|Year : 2019 | Volume
| Issue : 2 | Page : 107-113
Role of community health worker in a mobile health program for early detection of oral cancer
N Praveen Birur1, Keerthi Gurushanth2, Sanjana Patrick3, Sumsum P Sunny4, Shubhasini A Raghavan2, Shubha Gurudath2, Usha Hegde5, Vidya Tiwari6, Vipin Jain7, Mohammed Imran3, Pratima Rao3, Moni Abraham Kuriakose8
1 KLE Society's Institute of Dental Sciences; Oral Cancer Screening Program, Biocon Foundation, Bengaluru, India
2 KLE Society's Institute of Dental Sciences, Bengaluru, India
3 Oral Cancer Screening Program, Biocon Foundation, Bengaluru, India
4 Head and Neck Oncology, Mazumdar Shaw Cancer Center, Bengaluru, India
5 JSS Dental College, Mysore, Karnataka, India
6 Oral Medicine Specialist, KLE Society's Institute of Dental Sciences, Bengaluru, India
7 Department of Public Health Dentistry, KLE Society's Institute of Dental Sciences, Bengaluru, India
8 Cochin Cancer Research Center, Kerala, India
|Date of Web Publication||2-May-2019|
N Praveen Birur
KLE Society's Institute of Dental Sciences; Oral Cancer Screening Program, Biocon Foundation, Bengaluru
Source of Support: None, Conflict of Interest: None
Background: The global incidence of oral cancer occurs in low-resource settings. Community-based oral screening is a strategic step toward downstaging oral cancer by early diagnosis. The mobile health (mHealth) program is a technology-based platform, steered with the aim to assess the use of mHealth by community health workers (CHWs) in the identification of oral mucosal lesions.
MATERIALS AND METHODS: mHealth is a mobile phone-based oral cancer-screening program in a workplace setting. The participants were screened by two CHWs, followed by an assessment by an oral medicine specialist. A mobile phone-based questionnaire that included the risk assessment was distributed among participants. On specialist recommendation an oral surgeon performed biopsy on participants. The diagnosis by onsite specialist that was confirmed by histopathology was considered as gold standard. All individuals received the standard treatment protocol. A remote oral medicine specialist reviewed the uploaded data in Open Medical Record System. Sensitivity, specificity, positive and negative predictive values were calculated. Inter-rater agreement was analyzed with Cohen's kappa coefficient (κ) test, and the diagnostic ability of CHWs, onsite specialist, and remote specialist was illustrated using receiver operating characteristic curve.
RESULTS: CHWs identified oral lesions in 405 (11.8%) individuals; the onsite specialist identified oral lesions in 394 (11.4%) individuals; and the remote specialist diagnosed oral lesions in 444 (13%). The inter-rater agreement between the CHW and the onsite specialist showed almost perfect agreement with the κ score of 0.92, and a substantial agreement between CHW and remote specialist showed a score of 0.62. The sensitivity, specificity, positive and negative predictive values of CHWs in the identification of oral lesion were 84.7, 97.6, 84.8, and 97.7%, respectively.
CONCLUSION: The trained CHWs can aid in identifying oral potentially malignant disorders and they can be utilized in oral cancer-screening program mHealth effectively.
Keywords: Early detection, oral cancer screening, oral potentially malignant disorders, oral medicine specialist, telemedicine
|How to cite this article:|
Birur N P, Gurushanth K, Patrick S, Sunny SP, Raghavan SA, Gurudath S, Hegde U, Tiwari V, Jain V, Imran M, Rao P, Kuriakose MA. Role of community health worker in a mobile health program for early detection of oral cancer. Indian J Cancer 2019;56:107-13
|How to cite this URL:|
Birur N P, Gurushanth K, Patrick S, Sunny SP, Raghavan SA, Gurudath S, Hegde U, Tiwari V, Jain V, Imran M, Rao P, Kuriakose MA. Role of community health worker in a mobile health program for early detection of oral cancer. Indian J Cancer [serial online] 2019 [cited 2020 May 26];56:107-13. Available from: http://www.indianjcancer.com/text.asp?2019/56/2/107/257547
| » Introduction|| |
The global burden of oral cancer accounts for 200,000 deaths annually, encompassing 80% of occurrence in low- and middle-income countries. Globally, India accounts for the highest number of oral cancer cases, reporting 83,000 new cases and 46,000 deaths annually.,, Oral cancer is largely preventable, and the long preclinical course makes it amenable to screening. Late presentation of the disease has resulted in poor treatment outcome with increased cost. Strategic communication with the population is vital in downstaging oral cancer., An approach that is out of clinical base focusing on early detection and surveillance of oral lesions will play a critical role in the oral cancer control program.
Community-based oral cancer screening has proved to be one of the most effective methods to reduce the high disease burden. In this prospect, community health workers (CHWs) can be effectively deployed for large-scale screening in high-risk individuals, as they bridge the health system with the community effectively. Several studies have supported the concept of using trained CHWs in early detection and prevention of oral cancer and oral potentially malignant disorders (OPMDs).,,,
Owing to limitations in conventional screening, we optimized the existing module by integrating health and technology to improve public health surveillance. In earlier studies, CHWs were empowered with a mobile phone with a decision-based algorithm in the oral cancer-screening program., The mobile phone-based approach is an evidence base because of electronic data captured and used, which has an advantage of rapid transmission of data to a specialist. The benefits of mobile health (mHealth) include maintenance of electronic medical records, record of community data collection, public health surveillance, remote consultation, patient monitoring, appointment compliance, cloud-based storage service, and autogenerated reporting.,,,,, The mobile phone used by CHWs would have improved access to the rural population by connecting with the specialist, and can identify the high-risk group and facilitate timely referral. However, there are very few studies to assess the utility of CHW for early oral cancer detection in a technology-based platform.
The present study was a mHealth-based oral cancer-screening program in a workplace setting. The objectives of the study were to evaluate the ability of CHWs in the identification of oral mucosal lesions using mobile technology and determine the agreement in the identification of oral lesion between CHWs using mHealth and onsite oral medicine specialists vis-à-vis remote oral medicine specialists.
| » Materials and Methods|| |
A cross-sectional analytical study was conducted on the workforce in a pipeline factory due to high risk of tobacco usage reported in baseline data. The sample size was estimated based on the pilot study conducted in the same cohort. A convenient sampling of 50 individuals was selected to assess the feasibility of mobile phone use by CHW in identification of oral lesions in comparison with onsite specialist. Based on the ability of CHW in identification of oral lesion using mHealth, the minimum sample size was calculated using the following formula:
By substituting the values (P = 75.4%, D = 3, E = 0.0754) in the above formula, the minimum sample size was found to be 652. However, the entire cohort in the workplace was screened.
This oral cancer-screening program was a mHealth program conducted for 8 months from June 2015 to January 2016. The early detection and prevention of oral cancer screening program actively involved two trained CHWs, two oral medicine specialists, an oral surgeon, and a pathologist. An oral medicine specialist performed onsite diagnosis and another provided remote consultation. Institutional Ethics Committee approved the study protocol. The eligible participants were included in the study after written informed consent.
[Box 1] illustrates the workflow of the screening program. Two CHWs were high school educated. The training for 3 days included a power point presentation, one-to-one discussion, and use of education modules through a clinical manual. The content of training module comprehensively included information on oral cancer disease burden, awareness, and prevention. They were trained to identify suspicious oral lesions and were educated on risk factors and importance of habit cessation. CHWs underwent training in a clinical setup that involved chair-side clinical examination of various subsites of oral cavity in identification of normal and tobacco-associated mucosal lesions by an oral medicine specialist. One of our studies is under review, which showed 84% improvement in posttest assessment following training and showed a 20% reduction in knowledge retention following 3 months (unpublished observations). The two android phones with a Poi mapper mobile App was used in training. The Poi mapper app on the mobile phones of CHW was used for training and data-capturing purposes. The images were captured in autofocus and autoflash mode.
The early detection and prevention of oral cancer-screening program were conducted in a workplace setting. Before screening all employees attended an educative talk on the importance of prevention and early detection of oral cancer and were encouraged to participate in the screening program. All employees were potential participants in the screening. Eligible participants from the workplace setting were issued a medical ID and were de-identified for the study. Age range of employees who participated in the study was between 18 and 57 years and all were men.
The entire population screening (n = 3445) was undertaken in batches of 30–45 participants per day for about 4 months. Participants were screened by CHWs using mobile phone-based questionnaire [Box 2]. Capture and uploading of the photographs of the oral lesions and normal mucosa of the participants was performed using the mobile phone camera. An onsite specialist re-screened the same participant, collected the data using case report form [Box 3], provided the appropriate diagnosis, and recommended biopsy if required. Histopathological examination based on clinical diagnosis of onsite specialist was done. An oral surgeon performed punch biopsy using 5 mm disposable punch on the suspicious lesions if recommended by the onsite specialist. Histopathological specimens were classified as per WHO, i.e., hyperkeratosis, mild, moderate, severe dysplasias, carcinoma in situ, and squamous cell carcinoma. Observations entered on Microsoft Excel were analyzed using Statistical Package for the Social Sciences (SPSS) version 17 for Windows. Sensitivity, specificity, positive and negative predictive values were calculated. Inter-rater agreement was analyzed with Cohen's kappa coefficient (κ) test, and the diagnostic ability of CHWs and onsite specialist vis-à-vis remote specialist was illustrated using receiver operating characteristic (ROC) curve.
All normal individuals without oral lesion were discharged with education and awareness on the ill-effects of tobacco. The diagnosis by onsite specialist was confirmed with histopathology. Individuals who were positive for tobacco usage and had OPMDs underwent individual risk-mitigation program. The nondysplastic and mild dysplastic lesions were managed conservatively with the use of health education on adopting healthier lifestyles by quitting tobacco and alcohol in any form and topical vitamin A prescription. Surgical excision and advise on tobacco consumption cessation along with chemopreventive medication were given to individuals with severe dysplastic lesions. A remote oral medicine specialist reviewed the uploaded data in Open Medical Record System (MRS) and made the necessary recommendations [Box 4]. All the observers and the interpretations were blinded.
| » Results|| |
A total of (n = 3445) individuals were screened by trained CHWs followed by screening by onsite oral medicine specialist. Among the target population screened, 1025 (29.8%) individuals did not consume alcohol or tobacco in any form, 1233 (35.8%) individuals used tobacco either in the form of smoking cigarettes/beedi or chewing paan with/without tobacco or any other commercially available tobacco products, and 2420 (70.2%) individuals had a combination of tobacco habit and alcohol use. [Table 1] shows identification of individuals with oral lesions and normal individuals by an onsite specialist, remote specialist, and CHWs. [Box 5],[Box 6],[Box 7] show a schematic representation of target population screened by CHWs, onsite specialist, and remote specialist.
|Table 1: Identification of oral lesion by onsite specialist, CHW, and remote specialist in the population screened|
Click here to view
Identification of oral lesions by CHW using mHealth and onsite oral medicine specialist diagnosis
Among 3445 screened individuals, an onsite oral medicine specialist found 3051 (88.6%) individuals as normal and identified 394 (11.4%) positive oral lesions. CHWs identified 3040 (88.2%) individuals as normal and 405 (11.8%) individuals as having positive lesions [Graph 1]. The positive oral lesions included 334 (84.8%) true positive oral lesions, false positive in 71 (2.3%) individuals, and false negative in 60 (15.2%) individuals identified by CHWs [Table 2].
|Table 2: Identification of oral lesion by CHW and an onsite specialist in the screened population|
Click here to view
The positive predictive value of screening test in early detection of oral cancer was 84.8%, and negative predictive value was 97.7%. Sensitivity and specificity were 84.7 and 97.6%, respectively [Graph 2]. The inter-rater agreement between CHWs and onsite specialist was calculated using the formula
For identification of oral lesions, there was 96% agreement between CHW and onsite specialist in identification of oral lesions with a κ score of 0.92, which showed almost perfect agreement. The cutoff value was set at 0.5, with a specificity of 97.6% and sensitivity of 84.7%. The area under curve (AUC) for the ability in identifying oral lesion was 0.764, demonstrating fair discriminatory power (confidence interval = 0.691–0.837; P value < 0.001) [Graph 3] and [Table 3]. There was statistically significant association between onsite specialist and CHW in identification of oral lesions (χ2-value = 85.151; P value < 0.001) [Table 2].
|Table 3: ROC curve with 95% confidence interval for CHW and onsite specialist and CHW and remote specialists|
Click here to view
Identification of oral lesions by CHW using mHealth and remote oral medicine specialist diagnosis
[Table 4] shows agreement between CHW and remote specialist for 334 (75.2%) individuals in the identification of oral lesions. The positive predictive value of screening test in early detection of oral cancer was 75.2%, and negative predictive value was 97.6%. The sensitivity and specificity were 75.2 and 97.6%, respectively. CHW identified 71 (2.3%) false positive among individuals. There was 96% agreement between CHW and remote specialist in identification of oral lesions with a κ score of 0.62, which showed substantial agreement. The cutoff value was set at 0.5, with a specificity of 97.6% and sensitivity of 75.2% (CI = 0.491–0.632; P value = 0.07) and the AUC was estimated to be 0.562 [Graph 3] and [Table 3]. There was statistically significant association (χ2-value = 25.457; P value < 0.001) between remote specialist and CHW [Table 4].
|Table 4: Identification of oral lesion by CHWs and remote specialist diagnosis|
Click here to view
Among 394 positive oral lesions identified by an onsite specialist, 302 (76.6%) individuals underwent biopsy; 92 (23.3%) individuals refused for biopsy. Biopsy correlated with the onsite diagnosis. Among 302 biopsied individuals, the most common diagnosis was hyperkeratotic epithelium in 167 (55.3%) individuals; probably the ease of identifying oral lesions could be due to the distinct clinical appearance and its correlation to the site of quid placement. The other premalignant lesions showed mild epithelial dysplasia in 53 (17.5%) individuals, moderate epithelial dysplasia in 6 (2%) individuals, and 9 (3%) individuals showed severe epithelial dysplasia. Moreover, 50 (16.6%) individuals had a lichenoid reaction, 8 (2.6%) individuals showed lichenoid dysplasia, and 9 (3%) cases were indeterminate [Graph 4]. [Table 5] shows the lesions identified by CHW and remote specialist in biopsy-confirmed lesions. A statistically significant association was found between CHW and remote specialist in biopsy-confirmed lesions (χ2-value = 50.33; P value < 0.001). The error rates with histopathology as gold standard between CHW and onsite specialist was 0.84, between CHW and remote specialist was 0.43, and 0.22 between onsite and remote specialist.
|Table 5: Comparison of diagnosis given by HCW and remote specialist in biopsy-confirmed lesions|
Click here to view
Comparison of remote oral medicine specialist with an onsite oral medicine specialist
Of the 3445 population screened, Open MRS data were retrievable on 3414 (99.1%); 31 (0.9%) cases were missing at the backend. There were no images for 27 (0.8%) individuals, and 36 (1.1%) images were of poor quality. The remote specialist identified a total of 444 (13%) individuals as having oral lesions and 2970 (87%) individuals without lesions. Among 444 identified lesions by remote specialist, 68 (15.31%) individuals were false positive and 12 (2.7%) were false negative [Graph 5].
Among 394 individuals having oral lesions, there were only 391 (99.2%) cases uploaded in Open MRS. Three (0.8%) individuals having oral lesions were missing at the backend. Among these 391 positive uploads, remote specialist identified 376 (96%) individuals as having oral lesions and had missed 15 (3.8%) oral lesions [Table 6]. The positive predictive value in early detection of oral cancer was 85.2%, and negative predictive value was 99.5%. Sensitivity and specificity were 96.2 and 97.8%, respectively. The diagnostic accuracy of a remote specialist was 96.4%. There was 97% agreement between remote specialist and onsite specialist in identification of oral lesions with a κ = 0.94, which had almost perfect agreement. A statistically significant association was found between remote specialist and onsite specialist in identification of oral lesion (χ2-value = 119.01; P value <0.001).
|Table 6: Identification of oral lesion by remote specialist and an onsite specialist in the population screened|
Click here to view
| » Discussion|| |
We found that there are no studies with this approach of using technology, CHW, an onsite specialist, and remote oral medicine specialist for oral cancer-screening program and to assess the use of CHW in identifying oral lesions. This study is the first of its kind to assess the usefulness of CHW in identifying oral lesions and use of mHealth in oral cancer screening. This module is unique as onsite specialist re-screened all participants and provided necessary care. Studies,, have shown that trained CHWs can effectively identify early oral lesions in the oral cancer-screening program.
Shankarnarayanan et al. and Mathew et al. have proved the feasibility of trained CHW in identifying oral lesions by visual screening and their utilization in oral cancer-screening program. They determined that oral screening of high-risk individuals should be established in routine health services in India, given the high disease burden. The study found that the sensitivity of oral cancer screening visually by CHW was 94.3%, and specificity of 99.3% when referenced to the physician. Several studies found the sensitivity of oral cavity inspection visually varied from 57.1 to 61.4%, and the specificity ranged from 98.6 to 98.8%.,, However, opportunistic screening by dental professional in targeted cohort found 100% concordance with specialist.
Conventional screening has several limitations. Some of the drawbacks include lack of permanent portable record of data, poor patient compliance to follow-up after screening, and delayed communication between the health care provider and the specialist. It is just not the early detection, but monitoring the disease and surveillance for the progression to OPMDs is important in downstaging oral cancer.
In the present study, the shortcomings of conventional screening methods were reduced by integrating health and technology. In this prospect, CHW trained with mobile phone-based technology was used as a decision-based algorithm for screening. Use of mHealth-based approach by CHW aids in remote diagnosis through remote consultation. However, remote consultation is a time-consuming process and we had faced a few challenges. One of the limitations of the remote specialist opinion depends on data provided by the CHW. During the screening process, CHW might have captured image of an inappropriate subsite of the oral cavity, probably missed imaging of the lesion and poor-quality images. However, remote consultation reduces the cost of travel, time, unnecessary referral of the patients. Triaging of patients is the priority through remote consultation. CHW plays a role of the first contact with the high-risk group in follow-up and referrals. If implemented this strategy can downstage oral cancer and improve survival rates.
The remote specialist discharged the normal individuals as closed cases and generated a follow-up queue based on the presence of oral lesions. If there is doubt regarding diagnosis, the individuals were directed to the follow-up queue. The remote specialist recommendation is reliable; however, it depends on the uploaded data and images provided by CHW. Opportunistic screening by general dentist was better than remote consultation in terms of lesion detection, image quality, and diagnostic accuracy. The false positive results (2.2%) by a remote specialist could be due to over diagnosis by the specialist.
The technical challenge in rural India is poor connectivity and low bandwidth, which poses a challenge in uploading files of large size. Hence the protocol was to capture the altered mucosa and a normal mucosa rather than taking images of entire oral cavity. However, if multiple suspicious lesions were found, all the subsites of oral mucosa were captured and the size of the images was optimized (image size ranged about 214–245 kb with a resolution of 1880 × 1056).
Braun et al. and Källander et al. in their systematic review have shown that CHWs have used mobile technology effectively in the house-to-house screening, person-to-person communication, data collection, promoting health education, counseling, and referral for further care in various domains of health. Our earlier studies, have proved the novelty of mobile health-based approach and aided remote early detection of oral cancer by CHW in a resource-constrained setting.
Advantages of this module
The system facilitates electronic data capture, assists remote consultation, connects the population to the specialist, and provides geo tagging for high-risk group by aiding in surveillance. This module assists in capacity building and empowerment of local resources (CHW). For patients it helps in educating, improves self-oral care, and creates awareness of the harmful effects of tobacco usage individually. It addresses the patient care most optimally ensuring timely referral.
This study did not assess the agreement between two CHWs as individual CHW performed screening on different subgroups. Thirty-one cases were missing in Open MRS among 3414 uploads, which could be due to voluminous screening and an incorrect entry of alphanumeric IDs; and 36 images were of poor quality due to poor retraction and instability to focus on lesions because of subject movement and/or phone movement, zooming with autoflash. The error rates with histopathology as gold standard among the three observers were 0.84 between CHW and onsite specialist, 0.43 between CHW and remote specialist, and 0.22 between onsite and remote specialist.
Future plan includes optimizing image capture, triage patients, facilitate image upload in low-band connectivity, use of a magnifying lens, good lighting, and tissue retractors for better outcomes. To address the missing data in a voluminous screening, we are introducing quick response code in our further programs to reduce the human error during manual entry of IDs.
| » Conclusion|| |
CHW can be utilized for early detection and screening of oral lesions in oral cancer program. There was good agreement between onsite specialist diagnosis and CHW. This module facilitates remote consultation and enables health care services far from clinical setting at remote places.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| » References|| |
Agrawal M, Pandey S, Jain S, Maitin S. Oral cancer awareness of the general public in Gorakhpur City. India Asian Pacific J Cancer Prev 2012;13:5195-9.
Birur PN, Sunny SP, Jena S, Uma K, Raghavan SA, Bhanushree R, et al
. Mobile health application for remote oral cancer surveillance. J Am Dent Assoc 2015;146:886-94.
Sankaranarayanan R, Ramadas K, Thomas G, Muwonge R, Thara S, Mathew B, et al
. Effect of screening on oral cancer mortality in Kerala, India: A cluster-randomized controlled trial. lancet 2005;365:1927-33.
Desai R, Birur P, Bajaj S, Shubhasini AR, Bhanushree R, Shubha G, et al
. Smokeless tobacco associated lesions: A mobile health approach. J Contemp Dent Pract 2015;16:813-8.
Ford PJ, Farah CS. Early detection and diagnosis of oral cancer: Strategies for improvement. J Cancer Policy 2013;1: e2-7.
Braun R, Catalani C, Wimbush J, Israelski D. Community health workers and mobile technology: A systematic review of the literature. PLoS One 2013;8:e6572.
Sankaranarayanan R. Health care auxiliaries in the detection and prevention of oral cancer. Oral Oncol 1997;33:14954-6.
Mathew B, Sankaranarayanan R, Sunilkumar KB, Kuruvilla B, Pisani P, Nair KM. Reproducibility and validity of oral visual inspection by trained health care workers in the detection of oral precancer and cancer. Br J Cancer 1997;76:390-4.
Ganesan M, Prashanta S, Jhunjhunwala A. A review on challenges in implementing mobile phone based data collection in developing countries. J Health Inform Dev Ctries 2012;6:366-74.
Lee Ventola MS. Mobile devices and apps for health care professionals: Uses and benefits.P&T 2014;39:356-64.
Hall CS, Fottrell E, Wilkinson S, Byass P. Assessing the impact of mHealth interventions in low- and middle-income countries what has been shown to work. Glob Health Action 2014;7:25606.
Leon N, Schneider H, Daviaud E. Applying a framework for assessing the health system challenges to scaling up mHealth in South Africa. BMC Med Infor Decis Mak 2012;12:123.
Celi LA, Sarmenta L, Rotberg J, Marcelo A, Clifford G. Mobile care (Moca) for remote diagnosis and screening. J Health Inform Dev Ctries 2009;3:17-21.
Bajwa M. mHealth security. Pak J Med Sci 2014;30:904-7.
Mathew B, Sankaranarayanan R, Wesley R, Joseph A, Nair MK. Evaluation of utilization of health workers for secondary prevention of oral cancer in Kerala, India. Eur J Cancer B Oral Oncol 1995;31:193-6.
Sankaranarayanan R, Black RJ, Parkin DM. Cancer Survival in Developing Countries. Lyon: IARC; 1998.
Yeole B, Sankaranarayanan R, Sunny L, Swaminathan R, Parkin DM. Survival from head and neck cancer in Mumbai (Bombay), India. Cancer 2000;89:437-44.
Dikshit R, Gupta PC, Ramasundarahettige C, Gajalakshmi V, Aleksandrowicz L, Badwe R, et al
. Cancer mortality in India: A nationally representative survey. Lancet 2012;379:2343-51.
Källander K, Tibenderana JK. Mobile health (mHealth) approaches and lessons for increased performance and retention of community health workers in low- and middle-income countries: A review. J Med Internet Res 2013;15:e17.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]