Feasibility of Training Community Health Workers in the Detection of Oral Cancer

Key Points Question Is it feasible for community health workers to screen for oral cancer using a mobile application capturing system? Findings In this cross-sectional study, 1200 participants were screened and the findings were compared with those of a subgroup who were also screened by dentists. There was near-perfect agreement between the findings of the community health workers and dentists in identifying positive or negative cases of oral lesions. Meaning The findings of this study suggest that it is feasible to train community health workers to perform oral cancer screening in resource-limited settings.

The lower-cost, scalable option of community health worker-led oral cancer screening may be an important alternative. 7 For example, accredited social health activists in India may be able to undertake these services; there is 1 accredited social health activist per 1000 residents who currently undertakes a variety of tasks including screening for noncommunicable diseases. 8 Furthermore, community health workers contributed substantially to a policy program (Ayushman Bharat) that includes the National Programme for Prevention and Control of Cancer, Diabetes, Cardiovascular Diseases, and Stroke that is being gradually implemented nationally in India. 9 Community health workers may be crucial in facing the growing challenge of health care access and delivery, but their role in low-and middle-income countries is less well studied. 10 Earlier studies in which community health workers were provided a mobile phone with an application for oral cancer screening support the concept of integrating health and technology to improve public health surveillance. 7 The electronic data captured have the advantage of maintenance of medical records, patient monitoring, and cloud-based storage service. 11 However, there are few studies assessing the utility of community health workers in early oral cancer detection through a technology-based platform with follow-up and smoking counseling. 7 Hence, the present study was undertaken to assess the feasibility of community health workers in screening and early detection of oral cancer using a mobile application capturing system.

Study Design and Setting
We conducted a cross-sectional analytical pilot study with a sample size of 1200 participants from 10 areas of Gautam Budhnagar district, Uttar Pradesh, India. Uttar Pradesh reported a tobacco use prevalence of 35.5% as per the Global Adult Tobacco Survey in 2015-2016, which is higher than the national average of 28.6% in India. 12 Gautam Budhnagar district was chosen from Uttar Pradesh as the study area owing to its close proximity to our organization (Indian Council of Medical Research, National Institute of Cancer Prevention and Research, referred to as HPC from here on) for accessing the screening services by the participants with ease when referred. The study was performed from January 31, 2020, to March 31, 2021, and screening was conducted in the same period. Ethics approval for the study was obtained from the University of Warwick, Coventry, UK, and the Indian Council of Medical Research in India. Written and verbal informed consent were obtained in the local language (Hindi); participants did not receive financial compensation. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Training of Community Health Workers
A 1-day hands-on training workshop was conducted at the HPC for the 10 selected community health workers to (1) systematically enumerate households and individuals using a household form; (2) A kit was provided to the community health workers containing tongue depressors, disposable masks, disposable gloves, flashlight, red medical waste bags, hand sanitizers, mobile phone for data entry, an easy-to-use reference manual for identification of oral lesions, and consent forms and referral cards in the local language. At the end of the training session, written tests were conducted to evaluate the level of understanding in identifying normal and abnormal lesions, use of the mobile application, and the knowledge gained by the community health workers. The top 5 ranked community health workers were first chosen to conduct screening and the remaining 5 community health workers received further training before deployment in the field. They were all paid for undertaking this extra screening activity.

Study Population and Approach
All homes in the 10 areas were eligible for the screening program. As per the national guidelines for population cancer screening, all men and women aged 30 years or older, as well as tobacco users younger than 30 years, were eligible for screening. 13 At the start of the program, a health awareness event was organized to encourage uptake. Then, the community health workers identified eligible individuals with the information provided in the household listing forms. Screening was done primarily by community health workers, with assistance from the dentists and a nurse during the initial days of the study. Specifically, the community health workers screened individuals using a mobile phone-based questionnaire (eAppendix 1 in the Supplement). During the first 3 months of the study, an average of 10 to 12 participants were screened per day and screening was stopped due to the COVID-19 lockdown in India from March 25 to May 31, 2020. Screening was gradually restarted in June 2020, with 5 to 6 participants screened per day with due consideration of adherence to COVID-19 guidelines. Capturing and uploading the photographs including normal mucosa and abnormal lesions in the oral cavity of the participants were undertaken using the oral cancer screening mobile application in the provided android phone (Redmi Note 9; Xiaomi Inc). The images were captured in autofocus and auto flash modes. If the community health worker noted the presence of any oral potentially malignant disorders (eg, white, red, or mixed lesions; ulcers; and growth masses), apparently normal mucosa, and apparently normal mucosa, they informed the participant. Yellow referral cards were provided to the participants with normal findings. Red cards were provided to participants with suspected oral lesions, and they were encouraged to visit the HPC for further examination by the dentists and recording of appropriate details in the online clinic (eAppendix 2 in the Supplement). The participants were provided with the appropriate diagnosis, and the photographs of the oral cavity were taken with a high-resolution digital single-lens reflex camera (Nikon D5600; Nikon Inc) by the dentists. Punch biopsies were performed on suspicious lesions by trained dentists and reported as per World Health Organization guidelines (ie, hyperkeratosis; mild, moderate, or severe dysplasia; carcinoma in situ; and squamous cell carcinoma). 13 Individuals diagnosed with squamous cell carcinoma were referred to a tertiary care center for further investigation and appropriate management.
All participants were informed of the ill effects of tobacco and the importance of routine oral self-examination. They were advised to undergo screening every 5 years per the national guidelines or whenever they noticed any abnormality in their oral cavity. 13 In addition, smokers were referred to the tobacco cessation clinic. Individuals who were tobacco users and had oral potentially malignant disorders underwent tobacco counseling in the tobacco cessation clinic. They were offered 4 to 8 counseling sessions lasting 30 to 60 minutes. Patients with potentially malignant oral disorders were followed up for up to 3 months depending on the size and site of the lesion, smoking cessation, presence of underlying disease, and change in the size of the lesion. Reasons for nonparticipation were also noted.

Statistical Analysis
The data collected in the oral cancer screening mobile application were exported to a Microsoft Excel 2018 (Microsoft Corp) spreadsheet and analyzed using Stata, version 14 (StataCorp LLC). Sensitivity, 95% CI, 33.5%-36.5%) was consumed more commonly than smoked tobacco (35%; 95% CI, 32.9-37.0) products in both sexes, with a high predilection of both tobacco and alcohol use among men.

Reliability in Identification by Community Health Workers
The reproducibility of community health workers' examinations and their agreement with the dentists' outcomes was measured using κ statistics. 14 The sensitivity and specificity of the oral visual examination by the community health workers were calculated by pooling all the results as well as by stratification according to participant sex and age.
The interrater agreement of the oral visual examination findings among the 3 dentists at our organization was done in 100 participants and showed near-perfect agreement between them as indicated by the overall κ value of 0.9 (P < .001). There also was near-perfect agreement between the findings of the 3 dentists as indicated by the overall κ value of 0.9 (P < .001).   A study done in Ernakulam district, Kerala, India, to assess the use of community health workers in performing the oral visual inspection showed a lower sensitivity of 56% and comparable specificity of 98% compared with our study. 18 In an earlier Sri Lankan study, 29 215 (33.5%) of the 87 277 eligible participants aged 20 years and older were examined by community health workers. 19  had either benign or no lesions. Thus, the false-positive referrals in the reported studies of oral cancer screening varied from 20% to 55%. 22 Community health workers have also played roles in previous studies to address the feasibility of using health care auxiliary personnel in the control of oral cancer in India and Sri Lanka. [18][19][20] These studies suggest that trained community health workers can undertake a systematic mouth examination and identify lesions.
Although oral cancer and oral visual examination meet most of the criteria of Wilson and Jungner 23 for a suitable disease and screening test, respectively, there are few randomized clinical trials to support the usefulness of screening for oral cancer as being effective in reducing the incidence and mortality associated with this disease. 24  analysis of clustering, blinding of outcome, and loss to follow-up. 15,16,24 In our study, the agreement between the results of oral examination by community health workers and the dentists was notable because, as the number of participants recruited in the study increased, the community health workers accumulated further experience in identifying oral mucosal lesions. Also, the retraining provided by the dentists along with additional monetary incentives improved the performance of the community health workers in screening and referrals of participants by them to our organization.
Our study builds on the work of Braun et al 25  subsite of the oral cavity, missing imaging of the lesion, and poor-quality images. These shortcomings occurred during the initial part of the study; with the passage of time and increasing experience in handling the devices, these issues were resolved.

Strengths and Limitations
Strengths of our study suggest that community health workers can undertake oral screening. We were able to follow up participants with potentially malignant oral disorders and, in 22 participants, there was a decrease in the size of the lesion and, through tobacco cessation counseling, 26 participants stopped this habit. Furthermore, electronic data collection using android mobile phones with data transfer when connected to the internet increased data accessibility for researchers. The high-quality oral images of the participants screened by dentists may serve as a repository for further researchers. None of the community health workers dropped out of the oral cancer screening in the community program throughout the duration of our study.
The study also has limitations. The intraobserver agreement among 10 community health workers in identifying normal and abnormal oral lesions was not calculated and the study was not powered to detect differences. In addition, participants were selected from a single geographic location, so the findings cannot be generalized.
The 9 potentially malignant oral disorders and 2 recurrent aphthous ulcers missed by the community health workers were present in the lower vestibular mucosa, buccal mucosa, and lower lip. This lack of identification indicates that these areas in the oral cavity may have been inaccessible owing to improper retraction.

Conclusions
The findings of this study suggest feasibility of training community health workers to perform oral cancer screening in a low-or middle-income country. However, a randomized clinical trial is needed to determine the effectiveness and cost-effectiveness of this approach.