Clinical prediction models are used frequently in clinical practice to identify patients who are at risk of developing an adverse outcome so that preventive measures can be initiated. A prediction model can be developed in a number of ways; however, an appropriate variable selection strategy needs to be followed in all cases. Our purpose is to introduce readers to the concept of variable selection in prediction modelling, including the importance of variable selection and variable reduction strategies. We will discuss the various variable selection techniques that can be applied during prediction model building (backward elimination, forward selection, stepwise selection and all possible subset selection), and the stopping rule/selection criteria in variable selection (p values, Akaike information criterion, Bayesian information criterion and Mallows’ Cp statistic). This paper focuses on the importance of including appropriate variables, following the proper steps, and adopting the proper methods when selecting variables for prediction models.

Background: Hypertension, itself being a major chronic condition, is one of the most significant risk factors for premature cardiovascular diseases and mortality. Hypertension is responsible for 13% of global deaths and threequarters of the world's hypertensive population reside in low-and middle-income countries. Bangladesh is one of those countries that experiencing an epidemiological transition from communicable to non-communicable diseases, a nutritional transition from a traditional diet to process and fast food, and an increase in a sedentary lifestyle, resulting in increased hypertension prevalence. We carried out a systematic review and meta-analysis to identify existing research on hypertension prevalence in Bangladesh, summarize findings and assess its temporal change. Methods: We searched MEDLINE, EMBASE and PubMed and relevant references to identify studies on the prevalence of hypertension in Bangladesh. We used Random-effects meta-analysis to pool the prevalence estimates and performed subgroup analyses. We assessed heterogeneity, a trend in prevalence of hypertension and publication bias in selected studies. Results: Our search initially identified 735 articles and after removing duplicates, reviewing titles and abstracts, and screening full texts, 53 studies were finally selected. The studies comprised 305,432 subjects and reported overall, gender-specific, geographical location specific and criteria specific prevalence of hypertension. We identified the range of hypertension prevalence is from 1.10% to 75.0% and the overall weighted pooled prevalence of hypertension is 20.0%. An extremely high heterogeneity (I 2 = 99.53%; Cochran Q-statistic p < 0.001) was observed in the prevalence of hypertension. Consequently, we performed subgroup analysis based on gender, age group and geographical location of the study participants, the cutoff level used to define hypertension, and the types of hypertension reported and presented our findings accordingly. An overall increasing trend of hypertension prevalence is also observed. Conclusions: The prevalence of hypertension is high and rising in Bangladesh. Strategies targeting prevention are required to mitigate a further increase in the prevalence and reduce the morbidity and mortality associated with it.

BackgroundMetabolic syndrome (MS) is a cluster of health problems that set the stage for serious health conditions and places individuals at higher risk of cardiovascular disease, diabetes and stroke. The worldwide prevalence of MS in the adult population is on the rise and Bangladesh is no exception. According to some epidemiological study, MS is highly prevalent in Bangladesh and has increased dramatically in last few decades. To provide a clear picture of the current situation, we conducted a systematic review and meta-analysis with an objective to assess the prevalence of metabolic syndrome among the Bangladeshi population using data already published in the scientific literature.MethodsWe searched MEDLINE, EMBASE and PubMed and manually checked references of all identified relevant publications that described the prevalence of MS in Bangladesh. Random effects meta-analysis was used to pool the prevalence. Heterogeneity was explored using formal tests and subgroup analyses. Study quality and publication bias was also explored.ResultsElectronic and grey literature search retrieved 491 potentially relevant papers. After removing duplicates, reviewing titles and abstracts and screening full texts, 10 studies were finally selected. Most of the studies were conducted in rural populations and study participants were mostly females. The weighted pooled prevalence of metabolic syndrome regardless of gender and criteria used to define metabolic syndrome, was 30.0% with high heterogeneity observed. Weighted pooled prevalence of metabolic syndrome is higher in females (32%) compared to males (25%) though not statistically significant (p = 0.434). Prevalence was highest (37%) when Modified NCEP ATP III criteria was used to define MS, while it was lowest (20%) when WHO criteria was used. In most cases, geographical area (urban/rural) was identified as a source of heterogeneity between the studies. Most of the studies met study quality assessment criteria’s except adequate sample size criteria and evidence of small study effect was also detected.ConclusionsThe prevalence of metabolic syndrome is high and rising in Bangladesh. Strategies aimed at primary prevention are required to mitigate a further increase in the prevalence and for the reduction of the morbidity and mortality associated with metabolic syndrome.Electronic supplementary materialThe online version of this article (10.1186/s12889-018-5209-z) contains supplementary material, which is available to authorized users.

No abstract

BackgroundCardiovascular disease (CVD) is a group of conditions affecting the functioning of the heart or blood vessels and is one of the leading causes of death globally. Like other countries, CVD prevalence is also rising among the adults in Bangladesh. Epidemiological studies have shown not only a high CVD prevalence but also a significant increase in its prevalence in Bangladesh in the last few decades. To have a better understanding of the current CVD prevalence scenario, we conducted this systematic review and meta-analysis. Our objective was to assess the prevalence of CVD among the Bangladeshi adult population using evidence from the published scientific literature.MethodsElectronic databases such as MEDLINE, Embase and PubMed were searched. We also manually checked the references of all relevant publications that describe the prevalence of CVD in Bangladeshi adults. To pool the CVD prevalence, we used random-effects meta-analysis. We assessed heterogeneity using both the formal tests and the subgroup analyses. We also assessed study quality and examined publication bias.ResultsWe retrieved 755 potentially relevant papers through searches of electronic and gray literature, of which only 13 met inclusion criteria after the screening and were included in this review. Of the studies that met inclusion criteria, three were carried out in rural populations, five in both urban and rural populations and two in strictly urban populations. Male and female participation in the studies was almost equal. The weighted pooled prevalence of CVD was 5.0%, regardless of the types of CVD, gender and geographical location of the study participants. There was also a high heterogeneity in the observed CVD prevalence. Weighted pooled prevalence of overall CVD in the Bangladeshi population was higher in urban areas (8%) compared to rural areas (2%). However, no such difference was observed in terms of gender (3% for both males and females). The highest reported prevalence (21%) was for heart disease, while the lowest reported prevalence (1%) was for stroke. Sources of heterogeneity were often unexplained. The criteria used to assess study quality were fulfilled by only a few studies, and adequate sample size criteria was missed by almost all of them. In addition, there was evidence of small-study effects.ConclusionA high CVD prevalence along with an upward trend was observed in Bangladeshi adults. Proper strategies are required for primary prevention of CVD so that a further increase can be alleviated and the morbidity and mortality associated with it can be reduced.