Cervical cancer accounts for 21.7% of all cancer deaths in the sub-Saharan Africa with a case fatality rate of 68%. Nigeria’s Federal Ministry of Health has adopted cervical cancer screening (CCS) using visual inspection with acetic acid or Lugol’s iodine (VIA/VILI) and cryotherapy treatment for precancerous lesions as the preferred screening and treatment strategy. Using the Exploration, Preparation, Implementation and Sustainment Framework, our study documents our experience during the development, piloting and roll-out of the APIN Public Health Initiatives (APIN)-developed VIA Visual Application (AVIVA) for CCS using the VIA method in 86 APIN-supported health facilities across 7 states in Nigeria. Between December 2019 and June 2022, with the aid of 9 gynaecologists and 133 case finders, a total of 29 262 women living with HIV received VIA-based CCS and 1609 of them were VIA-positive, corresponding to VIA positivity rate of 5.5%. Over the 30 months duration and the 5 phases of CCS scale-up, AVIVA development and expansion, a total of 1247 cases were shared via the AVIVA App (3741 pictures), with 1058 of such cases undergoing expert review, corresponding to a reviewer rate of 84.8%. Overall, the use of the AVIVA App improved both the VIA-positive and VIA-negative concordance rates by 16 percentage points each (26%–42% and 80%–96%, respectively) from baseline to the end of the study. We concluded that the AVIVA App is an innovative tool to improve CCS rates and diagnostic precision by connecting health facility staff and expert reviewers in resource-limited settings.
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Approximately half of the 600,000 women who develop cervical cancer succumb to the disease, and 90% of these deaths occur in low- or middle-income nations (LMICs). In LMICs, the Ministry of Health adopts visual inspection with acetic acid or Lugol's iodine (VIA/VILI) and cryotherapy as the preferred cervical cancer screening and treatment technique. Literature demonstrates that the VIA is a subjective screening test with a propensity to overdiagnose precancerous cervical lesions due to its relatively low positive predictive value.
APIN Public Health Initiatives introduced multiple levels of review for stained images of the cervix during VIA using the APIN-developed VIA Visual Application (AVIVA). This was done to improve the accuracy of VIA-based screening tests and reduce the social and psychological effects of misdiagnosis.
Using the Exploration, Preparation, Implementation, and Sustainment (EPIS) framework, our study documents the development, piloting, and rollout of the APIN Public Health Initiatives (APIN)-developed VIA Visual Application (AVIVA) for cervical cancer screening (CCS) among women living with HIV (WLHIV).
Globally, half of the estimated 600 000 women who develop cervical cancer die from the disease, and 90% of these deaths occur in low-income and middle-income countries.1 2 An estimated >60% of women eligible for cervical cancer treatment in sub-Saharan Africa (SSA) live in rural and sub-urban areas where screening rates are below 20%.3–5 In 2018, cervical cancer accounted for 21.7% of all cancer deaths in SSA with a case fatality rate of 68% and 33% survival rate at 5 years postdiagnosis.1 6 Some of the major drivers of high cervical cancer incidence in SSA include the high prevalence of HIV and high-risk Human papillomavirus.11
In Nigeria, where cervical cancer screening (CCS) is five times lower in rural vs urban communities,4 5 women ages 15–49 are more than twice as likely as their male counterparts to be living with HIV (1.9% vs 0.9%).2 There is also a higher prevalence of high-risk HPV and cervical precancerous lesions among women living with HIV (WLHIV) compared with HIV-negative women.6–8 Furthermore, less than 15% of the women ages 15–49 have ever had CCS.6 9
The Federal Ministry of Health (FMOH) in Nigeria adopted CCS using visual inspection with acetic acid or Lugol’s iodine (VIA/VILI) and cryotherapy treatment for precancerous lesions as the preferred screening and treatment strategy.10 The WHO recommended that VIA alone should be the initial test to start implementation of CCS programme in a low-resource settings, and where resources are available, triage should be done to confirm the diagnosis before treatment. However, available literature has shown that VIA is a subjective screening test with a tendency for overdiagnosis of precancerous cervical lesions due to its relatively low positive predictive value.11–14 A recent study in a primary health setting in the Democratic Republic of Congo showed a higher screening sensitivity at cervical cancer suspicion and threshold level using VIA or VILI done by the physician (90.3%) compared with those done by nurses (7%) following 10 days of colposcopy and VIA screening training for a cohort of physicians and nurses.15 The VIA screening sensitivity was also higher when done by a gynaecological oncologist compared with those done by a general medical practitioner.15 Similarly, the specificity and negative predictive value were higher with physicians compared with non-physicians.15 The subjectivity of VIA can be improved when the screening is performed or reviewed by an experienced specialist.15
Overall, 85% of the health facilities in Nigeria are primary health centres (PHCs),16 and up to 75% of women in rural areas and urban slums use the PHCs,17 making them veritable sites for screening of most Nigerian women for cervical cancer. Most of the PHCs in Nigeria are manned by community health extension workers (CHEWs), with few doctors and nurses seen at the comprehensive PHCs which are higher level PHCs.18 CHEWs are a lower cadre of health workers with less expertise in the use of diagnostic equipment for disease management, hence, the need for supervision and work review by a higher cadre of staff for improved diagnostic accuracy of diseases. Data from the five largest PEPFAR-supported implementing partners (IPs) in Nigeria shows that 66% of health facilities rendering HIV services are PHCs. With the roll-out of CCS using VIA by the PEPFAR HIV programme in Nigeria in 2019, there was a need to provide a higher-level review of cases diagnosed as precancerous lesions or cervical cancer at these levels of healthcare.
In response to the FMOH call to lower the incidence, morbidity and mortality of cervical cancer in Nigeria and among at-risk population of HIV infected women, APIN Public Health Initiatives (APIN) introduced CCS programme using VIA among HIV infected women in 2019.3 10 Furthermore, to improve VIA-based screening test accuracy and ameliorate the accompanying psychosocial implications of misdiagnosis, APIN introduced multiple levels of review for stained images of the cervix during VIA using the APIN-developed VIA Visual Application (AVIVA). AVIVA is a lightweight android mobile application designed using Google’s front-line programming language Angular version 11 and packaged for mobile deployment using Ionic and running on the Apache Cordova mobile application development framework. The framework implements several services that enhance the VIA screening service both for the case finder and the specialist reviewer. One of the services is the Hypertext Transfer Protocol Secure Application Programming Interface (API). This service runs a real-time handshake with a cloud-based data warehouse system for the storage of all captured images and analytics to ascertain the level of concordance between case finder diagnosis and reviewer diagnosis. It also tracks the progress of the case finder in bridging the concordance gap and displays performance at case finder, facility, state and programme levels, respectively. The system also has a notification service that ensures real-time alerts and feedback mechanism, including short message service (SMS) for case finders when their findings are reviewed. In this article, we document our experience during the development, piloting and roll-out of the AVIVA for CCS using the VIA method across 86 APIN-supported health facilities in Nigeria.
We use the Exploration, Preparation, Implementation and Sustainment (EPIS) framework,19 20 to report on the creation of the AVIVA, strategy development and implementation to enhance the screening and diagnosis of cervical precancerous lesions among WLHIV. The EPIS comprises four distinct phases that describe the implementation process of AVIVA and enumerates common and unique factors within and across levels of outer context (system) and inner (organisational) context across the phases.19 20 The exploration phase looks at the new or ongoing health requirements of the patients or communities and works to find the appropriate evidence-based practices (EBPs) to address these needs and decide their subsequent adoption. In addition, within the EBP itself, adaptation takes place at the system, organisational environment and individual levels.19 20 The key objectives of the preparatory phase were to identify potential implementation barriers, facilitate the external and internal settings, develop a strategy to optimise implementation of facilitators and minimise potential barriers. The inner contexts focus on the culture and systems of organisations implementing EBPs,19 20 while the outer contexts focus on external factors such as government policies, legislation, funding and grants that may facilitate or hinder implementation.19 20 The EBP is initiated during the implementation phase, with continuous monitoring and evaluation during the implementation process guiding needed adjustments.19 20 The outer and inner context structures, processes and supports are updated during the sustainment phase so that the EBP can be maintained, with or without adaptation, to realise the public health impact that results from the implemented EBPs.19 20
Overall needs assessment and stakeholder engagement
Nigeria disproportionately contributes to the burden of cervical cancer cases worldwide, with an estimated 14 089 cases diagnosed annually.3 In response, the National Strategic Plan for Prevention and Control of Cancer of the Cervix was introduced in 2017 by the FMOH to lower cervical cancer incidence, morbidity and mortality, especially among HIV-infected women.10 21 22 APIN began integrating VIA-based CCS into its regular HIV programme in phases from December 2019 with support from PEPFAR. We identified key stakeholders in the CCS programme’s implementation across these four domains: (1) Policy-makers—Ministry of Health; (2) Providers—healthcare providers at primary, secondary and tertiary facilities; (3) Payers—PEPFAR-supported implementing partners and (4) Patients—WLHIV. We held multiple stakeholder discussion forums on reducing the risk of misdiagnosis associated with VIA screening, especially among lower cadre of health professionals.23 24 At each meeting, we conducted key interviews with the stakeholders focusing on the programme needs, the readiness for service delivery and existing strategies. The key interviews revealed barriers to an effective VIA-driven screening and treatment programme including inadequate skills among healthcare workers, insufficient equipment and consumables to support CCS, and the subjectivity of VIA screening results. Furthermore, a pre-AVIVA programme review was done 1 month following the phase 1 expansion which revealed that only 9 out of 35 (25.7%) people initially classified as VIA-positive and eligible for thermal ablation were truly VIA-positive, after confirmation by a gynaecologist. The solutions proffered included (1) the need for site assessment and service readiness; (2) the development of a real-time app that could share high-quality cervical images with an experienced reviewer to make an instant decision to guide treatment following WHO see- and- treat guidelines25 and (3) ongoing mentoring and supportive supervision.
Programme pilot phase:
During the first phase (pilot phase), two tertiary centres were selected—University College Hospital, Ibadan and Jos University Teaching Hospital, Jos. These facilities were selected to provide a learning experience and guide future scale up. The two sites were selected because they (A) have a gynaecologist who is involved in the day-to-day running and oversight of the HIV programme, (B) the gynaecologists were members of the National Prevention of Mother to Child Transmission of HIV (PMTCT) technical working group and cervical cancer technical working group and (C) the specialists were willing to train other healthcare workers on CCS and management. The learning experience led to the development of a client flow map for integrating the CCS programme into the Antiretroviral treatment (ART) services (figure 1) and a health talk guide that incorporated CCS and programme messages into the usual HIV programme health talk. The health talk themes focus on CCS procedures, benefits and follow-up care after screening, HPV vaccination and cervical cancer, and the relationship between HIV and cervical cancer. We had flyers and posters with educational information in the supported health facilities.
Site needs and implementation readiness assessment for CCS programme
Following the pilot phase, APIN conducted site assessment and implementation readiness for the CCS programme in 2020. Each proposed site was assessed using the WHO Service Availability and Readiness Assessment,26 tool which has been used widely in Nigeria.27 28 The sites were assessed across the following domains: human resource, adherence to protocol, equipment and commodities, diagnostics, infrastructure, and service availability. Domain weighted averages were computed for each site and the domain scores were summed up to obtain site composite service readiness scores. The readiness scores were used to categorise which facilities would be activated during each phase, as well as the supplies needed to ensure quality service delivery. An additional criterion was used to select initial facilities to be activated based on (1) HIV treatment availability for women aged 25–49, (2) geographical distribution within the state, (3) absence of a CCS programme in the ART clinic and (4) the site’s commitment to drive and own the programme.
Site selection and engagement of selected sites
To ensure room for CCS and treatment programmes in the ART clinics and the commitment of site leadership to drive the process, APIN engaged the sites to perform an informal leadership climate assessment. We computed the minimum sample size of 136, as the sample size required to infer on the cervical cancer VIA screening accuracy using the Buderer et al,29 30 2006 sample size formula at a 95% confidence limit, with a sensitivity of 50%, specificity of 85% and prevalence of 10.75. The programme team met to select sites to be activated in phases based on the criteria mentioned above and funding availability, ensuring that the number of study participants is more than the computed sample size. Figure 2 shows the number of facilities activated per implementation phase. Understanding patient flow for CCS among WLHIV (see figure 3 for adopted clinic flow chart) at the clinics and including CCS health talks within the general HIV health talks were two key takeaways from the pilot phase.
Identification and training of designated healthcare workers
State cervical cancer focal persons from APIN had key informant interviews with the leadership of the health facilities to select healthcare workers with the training and aptitude for preventive services. The selection criteria included: (A) previous experience in CCS; (B) experience in HIV programme; (C) interest in preventive services and (D) requisite educational qualification (trained as either nurses, CHEWs or physicians). The APIN team conducted training for the selected healthcare workers using the national training manual for cervical cancer secondary prevention,25 and a WHO CCS agenda. The number of healthcare professionals trained in each supported state is shown in table 1.
Supply of equipment and consumables
APIN provided the needed supplies and tools to commence the CCS programme. Consumables are monitored and periodically replenished to prevent stockouts and ensure continuity of the free CCS programmes for WLHIV. The list of tools and supplies is in table 2.
Phase 1 scale up: pre-AVIVA phase 1
During the pre-AVIVA phase 1, we expanded from the two tertiary sites to the secondary level health facilities and included faith-based health facilities. The facilities selected at this phase had ≥1500 WLHIV (>14 years old) on treatment. A total of 17 health facilities were activated to offer CCS programmes and use AVIVA in June 2020. These facilities were selected across seven states in Nigeria, with Benue having 11 facilities and Oyo, Ondo, Osun, Ogun, Ekiti and Plateau states having one facility each.
Interim programme evaluation
Within 1 month of the pre-AVIVA phase 1 expansion, we identified misdiagnosis of WLHIV who consented for VIA screening. Out of 35 women who were randomly selected from the total of 117 identified as VIA positive and classified as eligible for thermal ablation, only 9 (25.7%) were later determined by a supporting trained gynaecologists to be positive. This discovery resulted in the institutionalisation of ongoing supportive supervision and mentoring of trained healthcare workers providing services in lower-level health facilities by gynaecologists and the development of AVIVA. In total, during this phase, 2892 women were screened for cervical cancer with 117 identified as VIA positive by trained healthcare workers.
Development of the APIN VIA Visual Application
To improve the diagnostic precision of the VIA done at the secondary-level facilities, APIN introduced multiple levels of reviews of VIA cervix-stained images using the APIN-developed mobile application named AVIVA. AVIVA works by capturing pictures of VIA-stained cervix and transmitting them in real time to the expert reviewers (trained gynaecologists) who provide diagnostic and treatment guidance. AVIVA operates in the following way: a health worker, known as a case finder, uses the AVIVA (installed on a phone) to capture images of the cervix, adds relevant patient information and uploads them to the online server (the application can automatically trigger an offline mode if there is no internet network). The trained gynaecologists assigned to that facility or state then receives a notification within the app and through SMS prompting them to open the app, download the images, provide a diagnosis and recommend a treatment option. After this, the case finder gets a notification with the reviewer’s diagnosis and makes appropriate documentation. At the end of the review, the images are committed to the server and cleared from the AVIVA App. To support the process, APIN provided phones (Samsung Galaxy A20s) and monthly phone credits to the expert reviewers and the case finders. Our team chose the Samsung Galaxy A20s due to its high-resolution camera with a 13 MP primary lens, autofocus capabilities and low-light performance, real-time image preview on its large 6.5-inch display and its image stabilisation mechanisms. These features are crucial for capturing near-real images of the cervix during the VIA examination and reducing the difference between seeing the stained cervix in person and a virtual interpretation.
Gynaecologist expert reviewers
We selected nine gynaecologists as expert reviewers to review AVIVA app cervical images across the seven states (Benue-2, Oyo-2, Plateau-1, Ondo-1, Osun-1 and Ekiti-1). The selection criteria were: (1) previous experience working with WLHIV, such as providing prevention of mother-to-child transmission of HIV services; (2) specialisation in gynaeoncology and (3) interest in committing time and effort to reviewing cervical images. The gynaecologist expert reviewers are either fellows of the West African College of Surgeons or the National Postgraduate Medical College of Nigeria. In total, they have an average of 8 years postspecialisation qualification experience. The reviewers have remained the same during the pre-AVIVA and post-AVIVA phases. They continue to train and provide ongoing mentorship and supportive supervision to our site case finders from inception until date.
Phase 1 Scale-up: post-AVIVA Phase 1
Following the launch of AVIVA in November 2020, we piloted the app at the ART clinic at the University College Hospital Ibadan. AVIVA and CCS was then scaled up to the 19 existing sites (2 pilot sites and 17 phase 1 sites) by January 2021. However, only 10 of the sites shared cervical images of all patients screened for VIA through the AVIVA app. In total, during this phase, 5184 WLHIV were screened for cervical cancer and 432 were identified as VIA-positive by trained healthcare workers in the 19 sites.
A total of 296 cervical images were shared by the case finders using the AVIVA app from the 10 sites that uploaded images (152 VIA-positive and 144 VIA-negative). Sixty-five of the 152 images classified as VIA-positive by trained healthcare workers were marked positive after the gynaecologist’s review giving a positive concordance rate of 42.8%. This is a 17% increase from the pre-AVIVA finding. Furthermore, 115 out of the 144 images that were classified as VIA-negative by case finders were marked negative after review by the gynaecologist giving a negative concordance rate of approximately 80%. Table 3 and figure 4A,B show the positive and negative concordance rates for each of the phases of implementation.
We had only one reviewer per case images shared by trained healthcare workers. Hence, we could not determine the level of agreement between reviewers which may also improve the validation of their diagnosis if more than one reviewer agreed. Figure 5 shows the flow of cases images from trained healthcare workers to expert reviewers.
Phase 2 scale up: post-AVIVA
In 2021, APIN expanded CCS to an additional 42 health facilities (38 secondary and 4 primary health facilities) across 7 states in Nigeria bringing the total number of sites to 61. Thirty facilities shared cervical images of all patients screened for cervical cancer through the AVIVA app. The key consideration at this phase for site selection was the geographical spread of sites offering CCS and improving access within the seven supported states. During this phase, 16 465 WLHIV were screened for cervical cancer and 907 were identified as VIA-positive. A total of 902 images of the cervix were uploaded using the AVIVA App (180 VIA-positive and 722 VIA-negative). By the end of phase 2, 38.9% (70/180) of images uploaded on AVIVA classified as VIA-positive were marked positive after review while 92.9% (671/722) of the VIA-negatives were marked negative after review by the gynaecologist (see table 3).
Lessons learnt from phase 2
One important lesson from this phase was that primary healthcare facilities can provide high-quality CCS services if they are supported by mentorship and technology like the AVIVA App. The activated primary healthcare facilities had a 91% positive concordance rate and 100% negative concordance rate. We also identified the following gaps with the AVIVA App in the field during this phase:
There was a delay in reviewing the pictures sent on the AVIVA App because there was no notification beep.
There was a challenge in tracking patients because of the absence of patient identifiers and the form does not allow the case finder to see baseline and follow-up pictures on the same layout.
There was inadequate technical guidance on treatment options for VIA-positive cases.
Our team then conducted user-group consultation meetings with the specialist reviewers to ensure regular feedback, facilitate review of the AVIVA App interfaces, and adapt the technology to programme needs and local context to make it more user friendly. Key information obtained include: (A) the inclusion of patient IDs in line with the Nigeria HIV patient registration numbering system. Prior to this, it was difficult to link the CCS and treatment records to the women’s HIV clinic records; (B) the need to include an image-uploaded notification beeping service to enable expert reviewers to respond quickly to review requests from healthcare workers and (C) the need to allow for expert reviewers to make comments and offer treatment options via the app other than just confirming whether the stained cervix is VIA-positive or negative. With these in mind, the following updates were made to improve the AVIVA App:
Creation of an in-App notification service for uploaded images. This service notifies the reviewer of a new case uploaded to the server by the case finders using SMS notification to reduce the lag time between screening and treatment.
Update of the form layout for the collection of patient-specific information to include a patient identifier (patient ID). This enables the tracking of patient history for any specific patient, from baseline to follow-up.
Update of the form layout for both case finders and reviewers to include treatment options for VIA-positive cases.
Phase 3 scale up: post-AVIVA
The phase 3 scale up saw the addition of 25 new sites in 2022 with 8 sites in Benue, 4 in Plateau, 4 in Oyo, 3 in Ondo, and 2 each in Osun, Ogun, and Ekiti. These facilities were selected from the three levels of care namely: primary, secondary and tertiary. At this stage, the main criterion was to ensure that each of the three senatorial districts in the seven supported states had at least one screening site for cervical cancer. A total of 22 sites uploaded cervical images of all patients screened for cervical cancer through the AVIVA app. During this phase, 4539 WLHIV were screened for cervical cancer and 144 were identified as VIA positive. A total of 192 images of the cervix from the 22 sites that used AVIVA app were uploaded and positive and negative concordance rates were 41.7% (5/12) and 95.6% (172/180), respectively.
As of June 2022, when phase 3 scale up ended, there were 86 sites offering CCS programmes for WLHIV and using the AVIVA App. A total of 29 262 WLHIV have been screened for cervical cancer and 1609 of them were VIA-positive, corresponding to VIA positivity rate of 5.5%. Online supplemental table 1 shows the list of HIV clinics where CCS and treatment programmes have been integrated.
Lessons learnt from phase 3
The addition of SMS notification for case finders when their uploaded images have been reviewed significantly reduced delay in attending to patients. Furthermore, the introduction of a new module for treatment documentation provided a learning repository for the case finders.
As part of the lessons from this scale up, APIN established a structure to support CCS programme implementation using the AVIVA app. Within the seven states, the team created three levels of supportive supervision:
A designated APIN programme staff was trained and assigned to be the cervical cancer programme focal person in each of the seven supported states. These focal persons interact with both the state reviewer (gynaecologist) and site-level healthcare workers. The programme focal persons had access to the contents of the AVIVA App and facilitated monitoring, nudged reviewers and facility staff, when necessary, to guarantee efficient flow of information between both parties. Additionally, they conducted technical assistance (virtual and on-site) to healthcare workers regarding cervical cancer management and screening.
The expert reviewers (nine gynaecologists across seven states) were involved in the training of healthcare workers and continue to provide ongoing mentoring and supportive supervision to the trained staff at the site level.
We developed a fine tuned process where a case finder takes three pictures of the cervix for each case, as follows: (A) plain cervix, (B) cervix stained with acetic acid and (C) cervix stained with Lugol’s iodine. These three pictures are sent to the expert reviewer who are trained to use the AVIVA app for review and feedback. A total of 1247 cases have been shared via the app, which corresponds to 3741 pictures, with 1058 of such cases undergoing expert review, corresponding to a reviewer rate of 84.8%. Figure 6A,B shows the current AVIVA App landing page and a sample consultation on the AVIVA App.
Overall, across the AVIVA phases of implementation, we observed a 16% increase in the positive concordance rate between primary case finder and expert reviewer (p=0.32), and 16% increase in negative concordance rate (p<0.01). This translates to a 17% reduction in programme wastages from false positivity and 15% in false negative misdiagnosis achieved through use of AVIVA technology.
Dissemination and sustainment
As part of our dissemination plan, our team has reached out to the agencies of the FMOH including the National AIDS, Viral Hepatitis, and Sexually Transmitted Infectious Control Programme (NASCP), and the National Cancer Control Programme (NCCP). We have also reported early results to funding agencies such as PEPFAR, and the Global Fund to show how AVIVA works and to get their buy-in. Working with the FMOH and other partners, we hope to deploy AVIVA as a national tool to support CCS.
While AVIVA is currently exclusively utilised in APIN-supported health facilities across seven states in Nigeria, we intend to expand its reach and adoption across Nigeria through the Nigeria Implementation Science Alliance (NISA) platform. NISA is a research collaboration of local PEPFAR-supported implementing partners, government health agencies and universities. NISA has established and maintained research infrastructure for implementation science in high volume health facilities and their surrounding communities across Nigeria’s six geopolitical regions.31 NISA’s research base is anchored at the Center for Translation and Implementation research at the University of Nigeria.
Patient and public involvement
Within 4 weeks of the expansion of CCS for WLHIV to secondary and primary levels of care, we discovered that of the 35 women initially classified as VIA-positive and eligible for thermal ablation, only 9 (25.7%) were later determined to be positive by a supporting trained gynaecologist. Our team convened a meeting with patients, case finders, gynaecologists, APIN Public Health technical experts and Ministry of Health representatives to discuss the high misdiagnosis rate and propose solutions. This meeting resulted in the creation of AVIVA. Various end-users (case finders) feedback has led to the revision of the AVIVA app, as highlighted under the implementation phase of the EPIS framework. Our team used various facility-level review meetings, meetings with NASCP, NCCP and funding agencies such as PEPFAR and the Global Fund to disseminate the improvement in diagnostic precision of acetic acid stained cervix using AVIVA and expert reviewers.
Data security and ethics
We protected patient privacy by anonymising sensitive personal information. Identifying patient details, such as names and other biographical data, are replaced with unique identifiers (Patient ID), minimising the risk of reidentification while still enabling accurate data analysis and programme review. AVIVA employs encryption techniques to protect data during transmission from the device to the central server. Transport Layer Security or Secure Sockets Layer protocols were implemented to establish a secure and encrypted connection. The app enforces strong user authentication mechanisms to prevent unauthorised access to the data. User credentials, such as usernames and passwords, are securely stored using industry-standard algorithms. Role-Based Access Control is implemented to manage user permissions and restrict access to specific functionalities and data within AVIVA. Different roles, such as case finders, reviewers and IP staff, are defined with varying levels of access privileges. The data stored on the central server is protected using various security measures. AVIVA and its central server undergo regular security audits to identify and address any vulnerabilities or potential risks. This includes monitoring software vulnerabilities, applying security patches and updates, and adhering to industry standards and best practices for data security.
Various studies have shown that the subjectivity of VIA reduces, and diagnostic precision improves, when screening is performed by higher cadres of healthcare workers with experience compared with when it is performed by lower cadre healthcare workers in low-resource settings.16 32–34 In this study, our team reports findings that demonstrated that training, mentorship support and use of optical imaging technology such as the AVIVA app, can lead to high-quality CCS services. Our implementation strategy used a second layer of review by gynaecological specialists for stained cervical images sent by healthcare workers using the AVIVA app and showed an increase in both the positive and negative VIA concordance rates.
HPV DNA testing, pap smears, colposcopy and biopsy are not widely available and are out of reach for poor and rural populations in resource limited settings such as ours.3–5 As such, VIA testing coupled with optical imaging and technical support has the potential to improve access to CCS and treatment services. We also used a collaborative approach during design, development and implementation of the AVIVA support programme that included patients, clinicians, researchers and policy-makers from the Ministry of Health which could inform policy changes to current CCS algorithm.
Our finding showed an increase of 16 percentage points from 26% to 42% positive concordance rate and from 80% to 96% negative concordance rate between the healthcare level case finder with VIA diagnosis and expert reviewers from baseline to the end of the study. This agrees with previous studies that reported improvement in the technical competencies of lower cadre staff in the diagnosis and treatment of health conditions following the use of technology and a series of onsite coaching and mentoring activities as part of the health system strengthening activities for CCS and management programmes.32 34
During implementation, our research team noted the need for SMS notification for case finders when their uploaded images have been reviewed. This strategy led to a reduction in patient waiting time which was a significant factor in delays early in the implementation process. As in previous studies introducing new technology, our use of AVIVA optical imaging technology combined with quality improvement cycles was critical to enhance CCS accuracy.35 36 The on-site coaching, mentoring and training during the implementation phase were critical to surmounting the challenges of low digital and smart-phone literacy among the healthcare workers. The large network of gynaecologists and case finders currently using the AVIVA app across 86 sites and generating over 4000 cervical images buttresses previous assertions that effective coaching and on-site mentoring ensure technological competence in healthcare settings when a new technology is introduced.36
Our team learnt several lessons from the implementation of this project including a potential policy implication and impact, as many institutions sought to use our imaging database images for training resident doctors, nurses and other healthcare workers who are undergoing specialised training in oncology. However, there are several limitations to our study. First, we did not measure the VIA-negative concordance rates prior to pre-AVIVA phase 1. As such, the expert gynaecologists reviewed only the cases diagnosed as positive on-site by the case finders and no baseline was available to compare VIA-negative concordance prior to introduction of AVIVA. Second, although the use of trained gynaecologists as expert reviewers has been found to be comparable to the use of site-level colposcopy,32–34 diagnosis by case finders and reviewers was not compared with the gold standard of tissue biopsy. Third, we had only one reviewer per case images and as such could not determine the level of agreement between reviewers which could improve the validation of diagnosis. Fourth, while our pilot was conducted in health facilities supported by PEPFAR, it is unclear how this programme will perform in facilities without PEPFAR support. However, we are working with the government to integrate AVIVA as a platform for training of gynaecologists in academic institutions. Finally, because we used the EPIS framework for reporting and not throughout the study planning, there was limited consideration given to sustainability. This oversight in programme design denied us the opportunity to explore clients’ acceptability of technology and their perceptions about taking visual images of the cervix for improved service delivery. Notwithstanding the EPIS framework helped us break down and explore the complex, multilayered and highly interactive nature of the project implementation, which might have been lost with other implementation frameworks.
The use of technology applications such as AVIVA can reduce subjectivity associated with VIA and improve diagnostic precision of VIA. A well-designed clinical trial will be needed to validate our pilot findings. A platform that improves diagnostic precision will enhance access to CCS in resource-limited settings where HPV DNA, pap smear and tissue biopsy are not easily accessible.
Data availability statement
Data are available on reasonable request.
Patient consent for publication
All the 86 health facilities—administrators, clinicians, nurses, case finders, women living with HIV and our AVIVA expert reviewers.
Handling editor Seye Abimbola
Contributors All the authors listed contributed in the following ways to the manuscript development: Conceptualisation: OOO. Technical development: OOO, OA, EO, CI and YTA. Implementation: YTA, OOO, CI, EO, PO and OA. Data collection: EO and CI. Data analysis: CI, EO, OOO, JOO and EEE. Manuscript development: OOO, OA, YTA, EO, CI, JOO and EEE. Expert reviews: JOO, CC and EEE.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.