Page 85 - ISMCON souvenir 2021

ISMSCON - 2021

Methods- Age and country wise incidence and mortality estimates for cervical cancer were obtained for
11 SEAR countries using data available from Globocan 2020. Age specific disease burden was analysed
using incidence, mortality and MI ratio (mortality/incidence).Data on human development index (HDI)
was extracted from United Nations Development Programme report. Bivariate correlation analysis was
done for HDI in relation to Incidence, Mortality and MI ratio separately.
Results – In SEAR cervical cancer estimates for the year 2020 were 190874 cases and 116015 deaths.
Age-standardized incidence rate (ASIR) was 18.1/100,000 and Age-standardized mortality rate (ASMR)
was 11.1/100,000. India was the leading country with 123907 cases and 77348 deaths. ASIR was highest
among 50-55 years age group whereas ASMR was maximum in 55-59 years age group. Association
between HDI and MI ratiowas negativewith significant correlation (r=-0.664, p=0.026). Association
between ASIR and HDI as well as ASMRand HDI though negative butwas found to be non-significant.

Conclusion- To reduce cervical cancer burden in Southeast Asia region, main challenges are lack of
cancer awareness, lack of systematic screening, late stage at disease presentation and poor treatment
facilities. Mortality is higher in older age group. There was an inverse relationship between HDI and
cervical cancer incidence and mortality in the SEAR.
Keywords: Cervical Cancer, Trend, Incidence, Mortality, Mortality-Incidence Ratio (MIR)

OS59: COMPARATIVE STUDY OF METHODS TO ESTIMATE

USUAL INTAKE DISTRIBUTION USING MEASUREMENT ERROR
MODEL

Smitha Joseph , Tinku Thomas , Sumathi Swaminathan 3
2
1
1,2Division of Epidemiology and Biostatistics, 3Division of Nutrition, St. Johns Research Institute, Bangalore.
Email ID: smitha.j@sjri.res.in

There is a strong relationship between diet and disease of an individual. To understand this relation
correct capturing of dietary data is essential. In dietary assessment using 24-hour recalls it has been
observed that the variability within individual is very high as the intake of individuals vary from time to
time. A simple estimate of usual intake is the average of individual intakes if a sufficiently large number
of recalls are available. Collecting multiple recalls are infeasible due to constraints in the resources and
the time to be spend by the respondents. The distribution of many nutrients found to be skewed due to
the infrequent consumption and variability in the data. But the existing methods used for estimation of
usual intake requires data to be symmetric. Due to these challenges better methods are required for the
estimation of usual intake distribution
Objectives: To estimate usual intake distribution of nutrients with the help of measurement error model.

Method : Data was collected to evaluate the dietary pattern and estimate the nutrient intake of
complementary feeds in 6 to ≤ 60 months children. This data includes intake from 2 non-consecutive
diet recalls measured during two seasons. Usual intake distributions were estimated using Iowa State
University (ISU method). As a first step nutrient intakes were adjusted for covariates related to recall, such
as day of recall, season etc. Following this the data were power transformed. A continuous piecewise
linear estimate of the power transformed values were obtained and measurement error model applied
on these transformed values. Using polynomial regression, curve fititng was used to back transform the
usual intake value to original scale of nutrient intake.

Result : The nutrient intake data could be transformed to normality by ISU method except vitamin B12
which is positively skewed and infrequently consumed with high proportion of extremely low intakes. The
Median (IQR) from ISU method was 0.4 mcg (0.1, 0.7) vs 0.3 mcg (0.1, 0.8) by averaging intakes.

CONFERENCE SOUVENIR 83