Assessing the progress and gaps in strategic health purchasing in Kenya [version 1; peer review: 1 approved with reservations]

Background: Kenya has committed to achieve universal health coverage (UHC) and prioritised purchasing reforms. Existing evidence indicates that purchasing can be leveraged to achieve health system objectives when done strategically. However, no single study has collated evidence on the progress and bottlenecks on strategic health purchasing (SHP) in Kenya to inform requisite SHP reforms in moving towards UHC. Against this backdrop, we assessed the progress and gaps in SHP in Kenya. Methods: We used a cross-sectional qualitative study design and collected data using document reviews, key informant interviews, and a stakeholder engagement workshop focussing on the three main purchasers in Kenya: the National Hospital Insurance Fund (NHIF), county departments of health (CDOH) and the Ministry of Health (purchaser for vertical programmes). Data was analysed following a framework analysis approach. Results: Kenya is making progress towards developing, implementing and institutionalising SHP practices that should lead to the attainment of health system goals. Progress has been made in 1) strengthening the process of defining a health benefits package available for the population, 2) engaging in selective contracting and integrated care contracting, 3) implementing multiple provider payment mechanisms, 4) designing legal and regulatory frameworks that provide direction for purchasing, and 5) increasingly utilising information systems. However, several gaps remain. First, there remain inadequate linkages between MOH policies and the NHIF that can facilitate adequate stewardship/oversight. Second, the purchasers lack adequate capacity to ensure quality assurance and enforce contracts. Third, provider payment methods (PPM) remain poorly designed, implemented, and are not adequately linked to health system goals. Finally, purchasers have weak or unclearly defined communications Open Peer Review Reviewer Status Invited Reviewers 1 version 1 13 Apr 2021 report Brendan Kwesiga , World Health Organization, Nairobi, Kenya 1. Any reports and responses or comments on the article can be found at the end of the article. Page 1 of 13 Wellcome Open Research 2021, 6:81 Last updated: 02 MAY 2021


Introduction
Universal health coverage (UHC) is a global health agenda that countries have committed to achieve by 2030 1 . UHC requires the whole population to have access to good quality preventative, promotive, curative and rehabilitative services relative to need while protecting them from experiencing any financial hardship as a result 2, 3 . Consequently, countries, including Kenya, are increasingly reforming their health systems in moving towards UHC 4 .
Kenya has committed to achieve UHC and prioritised purchasing reforms 5 . Purchasing refers to the transfer of pooled resources to healthcare providers to provide a defined set of services to the population 6,7 . Existing evidence indicates that purchasing can be the critical cause of inefficiencies and a barrier to ensuring the population can equitably access good quality services (if any at all) but can be used as a lever to achieving health system objectives 2 . Purchasing involves three decisions: 1) what to buy -the set of services or benefits package to be purchased, 2) whom to buy from -the set of healthcare providers to be selected to offer the defined set of services, and 3) how to buy -the provider payment methods and contractual arrangements 8 .
As opposed to passive purchasing -using the previous budget or simply paying bills when presented -countries are adopting strategic health purchasing (SHP) as a reform to move them closer to achieving UHC 7,8 . Strategic purchasing requires a continuous pursuit for better ways to implement the three purchasing decisions in a way that promotes health system performance.
Public sector purchasing in Kenya is conducted by three main purchasers including the National Hospital Insurance Fund (NHIF), the national ministry of health (MOH) and the county departments of health (CDOH) across 47 semi-autonomous counties ( Table 1). The NHIF contracts with public and private providers and uses various provider payment mechanisms. For instance, NHIF uses 1) capitation for outpatient services where providers are reimbursed different rates depending on both their facility level (primary or secondary or tertiary) and the scheme a member belongs (national or civil servants), 2) casebased payments for renal dialysis services, kidney transplant, free maternity program, oncology treatment, surgical services, and rehabilitation services and 3) fee-for-service for radiological, dental, optical services, inpatient and outpatient care for enhanced scheme members 9 . The national MOH owns tertiary hospitals and pays through global budgets whereas the CDOH owns both primary and secondary public health facilities and pays using salaries to healthcare workers, line-item budgets and supply of commodities.
No single study has collated evidence on the progress and bottlenecks on strategic purchasing (SHP) in Kenya to inform requisite SHP reforms in moving towards UHC. Consequently, this study describes the progress made on SHP and the gaps hindering progress in SHP in Kenya to also act as a baseline for future monitoring.

Study setting
Kenya is a lower-middle-income country (LMIC) with a governance structure comprising a central government and 47 semi-autonomous county governments that enjoy devolved fiscal, political and administrative authority 10, 11 . Kenya's health system's performance is lower than expected when compared against historical trends and its peers in areas such as equitable access to key services (such as reproductive, maternal,  [12][13][14] .
The health sector is managed at the national and county government levels; the national government is in charge of policy and regulatory roles and management of the national referral hospitals whereas county governments are responsible for health service delivery and ownership of public primary and secondary facilities.
Like many other LMICs, Kenya's total health expenditure (THE) is characterised by large proportions of out-of-pocket (OOP) payments (39.6% in 2016) and donor funds (23.4% in 2016) 15 . Nearly 20% of the population in Kenya are covered by some form of health insurance with 89% of these covered by the NHIF 16 .

Study design and data collection
We used a cross-sectional qualitative study design, employing a framework analysis approach, to assess the progress and gaps in SHP in Kenya. Three approaches were used to collect data: 1) document reviews (conducted in December 2019), 2) key informant interviews (conducted in November to December 2019), and 3) a stakeholder engagement workshop (conducted in January 2020). The consolidated criteria for reporting qualitative research (COREQ) checklist was used 17 .

Document review.
We searched, identified and reviewed documents on 1) previous assessments of healthcare purchasing in Kenya, and 2) health policies, roadmaps, strategic plans, and reports from MOH reforms task forces such as the health financing reforms expert panel (HEFREP). These documents were identified from 1) simple web searches in Google, 2) a search in PubMed, 3) references of the obtained documents, 4) the MOH and NHIF websites, and 5) grey literature. Web and PubMed searches included search terms such as "Healthcare Purchasing" OR "Strategic Purchasing" AND "Kenya" and was essential in retrieving peer-reviewed literature.

Key informant interviews (KIIs).
Using a semi-structured interview guide 18 developed from the study's conceptual framework (detailed below), we conducted five out of eight possible key informant interviews (KIIs) with purposefully selected participants based on their roles in organisations and/or in the MOH working on strategic health purchasing or related health financing activities) from the national ministry of health and other partners working on health financing and/or strategic health purchasing in Kenya. A list of organisations/partners working on health financing in Kenya was drawn and the number of potential respondents determined following the structure of the organisation/partner and previous engagements ( Table 2). Potential study participants were approached via email and then by telephone to notify them about the study and request their participation. Approval for participants' participation in the interviews was first sought from relevant organisational heads where applicable. The interviews were conducted in November to December 2019 by JK and EK in the respondents' offices. Before starting the interviews, respondents were presented with information regarding the study objectives, why they were selected, that their participation was voluntary and provided written informed consent. The interview questions focussed on the SHP areas where Kenya is making progress and where challenges exist across the conceptual framework's domains.
Workshop. The findings from the document review and key informant interviews were mapped onto a progress matrix informed by the conceptual framework for the study. The progress matrix presents in tabular form key purchasing functions as well as key steps that should be taken to move towards the most optimal arrangements for those functions (Table 3).  Table 3 summarises the conceptual framework and the key steps for making progress towards strategic purchasing in Kenya.
The framework comprises of three main categories: 1) governance arrangements, 2) purchasing functions and 3) other capacities. Governance arrangements focus on the 1) mandate and autonomy of the purchaser with regards to, for instance, decisions/functions the purchaser can carry out; 2) legal or regulatory frameworks -comprising of the laws and regulations available to govern the operations of the purchaser; and 3) governance structures that direct the day-to-day management of the purchaser and both the oversight and internal accountability lines of the purchaser.
Purchasing functions detail the key action areas that a purchaser should decide on including the set of service entitlements (benefits package) that the citizens/members will benefit, the set of providers the purchaser can buy from and the payment methods used to reimburse the providers. Other capacities in the framework relate to how the purchaser communicates with stakeholders and the use of information systems to improve service delivery and the relational aspects of provider contracts.
In this study, we focussed on the three purchasers in the public sector in Kenya: the National Hospital Insurance Fund (NHIF), the Ministry of Health (purchase for vertical programmes), and the county departments of health (CDOH). The three were selected because they are the purchasers with the greatest influence on Kenya's health system and whose reforms can easily be leveraged towards achieving UHC.

Data analysis
We adopted the framework analysis approach 23 to analyse the gathered data from KIIs and document reviews. Detailed notes indicating the feedback and comments from the workshop were used to further identify the areas where progress was made and where gaps remain -notes can be accessed here 20 . Two researchers (JK and EK) independently coded the data based on the conceptual framework 24,25 . Discussions with EB and KM helped build where variations were observed. The coded data were then interpreted and mapped onto the progress matrix to provide an indication of the stage of progress of the specific domain. The coding of the data was conducted in Word for Microsoft 365 and the mapping onto the progress matrix in Excel for Microsoft 365.

Ethics statement
We received ethical approval from the Kenya Medical Research Institute (KEMRI) scientific and ethics review unit (SERU) under SERU number SSC 2795. Additionally, the study was permitted by the National Commission for Science, Technology and Innovation (NACOSTI) as well as the Council of Governors. We also engaged relevant organisational heads to discuss the study prior to seeking consent from and interviewing their staff. We obtained written informed consents from all participants who agreed to participate in the study.

Results
Progress and remaining challenges in strategic health purchasing across key purchasing functions in Kenya 1. Benefits package A continuous pursuit for an explicitly designed benefits package that reflects the health and needs of the population is crucial for making progress in strategic health purchasing in any health system. To this end, Kenya has made progress on strengthening the process of defining a health benefits package that reflects the health needs of the population. Specifically, the MOH's established a health benefits package advisory panel (HBAP) in recognition of the need for an evidence-informed, inclusive and transparent process using explicit criteria, to propose an essential benefit package for UHC for Kenya. The HBAP proposed an essential benefits package that has since informed pilots of health system reforms and may be used for planned expansions of coverage 26 . The National Medicines and Therapeutics Committee (NMTC) has updated the Kenya Essential Medicines And Supplies List (2019) to expand the available medicines list by 419 to 1,072 27 . The KEML is intended to complement the essential benefits package. The KEML review was based on nine key selection criteria, including relevance, safety, comparative efficacy, quality, performance and comparative cost-benefit.
However, two key gaps still exist in the benefits package development process. First, the benefits package development process is yet to be institutionalized. The HBAP's term expired in June 2020 while that of the NMTC is poorly defined. Second, the benefits package development process remains fragmented with all three purchasers offering multiple benefit packages that are not harmonized. For instance, the NHIF still uses multiple fragmented benefit packages across the over 70 different schemes it runs. These benefit packages have also not been developed using explicit criteria, and their development has not been transparent and inclusive.

Whom to buy from
Purchasers are required to actively identify and contract with a set of providers based on explicit criteria that meet strategic purchasing goals such as the comprehensiveness and quality of services offered, and their geographical location.
Several advancements have been made. First, the NHIF engages in selective contracting (primarily among private providers). The utilisation of selective contracting by NHIF and CDOH indicates some progress in strategic purchasing as it promotes health system goals including equity in geographical access to providers and services, quality of services delivered, and efficiency. Second, one CDOH (Kisumu county) has started piloting the contracting of primary care networks aiming to contract a group of facilities (given a single contract) comprising of a sub-county hospital, all public health centres and dispensaries as a bid to enhance efficiency in contracting and service delivery through adherence to, for instance, appropriate referral channels.
Several challenges that hinder provider identification and contracting persist. First, despite the NHIF engaging in selective contracting mainly with private providers, the MOH and CDOH performs little if any selective contracting. Second, the NHIF's processes are hindered by the limited choice in areas where certain providers, usually public, are the only option. Third, the NHIF, as a state-owned agency, cannot easily reject a public health facility. The identification and selection function is poorly defined in the CDOH and MOH, in part because they own and operate many of the facilities from which they purchase health services. This, therefore, may hinder action against poor performing facilities and does not guarantee a citizen's access to service entitlements. Fourth, all three purchasers demonstrate gaps in monitoring and enforcing contractual terms. For instance, although NHIF is required to conduct periodic monitoring of facilities to ensure adherence to contractual agreement, there is a lack of adequate capacity to implement this role. The MOH and CDOH's monitoring activities are also limited to sporadic inspections for compliance with licensing requirements and routine work supervision. Progress to institutionalize accreditation, quality assurance and inspection as part of monitoring has been slow.

Provider payment systems
Provider payment mechanisms are crucial for aligning healthcare provider behaviour with health system goals. Adequately designed provider payment mechanisms can incentivize providers to offer high-quality services, efficiently and equitably. In moving towards these, all three key purchasers have primarily incorporated a mix of provider payment mechanisms (Table 1). For instance, the NHIF pays providers on a capitation basis for outpatient services but also uses fee for service for diagnostic services. The MOH uses global budgets for the vertical programmes and the national referral hospitals that it manages whereas CDOH uses a line-item budget for commodities and salaries, as well as a fee for service for counties with their universal health coverage plans e.g. Makueni.
However, several gaps remain. First, although a range of provider payment systems is in use by the three purchasers and especially by the NHIF, existing evidence shows that they are either poorly designed, poorly implemented, do not protect against fraud/corruption and/or are not adequately linked to health system goals. For instance, the NHIF's implementation of capitation for outpatient services has experienced challenges including misunderstanding by providers 28 , lack of rate negotiation and revision 28 , and inadequate mechanisms to guard against adverse consequences such as fraud by providers 28 . The MOH and CDOH mainly utilize line item budgets even where budgets are programme based, budget format would not permit the use of global budgets.
Second, we found that purchasers (particularly NHIF) have multiple payment mechanisms that are incoherent across different schemes. For instance, for normal delivery for members in public facilities, the NHIF pays KES 10,000 (USD 100) under the general scheme but pays KES 5,000 (USD 50) under the free maternity scheme. This fragmentation does not promote strategic purchasing as it sends different signals to providers that may hinder equity, efficiency and quality of services. Additionally, payment rates are often inadequate and not informed by evidence e.g. costing. Furthermore, across the three purchasers, existing evidence shows that payments are often unpredictable both in amounts and timeliness 28 which influences provider behaviour and impacts patients access to services.

Governance arrangements
Governments are required to establish clear frameworks and accountability mechanisms that guide purchasers' actions 8 . The three purchasers have legal frameworks that guide their operation. For instance, the NHIF is regulated by the National Hospital Insurance Fund Act of 1998 29 , and the MOH and CDOH by the Health Act of 2017and the Public Finance Management Act of 2012. The CDOH may also be regulated by various subnational level acts such as the County Finance Act.
There has been some progress in addressing gaps in the legal and regulatory environment in a bid to improve strategic purchasing practices. For instance, the HEFREP review included a review of the governance structure of NHIF that will explicitly restructure around members and "employer enrolment, beneficiary relations, provider empanelment, contracting, claims management" and adequate oversight from the MOH 5 . Additionally, proposed amendments to the NHIF Act currently being considered by parliament require regular reporting on health service purchasing performance by the NHIF to the MOH and the public.
Besides, some counties have passed county-level public financial management legislation to enhance the financial autonomy of public health facilities which improves the functioning of provider payment methods. The MOH and county governments have prioritised greater intergovernmental collaborations to achieve coherence in the legal and policy frameworks. For instance, the MOH is supporting counties to generate local solutions (such as local legislation) to PFM bottlenecks.
Despite these advancements, three key challenges remain. First, we found the government, through the MOH, lacking effective oversight strategies specifically on the NHIF even though the NHIF's CEO and board of directors are required to report to the cabinet secretary of health. Further, the MOH has insufficient oversight of the semi-autonomous or vertical programmes that purchase healthcare services on its behalf. Second, there remain inadequate linkages between MOH policies and the NHIF that can facilitate adequate stewardship/oversight. For instance, MOH policies do not yet identify the NHIF as a strategic purchaser of healthcare services in Kenya and do not clearly define the role of NHIF. Third, the accountability of the NHIF to the government appears to be biased towards financial accountability and neglecting operational accountability. Therefore, while the government has clear mechanisms to hold the NHIF accountable for financial issues, they do not have clear mechanisms to hold the NHIF accountable to strategic purchasing objectives.

Information systems
Information systems are a key building block in all efficient healthcare systems. Purchasers are required to develop, manage and use information systems to purchase strategically. Additionally, providers are required to communicate with stakeholders to share information, raise awareness about, inter alia, the benefits package. To this end, there is a continued emphasis on the utilisation of health information systems across purchasers in Kenya. For instance, the Kenya Health Information System (KHIS) is now widely used by the public sector with a continued improvement in data quality. Additionally, the recent HEFREP review recommended revamping the NHIF's information systems such as the claims management system into a fully automated and paperless claims process with well-developed information technology (IT) infrastructure to enable offline interaction, smooth user experience, verification, pre-authorization and vetting systems.
However, Kenya still lacks a unified national database that can permit clinical decision-making, facility performance monitoring and an examination of hospital performance at a system level. Besides, all three key purchasers have weak or unclearly defined communication strategies useful for patients to understand their service entitlements and feedback mechanisms for users to share experiences and raise complaints. For instance, while NHIF advertising on Supa Cover (one of the NHIF schemes) has increased, there remain weaknesses in communication to the user (e.g. about entitlements, responsibilities, and key processes) with the modes used (such as television and social media) favouring the rich and people in urban settings. Additionally, the MOH and CDOH have fragmented communication strategies with key stakeholders where for example the key health programmes such as community health services and key policy documents such as the Kenya health sector strategic plan have separate and incoherent communication strategies.

Discussion
This study aimed to assess the progress and gaps in strategic health purchasing using SPARC's progress matrix framework across three main purchasers in Kenya: the NHIF, the MOH and CDOH. We utilise a progress matrix approach 20 to assess the progress and gaps in strategic health purchasing in Kenya which is similar to the WHO's health financing progress matrix 22 . The current assessment demonstrates that progress is being made on moving towards more strategic health purchasing practices (what to buy, from whom, and how) that should lead to the attainment of health system goals.
On what to buy (benefit package), the assessment highlights the initiative by the government to introduce and institutionalize an evidence-based benefit package development process. A strategic purchaser is required to adequately identify and define the set of services or interventions to purchase for the population 8 .
Having an evidence-based health benefits package is the first step into making the NHIF a strategic purchaser, however, this needs to be implemented to remove the fragmentation in benefit package across the 74 NHIF schemes 5 .
On whom to buy from, as purchasers are required to adequately identify a set of providers to purchase services from, the NHIF's engagement with providers in selective contracting is a positive step into promoting provider competition and the delivery of good quality and timely services at the least cost 30 . Evidence from other settings has shown that selective contracting can be leveraged as a mechanism to control healthcare costs 31,32 . Additionally, although only piloted in one county (Kisumu), the assessment highlighted a change in contracting from contracting individual providers to contracting a network of providers. This is in line with the HEFREP recommendation on NHIF contracting reforms to introduce an integrated care contracting 5 . Integrated care contracting enhances the integration of healthcare service provision whilst reducing inefficiencies in contracting resulting from contracting multiple providers individually.
On how to buy, our findings show that purchasers utilise multiple provider payment mechanisms, however, these are not aligned to incentivize providers to enhance efficiency and/or quality of services. Existing evidence from a purchasing study in Kenya revealed that the differences (such as predictability and sufficiency) in the multiple funding flows to public health facilities lead to undesired provider behaviour such as inequitable prioritisation of resources and services in favour of patients from schemes with higher payment rates 33 . Similar findings have been reported from studies in Nigeria 34 . It is essential to consider the economic signals that different provider payment mechanisms have on provider behaviour to both develop and implement coherent provider payment mechanisms that elicit the required incentives for meeting the health system goals [35][36][37] .
Other than the above three strategic purchasing decision areas, the role of governance and information systems in aligning systems in moving towards UHC cannot be overemphasised. On governance arrangements, while the primary finding is that robust legal and regulatory frameworks that provide direction for purchasing exist on paper, there is poor implementation and oversight or stewardship from the government. Well developed and implemented regulatory frameworks are levers for governments and purchasers to affect the equity, quality, effectiveness, quantity and safety of healthcare services across the health system thus improving the performance of health systems 38,39 .
Lastly, on health information systems, the assessment showed an increased shift to the utilisation of information systems such as the Kenya Health Information System. Information systems are crucial to providing the evidence required to monitor the performance of both purchasers and providers and as a lever for an active engagement of the purchaser with the providers, citizens and the government 8 . Reid et al. refers to information and communication systems as the "backbone of health care delivery" for all functioning health systems 40 and thus the need for investment into the infrastructure and capacity of information systems.
Our findings should be interpreted considering the following limitations. The document review was not as systematic as would be expected in systematic reviews rather was used to supplement web searched from the key purchasers' websites. Consequently, we don't present a PRISMA flow diagram in this manuscript 41 . Additionally, whereas our findings focus on the three key purchasers in Kenya, other purchasers such as private health insurance and community-based health insurance exist but were not assessed. However, their contribution to overall coverage is low 42 .
However, based on these findings, we make several recommendations. First, there is a need to implement the benefits package developed by the health benefits advisory panel, and to institutionalize health technology assessment as an evidencebased approach to healthcare priority setting. There is also a need to accompany these reforms by the harmonization of benefit packages while removing the fragmentation in the current duplicated and incoherent benefit packages across the three purchasers. Additionally, adopted benefit packages should reflect current and projected resources available. Second, there is a need for the MOH to embrace the stewardship role and provide oversight to the NHIF and the vertical programmes. Good enforcement of the established laws and policies is essential in improving health system performance 39 . A specific improvement would be a clear articulation of strategic purchasing objectives and accountability requirements between the MOH and the NHIF. Third, all purchasers should have explicit requirements for engaging with providers to ensure that health facilities improve on the quality, equity, and efficiency of the services they deliver while enhancing provider accountability and that their selection and contracting is pro-equity. Fourth, purchasers should tailor their provider payment mechanisms to ensure that they are coherent, elicit the required provider behaviour and promote health system goals. For instance, provider payment rates should be based on costing evidence. Fifth, purchasers should enhance the adoption of information systems by making investments and building capacity for information systems such as the Kenya Health information system (KHIS). Sixth, purchasers should explicitly communicate the service entitlements to citizens and action on their complaints and feedback. Poor communication hampers public accountability and the responsiveness of the purchasers to citizens needs and requirements. Effective communication of purchasers builds awareness of service entitlements and trust among citizens which promotes strategic purchasing 39 .

Conclusion
In conclusion, although purchasers in Kenya have made progress in, inter alia, establishing structures such as legal frameworks, contractual agreements with providers, provider payment mechanisms and the use of information systems that are in line with the ideals of strategic purchasing actions, the design and implementation of these structures remain poor and fragmented. Given the role that the three purchasers in Kenya play and the government prioritisation of strategic purchasing as a lever for moving towards UHC, there is a need for reforms that align the purchasers' actions to the ideal strategic purchasing actions.

Data availability
Underlying data Documents reviewed in this analysis are publicly available. Transcripts from the interviews are not publicly available due to them containing personal identifying information that could compromise research participant privacy. The transcripts can be obtained upon reasonable request from the authors (kjacob@kemri-wellcome.org).

Introduction
In describing public sector purchasing in Kenya, you mention the health system context of devolution in Kenya and how it is likely to affects success/failure of strategic purchasing initiatives for the purchasing arrangements presented. It would also be important as part of the background to mention the percentage composition of the different resource pools used for purchasing. This is an important component of the background because while it is important to be strategic with 20% of our resources, it would be even more important if we achieve this with 80% of the resources. This should emphasize the need for ensuring strategic purchasing with government resources. No need to mention the provider payment mechanisms used by the purchasers in this section as it should be part of your results.

Methods
I suggest that the study setting focuses on describing how the health system is organized. The health financing context specific to strategic health care purchasing should be moved to the background/introduction as recommended above.