Facility Management Information System (FMIS)

Background:

The recent history of Sierra Leone includes significant shocks from man-made and natural causes, including years of civil war, natural disasters as well as the 2014-2016 West African Ebola outbreak and the ongoing COVID-19 pandemic, all of which have hampered economic recovery and development. As a consequence, the country continues to report the world‘s highest maternal mortality rate, while one out of six children is bound to die before the age of five and close to four fifth of the country‘s population live in poverty. Communicable diseases remain the leading cause of death, and social structures bear witness to pervasive gender inequality. 

As a consequence of the Ebola crisis, there has been newly found attention on, and funding of, health care, which is reflected in national strategies and policies, culminating in the publication of the Universal Health Coverage Roadmap for Sierra Leone 2021-2030 in late 2020.

Even so, key health indicators such as maternal and child mortality have not improved significantly, due to incomplete implementation of ambitious strategies, policies and guidelines, and patchy alignment of key elements of the health care system to the needs of the average people.

The Problem:

While there is an extensive and, for most part, well-managed Health Information Management System (HMIS) in place, it demands significant resources that are prone to divert the health care provider's attention from patient care. Especially in the domains of child health and maternity, there are multiple forms to be completed, with duplicate data entry requirements, all of which need to be summarised every month for monthly summaries due to be submitted to the district and forwarded to the central level. Even so, the amount of truly actionable data remains small, while the upkeep of resources to keep the HMIS running has become increasingly expensive.

What is more, there remains a dire lack of affordable, accessible and up-to-date healthcare performance data that can efficiently and effectively drive informed decision-making.

Healthcare staff spend the majority of their care time completing registration forms, and the necessarily encountered interruptions and diversions so common to the clinical domain do not help with maintaining adequate reporting quality.

While this situation affects most patients presenting at a governmental health facility in sub-Saharan Africa, where a HMIS has been implemented as manual reporting tool, those patients with multiple reporting trajectories are most severely affected. They include children under five years-old as well as maternity patients, whose data are routinely captured by a multitude of registers, each of them demanding duplicate entry of core patient data. For example, there are treatment registers, Expanded Program on Immunization (EPI) registers, nutrition registers, delivery registers, mother and neonate registers, family planning registers, under-two registers, under-five registers to name but a few. None of these registers provide a longitudinal approach to patient care, which is why the majority of mothers carry a notebook for each child that the health care provider is asked to complete, constituting yet another piece of paperwork to deal with.

At the end of each month, health care providers are to manually aggregate patient register data and report them in accordance with national standards. The assembly of these monthly summaries has necessitated considerable and costly training efforts across Sierra Leone and, by extension, sub-Saharan Africa, where a similar system is in place. This system is hampered by high printing costs for HMIS forms, form stock-outs as well as busy clinical schedules that interrupt monthly reporting efforts and therefore endanger data consistency.

Over time and after multiple up-skilling efforts, it has become clear that it would be neither cost-effective nor sustainable to throw even more capacity building at a system that is short on usability and lacks a fail-proof way of data handling. Instead, DPPI as lead agency in Sierra Leone called for an appropriate digital solution that address the above challenges, while maximising robustness and sustainability in environments that more common than not lack reliable power, Internet access and computer savvy staff.

Summary:

While there is a huge need for improved evidence-based health care provision, the current manual efforts to provide this very evidence have diverted valuable resources from the core of the healthcare encounter and affected overall healthcare quality negatively. There is an urgent need for a harmonised, patient-centered approach that marries evidence-based patient management with the demands of automated, efficient and relevant aggregate reporting.

The obvious solution to the problems outlined above is to (a) reduce the paperwork required during patient encounters by capturing both patient core data, such as name, sex and date of birth, as well as encounter details in a way that remains fully aligned to the standard clinical circuit, and to (b) automate any summary reporting by aggregating existing data after automatic validation without any human interaction.

The FMIS does exactly that, while at the same time extending the scope of a mere electronic medical record (EMR) system:

For example, by managing all resources, be they human, finance, infrastructure or time, the FMIS translates prescribed and dispensed drugs into automatic stock updates, without the need of human interference. Likewise, attendance monitoring is closely linked to activity monitoring of all logged-in users in order to comply with standard safety and security requirements of mission-critical software. In a similar fashion, patient categories, set and updated at patient registration time, set applicable price lists, in full accordance with national guidelines, and are ready to inform insurance providers automatically, once given appropriate permissions.

The FMIS is part of a bigger infrastructure that honours the following design principles:

The data flow is summarised in the following illustration:

1. Clinical data are entered via mobile devices, preferably 7” or 8” tablets adequately protected by screen protector and and children’s bumper case.

2. Data flow by local WLAN to the FMIS miniserver, where automatic data validation, cleansing, aggregation and local reporting are effected.

3. The FMIS miniserver synchronises summary data automatically to the in-charge’s tablet, where they will be stored in an encrypted manner.

4. The in-charge travels to the district capital at least once a month.

5. At the district health management team (DHMT)’ office, the in-charge’s tablet will have access to the Internet.

6. The encrypted data summaries contained on the in-charge’s tablet are automatically synchronised to the MoHS’s dashboard server.

7. The MoHS dashboard server is capable of further aggregating, analysing and preparing a wide variety of standards-compliant and custom data formats, including automated data feeds understood by the de facto standard medical data warehousing software, the DHIS2.

8. While at the DHMT, and because the synchronisation mechanism works in a bidirectional way, the in-charge’s tablet can also receive data and media from the MoHS’s dashboard server that are destined for the PHU, such as updated clinical guidelines, performance feedback or order fulfillment data.

9. As before, these data are safely encrypted on the tablet to avoid non-privileged access by loss, theft or accident, and are pushed to the FMIS miniserver, as soon as the in-charge’s tablet has access to the PHU’s WLAN.

10. Remote maintenance of the PHU-based FMIS miniserver is possible via two avenues, either by available Internet on-site, as seen in bigger PHUs, or by bringing the (extremely portable) FMIS miniserver to the nearest point with adequate Internet access. Connection is only possible via a virtual private network (VPN).

Summary:

The FMIS offers a full replacement to time-consuming, manual record keeping and data aggregation for summary reporting that at it current HMIS state is riddled with the need to enter duplicate data sets, which in turn fail to go beyond mere snapshot capacity due to their failure to link successive patient encounter. In contract, the FMIS follows the best-practice workflow of clinical history taking, clinical examination, recording of investigative evidence and diagnosis, without duplicate data entry, while fully preserving HMIS compliance. Previous, fully manual data aggregation procedures that lacked any means of built-in validation have been replaced with automated algorithms that use the above EMR data, digitally validated at the time of data entry, as source. The same applies to any applicable data set handled by the FMIS, such as supply chain management (SCM) data, human resources (HR) data and accounting.

The current phase of FMIS pilot targets a total of six community health centres (CHCs), which present the highest level of PHUs in the country. The CHCs are located in Kambia, a district at the border with Guinea, and Kailahun in the East, which calls Guinea and Liberia its neighbours. The selection of CHCs include basic emergency obstetric and neonatal care (BEmONC) sites that are facilities upgraded to cope adequately with common emergencies encountered in maternity and neonatal care, such as (pre-)eclampsia, postpartum haemorrhage, postpartum infection and neonatal resuscitation.

As elaborated earlier, the highest reporting burden and therefore most negative impact on actual time available for patient care is seen in maternity (including neonatal care) and in paediatric care of under-five year-olds. While women and children are most affected, presumably through well-meant efforts by both governmental agencies and development partners in an effort to diligently monitor the quality of care, the overburdening reporting chores do affect all areas of healthcare provision.

It is expected that the higher the current paper-based reporting requirements are, the more negatively the latter impact on the time available for actual health care provision. Reversely, as the digital FMIS obsoletes duplicate reporting and greatly aligns the flow of required information to the clinical circuit, the biggest impact will be seen with maternity and under-five year-old patients. However, it is expected that the FMIS will also strengthen currently under-served areas in general medicine, such as non-communicable diseases, such as the increasingly prevalent entities of hypertension, type II diabetes and hyperlipidaemia, as well as mental health.

Apart from streamlined medical record keeping during a patient encounter, there is no more need to manually assemble monthly summary reports, because they are done automatically in real-time based on existing patient records. This not only makes valuable healthcare provider time available for immediate patient care, but also avoids the commonly encountered data aggregation and transcription errors seen with manual procedures. Improved data quality is expected to greatly assist evidence-based planning at the central and district levels, which are in turn expected to feed back to improved, data-based resource allocation at the healthcare provider level.

The Directorate of Policy, Planning and Information (DPPI) of the Sierra Leonean Ministry of Health and Sanitation (MoHS) has been tasked to develop and oversee the appropriate implementation of health-related policies, including improved access to and quality of relevant health data, which are at the heart of evidence-based decision-making for needs-based planning.

The reach of MoHS, including DPPI, extends from the central (or national) level all the way down to the PHUs, with DHMTs acting as crucial coordination and management entities at district level. This implementation and reporting hierarchy greatly leverages the efficiency and efficacy of the national level and allows for rapid deployment and coordination of corrective actions or emergency preparedness (as seen, for instance, in outbreak response management).

What is more, development partners cooperate actively not only at the national, but also district levels with their respective MoHS counterparts to forge harmonised policies, implementation strategies and to avoid duplication as much as possible at all levels.

Being at the core of the policy, planning and information system of MoHS, DPPI is therefore ideally placed to oversee and coordinate the FMIS pilot, growth and scale-up, both by its very mission as well as by its authority and competence at all relevant decision-making levels.

DPPI has guided and overseen the development of the FMIS. With funding by the German Development Cooperation (GIZ), the FMIS team has completed a first pilot phase in two CHCs in Kychom in Kambia and Daru in Kailahun, the main purpose of which was to assess the human factor and confirm that the FMIS is indeed as user-friendly as hoped for during actual field deployments. Not only has this field deployment confirmed usability with great success, but also seen bugs removed and additional features added. There are four more sites due to be included in the pilot stage, which are, like the above first phase, covered by German Development Collaboration (GIZ) funding till the end of 2022. For any further growth and scale-up, DPPI has been obliged to search for alternative funding. In this context, it is hoped for that participation in the MIT Challenge contributes to visibility and may attract valuable feedback and comments from like-minded innovative parties, keeping in mind that the FMIS is a somewhat unusual product of conventional development, which usually does not favour incubators, developers and innovators. With added visibility and support, it is expected that DPPI can more effectively approach other donors to make the FMIS scale across the nation, both at PHU and hospital levels.

At the present stage, patient registers are paper-based and aggregated manually for end-of-month reporting. Although there have been a few attempts to implement EMR systems in a variety of hospitals in the country, the focus with the latter has been mainly on the EMR side, and less so on streamlined HMIS reporting. However, the majority of health data stems from PHUs, i.e. levels of care below government hospitals, including, in descending priority, community health centres (CHCs), community health posts (CHPs) and maternal and child health posts (MCHPs).

The proposed (and successfully) piloted system has been designed from the bottom up to be fully functional in remote, offline environments, with hardware optimised for cost and energy efficiency, compatible with very modest solar design requirements. The FMIS has been piloted at the CHC level because of donor conditions. However, the very same system can be readily deployed to CHPs and MCHPs, with an easily accomplished adaptation to the reduced range of drugs and tests available at those lower levels of care. By activating additional modules and configuring them to align to national HMIS requirements, the FMIS can readily scale to government hospitals (GHs, at district level) and national referral hospitals (NRHs). In the latter case, the hardware would have to be upgraded accordingly to be able to cope with the higher expected data volume. Still, the software architecture remains extremely efficient and therefore accounts for the power challenges seen in many GHs that rely on generator power either entirely, or as backup for highly unreliable mains power: With ever-increasing energy prices, generators have long become an unsustainable power source, whilst the energy-efficient FMIS can entirely run on a modest photovoltaic system, even when scaled-up to hospital level.

It is noteworthy that the entire FMIS consists of standard, well-supported FLOSS elements. The core FMIS software has been customised to be fully aligned to the national HMIS without any programming efforts. The FMIS presents a modular, yet full-blown facility management information system and conveys added value by not only handling EMRs, but also all other aspects of facility management, currently activated in a step-wise fashion to facilitate adaptation, for HR, automatic inventories for SCM and patient profiling for accounting. Thanks to its underlying ERP logic, it is able to easily yield business intelligence, and allow database queries programmatically by third-party tools. This, and its capability to scale from small, remote health units to hospitals in a secure offline fashion, while maintaining utmost resource efficiency, make it stand out among classical EMR.

Summary: Key advantages of the FMIS include:

• FLOSS throughout, including operating system and all auxiliary tools and utilities

• Extensive, long-standing community support (15 years)

• Full ERP logic included, covering all management domains of health facilities

• Extensive customisation possible without programming

• Highly resource-efficient software and hardware infrastructure

• Fully national HMIS-aligned

• Automatic offline-enabled dashboarding facility for performance data, feedback, data mining, event flagging

• High usability, as confirmed in field pilot

The current intervention is part of a GIZ-funded project to improve the decentralised management of health services, by building adequate capacity to achieve satisfactory data quality for ready use in evidence-based sector planning built at district level in the two donor target districts Kailahun and Kambia districts in Sierra Leone. The underlying assumption is that while evidence-based planning is likely to yield better (health) outcomes, it also calls for improved data quality to inform appropriate action.

Planning for health services at district level involves the district council as local government, the district health management team (DHMT), all health facilities and development partners.

Consequently, to effect evidence-based planning, the data to inform the latter have to be presented in a highly usable way that enables, other than DHMT staff, less technical decision-makers in the local government, as well as frontline healthcare workers to understand and apply key information to design corrective action.

Note: The project team has also digitised the previously manual Integrated Supportive Supervision tool for PHUs, and created a self-hosted dashboard feed for data analysis, aggregation, presentation and mining.

The implementation therefore applies the following logic:

Digital FMIS

↓

Improved usability (proof of work accomplished by pilot, growth phase planned for 2023 and 2024)

↓

More time for patient care (proof of work accomplished by pilot, growth phase planned for 2023 and 2024)

↓

Improved data quality at source by digital validation (proof of work accomplished by pilot, growth phase planned for 2023 and 2024)

↓

Improved (automated) summary data (proof of work accomplished by pilot, growth phase planned for 2023 and 2024)

↓

If presented appropriately by “need to know”: Improved visibility of actionable data (proof of work accomplished by pilot and established link to digitised integrated supportive supervision, growth phase planned for 2023 and 2024)

↓

Improved visibility of performance, easy pinpointing of need for follow-up (proof of work accomplished by pilot and established link to digitised integrated supportive supervision, growth phase planned for 2023 and 2024)

↓

Emergence of a comprehensive quality management tool for all domains of healthcare provision (long-term impact goal: pending further scale up to more PHUs and possibly hospitals)

↓

Improved health outcomes by evidence-based resource allocation

All work in Sierra Leone is guided by the United Nations Sustainable Development Cooperation Framework (UNSDCF) Sierra Leone 2020-2023. It features four key priority areas of intervention: (1) Sustainable Agriculture, Food and Nutrition Security; (2) Transformational Governance; (3) Access to Basic Services; (4) Protection and empowerment of the most vulnerable.

The FMIS does not introduce new indicators, but re-uses existing ones and attempts to render them more accessible, reliable and usable.

With regards to data quality, DPPI has been using a well-established indicator framework for data quality audits (as shown below), which is as applicable to the novel digital FMIS monthly summary reports, as it has been to manual data quality audits and checks in the past. The persistence of indicators guarantees comparability and full alignment of FMIS summary reports to existing HMIS requirements.

Likewise, healthcare performance indicators are firmly embedded in the national HMIS indicator framework using DHIS2 as national data warehouse. As before, the digital FMIS and aggregate dashboard facilities are not to re-invent the wheel and use existing, national HMIS indicators throughout.

DPPI, together with the in-country consulting team funded by GIZ, concluded that the highly manual and resource-intensive nature of data handling and aggregation at facility level could not be improved any further by even more training, but rather require a radical re-design of how (health) data are collected and managed in the first place during the patient encounter. By questioning the existing approach, it was agreed that a more patient-centred, streamlined strategy, if implemented appropriately, could not only provide better data quality at source, but also free up valuable clinical encounter time for patient care, and provide an automated infrastructure for summary reporting, all of which had to be done manually previously.

There are over 1,130 PHUs in Sierra Leone, with 209 CHCs, 345 CHPs and 577 MCHPs as of early 2022. The present pilot has successfully included an initial two CHCs, with four more following suit by autumn 2022. While there is already strong evidence not the hugely improved usability and data quality of the FMIS coherence, the ensuing growth and scale-up phases are expected to include significant amounts of time and resources to digitise the primary health care landscape in Sierra Leone in the affordable and sustainable way paved by the digital FMIS. Likewise, each and every additional facility that will join the digital FMIS effort will not only yield improved aggregate data for solid evidence-based decision-making, but is also hoped to improve healthcare quality by providing more time and resources for patient care by obsoleting time-consuming manual record-keeping.

The current donor work plan sets the logical framework for the present (pilot) phase, and is extended as indicated:

Development Objective: The healthcare system in selected districts in Sierra Leone is more suited to serve the health needs of the population

Output: The decentralized management of health services is improved

Expected Result: Adequate capacity to achieve satisfactory data quality for ready use in evidence-based sector planning built at DPPI, district councils, DHMTs and selected health facilities in Kambia and Kailahun (Sierra Leone)

Activity 1: Install, configure and populate draft digital FMIS

Activity 2: Procure, install required or repair/replace defective solar equipment

Activity 3: Provide e-training in the FMIS to DPPI system administrators, application experts

Activity 4: Groom super-end-users as digital FMIS trainers

Activity 5: Train PHU as digital FMIS end-users

Apply the above iteratively for a total of six pilot facilities in Kailahun and Kambia, the subsequent growth phase, and for nationwide scale-up, with the following envisaged timeline:

• Completion of six pilot sites in Kailahun and Kambia (Sierra Leone), including grooming of FMIS instructors and further fine-tuning of the FMIS – by end 2022

• Completion of an early grow phase of at least 12 additional sites in Kailahun and Kambia (Sierra Leone), including further grooming of FMIS instructor team and fine-tuning of the FMIS – by end 2023

• Completion of a first-phase scale-up phase totalling 30% of all CHCs, 10% of CHPs and 10% of MCHPS in ten key districts using the experience and instructor teams groomed in 2022 and 2023, with full alignment to the lesser CHP and MCHP requirements, when compared to CHCs – by end 2025

Note: All FMIS installations depends on reliable power on-site, either exclusively by a 24/7 photovoltaic system, or by a backup solar system that complements (unreliable) mains power, which is not available in Kailahun nor Kambia.

1. Data capture by robust Android™ 7 or 8 inches tablets (1), locked down to avoid user interference and the incidental installation of malware, connect via WLAN to on-site FMIS miniserver (2)

2. FMIS miniserver scoped to serve PHUs, including busy CHCs with BEmONC services, with: 1 x Raspberry Pi 4 8 GB single board computer (SBC) with 128 GB microSD card mainly used for read-only access, and an external 250 GB SSD drive for database use, connected to a 60,000 mAh power bank capable of pass-through charging to serve as UPS in the unlikely event of solar backed-up power failure USB WLAN adapter for small PHUs, or external access point (EAP) with power over Ethernet (PoE) for installations, where the USB WLAN is insufficient The FMIS miniserver runs Syncthing and a customised version of ERPNext on Linux

3. In-charges tablet, constituting of the standard staff Android™ device as seen in (3), with active Syncthing FLOSS to automatically synchronise the FMIS miniserver’s summary data to the tablet in an encrypted fashion

4. Data synchronisation via Syncthing to the DPPI dashboard server occurs, as soon as the in-charge’s tablet reaches (4) the DHMT Internet (5)

5. Automatic synchronisation by the FLOSS Syncthing running on the in-charge’s tablet and the DPPI dashboard server (6)

6. DPPI dashboard server running FLOSS InfluxDB, Telegraf and Grafana stack on Linux

7. Optional remote maintenance via Internet, whenever accessible, and VPN (10)

The FMIS does not add any additional data entry avenue, but bundles existing (manual) data entry needs into one simplified channel. Therefore, the existing responsibilities remain the same, and PHU health care staff remain fully responsible for all related data entry, even though the load of work is expected to be greatly reduced by virtue of the digital FMIS properties as described earlier. Their capacity as health care professionals is verified and their on-system activities are audited in real-time by the FMIS.

The MoHS, including DPPI, is committed to diversity, equity and inclusion, and very actively promotes gender equality. The consulting team that has developed the technical solution consist of two national experts and a foreign consultant, all being supported by the donor via a subcontracted consulting firm, while working hand-in-glove with MoHS. To date, there have been challenges in ensuring gender balance within the project team.

Open access to the innovative technology deployed remains a cornerstone of sustainability and free accessibility, and the project has created a DPPI-owned e-learning platform to train system administrators and application experts from the bottom up to install, maintain and support the entire system, while creating additional online training material as the need arises.

Likewise, end-user training, which addresses each and every end user individually, is an all-inclusive exercise that involves the entire PHU, with “star learners” being promoted to facilitators and multiplicators.

Finally, the evident lack of solar engineering and maintenance expertise at the district level, which is essential to keep the digital FMIS running, has been addressed in both Kailahun and Kambia. Extra effort has been made to involve a diverse group of community members from remote chiefdoms to conduct hands-on, highly practical training courses intended to grow a diverse community of skills and knowledge including highly motivated trainees ranging from car mechanics and electricians to solar mechanics and community health officers.

The Directorate of Policy, Planning and Information, under the umbrella of the Sierra Leonean Ministry of Health and Sanitation, drives and coordinates the development of comprehensive and strategic health legislations and policies, strategic plans, health information systems, programmes and budgets. It monitors and evaluates their implementation in collaboration with other programmes, ministries, departments and agencies (MDAs), as well as development partners and donors. As part of its responsibility for health information, it leads information, communication and technology (ICT).

The above terms provide a concise view of DPPI’s general business concept: Whilst taking responsibility for policy, planning and information in the health domain, DPPI coordinates and balances national planning and finances with available partner and donor resources to ascertain the best value for resources spent. The governmental health sector is not designed to be for-profit. While self-sustenance remains an ultimate goal, dependence on donor funds remains at present significant, which further underlines the need to optimise both performance as well as operational efficiency.

More specifically, the pilot phase of the FMIS implementation in three CHCs each in Kailahun and Kambia districts in the East and Northwest of the country, respectively, is covered by current donor funding. The ensuing growth phase is presently unsecured, yet major donors, such as GFATM or WB, have confirmed considerable interest in scaling up fully working solutions.

DPPI as part of MoHS is tasked to optimise planning and project execution processes in collaboration with other MDAs and partners to maximise efficacy, efficiency and sustainability, last not least through alignment with existing policies and projects. Financial requirements for health interventions are ascertained as part of the annual planning done in close collaboration with districts in a process overseen by the Sierra Leonean Ministry of Finance (MoF).

In the given case, the digital FMIS is not introducing any additional processes, and has been designed to actually increase the efficiency of laborious, manual processes, and to render them more robust, last not least in financial terms. The opportunity costs for digitisation, which involve not only ICT equipment, but also solar power infrastructure, serve an entire mix of purposes: Priority healthcare intervention areas, such as maternity, under-five child health and emergency care will greatly benefit from reliable access to power, whereas clinical data exchange and management, including continuity of care, benefit significantly from the ICT investment.

Equipment selection is key in this regard, and the entire FMIS intervention has been designed from the ground up with total cost of ownership (TCO) and durability in mind.

DPPI as well as the entirety of the Sierra Leonean MoHS have managed critical health programmes for decades, including highly efficient and effective emergency preparedness and outbreak response interventions during the West African Ebola outbreak, the introduction of a well-circumscribed, succinct HMIS, successfully managed by the de facto standard DHIS2 data warehouse, as well as the emerging Sierra Leonean Social Health Insurance (SleSHI) to serve all Sierra Leonean citizens.

The development and implementation of comprehensive health care in a country like Sierra Leone is challenging, however, DPPI’s efforts to coordinate and align donors and development partners through improved, evidence-based planning at both national and district levels have been a key success factor so far.