DIMTECH Project.

In Uganda, cervical cancer (CC) is the most common cancer among women. The incidence of CC among women aged 15-49 years is high, with 48.5 cases per 100,000 people [1]. Worse still, women living with HIV are faced with an increased risk of acquiring the disease. However, primary healthcare providers tend to miss symptoms and rarely detect cancer in the early stages, thus leading many to develop the advanced untreatable disease. Therefore, CC disease especially in its advanced stages exerts a lot of economic burdens, especially among low-income families.

The World Health Organization (WHO) and the Ministry of Health Uganda (MOH) have rolled out new guidelines for CC screening, including HPV self-sampling screening tests and visual inspection with acetic acid (VIA) triage tests [2]. Unfortunately, in rural HIV clinics in Uganda CC screening services are largely non-existent. In instances where it is available, several limiting factors exist including; 1) a knowledge gap on how to administer the treatment; 2) low community demand; and 3) unkept appointments. Additionally, medical records of women screened are not kept in a well-established data management system, but rather in an outdated paper-based that results in inefficiency.

Paper-based medical record systems are known to have major problems of inaccuracy, incomplete data, poor accessibility, and challenges to patient follow-up appointments, which have resulted in an increase in CC disease cases and family burden [3]. The system also serves as an inefficient mechanism of record-sharing for interdisciplinary patient assessment and management. Fundamentally, the system represents a major problem for keeping current and monitoring quality control to facilitate improvement, especially in areas with low doctor-to-patient ratios as is the case for Uganda [4]. To address those concerns, oncology information management systems have been proposed, including customized hospital patient information management systems, and attempts to customize deaf and hearing impaired services (DHIS) for cancer information [5].

However, answering information requests in many oncology units in Uganda is still a time-consuming task since the physicians have to track down patient records stored in paper format from different districts and medical facilities. This vital information includes hospital information, laboratory/pathology results, diagnostic imaging, hospital notes, clinical letters, and the patient medical record. In order for information to flow seamlessly, there is a need to implement an integrated Oncology Information Management Suite. Furthermore, Uganda has one central cancer unit (Uganda Cancer Institute), with several cancer service centres in different parts of the country. To respond to this challenge, we have implemented an integrated oncology information management system which has been successfully piloted at a regional cancer screening centre. Our intention is to extend this platform to HIV clinics in Uganda and provide real-time cancer information to MOH; and other key stakeholders.

The Ugandan health system is highly decentralized and geographically dispersed, composed of thousands of health facilities across 147 districts. A limited capability for information sharing among clinics and other health care facilities has resulted in lost patient records, inadequate responses to CC management, and a poor understanding of whether patients are making routine appointments to receive screening or medication. With support from UNICEF and WHO, Uganda introduced an SMS-based health reporting program called mTRAC. According to the Global Partnership for Sustainable Development Data, out of the 112 districts in Uganda, 70% are above the suggested reporting rates, but 17 of the more remote districts still experience difficulty. MTRAC has offered an alternative approach but has not fully replaced the existing paper documentation system and falls short of identifying gaps in and public perception of public health performance. Hence further solutions are needed to understand gaps in primary care access and gauge public response to health care performance and interventions in CC screening and management.

References

[1] Bruni, L., et al., Human Papillomavirus and Related Diseases in Uganda. Summary Report 17 June 2019. 2019, ICO/IARC Information Centre on HPV and Cancer (HPV Information Centre).

[2] L. Akhu-Zaheya, R. Al-Maaitah, and S. Bany Hani, “Quality of nursing documentation: Paper-based health records versus electronic-based health records,” J. Clin. Nurs., 2018, doi: 10.1111/jocn.14097.

[3] Y. Chen, S. Ding, Z. Xu, H. Zheng, and S. Yang, “Blockchain-Based Medical Records Secure Storage and Medical Service Framework,” J. Med. Syst., 2018, doi: 10.1007/s10916-018-1121-4.

[4] W. K. Evans, F. D. Ashbury, J. Pun, G. L. Hogue, and A. Smith, “Implementing a regional oncology information system: Approach and lessons learned,” Curr. Oncol., 2014, doi: 10.3747/co.21.1923.

[5] E. N. Mupela, P. Mustarde, and H. L. C. Jones, “Telemedicine in Primary Health: The Virtual Doctor Project Zambia,” Philosophy, Ethics, and Humanities in Medicine. 2011, doi: 10.1186/1747-5341-6-9.

[6] American Cancer Society, “Cancer Facts and Figures 2017,” Genes Dev., 2017, doi: 10.1101/gad.1593107.

Our DIMTECH idea includes three solutions including; 1) a novel screening method for CC for women living with HIV; 2) a successfully proven oncology system that will increase medical data management, and 3) media monitoring of public-health related Ugandan traditional news and social media content to identify gaps in primary care relating to human papillomavirus and cervical cancer care.

DIMTECH proposes to employ a novel screening method and data monitoring approach, where women living with HIV (WLWH) are also screened for HPV and CC as an additional and beneficial service. Using trained female nurses, a simple, fast and low-cost technique will be administered that requires only visual inspection of CC lesions on the cervix upon application of 5% acetic acid and gene Xpert for HPV 16 and 18 tastings. This approach reduces or eliminates self-selection for the receipt of such a service by all WLWH attending anti-retroviral treatment (ART) clinics. Further, the novel treatment eliminates the transport costs of seeking screening alone if such a service is not integrated into local clinics. In low-income countries, HIV care uptake is relatively expensive, therefore services are not widely distributed within primary health care.

Health workers involved with CC screening will be introduced to the Oncology Information Management Platform (ONCIMS) to enhance data capture, monitoring, treatment and tracking of women screened for CC. ONCIMS is a series of next-generation integrated Cancer Data Management AI-powered digital Suit for resource-constrained areas. The suit consists of a web platform and mobile app for CC cancer and HIV management in resource-constrained settings. The platform also has a media section for patient and HW worker interactions for consultations, feedback, advice and next appointments.

To understand community sentiments from the media platform to improve primary health care CC management performance, we will use AFS Next-Generation Mission Analytics (NMA), a web-based data aggregation, enrichment, and analytics platform. Each day the AFS public health team monitors and collects coverage from more than 400 African traditional news outlets, as well as hundreds of thousands of posts from African social media users across multiple platforms. NMA will perform as a unitary data storage hub for all public-health-related Uganda-based open-source information, including auxiliary data such as translated transcriptions of radio and TV broadcasts. Using native language machine learning algorithms, media content that is public-health related will automatically be pulled into the NMA database, and then enriched in the NMA environment. Each public-health-related germane quote or statement identified in the content will be assigned a direct sentiment and a key topic or public health concern. Additionally, key misinformation narratives, which may harm adherence to public health and safety measures or attendance of appointments will be identified. This media data will provide insight into key narratives and shifts in sentiment toward specific areas of HPV and CC healthcare performance.

The three-tiered approach of the novel treatment, state-of-the-art oncology system, and public-health media monitoring will benefit women of reproductive age (15-49 years) who visit ART clinics in Uganda. Incidence of HPV and CC still remains high among WLWH; yet usually, individuals who are poor, none educated, and economically challenged do not consider CC screening a priority, and CC screening services are often non-existent in rural ART clinics. Therefore, WLWH ART facilities will also be screened for ART and CC as an additional service at no cost. This integration will help women avoid extra travel costs, and potential stigma for seeking CC screening services, and may not need a male decision to allow them to visit the facility.

The secondary beneficiaries will include health workers to reduce the burden of paper-based medical records. Ministry of Health to: (i) register all the screening centres and provide online collaboration (ii) view patient statistical reports from screening centres, (iii) manage human resources at each centre, (iv) manage resource allocation, and demand forecasting based on statics, and (v) provide instant communication to the public about cancer burden and other relevant information. Cancer screening centres to: (i) manage cancer patients' information including automated diagnosis; automated patient clerking, protocol management, tumour board meetings management, follow-up, imaging with automated image analysis, (ii) communicate among themselves including telemedicine support, and picture archive communication with real-time online collaboration and diagnosis, (iii) communicate to Uganda Cancer Institute, (iv) share images (Biopsy and Pap smears) for telemedicine support with support for automated analysis, digitize pap-smear images with low-cost digital microscope slide scanner, (v) post anonymous data through the platform wise data module for public research and (v) request resources from Ministry of Health; and other private development partners. 

The project will be led by Makerere University, School of Public Health (MakSPH). MakSPH is a leading public health training institution in Uganda and East Africa whose main responsibility is to train, carry out research, and provide service delivery in public health. The school has strong local, regional, and international partners, and a successful track record regarding the management of funds from renowned organizations including but not limited to the U.S. Department for International Development, European Union, Save the Children, UNICEF, Gates Foundation, and Comic Relief. Agnes Nyabigambo, a Research Associate as well as Health Services Research Specialist and innovation expert at MakSPH will coordinate DIMTECH project activities.

 MakSPH through funding from USAID supported a team at the Department of Biomedical Sciences and Engineering at Mbarara University in Uganda to develop an oncology Information Management Platform (ONCIMS) to enhance data capture, monitoring, treatment and tracking of cancer. William Wasswa, a Senior Lecturer at Mbarara University, will support the updating and monitoring of the ONCIMS system.

 MakSPH is partnering with Accenture Federal Services (AFS) to create a novel, low-cost, fit-for-purpose solution to measure primary health care performance around human papillomavirus (HPV) and CC care in HIV/AIDS clinics in Uganda. Since the Ebola outbreak in Western Africa in 2014, the AFS public health team has compiled, enriched (coded), and analyzed local, national, and international traditional news and social media content to enable African governments and health institutions to monitor perceptions of public health interventions and identify key misinformation narratives that harm the success of health campaigns. Since the Ebola outbreak, AFS has worked with and provided weekly reports to the U.S. National Institutes of Allergy and Infectious Disease (NIAID), Africa Centres for Disease Control and Prevention (Africa CDC), African Field Epidemiology Network (AFNET), and the Joint United Nations Programme on HIV/AIDS (UNAIDS) to inform a wide range of stakeholders including African ministries of health regarding trends in public opinion and gaps in public health performance identified in open-source media content. Neil Edwards, an Open-Source African Media Specialist, will support building capacity and media monitoring of public-health-related Ugandan traditional news and social media content to identify gaps in primary care relating to human papillomavirus and cervical cancer care.

First and foremost, our interdisciplinary team hopes to bear our three-tier novel solutions to address gaps in HPV and CC screening in Ugandan ART clinics. On the medical treatment front, we would like to develop and test the effectiveness of the DIMTECH solution. The solution will be implemented in HIV clinics providing HPV and CC screening services. This approach will be implemented in four phases.

Phase 1: streaming HPV and CC screening services in 6 ART clinics. We will work with the health workers (HWs) within the ART clinics to standard the CC screening services in terms of setting up data capture spaces and patient flow.

Phase 2: updating the ONCIMS platform including the media interactions sections for toll-free calling or text messages. This will require the following: (1) Project design. The study will use an explanatory project design. It is from this basis that we will be able to get data in form of barriers, enablers, stakeholders and clinical workflow best practices for the adoption of digital technologies for oncology management in Uganda. (2) Study population. Oncology is integrated care involving several health care workers and other stakeholders. For this project, information will be obtained from healthcare workers from all HIV clinics in Uganda. Information will also be obtained from patients attending the cancer screening to obtain feedback on how our technology can be improved to better service delivery even to those in rural areas with no power and limited internet connectivity. (3) Sampling. Sample size and determination will be achieved using the saturation point approach. Purposive sampling will be used to select the health care workers to participate in these sessions. On the other hand, stratified random sampling will be used to select the patients for their views on using digital technologies for oncology services. (4) Ethical considerations. Informed consent will be obtained from all respondents willing to participate in the study. The innovator will also seek for Project ethics clearance from MakSPH-HDREC.

We will then test the solution within the selected ART clinics. In this project, we shall develop and implement digital technology based on the results in Step 1. In step 1, we shall deeply investigate the ideal modules and digital workflow to be deployed for digital oncology in resource-constrained settings. Once these are established, the platform will be developed using a socio-technical approach coupled with a user-centred design. As shown in figure 1  https://ibb.co/tx3r4jL. 

Social-technical approach arose primarily to explain why "technically excellent" software packages were often rejected by their users. It became apparent that users of the system played a very significant role in determining whether or not the system could be used successfully. Introducing a computer system can inevitably alter the structure of the organization that receives it. This alteration leads to changes in the allocation of manpower, resources or workload within that organization hence any attempts to alter the structure of the organization through changes in clinical workflows are likely to result in, at the very least, social inertia from those who see their influence is reduced or rendered useless. However, digital technologies like Artificial Intelligence should only enhance human skills and capability does not eliminate them. The user-centred design will ensure that the potential end users are part of the design process. The web-based platform will be developed using Python programming language, locally stored on NITA-U data centres, with limited data stored in the cloud via a custom secure gateway. The mobile App will be Android based with plans of having an apple version.

Phase 3: The media metrics generated by AFS Next-Generation Mission Analytics (NMA) will identify gaps in and sentiment toward public health care performance. Additionally, metrics will identify key pieces of misinformation that may harm public health response and performance. Analysis of these metrics, reported to the Ugandan Ministry of Health, will provide actionable insights into changes needed at specific clinics and in the Ugandan public health system at large.

Specific Tasks for Developing Public Health Media Monitoring Solution

1.     Expand Open-Source Media Collection in Uganda

In addition to the already extensive Uganda-focused traditional and social media ingested into NMA, MakSPH will contact Ugandan radio stations, broadcast television networks, Telcom companies, Pulse Lab, and The Medical Concierge Group (TMCG), and other relevant media houses to collect as much data as possible. The data will include both digitally formatted and hard copies. It will be the responsibility of MakSPH to collate these data and upload relevant digitally formatted copy materials into the NMA platform.

2.    Develop a customized taxonomy and operational research design

MakSPH and AFS will work together to finalize a custom research taxonomy detailing how media content will be tagged, including key topics and sentiment. As shown in figure 2 https://ibb.co/Tw5G48Y. In addition to these metrics, all media data will be geolocated to the national level. For radio and social media data, geolocations often occur at the subnational level, e.g. to a province, city/town, or village, allowing trends to be observed hyperlocally.

1. In-Person Training of Coding Team: AFS will conduct initial, in-person training for the enrichment and coding team comprised of MakSPH graduate students. The purpose of this training is to ensure that all coders are fully aware of how to enrich content using the taxonomy jointly developed by MakSPH and AFS teams.

 AFS will be responsible for training the following:

·      How to use NMA to code media content to reflect agreed-upon customized taxonomy.

●      How to upload articles and media content to NMA;

●      Training team to use Google Sheets and V Lookup Tabs to produce sub-national level Twitter data.

MakSPH will be responsible for training the following:

●      Training on the translation of qualitative data from local languages to English

4. Pretesting and piloting of tools

With assistance from AFS, MakSPH resources will upload relevant auxiliary data and media content to NMA. In addition to the currently monitored Ugandan traditional and social media content provided by AFS, the NMA platform will be optimized to ensure coding can be conducted according to the agreed-upon custom taxonomy.

5. Analysis and presentation

The analytical framework will be co-created by AFS and MakSPH. The MakSPH team, with support from AFS, will conduct data analysis to be used in a final report. AFS will provide consultation and guidance on media analytics using its decades of experience, and both organizations will be involved in the presentation of findings. 

Phase 4: Another objective will determine the effectiveness of the DIMTECH approach. Identified healthcare workers will be trained on how to use the platform and will be monitored while using the platform. The outcome measure will be the usability and usefulness of the platform. The selection of the 44 health workers to be part of the survey will entirely be purposive. We shall also evaluate patients throughout to gauge their adherence to appointments/follow-ups. We will further determine the cost-effectiveness of the DIMTECH approach.

We will develop a Microsoft Excel spreadsheet model of the cost-benefit analysis and cost-effectiveness of the DIMTECH approach. The main outcomes of the economic evaluation will be the cost per woman screened and the cost per early CC detected. Depending on the effectiveness results, we will perform a comparison of standard practice and DIMTECH approaches. We also propose to perform a static aggregate budget impact analysis (BIA) to examine the potential economic value of scaling the DIMTECH approach in the HIV clinics and the fiscal implications to the government and the individual taxpayer of implementing this approach.

We will perform a primary micro-costing of the DIMTECH approach at each of the HIV clinics from the modified societal perspective i.e. include all direct medical, all direct non-medical costs and the opportunity cost of lost patient time. The direct medical costs will include all resources used in screening including equipment, supplies and personnel. We will perform a detailed inventory of CC screening resource inputs. The direct non-medical costs will include transport and upkeep while attending screening visits. Data on direct non-medical costs will be collected in the exit interview. Data on personnel resource use and patient waiting time will be collected in the time-motion survey.

The model will be validated by varying inputs to check if the expected values are logical and set costs and outcomes to 0 separately to check if identical expected values are obtained for the comparators. The baseline analysis will consist of a calculation of the costs, the number of women screened and the number of women with CC lesions identified, and the incremental cost-effectiveness ratios—cost per woman screened and cost per case detected. A cost per case accurately detected will be calculated depending on the results of specific objectives 1 & 2 concordance of DIMTECH against a standard will be estimated to investigate the impact of parameter uncertainty on payoffs. Empirically derived confidence intervals will be used when available.

One-way sensitivity analyses will be performed on all parameter inputs to generate tornado diagrams representing the most influential parameters. Distributions will be assigned to all input parameters and probabilistic sensitivity analysis will be performed using a second-order Monte Carlo simulation. This will enable the generation of a cost-effectiveness acceptability curve to represent the impact of parameter uncertainty on estimates of cost-effectiveness. Budget impact analysis is an essential part of a comprehensive economic assessment of a health care technology or program. We propose to estimate the aggregate national budget impact as well as the cost per Ugandan taxpayer per year of implementing DIMTECH the approach in the HIV clinics. The BIA will involve the following steps: 1) characterization of the population of Uganda to estimate the potential number of people to be impacted by a scaled-up DIMTECH, 2) selection of a time horizon between 1 and 3 years.

Our DIMTECH idea is a novel screening method and data monitoring approach, where women living with HIV (WLWH) are also screened for HPV and CC as an additional and beneficial service. Using trained female nurses, a simple, fast and low-cost technique is administered that requires only visual inspection of CC lesions on the cervix upon application of 5% acetic acid and gene Xpert for HPV 16 and 18 testings. This approach reduces or eliminates self-selection for the receipt of such a service by all WLWH attending ART clinics. Further, the solution eliminates the transport costs of seeking screening alone if such a service is not integrated into local clinics. Health workers involved with CC screening will be introduced to the Oncology Information Management Platform (ONCIMS) to enhance data capture, monitoring, treatment and tracking of women screened for CC. ONCIMS is a series of next-generation integrated Cancer Data Management AI-powered digital Suit for resource-constrained areas. The suit consists of a web platform and mobile app for CC cancer and HIV management in resource-constrained settings. To understand community sentiments from the media platform to improve PHC CC management performance, we will use AFS Next-Generation Mission Analytics (NMA), a web-based data aggregation, enrichment, and analytics platform. Media analytics from NMA will provide insight into key narratives and shifts in sentiment toward specific areas of HPV and CC healthcare performance. We believe this three-tier solution will increase the demand for CC screening services in HIV clinics, create a better data management system, and identify gaps in HPC and CC primary care performance. In turn, the novel solutions will produce better prevention, treatment, and early detection of CC.

In the initial phases (1-3), our team hopes to validate the DIMTECH approach from one HIV clinic to 6 ART clinics among WLWH 15 - 49 years attending routine ART care. The main aim of this phase is to determine the effectiveness of the DIMTECH approach among WLWH as a strategy to increase access and uptake of the CC screening as well as monitoring and tracking women with the CC disease. Effectiveness will be determined using prospective Impact Evaluation and full health economic evaluation in 9 months. In phase 4, our team aims to break through the proof of concept phase and find other African countries to implement our tier-tier solution. Makerere University, School of Public Health, leads the Resilient Africa Network (RAN), a consortium of 26 African university stakeholders from 16 countries that strengthen capacity and local innovative evidence-based solutions in the field of HIV/AIDS. Implementing these solutions in the countries of the RAN network that would fit the model would be the next step.   

Upon integration of the DIMTECH into the ART clinics, the following indicators will be monitored;

1. Number of HWS trained

2. Number of WLWH screened for CC and HPV.

4. Number of WLWH identified to have CC and HPV

5. Number of WLWH with HPV and CC followed up

Ugandan media data ingested into the Next-Generation Analytics platform will automatically be tagged with basic information, including news organization, username, media type, geolocation, date, and the text. Data relevant to HPV and cervical cancer, which will be automatically sorted using native language machine learning algorithms, will then be coded by Makerere University graduate students. Key indicators in the coding process include directed sentiment (positive, negative, or neutral toward specific entities including medical facilities and the government), key topics/concerns, and misinformation. See the attached link for the full list. 

The Unified Theory of acceptance and use of technology model (UTAUT) will be used [7]. UTAUT is the unification of several frameworks that help to explain the acceptance and use of technology. These frameworks include, Diffusion of Innovation Theory, Theory of Reasoned Action (TRA), Technology Acceptance Model (TAM), Theory of Planned Behavior (TPB), Combined Technology Acceptance Model and Theory of Planned Behavior, Model of PC utilization and Social Cognitive Theory.

Reference

[7] A. M. Momani, “The Unified Theory of Acceptance and Use of Technology,” Int. J. Sociotechnology Knowl. Dev., 2020, doi: 10.4018/ijskd.2020070105.

AFS Next-Generation Analytics (NMA) delivers near real-time pattern and behavioural analysis of social, print, and broadcast media content and incident data. NMA’s use of machine learning and human curation processes creates structured data from multiple streams of publicly available information.

Using native-language analysis of messaging and sentiment, NMA enables the discovery of trends and shifts in behaviour of correlated entities, providing quantitative viability into emerging threats or opportunities in a target information environment.

By fusing multiple data types, accurately curating the data using a methodology customized to each customer’s use case, and making the data available in a web-based data exploitation platform, NMA provides analysts, policymakers, and data scientists access to a common operating picture with insights into their problem set.

AFS specializes in supporting international security and crisis response in non-permissive environments, where unbiased data collection is challenging, and disinformation is rampant. Customers use NMA to inform strategy and tactics in support of security, aid, and crisis response.

The health Workers will be screening women for CC in the ART clinics and will collect the data for the DIMTECH solution.

Our DIMTECH approach will initially be tested in the rural ART clinics where there are barely any HPV or CC screening services. The approach mainly targets WLWH in reproductive age (15-49 years) however, we will also include HPV screening during the scale-up phase. 

Our business model is to scale our operations past Uganda and into other African public health spaces. Following the success of our pilot program in Uganda, our three-tier approach can easily be scaled to other health care systems. Our plan is to target African Ministries of Health.

We require an initial investment of USD 100,000 to develop and test DIMTECH. Once the pilot program provides proof of concept, our team will reach out to the Africa CDC, UNAIDS, and other African Ministries of Health to showcase its feasibility in other African localities. We hope to secure funding from one of these agencies.

In addition to securing outside funding, we will use a free-for-service approach, where we will sell our services directly to the African health ministries. In each case, an appropriate fee structure will be established to cover data and labour upkeep. 

With support from Wellcome Trust of GBP 10000, we have tested a novel clinic-based HPV self-sampling approach among 372  women attending ART clinic at Luweero district hospital.  Through MakSPH-RAN, USAID supported the development of the ONCIMS with USD 5000. This approach has been tested at Mbarara Regional Referral Hospital among 150 patients. However, the novel 3-tier DIMTECH approach has not yet received any funding.