An Implementation Evaluation of the Smartphone-Enhanced Visual Inspection with Acetic Acid (SEVIA) Program for Cervical Cancer Prevention in Urban and Rural Tanzania

Introduction: The World Health Organization (WHO) recommends visual inspection with acetic acid (VIA) for cervical cancer screening (CCS) in lower-resource settings; however, quality varies widely, and it is difficult to maintain a well-trained cadre of providers. The Smartphone-Enhanced Visual Inspection with Acetic acid (SEVIA) program was designed to offer secure sharing of cervical images and real-time supportive supervision to health care workers, in order to improve the quality and accuracy of visual assessment of the cervix for treatment. The purpose of this evaluation was to document early learnings from patients, providers, and higher-level program stakeholders, on barriers and enablers to program implementation. Methods: From 9 September to 8 December 2016, observational activities and open-ended interviews were conducted with image reviewers (n = 5), providers (n = 17), community mobilizers (n = 14), patients (n = 21), supervisors (n = 4) and implementation partners (n = 5) involved with SEVIA. Sixty-six interviews were conducted at 14 facilities, in all five of the program regions Results SEVIA was found to be a highly regarded tool for the enhancement of CCS services in Northern Tanzania. Acceptability, adoption, appropriateness, feasibility, and coverage of the intervention were highly recognized. It appeared to be an effective means of improving good clinical practice among providers and fit seamlessly into existing roles and processes. Barriers to implementation included network connectivity issues, and community misconceptions and the adoption of CCS more generally. Conclusions: SEVIA is a practical and feasible mobile health intervention and tool that is easily integrated into the National CCS program to enhance the quality of care.


Introduction
Sub-Saharan Africa has a cervical cancer incidence rate of 50.9 cases per 100,000 women, the highest incidence rate in the world [1].The burden of disease is disproportionately high in low-and middle-income countries (LMIC), which account for 85% of cases and nearly 90% of cervical cancer deaths worldwide [2].In Tanzania, cervical cancer is the most common cause of death from cancer for women [3].Sporadic uptake cervical cancer screening (CCS), prevalence of high-risk oncogenic human papillomavirus (HPV) subtypes, and relatively high rates of human immunodeficiency virus (HIV) co-infection compound risks of infection [4].Survival rates are low, with more than half of women diagnosed in Tanzania dying of the disease, as they receive their diagnosis at an advanced stage when curable treatment options are limited [5].In 2015, 82% of women in Kilimanjaro, Tanzania reported they had knowledge of cervical cancer, although only 6% had ever been screened [6].
It is well documented that in order to avoid progression to later stages of disease, effective prevention strategies must be employed widely and routinely.Screening and treatment for pre-cancer of the cervix is a secondary prevention strategy used globally in many low-resource health care settings to prevent cervical cancer.In these contexts, the main method of secondary prevention for cervical cancer is visual inspection of the cervix with the naked eye after the application of 3% to 5% acetic acid solution.Visual inspection with acetic acid (VIA) is regarded as the best approach in most low and some middleincome income countries and has been endorsed by the World Health Organization (WHO) when combined with ablative 'see and treat' approaches at the time that a woman receives VIA [7,8].VIA can be performed by nurses and other skilled health practitioners and is inexpensive and non-invasive.It can be practiced broadly in lower-level health facilities as well as in HIV care and treatment programs, as part of regular screening services for HIVrelated malignancies.More importantly, VIA provides instant results, and those eligible for treatment can receive treatment with ablative methods (thermal ablation or cryotherapy) on the very same day in the same health facility.This "see and treat" method promotes adherence to treatment, as it is offered immediately after diagnosis, thus minimizing the likelihood of loss to follow-up associated with patient referrals for treatment at an alternate facility or higher level of care [9].Since 2020, the WHO has recommended HPV DNA testing as a primary screening method in preference to cytology and visual inspection with acetic acid (VIA).The absence of carcinogenic HPV types indicates an extremely low immediate risk of precancer/cancer and a reassuring low risk of cervical cancer for longer subsequent periods [10].The superior sensitivity of HPV testing as compared to cytology and VIA and the reassurance against cervical cancer following a negative HPV test result is increasingly leading to its adoption as the main primary screening method in many countries worldwide [10].VIA still plays a role as a triage step among HPV-positive women, but efforts to improve its performance are needed.In fact, as the interpretation of VIA is highly subjective, quality control has proven to be highly variable, impacting profoundly on both sensitivity and specificity [11][12][13].A triage strategy is important to reduce the number of women referred to colposcopic biopsy in search of precancer or treatment.In many settings, the HPV prevalence is too high and the healthcare capacity too low to refer all HPV-positive women to colposcopic biopsy or to treat all.
Informed by the 2020 WHO cervical cancer prevention guidelines, Tanzania's national cervical cancer prevention program (CECAP) is piloting the use of HPV DNA testing as a primary screening tool across four regional sites in 2024.Additional research programs [14] have been implemented in several sites to evaluate effective strategies for larger-scale implementation of HPV DNA testing through self-sampling in health facilities and in rural and urban communities.At present, and while further research to inform the implementation and scale-up of HPV DNA testing as a primary screening strategy is completed, the CECAP program continues to deliver screening services that are underpinned by the previous WHO guidelines with the use of VIA and the "see and treat" method and utilizes nurses and other clinicians providing frontline, non-specialist care.The program consists of six days of competency-based training, and continued mentorship by experienced senior CCS trainers to complete screening with VIA in the field under supervision, as well as routine follow-ups.Due to resource constraints, available training staff, and geographic logistics, follow-up training and supportive supervision opportunities do not always occur, and maintaining competency and quality of screening among CCS providers in the CECAP program has been a notable challenge.There has been a significant decline in skill retention and VIA quality, as a result of the lack of post-training mentorship and continuous technical support for CCS providers, combined with the frequent turnover of trained staff [15,16].
Digital cervicography has been known to improve the quality of VIA [17,18].It uses a digital camera to transmit an image of the cervix to a screen/monitor, using higher clarity and resolution to review the image, than is possible with the naked eye [17,18].Images are transmitted to experts at coordinating sites for review [19][20][21].However, the program requires particular infrastructure and equipment, making it less feasible in many low-resource settings, including Tanzania.During the study period, a digital cervicography program was piloted at a single site in northern Tanzania but did not expand to other locations due to resource constraints.However, the growing prevalence of smartphones and dependable mobile networks across Tanzania presented a hopeful resolution to the challenges encountered during the program's attempted scaling.Program developers believed that smartphone cameras could provide a feasible avenue to enhance supervision, oversight, and measurable quality assurance, especially in remote areas [15].
The Smartphone-Enhanced Visual Inspection with Acetic acid (SEVIA) program included the dissemination of smartphones to CCS providers, together with a smartphone application permitting real-time, secure sharing of de-identified VIA cervical images and relevant clinical information by health providers to expert 'reviewers'.Image reviewers are senior VIA providers (i.e., gynecologists, and other skilled physician/non-physician VIA trainers within Tanzania's CECAP program who review images in real-time, providing supervision and mentorship to nurses and non-physician clinicians [15].When there is discordance between the health provider and the reviewer, the application sends a notification, allowing the reviewer to provide secure in-app feedback and recommendations on the diagnosis and treatment plan.This is meant to ensure the most accurate diagnosis and treatment plan for the client, while also promoting continued learning and high-quality supervision for the provider [15].
The program efficacy of the SEVIA concept was tested in a pilot study in the Kilimanjaro region from June 2014-March 2015 [16] and transitioned to scale in a pre-post study to evaluate the effectiveness of the intervention at existing CECAP sites using VIA.The program was delivered intensively for 6 months at 24 health facilities, followed by a 6-month maintenance phase.This comprehensive implementation evaluation was completed to evaluate client and health provider-related implementation outcomes, and reports on experiences, perspectives, and general acceptance of the program by both groups.

Methods
The study was conducted from September to December 2016 in 5 regions of Northern Tanzania (Kilimanjaro, Arusha, Kigoma, Bukoba, and Tanga).Measures included a thorough evaluation completed during the implementation phase, in order to understand barriers and enablers of the intervention, and early learnings from program consumers, providers, and higher-level stakeholders.The primary author, holding an MPH, trained in qualitative research epidemiology, and possessing Swahili language skills, completed observational activities and open-ended interviews with image reviewers (n= 5), providers (n = 17), community mobilizers (n = 14), patients (n = 21), supervisors (n = 4) and implementation partners (n = 5) involved with the SEVIA program.Facilities were selected for assessment based on proportional calculations of regional representation in the program, and by inclusion of rural/urban and higher/lower performing facilities.Image reviewers, providers, supervisors, and implementation partners were selected by consecutive sampling, and contacted in advance by telephone.Patients and community mobilizers were sampled by convenience sampling on-site.A total of 66 interviews were conducted with program informants at 14 facilities, in all 5 of the program regions.Additional interviews conducted at program sites run by an international nongovernmental organization were excluded from this analysis, as these clinics functioned on a private for-profit model, and the patient population was of a higher socioeconomic status.
The Reach, Efficacy, Adoption, Implementation, and Maintenance (RE-AIM) Framework [22,23] and the Consolidated Framework for Implementation Research (CFIR) [24,25] are implementation science frameworks that were developed to guide systematic assessment of multilevel implementation contexts and assist in translating research to evidencebased practice.An implementation evaluation framework was designed by adopting similar implementation outcome variables [26].Implementation outcome measures included acceptability, adoption, appropriateness, feasibility, fidelity, implementation costs, coverage, and sustainability.Data collection was performed iteratively, with continual refinement of the protocol.A relationship was established prior to study commencement.Participants were aware of the research team's goals in performing the research.There was no presence of non-participants.
One female researcher conducted, transcribed, and coded all participant comments and interviews, allowing for data immersion and obtaining an overall sense of the data.Interviews were conducted in-person at the sites, with a few implementation partner interviews being conducted over the phone.Most interviews lasted 20-30 min.Repeat interviews were not carried out.The questions, prompts, and guides were provided by the study authors.Field notes were made during and after the interviews were conducted.Transcripts were not returned to participants for comments and/or corrections.Using the outcome variable framework [26], content inductive analysis was used for each variable [27].An open coding approach was adopted, forming a general description of the research topic by generating categories and subcategories as they emerged [27].This systematic approach was appropriate for open-ended interviews to determine trends and patterns.Discussion with a second researcher familiar with the data confirmed emerging categories.Lastly, to minimize confirmation bias, two external female researchers, trained in qualitative research methods, who were unfamiliar with the data and preliminary findings, independently coded all participant interviews inductively with QSR-NVivo [28].Secondary researchers undertook a thematic analysis, established consensus on emerging themes, and resolved conflicts to determine final results once data saturation was met.Findings were confirmed with the preliminary researcher, and all three collaborated to determine the final themes presented.Participants did not provide feedback on the findings.

Results
A full extrapolation of results by implementation outcome is provided in Table 1.Overall, our findings indicated that the intervention was trusted by both patients and providers, was easily implemented within regular/routine practice and care and assisted in both the quality of diagnosis/care offered to patients, and in the oversight and ongoing training of providers.Women appreciated seeing an image of their cervix as it provided an immediate understanding of their health and body, and providers expressed increased confidence in the care they were able to provide.Community-level knowledge or trust in CCS in general was seen as a barrier impacting the overall acceptability of the program's desired outcome.The most common misconceptions of screening were that the procedure would be intrusive and painful/uncomfortable, and that the reproductive parts would be removed for examination.Women additionally expressed a reluctance to screen for fear of receiving a positive result or prognosis.The SEVIA application was seen to be quickly understood by providers, but issues related to mobile network coverage impacted the speed and regularity with which providers could share images with reviewers in real-time.While the intervention appeared to reach both women of middle and lower socioeconomic statuses, it was difficult to determine the true reach of the program, as interviews were generally conducted with respondents who had already reached the centre.Community mobilizers did provide some useful information in regard to coverage/reach of the program, but their position was inherently biased as they were tasked with mobilizing within a particular community.The extent to which more rural communities not encompassed within the study had an understanding of screening or cervical cancer in general could not be established.Motivations for participation were unclear, as providers and image reviewers were provided small per diems to see patients and review images.In the absence of compensatory incentives in the long term, program sustainability could not be sufficiently determined.

Implementation Outcome Working Definition Related Terms Results
Reasons included: (1) taking a picture of the cervix with the smartphone allowed them to visualize it better and make a more accurate assessment which enhanced confidence in their role as a provider; (

Implementation Outcome Working Definition Related Terms Results
For example, one patient expressed there was a need for "more education to women in the villages: in the interior.
They don't get information, so it is not easy to convince them to come [for screening]".

Sustainability
The

Discussion
It is critical to acknowledge the amount of time that has passed since data collection.These data on cervical cancer were collected between 2014-2016; however, it retains significant relevance in 2024, offering critical insights into the disease's impact and the efficacy of interventions.Despite efforts to mitigate challenges such as screening and treatment deficiencies, cultural barriers, and limitations in healthcare infrastructure, cervical cancer incidence and mortality rates have exhibited minimal change, highlighting its persistent public health importance in Tanzania.This study enables the evaluation of preventive strategies and identification of current gaps, and emphasizes the ongoing imperative for sustained global and local efforts to address disparities in healthcare access and outcomes.Furthermore, these data serve as a pivotal reference for ongoing initiatives aimed at reducing the burden of cervical cancer and enhancing women's health in Tanzania and globally.
A number of key findings were highlighted across all settings and by numerous respondent types.A major challenge was network connectivity issues and interruptions especially in rural health facility settings, which was discussed by providers, reviewers, and other program stakeholders.While Tanzania has seen significant improvements in the speed and reliability of mobile telephone networks in recent years, coverage issues persist in many, primarily rural areas.Certain cellular providers have strong networks in particular geographic regions but not in others.As the program spans the country, program implementers were challenged to select a carrier to provide reliable coverage at all sites, and some sites struggled more than others.A more comprehensive coverage plan with multi-providers may be required to ensure every facility is equipped with a phone with sufficient network coverage.This would dramatically decrease the necessity of providers saving images, reduce any delays of image reviewers responding to providers, and allow the intervention to function in real-time as intended.
An overwhelming majority of providers and reviewers reported that SEVIA integrated seamlessly into their general practice of screening, giving it a high feasibility and probability of long-term integration in standard practice.However, it was unclear whether study participants would be equally motivated to continue using SEVIA once study per diems ceased.A significant portion of the clients who were screened during the intervention phase of the study were acquired at outreach campaigns (small mobilization efforts conducted at facilities in rural villages) where providers were given per diems.Campaigns were funded by the research study in order for newly trained providers to reach their minimum number of screens for certification, as well as to increase the number of study participants observed in a short period of time.It was later observed that a consistently low number of women reported to facilities for screening after the intervention phase of the program.
The intervention was also found to increase the knowledge and skills of providers with limited training.SEVIA permitted providers to better visualize the cervix, and consequently increased confidence in their diagnosis and role as screeners in general.These findings are supported by other mHealth interventions in resource-limited countries which demonstrate the value of mobile phones in tackling barriers to service provision and improving both the range and quality of services offered by community-level health providers [29][30][31].For example, in a qualitative study evaluating the acceptability and usability of a mobile phone-based ophthalmic testing system to perform comprehensive eye examinations in Nakuru, Kenya, healthcare providers reported that the tool aided them in detection and diagnosis, provided decision support, improved communication among providers, and assisted in education and training [31,32].
Improved health education and enhanced health literacy are essential for SEVIA or similar mobile health programs and/or applications supporting cervical cancer screening services to be broadly accepted and sustainably implemented at the community level, especially in rural and remote communities.While there appeared to be widespread acceptability for the use of the smartphone to capture images of the cervix by women who had already agreed to screen, there appeared to be low acceptability at the community level of cervical cancer screening in general.For example, a number of patients and providers commented there was a misconception that the reproductive parts would be removed for examination during the procedure.Others expressed fears that screening would be painful and uncomfortable, and feared receiving a positive result (especially among HIV+ women).As CCS is paired with HIV testing in Tanzania, this may especially deter women most at risk of receiving a positive HIV diagnosis.Additionally, results indicated that the majority of women were not aware that early-stage detection and treatment by ablative methods could be performed on-site.These beliefs are aligned with other studies reporting barriers to cervical cancer screening uptake in the region.For example, a qualitative crosssectional study in Lilongwe, Malawi, found barriers to CCS with VIA uptake to include misconceptions of screening procedures and fatalistic views on cancer in general [33].In this study, most participants reported that prior to undergoing cervical cancer screening they had limited understanding of the process.Myths and misconceptions of the screening process included expectations that the exam would be painful, fear of receiving a positive screening result, distrust in healthcare workers and suspicion of specimen collection and removal of the uterus [33].In another cross-sectional study assessing factors associated with cervical screening uptake among HIV-infected women at Mildmay, Uganda, where CCS with VIA was integrated into HIV care, respondents reported similar misconceptions related to screening, such as removal of their ovaries and/or uterus and "cutting off of flesh" [34].While our study did not specifically seek to ascertain the acceptability of CCS in the general sense, these findings have implications for the potential impact SEVIA on target populations in the future.The distinct lack of knowledge about cervical cancer at the community level and primarily at rural community sites where our evaluation took place signals the need for more widespread education on cervical cancer, screening, and treatment.Our findings, in addition to others [11] demonstrate that much of the fear, mistrust, and misconceptions can be alleviated with targeted health education.Specifically, increasing awareness about the importance of regular screening, dispelling myths and misconceptions about cervical cancer and the HPV vaccine, and providing culturally sensitive information will reduce fear and mistrust.Results suggest that in order for SEVIA to reverse the trajectory of cancer diagnoses in the country, efforts must also be placed on community education, increasing health literacy, and general promotion of CCS.Educational initiatives should be delivered through community outreach, integration into school curricula, and leveraging media campaigns to ensure widespread dissemination.
Patients generally expressed positive experiences with SEVIA and appreciated the addition of smartphone technology to screening.The data suggest SEVIA may even serve to empower patients with respect to health education and individual/personal health literacy.Being able to directly visualize one's own cervix and any associated lesions provides immediate reassurance and information about one's health, thereby providing an opportunity for improved individual health literacy and understanding of self.Increased health literacy has been shown to improve one's knowledge and self-care behaviours among individuals with various health conditions across socioeconomic and cultural settings [35][36][37].Desire to know one's health status was reported as a key outcome in the same study of VIA clients in Lilongwe noted above, "I did that [screening] because I wanted to know the condition of my body, you can just be staying and never be certain you are okay or not.So, this time I thought it wise to go get screened" [33].
As Tanzania moves towards more widespread implementation of HPV DNA testing as a primary screening strategy that can provide more broad population coverage (e.g., women can perform a vaginal self-swab in a rural community with the guidance of a trained community health worker and results can be quickly delivered back to a local screening nurse to communicate with the woman about the test result and next steps for follow-up screening and directing the woman to a local screening site for visual assessment for treatment if they are HPV DNA positive.The role of a mobile health platform such as SEVIA has the potential to support the rapid roll out and scale-up of new strategies that utilize HPV self-sampling.
As this study employed a cross-sectional design, temporality cannot be inferred.Additionally, given the challenges of sampling hard-to-reach populations, the present results from patients could underrepresent more marginalized women.All of the patient testimonies were obtained from women who reached health facilities in the program and agreed to be screened, thus a great deal of information is missing on the challenges of increasing uptake of CCS more generally.Finally, the variables in this analysis by all respondent types were self-reported and thus may be subject to social desirability bias.

Conclusions
Findings from this semi-structured qualitative implementation evaluation indicated that the SEVIA program was a highly regarded tool for the enhancement of CCS services in Northern Tanzania.Acceptability, adoption, appropriateness, and feasibility of the intervention were highly recognized.It proved an effective means of improving good clinical practice among providers and fit seamlessly into existing roles and processes.While technical restraints caused adaptations to the intervention protocol, these are to be expected in the preliminary development of a technology.Alterations to the app were ongoing at the time of the study, and the usability of the tool was increasing.Network connectivity issues were a persistent challenge to program adherence in a number of locations and will need to be overcome for the program to be effective in the future.Allocation of permanent funds for outreach activities and more comprehensive community education and mobilization approaches are recommended, in order to increase the regularity of screening in general and access the most vulnerable women.In 2018, as a consequence of SEVIA supporting the development of good clinical practice among CCS providers through oversight and continuous training, Tanzania's CECAP program integrated the SEVIA model into their Cervical Cancer Prevention Strategic Plan for 2020 to 2024 [37].Funding restrictions and resource shortages have remained intermittent challenges, as has the COVID-19 pandemic, with health system resources re-directed to public health.At the time of writing, the SEVIA program was undergoing further scale-up and evaluation of a new version of the SEVIA mobile app that includes integration of HPV DNA test results and additional tracking functions and follow-up indicators to reduce loss to follow-up and support navigation of women to improved linkage to follow-up screening services.

Table 1 .
Results by Implementation Outcome Variable.