Analysing outlier communities to child birth weight outcomes in Malawi: application of multinomial logistic regression model diagnostics

A bivariate Poisson regression to analyse impact of outlier women on correlation between female schooling and fertility in Malawi

BACKGROUND

Women's levels of education and fertility are commonly associated. In Sub-Saharan Africa, the pace of decreasing fertility rates varies greatly, and this is linked to women's levels of education. However, this association may be influenced by unusual females who have uncommon measurements on both variables. Despite this, most studies that researched this association have only analysed the data descriptively, without taking into account the effect of potential outliers. This study aimed to examine the presence and impact of outlier women on the relationship between female education and fertility in Malawi, using regression methods.

METHODS

To analyse the correlation between women's schooling and fertility and evaluate the effect of outliers on this relationship, a bivariate Poisson model was applied to three recent demographic and health surveys in Malawi. The R software version 4.3.0 was used for model fitting, outlier computations, and correlation analysis. The STATA version 12.0 was used for data cleaning.

RESULTS

The findings revealed a correlation of -0.68 to -0.61 between schooling and fertility over 15 years in Malawi. A few outlier women were identified, most of whom had either attended 0 or at least 9 years of schooling and had born either 0 or at least 5 children. The majority of the outliers were non-users of modern contraceptive methods and worked as domestic workers or were unemployed. Removing the outliers from the analysis led to marked changes in the fixed effects sizes and slight shifts in correlation, but not in the direction and significance of the estimates. The woman's marital status, occupation, household wealth, age at first sex, and usage of modern contraceptives exhibited significant effects on education and fertility outcomes.

CONCLUSION

There is a high negative correlation between female schooling and fertility in Malawi. Some outlier women were identified, they had either attended zero or at least nine years of schooling and had either born zero or at least five children. Most of them were non-users of modern contraceptives and domestic workers. Their impact on regression estimates was substantial, but minimal on correlation. Their identification highlights the need for policymakers to reconsider implementation strategies for modern contraceptive methods to make them more effective.

Bivariate logistic regression model diagnostics applied to analysis of outlier cancer patients with comorbid diabetes and hypertension in Malawi

The joint occurrence of diabetes and hypertension conditions in a patient is common. The two diseases share a number of risk factors, and are hence usually modelled concurrently using bivariate logistic regression. However, the postestimation assessment for the model, such as analysis of outlier observations, is seldom carried out. In this article, we apply outlier detection methods for multivariate data models to study characteristics of cancer patients with joint outlying diabetes and hypertension outcomes observed from among 398 randomly selected cancer patients at Queen Elizabeth and Kamuzu Central Hospitals in Malawi. We used R software version 4.2.2 to perform the analyses and STATA version 12 for data cleaning. The results showed that one patient was an outlier to the bivariate diabetes and hypertension logit model. The patient had both diabetes and hypertension and was based in rural area of the study population, where it was observed that comorbidity of the two diseases was uncommon. We recommend thorough analysis of outlier patients to comorbid diabetes and hypertension before rolling out interventions for managing the two diseases in cancer patients to avoid misaligned interventions. Future research could perform the applied diagnostic assessments for the bivariate logit model on a wider and larger dataset of the two diseases.