Year: 2014
Type: dissertation
Study design: Observational Study
Abstract: Background: Oral cancer is a malignant disease contributing to one third of the total cancer burden in India. There is a worldwide social disparity in oral cancer incidence and survival. Life course epidemiology has shown that early-life socioeconomic conditions (SEC) could influence adult health through various pathways. Thus, the socioeconomic disparities in the occurrence of oral cancer underscore the importance of understanding the ‘life course processes’ operating between SEC at different stages in life and oral cancer. In addition to understanding socioeconomic disparities, practical solutions are required to reduce the burden of the disease. Early diagnosis and prompt treatment could reduce morbidity and mortality. Though visual screening helps in early diagnosis, it requires training and calibration of the screeners. Developing a simple screening model that can be utilized by untrained health care workers will be helpful in triaging asymptomatic adults with oral cancer. Therefore a study was designed with the following hypotheses. Research hypotheses 1. Accumulation of socioeconomic disadvantage over the life course is associated with oral cancer in the Indian population. 2. Early-life socioeconomic disadvantage has a lasting effect on the oral cancer outcome in adulthood in the Indian population. 3. An oral cancer screening model developed for the Indian population, to screen high-risk people from rural/remote areas, has good predictive ability. Methods: A multicentre hospital based case-control study was conducted between July 2011 and August 2012 in Karnataka, India. Cases were newly diagnosed oral and oropharyngeal cancer patients and controls were patient-visitors or patients seeking care for other reasons. Data were collected through direct interviews, oral examination and record abstraction. Cases were ascertained from hospital records. A semi-structured questionnaire was designed to collect life course information on SEC, family structure, housing conditions, parental habits of tobacco, quid and alcohol use, parental education, family history of malignancy, participants’ own diet, tobacco, quid, alcohol use and oral hygiene behaviour. A life-grid was used to improve recall accuracy. All consenting participants underwent an oral examination following an interview by a trained examiner. Oral soft and hard tissues were examined for the presence of any oral mucosal lesions, teeth present and oral hygiene status. Data were analysed using SAS v 9.2. Conventional logistic regression models were used to determine the associations between life course SEC and oral cancer. Marginal structural model (MSM) was built to estimate the controlled direct effect of childhood SEC on oral cancer in adulthood. The validity of effect measures was checked with sensitivity analysis. A multivariable logistic regression model was used to develop a screening model for identifying individuals at high-risk for developing oral cancer. The development of the model involved deriving risk scores for the predictors of oral cancer. The predictive ability of the screening model was examined with c statistics, sensitivity, specificity and predictive values. Results: A total of 180 incident cases and 272 controls participated in the study. Of them, 163 cases and 264 controls had complete information on SEC at all three stages. Nearly two-thirds (65%) of participants were stable in low SEC across all stages. Low SEC at all the three stages (childhood, early adulthood and later adulthood) was associated with oral cancer after adjusting for age and sex. The association was strongest for those who remained in the low SEC at all the three stages. Odds ratios (OR) for oral cancer in socially mobile groups were intermediate to that of the stable groups. The largest differences in OR for oral cancer were observed between the stable groups. The total effect model showed that the risk was 63% [Risk ratio (RR) = 1.63 (95% CI = 1.38– 1.92)] higher for those who lived in low SEC in childhood than for those in high SEC. From the MSM, the estimated risk for developing oral cancer for those in low SEC during early-life was 48% [RR = 1.48 (95% CI = 1.43–1.53)], 24% [RR = 1.24 (95% CI = 0.88–1.74)] and 94% [RR = 1.94 (95% CI = 1.66–2.27)] greater than those in the high SEC after controlling for smoking, chewing and alcohol respectively. However, the adjusted effect of low SEC on oral cancer was null when analysed using conventional regression. A screening model was developed using statistical methods that involved smoking, chewing quid and/or tobacco, alcohol, family history of upper aero-digestive tract (UADT) cancer, diet and oral hygiene behaviour as predictors. Total risk score that was derived from odds ratio ranged from 0 to 28. Area under the curve of the Receiver Operating Characteristic (ROC) curve for risk scores was 0.866. The sensitivity (0.928) and negative predictive value (0.927) were higher while specificity (0.603) and positive predictive value (0.607) were lower for risk scores cut-off of 6. Conclusions: Low SEC in childhood and early adulthood are important in determining oral cancer in later adulthood. Early-life socioeconomic disadvantage increases the risk for oral cancer that is not mediated by later life risk factors when MSM was used. The developed screening model using risk scores had satisfactory predictive ability in the study population. However, validation of the model in other settings is necessary before it can be recommended to identify subgroups of the people to be referred for further clinical evaluation.
Authors: Krishna Rao, Sree Vidya
Language: English
Topic: Head and Neck Cancer Studies
Subfield: Otorhinolaryngology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W2586575522