Classification tree analysis: a statistical tool to investigate risk factor interactions with an example for colon cancer (United States)

Development and application of a weighted change score to evaluate interventions for vasomotor symptoms in patients with breast cancer using regression trees: a cohort study

PURPOSE

Vasomotor symptoms (VMS) are common among individuals with breast cancer (BC) and poorly managed symptoms are associated with reduced quality of life, treatment discontinuation, and poorer breast cancer outcomes. Direct comparisons among therapies are limited, as prior studies evaluating VMS interventions have utilized heterogeneous change measures which may not fully assess the perceived impact of change in VMS severity.

METHODS

We performed a prospective study where BC patients chose one of four categories of interventions to manage VMS. Change in VMS severity at 6 weeks was assessed using the validated Hot Flush Rating Scale (HFRS). A novel weighted change score integrating baseline symptom severity and directionality of change was computed to maximize the correlation between the change score and a perceived treatment effectiveness score. Variables influencing change in VMS severity were included in a regression tree to model factors influencing the weighted change score.

RESULTS

100 baseline and follow-up questionnaires assessing VMS were completed by 88 patients. Correlations between treatment effectiveness and VMS outcomes strengthened following adjustment for baseline symptoms. Patients with low VMS severity at baseline did not perceive change in treatment effectiveness. Intervention category was predictive of change in HFRS at 6 weeks.

CONCLUSION

Baseline symptom severity and the directionality of change (improvement or deterioration of symptoms) influenced the perception of clinically meaningful change in VMS severity. Future interventional studies utilizing the weighted change score should target moderate-high baseline severity patients.

Demographic and clinical characteristics associated with advanced stage colorectal cancer: a registry-based cohort study in Saudi Arabia

BACKGROUND

In Saudi Arabia, approximately one-third of colorectal cancer (CRC) patients are diagnosed at an advanced stage. Late diagnosis is often associated with a worse prognosis. Understanding the risk factors for late-stage presentation of CRC is crucial for developing targeted interventions enabling earlier detection and improved patient outcomes.

METHODS

We conducted a retrospective cohort study on 17,541 CRC patients from the Saudi Cancer Registry (1997-2017). We defined distant CRCs as late-stage and localized and regional CRCs as early-stage. To assess risk factors for late-stage CRC, we first used multivariable logistic regression, then developed a decision tree to segment regions by late-stage CRC risk, and finally used stratified logistic regression models to examine geographical and sex variations in risk factors.

RESULTS

Of all cases, 29% had a late-stage diagnosis, and 71% had early-stage CRC. Young (< 50 years) and unmarried women had an increased risk of late-stage CRC, overall and in some regions. Regional risk variations by sex were observed. Sex-related differences in late-stage rectosigmoid cancer risk were observed in specific regions but not in the overall population. Patients diagnosed after 2001 had increased risks of late-stage presentation.

CONCLUSION

Our study identified risk factors for late-stage CRC that can guide targeted early detection efforts. Further research is warranted to fully understand these relationships and develop and evaluate effective prevention strategies.

Factors that contributed to Ontario adults' mental health during the first 16 months of the COVID-19 pandemic: a decision tree analysis

The COVID-19 pandemic has negatively impacted the mental health of individuals globally. However, less is known about the characteristics that contributed to some people having mental health problems during the pandemic, while others did not. Mental health problems can be understood on a continuum, ranging from acute (e.g., depression following a stressful event) to severe (e.g., chronic conditions that disrupt everyday functioning). Therefore, the purpose of this article was to generate profiles of adults who were more or less at risk for the development of mental health problems, in general, during the first 16-months of the COVID-19 pandemic in Ontario, Canada. Data were collected via online surveys at two time points: April-July 2020 and July-August 2021; 2,188 adults (Mage = 43.15 years; SD = 8.82) participated. Surveys included a demographic questionnaire and four previously validated tools to measure participants' mental health, subjective wellbeing, physical activity and sedentary behaviour, and sleep. A decision tree was generated at each time point for those with mental health problems, and those with no mental health problems. Results showed that subjective wellbeing was the biggest contributor to mental health status. Characteristics associated with no mental health problems among adults included having good wellbeing, being a good sleeper (quantity, quality, and patterns of sleep), and being over the age of 42. Characteristics associated with mental health problems included having poor wellbeing and being a poor sleeper. Findings revealed that specific characteristics interacted to contribute to adults' mental health status during the first 16 months of the COVID-19 pandemic. Given that wellbeing was the biggest contributor to mental health, researchers should focus on targeting adults' wellbeing to improve their mental health during future health crises.

Patterns of physical activity among nursing home residents before and during the Covid 19 pandemic-a systematic observation

BACKGROUND

The Covid-19 outbreak in spring of 2020 posed an array of challenges for nursing homes, including promoting resident physical activity (PA). Given the diversity of factors affecting resident PA, we explored how activity patterns outside weekly-scheduled structured activities changed during the pandemic and what factors promoted or inhibited PA during the pandemic.

METHODS

We conducted systematic direct observations over 823.5 h in eight nursing homes in Southern Germany in 2020 and 2021.

RESULTS

In 2020, 84.7% of person observation units were classified as sedentary (average activity level: 1.14 MET). In 2021, the percentage increased to 91.6% of observed person units (average activity level: 1.08 MET) (t = 6.947; p = .000). According to tree classification, influencing factors of PA included mealtime and daytime in 2020 and 2021, as well as presence of men residents only in 2020 and guided low threshold activities in 2021.

CONCLUSIONS

Nursing homes constitute highly sedentary places-an issue exacerbated by access restrictions for external activity experts and significant others as well as behavioural restrictions for residents during the Covid-19 pandemic. Staff could not compensate due to existing time restraints and lack of training in PA promotion. Based on our findings, we recommend future studies to develop feasible and resource-low activities to be integrated into the daily routines of nursing homes.

Which Intervention Synergies Maximize AGYW's HIV Outcomes? A Classification and Regression Tree Analysis of Layered HIV Prevention Programming

INTRODUCTION

Intersecting behavioral, social, and structural factors increase adolescent girls' (AG) and young women's (YW) HIV vulnerability. Yet, understanding of optimal intervention synergies remains limited. We identified intervention combinations that statistically maximized reductions in AGYW's HIV-related risk.

METHODS

Using data collected in 2018 with Zambian AG (n = 487, aged 15-19 years) and YW (n = 505, aged 20-25 years) after 12-14 months exposure to Determined, Resilient, Empowered, AIDS-free, Mentored, and Safe (multisectoral HIV program), we used classification and regression trees to explore relationships between interventions (safe space/social asset building [SAB] and provision of/linkage to youth-friendly health services [YFHS], education social protection [Educ], economic social protection [Econ]) and HIV-related outcomes (HIV testing, consistent condom use, transactional sex, and sexual violence experience from partners and nonpartners).

RESULTS

Overall, 59.9% completed SAB and 81.5%, 35.4%, and 29.6% received YHFS, Educ, and Econ, respectively. For AG, HIV testing improved (from 73% to 83%) with exposure to all interventions, condom use improved with Econ (from 33% to 46%), transactional sex reduced with SAB + Educ, and sexual violence from partners and nonpartners reduced with Educ and SAB, respectively. For YW, HIV testing increased with Educ (from 77% to 91%), condom use increased with SAB + YFHS (from 36% to 52%), transactional sex reduced with combinations of all interventions, and sexual violence from partners reduced with YFHS and from nonpartners with SAB + Econ.

CONCLUSIONS

Tailored interventions might be more effective than uniform combination intervention packages in reducing AGYW's HIV risk. AG benefitted most from SAB and/or Educ while YFHS, Educ, and/or SAB reduced YW's HIV-related risk. Educational and asset-building interventions could have the greatest impact on AGYW's HIV risk.

Risk Assessment Score and Chi-Square Automatic Interaction Detection Algorithm for Hypertension Among Africans: Models From the SIREN Study

BACKGROUND

This study aimed to develop a risk-scoring model for hypertension among Africans.

METHODS

In this study, 4413 stroke-free controls were used to develop the risk-scoring model for hypertension. Logistic regression models were applied to 13 risk factors. We randomly split the dataset into training and testing data at a ratio of 80:20. Constant and standardized weights were assigned to factors significantly associated with hypertension in the regression model to develop a probability risk score on a scale of 0 to 1 using a logistic regression model. The model accuracy was assessed to estimate the cutoff score for discriminating hypertensives.

RESULTS

Mean age was 59.9±13.3 years, 56.0% were hypertensives, and 8 factors, including diabetes, age ≥65 years, higher waist circumference, (BMI) ≥30 kg/m2, lack of formal education, living in urban residence, family history of cardiovascular diseases, and dyslipidemia use were associated with hypertension. Cohen κ was maximal at ≥0.28, and a total probability risk score of ≥0.60 was adopted for both statistical weighting for risk quantification of hypertension in both datasets. The probability risk score presented a good performance-receiver operating characteristic: 64% (95% CI, 61.0-68.0), a sensitivity of 55.1%, specificity of 71.5%, positive predicted value of 70.9%, and negative predicted value of 55.8%, in the test dataset. Similarly, decision tree had a predictive accuracy of 67.7% (95% CI, 66.1-69.3) for the training set and 64.6% (95% CI, 61.0-68.0) for the testing dataset.

CONCLUSIONS

The novel risk-scoring model discriminated hypertensives with good accuracy and will be helpful in the early identification of community-based Africans vulnerable to hypertension for its primary prevention.

A wide range of missing imputation approaches in longitudinal data: a simulation study and real data analysis

BACKGROUND

Missing data is a pervasive problem in longitudinal data analysis. Several single-imputation (SI) and multiple-imputation (MI) approaches have been proposed to address this issue. In this study, for the first time, the function of the longitudinal regression tree algorithm as a non-parametric method after imputing missing data using SI and MI was investigated using simulated and real data.

METHOD

Using different simulation scenarios derived from a real data set, we compared the performance of cross, trajectory mean, interpolation, copy-mean, and MI methods (27 approaches) to impute missing longitudinal data using parametric and non-parametric longitudinal models and the performance of the methods was assessed in real data. The real data included 3,645 participants older than 18 years within six waves obtained from the longitudinal Tehran cardiometabolic genetic study (TCGS). The data modeling was conducted using systolic and diastolic blood pressure (SBP/DBP) as the outcome variables and included predictor variables such as age, gender, and BMI. The efficiency of imputation approaches was compared using mean squared error (MSE), root-mean-squared error (RMSE), median absolute deviation (MAD), deviance, and Akaike information criteria (AIC).

RESULTS

The longitudinal regression tree algorithm outperformed based on the criteria such as MSE, RMSE, and MAD than the linear mixed-effects model (LMM) for analyzing the TCGS and simulated data using the missing at random (MAR) mechanism. Overall, based on fitting the non-parametric model, the performance of the 27 imputation approaches was nearly similar. However, the SI traj-mean method improved performance compared with other imputation approaches.

CONCLUSION

Both SI and MI approaches performed better using the longitudinal regression tree algorithm compared with the parametric longitudinal models. Based on the results from both the real and simulated data, we recommend that researchers use the traj-mean method for imputing missing values of longitudinal data. Choosing the imputation method with the best performance is widely dependent on the models of interest and the data structure.

Depressive disorders in older Chinese adults with essential hypertension: A classification tree analysis

Background

Although there has been accumulating evidence on the elevated risk of depression in hypertensive patients, data regarding depressive disorders in older adults with hypertension and the interplay between factors associated with depression in this population are very limited. Disentangling the mutual influences between factors may help illuminate the pathways involved in the pathogenesis of the comorbidity of depression in hypertension. This study investigated the prevalence of depressive disorders in older Chinese adults with hypertension and examined major correlates of depressive disorders and the interactions between correlates by using classification tree analysis (CTA).

Methods

In total, 374 older adults with essential hypertension were enrolled from seven urban and six rural primary care centers in Wuhan, China, and interviewed with the Chinese Mini-international Neuropsychiatric Interview 5.0. Family relationship and feelings of loneliness were assessed with standardized questions. A checklist was used to assess the presence of six major medical conditions: diabetes mellitus, heart disease, cerebrovascular disease, chronic obstructive pulmonary disease, chronic gastric ulcer, and arthritis.

Results

The 1-month prevalence rate of depressive disorders was 25.7%. The CTA model identified four major correlates of depressive disorders: loneliness was the most salient, followed by arthritis, family relationship, and heart disease. There were statistically significant interactions between loneliness and arthritis, loneliness and family relationship, and arthritis and heart disease.

Conclusion

Over one out of every four older Chinese adults with hypertension suffer from depressive disorders. Collaborative multidisciplinary management services are needed to reduce the burden of depression in hypertensive older adults, which may include social work outreach services to promote family relationship, mental health services to relive loneliness, and primary care services to manage arthritis and heart disease.

Prevalence and Correlates of Insomnia Symptoms in Older Chinese Adults During the COVID-19 Outbreak: A Classification Tree Analysis

OBJECTIVE

To examine prevalence and correlates of insomnia symptoms in older Chinese adults (OCAs) during the COVID-19 outbreak.

BACKGROUND

During the COVID-19 pandemic, insomnia is a major health concern of elderly individuals, but its subtypes have not been investigated.

METHODS

Altogether, 590 OCAs (50+ years) were recruited via snowball sampling during the COVID-19 outbreak. Standardized self-report questions were used to assess the presence of difficulty initiating sleep (DIS), difficulty maintaining sleep (DMS), and early morning awakening (EMA). Classification tree analysis (CTA) was used to identify correlates of insomnia.

RESULTS

The one-month prevalence (95% confidence interval) of any subtype of insomnia symptoms was 23.4% (20.0-26.8%), with DIS, DMS, and EMA being 15.4% (12.5-18.3%), 17.1% (14.1-20.2%), and 11.2% (8.64-13.7%), respectively. Worry about being infected with COVID-19 emerged as the most salient correlate of insomnia (P < .001); compared to participants who were not worried about being infected, those who were worried and very worried were 3.2-fold (24.3% vs 7.5%) and 5.5-fold (24.3% vs 7.5%) more likely to have insomnia, respectively. Among participants in the "very worried" branch, those residing in Wuhan were 1.8-fold more likely to have insomnia than those residing in other places (50.0% vs 27.5%, P = .011). Among participants in the "worried" branch, unemployed persons were 2.0-fold more likely to have insomnia than employed persons (37.0% vs 18.1%, P < .001).

CONCLUSIONS

Insomnia symptoms were prevalent among OCAs during the COVID-19 outbreak. Selective intervention programs targeting elderly individuals who are worried about being infected, living in the epicenter of COVID-19, and unemployed might be effective.

Do we need a more flexbible use of Team Timeout calling? Evidence from the Handball Bundesliga

Calling a Team Timeout (TTO) is one of the coaches' most important tools. Given the key competitive advantage to determine your own timing, it is crucial to make a good decision when to use a TTO. Existing research shows that teams can benefit in general from TTOs and that they are called at the end of the game and when trailing (Gomes et al., 2014; Gutiérrez-Aguilar et al., 2016; Prieto et al., 2016). However, to generate relevant findings, situational variables must be included (Fernandez-Navarro et al., 2020; Gómez, Lago-Peñas et al., 2015). By integrating situational variables like scoring streak and player difference and higher-order interactions, this study aims to identify specific game situations where TTOs are most effective. Based on 850 games of the German Handball Bundesliga, game situations are identified by Classification Tree Analysis and efficacies are evaluated. Findings indicate a strong impact of timing. Frequently used TTOs, e.g., at the end of periods, are beneficial to the teams. However, strongest effect occurs for TTOs taken at the early stages of the game and with a positive run. Results indicate that TTO is a powerful tactical tool and an application at uncommon timings may even enhance the success rate.

Data Analytics Determines Co-occurrence of Odorants in Raw Water and Evaluates Drinking Water Treatment Removal Strategies

A complex dataset with 140 sampling events was generated using triple quadrupole gas chromatography-mass spectrometer to track the occurrence of 95 odorants in raw and finished water from 98 drinking water treatment plants in 31 cities across China. Data analysis identified more than 70 odorants with concentrations ranging from not detected to thousands of ng/L. In raw water, Pearson correlation analysis determined that thioethers, non-oxygen benzene-containing compounds, and pyrazines were classes of chemicals that co-occurred, and geosmin and p(m)-cresol, as well as cyclohexanone and benzaldehyde, also co-occurred, indicating similar natural or industrial sources. Based on classification and regression tree analysis, total dissolved organic carbon and geographical location were identified as major factors affecting the occurrence of thioethers. Indoles, phenols, and thioethers were well-removed through conventional and advanced treatment processes, while some aldehydes could be generated. For other odorants, higher removal was achieved by ozonation-biological activated carbon (39.3%) compared to the conventional treatment process (14.5%). To our knowledge, this is the first study to systematically identify the major odorants in raw water and determine suitable treatment strategies to control their occurrence by applying data analytics and statistical methods to the complex dataset. These provide informative reference for odor control and water quality management in drinking water industry.

Multi-Statistical Approach for the Study of Volatile Compounds of Industrial Spoiled Manzanilla Spanish-Style Table Olive Fermentations

Table olives can suffer different types of spoilage during fermentation. In this work, a multi-statistical approach (standard and compositional data analysis) was used for the study of the volatile organic compounds (VOCs) associated with altered (butyric, sulfidic, and putrid) and non-altered (normal) Manzanilla Spanish-style table olive fermentations. Samples were collected from two industrial fermentation yards in Seville (Spain) in the 2019/2020 season. The VOC profiles of altered (n = 4) and non-altered (n = 6) samples were obtained by headspace solid-phase microextraction combined with gas chromatography-mass spectrometry (HS-SPME-GC-MS). Ninety-one VOCs were identified and grouped into alcohols (30), esters (21), carbonyl compounds (12), acids (10), terpenes (6), phenols (6), sulfur compounds (2), and others (4). The association of the VOCs with spoilage samples depended on the standard or compositional statistical methodology used. However, butyric spoilage was strongly linked by several techniques to methyl butanoate, ethyl butanoate, and butanoic acid; sulfidic spoilage with 2-propyl-1-pentanol, dimethyl sulfide, methanol, 2-methylbutanal, 2-methyl-2-butenal, ethanol, 2-methyl-3-buten-2-ol, and isopentanol, while putrid was mainly related to D-limonene and 2-pentanol. Our data contribute to a better characterisation of non-zapatera spoiled table olive fermentations and show the convenience of using diverse statistical techniques for a most robust selection of spoilage VOC markers.

Identifying factors that conjointly influence nicotine vaping product relative harm perception among smokers and recent ex-smokers: Findings from the 2016 ITC Four Country Smoking and Vaping Survey

BACKGROUND

Use of nicotine vaping products (NVPs) to replace smoking is often influenced by perceived harmfulness of these products relative to smoking. This study aimed to identify factors that conjointly influenced NVP relative harm perception among smokers and ex-smokers.

METHODS

Data (n = 11,838) from adult smokers and ex-smokers (quit < 2 years) who participated in the 2016 ITC 4 Country Smoking and Vaping Surveys in Australia, Canada, England and the US were analyzed. Decision tree models were used to classify respondents into those who perceived vaping as less harmful than smoking ("correct" perception) versus otherwise ("incorrect" perception) based on their socio-demographic, smoking and vaping related variables.

RESULTS

Decision tree analysis identified nicotine replacement therapy (NRT) harmfulness perceptions relative to smoking, perceived vaping portrayal in the media and other sources as positive, negative or balanced, recency of seeking online vaping information, and age as the key variables that interacted conjointly to classify respondents into those with "correct" versus "incorrect" harm perceptions of vaping relative to smoking (model performance accuracy = 0.70-0.74). In all countries, NRT relative harmfulness perception and vaping portrayal perception were consistently the two most important classifying variables, with other variables showing some country differences.

CONCLUSIONS

In all four countries, perception of NVP relative harmfulness among smokers and recent ex-smokers is strongly influenced by a combination of NRT relative harmfulness perception and vaping portrayal in the media and other sources. These conjoint factors can serve as useful markers for identifying subgroups more vulnerable to misperception about NVP relative harmfulness to benefit from corrective intervention.

Canadian children's and youth's adherence to the 24-h movement guidelines during the COVID-19 pandemic: A decision tree analysis

PURPOSE

The purpose of this study was to use decision tree modeling to generate profiles of children and youth who were more or less likely to meet the Canadian 24-h movement guidelines during the coronavirus disease-19 (COVID-19) outbreak.

METHODS

Data for this study were from a nationally representative sample of 1472 Canadian parents (Mean_age = 45.12, SD = 7.55) of children (5-11 years old) or youth (12-17 years old). Data were collected in April 2020 via an online survey. Survey items assessed demographic, behavioral, social, micro-environmental, and macro-environmental characteristics. Four decision trees of adherence and non-adherence to all movement recommendations combined and each individual movement recommendation (physical activity [PA], screen time, and sleep) were generated.

RESULTS

Results revealed specific combinations of adherence and non-adherence characteristics. Characteristics associated with adherence to the recommendation(s) included high parental perceived capability to restrict screen time, annual household income of ≥ $100,000, increases in children's and youth's outdoor PA/sport since the COVID-19 outbreak began, being a boy, having parents younger than 43 years old, and small increases in children's and youth's sleep duration since the COVID-19 outbreak began. Characteristics associated with non-adherence to the recommendation(s) included low parental perceived capability to restrict screen time, youth aged 12-17 years, decreases in children's and youth's outdoor PA/sport since the COVID-19 outbreak began, primary residences located in all provinces except Quebec, low parental perceived capability to support children's and youth's sleep and PA, and annual household income of ≤ $99,999.

CONCLUSION

Our results show that specific characteristics interact to contribute to (non)adherence to the movement behavior recommendations. Results highlight the importance of targeting parents' perceived capability for the promotion of children's and youth's movement behaviors during challenging times of the COVID-19 pandemic, paying particular attention to enhancing parental perceived capability to restrict screen time.

Identification of production chain risk factors for slaughterhouse condemnation of broiler chickens'

Slaughterhouse condemnation of broiler chickens results from identification of polymorphic pathological conditions during meat inspection from arrival and on the slaughter line. While conditions that result in condemnation are multifactorial, identification of factors that are common for a number of categories could be valuable for developing strategies to reduce total condemnation. This study aimed to identify those condemnation categories that were most common in batches of broiler chickens and to determine and compare associated risk factors. In the first step, retrospective meat inspection records for 55,918 broiler batches from one large broiler integrator for 2015-2017 were used for association rules analysis. Results identified a network of nine associated condemnation categories: whole carcass condemnation for ascites, abnormal colour, perihepatitis, cellulitis, hard breast, tumours and dead on arrival, and liver only and heart only most often associated with hepatitis and pericarditis, respectively. Secondly, a longitudinal study collected data on 109 explanatory variables from broiler parental flocks to slaughterhouse characteristics between January 2015 and December 2017. Condemnation outcome data were obtained from meat inspection records for 539 broiler batches participating in the study. Parental flock-, rearing farm-, shed- and transport-level risk factors were assessed for each outcome using mixed-effects multivariable Poisson regression including shed and farm as random effects. A Poisson regression tree method was used as the first step to identify variables most relevant for analysis and comparison across the outcomes. No single production factor was associated with all nine of the condemnation outcomes investigated in this study, although some were shared across multiple outcomes: age of parental flock at time of lay, flock-level Campylobacter spp. frequency, broiler chick weight at seven days of age, weight at slaughter, type of broiler removal (i.e. thinning, final depopulation), catcher team, number of birds per transport crate, slaughterhouse shift number, and type of slaughterhouse line. Broiler chickens removed during final depopulation were at greatest risk of condemnation. Condemnation rates for cellulitis and tumours were found to be higher in broilers inspected by night shift at the slaughterhouse. Discovery of an apparent protective effect of a higher number of broilers per transport crate was unexpected. These findings provide information for the broiler industry on production chain factors that might be amenable to targeted intervention to improve future efforts for control of condemnation.

Dietary patterns and cancer risk

Over the past decade, the search for dietary factors on which to base cancer prevention guidelines has led to the rapid expansion of the field of dietary patterns and cancer. Multiple systematic reviews and meta-analyses have reported epidemiological associations between specific cancer types and both data-driven dietary patterns determined by empirical analyses and investigator-defined dietary indexes based on a predetermined set of dietary components. New developments, such as the use of metabolomics to identify objective biomarkers of dietary patterns and novel statistical techniques, could provide further insights into the links between diet and cancer risk. Although animal models of dietary patterns are limited, progress in this area could identify the potential mechanisms underlying the disease-specific associations observed in epidemiological studies. In this Review, we summarize the current state of the field, provide a critical appraisal of new developments and identify priority areas for future research. An underlying theme that emerges is that the effectiveness of different dietary pattern recommendations in reducing risk could depend on the type of cancer or on other risk factors such as family history, sex, age and other lifestyle factors or comorbidities as well as on metabolomic signatures or gut microbiota profiles.

Identifying At-Risk Subgroups for Acute Postsurgical Pain: A Classification Tree Analysis

Objective

Acute postsurgical pain is common and has potentially negative long-term consequences for patients. In this study, we evaluated effects of presurgery sociodemographics, pain experiences, psychological influences, and surgery-related variables on acute postsurgical pain using logistic regression vs classification tree analysis (CTA).

Design

The study design was prospective.

Setting

This study was carried out at Chongqing No. 9 hospital, Chongqing, China.

Subjects

Patients (175 women, 84 men) completed a self-report battery 24 hours before surgery (T1) and pain intensity ratings 48-72 hours after surgery (T2).

Results

An initial logistic regression analysis identified pain self-efficacy as the only presurgery predictor of postoperative pain intensity. Subsequently, a classification tree analysis (CTA) indicated that lower vs higher acute postoperative pain intensity levels were predicted not only by pain self-efficacy but also by its interaction with disease onset, pain catastrophizing, and body mass index. CTA results were replicated within a revised logistic regression model.

Conclusions

Together, these findings underscored the potential utility of CTA as a means of identifying patient subgroups with higher and lower risk for severe acute postoperative pain based on interacting characteristics.

Florida Populations Most at Risk of Not Being Up to Date With Colorectal Cancer Screening

INTRODUCTION

The purpose of this study was to examine the characteristics of populations at risk of not being up to date on colorectal cancer screening in Florida.

METHODS

We used Exhaustive Chi-squared Automatic Interaction Detection, a classification tree analysis, to identify subgroups not up to date with colorectal cancer screening using the 2013 Florida Behavioral Risk Factor Surveillance System. The data set was restricted to adults aged 50 to 75 years (n = 14,756).

RESULTS

Only 65.5% of the sample was up to date on colorectal cancer screening. Having no insurance and having a primary care provider were the most significant predictors of not being up to date on screening. The highest risk subgroups were 1) respondents with no insurance and no primary care provider, regardless of their employment status (screening rate, 12.1%-23.7%); 2) respondents with no insurance but had a primary care provider and were employed (screening rate, 32.3%); and 3) respondents with insurance, who were younger than 55 years, and who were current smokers (screening rate, 42.0%).

CONCLUSION

Some populations in Florida are at high risk for not being up to date on colorectal cancer screening. To achieve Healthy People 2020 goals, interventions may need to be further tailored to target these subgroups.

Risk factors for race-day fatality in flat racing Thoroughbreds in Great Britain (2000 to 2013)

A key focus of the racing industry is to reduce the number of race-day events where horses die suddenly or are euthanased due to catastrophic injury. The objective of this study was therefore to determine risk factors for race-day fatalities in Thoroughbred racehorses, using a cohort of all horses participating in flat racing in Great Britain between 2000 and 2013. Horse-, race- and course-level data were collected and combined with all race-day fatalities, recorded by racecourse veterinarians in a central database. Associations between exposure variables and fatality were assessed using logistic regression analyses for (1) all starts in the dataset and (2) starts made on turf surfaces only. There were 806,764 starts in total, of which 548,571 were on turf surfaces. A total of 610 fatalities were recorded; 377 (61.8%) on turf. In both regression models, increased firmness of the going, increasing racing distance, increasing average horse performance, first year of racing and wearing eye cover for the first time all increased the odds of fatality. Generally, the odds of fatality also increased with increasing horse age whereas increasing number of previous starts reduced fatality odds. In the 'all starts' model, horses racing in an auction race were at 1.46 (95% confidence interval (CI) 1.06-2.01) times the odds of fatality compared with horses not racing in this race type. In the turf starts model, horses racing in Group 1 races were at 3.19 (95% CI 1.71-5.93) times the odds of fatality compared with horses not racing in this race type. Identification of novel risk factors including wearing eye cover and race type will help to inform strategies to further reduce the rate of fatality in flat racing horses, enhancing horse and jockey welfare and safety.

Why are some groups physically active and others not? A contrast group analysis in leisure settings

BACKGROUND

This field study aims to investigate the determinants of physical activity of particularly active and inactive groups in their leisure environments. In order to consider the context in which physical activity occurs and to investigate whether cultural settings may influence physical activity, we conducted the study at pools in different cultural environments - Hawai'i and Germany.

METHODS

This study presents the quantitative data of a systematic (covert) participant observation. We recorded the physical activity of face-to-face interacting groups and analysed categories such as group size, estimated age of the group members, and verbal communication patterns. Total observation period was eight and a half months. In total, we observed 907 groups with the groups' size varying between 2 and 8 members. For the general statistics, we accessed the significance of differences regarding the degree of physical activity dependent on the target variables. To better understand activity promoting and hindering mechanisms, special attention is given to the identification of contrasting factors that characterise groups which are very active or very inactive. For this, we conducted a classification tree analysis.

RESULTS

General statistical analysis shows that, overall, the most differentiating factor regarding physical activity was the body shape of the group members. While obese groups had the lowest average activity level, groups mainly consisting of people with an athletic body shape were the most physically active. Yet, classification tree analysis reveals that obesity itself does not necessarily determine physical inactivity levels. The identification of six contrasting clusters highlights that besides the body shape several factors interact regarding a group's physical level. Such interacting factors were for example the degree of communication within the group, the gender- and age-related composition of the group, but also the equipment that had been brought to the beach/pool. Obese people were particularly inactive when they were members of frequently communicating, age-heterogeneous groups.

CONCLUSIONS

Our study shows that several social factors determine the physical activity of very active and very inactive groups. In order to promote physical activity, future health initiatives should target these factors of a person's network.

External validation of risk prediction models for incident colorectal cancer using UK Biobank

BACKGROUND

This study aimed to compare and externally validate risk scores developed to predict incident colorectal cancer (CRC) that include variables routinely available or easily obtainable via self-completed questionnaire.

METHODS

External validation of fourteen risk models from a previous systematic review in 373 112 men and women within the UK Biobank cohort with 5-year follow-up, no prior history of CRC and data for incidence of CRC through linkage to national cancer registries.

RESULTS

There were 1719 (0.46%) cases of incident CRC. The performance of the risk models varied substantially. In men, the QCancer10 model and models by Tao, Driver and Ma all had an area under the receiver operating characteristic curve (AUC) between 0.67 and 0.70. Discrimination was lower in women: the QCancer10, Wells, Tao, Guesmi and Ma models were the best performing with AUCs between 0.63 and 0.66. Assessment of calibration was possible for six models in men and women. All would require country-specific recalibration if estimates of absolute risks were to be given to individuals.

CONCLUSIONS

Several risk models based on easily obtainable data have relatively good discrimination in a UK population. Modelling studies are now required to estimate the potential health benefits and cost-effectiveness of implementing stratified risk-based CRC screening.

Heterogeneities in utilization of antenatal care in Uttar Pradesh, India: the need to contextualize interventions to individual contexts

BACKGROUND

This paper explores the heterogeneities in antenatal care (ANC) utilization in India's most populated state, Uttar Pradesh. Taking an intersectionality lens, multiple individual- and district-level factors are used to identify segments of any antenatal care usage in Uttar Pradesh Objective: This paper seeks to understand the multilevel contexts of ANC utilization. The planning and programming challenge is that such knowledge of contextual specificity is rarely known upfront at the initial stages of planning or implementing an intervention. Exploratory data analysis might be needed to identify such contextual specificity.

METHODS

Tree-structured regression methods are used to identify segments and interactions between factors. The results from the tree-structured regression were complemented with multilevel models that controlled for the clustering of individuals within districts.

RESULTS

Heterogeneities in utilization of any ANC were observed. The multiple segments of ANC utilization that were developed went from a low utilization of 23.7% for respondents who were not literate and did not have home ownership to a high of 82.4% for respondents who were literate and at the highest level of wealth. Key variables that helped define the segments of ANC utilization include: woman's literacy, ownership of home, wealth index, and district-level sex ratio. Based on the multilevel model of any ANC utilization, cross-level interactions also were obtained between sex ratio and ownership of home as well as between sex ratio and literacy. Increases in sex ratio increased the influence of ownership of home on any ANC, while increases in sex ratio reduced the impact of woman's literacy on receiving any ANC.

CONCLUSION

We argue that a focus on heterogeneous segments of utilization can help build knowledge of the mechanisms that underlie inequities in maternal health utilization. Such knowledge of heterogeneity needs to be incorporated in contextualizing interventions to meet a variety of recipients' needs.

Risk factors for race-day fatality, distal limb fracture and epistaxis in Thoroughbreds racing on all-weather surfaces in Great Britain (2000 to 2013)

The incidence of race-day injuries in Great Britain (GB) is higher on all-weather (AW) surfaces than on turf. However, to date no studies have focused on identifying risk factors for injury specific to AW racing. Therefore, the objective of the current study was to determine risk factors for fatality, distal limb fracture (DLF) and episodes of epistaxis in flat racing Thoroughbreds racing on AW surfaces in GB. Data included all flat racing starts on AW surfaces (n=258,193) and race-day veterinary events recorded between 2000 and 2013. Information on additional course-level variables was gathered during face-to-face interviews with racecourse clerks. Horse-, race- and course-level risk factors for each outcome were assessed using mixed-effects multivariable logistic regression including horse as a random effect. A classification tree method was used to identify potential interaction terms for inclusion in the models. During the study period, there were 233 fatalities resulting in a fatality incidence of 0.90 per 1000 starts; 245 DLF with a resultant DLF incidence of 0.95 per 1000 starts and 410 episodes of epistaxis resulting in an epistaxis incidence of 1.59 per 1000 starts. Risk factors varied for each outcome, although some factors were similar across models including the going, racing intensity, horse age, age at first race start, horse and trainer performance variables. Generally, older horses and those that had started racing at an older age were at higher risk of an adverse outcome, albeit with an interaction between the two variables in the fatality model. Faster going increased the odds of epistaxis and DLF but not fatality. Increasing race distance increased the odds of fatality but reduced the odds of epistaxis. Epistaxis was associated with type of AW surface (Fibresand versus Polytrack^®), but DLF and fatality were not. This study provides further evidence of the association between the risk of race-day injuries and fatalities and current age, age at first start, race distance, going and horse performance. These findings provide the racing industry with information to develop strategies to reduce the occurrence of race-day events on AW surfaces.

Identification of immune correlates of protection in Shigella infection by application of machine learning

BACKGROUND

Immunologic correlates of protection are important in vaccine development because they give insight into mechanisms of protection, assist in the identification of promising vaccine candidates, and serve as endpoints in bridging clinical vaccine studies. Our goal is the development of a methodology to identify immunologic correlates of protection using the Shigella challenge as a model.

METHODS

The proposed methodology utilizes the Random Forests (RF) machine learning algorithm as well as Classification and Regression Trees (CART) to detect immune markers that predict protection, identify interactions between variables, and define optimal cutoffs. Logistic regression modeling is applied to estimate the probability of protection and the confidence interval (CI) for such a probability is computed by bootstrapping the logistic regression models.

RESULTS

The results demonstrate that the combination of Classification and Regression Trees and Random Forests complements the standard logistic regression and uncovers subtle immune interactions. Specific levels of immunoglobulin IgG antibody in blood on the day of challenge predicted protection in 75% (95% CI 67-86). Of those subjects that did not have blood IgG at or above a defined threshold, 100% were protected if they had IgA antibody secreting cells above a defined threshold. Comparison with the results obtained by applying only logistic regression modeling with standard Akaike Information Criterion for model selection shows the usefulness of the proposed method.

CONCLUSION

Given the complexity of the immune system, the use of machine learning methods may enhance traditional statistical approaches. When applied together, they offer a novel way to quantify important immune correlates of protection that may help the development of vaccines.

A cross-sectional study for predicting tail biting risk in pig farms using classification and regression tree analysis

Tail biting in pigs has been an identified behavioural, welfare and economic problem for decades, and requires appropriate but sometimes difficult on-farm interventions. The aim of the paper is to introduce the Classification and Regression Tree (CRT) methodologies to develop a tool for prevention of acute tail biting lesions in pigs on-farm. A sample of 60 commercial farms rearing heavy pigs were involved; an on-farm visit and an interview with the farmer collected data on general management, herd health, disease prevention, climate control, feeding and production traits. Results suggest a value for the CRT analysis in managing the risk factors behind tail biting on a farm-specific level, showing 86.7% sensitivity for the Classification Tree and a correlation of 0.7 between observed and predicted prevalence of tail biting obtained with the Regression Tree. CRT analysis showed five main variables (stocking density, ammonia levels, number of pigs per stockman, type of floor and timeliness in feed supply) as critical predictors of acute tail biting lesions, which demonstrate different importance in different farms subgroups. The model might have reliable and practical applications for the support and implementation of tail biting prevention interventions, especially in case of subgroups of pigs with higher risk, helping farmers and veterinarians to assess the risk in their own farm and to manage their predisposing variables in order to reduce acute tail biting lesions.

Testing for interactive and non-linear effects of risk factors for binge eating and purging eating disorders

Almost no research has tested whether risk factors interact in the prediction of future eating disorder onset, which might suggest qualitatively distinct etiologic pathways. Accordingly, this prospective study tested for possible interactions between risk factors in the prediction of binge eating and purging eating disorders in adolescents. It also examined sex differences in pathways to risk. Two analytical approaches were used: (1) classification tree analysis (CTA), which is ideally suited to identifying non-linear interactions and the optimal cut-points for defining risk, with follow-up random forest analyses; and (2) two-way interaction terms in a series of logistic regression models. Data were drawn from the Western Australian Pregnancy Cohort (Raine) Study, a population-based study that followed participants from pre-birth to young adulthood. This study involved 1297 adolescents (49% male), 146 (11%) of whom developed bulimia nervosa, binge eating disorder or purging disorder in late adolescence. In CTA, sex was the first and most potent predictor of eating disorder risk with females showing a 5-fold increase in risk relative to males. For males and females, weight and eating concerns were the next most potent predictor of risk and three risk groups emerged, reflecting non-linear risk. For females with intermediate weight and eating concerns, externalizing problems emerged as an additional predictor. Interaction terms in logistic regression models did not produce significant results after correcting for multiple testing. Findings advance knowledge on risk pathways to eating disorder onset, highlight non-linear risk processes, and provide cut-points for prospectively identifying high-risk youth for prevention programs.

Analysis of risk factors and their interactions in type 2 diabetes mellitus: A cross-sectional survey in Guilin, China

Aims/Introduction

Type 2 diabetes is a metabolic disease characterized by insulin resistance, and is associated with the effects of genetic and environmental factors. The present study aimed to not only analyze the influence of a single factor for type 2 diabetes, but also to investigate the interaction effects between risk factors.

Materials and Methods

A total of 6,660 individuals selected by the method of cluster random sampling accepted a cross‐sectional survey (questionnaire investigation, physical measurement, laboratory examination and liver ultrasound examination). The classification tree was used to analyze the risk factors and their interactions in type 2 diabetes. The clinical and metabolic characteristics were compared between type 2 diabetes patients and controls, and the non‐conditional logistic regression model was used to quantitatively analyze the interactions.

Results

A total of 338 participants were classified as type 2 diabetes (217 men and 121 women), the classification tree model showed three variables with close associations with type 2 diabetes: age, triglycerides (TG) and non‐alcoholic fatty liver disease (NAFLD). Type 2 diabetes patients had higher age and incidences of high TG, NAFLD, hypertension, high body mass index, high uric acid, high total cholesterol, high low‐density lipoprotein cholesterol and low high‐density lipoprotein cholesterol. The multivariate logistic regression analysis showed that the following factors had interactions in type2 diabetes: high TG × advanced age (odds ratio 2.499, 95% confidence interval 1.868–3.344, P = 0.000), NAFLD × advanced age (odds ratio 1.250, 95% confidence interval 1.048–1.491, P = 0.013) and NAFLD × high TG (odds ratio 1.349, 95% confidence interval 1.144–1.590, P = 0.000).

Conclusions

The present study showed that type 2 diabetes resulted from the interactions of many factors; the interactions among age, TG and NAFLD are important risk factors for type 2 diabetes.

Application of classification trees in determining the impact of phenotypic factors on conception to first service in Holstein cattle

The aim of the study was to evaluate the impact of phenotypic factors on the probability of conception to first service. The study analysed 49 317 lactation records gathered between 1995 and 2007 in Iran, detailing six consecutive lactations in 13 herds of Holstein cows. The impact of phenotypic factors on the conception to first service was evaluated using the classification tree technique. The CHAID algorithm with the Pearson Chi-squared test was used as the division criterion. The resulting tree was six levels deep and consisted of 22 leaves. The greatest divisions in the tree occurred based on the year of calving (five divisions), herd and parity (four divisions each), level of milk production and age at calving (three divisions each). Fewer divisions occurred based on the month when the first insemination was performed (two divisions). The tree structure obtained in the present study allowed identification of combinations of levels of factors associated with highest probabilities of conception. The highest rate of conception to first service was recorded for cows that: while calving were younger than 5 years old; were in herds 2, 4, 5, 7, 8, 12 and 14; and were inseminated in January, February, March, April, May and November. Those cows were at least in their third lactation.

Risk Prediction Models for Colorectal Cancer: A Systematic Review

Colorectal cancer is the second leading cause of cancer-related death in Europe and the United States. Survival is strongly related to stage at diagnosis and population-based screening reduces colorectal cancer incidence and mortality. Stratifying the population by risk offers the potential to improve the efficiency of screening. In this systematic review we searched Medline, EMBASE, and the Cochrane Library for primary research studies reporting or validating models to predict future risk of primary colorectal cancer for asymptomatic individuals. A total of 12,808 papers were identified from the literature search and nine through citation searching. Fifty-two risk models were included. Where reported (n = 37), half the models had acceptable-to-good discrimination (the area under the receiver operating characteristic curve, AUROC >0.7) in the derivation sample. Calibration was less commonly assessed (n = 21), but overall acceptable. In external validation studies, 10 models showed acceptable discrimination (AUROC 0.71-0.78). These include two with only three variables (age, gender, and BMI; age, gender, and family history of colorectal cancer). A small number of prediction models developed from case-control studies of genetic biomarkers also show some promise but require further external validation using population-based samples. Further research should focus on the feasibility and impact of incorporating such models into stratified screening programmes.

Obesity and cancer: at the crossroads of cellular metabolism and proliferation

Obesity is associated with an increased risk of cancer. The mechanisms underlying this association include but are not limited to increased systemic inflammation, an anabolic hormonal milieu, and adipocyte-cancer crosstalk, aberrant stimuli that conspire to promote neoplastic transformation. Cellular proliferation is uncoupled from nutrient availability in malignant cells, promoting tumor progression. Elucidation of the mechanisms underlying the obesity-cancer connection will lead to the development of novel metabolism-based agents for cancer prevention and treatment.

Health in elite sports from a salutogenetic perspective: athletes' sense of coherence

OBJECTIVE

Considering the high number of stressors encountered in the context of elite sports, a high sense of coherence (SOC) is crucial to allow athletes to maintain their health from both short- and long-term perspectives. The aim of this cross-sectional study was to investigate SOC in a population of elite athletes, focusing on identification of subsets of athletes with particularly high and low SOC scores, and any related predictors. The elite athletes' SOC scores were also evaluated for differences with those of the general population of Germany; whether a correlation between SOC and subjective health existed was additionally examined.

METHOD

In total, 698 male and female elite athletes, drawn from Germany's highest-level national track and field squads, and first and second division handball teams, completed a survey that included the SOC-L9 Scale and measures of subjective health, sociodemographic information, and the number of injury lay-offs experienced during the athletes' careers to date.

RESULTS

Classification tree analysis reveals six contrast groups with varying SOC scores. Several interacting factors determine the group to which an athlete belongs. Together with overuse injuries, additional factors are age, gender, and completed/not completed apprenticeship/degree. Female athletes aged between 19 and 25, who had already been subject to lay-offs due to overuse injuries, comprise the group with the lowest SOC scores. Overall, the SOC of elite athletes is slightly lower than in the general population. In accordance with other studies, a stronger SOC is also correlated significantly with better global subjective health.

CONCLUSION

The identification of contrast groups with varying SOC scores contributes to the development of more targeted salutogenetic health promotion programs. Such programs would ideally include learning modules pertaining to coping with overuse injuries, as well as social support systems aiming to effectively combine education and elite sport.

Classification tree analysis of race-specific subgroups at risk for a central venous catheter-related bloodstream infection

BACKGROUND

Studies of racial disparities in patient safety events often do not use race-specific risk adjustment and do not account for reciprocal covariate interactions. These limitations were addressed by using classification tree analysis separately for black patients and white patients to identify characteristics that segment patients who have increased risks for a venous catheter-related bloodstream infection.

METHODS

A retrospective, cross-sectional analysis of 5,236,045 discharges from 103 Florida acute hospitals in 2005-2009 was conducted. Hospitals were rank ordered on the basis of the black/white Patient Safety Indicator (PSI) 7 rate ratio as follows: Group 1 (white rate higher), Group 2, (equivalent rates), Group 3, (black rate higher), and Group 4, (black rate highest). Predictor variables included 26 comorbidities (Elixhauser Comorbidity Index) and demographic characteristics. Four separate classification tree analyses were completed for each race/hospital group.

RESULTS

Individual characteristics and groups of characteristics associated with increased PSI 7 risk differed for black and white patients. The average age for both races was different across the hospital groups (p < .01). Weight loss was the strongest single delineator and common to both races. The black subgroups with the highest PSI 7 risk were Medicare beneficiaries who were either < or = 25.5 years without hypertension or < or = 39.5 years without hypertension but with an emergency or trauma admission. The white subgroup with the highest PSI 7 risk consisted of patients < or = 45.5 years who had congestive heart failure but did not have either hypertension or weight loss.

DISCUSSION

Identifying subgroups of patients at risk for a rare safety event such as PSI 7 should aid effective clinical decisions and efficient use of resources and help to guide patient safety interventions.

Classification and regression trees for epidemiologic research: an air pollution example

BACKGROUND

Identifying and characterizing how mixtures of exposures are associated with health endpoints is challenging. We demonstrate how classification and regression trees can be used to generate hypotheses regarding joint effects from exposure mixtures.

METHODS

We illustrate the approach by investigating the joint effects of CO, NO2, O3, and PM2.5 on emergency department visits for pediatric asthma in Atlanta, Georgia. Pollutant concentrations were categorized as quartiles. Days when all pollutants were in the lowest quartile were held out as the referent group (n = 131) and the remaining 3,879 days were used to estimate the regression tree. Pollutants were parameterized as dichotomous variables representing each ordinal split of the quartiles (e.g. comparing CO quartile 1 vs. CO quartiles 2-4) and considered one at a time in a Poisson case-crossover model with control for confounding. The pollutant-split resulting in the smallest P-value was selected as the first split and the dataset was partitioned accordingly. This process repeated for each subset of the data until the P-values for the remaining splits were not below a given alpha, resulting in the formation of a "terminal node". We used the case-crossover model to estimate the adjusted risk ratio for each terminal node compared to the referent group, as well as the likelihood ratio test for the inclusion of the terminal nodes in the final model.

RESULTS

The largest risk ratio corresponded to days when PM2.5 was in the highest quartile and NO2 was in the lowest two quartiles (RR: 1.10, 95% CI: 1.05, 1.16). A simultaneous Wald test for the inclusion of all terminal nodes in the model was significant, with a chi-square statistic of 34.3 (p = 0.001, with 13 degrees of freedom).

CONCLUSIONS

Regression trees can be used to hypothesize about joint effects of exposure mixtures and may be particularly useful in the field of air pollution epidemiology for gaining a better understanding of complex multipollutant exposures.

Diet and lifestyle factors and risk of subtypes of esophageal and gastric cancers: classification tree analysis

PURPOSE

Although risk factors for squamous cell carcinoma of the esophagus and adenocarcinomas of the esophagus (EA), gastric cardia (GC), and other (noncardia) gastric (OG) sites have been identified, little is known about interactions among risk factors. We sought to examine interactions of diet, other lifestyle, and medical factors with risks of subtypes of esophageal and gastric cancers.

METHODS

We used classification tree analysis to analyze data from a population-based case-control study (1095 cases, 687 controls) conducted in Connecticut, New Jersey, and western Washington State.

RESULTS

Frequency of reported gastroesophageal reflux disease symptoms was the most important risk stratification factor for EA, GC, and OG, with dietary factors (EA, OG), smoking (EA, GC), wine intake (GC, OG), age (OG), and income (OG) appearing to modify the risk of these cancer sites. For esophageal squamous cell carcinoma, smoking was the most important risk stratification factor, with gastroesophageal reflux disease, income, race, noncitrus fruit, and energy intakes further modifying risk.

CONCLUSION

Various combinations of risk factors appear to interact to affect risk of each cancer subtype. Replication of these data mining analyses are required before suggesting causal pathways; however, the classification tree results are useful in partitioning risk and mapping multilevel interactions among risk variables.

Diagnostic cellular abnormalities in neoplastic and non-neoplastic lesions of the epidermis: a morphological and statistical study

BACKGROUND

Distinguishing cellular abnormalities in reactive and malignant lesions is challenging. We compared the incidence and severity of cytological abnormalities in malignant/premalignant and benign epidermal lesions.

METHODS

One hundred fifty-two biopsies representing 69 malignant/premalignant squamous lesions and 83 benign conditions were studied. Cytological features, including nuclear hyperchromasia, nuclear overlap (crowding), irregular nuclei, high nuclear/cytoplasmic (N/C) ratio, conspicuous nucleoli, delicate inconspicuous nucleoli, clumped chromatin, pleomorphic parakeratosis, normal and abnormal mitotic figures and necrotic keratinocytes, were evaluated and graded. Statistical analysis was performed.

RESULTS

Irregular nuclei, increased N/C ratio, conspicuous single prominent nucleoli, nuclear overlap (crowding), pleomorphic parakeratosis, nuclear hyperchromasia, necrotic keratinocytes, normal and abnormal mitotic figures and coarse chromatin were seen more frequently in malignant neoplasms (p < 0.05). Abnormal mitotic figures, although uncommon (20.3%), were only noted in the malignant/premalignant group. Certain cytological features were common among both malignant and benign lesions, suggesting that they are of little value.

CONCLUSION

In the setting of an atypical cutaneous squamous proliferation, nuclear irregularity, increased N/C ratio, conspicuous nucleoli, crowding and hyperchromasia are the most useful indicators of malignancy. In contrast, mitotic figures, necrotic cells and coarse chromatin are less useful. The presence of abnormal mitotic figures is very helpful when present; however, their overall rarity limits their utility.

Evaluation of methodologies for assessing the overall diet: dietary quality scores and dietary pattern analysis

This paper aims to describe different approaches for studying the overall diet with advantages and limitations. Studies of the overall diet have emerged because the relationship between dietary intake and health is very complex with all kinds of interactions. These cannot be captured well by studying single dietary components. Three main approaches to study the overall diet can be distinguished. The first method is researcher-defined scores or indices of diet quality. These are usually based on guidelines for a healthy diet or on diets known to be healthy. The second approach, using principal component or cluster analysis, is driven by the underlying dietary data. In principal component analysis, scales are derived based on the underlying relationships between food groups, whereas in cluster analysis, subgroups of the population are created with people that cluster together based on their dietary intake. A third approach includes methods that are driven by a combination of biological pathways and the underlying dietary data. Reduced rank regression defines linear combinations of food intakes that maximally explain nutrient intakes or intermediate markers of disease. Decision tree analysis identifies subgroups of a population whose members share dietary characteristics that influence (intermediate markers of) disease. It is concluded that all approaches have advantages and limitations and essentially answer different questions. The third approach is still more in an exploration phase, but seems to have great potential with complementary value. More insight into the utility of conducting studies on the overall diet can be gained if more attention is given to methodological issues.

The development of a population-level clinical screening tool for self-harm repetition and suicide: the ReACT Self-Harm Rule

BACKGROUND

Self-harm is a common reason for Emergency Department (ED) attendance. We aimed to develop a clinical tool to help identify patients at higher risk of repeat self-harm, or suicide, within 6 months of an ED self-harm presentation.

METHOD

The tool, the ReACT Self-Harm Rule, was derived using multicentre data from a prospective cohort study. Binary recursive partitioning was applied to data from two centres, and data from a separate centre were used to test the tool. There were 29 571 self-harm presentations to five hospital EDs between January 2003 and June 2007, involving 18 680 adults aged ⩾16 years. We estimated sensitivity, specificity and positive and negative predictive values to measure the performance of the tool.

RESULTS

A self-harm presentation was classified as higher risk if at least one of the following factors was present: recent self-harm (in the past year), living alone or homelessness, cutting as a method of harm and treatment for a current psychiatric disorder. The rule performed with 95% sensitivity [95% confidence interval (CI) 94-95] and 21% specificity (95% CI 21-22), and had a positive predictive value of 30% (95% CI 30-31) and a negative predictive value of 91% (95% CI 90-92) in the derivation centres; it identified 83/92 of all subsequent suicides.

CONCLUSIONS

The ReACT Self-Harm Rule might be used as a screening tool to inform the process of assessing self-harm presentations to ED. The four risk factors could also be used as an adjunct to in-depth psychosocial assessment to help guide risk formulation. The use of multicentre data helped to maximize the generalizability of the tool, but we need to further verify its external validity in other localities.

Collaborative evaluation and management of students' health-related physical fitness: applications of cluster analysis and the classification tree

Health-related physical fitness has decreased with age; this is od immense concern to adolescents. School-based health intervention programs can be classified as either population-wide or high-risk approach. Although the population-wide and risk-based approaches adopt different healthcare angles, they all need to focus resources on risk evaluation. In this paper, we describe an exploratory application of cluster analysis and the tree model to collaborative evaluation of students' health- related physical fitness from a high school sample in Taiwan (n=742). Cluster analysis show that physical fitness can be divided into relatively good, moderate and poor subgroups. There are significant differences in biochemical measurements among these three groups. For the tree model, we used 2004 school-year students as an experimental group and 2005 school-year students as a validation group. The results indicate that if sit-and-reach is shorter than 33 cm, BMI is >25.46 kg/m2, and 1600 m run/walk is >534 s, the predicted probability for the number of metabolic risk factors ≥2 is 100% and the population is 41, both results are the highest. From the risk-based healthcare viewpoint, the cluster analysis can sort out students' physical fitness data in a short time and then narrow down the scope to recognize the subgroups. A classification tree model specifically shows the discrimination paths between the measurements of physical fitness for metabolic risk and would be helpful for self-management or proper healthcare education targeting different groups. Applying both methods to specific adolescents' health issues could provide different angles in planning health promotion projects.

Applying multilevel model to the relationship of dietary patterns and colorectal cancer: an ongoing case-control study in Córdoba, Argentina

PURPOSES

Scientific literature has consistently shown the effects of certain diets on health but regional variations of dietary habits, and their relationship colorectal cancer (CRC) has been poorly studied in Argentina. Our aims were to identify dietary patterns and estimate their effect on CRC occurrence and to quantify the association between family history of CRC and CRC occurrence by applying multilevel models to estimate and interpret measures of variation.

METHODS

Principal components factor analysis was performed to identify dietary patterns that were then used in a multilevel logistic regression applied to an ongoing case-control data about dietary exposure and CRC occurrence taking into account familiar clustering.

RESULTS

Three dietary patterns were identified: "Southern Cone pattern" (red meat, wine, and starchy vegetables), "High-sugar drinks pattern", and "Prudent pattern". The study considered 41 cases and 95 controls. There was a significant promoting effects on CRC of "Southern Cone" (OR 1.5, 95%CI 1.0-2.2) and "High-sugar drinks" (OR 3.8, 95%CI 2.0-7.1) patterns, whereas "Prudent pattern" (OR 0.3, 95%CI 0.2-0.4) showed a significant protective effect at third tertile level. BMI, use of NSAIDs, and to have medical insurance showed significant effects. Variance of the random effect of family history of CRC was highly significant.

CONCLUSIONS

This novel approach for Argentina showed that Southern Cone and High-sugar drinks patterns were associated with a higher risk of CRC, whereas the Prudent pattern showed a protective effect. There was a significant clustering effect of family history of CRC.

Variation in the CYP19A1 gene and risk of colon and rectal cancer

CYP19A1, or aromatase, influences estrogen-metabolizing enzymes and may influence cancer risk. We examine variation in the CYP19A1 gene and risk of colorectal cancer using data from population-based case-control studies (colon n = 1,574 cases, 1,970 controls; rectal n = 791 cases, 999 controls). Four SNPs were statistically significantly associated with colon cancer and four were associated with rectal cancer. After adjustment for multiple comparisons, the AA genotype of rs12591359 was associated with an increased risk of colon cancer (OR 1.44 95% CI 1.16-1.80) and the AA genotype of rs2470144 was associated with a reduced risk of rectal cancer (OR 0.65 95% CI 0.50-0.84). Variants of CYP19A1 were associated with CIMP+ and CIMP+/KRAS2-mutated tumors. CT/TT genotypes of rs1961177 were significantly associated with an increased likelihood of a MSI+ colon tumor (OR 1.77 95% CI 1.26-2.37). We observed statistically significant interactions between genetic variation in NFκB1 and CYP19A1 for both colon and rectal cancer. Our data suggest the importance of CYP19A1 in the development of colon and rectal cancer and that estrogen may influence risk through an inflammation-related mechanism.

Calcium, vitamin D, VDR genotypes, and epigenetic and genetic changes in rectal tumors

Calcium, vitamin D, exposure to sunshine, and vitamin D receptor (VDR) genotypes have been associated rectal cancer. We used data from 750 rectal tumors and 1,205 population-based controls examine associations with TP53, KRAS2, and CpG Island methylator phenotype (CIMP) markers. Rectal tumors were associated with high levels of calcium overall and with TP53 tumor mutations specifically (OR = 0.6, 95% CI = 0.42-0.84). High levels of sunshine exposure had a borderline protective effect for TP53 tumor mutations (OR = 0.78, 95% CI = 0.59-1.03). A mutation at codon 248 was significantly associated with dietary calcium intake (OR = 0.26, 95% CI = 0.09-0.77); having the Ff/ff genotypes of the FOK1 VDR polymorphism significantly increased the odds of a mutation at codon 245 (OR = 4.74, 95% CI = 1.05-21.39); high levels of dietary vitamin D (OR = 3.42, 95% CI = 1.15-10.17) and the Ff/ff genotypes of FOK1 (OR = 3.34, 95% CI = 1.11-10.02) and the GA/AA genotypes of the CDX2 VDR polymorphism (OR = 2.85, 95% CI = 1.23-6.58) increased the odds of a TP53 mutation at codon 273. These data support an association between calcium and rectal tumors overall as well as specifically with TP53 mutations. However, given the number of comparisons, findings need to be confirmed in other studies.

Genetic variation in a metabolic signaling pathway and colon and rectal cancer risk: mTOR, PTEN, STK11, RPKAA1, PRKAG2, TSC1, TSC2, PI3K and Akt1

Serine/threonine protein kinase 11 (STK11) and phosphatase tensin homolog deleted on chromosome 10 (PTEN) link insulin sensitivity and metabolic signaling to inflammation and other hormonal factors and colorectal cancer. We evaluate genetic variation in nine genes in a candidate pathway as follows: STK11 (3 tagSNPs), PTEN (9 tagSNPs), FRAP1 (mTOR) (4 tagSNPs), TSC1 (14 tagSNPs), TSC2 (8 tagSNPs), Akt1 (2 tagSNPs), PIK3CA (7 tagSNPs), PRKAA1 (13 tagSNPs) and PRKAG2 (68 tagSNPs) in two population-based case-control studies of colon (n = 1574 cases, 1940 controls) and rectal (n = 91 cases, 999 controls) cancer. FRAP1, PRKAA1, PRKAG2 and TSC2 genes were significantly associated with colon cancer; risk estimates ranged from 1.21 [95% confidence interval (CI) 1.05-1.38] for FRAP1rs1057079 for the AG/GG genotype to 1.51 (95% CI 1.09-2.09) for PRKAG2rs9648723 CC genotype. PIK3CA, PRKAG2, PTEN, STK11 and TSC1 were significantly associated with rectal cancer overall. The strongest association was observed for PIK3CA rs7651265 GG genotype (odds ratio 2.32 95% CI 1.02-5.30). FRAP1 was associated with microsatellite instability (MSI)+ colon tumors; PRKAA1, CpG island methylator phenotype (CIMP)+ and MSI+ colon tumors; PRKAG2 and KRAS2 colon tumors; TSC1 and CIMP+ and MSI+ colon tumors; TSC2 with MSI+ colon tumors; PIK3CA with KRAS2-mutated rectal tumors; PRKAG2 (rs6964824) with KRAS2- and TP53-mutated rectal tumors and with PRKAG2 (rs412396 and rs4725431) with CIMP+ rectal tumors. These data suggest that genetic variation in a predefined candidate pathway for colorectal cancer contributes to both colon and rectal cancer risk. Associations appear to be strongest for CIMP+ and MSI+ tumors.

Comparing 3 dietary pattern methods--cluster analysis, factor analysis, and index analysis--With colorectal cancer risk: The NIH-AARP Diet and Health Study

The authors compared dietary pattern methods-cluster analysis, factor analysis, and index analysis-with colorectal cancer risk in the National Institutes of Health (NIH)-AARP Diet and Health Study (n = 492,306). Data from a 124-item food frequency questionnaire (1995-1996) were used to identify 4 clusters for men (3 clusters for women), 3 factors, and 4 indexes. Comparisons were made with adjusted relative risks and 95% confidence intervals, distributions of individuals in clusters by quintile of factor and index scores, and health behavior characteristics. During 5 years of follow-up through 2000, 3,110 colorectal cancer cases were ascertained. In men, the vegetables and fruits cluster, the fruits and vegetables factor, the fat-reduced/diet foods factor, and all indexes were associated with reduced risk; the meat and potatoes factor was associated with increased risk. In women, reduced risk was found with the Healthy Eating Index-2005 and increased risk with the meat and potatoes factor. For men, beneficial health characteristics were seen with all fruit/vegetable patterns, diet foods patterns, and indexes, while poorer health characteristics were found with meat patterns. For women, findings were similar except that poorer health characteristics were seen with diet foods patterns. Similarities were found across methods, suggesting basic qualities of healthy diets. Nonetheless, findings vary because each method answers a different question.

Convergence of hormones, inflammation, and energy-related factors: a novel pathway of cancer etiology

Colorectal cancer (CRC) is a multifactorial disease with several hypothesized etiologic factors including inflammatory processes; hormones such as estrogen, androgen, and insulin; and energy-related factors. We present evidence that integrates these elements in a pathway we call the convergence of hormones, inflammation, and energy-related factors (CHIEF). First, given the physiology of the gut, substantial epidemiologic and molecular data support the hypothesis that activation of innate immunity in the normal gut mucosa by various environmental agents (commensal bacteria, dietary antigens, mucosal irritants, pathogens) and endogenous factors such as estrogen, androgens, and insulin levels provokes basal inflammation as an underlying factor of the association of insulin, estrogen, and energy-related factors with CRC. Second, critical genes involved in this pathway, e.g., phosphatase tensin homologue on chromosome 10 (PTEN) and serine threonine kinase 11 (STK11)/LKB1, are tumor suppressor genes often mutated in intestinal cancer or CRC. Third, laboratory experiments show that cellular PTEN and STK11/LKB1 tumor suppressor enzymes are vulnerable to inactivation by redox-active species, especially chemically reactive lipid mediators of inflammation and redox stress. Epidemiologic data further support the underlying proposal that CHIEF comprises important elements of CRC risk. Although this discussion of the CHIEF pathway focuses on CRC, we believe that this pathway may play an important role in the etiology of other cancers as well.