Vikelani - Datacentric Healthcare Management

Open-source, free-to-use practice-support tools for doctors and veterinarians that generates rich anonymised datasets on antimicrobial use and effectiveness in the community.

Dieter van der Westhuizen

While hospitals benefit from formal antimicrobial surveillance, up to 80% of antibiotics in South Africa are prescribed at the community level where such surveillance is rare (Brink, 2016). Communities are thus sorely underrepresented when it comes to antimicrobial use data, a concerning situation as antimicrobial use for human consumption increased by 35% between 2000 and 2010 with BRICS countries accounting for 76% of this change (Van Boeckel, 2014). It is additionally estimated that 60% of acute respiratory tract infections are unnecessarily treated with an antibiotic, again highlighting the importance of accurate use data in this setting.

The disparity in available data between hospital- and community-based health practice is because community health data is rarely electronic or centralised, particularly in LMICs. Currently, collection of community antibiotic use data relies on dedicated programs including surveys, exit interviews and antibiotic sales data (Do, 2021). However, such programs require staff, funding, sustained healthcare system engagement and rarely generate longitudinal data.

Our solution seeks to provide tools that healthcare workers will use for their own sake during routine clinical care to reduce workload and costs, but have the added benefit of generating community-based antibiotic use data.

The platform's target audience is aimed at:

• Human healthcare providers (all phases)
• Healthcare seekers (all phases through generation of antimicrobial resistance (AMR) data but phase 2 onwards will provide specific tools such as a patient application)
• Animal healthcare providers (phase 3 onwards)
• Health policy makers (all phases)
• Researchers (all phases)

Healthcare providers are supported through free-to-use practice management tools. We have already had a focus group with local healthcare workers to better understand their needs. From phase 2 onwards healthcare seekers are given agency through downloadable healthcare record tools that provide health information and are given a platform to report treatment effectiveness. Policy makers and researchers are supported through the integration of data collection tools in the practice management suite, offering them access to healthcare surveillance data that is otherwise largely unavailable.

Throughout development and scaling, feedback on the app's useability and utility will be sought. Feedback on the proof-of-concept application has so far been sought from a small and diverse group of primary health care providers in the Cape Town area, South Africa. 

The primary public goods we will provide are: Open-source software and data from users who choose to make their anonymised data publicly accessible

Key initiatives that will be pursued or developed beyond the primary public goods provided include:

• A free-to-use dashboard visualizing antimicrobial use-related data, offering public insights and supporting global monitoring after further project phases on aggregated data 
• White paper/publication opportunities to share research and insights in peer-reviewed journals, enhancing strategies for similar implementations, especially in under-resourced regions   
• Free educational resources and workshops to facilitate the app's adoption by medical facilities, raising awareness on antimicrobial resistance among healthcare professionals and the public
• An open API for accessing comprehensive data on antimicrobial use, enabling the development of new applications and research  
• Open-source predictive models to aid in understanding and predicting resistance trends, supporting informed healthcare and policy decisions  

This underscores our commitment to leveraging open-source software for healthcare improvement, particularly in combating antimicrobial resistance.

Our software solution will:

• Provide free practice management tools which can reduce costs for healthcare providers and facilities. Our software will also expand access to such tools to facilities or providers which may not have the resources to acquire currently availble commercial tools.
• Provide a patient app that can be used for personal health record management and to interact with healthcare facilities.
• Provide analytics data to providers and facilities to evauluate multiple components of their practice, including AMR-related data.
• Generate antimicrobial usage data (which if made publicly accessible by providers or facilities) can facilitate the identification of inappropriate usage or establish baseline data before interventions at a population level to gauge their impact. For instance, if a clinical trial aims to assess the effectiveness of prophylaxis using a specific drug within a population, the platform's data can offer insights into baseline drug usage and characteristics of the patient population most affected.
• Empower the healthcare community by allowing them to monitor trends in their served populations, potentially incentivizing compliance with health protocols.
• Enable researchers to infer AMR data from available data points, such as patient prompted treatment responses or health-seeking behaviour

Our software solution is built for scalability from the ground up. Within the Django framework, we've crafted flexible data models allowing seamless integration of new software components without disrupting the application's core functionality.

Imagine a healthcare facility wanting to incorporate additional laboratory data for analysis; our system enables the addition of new models to import this data directly into the same Postgres database. This data can then be analysed using the tools we provide as part of the solution.

The implementation strategy follows a phased approach, both in software development and geographical expansion:

• Year 1: Phase 1 and rollout in one country
• Year 2: Phase 2 and expansion across two countries
• Year 3: Phase 3 and deployment covering three countries

This methodical approach ensures controlled growth and effective adaptation to varying contexts over time.  Only three countries are targetted, considering the budget allowed, we don't foresee this as limitation, since we believe the solution has the capability to scale well beyond this, given more resources and adoption.

Moreover, our approach to web hosting and application packaging ensures smooth scalability into diverse geographical settings. Whether in rural or urban environments, our solution can adapt fairly easily, requiring minimal internet connectivity. 

Current measurement:

• Testimonials and reviews - ongoing with focus group discussions of proof-of-concept

Planned analytics:

• Google Analytics (or similar platform) for monitoring web traffic
• New user (healthcare provider) acquisition count
• New institution / facility onboarding count
• Doctor and facility retention rate (the data generation trends)
• Peer reviewed publications using solution generated data
• Policy documents that reference the solution
• Treatment Guidelines referencing the solution
• Open source fork counts
• Third-party generation of tools using the solution platform
• Use in countries not specifically included in planned expansion (measurable with above data points)
• Funding acquisition

• Funding: This is essential due to our open-source, free model. Initial scaling aims for minimal maintenance costs, with further software component development contingent on additional, fiscally sensible funding.
• Poor community engagement: User-friendly design and responsiveness to feedback ensure the app meets user needs, enhancing adoption and ease of use.
• Rapid technological advancements: Although this could benefit our solution, to prevent obsolescence, our flexible architecture adapts to new technologies, supported ideally by a development team monitoring tech trends. We do however envisage the Django web framework to be in use for the foreseeable future.
• Regulatory compliance: Implementing solutions in new regions involves due diligence to meet legal requirements, with transparent data handling - ideally monitored by a dedicated team of consulting legal representatives.
• Healthcare professional acceptance: Early change management is critical to overcoming resistance. Continuous feedback and a focus on simplicity and utility aim to showcase the solution's benefits, encouraging use. We plan to highlight the many benefits of using the solution at each opportunity in order to incentivise its use.

• Funding: We are applying to the Trinity Challenge because direct funding to address issues such as a lack of comprehensive community healthcare data is scarce. 
• Support: While awarded funding would advance our solution, the Trinity Challenge also offers access to resources beyond funding such as expertise, collaborators, implementation support and regulatory advice from affiliated institutions. 
• Visibility: The collection of like-minded teams from diverse backgrounds viewing and interacting with presented solutions can foster collaboration through increased visibility. This may result in novel solutions regardless of whether funding is awarded.
• Public Awareness: Through its platform and resources, The Trinity Challenge can expand the reach of our solution to the general public, who can drive use of our solution through personal health advocacy. 

In general, our platform needs to be used by as many human healthcare providers, researchers, clinical trial facilitators and animal healthcare providers to be successful. It is also an open platform with the option to add functionality based on local or evolving needs. The Trinity Challenge provides the opportunity to generate awareness, whether we win or not.

We have identified 3 collaborators in line with the 3 major components of our solution that we would like to collaborate with. These components are information technology, local accessibility and big data analysis.

• The first collaborator we would like is Amazon Web Services, Google Cloud Services (or a similar cloud services provider) to help us build and expand our infrastructure.
• The second is an expansive collaboration which should include local healthcare and community partners in the intended area of expansion. Examples include the Faculty of Veterinary Science University of Pretoria as a partner to aid in building an animal-focused version of the app.
• The third partner would be the University of Cape Town School of Public Health to support the development of data analytics tools to generate the most benefit from our data.