.6% using BSS and 0.2% using RDS) with no significant change over the five rounds of surveillance undertaken between 2001 and 2006. HIV prevalence among IDUs was less than 2 percent in all sites with the exception of Dhaka. In Dhaka, prevalence rose from 1.7 % in 1999 to 4 % in 2002 and in 2006 to 7 percent marking the first concentrated epidemic among any high risk group in Bangladesh. The epidemic was localized to a sub-region in Dhaka where prevalence has risen to 10.4 percent as compared to prevalence of 1 percent in other part of the city. Overall levels of prevalence among the general reproductive age population is less than 0.1 percent as would be expected given the low prevalence among most at risk populations and the protective effect of high levels of male circumcision.
Syphilis prevalence among high risk groups: Active syphilis rates have fallen among most high risk populations. Between 2003-04 and 2006 a significant decline in active syphilis rates was noted in brothel based sex workers, street based female sex workers in Dhaka and among IDUs in Dhaka and Rajshahi. Active syphilis rates among brothel workers ranged between 3.2 and 12.2 percent in 2003-04 while the comparable range for 2006 was 0.4 to 6.3 percent. Active syphilis rates also declined among Hijras, but remained similar in MSM and MSW. There has been little change in syphilis prevalence of street based workers in Chittagong and hotel based workers in all three major cities.
Behavioural Indicators: Progress on behavioural outcomes is mixed and varies by high risk groups and location (Tables 1&2). Needle sharing among IDUs worsened or remained unchanged in the cities sampled for BSS with the exception of Dhaka. In Dhaka which has the highest concentration of IDUs and is the main centre of the epidemic there was considerable reduction in reported sharing from 85 to 55 percent. Despite the improvement the majority was still using unsafe injecting practices, i.e. the reduction achieved is not enough for lasting program impact. Condom use rose substantially between 2003-04 and 2006-07 in most groups of female sex workers. The increase is particularly notable among street and brothel based FSWs while modest or little change was observed among FSWs working in hotels in Dhaka and Chittagong. Hotel based workers are specially vulnerable to HIV as they have the highest number of clients; mean number of clients in the last week for hotel based workers in Chittagong and Dhaka was 61 and 42 respectively as compared to 19 for brothel based workers and between 8 to 15 for street based workers in Dhaka, Chittagong and Khulna.
The evidence does not in some cases indicate the expected linkage between condom use and changes in syphilis prevalence. While a falling trend in syphilis prevalence is noted for brothel based workers and street based workers in Dhaka, substantially higher condom use particularly among street based workers in Chittagong was not associated with a significant decline in syphilis prevalence.
Coverage: The BSS monitors coverage as reported by high risk populations in select cities with the highest concentration of particular risk groups. A comparison of BSS data for 2004 and 2006 indicates that exposure to interventions reported by IDUs and FSWs declined in all cities and subgroups with the exception of IDUs in Chandhpur where an overwhelming majority was involved in the needle exchange program and street based workers in Chittagong where reported coverage increased dramatically from 42 to 87 percent (Tables 1 & 2). Lower levels of exposure could be due to the timing of the survey from November 2006 to February 2007 that overlapped with the transition period between phase two and three of the UNICEF contract from June to December 2006. The transition and the associated interruption in financing would have adversely impacted on field level activities. An additional problem could be the fact noted in the BSS report that the surveillance was preceded by a clean-up drive as well as political unrest that would likely have lead to dispersion of street based high risk populations such as IDUs and FSWs making them more difficult to reach.
Exposure to needle exchange programs reported by IDUs was still reasonably high ranging between 73 and 98 percent. Mean number of contacts with the interventions over the past month varied more markedly across sites from 15 and 11 in Chandhpur and Dhaka, respectively, to less than 3 each for Rajshahi and Chapinawabganj. The latter two cities also saw an increase in needle sharing. In 2006/2007, even in the better served cities, the impact on needle sharing is not evident, reflecting on the poor quality of the contacts, particularly the adequacy of syringe and needle distribution. Estimates derived in the NGO Performance Audit indicate that roughly 23 needles/per month/per IDU were supplied during 2007 on the assumption that these were only provided to IDUs and not the larger population of drug users which is also covered by the interventions (Performance Audit 2007). The average estimates hide irregular supply of needles and syringes over the year. According to NGO monitoring reports the supply of needle/syringes was significantly lower in the first quarter of December to Februray 2007, the initial start-up period of the third phase, more than doubled over the second quarter and increased by another 40 percent in the third quarters (UNICEF 2007).
Despite lower reported coverage, condom use improved remarkably among brothel based and street based FSWs in Dhaka as well as Khulna. Although exposure to interventions had decreased over the previous round, more than half of street based workers (56 percent in Dhaka and 53 percent in Khulna) and three fourths of brothel workers reported contact with prevention efforts. In general, participation of FSWs was highest in the education component of the program on HIV-AIDS, safe sex and correct use of condoms but they were not so dependent on NGOs for condom supplies;in the previous month the majority of brothel based (77 percent) and street based workers in Dhaka (61 percent) relied on shops or pharmacies for their supply of condoms.
An assessment of country-wide coverage of interventions of high risk groups has been initiated on the basis of a consultative process involving NGOs, government organizations and research institutions. In the absence of comprehensive mapping information, the likelihood of error in size estimates is considerable. Preliminary estimates suggest 100 percent coverage of brothel based workers, 51 percent for IDUs while the scale of preventive services for street based FSWs, MSM/MSW and transgender is limited at 18, 15, and 10 percent respectively. Nationwide coverage of hotel based workers at 90 percent is contrary to findings of BSS of very low exposure to interventions even in the main urban centres of Dhaka and Chittagong. The estimates are being re-examined currently by a group lead by UNAIDS and including the Bank.
3.3. Efficiency (Net Present Value/Economic Rate of Return, cost effectiveness, e.g., unit rate norms, least cost, and comparisons; and Financial Rate of Return):
No economic analysis was undertaken during project preparation. Instead of ex-ante cost effectiveness analysis, the PAD indicated that the project would finance monitoring and evaluation of costs and effects of specific interventions, which would contribute to a meaningful economic analysis in the future. This was not undertaken during the course of the project.
Efficiency is rated as modest. The project appropriately prioritized preventive interventions targeted at high risk populations that are most cost effective in a concentrated epidemic. Targeted interventions accounted for 46 percent of total project costs. Expenditure by high risk groups are not readily available but rough estimates for 2007 provided in the Performance Audit of NGOs indicate that the package for injecting drug users/heroin users accounted for roughly 50 percent of total expenditures on preventive programs for high risk groups in 2007 (HLSP 2008). Heroin users who occasionally inject and are at much lower risk comprised roughly two-third of the targeted population covered by the services. The delay in initiating the recruitment of NGOs, short term contracts with gaps in funding and associated uncertainties and fluctuations in field activities diluted the impact of preventive efforts on behavioural outcomes particularly among groups that were most at risk including IDUs and hotel based sex workers.
3.4. Justification of Overall Outcome Rating (combining relevance, achievement of PDOs, and efficiency):
Rating: Moderately Unsatisfactory
Although the PDO remains highly relevant to the country context, corporate goals and global priorities and the project's development objective has been met, the evidence on intermediate indicators of behavioural change and coverage suggest that it is impossible to ascribe the outcome to project interventions alone. Prevention efforts have had limited effect on changing behaviours of groups of the greatest epidemiological importance to the epidemic such as IDUs, hotel based sex workers and transgender workers. Additionally the evidence points to inefficiencies due to short term contracts particularly for IDUs that not only led to fluctuations in coverage but also negatively affected quality particularly of harm reduction services through irregular supply of needles and syringes. The project however did contribute to safer behaviours among select groups of sex workers. The increase in reported condom use along with data on declining STI rates for brothel based workers nationally and street based workers in Dhaka suggest that interventions have been effective in reducing risk behaviours in these populations. As noted earlier interventions for street based sex workers were also implemented by FHI.
3.5. Overarching Themes, Other Outcomes and Impacts (if any, where not previously covered or to amplify discussion above):
(a) Poverty Impacts, Gender Aspects, and Social Development
See section 3.2.
(b) Institutional Change/Strengthening (particularly with reference to impacts on longer-term capacity and institutional development):
It had been envisaged that the project would be implemented as an integral part of the HPSP, through government entities, NGOs and other private sector organizations. No separate implementation unit was to be created (PAD; pp 18). However, as part of HPSP, the implementation concerns of the larger sector-wide program completely overshadowed the objectives and challenges of the HIV program; and a separate project management unit was set up. Institutional capacity has been an on-going area of concern: in December 2002, the IDA review team (PSR#5) pointed out that "Weak management, low implementation capacity and lack of motivated staff have been the most critical underlying causes for lack of implementation progress". This situation did not change substantially throughout the project (A-M, November 2007). The expectation that HAPP could manage the contracting of NGOs was unrealistic: the experience under the Bangladesh NNP had also shown that this was a difficult area for which GOB had limited capacity. The capacity of NASP/GOB to provide on-going oversight and monitoring of NGO contracts was also limited, and there is some evidence that this could have lead to mismanagement of funds in some cases (see note: BWHC Corruption Issues 060508, internal memo, Bangladesh HNP team).
Associated with this are questions regarding the political commitment to the program within GOB (specifically MOHFW): although some major policy decisions were taken, delays in granting clearances for key activities such as the extension of the UNICEF contract, and frequent transfer of key officials, seem to indicate a lack of ownership of the program. Repeated requests (A-M, June 2003; April 2006) that GOB demonstrate long-term commitment to the program by undertaking an Functional Task Analysis (FTA) have not led to the desired results. The recommendations of the FTA undertaken by UNFPA have not yet been accepted or implemented by the NASP/MOHFW.
With specific reference to the UNFPA efforts towards capacity building of NASP, from September 2003 until March 2007, under the contract (i) about 252 person months of local TA and 4 person months of international TA were provided; (ii) 13 pieces of technical work were produced very largely with their support, including technical guidelines and support materials essential to the project and program; and (iii) a FTA was completed, which provided an analysis of the capacity building needs of NASP and recommendations. However, an assessment of UNFPA's efforts conducted by the Bank in April 2007 concludes that very little was accomplished by way of long-term institutional strengthening of NASP, although this was repeatedly raised as an issue by various Bank/DfID missions.
(c) Unintended Outcomes and Impacts (positive and negative):
HAPP laid the foundation for the development of the national program, managed through the NASP office. This resulted in the development of the national response to HIV/AIDS, as well as the beginnings of some stewardship capacity; and provided an opportunity to develop the capacity to work with international agencies, particularly the UN, in a coordinated manner. The capacity built through the project was instrumental in drafting the successful GFATM proposals.
3.6. Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops (optional for Core ICR, required for ILI, details in annexes):
Table 1: Trends in Needles/Syringes Sharing and Reported Coverage 2002-03 2003-04 2006-07 % of IDUs not sharing needles/syringes in last week Dhaka 34 14 45 Rajshahi 73 79 51 Chapinawabganj 53 36 26 Chandhpur 26 37 37 2002-03 2003-04 2006-07 % covered by needle exchange programs last year Dhaka 45 88 80 Rajshahi 88 93 73 Chapinawabganj 53 88 75 Chandhpur 99 Source: Behavioural Surveillance Surveys-Round 4-6 Table 2: Trends in Condom Use and Reported Coverage of FSWs 2002-03 2003-04 2006-07 % of FSWs who used condoms during last sex with new client Brothel Workers 36 40 70 (National) Street Based Workers 36 38 81 Dhaka Chittagong 22 14 91 Khulna 21 51 Hotel Based Workers 24 30 40 Dhaka Chittagong 38 36 Sylhet 71 2002-03 2003-04 2006-07 % of FSWs covered by interventions Brothel Workers 86 88 75 (National) Street Based Workers 51 96 56 Dhaka Chittagong 21 42 87 Khulna 71 53 Hotel Based Workers 72 44 20 Dhaka Chittagong 64 9 Sylhet 37 Source: Behavioural Surveillance Surveys-Round 4-6
|Printer friendly Cite/link Email Feedback|
|Publication:||Bangladesh - HIV/AIDS Prevention Project|
|Date:||Jun 10, 2008|
|Previous Article:||2. Key factors affecting implementation and outcomes.|
|Next Article:||4. Rationale for rating of risk to development outcome.|