Paul H. Lee

Prevalence and impacts of poor sleep on quality of life and associated factors of good sleepers in a sample of older Chinese adults

BackgroundSleep disturbance is a complex health problem in ageing global populations decreasing quality of life among many older people. Geographic, cultural, and ethnic differences in sleep patterns have been documented within and between Western and Asian populations. The aim of this study was to explore sleep problems among Hong Kong seniors by examining the prevalence of poor sleep quality, the relationship between sleep quality and health-related quality of life, and associated factors of good sleepers in different age groups.MethodsThis cross-sectional study used convenience sampling and gathered data during face-to-face interviews. Older community-dwelling individuals (n = 301) were recruited in community centres in 2010. The Pittsburgh Sleep Quality Index and Medical Outcomes Study Short Form-36 were used to measure sleep quality and health-related quality of life. The Medical Outcomes Study Short Form-36 domain scores were compared between good and bad sleepers and between long and short sleepers using Hotelling’s T-Square test. SF-36 domain scores were placed into a logistic regression model that controlled for significant demographic variables (gender, educational level, perceived health).ResultsMost (77.7%) participants were poor sleepers. Participants who had global Pittsburgh Sleep Quality Index scores <5 and slept ≥5.5 h/night had better health-related quality of life. Vitality, emotional role, physical functioning, and bodily pain domain scores were associated factors of good sleepers in different age groups.ConclusionsThis study found a strong negative association between sleep deprivation (poor quality, short duration) and health-related quality of life. Associated factors for good sleep quality in later life differ among age groups in relation to universal age-related changes, and should be addressed by social policies and health-care programmes.

Is a Cutoff of 10% Appropriate for the Change-in-Estimate Criterion of Confounder Identification?

Background When using the change-in-estimate criterion, a cutoff of 10% is commonly used to identify confounders. However, the appropriateness of this cutoff has never been evaluated. This study investigated cutoffs required under different conditions. Methods Four simulations were performed to select cutoffs that achieved a significance level of 5% and a power of 80%, using linear regression and logistic regression. A total of 10 000 simulations were run to obtain the percentage differences of the 4 fitted regression coefficients (with and without adjustment). Results In linear regression, larger effect size, larger sample size, and lower standard deviation of the error term led to a lower cutoff point at a 5% significance level. In contrast, larger effect size and a lower exposure–confounder correlation led to a lower cutoff point at 80% power. In logistic regression, a lower odds ratio and larger sample size led to a lower cutoff point at a 5% significance level, while a lower odds ratio, larger sample size, and lower exposure–confounder correlation yielded a lower cutoff point at 80% power. Conclusions Cutoff points for the change-in-estimate criterion varied according to the effect size of the exposure–outcome relationship, sample size, standard deviation of the regression error, and exposure–confounder correlation.

Performance of the international physical activity questionnaire (short form) in subgroups of the Hong Kong chinese population

BackgroundThe International Physical Activity Questionnaire (IPAQ-SF) has been validated and recommended as an efficient method to assess physical activity, but its validity has not been investigated in different population subgroups. We examined variations in IPAQ validity in the Hong Kong Chinese population by six factors: sex, age, job status, educational level, body mass index (BMI), and visceral fat level (VFL).MethodsA total of 1,270 adults (aged 42.9 ± SD 14.4 years, 46.1% male) completed the Chinese version of IPAQ (IPAQ-C) and wore an accelerometer (ActiGraph) for four days afterwards. The IPAQ-C and the ActiGraph were compared in terms of estimated Metabolic Equivalent Task minutes per week (MET-min/wk), minutes spent in activity of moderate or vigorous intensity (MVPA), and agreement in the classification of physical activity.ResultsThe overall Spearman correlation (ρ) of between the IPAQ-C and ActiGraph was low (0.11 ± 0.03; range in subgroups 0.06-0.24) and was the highest among high VFL participants (0.24 ± 0.05). Difference between self-reported and ActiGraph-derived MET-min/wk (overall 2966 ± 140) was the smallest among participants with tertiary education (1804 ± 208). When physical activity was categorized into over or under 150 min/wk, overall agreement between self-report and accelerometer was 81.3% (± 1.1%; subgroup range: 77.2%-91.4%); agreement was the highest among those who were employed full-time in physically demanding jobs (91.4% ± 2.7%).ConclusionsSex, age, job status, educational level, and obesity were found to influence the criterion validity of IPAQ-C, yet none of the subgroups showed good validity (ρ = 0.06 to 0.24). IPAQ-SF validity is questionable in our Chinese population.

Distance-based tree models for ranking data

Ranking data has applications in different fields of studies, like marketing, psychology and politics. Over the years, many models for ranking data have been developed. Among them, distance-based ranking models, which originate from the classical rank correlations, postulate that the probability of observing a ranking of items depends on the distance between the observed ranking and a modal ranking. The closer to the modal ranking, the higher the ranking probability is. However, such a model basically assumes a homogeneous population and does not incorporate the presence of covariates. To overcome these limitations, we combine the strength of a tree model and the existing distance-based models to build a model that can handle more complexity and improve prediction accuracy. We will introduce a recursive partitioning algorithm for building a tree model with a distance-based ranking model fitted at each leaf. We will also consider new weighted distance measures which allow different weights for different ranks in formulating more flexible distance-based tree models. Finally, we will apply the proposed methodology to analyze a ranking dataset of Ingleharts items collected in the 1999 European Values Studies.

Data imputation for accelerometer-measured physical activity: the combined approach

BACKGROUND Accelerometers are gaining popularity for the assessment of the physical activity level; however, compliance is a problem that results in missing data. Data from study days in which the accelerometer is not worn for a number of hours that are sufficient to reach a predetermined cutoff value are considered invalid and discarded. The problem of missing data is commonly handled by imputation; however, all traditional imputation methods ignore the available information from invalid days. OBJECTIVE In this study, I propose a new approach to the imputation of missing accelerometer data that takes into account the data available from invalid days. DESIGN A total of 4069 participants in NHANES waves 2003-2004 and 2005-2006 who provided 7 d of valid accelerometer data were used to illustrate this new approach. The method of imputation was a combined approach that combined the available data from valid days and invalid days to impute missing values. Simulation studies were carried out to compare this new combined approach with the traditional imputation method for 1) accuracy and 2) effect-size estimation of the sex-physical activity relation by using the root mean squared error (RMSE). RESULTS The combined approach performed significantly better than traditional imputation method (all t tests P < 0.001), with the percentage reduction of the RMSE for accuracy and effect-size estimation that ranged from 12.4% to 17.3% and 19.8% to 32.9%, respectively. CONCLUSION The combined approach significantly outperforms the traditional imputation algorithm.

Depressive symptoms in people with chronic physical conditions: prevalence and risk factors in a Hong Kong community sample

BackgroundDepression is predicted to become one of the two most burdensome diseases worldwide by 2020 and is common in people with chronic physical conditions. However, depression is relatively uncommon in Asia. Family support is an important Asian cultural value that we hypothesized could protect people with chronic physical conditions from developing depression. We investigated depressive symptom prevalence and risk factors in a Chinese sample with chronic medical conditions, focusing on the possible protective role of family relationships.MethodsData were obtained from the Hong Kong Jockey Club FAMILY Project cohort study in 2009–2011, which included 6,195 participants (age ≥15) with self-reported chronic conditions. Depressive symptoms were recorded using the Patient Health Questionnaire-9 (PHQ-9). Demographic and lifestyle variables, stressful life events, perceived family support and neighborhood cohesion were assessed. Factors associated with a non-somatic (PHQ-6) depression score were also examined.ResultsThe prevalence of depressive symptoms (PHQ-9 scores ≥5) was 17% in those with one or more chronic conditions, and was more prevalent in women than in men (19.7% vs. 13.9%; p < 0.001). In multilevel analyses, life stress, number of chronic conditions and satisfaction with family support explained 43% of the variance in PHQ-9 scores (standardized regression coefficients of 0.46, 0.15, and −0.12 respectively, all p <0.001). Body mass index, problem alcohol drinking, physical activity, and unmarried status were significantly associated with PHQ-9 scores, although these associations were weak. Variables associated with depression explained 35% of the variance in non-somatic (PHQ-6) depression scores. Satisfaction with family support played a stronger protective role against depressive symptoms (both PHQ-9 and PHQ-6 scores) among women than men (p < 0.05).ConclusionsAcute life stress and the number of chronic conditions, together with socio-demographic factors, explain most variance in depressive symptoms among chronically ill Chinese individuals. Somatic items in the PHQ-9 increased the depression scores but they did not alter the pattern of predictors. Family support appears to be an important protective factor in Chinese cultures for individuals with chronic conditions.

Mixtures of weighted distance-based models for ranking data with applications in political studies

Analysis of ranking data is often required in various fields of study, for example politics, market research and psychology. Over the years, many statistical models for ranking data have been developed. Among them, distance-based ranking models postulate that the probability of observing a ranking of items depends on the distance between the observed ranking and a modal ranking. The closer to the modal ranking, the higher the ranking probability is. However, such a model assumes a homogeneous population, and the single dispersion parameter in the model may not be able to describe the data well. To overcome these limitations, we formulate more flexible models by considering the recently developed weighted distance-based models which can allow different weights for different ranks. The assumption of a homogeneous population can be relaxed by an extension to mixtures of weighted distance-based models. The properties of weighted distance-based models are also discussed. We carry out simulations to test the performance of our parameter estimation and model selection procedures. Finally, we apply the proposed methodology to analyze synthetic ranking datasets and a real world ranking dataset about political goals priority.

Hypoadiponectinemia As an Independent Predictor for the Progression of Carotid Atherosclerosis: A 5-Year Prospective Study

BACKGROUND Hypoadiponectinemia predicts the development of diabetes and hypertension, both being potent atherosclerotic risk factors. Whether adiponectin predicts the progression of early atherosclerosis remains unclear. In this 5-year prospective study, we examined the relationship between serum adiponectin and carotid intima media thickness (CIMT), a marker of subclinical atherosclerosis. METHODS A total of 265 subjects from the population-based Hong Kong Cardiovascular Risk Factor Prevalence Study, with no known cardiovascular disease, underwent CIMT measurement at baseline and at 5 years. RESULTS In all, 129 men and 136 women, aged 54.6±12.3 years, were studied. Median CIMT at baseline was 0.63 mm (interquartile range 0.52-0.73 mm) and increased to 0.67 mm (0.56-0.78 mm) after 5 years (P<0.001). CIMT increment correlated with baseline adiponectin, age, and smoking (all P<0.05) and baseline CIMT (P<0.001), but not with sex, fasting glucose, lipid profiles, hypertension, or diabetes. In multiple linear regression analysis, baseline serum adiponectin level was an independent predictor of CIMT increment β (standardized beta)=-0.17, P=0.015], after adjusting for age, smoking, baseline CIMT, hypertension, body mass index, fasting glucose, low-density lipoprotein cholesterol, and triglycerides. CONCLUSION Hypoadiponectinemia predicted CIMT progression, independent of known predictive factors such as age, smoking, hyperlipidemia, and hypertension.

An R package for analyzing and modeling ranking data

BackgroundIn medical informatics, psychology, market research and many other fields, researchers often need to analyze and model ranking data. However, there is no statistical software that provides tools for the comprehensive analysis of ranking data. Here, we present pmr, an R package for analyzing and modeling ranking data with a bundle of tools. The pmr package enables descriptive statistics (mean rank, pairwise frequencies, and marginal matrix), Analytic Hierarchy Process models (with Saaty’s and Koczkodaj’s inconsistencies), probability models (Luce model, distance-based model, and rank-ordered logit model), and the visualization of ranking data with multidimensional preference analysis.ResultsExamples of the use of package pmr are given using a real ranking dataset from medical informatics, in which 566 Hong Kong physicians ranked the top five incentives (1: competitive pressures; 2: increased savings; 3: government regulation; 4: improved efficiency; 5: improved quality care; 6: patient demand; 7: financial incentives) to the computerization of clinical practice. The mean rank showed that item 4 is the most preferred item and item 3 is the least preferred item, and significance difference was found between physicians’ preferences with respect to their monthly income. A multidimensional preference analysis identified two dimensions that explain 42% of the total variance. The first can be interpreted as the overall preference of the seven items (labeled as “internal/external”), and the second dimension can be interpreted as their overall variance of (labeled as “push/pull factors”). Various statistical models were fitted, and the best were found to be weighted distance-based models with Spearman’s footrule distance.ConclusionsIn this paper, we presented the R package pmr, the first package for analyzing and modeling ranking data. The package provides insight to users through descriptive statistics of ranking data. Users can also visualize ranking data by applying a thought multidimensional preference analysis. Various probability models for ranking data are also included, allowing users to choose that which is most suitable to their specific situations.

Decision tree modeling for ranking data

Ranking/preference data arises from many applications in marketing, psychology, and politics. We establish a new decision tree model for the analysis of ranking data by adopting the concept of classification and regression tree. The existing splitting criteria are modified in a way that allows them to precisely measure the impurity of a set of ranking data. Two types of impurity measures for ranking data are introduced, namelyg-wise and top-k measures. Theoretical results show that the new measures exhibit properties of impurity functions. In model assessment, the area under the ROC curve (AUC) is applied to evaluate the tree performance. Experiments are carried out to investigate the predictive performance of the tree model for complete and partially ranked data and promising results are obtained. Finally, a real-world application of the proposed methodology to analyze a set of political rankings data is presented.