Case Study: Rwanda’s Twubakane Decentralization and Health Program

a.
February 2017
This publication was produced for review by the United States Agency for International Development.
It was prepared by Cristina Sciuto for the Health Finance and Governance Project.
CASE STUDY: RWANDA’S TWUBAKANE
DECENTRALIZATION AND HEALTH
PROGRAM

The Health Finance and Governance Project
USAID’s Health Finance and Governance (HFG) project helps to improve health in developing countries by
expanding people’s access to health care. Led by Abt Associates, the project team works with partner countries to
increase their domestic resources for health, manage those precious resources more effectively, and make wise
purchasing decisions. The five-year, $209 million global project is intended to increase the use of both primary and
priority health services, including HIV/AIDS, tuberculosis, malaria, and reproductive health services. Designed to
fundamentally strengthen health systems, HFG supports countries as they navigate the economic transitions
needed to achieve universal health care.
February 2017
Cooperative Agreement No: AID-OAA-A-12-00080
Submitted to: Scott Stewart, AOR
Office of Health Systems
Bureau for Global Health
Recommended Citation: Sciuto, Cristina. February 2017. Case Study: Rwanda’s Twubakane Decentralization
and Health Program. Bethesda, MD: Health Finance & Governance Project, Abt Associates Inc.
Abt Associates Inc. | 4550 Montgomery Avenue, Suite 800 North | Bethesda, Maryland 20814
T: 301.347.5000 | F: 301.652.3916 | www.abtassociates.com
Avenir Health | Broad Branch Associates | Development Alternatives Inc. (DAI) |
| Johns Hopkins Bloomberg School of Public Health (JHSPH) | Results for Development Institute (R4D)
| RTI International | Training Resources Group, Inc. (TRG)

CASE STUDY: RWANDA’S TWUBAKANE
DECENTRALIZATION AND HEALTH
PROGRAM
DISCLAIMER
The author’s views expressed in this publication do not necessarily reflect the views of the United States Agency
for International Development (USAID) or the United States Government.

i
CONTENTS
Contents .................................................................................................................... i
Acronyms................................................................................................................. iii
Executive Summary ................................................................................................ v
1. Introduction .............................................................................................. 1
2. Methods ..................................................................................................... 3
2.1 Design and implementation ..............................................................................................3
2.2 Data collection and analysis..............................................................................................6
2.3 Cross-case analysis..............................................................................................................7
3. Findings...................................................................................................... 8
3.1 Pre-conditions.......................................................................................................................8
3.2 Pre-implementation...........................................................................................................18
3.3 Implementation...................................................................................................................19
3.4 Maintenance and evolution .............................................................................................23
3.5 Challenges............................................................................................................................25
3.6 Success Factors ..................................................................................................................25
3.7 Dissemination .....................................................................................................................26
3.8 Lessons learned..................................................................................................................26
4. Discussion and Synthesis ....................................................................... 28
4.1 Synthesis...............................................................................................................................28
4.2 Limitations ...........................................................................................................................29
4.3 Conclusion...........................................................................................................................29
Annex A: Combined Implementation Framework........................................... 30
Annex B: Key Informant Interview Guide.......................................................... 33
Annex C: Detailed Activity Table ....................................................................... 37
Annex D: Bibliography.......................................................................................... 38

ii
List of Figures
Figure 1 Outline of combined implementation framework........................................................3
Figure 2 Twubakane Decentralization and Health Program Qualifying HSS Criteria
............................................................................................ Error! Bookmark not defined.
Figure 3 Rwanda Health and Decentralization Diagram (USAID 2010:8)...................Error!
Bookmark not defined.
Figure 4 2006 Revised Twubakane Intervention Zones (Twubakane Program) ................14

iii
ACRONYMS
CFIR Consolidated Framework for Implementation Research
CHIS Community‐Based Health Information System
COP Chief of Party
DCOP Deputy Chief of Party
DIF District Incentive Funds
EONC Essential Obstetric and Newborn Care
F&A Finance & Administration
FP Family planning
GOR Government of Rwanda
HBM Home-based management of fever
HFG Health Finance and Governance Project
HRIS Human resources information system
HSS Health system strengthening
IMCI Integrated Management of Childhood Illness
JADF Joint Action Development Forums
MINALOC Ministry of Local Government
MINECOFIN Ministry of Finance and Economic Planning
MOH/MINISANTE Ministry of Health
NDIS National Decentralization Implementation Secretariat
NHA National Health Accounts
PAQ Partenariat pour l’Amélioration de la Qualité
PEPFAR President’s Emergency Plan for AIDS Relief
PMI President’s Malaria Initiative
RALGA Rwandan Association of Local Government Authorities

iv
REP Replicating Effective Programs
RFA Request for Applications
RH Reproductive health
SWOT Strengths, Weaknesses, Opportunities, Threats Analysis
TAG Technical Advisory Group
USAID United States Agency for International Development
WHO World Health Organization

v
EXECUTIVE SUMMARY
USAID’s Health Finance and Governance project (HFG) contributes to USAID’s assistance to countries
to deliver key health services and builds the evidence base around health systems strengthening
(HSS). Under HFG’s research portfolio, a series of retrospective, qualitative case studies were
undertaken to understand the dynamics of successful HSS interventions by focusing on how HSS
projects were implemented. This report presents the results for one of the five cases: Rwanda’s
Twubakane Decentralization and Health Program.
Twubakane was led by IntraHealth with support from RTI International, Tulane University, and
EngenderHealth as the subcontractors. Numerous Rwandan counterparts, local implementers, and
development partners supported smaller pieces of the contract. Originally funded at $24 million, the
project received a total obligation of $28,379,327 from USAID. The project implemented a two-pronged
health system strengthening approach where both health care delivery and governance were addressed
through six components:
 Family planning and reproductive health;
 Child survival, malaria and nutrition;
 Decentralization policy, planning and management;
 District-level capacity building;
 Health facilities management; and
 Community engagement and oversight.
The project was meant to support health system financing, decentralization, and human resources for
health. Only six months into the project in 2005, the geographic scope of the work changed when Mr.
Paul Kagame initiated Phase II of Rwanda’s decentralization efforts. Phase II aligned formerly unaligned
administrative and health districts with single unified districts. In response, the project reprioritized to
make district-level capacity-building its top priority. The projects’ ability to accommodate this change in
work plan facilitated project sustainability.
Critical features of the project were established early on. First, USAID engaged the Government of
Rwanda during the Request for Application (RFA) development process to understand GOR priorities
and establish a mutually respectful relationship. In doing so, Twubakane was able to align activities with
existing GOR policies and initiatives from the beginning. Consequently, Twubakane staff were invited to
participate in workgroups and steering committees to update outdated policies, which further
strengthened project-GOR relationships. When ad-hoc needs were identified, like training police officers
to raise awareness about gender-based violence or changing the geographic scope of the project,
Twubakane staff knew who to speak with and how to navigate the process successfully. One project
activity expanded an existing structure; community-provider partnerships, Partenariats pour
l’Amélioration de la Qualité (PAQs), which were ultimately deemed a best practice by the GOR before
the project ended. Second, IntraHealth promoted collaboration by making staff engagement a top
priority. Everyone from drivers to the Chief of Party were expected to participate in staff meetings and
annual retreats. This increased staff buy-in and ownership of the project, which ultimately increased
team spirit. Finally, district officials were offered financial resources and the opportunity to manage them
through District Incentive Funds. This increased district ownership over the planning, budgeting, and

vi
management of resources which increased their adherence to the capacity-building opportunities offered
by Twubakane.
The main challenges revolved around reporting requirements, data availability, and limited opportunities
for innovation. Over the course of the project, multiple reporting requirements were expected from
various GOR agencies at different times. Sometimes this limited staff availability to work on the
implementation itself. A second limitation was the lag in the Demographic and Health Survey (DHS)
data. While the project was able to conduct large-scale capacity assessments, their monitoring and
evaluation efforts to assess health outcomes were limited due to the data lag. Finally, GOR preferred
not to fund pilot activities unless they were evidence-based and scalable. This limited the amount of
innovation Twubakane could introduce.
Among other outcomes, the project contributed to 1) a reduction in the infant mortality rate from 107
to 62 per 1000 live births, 2) an increase in couple years of protection by almost five-fold, and 3) the
active engagement of a supervisor in 98% of the 136 PAQs.
Some lessons learned include the value in:
 Regular communication between team members and stakeholders;
 Building highly effective teams through regular communication and in-person retreats;
 Bottom-up stakeholder engagement in the planning and implementation process; and
 Convincing the target population of the importance of the activity.

1
1. INTRODUCTION
USAID’s Health Finance and Governance (HFG) project helps to improve health in developing countries
by expanding people’s access to health care. The project team works with partner countries to increase
their domestic resources for health, manage those precious resources more effectively, and make wise
purchasing decisions. HFG’s research portfolio enhances the ability of USAID to assist countries in
delivering priority health services while simultaneously contributing to the global pool of knowledge on
health systems strengthening (HSS).1
Under this research portfolio, the “Understanding the Dynamics of Successful Health System
Strengthening Interventions” study seeks to bring into better balance our focus on “what works” in HSS
with “how HSS works” to improve the performance of future HSS efforts. Our aim is to examine the
dynamics of HSS project implementation, not to examine the cases as models for HSS interventions. We
are pursuing this goal by initially conducting a set of five qualitative, retrospective case studies of
successful USAID-supported HSS interventions and then producing a cross-case analysis to draw
common patterns across cases.
The aim of this study to address four key questions:
1. How were a range of successful HSS interventions implemented in different countries?
2. What factors facilitated and constrained the successful implementation and documented
outcomes of the interventions?
3. What were important factors about implementation that emerged across the different cases?
4. What are the implications of this study for future of implementing HSS interventions?
We chose five cases to examine a small sample of successful HSS initiatives in different places under
different conditions and with different features in an attempt to tease out some of the policy setting,
adoption, and implementation factors and processes that matter. While we remain attentive to the
range of complex factors that affect success, we seek to distinguish those factors that decision-makers
and implementers can control or influence. In so doing, we hope to develop and provide
recommendations for adapting and sustaining HSS reforms in low-income countries.
This report presents one of the five case studies – on the Twubakane Decentralization and Health
Program project. In Section 2, we describe the study methods. In Section 3, we present the contours of
the context in which the intervention was implemented, basic information on the intervention, how it
was designed, and its outcomes; the implementation process for the intervention, including
implementation groundwork, key features of the implementation process; and how the intervention was
1 As defined by the World Health Organization, we define HSS interventions as those that implement “changes in policy
and practice in a country’s health system” and improve “one or more of the functions of the health system and that leads
to better health through improvements in access, coverage, quality, or efficiency” (WHO 2011: 9). HSS interventions are
horizontal approaches that can address the root causes of health system constraints and impact multiple issues, rather
than vertical service- or disease-specific interventions like health system support programs (Travis et al. 2004: 903).

2
sustained and disseminated. Finally, in Section 4, we present our synthesis of the primary factors that
influenced the intervention’s implementation and contributed to its success.

3
2. METHODS
The study, comprised of five case studies and cross-case analysis, was conducted in several phases, each
of which is briefly described in turn. For a more detailed explanation of our case selection process and
methods, please see the study design (Conrad et al. 2016).
2.1 Design and implementation
In the first phase of the study (October 2015-March 2016), we finalized the design and began
implementation, which involved engaging USAID and selecting the case studies.
2.1.1 Design
The aim of this study was to address four key questions:
1. How were a range of successful HSS interventions implemented in different countries?
2. What factors facilitated and constrained the successful implementation and documented
outcomes of the interventions?
3. What were important factors about implementation that emerged across the different cases?
4. What are the implications of this study for future of implementing HSS interventions?
To answer these questions, we designed a protocol to conduct retrospective, qualitative case studies.
We used an implementation framework to guide the case studies. Our primary aim for applying the
implementation framework was to determine which factors influence implementation that we needed to
collect data on and consider during analysis. We combined two implementation frameworks to apply in
this study – the Consolidated Framework for Implementation Research (CFIR) (Damschroder et al.
2009) and the Replicating Effective Programs (REP) framework (Kilbourne et al. 2007). Both CFIR and
REP are based on implementation theories and empirical evidence of what affects the successful
implementation of health interventions. We used CFIR to more broadly frame the intervention and we
used REP as a framework that focuses on
project implementation process. Figure 1
outlines the combined framework. See Annex A
for detail.
As we assessed each implementation domain
and factor, we also explored:
1. Decision-making processes associated
with design and adoption of the
intervention;
2. How the intervention was
implemented, including how potential
challenges or obstacles were addressed;
Figure 1: Outline of combined implementation
framework

4
3. Expected and unexpected outcomes of the intervention, both positive and negative; and
4. Prospects for sustainability of the intervention, such as the degree to which the project activities
have been institutionalized in the country.
Before we finalized the design, the team submitted the study design and data collection instruments to
Abt’s Internal Review Board (IRB) and JHSPH’s IRB for review. Abt’s and JHSPH’s IRB exempted the
study from review.
2.1.2 Sampling Process
To ensure that the case studies were of practical relevance, we set up a Technical Advisory Group
(TAG) composed of experts and representatives from inside and outside USAID Bureau of Global
Health to consult with on the study and provide expertise.
This case was selected for study from USAID’s 2014 Global Call for Health System Strengthening Cases
using a defined set of criteria and a systematic review and sampling process that we developed. The case
was purposively selected from the available pool and the case is not representative or necessarily the
most successful health system strengthening (HSS) project implemented in the region. Our objective in
the case selection was to purposively select 6 cases from the 143 cases submitted to USAID’s 2014
Global Call for Health System Strengthening Cases that are successful, robust examples of health system
strengthening interventions.2
The reviewers engaged in a multi-stage sampling process consisting of four sequential selection rounds
that excluded cases that did not meet the specified criteria in each round using the identified available
data and the predetermined review method. The 4 selection rounds were as follows:
1. Round 1: Reviewers considered only those interventions that were fully implemented before
the start of the selection process.
2. Round 2: Reviewers accepted the submitter’s self-reported definition of health systems
strengthening, labeled the intervention “provisional,” and sought a determination of an
“effective” intervention.
3. Round 3: Reviewers applied criteria to determine whether a provisional, effective health system
strengthening intervention could be confirmed as health system strengthening.
4. Round 4: Reviewers applied criteria to determine whether a confirmed, effective health system
strengthening intervention was robust (Figure 2)
Figure 2: Twubakane Decentralization and Health Program Qualifying HSS Criteria
Round Criteria Inclusion criteria How met criteria
1 (implementation
period)
Implementation completed Submission states
implementation period
was completed by
10/2015
2010
2 One of the six case studies was not completed, because we were unable to collect sufficient data to complete analysis.

5
2
(impact and
evidence)
Effective intervention One of 13 identified types
of interventions
referenced
 Accountability and
engagement
interventions
 Health insurance
 Health worker
training to improve
service delivery
 Pharmaceutical
systems strengthening
initiatives
 Service integration
 Strengthening health
services at the
community level
 Task sharing/task
shifting
Health systems outcome One of 4 health systems
outcomes referenced
 Improved service
provision/quality
 Uptake of healthy
behavior
Health impact Health impact referenced Reduced morbidity and
mortality
Both health system outcome
and health impact
At least one health system
outcome and health
impact referenced
Yes
Verification of health impact
and health system outcome
achieved
One type of
documentation is
referenced for at least
one health impact or
health system outcome
Project M&E data
3
(HSS)
Multiple primary disease
targets
At least 2 diseases
targeted referenced
Family planning,
reproductive health, child
health, malaria, nutrition
4
(Robust HSS)
Multiple health system
functions and sub-systems
targeted
At least 2 HSS WHO
building blocks targeted
and at least 2 sub-systems
functions targeted
Building blocks:
 Financing
 Health workforce
 Leadership and
governance
Sub-systems functions:
 Financing
 Health workforce
 Leadership and
governance
 Service delivery

6
Verification that intervention
was successful HSS
intervention
Intervention had health
system outcome, health
impact and targeted
multiple diseases and
health system functions
Yes
Category D for HSS
intervention type
Based on typology of HSS
we developed, case
addresses at least 2 health
system functions and at
least 3 sub-systems
Yes
Category E for HSS
intervention type (not
inclusive of D)
Based on typology of HSS
we developed, case
addresses at least 2
health system functions
and at least 4 sub-systems
Yes
2.2 Data collection and analysis
In the second phase, we conducted the case study research. We divided the case studies among our
team members so that no team members conducted research on a project that their organization
implemented. The case teams collected both primary and secondary data on retrospective (features 1-3
above) and prospective (feature 4 above) data that are described in more detail below. As applicable, we
collected primary and/or secondary data on each implementation factor and domain.
For primary data collection, we conducted individual or joint interviews with key informants who
possessed in-depth knowledge of the history and workings of the HSS intervention. We followed a
common semi-structured interview guide for the interviews, but adjusted the questions posed as
applicable for the respondent and their role in the project (see Annex B for the interview guide).
Interviews were conducted in English or French, depending on the respondent’s preference and
comfort. We documented each interview through verbatim notes in English and audio recordings. We
interviewed 11 key informants for this case study. Informants included representatives of USAID’s
implementing partners who sponsored the intervention, relevant Ministry of Health officials, and USAID
mission staff with knowledge of the intervention, as appropriate.
The research team imported the interview notes into NVivo 11, qualitative data analysis software
package, for coding and analysis. Analysts applied a single codebook developed prior to beginning the
coding process and refined by coding a small sample of interview notes from several cases. The codes
were informed by a priori concepts based on the domains and factors from the combined CFIR and REP
implementation frameworks. To accommodate unexpected or context-bound themes and concepts
emerging from the data, the codebook included a ‘family’ for each case to allow for inductive coding as
needed for each specific country or intervention. We applied this common codebook for the purposes
of reliability, quality control, and comparison across interview respondents and eventually across case
and country contexts. All data is saved to Abt’s secure server.
Once coding was complete, the analysts conducted iterative, exploratory analysis in NVivo using text
analysis techniques (e.g., repetition, similarities and differences, word frequency, word co-occurrence,
semantic network analysis, etc.) to explore themes, patterns, outliers, and trends, and conflicts between
and among data sources.

7
We reviewed secondary data to capture different features of the intervention and contextualize the
intervention. We conducted document review of the relevant published and unpublished documents
about the intervention that we were able to obtain. To review the documentation on each case, we
filled out a common document abstraction template (in an Excel spreadsheet) to systematically review
the documents and synthesize salient data. Abstraction categories reflected domains from our
combined CFIR and REP frameworks. We also conducted a focused literature review to identify the key
contextual factors (e.g. socio-cultural, political, economic, etc.) relevant to the case and existing
evidence about barriers to and success of health system strengthening and reform in the country. We
used the literature and document reviews to build on and verify the interview data where possible and
applicable (bearing in mind that written documentation represents the official record). We analyzed the
findings from the literature and document reviews in conjunction with analysis of the primary data. We
uploaded the document abstraction forms in NVivo for analysis with the interview data.
The research team ensured the reliability and validity (both external and internal) of our qualitative
research in several ways. We revised our semi-structured interview guide and record review forms
based initial use. We used experienced researchers and held team meetings to ensure that all team
members had a consistent and thorough understanding of the research goals and intent behind each
question and probe. We further used consistent data documentation procedures and structured,
systematic analysis techniques using qualitative analysis software (e.g., NVivo) to ensure reliability, quality
control, and cross case comparisons. Further, we triangulated primary qualitative data with secondary
data to improve the validity of findings from primary data. Finally, we conducted member checking by
asking a key informant, usually the project’s Chief of Party, to review and comment on the case
narratives regarding coherence and validity. We also had a TAG member review each case narrative to
provide further expert review. We then finalized the case narratives based on this feedback.
2.3 Cross-case analysis
In the third phase of the study, we analyzed this and the other four descriptive case study narratives
from Phase 2 to help generate explanations for successful HSS interventions. The cross-narrative
analysis of Phase 3 sought to build or strengthen the evidence base for the “how” and “why” of what
works in HSS by determining which implementation domains and factors from the implementation
framework influenced the success of the interventions. We looked for common and divergent factors
that were present or absent across cases and contexts, and we tried to determine the relationships
between the implementation factors and domains based on our findings. As an exploratory study, we
hope these findings can provide some comment on the factors that may be associated with successful
HSS implementation and inform future studies of HSS interventions.

8
3. FINDINGS
The report describes the implementation experience of the Rwanda Twubakane Decentralization and
Health Program supported by USAID, “to increase access to and the quality and use of family health
services in Rwanda” (USAID 2010).
In this Section, first we outline the relevant features of the context within which the intervention was
implemented, including key features of the socio-economic context, political system, and health system.
Second, we describe the basic features of the intervention, including its primary goals, activities, design,
and timeline. Third, we outline the main outcomes and impacts of the project. Fourth, we describe the
implementation process, beginning with the implementation groundwork, implementation itself, and then
how the project was sustained and disseminated.
3.1 Pre-conditions
3.1.1 Problem definition
The Twubakane project was, “designed to increase access to and the quality and use of family health
services in Rwanda.” While most respondents agreed that the project aimed to address, “pillars around
the health system including financing, decentralization, and human resources for health” (Rwanda 06,
MOH), respondents from different organization types differed in the weight they gave to various aspects
of the project. Ministry of Health (MOH) and subcontractor representatives, those most engaged in
capacity-building and governance, listed decentralization efforts first when describing the problems the
project was trying to solve. Respondents from the prime organization listed improving health outcomes
first.
While poor health outcomes (such as malaria as number one killer of children under five years of age
and high maternal, child and infant morbidity and mortality rates in general, along with low use of
modern contraception) were part of the problem, Twubakane first targeted system-level challenges,
which arose from the government’s rapid decentralization of the health system from the central to the
district level. Due to these changes, the project focused on building the capacity of the newly-designated
district officials to plan, budget for and deliver health services in order to positively impact health
outcomes.
“Rwanda’s a very centralized model. The problem they were also trying to solve was…improving the
functioning of the health system at the district level with clear accountabilities of people, clear
accountabilities to really provide the kinds [of] quality health services that populations needed.”
(Rwanda 09, Implementer)
3.1.2 Enabling environment
The year Twubakane began, 2005, was a time of rapid change in Rwanda. In 2003, President Paul
Kagame had come to power. With this shift in leadership came swift interest in the decentralization
strategy, among other changes. One report indicates that decentralization spanned from 2000-2015
(USAID 2009, Rwanda 2011). Multiple respondents indicated that the decentralization process began 6

9
months into the Twubakane project, which involved decentralizing the MOH and creating new
administrative districts. These replaced the former unaligned administrative and health districts with
single unified districts. This required a rapid readjustment of the geographic scope of the project which
involved multiple stakeholders. One implementer explained,
“Within 6 months of starting, the government announced that they wanted to fast-track
decentralization, which meant they created brand new districts. Instead of having separate
administrative districts and health districts, they were all aligned. Originally it was crazy. Every health
district would have like 2 or 3 administrative districts within it. They decided to create 30 districts for the
whole country, so every district had a locally elected mayor, which actually greatly facilitated the project
at being a decentralization and health project…From then on, the annual work plans were really done
with the national programs; whether it was the national malaria program, the family planning program,
child survival in general, maternal health, but also had district-level activity plans that were approved by
the district health officer and the district team. So it was very participatory.” (Rwanda 04,
Implementer)
Several respondents referred to the first year of decentralization as a “mess” accompanied by significant
staff turnover at the central government level and cultural shifts. Two tensions mentioned several times
were that 1) many of the new, decentralized roles were taken over by younger, less experienced
Rwandans and 2) the roles and responsibilities of these new positions were not always well defined. One
respondent described these tensions:
“I know that there were issues with the MOH; let’s just put it that way. You know, what their role was
as things evolved; their stewardship role versus their implementation role. I’m sure there were issues
there and as I mentioned, there [were] also issues of roles and responsibilities in the districts as the roles
shifted. We had to do a lot of participatory workshops just to clarify who was doing what; what they
thought their job was. You had the old, authoritarian hospital director versus this new kid on the block
[who] was coming in to manage the health office. It wasn’t always an easy transition, so that had to be
sorted for sure.” (Rwanda 01, Implementer)
Though decentralization began as a somewhat chaotic transition, many respondents viewed Twubakane
in a positive light. Given the focus of the project on health systems and governance, it “came at the right
time” in Rwanda (USAID 2006:6). One respondent described Twubakane as “responsive and timely”
(Rwanda 02, Implementer) to various opportunities in 2005 provided by health system decentralization.
Further, the government was actively working towards and receptive to change.
Many government policies supported decentralization and Twubakane’s work [Government of Rwanda
(GOR)’s Health Sector Policy and Strategic Plans (2005-2009), the Decentralization Strategy (2004–
2008), Vision 2020, and the EDPRS (Economic Development and Poverty Reduction Strategy) (USAID
2009)]. However, Twubakane’s lack of funding for and focus on HIV was unusual for the GOR and
resulted in minor tensions. One respondent is under the impression that in 2005, Twubakane was the
only US government-funded project not receiving PEPFAR funding (Rwanda 04, Implementer). When the
Executive Secretary of Rwanda’s National AIDS Control Commission became the Minister of Health
(NIH 2015), questions about why Twubakane was not collaborating with the National AIDS Control
Commission increased (Rwanda 04, Implementer). The same respondent described being asked multiple
times to remind officials why Twubakane was not focused on HIV. Each time, the respondent would
reply that there were many NGOs already working on these issues and that Twubakane did not have
PEPFAR funding, though it did receive some during the life of the project. Because it was only a nominal
amount of funding, it did not significantly change project activities.
The government of Rwanda had a, “desire to develop economically” (Rwanda 09, Implementer) in 2005 and
to rebuild after the 1994 genocide. The respondent further elaborated, “Rwanda was still going through a

10
lot of not ‘healing’ exactly, but you know what I mean”. This desire to develop and heal provided a healthy
participatory framework through which Twubakane was able to engage with project stakeholders and
execute the work.
Figure 3: Rwanda Health and Decentralization Diagram
Source: USAID 2010:8
3.1.3 Implementation setting
The impact of decentralization was felt from the very top to the local administrative levels of
governance and health systems. The staffing structure and job roles and responsibilities changed at the
ministerial, provincial, and district levels, which led to significant changes in not only leadership, but
overall who was in charge of what. One implementer explained,
“The MOH shrank by 30 people at the national level, which severely curtailed their ability to do a lot
and power was shifting to the local levels, de-franchising provinces and really improving and
strengthening the role of districts in all of this. The way health services were managed locally was [also]
shifting. Traditionally it had been the heads of the provincial and district hospitals that were in charge,
but because of the restructuring, they were creating new health offices and putting much younger people
in charge of those offices. They were kind of colliding with the traditional health managers, the facility
managers. All the while roles and responsibilities were really confused and confusing. Our project had to
help sort all of that out and make sure that everyone was involved in that process from top to bottom.”

11
The ministries Twubakane worked with had unique organizational cultures and internal politics to
navigate. In addition to the fast-paced staffing changes, there was a shift to doing business in English as
opposed to French. While not mentioned in project documents, this was confirmed by an NPR article
from 2008 which discusses the replacement of French with English as the, “language of business,
diplomacy, and scholarship” in Rwanda. The author notes that this is due to the fact that those in power,
including President Kagame, are English speakers because many were in exile after the genocide and
learned English abroad (NPR 2008). One implementer noted that with the “transition itself in terms of the
bilingual between the French, English, and Kinyarwanda… there was lots of government [Ministries] where you
had the Minister was [speaking] one language, the Deputy Minister came from another… they were really trying
to do a lot of balancing there” (Rwanda 09, Implementer).
The Rwandan culture is seen by many development partners as being one of compliance and a “zero
tolerance policy” towards corruption (Rwanda 11, Implementer). This was both a facilitator and a barrier
for the project. According to those interviewed, GOR staff largely complied with government policies.
This high level of compliance facilitated swift decentralization, and even the abolishment of plastic bags
almost overnight, according to one implementer. But this compliance seems to be a result of what some
development partners see as the authoritarian-like governance of the GOR, which had an impact on
implementing agencies, including the Twubakane program. For donors, being in compliance with the
GOR meant accepting that their ideas had lesser weight in project-related decision making, even to
project branding.
“I know there were some challenges in terms of the real, not just the buy-in, but the real ownership of
some of the directions of the project, some of the visibility of the project and I think all of us working in
this know of the challenge of too much attention called to a project versus the government’s help and
the project’s TA to make the health sector perform better. I think at different times in that project there
was tension around the whole branding. It was Twubakane and this USAID project as opposed to the
government sort of health system...I think there were some people in the government who didn’t like the
fact - I don’t know if they ever said it, that there was so much visibility around the project.“ (Rwanda 09,
Implementer)
Additionally, following the culture of compliance meant that implementers did not have the flexibility to
test out as many new creative and innovative ways to implement as they and USAID would have liked.
This also was a result of what the government sometimes called “pilotitis,” as many development
partners were eager to test out pilot approaches in specific geographic areas. As a respondent
explained,
“I think the hardest part was when you wanted to do something innovative and they were – the only
thing they didn’t like and were always forbidding was pilot projects…if it was a good idea and they
approved it, then you had to do [it] everywhere in the country and if it’s not a good idea then we won’t
do it. That I felt like was a problem when trying to do something innovative—which USAID likes—
because they [GOR] just didn’t think pilot projects were necessary or needed.” (Rwanda 05,
Implementer)
One respondent felt bureaucracy was an issue, citing there was, “more bureaucracy than I’ve seen in some
countries as far as getting approval… I felt it was hard to do any sort of research in the country because of the
bureaucracy around that” (Rwanda 05, Implementer). But overall, because the project “followed Rwandan
policy” (Rwanda 05, Implementer), most respondents did not feel the government was prescriptive or
restrictive.
According to one respondent, it was illegal to talk about the genocide in a way that one could be
accused of “genocide denial” and to mention, by name, the major ethnic groups in Rwanda, (Rwanda 11,

12
Implementer) but its impact could still be felt. The project used an inclusive staffing approach to comply
with this law and gain the trust of Rwandan stakeholders.
While the administrative and political changes underway complicated the implementation setting, the
GOR handled the changes relatively well. According to multiple respondents, the GOR is known for
making changes quickly, “I think Kagame has even used this in his speeches. He talks about Rwanda being a
country in a hurry. So when they would decide that something worked, they decided not to do sort of a gradual
kind of scaling up, but to move quickly to sort of a national roll out” (Rwanda 11, Implementer). As it turns out,
family planning innovations, the community health insurance scheme called “mutuelles de santé” and the
overall decentralization strategy, three of Twubakane’s initiatives, were among those the government
believed “worked,” facilitating political support for the project.
3.1.4 Project features
The Twubakane project was implemented in Rwanda from 2005 to 2010 by a consortium led by
IntraHealth International. The consortium included sub-contractors, (RTI International, Tulane
University, and EngenderHealth), as well as local
implementers RALGA, Pro-Femmes Twese Hamwe, and
VNG. The consortium worked in partnership with GOR
stakeholders, including the Ministry of Health, Ministry of
Local Government, and District Mayors, in addition to other
development partners such as GEZ and the Dutch SNV who
helped implement the decentralization activities.
In collaboration with the aforementioned groups, Twubakane
used 1) technical assistance, 2) capacity-building activities, and
3) grants to help the GOR decentralize its health system
down to the district level. While the project name,
“Twubakane Decentralization and Health” might suggest a
siloed approach, the Kinyarwanda meaning of “Twubakane” is
actually “let us build together.” Respondents agreed that the
project was collaborative, innovative and groundbreaking at
the time. A project document from 2009 confirms this,
"Rebuilding service delivery capacity of the state was a key
objective of the government following the genocide, and health
services figure prominently among the services that citizens
need and desire. The government’s interest in decentralization
is in part driven by the performance link between
decentralization and improved service delivery." (USAID
2009:5)
The following six components and associated objectives of
the project (see Annex C for detailed activities) balanced a
bottom-up approach to planning with close and consistent
central-level government engagement to ultimately increase
access to, quality, management and use of family health
services (USAID 2010).
1. Family Planning and Reproductive Health
Project Profile
Title: Twubakane Decentralization and
Health Program
Period: January 2005-January 2010
Funding: USAID
Budget: Over $34,871,226
Prime contractor: IntraHealth
Sub-contractors: RTI International,
Tulane University, EngenderHealth
Rwandan Counterparts: Government
of Rwanda, Ministry of Local
Government, Ministry of Health, District
Mayors, District Accountants, and other
local administrators
Local implementers: RALGA, Pro-
Femmes Twese Hamwe, VNG
Development Partners: GEZ [GED],
the German cooperation, the Belgian
technical cooperation, the Dutch SNV, B
and G international, DFID, the Swiss
technical cooperation, the Swedish
development organization, and the
Canadian Development Organization
Focus: Local district management and
financing to inform access to and quality
of healthcare services

13
a. Increase access to and the quality and utilization of FP and RH services in health facilities
and communities
2. Child Survival, Malaria and Nutrition
a. Increase access to and the quality and utilization of child health, malaria and nutrition
services in health facilities and communities
3. Decentralization Policy, Planning and Management
a. Strengthen central-level capacity to develop, support and monitor decentralization
policies and programs, with an emphasis on health services
4. District-Level Capacity Building
a. Strengthen capacity of districts to plan, budget, mobilize resources and manage services,
with an emphasis on health services
5. Health Facilities Management
a. Strengthen capacity of health facilities, including health centers and hospitals, to better
manage resources and promote and improve the functioning of mutuelles
6. Community Engagement and Oversight
a. Increase community access to, participation in, and ownership of health services
The project’s approach was to improve access to and the quality of services, while improving the
government’s ability to maintain these services on their own. In the words of one implementer, the
approach, “was to build the capacity of the government of Rwanda at all levels – the central level, MOH, local
governance, and Ministry of Finance quite honestly, as well as [the] district level to be able to prioritize health
needs, budget and plan for meeting the health needs, and work with strong community engagement to improve
health outcomes and improve governance” (Rwanda 04, Implementer).

14
Figure 4: 2006 Revised Twubakane Intervention Zones (Twubakane Program)
Twubakane was a cooperative agreement implemented in 12 districts (see Figure 4). Several
respondents felt that the nature of the cooperative agreement allowed Twubakane more flexibility and
closer collaboration with USAID and GOR than contracts might.
The project supported more than a third of Rwanda’s population (USAID 2010). According to one
implementer, Twubakane worked in 14 hospitals and 56 health centers (Rwanda 03, Implementer). This
was done with USAID funding, which increased over time. Twubakane was initially funded at $24 million,
but received a total obligation of $28,379,327 from USAID (USAID 2010). In Year 3, Twubakane
received PEPFAR and PMI funding. PMI funding was used to support more in-depth policy development
and integration with existing malaria prevention activities. In Year 4, Twubakane discovered that the full
anticipated funding from USAID was not available and thus revised 2009 workplans accordingly. As a
result, some planned activities were postponed or cancelled (USAID 2010). At one point, a $1 million
grant was awarded by the William and Flora Hewlett Foundation to engage parliamentarians and mayors
in the project’s family planning work (Rwanda 04, Implementer). In addition, a grant of $250,000 was
awarded by the David and Lucille Packard Foundation specifically to support adolescent reproductive
health. Ultimately, the project’s cost-share and leveraging added approximately $6,491,899 of associated
cash and in-kind programming to the budget for a total of $34,871,226 million over the course of the
project (USAID 2010).
Twubakane’s strategy was comprehensive yet targeted (IntraHealth, 2007). The project used a number
of guiding principles to achieve its cross-cutting objectives. These strategies encompassed stakeholder
engagement and criteria for improved decision-making, among others:

15
 Capacity building for existing systems, especially newly created districts.
 Partnership to privatize markets using DIF and engage stakeholders at all levels, from Community
Health Workers and religious leaders to President Kagame.
 Integration with GOR policies (both existing and those that were developed during the life of the
project). This involved a collaborative planning process to ensure district performance contracts
reflected and aligned with these policies.
 Creativity, innovation, and flexibility by adapting interventions to both central- and district-level
needs.
 Gender equity by including this as a criteria in District Incentive Fund (DIF) grant requests and
MINISANTE’s community health policy, requiring one female and one male community health
worker per community (IntraHealth 2007:10).
 Data-driven informed decision making by 1) using Rwanda’s Demographic Health Survey data
(Rwanda 04, Implementer) to inform project planning, 2) contributing to an assessment of the
national HMIS, 3) helping to computerize community health and mutuelles health information
systems, 4) building district-level capacity to make data-driven decisions, and 5) contributing to
the institutionalization of National Health Accounts (NHA).
 Accountability and transparency by building capacity for planning, budgeting, and financial and
programmatic analysis capacity and promoting GOR’s zero tolerance for corruption initiative in
partnership with RALGA.
 Community and individual participation through government “accountability days”, community
participation in Partenariat pour l’Amélioration de la Qualité (PAQ) teams, mutuelles, and
providing a platform for CHW elections.
Below we outline key events from the life of the project including major contextual or policy changes,
and project milestones. These include, but are not limited to, the redistricting process of 2006, the
evolution of the DIF, and the dates when major GOR policies were instated.

16
Table 1: Project timeline
The project asserts that it contributed to the health improvements documented at a national level for
the areas targeted by the project. For each relevant project component, we highlight these
improvements below (USAID 2010):
1. Family Planning and Reproductive Health
Year Event
2005 USAID awards Twubakane to IntraHealth and partners and launch of administrative reform
and redistricting process.
2006 MINISANTE identifies PAQ as best practice; MINISANTE conducts first Joint Health Sector
Review with Twubakane support; GOR’s redistricting and territorial reform begins; first
district mayoral elections held; Twubakane helps develop the first "imhigo", performance-
based contract between the district mayors and the President of Rwanda; District Incentive
Funds (DIF) grants initiative starts after initial delay; Rwanda joins the President's Malaria
Initiative (PMI) which Twubakane supports via strategy and work plan development; GOR
begins updating MPA/Complementary Package of Activities (CPA) for Family Health to expand
the coverage of services; and partners contributing technically and financially to the NHA
workplan sign Memorandum of Understanding (MOU).
2007 Government rolls out national health insurance coverage law; Twubakane holds workshop and
presents a revised strategic proposal to USAID and GOR - in response, USAID/Rwanda raises
budget ceiling to $30,689,199; Twubakane submits continuing application to GOR and USAID;
and Community‐Based Health Information System (CHIS) pilot tested in response to
MINISANTE request.
2008 Citizen report cards obtained for a number of services across sectors; Adam Smith
International conducts nationwide district capacity building needs assessment; Twubakane
focuses on sustainability strategies for each program-supported intervention; Twubakane
implements on-the-job training around FP services in hopes of making access to these
sustainable; Twubakane conducts assessment of the financial viability of mutuelles. Results
show about 70% of mutuelles from the Twubakane-supported districts risked bankruptcy in
2008 due to management problems or insufficient funding to pay health facilities’ bills;
Twubakane helps hospitals and health centers develop strategic plans and adapts health facility
management manuals to local context; Mutuelle financial viability assessment results presented
at national mutuelles workshop, prompting MINISANTE and development partners to conduct
nationwide audit.
2008, 2009 Twubakane continues to participate in Health Sector Cluster Group and Decentralization
Sector Cluster Group and supports initiation of new Health Development Partners Group.
2009 Full anticipated funding from USAID not available so Twubakane revises 2009 workplans
accordingly; some planned activities are postponed or cancelled so funding leveraged from
other development partners for some activities and a few districts take on the funding
themselves; MINISANTE incorporates PAQ into National Quality Assurance Policy;
Twubakane holds participatory stakeholder workshops with representatives from USAID,
MINISANTE, MINALOC, National Decentralization Implementation Secretariat (NDIS),
partners, and supported districts, hospitals, etc. where accomplishments and lessons learned
are reviewed and a commitment is made to maintain momentum and ensure sustainability;
End-of-Project Ceremony held with nearly 200 participants including the Minister of Health,
the US Ambassador, etc.; and Revised MPA/Complementary Package of Activities (CPA) for
Family Health still under review by MINISANTE.
Jan 2010 End of project

17
a. Couple years of protection increased by almost five-fold between 2005-2009
b. Nationally, the use of modern contraception increased from 4% in 2005 to 27% in
2007/2008
c. Use of partograms in health facilities increased over 60%
d. ANC visits almost doubled between 2006-2009 (from 52, 519 to 92, 027)
e. FP secondary posts had measurably increased access to and use of modern
contraception.
2. Child Survival, Malaria and Nutrition
a. Between 2005 and 2008, the infant mortality rate reduced from 107 to 62 per 1000 live
births and the under 5 mortality rate decreased from 196 to 103 per1000 live births
b. Home-based management of fever (HBM) went from not being offered at all to being
offered in 5 districts
c. The number of children under 12 months of age receiving DPT3 increased by 15%.
d. All 12 Twubakane-supported districts had trained and validated hospital training teams
in Essential Obstetric and Newborn Care (EONC).
e. 75% of health centers were actively implementing clinical Integrated Management of
Childhood Illness (IMCI).
3. Community Engagement and Oversight
a. CHWS began offering Child Survival/Malaria/Nutrition services in 5 focus districts
b. 84% of health centers had a PAQ team working to improve health services
c. The Community-based Health Information System was tested in 2 districts and was later
merged into the national community-PBF system
Health governance outcomes were also achieved. A 2009 health governance study led by RTI evaluated
Rwanda’s progress in terms of 1) responsiveness, 2) leadership, 3) voice, 4) accountability, 5)
transparency, 6) evidence-based decisions, and 7) efficiency & efficacy. These were assessed on a Likert
scale. The evaluation found that the largest impact on health governance outcomes came from increased
efficiency and effectiveness (USAID 2009:30). One of the respondents was actively involved in the health
governance study. They shared their own perspective on the study results:
"What we concluded is that the biggest impact on health governance fell into the category of increasing
efficiency and effectiveness … helping district mayors to be better at fulfilling their responsibilities,
helping people to meet requirements for funding and reporting. A lot of the interviewees cited things like
increased service utilization rates, higher immunization rates, better health insurance coverage; that was
the link to the health outcomes.…Then the second largest category was on the accountability. Here it
had to do with the accountability open house days…the training of auditors that essentially had to do
with financial accountability and the DIF. Those funds essentially forced the local governments to specify
how they were going to use the money, what the purposes were. It increased the accountability of
planning and budgeting more than had been there before. Then I guess in terms of the other areas, we
noted some increases in responsiveness around contributing to the enabling conditions of the health
system so management training, better facilities. The SWOT [strengths-weaknesses-opportunities-
threats] analysis helped people determine where they needed to make improvements.” (Rwanda 11,
Implementer)

18
Other study respondents felt differently. They felt Twubakane’s greatest impact on health governance
was increased citizen voice. They felt that “accountability days”, PAQs, mutuelles, etc. ensured that
health and government officials took citizens’ voices more seriously. As a result, citizens were more
empowered to contribute to health governance activities.
3.1.5 Project Design
As previously mentioned, Twubakane was a unique and creative project. One implementer asserted that
the project design was risky for USAID because it was innovative. According to the implementer, “as far
as I know it was one of the first projects in which that much money, 6 million dollars, was set aside for district
grants” (Rwanda 02, Implementer).
The project design was responsive to the needs of the GOR. Respondents from the Prime had the
impression that USAID worked closely with representatives of the GOR to draft the cooperative
agreements’ RFA (Rwanda 04, Implementer). The implementers refined the project design in response
to the RFA and during post-award and ongoing meetings with both the GOR and USAID. From large-
scale changes such as the actual project intervention zones, to adapting health facility training manuals to
local contexts (USAID 2010), interventions were modified and adapted as needed. Most stakeholders
from the district to the national level, USAID, and civil society organizations, were engaged in the
project design process (USAID 2010).
3.2 Pre-implementation
3.2.1 Implementation groundwork
The implementation groundwork was influenced by the characteristics of the implementing organizations
and project planning. Twubakane was IntraHealth’s first bilateral project. While many respondents
appreciated the sense of “family” and collaboration the prime created, several respondents recognized
how much IntraHealth learned from RTI. This was made possible especially due to the close
collaboration between the IntraHealth COP and the RTI senior staff member on the project team
(Rwanda 09, Implementer).
“I think we were all learning also as an organization, that being our first bilateral that we were really
leading and responsible for, drawing on the lessons that we learned from being a sub on other bilaterals
in terms of the partnership and how we engaged partners and counterparts at the district and central
level and what then were going to be the activities that were really going to move us farther along.”
Not only the project design (see section 3.1.5), but also the structure, was unique for a USAID project.
It appears no one was hired as DCOP, though one respondent did mention RTI’s “acting COP”
(Rwanda 05, Implementer). Further, there were only two expats on the RTI team; the American lead
and a Senegalese expert on health facility management (Rwanda 01, Implementer). One respondent felt
the structure was a key to success that should be replicated.
“I really liked this project and I think about it when I am writing proposals and designing projects. I liked
that there was a COP and then there were these 5 team leads. There wasn’t an actual DCOP. I really
liked that structure. I think we had weekly or bi-weekly senior team meetings with the COP and us 5
team leads. That got the 6 of us on the same page. I really liked that and I think USAID always requires
this COP and a DCOP and no one knows where to put the M&E team and nobody knows how finance

19
fits in or finance is sort of pushed to the side and it’s vertical, so I really liked that structure with these 5
senior team leaders.” (Rwanda 05, Implementer)
Planning was done as a larger project team, including MOH. The parliamentary liaison shared that the
MOH was the final step in approving Twubakane workplans, “because most the activities were implemented
supposedly with MOH so there would be a consultation about those activities” (Rwanda 08, Implementer).
“Team spirit” was cultivated and valued, from the annual work planning process down to individual
project activities. This facilitated the visibility and impact of the project.
“At one point we had more than 70 staff members and innovative things we did was once a month we
had an all staff meeting and that meant all staff including the drivers and the accountants…I remember
because the first time we had the all staff meeting all the drivers were in the parking lot and all the
accountants were back in their offices. It was just the technical team and I thought, ‘is this everyone?’
They were like, ‘no they are not’. They wanted the whole team and wanted the drivers to understand
the project. It was great because the drivers became so engaged that they asked for additional work. For
example we gave them all training in photography so that when they went out instead of sitting in their
car when we were in the health center, they could walk around and take pictures. I mean, not that they
were great pictures, but they felt like they were more engaged in the project and the goal was kind of
too, that everyone who worked there was an ambassador and spokesperson for the project and should
at least be able to describe to their friends, family, and neighbors what the project was all about. They
probably thought were really weird at first (laughs), you know what I mean because we tried to make it
less hierarchical than maybe Rwandan culture was used to, but in the end I think it really sort of pulled
the team together.” (Rwanda 04, Implementer)
One consistent perspective across respondents and project documents is the challenging environment
resulting from rapid decentralization efforts in 2005-2006. But Twubakane’s ability to implement the
project regardless, with only some delays to start-up of planned activities like the DIF, are what
respondents felt made it so successful.
3.3 Implementation
As indicated in section 3.1.4, IntraHealth was the prime, heading up the three components related to
health and performance improvement. RTI was the primary subcontractor, responsible for the three
components related to governance. They also conducted an evaluation of Twubakane’s health
governance outcomes in 2009 (Rwanda 11, Implementer). Tulane and EngenderHealth were also
engaged, though project documents and respondents did not elaborate much on their role. Our
research indicates that many additional local implementers were involved, but it was not possible to gain
a clear picture of every organization’s role and responsibilities. A clear theme across implementers was
strong collaboration with Rwandan counterparts and the ability to adapt and be flexible as the political
climate and GOR expectations changed.
Note, we interviewed staff who worked for the prime (IntraHealth), the sub (RTI) and the Ministry of
Health. We mostly learned about implementation of the decentralization process, and less about the
implementation of technical assistance and more targeted capacity-building activities.
3.3.1 Execution process
For most respondents, the theme of collaboration with both GOR and district officials was part and
parcel of each phase of implementation. In fact, a key strategy was that the project integrated its
activities into existing health system programs, as required by the government.

20
“The Twubakane activities had to be integrated into the country’s plan. We could not propose new activities,
but rather we integrated the activities into existing health system programs, those of the local administration,
and those of civil society. There was something that all implementers at the district level had to do;
something we call the ‘action plan’ or the ‘district development plan’. Using the district development plan,
we would integrate our activities into existing programs to further reinforce them.” (Rwanda 03,
Implementer)
According to one respondent, start-up involved implementing innovative training approaches as soon as
possible and relaunching some family planning policies which had not been updated since 1994 (Rwanda
04, Implementer). Twubakane staff did so by participated in government workgroups like the Health
Sector Cluster Group and Decentralization Sector Cluster Group (USAID 2010). They also developed a
Joint Health Sector Operating Plan together with other donors and partners (Rwanda 04, Implementer).
With the workgroups, Twubakane staff worked to ensure the project supported and aligned with
existing Rwandan policies like the Economic Development and Poverty Reduction Strategy, Vision 2020,
Health Sector Policy and Strategic Plans (2005-2009) and the Decentralization Strategy (2004–2008).
Part of this work was also informed by the Twubakane Program Steering Committee (USAID 2008b).
“The dialogue between the Twubakane project team and the counterparts at the central and local level,
for me was important. The management of the project was discussing the counterpart’s view on the
priorities…the government was separated and I was looking [at] how the project was in the national
plan and having a valid plan, some common agreement for [an] acceptable plan.“ (Rwanda 06, MOH)
At the local level, Twubakane staff worked hand-in-hand with district officials from planning to execution
of capacity-building activities to ensure ownership and sustainability of the work.
“We met once a month for just a debriefing about what is going on as the routine staff meeting, but
they [Twubakane staff] are fully there in case they need a technical representative from the program
level. So the opportunity was there, but there was technical support who was on a daily basis assisting
and identifying the key challenges in the districts to make sure that that issue, or planning at the district
level [was implemented] from [the] bottom up instead of what was used from the Ministry to the
districts.” (Rwanda 08, Implementer)
Similarly, the Partenariat pour l’Amélioration de la Qualité (PAQ) (an IntraHealth addition to the
proposal, not specified in the RFA), or community-provider partnerships approach, was particularly
valued by project implementers. This initiative engaged community stakeholders (i.e., teams of health
center managers, providers, and community representatives associated with community health centers)
to meet regularly to discuss solutions to gaps in their health care service delivery (USAID 2009a:3). One
respondent complimented the project explaining, “Twubakane used a bottom [up] approach to plan its
interventions, taking into account the real needs of end users” (Rwanda 10, Implementer).
“It did have a strong community component – there was the whole PAQ component that was very
critical in the whole design as well in terms of community engagement on the governance piece, but why
weren’t people accessing services if they were there? Well, they were treated poorly or the health center
was a mess…starting to have much more of that dialogue between community leaders. I know they put
in place those community groups that had teachers and religious leaders and strong women’s groups
that could then dialogue more with the health sector.” (Rwanda 09, Implementer)
While critical to project success and a buttress for the “bottom-up approach”, the numerous work
plans required of the district officials and the Twubakane project overall were a source of frustration for
many respondents. Not only were they numerous, but the requirements changed over time. Twubakane
had to report progress towards multiple operating plans for the MOH, while USAID PEPFAR and PMI
had completely different formats and expectations for their work plans. Further, few if any of the work

21
plan timelines aligned. Because many district officials were responsible for these work plans, they did not
always have time to fully engage in Twubakane’s capacity-building activities.
“At the national level there were always discussions. I don’t know if they were always before the work
planning or after the retreat and before submission there were discussions, but Rwanda had its own
work planning process and joint action work plan. One of my complaints is that there were too many
work plans. There was a work plan for the malaria unit and a work plan for the reproductive health
unit, and they needed all of our work to go in there, but that was usually after USAID needed to see a
version of a work plan.” (Rwanda 05, Implementer)
Feedback and refinement was highly valued by the project. Built into the RFA were not only
participatory SWOT and situation analyses used throughout the course of the project to align local
needs with national strategy resource allocation (USAID 2010), but also regular project activity
evaluations. Assessments of the PAQ approach and DIF were conducted. A final rapid facilities
assessment in all health centers and hospitals in the program’s partner districts were also conducted at
the project end (USAID 2010). An assessment of the financial viability of mutuelles resulted in a
nationwide audit by the Government of Rwanda of all mutuelles (USAID 2010). Furthermore, once the
rapid decentralization process had concluded, the project proposed to, and was approved by USAID, to
conduct an evaluation of its governance work. One respondent elaborated that the governance
evaluations’ respondents may not have been as honest as possible, given the culture of compliance.
“The issues around voice and transparency; because of Rwanda’s history, the citizen engagement is very
much structured by the state. If you want to talk about service delivery that’s ok, but anything that veers
into anything related to politics is kind of a no-no and as you’re probably aware, there’s a law that
makes any comments about genocide and all that against the law, so you can’t bring up any sort of that
stuff.” (Rwanda 11, Implementer)
Regular District-level capacity assessments were also conducted. The project responded to the
identified needs on multiple occasions. At times, this meant moving resources from a higher-performing
district to one with significant capacity-building needs, or obtaining the additional Packard funds to study
the family planning context (Rwanda 04, Implementer) and engage parliamentarians and mayors in the
family planning work (Rwanda 08, Implementer). Twubakane capitalized on close collaboration with all
stakeholders and sharing and dissemination of project documents so that those districts the project was
unable to reach with human resources were still able to benefit (Rwanda 01, Implementer).
3.3.2 Changes
Most respondents felt the project was executed as planned, with the exception of the change in
implementation zones due to decentralization. Two respondents mentioned project innovations that
may not have been part of the RFA; creating secondary posts to increase access to family planning
services and training district accountants.
According to a parliamentarian, “in Rwanda, 44% of health activities are owned by religious leaders who don’t
accept some contraceptive methods” (Rwanda 08, Implementer). To increase access to contraceptives,
Twubakane worked out an agreement with religious leaders where patients from religiously-funded
clinics would be referred to secondary posts when they needed contraceptive services (USAID 2010).
At one point, at the request of the MOH and USAID, Twubakane began training providers on a new
type of implant for female contraceptives and introduced non-scalpel vasectomies (Rwanda 05,
Implementer). The respondent seemed to think these were not outlined in the RFA or proposal.
One unanticipated need that arose was financial training for district accountants. As previously
mentioned, GOR was not entirely comfortable with the amount of money offered to the districts

22
through the DIF. Thus, MINILOC instituted a “fiscal decentralization” and Twubakane assigned district
accountants from the Ministry of Finance to, “understand how to finance the technical work at the
decentralized level” (Rwanda 03). In the beginning, it seemed accountants were almost afraid to spend
the DIF (Rwanda 04, Implementer), possibly due to fear of the zero tolerance policy towards
corruption. The project then identified a need to build their capacity for budget management and
resource allocation.
“Training accountants, we didn’t plan on it in the beginning, but we ended up training all of the district
accountants on how to do their whole budget because they didn’t have the skills necessary, so hopefully
that capacity remained. When we first gave the district grants the first year, every district didn’t know
what to do with $100,000, so they all wanted to do things like build latrines, which there’s nothing
wrong with that, but they gradually over the life of the project started putting that $100,000 into things
like training CHWs or things that had maybe more of a public health impact and not sort of just
building things.” (Rwanda 04, Implementer)
One respondent mentioned that trainings were conducted with police officers (Rwanda 05,
Implementer) and health center managers in Kigali to encourage police officers to collaborate with
health workers to raise awareness of and better respond to gender-based violence. This initiative, not in
the original RFA or proposal, was funded through PEPFAR funding awarded to the project in later years.
3.3.3 Actor engagement/dynamics
Overall, the project was collaborative and full of team spirit. “Mutual respect” and “understanding” were
frequently used to describe actor dynamics, especially in reference to IntraHealth and RTI. Only two
respondents described the dynamics with local partners in-depth. Both respondents agree that these
were strong partners, but the development partners’ competing agendas and funding often impacted
local partners’ commitment to Twubakane.
“In terms of local partners, RALGA at first was a very good partner, but they had a change in director
and some staffing and it became a little more difficult to work with them. It became a little sort of
distracted by other partners that they were working with like the Canadians, the Swedes were giving
RALGA certain directives in terms of their support and so we felt that RALGA kind of got distracted
towards the end of the project...” (Rwanda 02, Implementer)
“I would say locally it was challenging. We had two main local partners that actually had budgets. One
of them was called Pro-Femmes, which was sort of a women’s leadership organization and then the
other one is called RALGA an acronym for Rwanda Association of Local Governing Authorities…Pro-
Femmes was good at doing some community mobilization, but they were challenging to work with. I
think it’s always a struggle when a local organization sees that you are a 30-million dollar project and
they want more money and they want more resources, they think they should get them. RALGA was
challenging in its own way, although I think it was very creative and important to have that local
association of mayors as being on the project team, so with them we were able to do a lot of creative
work around sort of anti-corruption campaigns or training district mayors and their teams and planning,
budgeting, etc.” (Rwanda 04, Implementer)
That said, there was a sense of ownership of project tasks, especially for implementers involved in the
training and capacity-building of the districts. One implementer felt, “that this design of the project and
funds went into the districts and the districts were really accountable for them and had a lot of authority
around what those funds would be spent on” (Rwanda 09, Implementer). This ownership was partially
built through strong project communication. All but one respondent agreed that the prime did an

23
excellent job communicating. A respondent from the finance team working across all aspects of the
project felt that both communication and integration were lacking.
“An F&A team is often not well-integrated, so that was part of what I did, was hold these coordination
meetings every other Friday. I guess as far as implementation goes, part of the problem was that when
they brought me in, all of the different teams were doing all of the things in the work plan, but no one
was communicating about if the malaria team was in this region training people and the FP team is in
this region and the community health team goes out, you sort of inundate an area all at the same time
because people were not really coordinating…So one of the things I did was make everyone come back
from wherever they were in the field every other Thursday so that we could have a Friday meeting with
all of the team leads.” (Rwanda 05, Implementer)
Work plans, meetings, and annual retreats were used extensively to manage the project and facilitate
communication. One respondent felt planning meetings with central and decentralized-level
representatives effectively built capacity, especially for district-level administrators (Rwanda 02,
Implementer). Eventually, the time spent in meetings became burdensome for everyone involved in the
project, from implementers to the target populations. MINECOFIN tried to address this by issuing a
directive to all development partners that local counterparts should not be expected to participate as
frequently in meetings as it was limiting their availability for actual work (Rwanda 02, Implementer).
Similarly, the project Steering Committee3 had trouble finding times that aligned with everyone’s
schedules and thus did not meet as frequently as anticipated (IntraHealth2007:32).
Respondents overwhelmingly reported GOR support of the project; several named Jean-Damascène
Ntawukuriryayo4 a champion of the family planning work. However, like many countries, changes in
GOR leadership impacted GOR priorities. Twubakane was able to successfully navigate these shifts by
staying neutral and engaging all ministries to the extent possible.
“We tried to bring MINALOC and MINISANTE together for joint planning at meetings, but it never went
above a certain level of government…At the senior level they only wanted individual meetings; they
refused to go each other’s Ministry. That’s just the way it was…I think it was prestige, like “who owns
this meeting?” it was about ownership and who had to go where – MINISANTE has to go to MINALOC
or MINALOC has to go to MINISANTE, no neutral ground…There were political sensitivities between
the Ministries, that’s for sure. The MOH itself was trying to figure out what it’s role was because
suddenly [due to rapid decentralization and empowerment of the districts] it was no longer ‘in charge’ of
a health system, it was pretty drastic as I recall at the time and things were in turmoil in the government
because of this new drastic conceptualization. So we had to play it carefully and play a neutral role.”
3.4 Maintenance and evolution
3.4.1 Sustaining implementation
An important element of Twubakane was to ensure that eventually, the GOR would be able to sustain
the gains made by the project. To ensure this, Twubakane staff 1) aligned project activities with
3 The Twubakane Steering Committee was comprised of the MINALOC and MINISANTE permanent secretaries,
provincial representatives and USAID (IntraHealth, 2007).
4 Jean-Damascène Ntawukuriryayo preceded Agnes Binagwaho as Minister of Health. He was elected to parliament during
the course of the project.

24
government priorities and 2) planned for effective stakeholder engagement at all levels during the design
phase. Several interventions are still in place today, even though the project was only awarded for five
years. What may have contributed to these lasting effects is the fact that projects with similar objectives
were awarded to Rwandan development partners during Twubakane’s project period. Multiple
respondents mentioned the Maternal and Child Health Integrated Program (MCHIP) which began in
20085. Respondents suggested that MCHIP’s activities overlapped with Twubakane’s. Some respondents
worked on both projects, because they were hired by MCHIP when Twubakane ended, and had trouble
distinguishing project activities and outcomes between the two.
At the end of the project in 2010, over 70 local staff reintegrated back into either MINISANTE or local
government jobs and about half got absorbed into other donor projects or USAID programs (Rwanda
02, Implementer). Most local Twubakane staff still work for NGOs today, evidence of the lasting effects
of collaborative project management. While it is common for local staff to continue working with NGOs
after a large donor project ends, respondents reflected that activities related to community engagement
and policy and management have had a unique and lasting impact. Mutelles and PAQ (Rwanda 06, MOH)
remain intact today. PAQ was adopted in Senegal and Nigeria (Rwanda 03, Implementer) and was
declared a best practice by MINISANTE in 2006 (USAID 2010). Twubakane was an early adopter of
HRIS which several respondents felt was innovative and groundbreaking. Since Twubakane, the number
of HRIS users continues to increase (Rwanda 06, Implementer).
One respondent believes Twubakane’s district-level work has been leveraged by other projects since
Twubakane ended, but did not specify which activities (Rwanda 04, Implementer). USAID’s Democracy,
Human Rights and Governance (DRG Center) included Twubakane in a study on integrating governance
into sectoral projects (Rwanda 11, Implementer) and has since mentioned Twubakane to other missions,
“as a model to sort of emulate for the sort of integrated approach” (Rwanda 11, Implementer). Maybe
most importantly, health indicators continued to improve after the project ended (Rwanda 04,
Implementer).
The primary risk to sustaining impact was the lack of a follow-on award. The 1) desire by GOR that
international NGO-led projects be limited, 2) growing suspicion of NGOs (Rwanda 04, Implementer)
and 3) shift in USAID funding priorities from decentralization to HIV prevented a follow-on award
(Rwanda 05, Implementer). Further, the nascent nature of district-level governance and their frequent
staff turnover meant that there was limited staff and institutional capacity to continue activities. One
responded learned that since the project end there, “was a scandal around the community based health
insurance, that there had been a lot of local level embezzlement and all and I think partly that was a
function of having pushed through so fast, spreading things out soon without really without putting in
place controls and all…” (Rwanda 11, Implementer).
The regular celebration of successes and collaborative identification of challenges and room for growth
is evidence of the project’s emphasis on documentation (Rwanda 11, Implementer) and dissemination to
ensure sustainability. Engaging “the right ministries, including parliament” (Rwanda 07, MOH) was also
important, as was seeking regular feedback on process indicators and other outcomes and adapting
activities as needed. The number of system-wide evaluations conducted by the project is evidence of this
5 “MCHIP supports programming in maternal, newborn and child health, immunization, family planning, nutrition, malaria
and HIV/AIDS, and encourages opportunities for integration of programs and services when feasible.” USAID,
http://www.mchip.net/node/1

25
3.5 Challenges
The wider environment of decentralization was by far the most significant challenge cited by the eleven
respondents. Resulting shifts in the implementation zones and the ever-evolving GOR expectations due
to staff changes required flexibility and adaptability from the project team. GORs’ misunderstanding of
Twubakane funding structures, especially the why the DIF were so well-financed, created some tensions
in the beginning, but these appear to have subsided over time.
“Yes, we had challenges. The first was adapting to this new program that was unique to Rwanda where
everyone wanted to claim their stake and we had to clarify, ’no, I’m the one managing this‘, ’no, I’m the
one …’ I would explain to the government that this is the American system. We cannot add these funds
to our national budget. Initially, this was a challenge. The second challenge was at the decentralized
level, how to – we had to- how can I describe this? We owed $500,000 to each district. They were each
responsible for the use of these funds, but they also had to show us their spending plan based on the
district needs and performance. What ended up happening, however, was that in some districts we were
unable to confirm their performance during that period of time. Instead of returning that money to
USAID, we gave those districts’ funds to the highest-performing districts.” (Rwanda 03, Implementer)
GOR work plan expectations, especially the numerous indicators both Twubakane staff and local
government officials had to report on, cut into time that could have been spent on trainings and other
project activities. In terms of training, one respondent suggested lack of capacity to be the greatest
challenge the project addressed (Rwanda 07, MOH). Similarly, another responded felt the delay in DHS
data collection was one of the greater challenges in determining project success.
“What’s really working from an evidence service output/service quality was a challenge for the project.
You always monitoring data and things like that, but I remember waiting for the DHS to really know,
had modern contraceptive prevalence had gone up? And what was happening with home-based
management of fever and rolling out new protocols, because I think they were doing presumptive
treatment of malaria with the babies and the children and then the new protocol came out and it wasn’t
presumed, you had to validate, so you needed the RDTs. So I think some of the challenges are just the
right clinical advancements and rolling them out, but I think the project actually responded quite well too
many of them.” (Rwanda 09, Implementer)
Finally, Twubakane’s visibility contributed to GOR-Twubakane tensions, as did the government’s distrust
of NGOs towards the end of the project (Rwanda 04). As previously mentioned, these were also
contributing factors to Twubakane’s failure to be re-awarded.
3.6 Success Factors
As introduced above, respondents reported several success factors around collaboration,
communication, engagement, and implementation:
Collaboration, communication, and leadership:
 Strong, collaborative project leadership which was open to learning as it was their (IntraHealth
International’s) first major bilateral project as the prime.
 Excellent communication with “really high-functioning, high-performing teams and team leaders”
 Mutually respectful team-building approach integrated across project activities, from proposal-
writing to project implementation and closeout. This required flexibility and adaptability.
 Effectively communicating the importance of capacity-building activities to district-level officials.

26
Project planning and implementation:
 Twubakane’s bottom-up approach to project planning and implementation.
 Aligning the RFA with existing GOR initiatives and incorporating activities to engage consistently
in GOR policy development and implementation.
 Integrated health and governance approach; “health systems strengthening”.
 Allowing for innovation and at times, piloting activities at the encouragement of USAID, largely
due to the fact that this was a cooperative agreement (as opposed to a contract).
 Ability to seek out and secure additional funding from the Hewlett Foundation, the Packard
Foundation, and others.
 Integration of regular monitoring and evaluation, documentation, and dissemination.
Engagement:
 Successful engagement of key stakeholders, especially the highest levels of government, the most
local staff, including project drivers, and USAID.
 Identifying and supporting project champions at the all levels; from President Kagame to
journalists asked to engage in a photography contest to promote family planning activities
(Rwanda 04, Implementer). Mayors were especially important for DIF and family planning
activities.
3.7 Dissemination
The prime discussed Twubakane’s branding and dissemination at length with GOR in order to promote
government ownership of the policies and programs that facilitated positive changes. The project put
significant effort into documenting and distributing important project information to hand over not only
to USAID but also those reached by Twubakane. As noted earlier, USAID is said to have touted the
project as an example of health systems strengthening but none of the respondents named specific
USAID-funded projects that had emulated Twubakane. But at least one respondent has taken lessons
learned about building financial capacity and the “power of the cross-sectoral programming” through
DIF to other countries moving from a centralized to a decentralized model (Rwanda 09, Implementer).
To ensure Twubakane successes were recognized on a larger scale, the project held a close-of-project
presentation for the USAID mission and delivered multiple presentations to the Global Health Bureau in
Washington, DC (Rwanda11, Implementer). Finally, there was an End-of-Project (EOP) Ceremony with
nearly 200 participants, during which lessons learned were discussed. Participants included the Minister
of Health, the US Ambassador, and the USAID Mission Director, as well as the MOH’s Permanent
Secretary, the governor of the southern province, mayors, hospital directors, other representatives
from all 12 of the Twubakane-supported districts, and a variety of partner organizations (USAID 2010).
3.8 Lessons learned
Many respondents seemed to walk away with a better sense of how to implement health systems
strengthening (HSS). Though few used the term HSS, several mentioned “integration of health and
governance” and a recognition that the six WHO Health Systems Framework Building Blocks are
“necessarily intertwined” (Rwanda 04, Implementer). Yet another respondent came away realizing that it
was not only important to partner with GOR, but rather to follow and support the GOR’s goals.

27
“I do think that the biggest challenge was this, and I don’t want to over play it, is this project versus the
government’s program. That I think in some ways was a little bit painful for us afterwards, that we really
had to- maybe that was like the hardest lesson to learn, that the emphasis was that we were really
accompanying the government and partnering with government, but it’s really about the government’s
program and their goals.” (Rwanda 09, Implementer)
Some additional lessons learned that respondents reported include the value in:
 Regular communication between team members and stakeholders.
 Building highly effective teams through regular communication and in-person retreats. Most
respondents mentioned the positive impact retreats had on the team dynamic.
 Stakeholder engagement, from GOR officials to local community members, in the planning and
implementation process.
 Convincing the target population of the importance of the activity. Twubakane was able to
improve district-level financial planning and administrative capacity because district officials
understood and agreed that these were important. One respondent applied his experience from
Twubakane to his work creating Water User Associations in Senegal (Rwanda 02, Implementer).

28
4. DISCUSSION AND SYNTHESIS
In this section we discuss our results and synthesize the key factors that led to the successful
implementation of the project.
4.1 Synthesis
Here we discuss both facilitators and barriers to the success of Twubakane’s implementation. The
facilitators build on the integrated, system-level, bottom-up, adaptable, mutually respectful project
approach described in earlier sections.
4.1.1 Facilitators
Engagement of MINECOFIN (Ministry of Finance) in health issues. Family planning is often a
very political issue, thus requiring more stakeholder buy-in than other project activities. Twubakane’s
engagement of MINECOFIN in these issues not only increased their awareness, buy-in, and financial
support of family planning activities, but was also an innovative and unique approach.
“I think this integration, I mean you don’t see it that often, that a project is doing this decentralization, that’s
what we call it but a lot of it is health systems and health management capacity strengthening with the
health services all the way from hospitals to the CHWs. It was a big project, but I think that integration
made people have dialogues that we wouldn’t normally and I would say in most projects I’ve ever worked on,
you work with the MOH and I’ve not since or ever before engaged with the Ministry of Finance; it’s not in my
job description. But I think [with this] project you had to, there was engagement with MINECOFIN, which
means we had a better understanding of some of the issues between MINECOFIN and the MOH and some
of those higher levels systems that I personally have not been privy to.” (Rwanda 05, Implementer)
District-level capacity building. Twubakane’s cross-sectoral design and bottom-up approach was
especially useful at the district level. Using its cross-sectoral design, Twubakane’s district health teams
were able to work with their counterparts in other agencies to “really advance the particular challenges
of the district” (Rwanda 09, Implementer). To achieve district buy-in, Joint Action Work Plans were
described as time-consuming, but also required in order to be “truly meaningful to local actors and to
have any chance of actually being implemented” (USAID 2010).
Highly capable local staff. Many local project staff received either graduate-level education in Rwanda
or education abroad as refugees of the 1994 genocide. This enabled them to effectively implement
Twubakane activities (Rwanda11, Implementer).
4.1.2 Barriers
Reluctance to implement pilot projects. At some point, one respondent felt that it was unwritten
GOR policy that interventions were only allowed to be tested and implemented if they were evidence-
based [and could be scaled nationwide] (Rwanda 05, Implementer). Although Twubakane, with the buy-
in of the GOR, was able to introduce several innovations, the GOR’s reluctance to fully support pilot

29
activities did limit the level of innovation the project could introduce, despite the encouragement of
USAID to experiment pilot activities.
Substantial reporting requirements. Some respondents implied that they were not always able to
spend as much time implementing as they would have liked due to the extensive amount of time
necessary to write the reports required by USAID due, in large part, to the integrated nature of the
project and change in USAID/Rwanda staff throughout the five-year project.
The project’s introduction of best practices like PAQ, which were adopted and incorporated into GOR
quality assurance guidelines, reinforced Twubakane’s commitment to both influencing and aligning with
GOR policies and initiatives. This remained a strong facilitator for the projects’ success throughout its
duration. Stakeholder engagement from the RFA development through the end of the implementation
phase was critical to sustaining the projects’ gains once it ended. Not only were stakeholders engaged,
but they actively participated, communicated, and shared lessons learned. This created a family-like team
dynamic where leadership was respected, but everyone was valued. Over the course of the project,
roles and responsibilities became more defined. This further strengthened both project- and health
system-level capacity.
4.2 Limitations
This report is based upon the views of eleven respondents. Among these, most were implementers who
worked for either the prime or the sub-contractor during Twubakane’s implementation. The others
worked for MOH at the time. None of the district-level staff, health care providers, or community
leaders’ views are represented. Furthermore, there is a risk of recall bias. Interviews were conducted in
2016, but the project ended in early 2010.
4.3 Conclusion
The Twubakane project helped successfully decentralize the Rwandan health service delivery system and
subsequently strengthen district-level health service delivery capacity, resulting in improved health
outcomes.
By aligning early and often with GOR priorities, the project ensured sustainability. Using an informed
and respectful approach to the country’s unique history and cultural context, project staff were able to
collaborate closely with stakeholders at all levels. Twubakane’s commitment to ensuring staff
understood their roles and responsibilities allowed the project to implement with excellence.
In this way, the project helped create a more transparent and responsive health governance process to
ensure that women and children especially, are better served by a quality health care system.

30
ANNEX A: COMBINED IMPLEMENTATION FRAMEWORK
Phase Domain Factor Description
Unit of
analysis
1
Pre-condition
Enabling
environment
Wider
environment
Economic, political, social, and health
system context within which
intervention6 is implemented
National/regional
context
External
policies and
incentives
Strategies to spread intervention – policy,
regulations (not directly implemented by
project but (pre)existing)
Policies that constrained implementation
Other donor led initiatives that
complement intervention
National/regional
context
Implementation
setting
Characteristics
of organization
Structural characteristics of organization
such as social architecture, age, maturity,
and size of organization
Culture of organization such as norms,
values, basic assumptions of organization
Change
target/larger host
organization7
(identify for each
case; e.g. MOH)
Implementation
climate
Climate within organization, including
relative priority of project, readiness for
implementation, learning climate, and
policies, procedures, and reward systems
that inhibit or facilitate implementation
Change
target/larger host
organization
(identify for each
case; e.g. MOH)
Project design
Intervention
source
Stakeholder perception if intervention
internally or externally developed
As applicable for
each case (e.g.
MOH, local
partners, change
target)
Identification of
effective
intervention
Process for deciding intervention
approach and activities
Stakeholder perception of quality and
validity of evidence that intervention will
have desired effects
Perceived relative advantage and
complexity/perceived difficulty of
As applicable for
each case (e.g.
MOH, local
partners, change
target)
6 The total package of activities that is implemented by the project.
7 Institution within which activities are being implemented; may be MOH or other local organization (will focus on larger
organization like MOH rather than individual hospitals); depending on the case this organization may be more or less
involved in the actual implementation.

31
intervention
Adaptability Degree to which intervention was
adapted to local needs, including degree
to which beneficiaries’ needs were
understood and design was adapted to
meet their needs
Project
implementers8 (e.g.
prime + subs)
Draft package Perceived quality of how intervention is
presented
As applicable for
each case (e.g.
MOH, local
partners, change
target)
2
Pre-implementation
Implementation
groundwork
Structural
characteristics
of implementing
organization
Structural characteristics of implementing
organization such as social architecture,
age, maturity, and size of organization;
culture of organization such as norms,
values, basic assumptions of organization
Project
implementers (e.g.
prime + subs)
Implementation
climate
Climate within project including relative
priority of project, readiness for
implementation, learning climate, and
policies, procedures, and reward systems
that inhibit or facilitate implementation
Project
implementers (e.g.
prime + subs)
Planning Degree to which intervention is planned
in advanced, quality of methods;
refinement of draft package based on
pilot testing, stakeholder feedback
Project activities
Orientation and
logistics
Quality of initial planning and execution
of the project, including needs
assessment, pilot testing, leadership
engagement
Project activities9
3
Implementation
Implementation
Executing Fidelity of implementation Project activities
Engaging How the project attracted and involved
appropriate individuals throughout
project: opinion leaders, formally-
appointed internal implementation
leaders, champions, external change
agents
Project activities
Feedback and
refinement
Qualitative and quantitative feedback
about progress and quality of
implementation
Refinement of activities based on
feedback
Project activities
8 Prime contractor and sub-contractors (may include local subs) who implement the project. This does not include the
change target organization.
9 Specific activities directly implemented by the project implementers. These may or may not align with other activities in
the change target organizations. These individual activities make up the intervention as a whole.

32
Cost Costs of total intervention - planned and
actual
Intervention4
Maintenanceandevolution
Sustaining
implementation
Organizational,
financial
changes
Changes made to sustain the intervention Project
implementers (e.g.
prime + subs);
Project activities
Re-customize
delivery as need
arises
Adapting the intervention delivery as
circumstances change
Project
implementers (e.g.
prime + subs)
Dissemination
National
dissemination
Preparing refined package, training, and
TA program for national dissemination;
was project nationally disseminated
Project
implementers (e.g.
prime + subs);
Change target

33
ANNEX B: KEY INFORMANT INTERVIEW GUIDE
Instructions
First complete informed consent to conduct interview and ask permission to record.
Ask as many of the primary questions as is feasible given the time constraints and as are appropriate for the respondent
given their role in the project. Ask probe questions as applicable. Prioritize the most important questions if you do not have
sufficient time to ask all applicable questions.
Respondent’s role
1. Can you tell me about your involvement with (PROJECT)?
a. When were you involved with (PROJECT)?
2. Who were you working for during that time? (e.g. Implementing partner (specify); USAID Mission; USAID
HQ; government counterpart; other—specify)
a. What was your position or title with (PROJECT)?
b. Did you change organizations or positions during your time on (PROJECT)?
Pre-condition
3. What problem(s) was the (PROJECT) trying to solve?
a. Who felt this was an issue of concern? (e.g., MOH, US Mission, other stakeholders?)
b. Why did they see it as a concern?
PROBE: What evidence was this based on?
c. Was there a country/government initiative or reform targeting this issue that the (PROJECT) was
intended to support? Please describe briefly.
4. How did USAID decide to fund a project to address this problem? Who was involved in the decision?
a. What evidence was used to understand the issue?
PROBE: Evidence used by respondent or respondent’s organization, other partners, local
stakeholders, USG?
b. What approaches or activities did USAID specify in the RFA/RFP? (Skip if can answer from
documentation)
PROBE: Did other stakeholders contribute to what was specified in the RFA/RFP?
c. How did USAID decide what to include in the RFA/RFP? Did other stakeholders contribute?
5. How was this (PROJECT) selected to address (ISSUE)?

34
a. Who was involved in the selection?
6. Can you briefly describe the (PROJECT’s) approach and activities?
a. Which do you think were the most important activities?
7. During the work planning process, how were the specific activities used in (PROJECT) selected?
a. Who contributed to these decisions?
PROBE: Prime or subcontractors, US Mission, MOH, hospitals, (PROJECT) participants,
beneficiaries
b. What other information influenced the selection of the (PROJECT) interventions? (e.g. government
priorities, new USAID/USG initiative, existing policies/regulations, new financing, etc.)
c. Were other interventions considered but not selected?
d. How much consensus was there between stakeholders about the design of the interventions?
8. How were the intervention sites identified? (e.g. hospital, school of nursing, etc.)
a. Who contributed to these decisions?
9. How were the activities designed to be appropriate for the local health system context?
a. How were planned activities piloted?
b. How were planned activities adapted to existing conditions during the (PROJECT)?
Pre-implementation
10. Were there any individuals or organizations who provided strong support for the (PROJECT)?
a. How did they promote (PROJECT) implementation?
PROBE: Did they promote implementation at individual sites or for particular activities?
b. What are the reasons they supported the (PROJECT)? (e.g. specific to (PROJECT) or supportive to
larger country initiative?)
11. Were there any individuals or organizations who delayed or impeded implementation of (PROJECT)?
a. How did they impede (PROJECT) implementation?
b. What are the main reasons they impeded it?
12. Can you tell me about the dynamics of the individuals and organizations working on (PROJECT)?
a. How did these evolve over time?
Implementation

35
13. How were (PROJECT) activities implemented?
a. Were all the activities implemented in all of the project sites? (Skip if can answer from documentation)
b. Were activities implemented in phases? (If yes) What were the phases? (Skip if can answer from
documentation)
c. Did the (PROJECT) activities change over time? (If yes) Why? (Skip if can answer from documentation)
d. Were changes documented? (If yes) How? (Skip if can answer from documentation)
e. How did contextual factors affect implementation? (e.g. social, economic, political, technological,
etc.)
14. Was there consensus among different partners and stakeholders about how the (PROJECT) was
implemented?
15. Where did the resources for (PROJECT) implementation come from? (e.g. (PROJECT)/(PARTNER), USG,
government, others) (Skip if can answer from documentation)
a. Was there enough funding and other resources to support (PROJECT) implementation?
PROBE: financial, technical, human, technological.
b. (If there was a shortage of resources) How was the shortage addressed?
16. What challenges were faced during day-to-day (PROJECT) implementation?
a. Were there any issues with policies or regulations?
b. How did (PROJECT) address these challenges?
17. How were (PROJECT) activities monitored and/or evaluated? (Skip if can answer from documentation)
a. Who was responsible for monitoring implementation progress? Was this part of standard
implementing practices?
b. Was an evaluation conducted? By whom? Who requested it? Who paid for it?
c. How were findings from M&E incorporated into implementation?
d. What was the response to M&E findings?
18. What dissemination activities were undertaken during (PROJECT)? (e.g. small-scale meetings at (PROJECT)
sites, national workshops presenting findings, feedback sessions to USG, etc.) (Skip if can answer from
documentation)
a. How was feedback disseminated throughout (PROJECT)? (e.g. (PROJECT) participants, end-of-the-
line beneficiaries and policymakers)
Maintenance and evolution
19. What was done during (PROJECT) to support continuation of activities after (PROJECT) ended?

36
a. What role did (PARTNER) or others have in helping to sustain the activities?
b. What role did others play in sustaining the activities? (e.g. US Mission, MOH, intervention sites,
communities)
20. What is the current status of activities included in (PROJECT)?
a. Who has taken responsibility for sustaining the interventions? (e.g. financial, organizational, technical
responsibility)
b. What are the long-term prospects of the interventions?
c. What, if any, are the plans to scale-up/expand the interventions from (PROJECT)? (e.g. same
country, other settings)
Reflections
21. What do you think were the impacts of (PROJECT)? (e.g. changes in health status, improved service delivery,
increased quality of services.)
22. Were there any consequences from (PROJECT) that were unintended or unexpected?
23. What were some challenges to the overall implementation of (PROJECT)?
a. How could have these been addressed during the implementation period?
b. Do these challenges remain an issue today? Why?
24. What were the key factors that led to the success of (PROJECT)?
25. What are some lessons learned from implementing this intervention that you would take forward on other
projects of this nature?
26. Is there anything else we have not discussed that you would like to share about the implementation of
(PROJECT)?
27. Do you have any questions for us?

37
ANNEX C: DETAILED ACTIVITY TABLE
Intervention Area Components No. of sub-
activities
Family Planning 1. Revising policy
2. Increasing Access to and Use of Quality FP Services
3. On-the-Job Training for Family Planning
4. Introduction of Permanent Methods—non-scalpel
vasectomy
5. Creation of “secondary posts” near Catholic-run facilities
6. Integration of FP and HIV services
7. Improving quality of and access to safe motherhood
services
8. Improving Gender-Based Violence Prevention and
Response
9. Partnership with Pro-Femmes Twese Hamwe
9
Child survival,
malaria and
nutrition
1. Prevention and Control of Malaria
2. Integrated Management of Childhood Illness (IMCI)
3. Increasing awareness and number of immunizations for
vitamins
3
Decentralization
Planning, Policy,
and Management
1. Collaborate and support policy planning for MINALOC,
MINISANTE and RALGA
2. Health Sector Costing and Good Governance and Health
Studies
3. NHA
4. Health Management Information System
4
District planning,
budgeting and
management
1. DIF Grants
2. JADF
3. District Auditors’ Training and Orientation
4. Participatory Assessments of Districts’ SWOT
4
Health facilities
management
1. Revision of Health Care Policies, Norms and Standards
2. Strengthen Health Care Financing via Mutuelles
3. Capacity building of health Facilities Management
3
Community
access to,
participation in,
and ownership of
health services
1. Technical assistance for development of new national
Community Health Strategy
2. Community Health Worker Capacity Building
3. Community-Provider Partnerships (PAQ)
4. CHIS
3

ANNEX D: BIBLIOGRAPHY
Adam, Taghreed, Justine Hsu, Don de Savigny, John N. Lavis, John-Arne Rottingen, and Sara Bennett.
2012. Evaluating health systems strengthening interventions in low-income and middle-income
countries: are we asking the right questions? Health Policy and Planning 27:iv9-iv19.
Bernard, Russell H. and Ryan, Gery W. 2010. Analysing Qualitative Data: Systematic Approaches. Los
Angeles: Sage Publications.
Damschroder, Laura J., David C. Aron, Rosalind E. Keith, Susan R. Kirsh, Jeffry A. Alexander, and Julie C.
Lowery. 2009. Fostering implementation of health services research findings into practice: A
consolidated framework for advancing implementation science. Implementation Science 4: 50.
Conrad, Abigail, Joseph Naimoli, Sweta Saxena, Daniela Rodriguez, Catherine Connor, and Lauren
Rosapep. 2016. Understanding the Dynamics of Successful Health System Strengthening Interventions:
Study Design. Bethesda, MD: Health Finance & Governance Project, Abt Associates Inc.
“English To Become Official Language In Rwanda”. National Public Radio. November 20, 2008.
http://www.npr.org/templates/story/story.php?storyId=97245421
GHI. 2012. Global Health Initiative Principle Paper: Health Systems Strengthening, Draft 6. Washington,
DC: Global Health Initiative, the U.S. Government.
Hatt, Laurel, Ben Johns, Catherine Connor, Megan Meline, Matt Kukla, and Kaelan Moat. 2015. Impact of
Health Systems Strengthening on Health. Bethesda, MD: Health Finance & Governance Project, Abt
Associates.
IntraHealth. 2007. Twubakane Decentralization and Health Program Cooperative Agreement #623-A-
00-05-00020-00 Continuing Application. Chapel Hill, North Carolina.
Kilbourne, Amy M., Mary S. Neumann, Harold A. Pincus, Mark S. Bauer, and Ronald Stall. 2007.
Implementing evidence-based interventions in health care: Application of the replicating effective
programs framework. Implementation Science 2:42.
Miles, Matthew B. and A. Michael Huberman. 1994. Qualitative Data Analysis: An Expanded Sourcebook.
Thousand Oaks, CA: SAGE Publications, Inc.
“Rwandan Health Minister Dr. Agnes Binagwaho to Deliver the 2015 David E. Barmes Global Health
Lecture”. National Institute of Dental and Craniofacial Research. October 12, 2016.
http://www.nidcr.nih.gov/NewsAndFeatures/Announcements/Barmes-2015.htm
Rwanda Ministry of Local Government. 2011. Decentralization Implementation Plan (DIP) 2011-2015.
http://www.minaloc.gov.rw/fileadmin/documents/Minaloc_Documents/DIP_3rd_Phase_Draft_14.3.1
1_final_version.pdf
Strauss, Anselm, and Juliet M. Corbin, eds. 1997. Grounded Theory in Practice. Thousand Oaks, CA:
SAGE Publications, Inc.
Travis, Phyllida, Sarah Bennett, Andy Haines, Tikki Pang, Zulfiqar Bhutta, Adnan Hyder, Nancy
Pielemeier, Anne Mills, and Timothy Evans. 2004. Overcoming health-systems constraints to
achieve the Millennium Development Goals. Lancet 364:900-906.

39
USAID. 2006. 2005 Annual Report. Kigali: Twubakane Decentralization and Health Program.
USAID. 2006a. 2006 Annual Report. Kigali: Twubakane Decentralization and Health Program.
USAID. 2007. Twubakane Third Year Annual Report. Kigali: Twubakane Decentralization and Health
Program.
USAID. 2008. Study on the Readiness of the Political, Legal, Health and Community Systems to Respond
to Gender-Based Violence in Three Districts of the City of Kigali. Kigali: Twubakane Decentralization
and Health Program.
USAID. 2008a. Twubakane Fourth Year Annual Report. Kigali: Twubakane Decentralization and Health
Program.
USAID. 2008b. Quarterly Performance Monitoring Report #13 January-March 2008. Kigali: Twubakane
Decentralization and Health Program.
USAID. 2009. Good Governance and Health: Assessing Progress in Rwanda. Kigali: Twubakane
USAID. 2009a. Community-Provider Partnerships for Quality Improvement in Rwanda. Kigali:
Twubakane Decentralization and Health Program.
USAID. 2009b. Quarterly Performance Monitoring Report #17 January-March 2009. Kigali: Twubakane
USAID. 2010. District Incentive Fund (DIF) Grant Initiative Final Assessment. Kigali: Twubakane
USAID. 2010. Twubakane Final Report: Let's Build Together. Kigali: Twubakane Decentralization and
Health Program.
WHO. 2007. Everybody’s Business: Strengthening Health Systems to Improve Health Outcomes,
WHO’s Framework for Action. Geneva: World Health Organization.
WHO. 2011. Health Systems Strengthening Glossary. Geneva: World Health Organization.

Case Study: Rwanda’s Twubakane Decentralization and Health Program

More Related Content

Similar to Case Study: Rwanda’s Twubakane Decentralization and Health Program (20)

More from HFG Project (20)

Recently uploaded (20)

Case Study: Rwanda’s Twubakane Decentralization and Health Program