Document of The World Bank Report No: ICR00001452 IMPLEMENT ATION COMPLETION AND RESULTS REPORT (IDA-H1290) ONA GRANT IN THE AMOUNT OF SDR 6.90 MILLION (US$ 10.0 MILLION EQUIVALENT) TO THE GOVERNMENT OF THE UNITED REPUBLIC OF TANZANIA FORA AFRICAN REGIONAL CAPACITY BUILDING NETWORK FOR HIV/AIDS PREVENTION, TREATMENT AND CARE (ARCAN) November 1,2010 Health, Nutrition & Population (AFTHE) Human Development Network Regional Integration (AFCRI) Africa Region CURRENCY EQUIVALENTS Exchange Rate: as of 0 I-Nov-201O (Exchange Rate Effective, November 1,2010) Currency Unit = XDR 1 XDR = 1.577420 USD FISCAL YEAR July 1- June 30 ABBREVIA nONS AND ACRONYMS ARCAN African Regional Capacity Building Network for HIV/AIDS Prevention, Treatment and Care ALERT All Africa Leprosy, Tuberculosis and Rehabilitation Training Centre ATG ARCAN trained graduate CAS Country Assistance Strategy ENHRI Ethiopian Health and Nutrition Research Institute ESAMI Eastern and Southern Africa Management Institute EU European Union FMA Financial management and procurement agent FMOH Federal Ministry of Health (Ethiopia) HAPCO HIV/AIDS Prevention and Control Office (Ethiopia) HIV/AIDS Human immunodeficiency viruslAcquired immune deficiency syndrome or acquired immunodeficiency syndrome M&E Monitoring and evaluation MAP Multi-Country HIV/AIDS Program for Africa MoH Ministry of Health MTR Mid-Term Review NAC National AIDS Commission! Council NACP National AIDS Control Program NBS National Bureau of Statistics (Tanzania) NSGRP National Strategy for Growth and Reduction of Poverty, a.k.a. MKUKUT A (Tanzania) PAD Project appraisal document PASDEP Plan for Accelerated and Sustainable Development to Eradicate Poverty (Ethiopia) PCU Project coordination unit PDO Project development objective PEPFAR US President's Emergency Plan for AIDS Relief SC Steering committee TB Tuberculosis ToT Training of Trainers ZPRP Zanzibar Poverty Reduction Plan a.k.a. MKUZA Vice President: Obiageli Katryn Ezekwesili Country Director: Yusupha B. Crookes Sector Manager: Eva Jarawan Project Team Leader: Dominic S. Haazen ICR Team Leader: Onur Ozlu REGIONAL INTEGRATION (AFRCI) BORROWER: TANZANIA IMPLEMENTATION COUNTRIES: ETHIOPIA, KENYA AND TANZANIA AFRICAN REGIONAL CAPACITY BUILDING NETWORK FOR HIV/AIDS PREVENTION, TREATMENT AND CARE (ARCAN) PROJECT CONTENTS Data Sheet A. Basic Information B. Key Dates C. Ratings Summary D. Sector and Theme Codes E. Bank Staff F. Results Framework Analysis G. Ratings of Project Performance in ISRs H. Restructuring (if any) I. Disbursement Graph 1. Project Context, Development Objectives and Design ............................................... 1 2. Key Factors Affecting Implementation and Outcomes ............................................... 4 3. Assessment of Outcomes ............................................................................................ 8 4. Assessment of Risk to Development Outcome ......................................................... 16 5. Assessment of Bank and Borrower Performance ..................................................... 17 6. Lessons Learned ........................................................................................................ 18 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners ........... 19 Annex 1. Project Costs and Financing .......................................................................... 21 Annex 2. Outputs by Component. ................................................................................. 22 Annex 3. Economic and Financial Analysis ................................................................. 28 Annex 4. Bank Lending and Implementation Support/Supervision Processes ............. 32 Annex 5. Beneficiary Survey Results ........................................................................... 34 Annex 6. Stakeholder Workshop Report and Results ................................................... 36 Annex 7. Summary of Borrower's ICR and/or Comments on Draft ICR ..................... 37 Annex 8. Comments of Cofinanciers and Other Partners/Stakeholders ....................... 39 Annex 9. List of Supporting Documents ...................................................................... 40 Annex 10. Maps ............................................................................................................ 41 A. Basic Information African Regional Capacity Building Network for Country: Africa Project Name: HIV1 AIDS Prevention, Treatment, and Care Project ID: P080406 LlC/TF Number(s): IDA-H1290 ICR Date: 10104/2010 ICR Type: Core ICR UNITED REPUBLIC OF Lending Instrument: SIL Borrower: TANZANIA Original Total Commitment: XDR 6.9M Disbursed Amount: XDR6.9M Revised Amount: XDR 6.9M Environmental Category: C Implementing Agencies: Office of the Prime Minister B. Key Dates Reviscd 1 Actual Process Datc Process Original Date Datc(s) Concept Review: 01109/2003 Effectiveness: 02/28/2005 02/28/2005 Appraisal: 0411912004 Restructuring( s): Approval: 09/22/2004 Mid-term Review: 1010112007 1010112007 Closing: 06/30/2009 04/30/2010 C. Ratings Summary Col Performance Rating by ICR Outcomes: Moderately satisfactory Risk to Development Outcome: Moderate Bank Performance: Moderately unsatisfactory Borrower Performance: Moderately satisfactory C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Bank Ratings Borrower Ratings Quality at Entry: Moderately unsatisfactory Government: Moderately unsatisfactory Implementing Quality of Supervision: Moderately unsatisfactory Satisfactory AgencylAgencies: Overall Bank Overall Borrower Moderately unsatisfactory Moderately satisfactory Performance: Performance: QAG Assessments (if Implementation Performance Indicators Rating l:1~Y) . Potential Problem Project at Yes Quality at Entry (QEA): None any time (Yes/No): Problem Project at any time Quality of Supervision Yes None (Yes/No): (QSA): DO rating before Moderately Satisfactory Closing/Inactive status: - . .. D. Sector and Theme Codes Actual .~~~~~.~.~~.~.e.J~_~_~!~~!~~~~.!1~.!i_~.~~_c.~~J;~) . Central government administration 15 Health 85 .!~.e.me~ode (~ ..:o.~f total1J.~~~f!"'~~_~.~!I_gl. HIV/AIDS 100 V ice President: . 2~ i.~~~.!i.~~~!"X~_~~e.~.~_e.~~.~i __ Callisto Madavo ... <;()u.~tryDirector: . yu~~pha~: Crooke~ ... J udyQ'Connor Sector Manager: Eva Jarawan .Dzingai Mutumbuka Project Team Leader: Dominic S. Haazen Sheila Dutta ICR Team Leader: Mehmet Onur Ozlu N/A fCR Primary Author: Mehmet Onur Ozlu N/A F. Results Framework Analysis Project Development Objectives (from Project Appraisal Document) Project Objective: To expand access to comprehensive and evidence-based HIV/AIDS prevention, treatment, and care programs by supporting a network of subregional learning sites to expand training of health care practitioners. The project supported HIV/AIDS-related health sector human resource capacity building in Kenya, Ethiopia, and Tanzania using a training-of-trainers (TOT) model. The project components include (a) health care practitioner training; (b) monitoring, evaluation and knowledge-sharing; and (c) program coordination. It was expected that, as a result of increased capacity among health care practitioners in the subregion, this project would contribute to the enhanced delivery of HIVIA IDS-related services, stronger linkages between prevention and care interventions, and expanded South-South partnerships. Revised Project Development Objectives (as approved by original approving authority) Not Applicable (a) PDO Indicator(s) Original Formall Target y Baseline Actual Value Achie"'cd at Indicator Values (from Reviscd Value Completion or Targct Years~ approval Target documents)1 Values Increase in the percent/number of successful ARCAN "graduates" (i.e. those who conduct at least the minimum required level of cascade training in the 50 % of ATGs conducted at least first six months after their preliminary training) whose effectiveness was o 50% N/A one cascading training session (848 out of 1721) predicted on the basis of the project's " student selection criteria ratings, pre- and post-course tests/evaluations, and/or rating 0 ftraining work plan quality. 6.8% - 1006 service delivery points with ATGs out of 14,721 SDPs across three countries. Increase in percent of service delivery The percent is lower than the target points in ARCAN countries with one (or but the number of SDPs has greatly more) primary recipients of project- o 1()'% N/A increased since project design. For supported TOT training. example, between 2003 and 2006 the number of VCT centers in Tanzania more than tripled. The trend is similar in Ethiopia and Kenya. 74% of ARCAN graduates felt that the training greatly increased their Increase in level of job satisfaction appreciation and understanding of reported by ARCAN graduates one year N/A 30% N/A the job, and 72% felt that the after course completion. training greatly influenced their care delivery behavior. I Original target values were established in April 30, 2005 shortly after project launch. 2 Data from last ISR dated 04/29/2010 (b) Intermediate Outcome Indicator(s) Original Target Formally Actual Value Achieved at Values (from Revised Indkator Baseline Value Completion or Target approval Target Years4 documents) J Values Increase in the percent of ARCAN graduates who conduct one training session in the first 6 months after their 0 50'Y., N/A 50 ARCAN traini ng, as measured annually through submitted training reports 4/30/2010 US$ 3,599,9885 Total amount of financial resources 11/30/2009 US$ 3,398,988 mobilized by ARCAN graduates (from Annual increases 6/23/2009 US$ 3,019,278 national MAPs and other sources) to O. in resources N/A 9/19/2008 US$ 1,265,427 conduct their ToT-model training, as mobilized 5/2112007 US$ 13,300 measured annually 9/30/2006 US$ 0 4/30/2010 1083 Increase in the number of training Increase in the 11130/2009 1026 monitoring reports submitted to the number of report 6/23/2009 957 ARCAN Project Coordination Unit by 0 N/A submitted 9/19/2008 549 ARCAN graduates (compiled annualJ)' 5/2112007 35 annually) 9/30/2006 0 Health workers trained in-service 0 21.630 N/A 69,1146 3 Original target values were established in April 30, 2005 shortly after project launch. 4Data for cumulative indicators taken from last ISR dated 04/2912010. Data for annualized indicators taken from respective ISRs with the dates noted. S Values are cumulative, measured annually. 6 Total value includes 1,721 ATGs (directly trained by ARCAN), 12,870 cascade trainees (trained by ATGs) and 54,523 informal cascade trainees who received on the job training and/or awareness creation from ATGs. G. Ratings of Project Performance in ISRs Actual Disbu,"scment Date ISR No. DO IP s Archived (USD millions) 11/30/2004 Satisfactory Satisfactory 0.00 2 05106/2005 Unsatisfactory Unsatisfactory 2.00 3 11/23/2005 Unsatisfactory Unsatisfactory 2.32 4 05/31/2006 Unsatisfactory Unsatisfactory 2.32 5 11/21/2006 Moderately Unsatisfactory Moderately Satisfactory 2.32 6 06/1312007 Satisfactory Satisfactory 5.35 7 12/28/2007 Satisfactory Satisfactory 6.17 8 06/25/2008 Satisfactory Satisfactory 8.38 9 09/1112008 Satisfactory Satisfactory 8.38 10 12/30/2008 Satisfactory Moderately Satisfactory 9.22 II 06/30/2009 Satisfactory Satisfactory 10.00 12 12/23/2009 Satisfactory Satisfactory 10.55 13 04/29/2010 Moderately Satisfactory Satisfactory 10.55 H. Restructuring (if any) Not Applicable I. Disbursement Profile - Original ---- Formally Revised - - Actual 12~------------------------------------------------------------~ .,.,.-- 10 8 "" c :§ 6 i ~ Ion 4 ;:) 2 0 ,.., 0 ... ,.., ~ 0 ... 0 N 0 ,.., ~ Q ... ~ 0 ,.., 0 ~ ,.., N 0 ... Q 0 N 0 Q N 0 0 ~ 0 Q N ,.., ,.., ,.., ,.., 0 = = = = = = = = = = = = = = = = = ... ... lI' lI' = ~ ~ ~ ~ = = GO GO GO GO CI'I CI'I CI'I = = = = = = ..... = = = = = = = = = = = = N N N N N N N N N N N N N N N N N N 1. Project Context, Development Objectives and Design 1. HIV IAIDS is a major economic, social, and human challenge to Sub-Saharan Africa: Two-thirds of all people infected with HIV-22.5 million people with HIV live in the region. The impact of HIV I AIDS on households, human capital, the private sector, and the public sector undermines the alleviation of poverty. 1.1 Context at Appraisal 2. Country situations with respect to HIV/AIDS at appraisal were dire: Ethiopia, Kenya, and Tanzania were among the most heavily affected countries by the HIV/AIDS epidemic. A UNAIDS estimate as of December 2003 put the national adult prevalence rate in Ethiopia at 4.4 percent, indicating that 1.5 million Ethiopians (adults and children) were living with HIV I AIDS. There were also estimated to be 720,000 AIDS orphans (under the age of 17 ye'ars) in Ethiopia at that point. The national adult prevalence rate in Kenya was estimated to be 6.7 percent in December 2003, with 1.2 million individuals living with HIV/AIDS and 650,000 orphans. In Tanzania, the corresponding adult prevalence rate was 8.8 percent, with 1.6 million individuals living with HIV/AIDS and over 980,000 orphans.? 3. In responding to this challenge, three countries committed to containing the epidemic by developing' universal prevention, care, and treatment interventions for their populations. They placed a high priority on the provision of relevant and comprehensive training for health care practitioners involved in HIV/AIDS. 4. ,Similarly, addressing HIV/AIDS challenges and focusing on capacity issues were also prominent in the Bank's Country Assistance Strategies (CAS) for Ethiopia, Kenya and Tanzania. The common focus of the three countries' CASs was to improve human development outcomes by enhancing pro-poor growth and reducing vulnerability. The strategies indicated that the health sector was essential to achieving this goal, and pointed out HIV I AIDS as the biggest challenge faced by the sector. The Ethiopia CAS identified that the shortage of skilled human resources in HIVIAIDS as the foremost threat to national poverty and vulnerability reduction efforts. The Kenya CAS stated that HIV/AIDS was a top priority for the government. In the case of Tanzania, the CAS noted the emergence of HIVIA IDS as a multi-sectoral development issue posing a severe constraint to poverty alleviation efforts. 5. The essentials of stopping this trend are prevention, treatment and care. s The majority of country level responses in late 1990s/early 2000 aimed at addressing these pillars. Despite increased levels of political commitment at both national and international levels, substantial deficiencies remained in building the necessary capacity in Ethiopia, Kenya, and Tanzania. Part of the capacity gap resulted from the lack of financing by development partners and donors for capacity building projects at the time of appraisal. A more fundamental factor was that the additional skills and approaches required for an effective response to HIV I AIDS were not typically part of these countries' health systems. Also the health systems in these countries were relatively weak and under-resourced prior to the advent of AIDS, which has since served to exacerbate these institutional and financial vulnerabilities. During project appraisal, the maturing epidemic was placing greater strains on these already over-stretched systems, as the demand for health services grew concurrently with increasing morbidity and mortality. HIV/AIDS also has led to resurgenc~ in tuberculosis, a lethal synergy, which required yet more additional efforts and 7 UNAIDS Epidemic Report, July 2004 8 World Bank, The World Bank's Commitment to HIY/AIDS in Africa: Our Agenda for Action, 2007-2011 resources. This meant a dramatically increased workload for healthcare practitioners, among other strains on the healthcare system. 6. The result was that at a time when resources for HIV/AIDS were starting to increase, the capacity to translate the investments in equipment, drugs and other commodities for HIV/AIDS control was limited as low capacity among healthcare workers remained a significant problem. The African Regional Capacity Building Network for HIV / AIDS Prevention, Treatment and Care (ARCAN) project was designed to address this challenge in three large countries. 7. The Governments of Ethiopia, Kenya and Tanzania and the Bank jointly adopted the "training-of-trainers" (TOT) model to address the capacity problem through this operation, where health care practitioners from three countries were to be brought together at regional learning sites. Project designers envisioned that African Regional Capacity Building Network for HIV/AIDS Prevention, Treatment and Care (ARCAN) project trainees would then train other health workers upon return to their countries. In determining the content for the training, project design adopted, to a large extent, the findings of a 2003 USAID study that analyzed the capacity needs for HIV / AIDS-related training in 12 Eastern and Southern African countries. The subsequent choice to deliver the trainings in multiple countries a~d institutes, on the other hand, was made due to (i) the lack of a single country that possessed training programs with the requisite proven implementation capacity and expertise in all priority areas and (ii) the desire to promote south- south learning through exchange of experiences between health workers. Given the increasing challenges faced by under-resourced healthcare workers, their low technical capacity and the relative absence of financial support towards building it, the project's choice to invest in targeted human capacity development through training was appropriate. 1.2 Original Project Development Objectives (PDO) and Key Indicators 8. The original PD~ was "to expand access to comprehensive and evidence-based HIVIAIDS prevention, treatment, and care programs by supporting a network of subregional learning sites to expand training of health care practitioners ". The PD~ was formulated differently in the legal agreement to account for the full range of activities supported under the project: "to expand access to comprehensive and evidence based HIVIAIDS prevention, treatment, and care programs in Participating Countries by: (a) supporting a network oflearning sites for the training of health care practitioners; (b) increasing the capacity of health care practitioners; (c) enhancing the delivery of HIVIAIDS-related services; (d) strengthening linkages between prevention and care interventions; and (e) expanding South-South partnerships. " 9. At the time of project design, there was a sense that while physical access to HIV / AIDS services was starting to expand (thanks to investments in equipment and commodities), human resource capacity constraints caused effective access to lag (because physical availability of services was not being translated into functioning HIV/AIDS services). The project rationale and description suggests that the primary focus of the project was therefore to improve capacity of HIV /AIDS service providers as a means to translating the physical access to HIV / AIDS services into effective access where capacity was lacking. 10. Key performance indicators stated in the PAD were as follows: I. Increase in the percent/number of successful ARCAN "graduates" (i.e. those who conduct at least the minimum required level of cascade training in the first six months after their preliminary training) whose effectiveness was predicted on the basis of the project's student selection criteria ratings, pre- and post-course tests/evaluations, and/or rating 0 ftraining work plan quality. 2. Increase in percent of service delivery points in ARCAN countries with one (or more) primary recipients of project-supported TOT training. 2 3. Increase in level of job satisfaction reported by ARCAN graduates one year after course completion. 1.3 Revised PDO 11. The PDO and KPls remained the same throughout project implementation. (See 2.3 Monitoring and Evaluation (M&E) Design: Implementation and Utilization for more details). 1.4 Main Beneficiaries, 12. Primary beneficiaries were professionals in the health sector, including physicians, nurses, counselor supervisors, laboratory technicians, HIVI AIDS program managers as well as procureinent and supply chain management specialists drawn from government, private sector and NGOs, working in urban and rural areas of the participating countries. 13. The indirect beneficiaries were people with HIV/AIDS and TB, as well as cases where HIV I AIDS and TB were averted. A secondary group of beneficiaries expected to benefit from the project were institutions which provided the training. Also, those with HIV I AIDS and TB in the three countries are another set of beneficiaries. 1.5 Original Components 14. The project had three components. The components and their estimated costs remained unc;hanged throughout implementation: a. Health care practitioner training component (US$ 7.72 million; 80.4% of project costs) aimed to support the implementation of short-term (non-diploma) training courses for the primary project beneficiaries (professionals in the health sector, including physicians, nurses, counselor supervisors, laboratory technicians, HIV/AIDS program managers as well as procurement and supply chain management specialists). The courses were to follow the ToT approach, where graduates of the course would train their colleagues upon return to their work stations; also known as "cascading". The ToT approach and cascading model were designed to maximize sub-regional capacity building by training as many healthcare workers as possible, despite relatively modest resources. b. Monitoring and evaluation and knowledge sharing component (US$ 0.38 million; 4.0% of project costs) was designed to establish an M&E system that would enable the rapid synthesis and sharing of lessons learned from the project. Routine dissemination of HIV IAIDS-related program information would also be financed under this component to ensure greater knowledge sharing between and within counties. Under this component, the project would also explore options for distance learning to maximize resources and to ensure a broader regional approach for sharing experiences and enhancing skills. c. Program coordination component (US$ 1.5 million; 15.6% of project costs) aimed to support the development of strong coordination and capacity building between ARCAN counties and their implementing, coordinating, and governing agencies. A major focus of this component involved providing the ARCAN Project Coordination Unit (PCU) with support to conduct its administrative, fiduciary, and technical responsibilities. 1.6 Revised Components 15. Components remained unchanged 3 1.7 Other significant changes 16. The composition of the PCU was overhauled in May 2006, when the original PCU staff members were terminated and a new group of staff was hired. This was welcome as it (i) helped improve relations between the Bank and the three countries by addressing the clients' concern over the transparency of the hiring ofthe original PCU, and (ii) untangled a number of key operational issues, such as the drafting of the implementation manual or conclusion of contractual agreements with the training centers (see 2.2 Implementation). 17. The project was extended from its. original closing d!lte on June 30, 2009 to April 30, 20 10, in order to have a little more time to complete project activities and resolve outstanding tax issues due to the regional nature of the project (see 2.4 Safeguard and Fiduciary Compliance). 18. In 2009, training courses were expanded to include prevention and control ofHIV/AIDS and TB co-infections. This change was important in addressing the growing number of co- infections and was responsive to global concerns about multi-drug resistant TB (see 2.2 Implementation). 2. Key Factors Affecting Implementation and Outcomes 2.1 Project Preparation, Design and Quality at Entry 19. During project preparation, the team drew lessons learned from existing AIDS projects, most of which were in a stage of "emergency response". Lessons drawn for the preparation of ARCAN from these projects included: (i) the crucial need to build capacity at all levels of the healthcare workforce to develop a long-term, sustained response to HIVI AIDS; (ii) importance of having a strong and flexible M&E system to track performance and capture lessons learned; and (iii) an emphasis on partnership and collaboration among governments, public and private sectors, civil society, donors and specialized agencies. The regional approach of ARCAN was not widely tested at the time of project design, particularly in addressing HIV I AIDS issues. Nonetheless, the design features adopted by the team appeared sound (e.g., TOT model; involving reputable institutions; accomplishing efficiency gains by involving regional bodies and exploiting opportunities for south-south learning). 20. While acknowledging key design challenges, the PAD correctly identified most of the potential risks posed, and recommended close monitoring of implementation as the main way to mitigate the risks. In retrospect, the two key shortcomings of the risk assessment were: (i) the underestimation of the risks associated with the funding for the cascading element9 and (ii) failure to have the respective ministries of health to officially endorse the training and recognize the ARCAN graduation certification. 2.2 Implementation 21. The project implementation had four distinct phases:(i) the period between approval in September 2004 and November 2005, which was dominated by disputes among the clients and between the clients and the Bank, effectively bringing the project to a halt at the onset; (ii) a subsequent two year phase of strong team leadership that addressed many of the problems and got the project back on track; (iii) a period during which the project's progress slowed, and (iv) a final phase starting in September 2009 during which project performance picked up and all outstanding operational issues were addressed. 9Within the ARCAN context, cascading is the provision of training by ARCAN graduates to their fellow colleagues upon return to their workstations. 4 22. The project's regional nature, which brought implementation efficiency (section 3.3), high design relevance (section 3.1) and other gains to the project, posed challenges in practical issues such as the location and staffing of the PCU. Disagreements on these issues during project start up created friction among the clients and the TTL, which delayed the start of project activities. To help break the stalemate, the Bank appointed a new TTL to the project in May 2005. This improved relations with clients which allowed the team to address operational problems, namely: (i) termination of the original PCU which lacked the necessary human resources and hiring a new and competent PCU team; (ii) addressing procurement irregularities; (iii) approving the project operations manual, which was a condition of effectiveness; (iv) conducting a financial audit and fraud investigation; (e) approval of operational manual, administrative and financial management handbooks for the project; (t) start of the main project activity - the trainings, and (g) involvement of health specialists in country offices in implementation. As a result, the project was lifted out of problem status. 23. The number of trainees reached over 1,000 by mid-2007. Ofthese, 35 had held cascade training by that time. DO and IP ratings, which were unsatisfactory, through 2006, became satisfactory. At project mid-term review (MTR) held in October 2007, task team and clients agreed to involve health specialists in Bank offices in Ethiopia, Kenya and Tanzania in the project with the objective of improving, among other things, the link between ARCAN and National AIDS Commissions/ Councils (NAC). Specifically, it was decided that Bank health specialists in client country offices would help convene national forums where A TGs presented their achievements, challenges, and future plans ensure that NACs would provide funding for cascade training through national MAP projects they managed and promote effective use of A TGs through other programs. 24. Following the improvements in project performance, the TTL moved to another duty station and a Washington-based operations analyst became the de-facto team leader. During this period, implementation was weakened by the lack of TTL response to numerous no-objection requests for operational and procurement issues, tax reimbursement problems and an overdue external project evaluation. Sector management did not address these issues, while the clients' attempts to move activities along were limited. 25. In September 2009, task team leadership changed hands to a seasoned TTL based in Tanzania. During this period, the outstanding operational issues were addressed-including the resolution of tax reimbursement problem, and the much delayed external project evaluation was completed 10 and the project was extended to ensure full completion of critical activities. A regional dissemination and stakeholder workshop was also organized to discuss post implementation strategy. 26. ARCAN implementation produced the following commendable aspects: First is the flexibility shown during implementation to address changing training demands. Highlighted as a significant risk during project design, the HIV /AIDS response to is a rapidly changing field and rigid training content would have been to the detriment of meeting the project objectives. ARCAN successfully adapted to the changing circumstances of the field by (i) regularly updating its training curriculum to present the most up to date information, as well as by (ii) introducing new trainings as they became necessary. The decision to deliver the training in the specialized regional institutes of excellence contributed significantly to the achievement of the first point, while the proactivity of the task team enabled the latter. Launching of the new training program IOThe external project evaluation conducted in March 20lOby a third party (ITAD, UK) to assess the effects of ARCAN training by measuring the change in trainees' technical knowledge and professional behavior as a result of the training as well as the effects of increased trainee capacity in health institutions they work. 5 on combined diagnosis and care of tuberculosis (TB) and HfV/AIDS in May 2009, for instance, is a good practice example for content flexibility. The second noteworthy element is the attention to quality during implementation. This will be discussed in greater detail under the achievement of project objectives (section 3.2). 27. In sum, ARCAN implementation was flexible to changing training demands, paid special attention to quality and produced significant and concrete results. 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization 28. Design at appraisal stage chose indicators that are measurable, but did not include baselines and targets. It should be noted, however, that this was common practice at the time, particularly for HIV I AIDS projects due to the new nature of the interventions and absence of established benchmarks. Target estimate values, however, were established in April 30, 2005, shortly after project launch. During project revival phase in 2006 and later in project implementation, the task team paid close attention to M&E and highlighted M&E as an area of improvement in Aide-memoires and ISRs. PCU's work on monitoring progress was complemented by a third party project evaluation undertaken in March 2010. Conducted by IT AD Consultants, UK, the evaluation assessed the effects of ARCAN training by measuring the change in ATGs' technical knowledge and professional behavior as a result of the training as well as the effects of increased trainee capacity in health institutions they work. To this end, the study specifically analyzed behavioral change, knowledge and skills gained and retained and changes in HIV I AIDS health care provision practices among ARCAN graduates, as well as their satisfaction with the training material. The study conducted document reviews, administered web surveys (reaching out to JOI respondents) as well as interviews with 60 ATGs and 28 beneficiaries of cascade training, in the three countries. The data and findings of the evaluation are extensively used in this report. 29. Two following M&E design and implementation elements are of importance: (i) PDQ: The PDO should have been worded more carefully to explicitly match the PDO with the scope of the project. Nonetheless, the PAD acknowledged that while capacity building alone will not expand access to HfV I AIDS prevention and care, it is an important means to translate the physical access to HIV I AIDS services into effective access where capacity is lacking. (ii) Indicators: The project's M&E structure contained measurable indicators such as the number of people trained directly by the project, the amount of money raised for cascading or the number of people trained through cascading. During implementation, the project team rigorously tracked this essential information. While tracking output level data adequately, the results framework would have benefitted from indicators that measured output quality (i.e. training quality) and outcomes (i.e. the post-training change in A TG behavior in executing their professional duties). Seeing this gap, the team specifically focused the external project evaluation on analyzing training quality and behavior change, and used the data from the evaluation to complement the formal results framework. 30. Aide-memoires and ISRs, as well as interv"iews conducted for this report indicate that the team was well aware of these shortcomings in the M&E system, wanted to make the necessary improvements and began the work for a formal restructuring of the project. However, after the preliminary work and discussions, task team and Bank management decided not to move forward with a formal restructuring since the project had already disbursed more than 60% of the funding and would therefore still need to be evaluated primarily against the original M&E indicators. 6 2.4 Safeguard and Fiduciary Compliance 31. The project was rated category C and hence did not trigger any safeguards. 32. Fiduciary arrangements: Given the project's regional nature, the project envisioned Financial Management (FM) and procurement to be carried out by a Financial Management and Procurement Agent (FMA) who would be contracted by the PCU. 33. Fiduciary compliance during implementation: The project faced a number of FM and procurement challenges which the team addressed diligently, albeit sometimes with delays by the Bank. At the outset, hiring of the original PCU was contested by the project Steering Committee (SC) members from Ethiopia and Kenya who raised concerns about procurement irregularities. This issue was resolved in May 2005, and a competitive hiring process was completed by the FMA in May 2006. Secondly, it took a long time to refine the project financial and procurement manuals, which were completed in September 2007 (delayed mainly due to delays from the Bank). Thirdly, the project was implicated with a fraud investigation on the use ofVS$ 57,665 project funds by the project's commercial bank, which was later resolved. In October 2007, the task team diagnosed poor performance by and overpayment ofthe FMA, and cancelled the contract. After a failed attempt to recruit a specialist, the PCU was charged with procurement activities and all procurement came under the review of the Bank. 34. The sluggish pace of Bank's response to requests for ''No Objection" slowed down project procurement and the absence of a procurement specialist within the PCU at times caused procurement irregularities, such as the hiring of a SC member to perform consulting services for the project. This issue was promptly diagnosed and resolved by the project team. Another issue has been the payment of income taxes on PCV staff salaries to the Tanzania Revenue Authority for the period of 2006-08, which were ineligible under the legal agreement. Financing parameters were changed in December 2009, which resolved the issue by allowing these taxes to be paid out of the grant proceeds. 2.5 Post-completion OperationlNext Phase 35. ARCAN graduates established national networks in the three countries to ensure the continuation of (i) cascade training and (ii) exchange of best.practices beyond project closing. Each with official NGO status, the networks were not in the original project design and they emerged as a result of ATGs' concern over the possible erosion of cascading and loss of contact. Both of these concerns were accurately raised as possible challenges to sustainability in the original project PAD, and the project team aptly supported the formation of the national ATG network entities by providing VS$ 5,000 of seed funding for each network, helping organize their meetings and introducing the entity leaders to their national AIDS commission/council to encourage the inclusion of the networks with appropriate HIV/AlDS activities. In addition to working towards mitigating some of the risks to sustainability of ARCAN, these entities, with over 1,000 members across Ethiopia, Kenya and Tanzania, represent a significant constituency in the fight against HlV/AlDS. 36. The project convened a regional dissemination and stakeholder workshop in March 2010, among other reasons, to prepare for the next phase. Led by the PCV, the workshop was attended by SC members, partner training institutions, ATGs, national ATG networks, health facilities with ATGs, national ARCAN focal officers, training needs assessment consultant, external project evaluation consultant, heads of training institutions in participating countries and the core ARCAN task team members (Annex 6). SC members committed at the workshop to support post- ARCAN cascade training and expressed their governments' request for ARCAN II. Subsequently, ARCAN II proposal has been developed, endorsed by the SC and forwarded to the Bank. 7 3. Assessment of Outcomes 3.1 Relevance of Objectives, Design and Implementation 37. Relevance of objectives. The stated development objective of ARCAN as designed in 2003 is to "expand access to HIV / AIDS prevention, treatment, and care programs by supporting a network of subregional learning sites to expand training of health care practitioners". As mentioned previously, the PDO should have been worded more carefully to explicitly match the PDO with the scope of the project. Nonetheless, this objective was relevant at the start of the project, and remains substantially relevant in 20 10 terms. The relevance of the objective is further supported by the fact that it is consistent with 2010 priorities of Ethiopia, Kenya and Tanzania, namely the Plan for Accelerated and Sustainable Development to Eradicate Poverty (PASDEP), Ethiopia; Kenya National AIDS Strategic Plan (2009110 - 2012113); Tanzania National Strategy for Growth and Reduction of Poverty (NSGRP or MKUKUT A) and Zanzibar Poverty Reduction Plan (ZPRP or MKUZA). 38. The context in the three countries show that this focus is well placed: HIV is still the leading cause of morbidity and mortality in the most economically productive age cohort (15 to 49 years) in Ethiopia and Kenya, and it follows malaria with the second place in Tanzania. Yet there is not enough human resource capacity in these countries to respond to the increasing demand for health services. Increasingly, national strategies and analyses conducted by the Bank and other donors point to enhancing human resource capacity as a crucial necessity in responding to the epidemic. Expanding capacity is therefore a major priority today, as was the case at appraisal. On this basis, relevance of objectives is rated high. 39. Relevance of design. In addition to full alignment with the priorities of the three clients as of 20 I 0, the project is also relevant from a design aspect. The project components and organization matched what the project aimed to achieve. The choice of Specific Investment Loan (SIL) as the instrument was appropriate given that the investment was used for training programs aimed at enhancing healthcare worker capacity and improving institutional infrastructure of the regional training institutions. 40. The regional character of the project was particularly appropriate. ARCAN's design as a regional project and its focus on partnership with training institutions across three counties merit special mention because these design elements provided multi-country and multi-level technical exchanges and partnerships, as well as implementation efficiency. The six partner training institutes developed the training curriculum, executed the training, provided accommodation for participants and took part in follow up meetings of national ATG networks. These institutes were chosen as a result of rigorous screening process and each had a proven track-record of providing high quality training for the particular courses taught. 41. The design was also highly efficient. Cost savings were made as a result of (i) leveraging existing expertise of institutions and (ii) working with regional institutes. Implementation costs would have beeri higher if the project had selected only one institution to deliver all the courses, due to potentially high curriculum development costs. Similarly, high implementation costs would have been incurred if the project had selected institution(s) outside of the region, due to high travel and accommodation costs (more details in Section 3.3). 42. The ToT approach adopted by project design is particularly appropriate. Given the daunting human capacity shortage in the three client countries' healthcare sectors, ToT approach was the optimal design to maximize the impact of a modest operation of US$ 10 million. 43. The involvement of recognized and reputable institutions in the region was a further positive design aspect. The high quality training delivered through multi-country institutions to healthcare practitioners from three countries, the flexibility in incorporating new content, direct applicability of the content to trainees' work and the ToT model are key design features which are 8 highly relevant. Both the external project evaluation and interviews conducted for this report indicate that beneficiaries and stakeholders unanimously praise the regional nature, technical and practical strength of ARCAN. 44. Given that both project's design features and its objective of increasing the capacity of healthcare workers are highly relevant in 2010 the project design is rated substantial. 3:2 Achievement of Project Development Objectives 45. Quantity and scope of capacity building efforts. ARCAN achieved its targets towards the goal of increasing capacity of health workers: It fulfilled 90% of the target set for courses to be held by the project (112 courses held out of a target of 125) and realized 85% of the target for number of participants in trained (2463 11 participants trained out of 2884 target number). Overall, the project trained 2463 participants from Ethiopia, Kenya and Tanzania through 112 sessions in six institutions of excellence. 12 These trainees now work in 1,006 service delivery points across three countries and have also provided training to 67,393 other health workers through cascading (12,870 through structured training, 54,523 through on the job training or awareness creation). 46. In addition to direct outputs such as the number of A TGs trained, the PCV data on external resources mobilization by A TGs for cascade training is also useful in assessing progress towards achieving objectives. ATG self reporting shows that 848 ATGs (49% of the total ATGs) have mobilized $3,599,988 from various sources for cascade training i3 which were used to train the above"mentioned 67,393. This means that ATGs have raised an average of$3,988 and trained 87 people with a unit cost of $46. Moreover, beneficiary surveys conducted for this report with 50 A TGs, and the external project evaluation indicate that these trainings made tangible behavioral and institutional changes in healthcare service delivery. 47. Quality of capacity building efforts. The external evaluation conducted in March 2010 which analyzed project outcomes indicates a (self reported) 98% ATG satisfaction rate with ARCAN training due to their high quality. When asked by the external evaluators in 2010 to evaluate the ARCAN training, graduates gave an overwhelmingly positive response (Figure 1). Similarly, when asked if the ARCAN training improved their knowledge and skills, almost all " ATGs responded that it was "fully satisfactory" or better (Figure 2). While these measures of quality are based on self-assessment, a more objective measure of quality is the ability of the A TGs to mobilize resources for cascading. The competitive review processes for resources have their own quality screening and the successful resource mobilization efforts provide evidence of reasonable level. of quality of training. II 1,721 individual health care workers, 742 of whom received a refresher course with updated curricula" 12 The trainees were physicians, nurses, counselor supervisors, laboratory technicians, and HIV/AIDS program managers. Additional training added at the later part of the project include TB/HIV co-infection management, and procurement and supply chain management. Level I training (1,721 trainees) followed the training-of trainers (TOT) approach in order to maximize the impact of the capacity-building project, while level II (742 trainees) were refresher courses aimed at recalling and reinforcing previously acquired knowledge and skills. 13 Sources include the Tanzania MAP, the Tanzania National AIDS Control Program and Malaria Control Program [as part of The Global Fund to Fight AIDS, Tuberculosis and Malaria], and Family Health International (FHI) and Centers for Disease Control (CDC) in Ethiopia; [US] President's Emergency Plan for AIDS Relief [PEPFAR], Global Fund Kenya Country Coordinating Mechanism, UNAIDS and professional vocational organizations such as the National Nurses Association in Kenya. 9 48. In addition to self reporting by ARCAN trainees, the adoption of ARCAN curriculum by a number of higher training institutions, such as the Zanzibar College of Health Sciences and the Dar- Es Salaam based Aga Khan School of Nursing, also points to the quality of training. Figure 1: ATG satisfaction with training quality Overall Rating of the TOT Training by TOT Type Physician Nurses Management Counsel Lab TB/HIV Procure o 20 40 60 80 100 Percent Figure 2: Use of training skills and knowledge by ATGs ATG Response to Knowledge and Skills Acquisition of Knowledge and Skills In the TOT Training 100 PhysicIan 97 Nurses 98 Management 100 Counsel 100 Lab 1nn TB/HIV 100 Procure o 20 40 60 60 100 Percent Knowledge Skills Set 49. Capacity building efforts and service delivery. Self reporting by ATGs indicate that, ARCAN also contributed to increased quality in health service delivery of its graduates, namely through the introduction of various new forms o( counseling including adherence counseling, improvement of HfY case management, feedback mechanisms, non-judgmental attitudes, provision of information on rational use of ARVs, and treatment literacy. Increase in the number of patients receiving services has also been reported, where ATGs attributed this trend to the emergence of more positive attitudes by the service providers. While the project's contribution to 10 the overall increase in the quality of national healthcare systems is not measurable, ARCAN contributed to an increase in the quality of healthcare provided at the individual level. 50. Equally importantly, ARCAN caused significant behavior change among health care workers trained. Over 90% of the 300 ATGs contacted for the external evaluation agreed that ARCAN courses had influenced their behavior in health care delivery. Specifically, they noted that there has been stigma reduction, improved ability to abide by ethics and accept feedback. According to the external project evaluation report, "the ATG training did impact on A TGs' job satisfaction, behavior and attitude to work, service delivery, and job and organizational performance. The training also significantly improved the competency of ATGs with respect to training skills." These findings correlate with the interviews conducted with randomly selected A TGs for this report. One A TG reports that "the ARCAN TOT Nursing training which I attended was a turnaroundfor me. It enhanced my knowledge about HIV and AIDS and clarified so many things that I do not know before. I have since offered training on Basic HIVIAIDS Workplace Prevention at four factories and transport companies. I also participated in trainings organized by AMREF and have been given a reference letter for my professional capability and skills in using participatory training methods. " 51. The project also helped increase the capacity of regional training institutions. Discussed in detailed in the Institutional Change/Strengthening section of this report, all six training institutes which delivered ARCAN curricula reported capacity improvements. The Ethiopian Health and Nutrition Research Institute (EHNRI), for instance, used the ARCAN curriculum and model to conduct an international training for the WHO Afro-Malaria program. The All Africa Leprosy, Tuberculosis and Rehabilitation Training Centre (ALERT), Ethiopia changed a TBIHIV training program to TOT TB/HIV Training using the ARCAN curriculum, to promote cascading and widen the reach to other health workers. 52. When compared with other active programs in HIV/AIDS field in all three countries -to varying degrees- such as the Global Fund or The US President's Emergency Plan for AIDS Relief (PEPF AR), the project's size is dwarfed. Yet given the fact that none of these programs were active in the project countries when ARCAN became effective, stakeholders (i.e. national AIDS commission/councils, training institutions and trainees) observe that the project was one of the first efforts to increase the capacity of health workers and that it did a lot with a relatively small budget. 53. These substantial results, however, were achieved despite the following three moderate shortcomings: (i) Fundingfor cascade training by ARCAN trained graduates (ATGs) through national MAP projects. The ARCAN design placed special emphasis on cascading: it was through cascading that the multiplier effect would· be created and large numbers of healthcare workers would be trained on key HIV/AIDS issues in a cost effective way. Project design gave the national governments the responsibility to finance the vital cascade training and pointed to the national MAP projects as the source of financing. All three countries accepted this responsibility through letters they provided to IDA at appraisal. Yet, in reality, the funding never materialized except for a small number of cases in Tanzania, where 57 ATGs received $200,000 MAP funding for cascading and subsequently trained 855 other health workers. This amount is dwarfed by the $3,399,988 A TGs raised from non-MAP sources (Figure 3). 11 Figure 3: Composition of cascade funds Funds Raised by ATGs for Cascading $4,000,000 $3,000,000 $2,000,000 $1,000,000 $0 Total funds raised Non-MAP funding MAP funding The MAP financing remained low and came only from one of the three client countries (Tanzania) because by the time ARCAN produced its initial group of graduates who applied for cascade funds in late 2007, the three countries had already largely committed their MAP project funding. (ii) Coordination with emerging national healthcare ToT programs. At appraisal, ARCAN was the primary ToT program for HIV/AIDS healthcare workers in the project countries. However, a number of national programs run by national AIDS commissionsl councils (NAC) emerged during project implementation. In Tanzania, for instance, two ToT programs on HIV/AIDS counseling were offered following the launch of ARCAN: One by the Africa Medical Research Foundation and one through the National AIDS Control Program (NACP). Aside from the potential inefficiency created by similar projects running simultaneously, the increasing numbers of HIV I AIDS ToT projects also resulted in a competition for funds for cascading between graduates of ARCAN and those of other programs. (iii) The MoH endorsement ofATGs in a timely manner. The three countries committed themselves to involving ATGs in the national programs and linking them with the national human resource management plans. However, the project graduates did not receive Min istry of Health endorsements from the beginn ing because the project design did not include MoH endorsement or integration of ATGs into national human resource plans. The fact that ARCAN trainees were not endorsed by their respective Ministries of Health brought their ToT status under question by the various cascading financing sources. The SC attempted to address this challenge by having National AIDS Control Programs (NACP's) send ATG endorsement letters to regional public health managers in mid-200B. However, the regional public managers did not include the A TGs in human resource management plans as the plans were mostly complete by the time the letters reached their destination. Although not official MoH certificates, the letters nonetheless provided a seal of endorsement for ATGs and helped them address the ToT status question posed by the various cascading financing sources after mid-200B. 54. These limitations, although not significant enough to prevent the achievement of project objectives substantially, nonetheless had an impact .. Due to these moderate shortcomings, the project efficacy is rated modest. 12 3.3 Efficiency 55. As per standard practice in 2004, no formal economic or financial efficiency analysis was done at appraisal stage. The project documents referred to the economic analysis of the MAP, which justified an overall HIV/AIDS intervention scheme and did not single out capacity building. Given this background, this report considers four major factors for assessing ARCAN's efficiency; (i) implementation; (ii) project returns, (iii) gains to the sector and (iv) project execution. 56. First, the regional nature and the multi-country execution brought significant efficiency gains in the training, the biggest of which was attained though the use of existing traini"ng institutions located in the three project countries. This provided cost savings in logistics terms; i.e. instructor vetting, travel and lodging for ATGs. Another gain brought by the regional nature of ARCAN is the exchange of best practices across countries during training. These elements, which could not be attained through single country interventions, increased the project's effectiveness. Another aspect of execution worth noting is the curriculum templates given to the trainees. These templates are used in cascading training and they reduce the cost of tailoring cascade trainings for specific audiences. 57. Secondly, ARCAN's returns on investment are noteworthy: US$ 8.7 million out of the total project cost (US$ 10 million) was spent on trainings (Annex 1). Out of US$ 8.7 million, US$ 5.07 million was spent on curriculum development and course delivery by partner training institutions (less than the original US$ 5.8 million estimated in the PAD) and the remaining US$ 3.63 million was expensed towards ATGs travel and logistics. With US$ 8.7 million, 2,463 14 units of trainings were provided, short of the original target of2,884 15 • The reason for the actual value to be 85% of the targeted is the increase in training costs by about 32% from what was budgeted in 2004 to the project implementation period of 2006-20 1O. Given the increase in costs, the training output delivery was commendable. Out of the US$ 3,532 16 unit cost of training, US $2,058 was for curriculum development and course delivery, while US$ 1,474 was for travel, accommodation and per diem. Subsequently after their ARCAN training, ATGs raised a total of US$ $3,599,988 producing a 36% leverage ratio. With these funds raised, they trained 12,870 health professionals through cascading and reached out to 54,523 others through awareness creation activities and on the job training. When the number of workers reached through cascade and on the job training, the project unit cost expense is further reduced to US$ 596 and US$ 126 respectively (Figure 4). 14 1,721 Level I trainings, 742 Level II trainings. Level II trainings were refresher courses aimed at recalling and reinforcing previously acquired knowledge and skills. 15 This target was set in April 30, 2005 shortly after project launch. 16 Calculated by dividing US$ 8.7 million spent for 2,463 units of training. 13 Figure 4: Project cost perspectives 69114 $3,532 2463 ARCAN Training ARCAN training + formal cascade ARCAN training + formal cascade + informal cascade _ Unit cost per training Number of trainings held 58. Third element is the project's gains to the sector, namely the complementary nature of ARCAN's capacity building in the fight against HIV/AIDS. Most global and domestic funds for fighting the epidemic are allocated to more quantifiable projects such as ARV provision or hospitallclinic construction, while addressing human shortages remains underfunded. By investing in the capacity building of workers who will deliver the services with higher efficiency and quality, the project also played a role in helping build long-term sustainability of other HIVI AIDS programs and complemented the national and international efforts, thereby increasing the chances of success of other projects that rely on human capacity to provide HIVI AIDS healthcare services. 59. The fourth factor is project execution: Disbursement was on track. Despite initial delays, the disbursement profile suggests that in the first year of the project (end- 2005) disbursement caught up with projected; During later stages, actual disbursement amounts trailed closely the original amounts: By MTR in October 2007, actual disbursement ofUS$ 6.2 million was slightly behind the projection amount ofUS$ 6.9 million. 100% disbursement was reached in five years, and project was extended for 10 months, which allowed the completion of the external project evaluation. Project supervision was particularly efficient as total Bank management costs were about $470,000179 staff weeks for supervision. 60. Efficiency gains attained though (i) the use of existing regional training institutions, (ii) considerable returns on investment indicated by 2,463 direct trainings delivered with a unit cost ofUS$ 3,532 and US$ $3.6 million funds raised after the trainings, producing a 36% leverage ratio and helping train 67,393 other healthcare workers, (iii) project's role in complementing national and international efforts by investing in local human capacity and (iv) efficient project execution, efficiency is rated substantial. 3.4 Justification of Overall Outcome Rating 61. Rating: Moderately satisfactory. The rating is based on substantial relevance, modest achievement ofproject's objective and substantial efficiency. Relevance rating is based on objectives, design features and implementation modalities that are highly relevant in 2010. The rating on the achievement of project objectives is the result of high quality training provided to a significant group of healthcare workers and the subsequent change in their behavior in HIVI AIDS healthcare provision, moderately limited by implementation challenges. Finally, the efficiency rating is based on significant economic and human resource gains attained by the project. 14 Table 2: Breakdown of project outcome rating Project Objective Relevance Efficacy Efficiency Outcome Rating Expand access to comprehensive and evidence-based HIV/AIDS prevention, treatment, and care programs by Moderately Substantial Modest'7 Substantial supporting a Satisfactory network of subregional learning sites to expand training of health care practitioners. 3.5 Overarching Themes, Other Outcomes and Impacts (a) Poverty Impacts, Gender Aspects, and Social Development 62. The project was gender sensitive and nearly 50% of A TGs were females (844 female, 877 male). The external project evaluation also indicates that there has been an increase in the number of clients or patients obtaining services from institutions where ATGs serve and the quality of HIV /AIDS healthcare they receive. This is attributed by the trainees to the development of a more positive attitude, ability to tackle stigma and discrimination and enhanced capability to deliver qualitative services, obtained as a result of the ARCAN training, though it should be noted that some ATGs also attended other training. Another significant impact of the project is the increase in the job satisfaction and career advancement of its trainees. Interviews conducted for this report and the external project evaluation show that most ATGs pointed to ARCAN as a "turning point" in their careers. (b) Institutional Change/Strengthening 63. The project contributed significantly to the following areas of institutional strengthening: 64. Strengthening of regional institutions: ARCAN trainings were delivered to physicians, nurses, counselors, lab technicians, procurement and supply officers, and health program managers by the following entities: Kenya Medical Association; Aga Khan School of Nursing, Kenya; Kenya Association of Professional Counselors; Ethiopian Health and Nutrition Research Institute (ENHRI), and Eastern and Southern Africa Management Institute (ESAMI), Tanzania .. The All Africa Leprosy, Tuberculosis and Rehabilitation Training Centre (ALERT), Ethiopia also delivered the TB-HIV/AIDS course to a mix of these professionals. The project contributed to improvements in these institutions by (i) assisting them in the development of toolkits or operating manuals; (ii) establishing new relations among them, or reinforcing existing ones, that 17 Substantial achievements with moderate shortcomings 15 lasted beyond ARCAN activities, and (iii) paving the way for work with other donors. The Aga Khan School of Nursing, for instance, received EU funding after ARCAN to develop a masters in nursing curriculum, while Kenya Association of Professional Counselors (KAPC) was contracted to provide the course in Rwanda and Botswana. 65. Dissemination ofARCAN curriculum: A number of schools of medicine in the region adopted ARCAN curricula (with modifications). Among these, the College of Health Sciences, Zanzibar adopted the ARCAN HIV / AIDS curriculum in its nurse, midwife, clinical officer, dental assistant and public health nurse trainings, while the Aga Khan School of Nursing, in cooperation with ESAMI, developed the ARCAN course for nurses to a bachelor degree level. (c) Other Unintended Outcomes and Impacts (positive or negative) 66. The two key un intended outcomes include (i) formation of national ATG networks, and (ii) significant institutional changes (sections 2.5 and 3.Sb). 3.6 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops 67. A series of meetings in Addis Ababa, Nairobi, Dar Es Salaam and Zanzibar organized as part of the ICR mission discussed the project with a total of 50 ATGs. Key findings include: (i) satisfaction with the training material content and equally overwhelming frustration with the lack of cascade funding; (ii) impressive level of entrepreneurialism from A TGs in raising alternative funds for funding cascading; (iii) national ATG networks should be sustained to help raise funds for cascading; (iv) concern over the material "getting older" over time and loss of cutting edge knowledge, and (iv) significant willingness to engage in a potential follow on phase, ARCAN II. 68. Meetings with stakeholders (training institutions and SC members) found: (i) training content was cutting edge; (ii) failure to fund cascading and integrating ARCAN into national HIV / AIDS programs were caused by the lack of country ownership, high turnover among SC members and lack of strategic leadership from the Bank, and (iii) capacity building in HIV/AIDS service provision is still a major gap in the countries and there is need for an ARCAN phase II, with design adjustments. 4. Assessment of Risk to Development Outcome Rating: Moderate. The two key risks to the project's objectives are: 69. Cascade training slowing down: During implementation, ATGs provided training to 67,393 other health workers through cascading, mostly thanks to personal fund raising efforts. There is a risk that momentum for cascading will be lost over time. However, that risk is largely mitigated by the fact that cascading never relied on project funding and was self-sustaining from the onset. Financing for cascading was not dependent on the project in the first place, and this increases the change that cascading will continue beyond project closing. Also, the establishment of national ATG networks, whose primary goal is to assist its members in raising funds for cascading, is another factor mitigating this risk. 70. Outdated ARCAN content: Prevention, treatment and care ofHIV/AIDS is a fast evolving field, where new developments, research and methods emerge rapidly, often making former methods and teachings redundant. Hence, similar to most other training, particularly in the field of HIV / AIDS, ARCAN's content faces the risk of becoming. outdated over time. However, the risk is mitigated by the involvement of various training institutes in delivering the training beyond ARCAN which will ensure that the material will be updated regularly. Additionally, the intrinsic value of ARCAN training as a whole is not limited to the HIV/AIDS technical component. Skills acquired by A TGs such as proposal writing, fund-raising and exchange of best practices are 16 among elements which are included in the ARCAN training. Complemented by the behavior change observed among trainees, these elements will continue to be valid. 5. Assessment of Bank and Borrower Performance 5.1 Bank Performance (a) Bank Performance in Ensuring Quality at Entry 71. Rating: Moderately unsatisfactory. The human capacity shortage for HIV / AIDS service delivery was a significant bottleneck at project appraisal, and the limited number of global actors with know-how and financing brought the Bank under considerable pressure to deliver. While clients and the Bank held consultations during project appraisal, inadequate experience in managing regional HIV / AIDS projects undermined the quality of analysis for determining the implementation arrangements as well as for the risk assessment. The Bank also fell short in ensuring a robust M&E framework (as discussed in Section 2.3). (b) Quality of Supervision 72. Rating: Moderately unsatisfactory. The Bank supervision quality was erratic. At critical junctures, particularly at project start, the sector management failed to provide the needed strategic problem solving assistance in a timely fashion to address operational challenges and tensions between the Bank team and other project parties. In later stages, the Bank responded with -sometimes significant- delays to absence of objection requests, which delayed procurement and training activities. ISR ratings were generous and, in the case of the DO ratings, could have been better supported by empirical evidence. Although M&E shortcomings were identified and repeatedly highlighted during supervision, the Bank failed to bring solutions. The Bank could have done a better job in ensuring consistency of supervision quality through project implementation. 73. On the other hand, the Bank's reporting and follow-up of agreed actions were adequate. While problematic at project start, compliance with covenants became satisfactory in the later stages of implementation. Supervision quality improved first with the involvement of country office- based health specialists after the MTR, and then with the new task team leadership toward project closing. At this stage, the Bank improved its relations with the clients, conducted a third party evaluation of the project and supported national ATG networks to ensure adequate transition for post-ARCAN cascade training. 74. Quality of supervision is rated moderately unsatisfactory because task team leadership and sector management attention were not adequate and although supervision quality significantly improved, it came late in the project life. (c) Justification of Rating for Overall Bank Performance 75. Rating: Based on the moderately unsatisfactory quality at entry and moderately unsatisfactory quality of supervision, overall Bank performance is rated moderately unsatisfactory. 5.2 Borrower Performance (a) Government Performance 76. Rating: Moderately unsatisfactory. The three governments showed strong engagement, as evidenced by the level of representation at the project SC and the frequency of their meetings. However, the ability to effect change and/or deal with implementation issues was varied at best. 17 The three strategic challenges that tested the performance of ARCAN countries were: (i) Cascade funding; (ii) integration of ATGs in the national HIV/AIDS training programs, and (iii) coordination with other projects which emerged during ARCAN implementation ran similar training programs. 77. Firstly, out of the VS$ 3 million raised for cascading, only VS$ 200,000 came from government sources and that was only in Tanzania. Kenyan and Ethiopian authorities failed to provide cascade funding altogether. In most cases ATGs never received an official response to their application for cascade funding from their respective NACs. Secondly, although the governments (through representatives in the SC) committed themselves to improving the involvement of ATGs in the national programs by linking them with the national human resource management plans, this to a large degree did not happen. Although NACPs sent ATG endorsement letters to regional public health managers, most ATGs continue to suffer from non- recognition by the Ministries of Health. Finally, a number of projects similar to ARCAN emerged during project implementation, some of which were government initiatives. Governments failed to ensure that there was no replication of efforts among these projects, causing ATGs to compete with the graduates of other programs for cascade funding and recognition. 78. Collectively, these point out low government ownership of the project and less than full commitment to achieving its development objectives. (b) Implementing Agency or Agencies Performance 79. Rating: Satisfactory. The project coordination unit (PCV) was the implementing agency and there were no shortcomings in its performance. Since its entry into the project in mid-2005, the unit showed significant commitment to achieving the development objectives. It conducted regular and frequent consultations with the ATGs, training institutions, SC and the Bank team to ensure full coordination and proactive diagnosis of any challenges to come. The PCV signaled implementation challenges in a timely fashion and proposed appropriate and realistic choices for solutions. The M&E system maintained by the PCV captures important project information, and is crucial in supplementing the official results framework in assessing project performance. As acknowledged by all parties, the PCV "kept the project moving", and even at times while in an "orphan" status, with little support from Bank management and task team leadership. (c) Justification of Rating for Overall Borrower Performance 80. Rating: Moderately Satisfactory. The borrowers failed to tackle larger strategic issues, and did not demonstrate strong and consistent ownership of the project. However, due to high PCV performance, activities were implemented with high beneficiary satisfaction, operational issues were addressed and key project progress information was captured. 6. Lessons Learned • Enhance outreach to rural practitioners, especially in the public health offices. This group of health care workers was underrepresented among ATGs, due mainly to the lack of internet and other communication services. Similar projects that depend on applications for selecting beneficiaries should develop specific tools for this segment, such as radio or regional print media, where available, as well as mobile phone technology. • Heterogeneity in audience exposure to HIVIAIDS service delivery caused challenges. Some ATGs had extensive experience in delivering these services, whereas others were more novices, which occasionally caused frustration during trainings. A two tiered structure where novice trainees receive a period of intense "orientation" would help alleviate this challenge. 18 • Respond to changing technical landscape. At project design, the Bank was the leading actor in the field and national systems were not mature. Future projects in addressing human capacity issues should leverage the emerging national capacity and efforts of other partners such as PEPFAR or the Global Fund by (i) embedding operations in national targets, and (ii) harmonizing efforts. Namely, training programs supported by the Bank and other donors should be aligned with existing national human resource development targets, and project indicators should be fully harmonized with national HIV/AIDS indicators and targets. • Working with regional institutions of learning worked. This allowed the flexibility of updating course content with cutting edge information, kept the costs low and strengthened these institutions. Operations that aim to build human capacity should continue this good practice, where feasible. • The regional approach was a hig win. This allowed the exchange of best practices across different countries, established relations between practitioners and most importantly, responded appropriately in helping alleviate the regional epidemic. • Keep M&E tight. Project development objective should be focused and specific. M&E should be fully integrated with national efforts, and should capture information about the experiences of trainees. This will help indicate what works and direct emerging efforts on human capacity building in HIV /AIDS. • Human capacity in HIVIAIDS is still an important gap in the region. Although the project trained a substantial number of health workers, the changing nature of the epidemic and the continuing gaps in national capacity mean the Bank and other partners should continue to support these national efforts. . • Select Bank task team leader carefully. Although now a cliche, this issue was key in determining the quality of Bank contributions in the project. Over the course of implementation, the Bank team was led by three different TTLs, and the quality increased only when (i) TTLs focused efforts and energy and engaged the client, and (ii) were located physically in the project area. • Bank's internal structures discouraged the much needed restructuring. As indicated in numerous Aide memoires, ISRs and other reports, the PDO, KPI and intermediate indicators was in grave need of revision. Yet the already overstretched team could not afford the extensive time needed to restructure the project. Bank task teams, of particularly smaller operations, face this challenge frequently. The recent Bank-wide simplifications to the restructuring process are therefore welcome. 7. Comments on Issues Raised by Borrower/lmplementing Agencies/Partners (a) Borrower/implementing agencies Dr. Meshesha Shewarega, Executive Director, Christian Relief and Development, SC Member Association (CRDA), Ethiopia "[ ... ] the report looks appreciative of our joint efforts and contribution in capacitating various actors in fighting this disease and pioneering this innovative program. The overall verdict as to our performance is very clear: Moderately satisfactory. The lessons are clearly articulated. One pronounced suggestion made at different forum did not stand very clear: The need for ARCAN II. This suggestion has come out very clear in the final stakeholders meeting. This is the only serious comment that I would like to raise. Could this be captured in a very clear manner? [ ... J" 19 Prof. Alloys S. S. Orago, Director National AIDS Control Council, Kenya, SC Member "The ICR appears good and well written." Dr. Rowland Swai, Program Manager, National AIDS Control Program, Tanzania, SC Member "[ ... ] it' [the ICR] is a fair record of what was done during the lifetime of ARCAN, but the only concern is that it is silent about ARCAN II. Can this point be corrected - mention about the desire to have ARCAN II and also the report should give an opinion on the request about ARCAN II." 20 Annex 1. Project Costs and Financing (a) Project Cost by Component (in usn Million equivalent) Appraisal Estimate Percentage of Components (lJSn millions) Actual (lJSD millions) Appraisal Healthcare Practitioner 7.72 8.70 113 Training Monitoring and Evaluation and Knowledge Sharing 0.38 0.00 18 o Program Coordination 1.50 1.50 100 Unallocated 0.40 0.35 19 88 Total Baseline Cost 10.00 10.55 Physical Contingencies 0.00 0.00 0.00 Price Contingencies 0.00 0.00 ....__ ....... 0.00 Total Project Costs 0.00 10.55 Front-end fee PPF 0.00 0.00 .00 Front-end fee IBRD 0.00 0.00 .00 Total Financing Required 10.00 10.55 (b) Financing Appraisal Actual Type of Estimate Percentage of Source of Funds (USn Cofinancing (USn Appraisal millions) millions) Borrower 0.00 0.00 0.00 IDA GRANT FOR HIV /AIDS 10.00 10.55 105.50 18 M&E agency envisioned at appraisal was not hired. The PCU performed most M&E tasks and hired a consultant to maintain a database (costs in included in consultancy services). 19 US$ 28,660.27 was for goods and works for PCU operations and US$ 318,841.00 for PPF refund. 21 Annex 2. Outputs by Component Component 1: Health Care Practitioner Training (US$ 7.72 million) Principal activities Baseline Indicators Result (2010) and outputs (2004) Number of trainees attending first time 0 1721 (Level I) Number of trainees attending follow up 742 training session (Level II) Ethiopia Kenya Tanzania Number, 496 582 643 geographical distribution of Female 844 Male 877 ARCAN trainees and Healthcare workers sector represented at trained Government 1028 the time of primary Civil Society Org. 491 training Private Sector 202 Total number = 1006 294/4073S0Ps in Ethiopia = 7.2% Percent of service delivery points 393/5129 SOPs in Kenya=7.6% (SOP) in trainee's country with one (or 0 282/5379 SOPs in Tanzania ML =5.24% more) primary recipients of 371140 SOPs in Zanzibar = 26.4% ARCAN supported TOT training Data source: UNAIDS distribution of health infrastructure data 2007 Kenya and Ethiopia and Tanzania service provision survey 2008 (TSPA) for Tanzania-M and in Zanzibar data Total 67,393 Healthcare workers Number of trained through health workers o Ethiopia 20,467 cascading trained (cascad in g) Kenya 22,990 Tanzania 23,936 22 Component 2: Monitoring and Evaluation and Knowledge Sharing (US$ 0.38 million) Principal Baseline activities Indicators Result (2010) (2005) and outputs Increase in Increases in the number of training the number Project monitoring reports submitted to the of report 1083 monitoring PCV by ARCAN graduates (compiled submitted annually) annually Project Overall ARCAN Project adjustments or 0 N/A monitoring fine-tuning due to M&E inpueo Total 9 Number of national and regional Knowledge ATG National Forums 4 knowledge sharing workshops N/A sharing ATG Experience Sharing organized 21 Workshops 4 Stakeholder Workshop 1 Information on Knowledge Sharing Activities Representation Number of Type of Objectives of Forum and Attendance Forums Forum at Forum Held • Create practical mechanism for cascading Attended by 200 implementation at national level key ARCAN stakeholders 4 • Promote cascading and effective use of ATGs (through mainstreaming/ integrating their training (total): Jan-Feb '08 ATG activities in existing national systems/programs) • WB country National • Create awareness of importance of cascading, the officers Ethiopia Forum efforts made by ATGs to cascade and experienced • SC members Kenya challenges (to key stake holders) .ATGs Tanzania • Promote ownership and partnership in transferring .ATG (Mainland) the ARCAN courses skills and knowledge to a Employers Zanzibar wider reach of health facilities at all levels • Donor agents • Promote integration of ARCAN TOT trainings in • PTI trainers 20 This indicator was in the results framework, however, was not tracked by the project M&E system. 21This indicator was not part of the results framework. It is presented here to give the full picture of project's efforts for knowledge sharing 23 the national training and other capacity building programs • Bring together ATGs at country level to exchange information, share experience and learn from each other • Discuss and share information on progress' of implementation of planned ATGs training activities as planned during the TOT courses. (Follow up of implementation of cascading plans as prepared by 4 the ATGs) National Feb-Sept '08 ATG • Update ATGs on efforts made by ARCAN and national agents in promoting cascading Attended by 577 Experience Ethiopia Sharing • Explore existing cascading opportunities at country ATGs (total) Kenya level and agree on way forward Workshop Tanzania • Discuss and agree on key steps towards formation (Mainland) of National ATGs Network (NAN) at country level Zanzibar - Presentation on efforts made by the NAN Taskforce (per country) • Promote formation of national learning groups (thematic) at country level • Improve ATGs knowledge with regard to network management and resource mobilization Attended by 70 key ARCAN stakeholders (total): • SC members • PTls .ATGs • National ATG I Networks Stakeholder .ATG Workshop • Determine project's lessons learned Employers March ' 10 (Annex 6) • Discuss project's strengths and weaknesses • National • Determine the need for a similar project ARCAN focal Kenya officers • Training needs assessment consultant • External project evaluation consultant 24 Component 3: Program Coordination Component (US$ 1.5 million) Principal activities and Baseline Indicators Result (2010) outputs (2001) Number of audit Audit and monitoring reports performed 0 5 during the project Number of IFRs Reporting performed during the 0 12 project ADDITIONAL INFORMATION ON QUALITY OF CAPACITY BUILDING EFFORTS AND SERVICE DELIVERY Behavior change among ATGs directly trained: 1,721 1. Web survey conducted for the external project evalua~ion with a random sample of 300 ATGs across seven training categories shows (Table 3) that majority of ATGs report ARCAN as having changed their behavior in health care delivery, which specific reference to reducing stigma, abiding by ethics and accepting feedback from clients. Table 3: Participants of the Web survey reporting influence on change in behavior Factor No. (%) Fishers' Exact "l; (P value) Extent ARCAN Training has influenced or changed behaviour in health care delivery Physicians 56/59 (94.9) Nurses 63/65 (96.9) Programme Managers 56/61 (91.S) 4.91 (0.56) Counsellor Supervisor 36/37 (97.3) Laboratory Technician 35/35 (100.0) TBIHIV 29/30 (96.7) Procurement 6/6 (100.0) Extent ARCAN Training has benefited other health practitioners Physicians 51160 (92.0) Nurses 51/64 (SO.O) Programme Managers 47/57 (S1.S) 9.4 (0.15) Counsellor Supervisor 36/37 (90.9) Laboratory Technician 31135 (76.9) TB/HIV Alert 29/31 (91.7) Procurement 5/5 (100.0) 25 2. The survey also indicates that over 77% of respondents agreed that the knowledge acquired from the ARCAN training has benefited other health professionals in their work places. Similarly, all 50 ATGs randomly selected and surveyed for this work reported behavioral changes. The changes vary from the use of a new diagnostics system, to the formation of new professional associations. Behavior change among health workers trained through cascading: 67,393 3. Due to insufficient monitoring of cascading quality and effectiveness, the analysis of cascade trainings relies heavily on the external project evaluation and stakeholder accounts. External project evaluation reports that the cascading trainees noted significant benefits and positive impact on work output, job performance and behavior. Majority of cascade trainees sampled for the evaluation reported high rates of knowledge and skill acquisition (Figure 5). Figure 5: Acquisition of knowledge and skills though cascaded training Trainee Reporting Moderate/Fully Acquisition of Knowledge and Skills in the cascaded Training Nurses Management Counsel Lab TB/HIV o 20 40 60 80 100 Percent IC~~=__ ::l Knowledge _ Skills Set I 4. Sample of cascade trainees reported that the training received from A TGs helped them "appreciate their jobs", and that they "would recommend similar trainings to others" (Table 4). Table 4: Cascade trainees at the workshops reporting appreciation & understanding of job Factor No. (%) Fishers' Exact -i; (P value) Effect of Cascaded course on the appreciation and understanding of the job CASCADED for Nurses 9/9 (100.0) CASCADED programme Managers 111 (100.0) 1.B (1.00) CASCADED Counsellor Supervisor 9/i 0 (100.0) CASCADED Laboratory Technician 6/6 (100.0) CASCADED TBIHIV Alert 111 (100.0) 26 5. The same trainees also reported that they are actively applying the knowledge and skills acquired through cascading (Figure 6). Figure 6: Trainee Reporting Moderate/Fully Application of Knowledge and Skills Nurses Management Counsel Lab TB/HIV o 20 40 60 80 100 Percent In terms of behavioral change, the trainees reported improvements in the areas of training and facilitation, program development and management, HIV/AIDS patient care, feedback mechanism and non-judgmental attitudes. Further, they noted that the cascade training has benefited other health care professionals (Table 5). Table 5: Cascade trainees and change in behavior & benefits' to other health care providers Factor No. (%) Fishers' Exact "i; (P value)_ Extent cascaded Training has influenced or changed behaviour in health care delivery CASCADED for Nurses 9/9 (100.0) CASCADED programme Managers 111 (100.0) - CASCADED Counsellor Supervisor 9/1 0 (90.0) CASCADED" Laboratory Technician 6/6 (100.0) CASCADED TB/HIV Alert III (\00.0) Extent cascaded Training has benefited other health practitioners CASCADED for Nurses 919 (100.0) CASCADED programme Managers 111 (100.0) - CASCADED Counsellor Supervisor 10/10 (100.0) CASCADED Laboratory Technician 6/6 (100.0) CASCADED TB/HIV Alert 111 (100.0) 27 Annex 3. Economic and Financial Analysis Annex 3 provides some background information on larger HIV/AIDS context and· capacity building needs in the three countries to provide context for the discussion of efficiencies at the project level, as well as in the sector. HIVIAIDS Background in Ethiopia, Kenya, Tanzania and Expected Gains from ARCAN Ethiopia is among countries most affected by HIV and AIDS. The existence of HIV infection in Ethiopia was recognized in early 1980s with the first two reported AIDS cases in 1986. Since then, the epidemic has rapidly spread throughout the country. It peaked in mid-1990s and started to decline in major urban areas since 2000 while stabilizing in rural settings. According· to projections based on the single point estimate - a calibrated estimate of ANC sentinel surveillance data up to 2005 - the national adult HIV prevalence was at 2.3% in 2009 with an estimated 1,116,216 people living with HIV I AIDS. Out of the estimated 5.4 million orphans in Ethiopia, 886,820 were orphaned due to AIDS. There were an estimated 44,751 deaths due to AIDS in 2009. The number of AIDS -related deaths would have been much higher ifit had not been for the free ART program which has been scaled up by the government in collaboration with development partners since 2005. The estimated national adult HIV incidence of 0.28% in 2009 translates to over 131,145,000 new HIV infections on top of over 1 million Ethiopians living with HIV. With the current status it is evident that HIV and AIDS remain formidable development challenges to the country. . The Ethiopian Government established the HIV/AIDS Prevention and Control Office (HAPCO) to mobilize multi sectoral and grassroots efforts in the fight against the HIV/AIDS epidemic. The Government's 5-year national HIV and AIDS Strategic Plan (2004 - 2008) for· intensifying multisectoral HIV I AIDS response was built around six strategic issues: capacity building; community mobilization and empowerment; integration with health programs; .leadership and mainstreaming; coordination and networking; and targeted response . . World Bank Country Assistance Strategy (CAS) for Ethiopia (2003) focused on improving human development outcomes, enhancing pro-poor growth, and reducing vulnerability, and also placed emphasis on the need for· capacity building. The strategy indicated that constraints to improving health services lie primarily in the shortage of skilled human resources and identifies HIV I AIDS as the foremost threat to national poverty and vulnerability reduction efforts. Kenya also suffers from a severe and generalized HIV epidemic. In recent years the country has experienced a notable decline in HIV prevalence, attributed in part to significant behavioral change and increased access to ART. National adult HIV prevalence is estimated to have fallen from 10% in the late 1990s to about 6.1 % in 2005. This, however, rose to 7.4% among the age group 15-49 and at 7.1 % among the age group 15-64 based on the 2007 Kenya HIV and AIDS Indicator Survey Report. Women face considerably higher risk of HIV infection than men, and also experience a shorter life expectancy due to HIV I AIDS. The seventh AIDS in Kenya reports an HIV prevalence rate of 8% in adult women and 4% in adult men. Despite recent decline in prevalence, HIVI AIDS continues to be a major challenge in the country's socio-economic development. Since the first case was discovered in 1984, it is estimated that over 1.5 million people have died due to AIDS-related illnesses, resulting into 1.8 million children left as orphans. It is also estimated that 1.5 million people are living with HIV. 28 The Government of Kenya declared HIV/AIDS a national disaster and established the National AIDS Control Council (NACC) in 1999. It then developed the Kenya National HIV/AIDS Strategic Plan (KNASP) 2000-2005, which set out a multisectoral response to the epidemic. The plan identified human resource development and management as one of seven key areas that could create an obstacle for effective implementation of the plan and pointed to capacity building through training was as a vital process for addressing health sector staff need. The Bank's CAS for Kenya (2004) notes that HIV/AIDS is a top priority. In Tanzania, HIV/AIDS adult prevalence has been estimated at 7.0%, with 1.6 million people estimated to be living with the virus as of 2007. This figure may not be truly representative of the HIV epidemic in Tanzania as a high number of AIDS cases are under- reported. The Ministry of Health estimates that only one out of five cases is reported. According to Tanzania National Multisectoral Strategic Framework on HIV/AIDS, AIDS is the leading cause of death among adults in the country. There are 1.3 million adults aged 15 and over living with HIV, of which 760,000 are women. 140,000 children under the age of 14 are living with HIV. The country has an estimated 970,000 orphans due to AIDS under the age of 17. The health sector in particular is experiencing high demand for its services, as AIDS patients occupy an ever-increasing number of beds in hospitals. More than 60% of patients admitted in the general ward are suffering from TB, prolonged fever, headache, skin disease, and diarrhea. This results in increased workload for hospital workers, who are already understaffed. There is a marked deterioration observed in the referral systems due to extra demands for medical care and treatments by AIDS patients. These all have their own recurrent impact on the economy as there is an increase in the recruitment and training costs for health personnel as well as an increase risk of HIV infection in hospitals/HBC due to lack of protective equipment. Tanzania established the Tanzania Commission for AIDS (T ACAIDS) in 2001 and the Zanzibar AIDS Commission in 2002 to lead and coordinate multi-sectoral responses to the epidemic, monitoring and evaluation, including research, resource mobilization and advocacy. Both entities point to capacity building as the main focus of national HIV response. Yet only a fraction of the health personnel have been trained on HIV and AIDS. Against this backdrop of significant pressure on the human resources of three econom ies, ARCAN was developed to provide a model for facilitated knowledge and skills acquisition by personnel critical to HIV /AIDS control in the health sector. The training of nurses, counselors, physicians, laboratory technologists, program managers and procurement specialists was expected to contribute to addressing the human capacity need in the country and bringing significant economic gains. Implementation Efficiency Implementation efficiency refers to economic and financial gains attained by project design and execution. ARCAN's regional nature and multi-country execution brought efficiency gains in training through the partnership with training institutions located in the three project countries. The six partner training institutes developed the training curriculum, executed the training, provided accommodation for participants and took part in follow up meetings of national ATG networks. These institutes were chosen as a result of rigorous screening process and each had a proven track-record of providing high quality training for the particular courses taught (Table 6). Cost savings were made as a result of (i) leveraging existing expertise of institutions and (ii) working with regional institutes. Implementation costs would be considerably higher if the project had selected only one institution to deliver all the courses, due to potentially high curriculum development costs. Similarly, high implementation costs would have been incurred if 29 the project had selected institution(s) outside of the region, due to high travel and accommodation costs. Table 6: ARC AN Courses Delivered by Partner Training Institutes To enable HIV &AIDS program managers to Program Eastern and Southern African improve the design, management and Managers Management Institute (ESAMI) administration of their programs. To facilitate the development of a cohort of nurses with demonstrable HIV related prevention and care skills in the three Aga Khan University School of Nurses countries through cascading by A TGs in order Nursing to provide effective HIV and AIDS interventions for outcomes. To develop a cadre of counselor supervision trainers and enhance counselor supervisors' Kenya Association of Counselors ability to support, supervise and monitor HIV Professional Counselors and AIDS counselors and counseling services. To capacitate the physicians, and other allied health care providers to be able to design, Physicians Kenya Medical Association train, evaluate, and follow up effective ART s. To impart theoretical and practical knowledge Laboratory Ethiopian Health and Nutrition to trainers on laboratory aspects of ARV Technicians Research Institute (EHNRI) therapy monitoring. All Africa's Leprosy, To promote effective management ofTB-HIV Tuberculosis Rehabilitation, TBIHIV collaborative activities and enhance quality Research and Training Centre TB-HIV service delivery. Another significant gain brought by the regional nature of ARCAN is the exchange of best practices among participants from different countries, which ATGs reported as a significant value. A final dimension of implementation efficiency gain is attained through the follow up material given to trainees, such as curriculum templates, which are used in cascade training. These follow up material reduce the cost of tailoring cascade trainings for specific audiences. Project Returns Project returns refer to the financial cost of training. The estimated unit cost of each ARCAN training is US$ 3,532: Total project training cost (US$ 8.7 million), over the number of trainings conducted (1,721 A TGs with a total of 2,463 trainings [1,721 Level I and 742 refresher Level II training]). The project cost was leveraged by the US$ $3,599,988 raised by A TGs to conduct cascade training, producing a 36% leverage ratio. With these funds, ATGs trained 12,870 health professionals through cascade training and reached out to 54,523 others through informal cascade training, such as awareness creation activities and on the job training. With the cascade training, the unit cost of each training decreases to US$ 596 (US$ 8.7 million / [2,463 ARCAN trainings + 12,870 cascade training]). With informal cascade training, the unit cost is further reduced to US$ 126 (US$ 8.7 million / [2,463 ARCAN trainings + 12,870 cascade training + 54,523 informal cascade training]) (Table 7). 30 ARCAN train· ARCAN training + formal cascade 14591 $596 ARCAN training + formal 000 cascade + informal cascade 69114 $126 Gains to the sector Gains to the sector refer to the contribution of the project to larger efforts in the health sector, namely the complementary nature of the project's capacity building in the fight against HIV/AIDS. More challenging to quantify, this gain is acknowledged by most parties interviewed for this report. Most donor funds for fighting HIV /AIDS are allocated to more quantifiable projects such as ARV provision or hospital/clinic construction, while addressing human shortages remains underfunded. By investing in the capacity building of workers who will deliver the services with efficiency and high quality, the project contributed to the long-term sustainability of other HIV/AIDS programs and complemented the national and international efforts, thereby increasing the chances of success of other projects that rely on human capacity to provide HIV /AIDS health care services. Implementation Efficiency Implementation Efficiency is the pace with which the project was implemented, taking into account disbursement and any project extensions. ARCAN disbursement was on track. Despite initial delays, the disbursement profile suggests that in the first year of the project (end- 2005) disbursement caught up with projected. During later stages, actual disbursement amounts trailed closely the original amounts: By MTR in October 2007, actual disbursement ofUS$ 6.2 million was slightly behind the projection amount ofUS$ 6.9 million. 100% disbursement was reached in five years, and project was extended for 10 months. The main reason for the extension was to complete the external project evaluation. Project supervision was particularly efficient as total Bank management costs were less than $400,000179 staff weeks for supervision, which is commendable. 31 Annex 4. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members Responsibility/ Names Title Unit Specialty Lending - from ~@ ...................__. . ......... - ......... _....... . ... ........-_........- _ - -. _.....--........... .........._....... _............. .. ........__ ...... ..... _- .....- .......................__ ........- ... .... , Sheila Dutta Senior Health.Specialist AFTHI Task Team Leader Financial Management Specialist Luigi de Felice (Consultant) Senior Financial Management John Nyaga AFTHV Specialist A. Waafas Ofosu-Amaah Senior Gender Specialist PRMGE Lead Monitoring & Evaluation Susan Stout HDNGA Specialist Dirk Prevoo Operations Officer AFTR2 Elizabeth Lule Adviser HDNHE Albertus Voetberg Lead Health Specialist AFTHI Emmanuel Malangalila Senior Health Specialist AFTHI Anwar Bach-Baouab Lead Operations Officer AFTH3 Gebreselassie Okubagzhi Senior Health Specialist AFTH3 Karen Hudes Senior Counsel LEGAF Pascale Dubois Senior Counsel LEGAF Deputy General Counsel, Elizabeth Adu LEGVP Operations Financial Management Specialist John Cameron (Consul!allt) Institutional Arrangement Specialist Abbas Kesseba (Consultant) Steve Gaginis Finance Officer (LOAG2) _ Rogati Kayani Lead Procurement Specialist AFTPC Senior Financial Management Mercy Sabai AFTFM Specialist Donald Mneney Procu~ement Analy~t___ . AFTPC Monitoring & Evaluation Specialist Kelvin Billinghurst (Consultant) Keith Hansen Manager, ACTafrica AFTHV Shiyan Chao Senior Health Economist ECSHD Cassandra de Souza Operations Analyst AFTHV Therese Cruz Program Assistant AFTHV Nadege Thadey Program Assistant AFTHV Evelyne Kapya Program Assistant AFTHV Chris Walker Lead Specialist AFTH 1 Peer Reviewer Bachir Souhlal Lead Social Development Specialist MNSRE Peer Reviewer Debrework Zewdie Director, Global AIDS Unit HDNGA Peer Reviewer Jonathan Brown Operations Adviser AFTQK Peer Reviewer SupervisionlICR - from ISRs 32 Dominic Haazen . LeadHealth Policy Specialist AFTHE TTL Mehmet Onur Ozlu Operations Officer AFTRL ICR TTL AFTH3- Anwar Bach-Baouab Lead Operations Officer HIS Cassandra De Souza Operations Analyst AFTHE Bella Lelouma Diallo Sr Financial Management Specialist AFTFM Wachuka W. lkua Senior Operations Officer AFTHE Rogati Anael Kayani Consultant (Pro~urement). AFTWR Emmal1 uel G. Malarigalila Consultant AFTHE Prasad C. Mohan Lead IE<:: Specialist AFTRL Anne Muuna Team Assistant AFCEI EvaK. Ngegba Program i\ssistant AFTHE John Nyaga Sr ~inancial Management Specialist AFTFM Jabulani Nyenwa Consultant WBIHD Adyline Waafas Ofosu- Senior Gender Specialist PRMGE Amaah Gebreselassie Okubagzhi Consultant AFTTR Patrick Lumumba Osewe Lead Specialist AFTHE Krishna Pidatala Senior Operations Officer CITPO Dirk Nicolaas Prevoo Senior Operations Officer AFTEN Mercy Mataro Sabai Sr Financial Management Specialist AFTFM Albertus Voetberg Lead Health Specialist SASHN (b) Staff Time and Cost Staff Time and <:ost.(Bank Budget Only) Stage of Project Cycle No. of staff weeks usn :Thousands (including travel and consultant costs) Lending FY03 7 92.89 FY04 39 496.40 FY05 15 62.40 FY06 0.00 FY07 0.00 FY08 0.00 Total: 61 651.69 SupervisionlICR FY03 0.00 FY04 0.00 FY05 22 136.11 FY06 15 115.14 FY07 15 84.08 FY08 10 60.68 FYQ9 17 71.72 Total: 79 467.73 33 Annex 5. Beneficiary Survey Results (a) Meetings held with beneficiaries for ICR City Date Number of participating ATGs Addis Ababa November 23,2010 17 Nairobi November 26,2010 10 Dar Es-Salaam December 2,2010 8 Stonetown December 7, 2010 15 Total number of beneficiaries met 50 (b) Key outcomes of surveys Training Material Content: • 90% found the training content to be useful and applicable to their work. • 74% gained knowledge/skills from the project. • 74% recorded behavior change. • 72% found content different than other trainings.received. The following skills were frequently reported: • Adherence counseling • Joint HIV/AIDS - TB diagnosis • Strategic work planning • Budgeting • Proposal writing • Fundraising • Conducting needs assessment 6. 90% found the training cascade funding disappointing. None of the A TOs in Kenya and Ethiopia received cascade funding from their national AIDS commission/council and they did not receive an official response to their funding request. 7. Despite the lack of official support, 38 out of the 50 ATGs surveyed reported that they provided cascade trainings to a total of3275 other healthcare workers. 12 of the 38 ATGs who cascaded received MAP funding in Tanzania. Others raised funds from other sources, or convinced their institutions to provide space, time and supplies for the trainings. 8. 92% of those surveyed raised the cascading issues as a major risk moving forward and requested that the national A TO networks be sustained to help mitigate against the risk. 80% raised concern over the training material "getting older" over time and loss of cutting edge knowledge. 34 9. When asked if they thought a similar capacity program would be beneficial, 98% said there should be a potential follow on phase i.e. ARCAN II. 35 Annex 6. Stakeholder Workshop Report and Results (a) Organization of the dissemination and stakeholder workshop 10. The following summary was based on presentations made at the ARCAN Regional Dissemination and Stakeholder Workshop held on March 18-192010 in Nairobi, Kenya. Led by the PCU, the workshop was attended by Steering Committee members, partner training institutions, ATGs, national ATG networks, health facilities with ATGs, national ARCAN focal officers, training needs assessment consultant, external evaluation consultant, heads of training institutions in participating countries and the core ARCAN task team members. (b) Strengths and weaknesses 11. The interrelationship among the SC members, health facilities with ATGs and the ATGs attending the workshop, showed the broad satisfaction with this project. Their presentations highlighted the following strengths: ARCAN paved the way for effective collaboration with other facilities such as the Global Fund or PEPFAR, it highlighted capacity building as a focal point of attention for HIV Response, enhanced South-South cooperation and facilitated collective learning through exchange of best practices. 12. Directors of Amana Municipal Hospital (Tanzania), Kenyatta National Hospital (Kenya) and Johns Hopkins Center for Communication Programs (Ethiopia), who are employers of ATGs, reported marked improvement of service provided by their staff after ARCAN training. Specifically, they pointed out improved quality of care, reduced hospital stay, increased patients satisfaction, improved dissemination ofHIV prevention algorithm and introduction of the use of provider initiated testing and counseling (PITC) as correlated with ARCAN training; 13. Presentations converged on the following three points as the project's biggest shortcomings: (i) Cascading model has not worked well, (ii) recognition and of ATGs as ToTs due to lack of endorsement and certification has created additional obstacles before cascading, particularly from a fundraising perfective and finally (iii) lack of a holistic monitoring of ATG activities, (i.e. quality of cascade trainings, post- training behavior change and impact of trainings on healthcare quality) has prevented a fully accurate picture of the project's accomplishments. Partner training institutions have also highlighted flaws within trainee selection, particularly due to varied treatment of ATGs from backgrounds with the same level of instruction. (c) Lessons learned 14. In general, the teams who attended the workshop emphasized the following positive outcomes and lessons: (i) despite being critical of the failures of cascading, training of trainers model in increasing health care worker capacity is considered as a model in three countries, (ii) there is a grave need for in-service training, given high staff turn-over, and (iii) HIV capacity building strategies need to be based on specific priority needs with clear and well defined monitoring indicators. . (d) Main outcome 15. The continued need for the project (dubbed ARCAN II) was highlighted by participants. Source: Compiled from presentation made at the ARCAN Regional Dissemination Workshop, March 2010 36 Annex 7. Summary of Borrower's ICR and/or Comments on Draft ICR Summary Borrower's IeR Assessment of the project's development objective With a project disbursement rate of 100% as at April 2010; 2,463 participants (85% of project target of 2,884) in 112 courses (90% of end of project target of 125) have been trained utilizing 87% (VSD 5.07m. out of planned costs ofVSD 5.82m) of the total project plan for training costs. A network of six sub-regional 'learning sites' have been supported (Partner Training Institutions or PTIs), and over 50% of the ATGs have reported to the PCV that they are actively equipping health workers with HIV/AIDS knowledge gained from the training held by PTIs. As of December 2009,848 (55%) ofthe total 1543 ATGs supposed to have trained/cascaded, have submitted cascading reports which show that they have managed to mobilize and use in their various forms of training, VSD 3,599,988. The geographical coverage of ATGs in each country was over 90% serving a total of 1,006 health service delivery points in the three countries (that is over 280 service delivery points in each country- which is more than 7% of all service delivery points in Ethiopia and Kenya, and more than 5% in Tanzania) thereby contributing to enhanced service delivery of HIV and AIDS related services and forging stronger linkages between prevention and care interventions. South-South partnerships have been expanded directly amongst the ATGs, PTIs and the national AIDS institutions represented in the Steering Committee, and indirectly to other specialists who worked on project activities such as procurement and consultancies. Consequently, the project has been on track in meeting its project development objective. More detailed information on the PDO can be seen in the Annexes which all address a certain aspect of the PDO. Challenges The PDO was problematic. The PCV and SC recognized within the first year of project i"mplementation that parts of the PDO were too broad and not measurable with existing PAD indicators and were beyond the project mandate thereby requiring system-wide interventions, namely the objectives on "enhancing the delivery of HIV and AIDS-related services; strengthening linkages between prevention and care interventions; and expanding South-South partnerships"; The SC also noted that there was need to include the quality of the training in the PDO. This was discussed with the World Bank during the mid-term review but it was agreed that the PDO remains as it is as there would be a lengthy process to change it and the project had already spent more than 50% of the project expenditure and had less than two years to completion. Lesson Learnt There is need for the PDO to be realistic and in line with the project design. This needs to be done during project design and if challenges are experinced during project implementation, then there is need to change it through quick decision making and follow through. Project Operations Manual and Project Implementation PCV has executed the project in line with the Project Operation Manual (POM) which was in draft for two years. The POM was a condition for project effectiveness of the Development Grant Agreement. All the project manuals had not been cleared by the World Bank when the current pev joined in May 2006. The PCV had to update the draft manuals for clearance in December 2006. These were all cleared by July 2008. Challenges 37 Delay in the approval of the Project Operational Manual contributed to implementation challenges and confusion at the beginning in various areas such as what is the per diem payable to the SC and PCV. The delay of project implementation in 2005-06 was a missed opportunity to integrate the ARCAN cascading plan into the country MAP I programs in the three countries. The Project was also not launched and development of the communication strategies was delayed. As a result the Project was not widely known at national level and contributed negatively in the national cascading efforts Lessons learnt It is important that all conditions of project effectiveness are met before the start of project implementation. This will ensure that there is clear understanding by all parties on their roles and hence reduce the risk of project closure/delays and related implementation activities. 38 Annex 8. Comments of Cofinanciers and Other Partners/Stakeholders 39 Annex 9. List of Supporting Documents The United Republic of Tanzania, 2005. National Strategy Jor Growth and Reduction oj Poverty (NSGRP). Republic of Kenya, National AIDS Control Council, 2009. Kenya National AIDS Strategic Plan 2009110- 2012113. Nairobi. UNAIDS/WHO, 2005. AIDS Epidemic Update, Report No UNAIDS/05.19E, Geneva. Federal Democratic Republic of Ethiopia, Ministry of Finance and Economic Development September 2006. PlanJor Accelerated and Sustainable Development to Eradicate Poverty (PASDEP) Jor the Period 2005106-2009110. Addis Ababa. World Bank, AIDS Campaign Team for Africa (ACT Africa), 2008. TB and HIVIAIDS Integration in Ethiopia, Kenya, Tanzania and Eritrea. Washington DC. World Bank, 2006. The World Bank's Commitment to HIVIAIDS in Africa: Our AgendaJor Action, 2007-2011. Washington DC. World Bank, various years. A ide-memo ires a/supervisions missions/or ARCAN, between 2005-2010. World Bank, various years. Implementation Status and Results ReportsJor ARCAN, between 2005-2010. World Bank. 2004. Project Appraisal DocumentJor the African Regional Capacity Building NetworkJor HIVIAIDS Prevention, Treatment and Care, Report No 29846-AFR. Washington DC. ITAD, 2010. The Impact Evaluation ojAfrica Regional Capacity Building NetworkJor HIVand AIDS Prevention, Treatment and Care Project (ARCAN) in Tanzania, Kenya and Ethiopia. World Bank. 2008. Country Assistance Strategy Jor the Federal Democratic Republic oj Ethiopia Jor the Period FY08-FYII. Report No 43051. Washington DC. World Bank. 2010. County Partnership Strategy Jor the Republic oj KenyaJor the Period FYI 0- FY13. Report No 52521-KE. Washington DC. World Bank. 2007. Joint Assistance Strategy Jor the United Republic oJ TanzaniaJor the Period FY07-IO. Report No 38625-TZ. Washington DC. World Bank, 2005. The World Bank's Global HfV/AIDS Program of Action. Washington DC. 40 MAP SECTION 32°E 36°E 40°E 42°E 44°E ERIT REA To Keren R REP. ET H IOP IA e OF SELECTED CITIES AND TOWNS d ETHIOPIA To Humera Adigrat YEMEN REGION CAPITALS S Gedaref ek T 14°N eze Axum 14°N NATIONAL CAPITAL D D e T I G R AY en a RIVERS Mekele ak Ras Dashen Terara (4620 m) MAIN ROADS ki Atb il ara RAILROADS Gonder De De REGION BOUNDARIES AMHARA A FA R Lake INTERNATIONAL BOUNDARIES se Dinder Tana 12°N Debra 12°N Tabor r rt Weldiya Bahir Dar DJIBOUTI DJIB Asayita 46°E 48°E ue Bl Nile n Dese o f A d e Abay G u l f b S UD AN Awa Debre BENSHANGUL w sh er Markos ng Asosa Ethiopian Ha 10°N 10°N Plateau DIRE DAWA Dire Dawa Didesa To Harer Hargeysa Gimbi Nekemte ADDIS ABABA ADDIS HARARI Jijiga ABABA Awash IA S OM AL I A Nazret R Baro Aware am a Welkite Degeh Bur O R O M I YA is Gore Gambela 8°N Domo 8°N GAMBELA y Asela e Ak abe Shebele l l o Jima Hosaina W Oga d e n bo a Bonga V Shashemene Goba SOMALI Warder Sodo Awasa Dodola Kebri Dehar f t Wendo SOUTHERN NATIONS, W Imi R i NATIONALITES ab e Ge t 6°N AND PEOPLES e sr Wa 6°N ro be t Shebe le e a 0 50 100 150 200 Kilometers Genale G r Negele 0 50 100 150 Miles Ferfer To Yavello Dawa Mogadishu Dolo Odo This map was produced by the Map Design Unit of The World Bank. IBRD 33405 R1 4°N Lake Mega The boundaries, colors, denominations and any other information 4°N shown on this map do not imply, on the part of The World Bank Turkana Group, any judgment on the legal status of any territory, or any IN DIAN JUNE 2007 endorsement or acceptance of such boundaries. Moyale UGANDA KENYA To To To O CE AN 32°E 34°E 36°E 38°E 40°E Wajir 42°E 44°E 46°E 48°E Marsabit Mogadishu IBRD 33426R K E N YA SELECTED CITIES AND TOWNS MAIN ROADS PROVINCE CAPITALS RAILROADS NATIONAL CAPITAL PROVINCE BOUNDARIES RIVERS INTERNATIONAL BOUNDARIES 34°E 36°E 38°E 40°E 42°E S U DAN To Murle To Juba Lotikipi Plain Karungu ETHIOPIA Lokichokio al bi Ch alb i To 4°N 4°N D e se r t To Ramu Mandera Imi Kakuma Dila lls Lak e Hi North Horr Sololo Tur k an a sa Moyale is Lodwar an D el Turkw Buna El Wak U G A N DA Lokichar South Horr Marsabit Tarbaj R I F T VA L L E Y . tns 2°N Kangatet M 2°N to do Log am N aB Wajir Su C og her al Lak B an or Milgis Hil gan ls y SOM AL I A ’iro Bilesha Plain To Kitale Maralal Ng Mbale Kapedo Ewaso Mando Archer’s Post Garba Gashi NORTH To Kampala Eldoret Marigat Tula EASTERN WESTERN Mbalambala Kakamega Isiolo era To Nyahururu Lak D Kismaayo Butere Falls EASTERN 0° Nanyuki 0° Mt. Kenya a Kisumu Tan (5,199 m) Nakuru Nyeri Kericho Garissa N YA N Z A Gilgil CENTRAL Embu Nguni Karungu Lake a NAIROBI M ar AREA Thika Bura Victoria To Narok Kolbio To Bur Gavo Musoma Lolgorien Ma NAIROBI Ngang u E Kitui erab sc arp eli Machakos Thua Plain me nt Konza COAST Magadi Bodhei 2°S Ya t t Ikutha i 2°S To a Ath Seronera Pla tea Garsen Lamu Kibwezi u Namanga To Arusha Tsavo Galana Tsavo To Moshi Vol Malindi INDIAN KENYA TANZ ANIA Mackinnon Park O CEA N This map was produced by the Map Design Unit of The 4°S Mombasa World Bank. The boundaries, Kwale colors, denominations and 0 40 80 120 160 200 Kilometers any other information shown on this map do not imply, on the part of The World Bank Shimoni Group, any judgment on the 0 40 80 120 Miles To legal status of any territory, Dar Es Salaam or any endorsement or acceptance of such 34°E 36°E 38°E 40°E boundaries. MARCH 2008 IBRD 33494R1 TA N Z A N I A SELECTED CITIES AND TOWNS MAIN ROADS PROVINCE CAPITALS RAILROADS NATIONAL CAPITAL PROVINCE BOUNDARIES RIVERS INTERNATIONAL BOUNDARIES 30°E 32°E 34°E 36°E This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any To endorsement or acceptance of such boundaries. 0° Tororo 0° To GAN DA U GAND A Kampala To Kampala Lake To Ka ge r a Victoria Nakuru K E N YA Bukoba Musoma Mara To Nakuru Buoen RWANDA KAGERA MARA 2°S 2°S Lake Mwanza Natron T T To M WA N Z A Simi yu Ka A R U S H A Kilimanjaro ma (5895 m) a a Moshi BURUNDI Arusha To DEM. REP. O F CO NGO OF CONGO Yalova Lake Malindi Mo M S H I N YA N G A Eyasi Lake yow y Kibondo Pa P Shinyanga e Manyara KILIMANJARO n ga p o si g Kahama s 4°S ni ep i Nzega Babati Same St sa Masai KIGOMA PEMBA e Steppe St er NORTH Kasulu MANYARA mb Kigoma Singida Kondoa Kaliua PEMBA Tabora Iwe SINGIDA Wete SOUTH Tanga TA N G A Mkoani ZANZIBAR TA B O R A Lake Ugalla Manyoni NORTH Tanganyika DODOMA ts. Mkokotoni ZANZIBAR M SOUTH & u Zanzibar Koani CENTRAL ur i am Mpanda D O D O M A Ng ZANZIBAR W WEST Morogoro Kibaha R U K WA Dar es Salaam Rung wa Grea MOROGORO MO ROGORO DAR ES SALAAM t Rua ha P WA N I Lake Iringa Sumbawanga Rukwa M B E YA e 8°S ng IRINGA Utete 8°S Ra ro Mpui ya IN DI AN ji e ufi be Mb R Kilom Kilwa Mbeya du Kivinje t an Ma Tunduma Ki pe Njombe O CE AN To Kasama n LINDI ur u mk ge ge be Lindi re 10°S M 10°S Mtwara Ra To n ng Kasama A MB IA Z AM B I A e Songea Masasi To TANZANIA Kasungu MTWARA Lake RUVUMA a Tunduru vum Ru Malawi To Chiúre 12°S To To Lichinga Marrupa MO ZA MBIQ UE MOZAM BI QUE 0 50 100 150 200 Kilometers 32°E 34°E 36°E 0 50 100 150 Miles 40°E NOVEMBER 2007