Longitudinal Analysis Methods Research Articles

Derivation and validation of the Ottawa weight loss prediction model for patients on a low-calorie diet

Accurate weight predictions are essential for weight management program patients. The freely available National Institutes of Health Body Weight Planner (NIH-BWP) returns expected weights over time but overestimates weight when patients consume a low-calorie diet. This study sought to increase the accuracy of NIH-BWP predicted weights for people on low-calorie diets. People enrolled in a weight management program were included if they received meal replacements with defined caloric content for the 3 months of the weight loss phase of the program. The Ottawa Weight Loss Prediction Model (OWL-PM) modelled the relative difference between observed and NIH-BWP predicted weights using longitudinal analysis methods based on patient factors. OWL-PM was externally validated. 1761 people were included (mean age 46 years, 73.3% women) with a mean (SD) baseline weight in pounds and body mass index of 271.9 (55.6) and 43.9 (7.4), respectively. At the end of the program’s weight loss phase, people lost a median (IQR) of 17.1% (14.8–19.5) of their baseline weight. Observed weight relative to NIH-BWP predicted weights had a median value of − 4.9% but ranged from − 32.1% to + 28.5%. After adjustment, weight overestimation by NIH-BWP was most pronounced in male patients, people without diabetes and with increased observation time. OWL-PM returned expected weights at 3 months that were significantly more accurate than those from NIH-BWP alone (mean difference observed vs. expected [95% CI] 6.7lbs [6.4–7.0] vs. 12.6lbs [12.1–13.0]). In the external validation cohort (n = 106), OWL-PM was significantly more accurate than NIH-BWP (mean squared error 24.3 vs. 40.0, p = 0.0018). OWL-PM incorporated patient-level covariates to significantly increase weight prediction accuracy of NIH-BWP in patients consuming a low-calorie diet.

Monitoring mammary gland development in preweaned dairy heifers using ultrasound imaging

Existing literature on the long-term effects of nutritional manipulation on milk production is incomplete and inconsistent as animals are typically culled before lactation. In addition, longitudinal studies are rarely conducted due to high costs and constraints related to tissue sampling; therefore, additional research is necessary to better understand the effect of early nutrition and growth on subsequent growth and milk production. Thirty Holstein heifer calves were pair-fed high (H; 22% crude protein and 15% fat) or low (L; 27% crude protein and 20% fat) milk replacers. Mammary gland ultrasound images and body weights were collected twice weekly, and blood samples weekly, for the first 8 weeks of life. Calves fed H had higher average daily gain (ADG) and mammary gland growth through 8 weeks of life. Calves from the H group also had larger fat pads at 8 weeks of age. Glucose and insulin concentrations were increased in H calves compared with L calves, whereas NEFA concentrations were not different between treatments. Our results suggest that ultrasound is a practical, noninvasive tool to monitor udder development preweaning, although additional research is necessary to improve image analysis techniques to analyze tissue composition. Our preliminary work provides data suggesting that the use of mammary gland ultrasound could provide a method of longitudinal analysis of the effects of preweaning nutrition on milk production.

유창성 장애의 언어치료 관련 경시적 계수자료의 유연한 분석

In this study, the goal was to apply a flexible and easy-to-manage analysis method through a model that includes a dispersion parameter in the longitudinal analysis method of count data. In the analysis of count data, it may be inconvenient to change the statistical model depending on whether there is under- or over-dispersion. To compensate for these weaknesses, We have applied a statistical method including a dispersion parameter to the statistical model. The count data covered were included in the diagnosis of fluency disorders and the evaluation of the effectiveness of speech therapy. but research on speech-language therapy is limited by being conducted on small numbers of subjects for several reasons. Therefore, statistical analysis methods that work only on large samples, are complex, or require excessive calculations in the model fitting process were not appropriate for the field. With this background, we applied the Conway-Maxwell Poisson model to longitudinal data analysis of fluency disorders. We evaluated whether the model that reflects the level of data dispersion in the model was consistent with the speech therapy data for fluency disorders by comparing it to the existing Poisson mixed model and linear mixed model. In particular, the Conway-Maxwell Poisson model had the advantage of reflecting well when the degree of dispersion was different for each group or specific subgroup.

Pathway-guided monitoring of the disease course in bladder cancer with longitudinal urine proteomics

BackgroundMonitoring bladder cancer over time requires invasive and costly procedures. Less invasive approaches are required using readily available biological samples such as urine. In this study, we demonstrate a method for longitudinal analysis of the urine proteome to monitor the disease course in patients with bladder cancer.MethodsWe compared the urine proteomes of patients who experienced recurrence and/or progression (n = 13) with those who did not (n = 17). We identified differentially expressed proteins within various pathways related to the hallmarks of cancer. The variation of such pathways during the disease course was determined using our differential personal pathway index (dPPi) calculation, which could indicate disease progression and the need for medical intervention.ResultsSeven hallmark pathways are used to develop the dPPi. We demonstrate that we can successfully longitudinally monitor the disease course in bladder cancer patients through a combination of urine proteomic analysis and the dPPi calculation, over a period of 62 months.ConclusionsUsing the information contained in the patient’s urinary proteome, the dPPi reflects the individual’s course of bladder cancer, and helps to optimise the use of more invasive procedures such as cystoscopy.

Liver lesion changes analysis in longitudinal CECT scans by simultaneous deep learning voxel classification with SimU-Net.

The identification and quantification of liver lesions changes in longitudinal contrast enhanced CT (CECT) scans is required to evaluate disease status and to determine treatment efficacy in support of clinical decision-making. This paper describes a fully automatic end-to-end pipeline for liver lesion changes analysis in consecutive (prior and current) abdominal CECT scans of oncology patients. The three key novelties are: (1) SimU-Net, a simultaneous multi-channel 3D R2U-Net model trained on pairs of registered scans of each patient that identifies the liver lesions and their changes based on the lesion and healthy tissue appearance differences; (2) a model-based bipartite graph lesions matching method for the analysis of lesion changes at the lesion level; (3) a method for longitudinal analysis of one or more of consecutive scans of a patient based on SimU-Net that handles major liver deformations and incorporates lesion segmentations from previous analysis. To validate our methods, five experimental studies were conducted on a unique dataset of 3491 liver lesions in 735 pairs from 218 clinical abdominal CECT scans of 71 patients with metastatic disease manually delineated by an expert radiologist. The pipeline with the SimU-Net model, trained and validated on 385 pairs and tested on 249 pairs, yields a mean lesion detection recall of 0.86±0.14, a precision of 0.74±0.23 and a lesion segmentation Dice of 0.82±0.14 for lesions > 5mm. This outperforms a reference standalone 3D R2-UNet mdel that analyzes each scan individually by ∼50% in precision with similar recall and Dice score on the same training and test datasets. For lesions matching, the precision is 0.86±0.18 and the recall is 0.90±0.15. For lesion classification, the specificity is 0.97±0.07, the precision is 0.85±0.31, and the recall is 0.86±0.23. Our new methods provide accurate and comprehensive results that may help reduce radiologists' time and effort and improve radiological oncology evaluation.

Longitudinal modelling of housing prices with machine learning and temporal regression

Purpose The purpose of this paper is to model housing price temporal variations and to predict price trends within the context of land use–transportation interactions using machine learning methods based on longitudinal observation of housing transaction prices. Design/methodology/approach This paper examines three machine learning algorithms (linear regression machine learning (ML), random forest and decision trees) applied to housing price trends from 2001 to 2016 in the Greater Toronto and Hamilton Area, with particular interests in the role of accessibility in modelling housing price. It compares the performance of the ML algorithms with traditional temporal lagged regression models. Findings The empirical results show that the ML algorithms achieve good accuracy (R2 of 0.873 after cross-validation), and the temporal regression produces competitive results (R2 of 0.876). Temporal lag effects are found to play a key role in housing price modelling, along with physical conditions and socio-economic factors. Differences in accessibility effects on housing prices differ by mode and activity type. Originality/value Housing prices have been extensively modelled through hedonic-based spatio-temporal regression and ML approaches. However, the mutually dependent relationship between transportation and land use makes price determination a complex process, and the comparison of different longitudinal analysis methods is rarely considered. The finding presents the longitudinal dynamics of housing market variation to housing planners.

Eco‐innovations of firms: A longitudinal analysis of the roles of industry norms and proactive environmental strategy

AbstractThere is scant empirical longitudinal evidence in the extant eco‐innovation literature on how industry norms, standards and expectations influence firms' strategic choices and innovation activities in the context of environmental sustainability. This study fills this research gap by determining the causal effects of industry norms that promote environmentally sustainable business practices on developing a firm's proactive environmental strategy, which propels a firm's eco‐innovation outputs or those innovations considered environmentally sustainable. It utilises two waves of survey data involving 274 small and medium enterprises and the cross‐lagged panel method of longitudinal analysis to bring to light the explanatory mediating power of a firm's environmental strategy as a response to industry norms that push firms to pursue eco‐innovations. Drawing on institutional and strategic adaptation theories, this study's main findings on the causal and predictive temporal effects of industry norms on firm strategy and eco‐innovation offer robust empirical evidence that extends and overcomes the shortcomings of previous studies on firm‐level eco‐innovation and its institutional and strategy‐related determinants. This article presents the significant theoretical, empirical, methodological and practical implications of the study.

The features of the USSR population reproduction in 1920–1930 in the demographic stability parameters

The 1932–1933 Holodomor genocide caused a significant reduction of the population and distorted the main parameters of the country’s demographic stability. To analyze the changes in the population by the longitudinal analysis method, the survival rates are calculated for age groups of the population in the intercensal periods. Calculations based on the data of 1926 and 1939 censuses show substantial inconsistencies in the survival rate for five female age groups within the range of 33-57 years old. Despite the 1932–1933 demographic catastrophe and taking into account the natural mortality before and after the Holodomor, the number of people in these age categories changed slightly and even increased for the 43-47 category. The authors prove that migration in the intercensal period couldn’t have caused such growth, and the calculated coefficients confirm the questionable quality of the 1939 census. The paper reveals that the first stage of demographic transition should have been in 1932–1933, yet historical-political and socio-economic conditions of demographic reproduction aggravated considerably in early 1930. The recovery of the high birth rate in 1937–1938 shows that the first stage of demographic transition wasn’t finalized. Calculation of the number of women in the most active childbirth age and analysis of birth rate for 1924–1929 shows that in 1930–1933, the birth rate should have increased. The calculated indirect demographic losses account 310,000-430,000 annually for 1930, 1931, 1937, and 1938, and from 630,000 to 1 million for 1932–1936. Such insignificant volume of indirect losses substantially reduced demographic stability, in the first place of Ukrainian village. The incomplete and questionable nature of demographic data after 1932 does not allow complete reconstruction of the demographic reproduction processes in the USSR in the 1930s. Archive data on the natural movement of the population on lover administrative-territorial levels, current statistical recordings, and recordings of the civil status acts on birth and mortality, etc. can be the perspective sources of demographic information.

Under-Five Pneumonia Patients in Menz Geramidr Mehalmeda Hospital North Shewa, Ethiopia:Bayesian Parametric Survival Model

Background: Pneumonia is that the most wanted largest infectious reason for death in children worldwide. It’s most prevalent in South Asia and geographic region. In Ethiopia, pneumonia could be a leading single disease killing under-five children. Parametric survival analysis is defined as a group of longitudinal analysis methods for interrogating data having time as an outcome variable and Bayesian analysis is employed to boost the precision of the results by introducing external information in terms of the prior distribution. The aim of this study was to analyze the survival rate of under-five pneumonia patients in Menz Gera Mehalmeda Hospital using Bayesian survival analysis. Methodology: Retrospective study was conducted in Menz Gera Mehalmeda Hospital from September1,2002 up to August 30,2020. Children whose age between greater than 29 days and less than five year were included within the study and Patients with insufficient information were excluded from the study. The parametric classical AFT models and Bayesian AFT models were used for the analysis. Result: The results implied that patients whose residence were urban, male patients, age groups of patient at the age (12-23) months, (24-35) months, (36-47) months and (48-59) months, patients without comorbidity, patients without severe acute malnutrition (SAM), was prolonged timing death of under-five pneumonia patients, while female patients, age of patients at the age of(1-11) months, patients with comorbidity, patients whose residence was rural and patients with severe acute malnutrition (SAM) were statistically significantly shorten timing of death. Conclusion: Finally, the findings of this study implied that the sex of children, residence of children, age of children, Co-morbidity, Severe Acute Malnutrition(SAM) and weight were major factors related to survival time of under-five pneumonia patients in Menz Gera Mehalmeda Hospital.

A multiple robust propensity score method for longitudinal analysis with intermittent missing data

Longitudinal data are very popular in practice, but they are often missing in either outcomes or time-dependent risk factors, making them highly unbalanced and complex. Missing data may contain various missing patterns or mechanisms, and how to properly handle it for unbiased and valid inference still presents a significant challenge. Here, we propose a novel semiparametric framework for analyzing longitudinal data with both missing responses and covariates that are missing at random and intermittent, a general and widely encountered situation in observational studies. Within this framework, we consider multiple robust estimation procedures based on innovative calibrated propensity scores, which offers additional relaxation of the misspecification of missing data mechanisms and shows more satisfactory numerical performance. Also, the corresponding robust information criterion on consistent variable selection for our proposed model is developed based on empirical likelihood-based methods. These advocated methods are evaluated in both theory and extensive simulation studies in a variety of situations, showing competing properties and advantages compared to the existing approaches. We illustrate the utility of our approach by analyzing the data from the HIV Epidemiology ResearchStudy.

Late, Persistent, Substantial, Treatment-Related Symptoms After Radiation Therapy (LAPERS): A New Method for Longitudinal Analysis of Late Morbidity—Applied in the EMBRACE Study

Current incidence methods for reporting mild or moderate symptoms capture the (first) occurrence of an event and do not allow distinguishing between patients who suffer from long-lasting versus transient morbidity. This paper introduces a new methodological approach that identifies cancer survivors who have clinically relevant, long-lasting symptoms (patients with late, persistent, substantial and treatment-related symptoms, [LAPERS]). LAPERS can be evaluated in patients with baseline information and at least 3 late follow-up assessments after treatment. LAPERS identifies individual patients with a given symptom that is substantial (above a predefined clinically relevant threshold) and must be present in at least half of the follow-ups. Baseline morbidity is accounted for by requiring the median of the late symptom score to be worse than the baseline condition. The LAPERS approach was applied to 4 relevant patient-reported genito-urinary/gastrointestinal symptoms within the prospective, longitudinal EMBRACE study (An intErnational study on MRI-guided BRachytherapy in locally Advanced CErvical cancer, www.embracestudy.dk). LAPERS was compared with crude incidence and prevalence rates. Within the EMBRACE cohort, 651/1044 patients (62%) had baseline and long-term follow-up available (median follow-up: 42 months). There was a considerable gap between LAPERS, crude incidence, and prevalence rates. The proportion of patients with LAPERS events was 3.8-4.8 times lower than crude incidences. The highest prevalence rates across follow-up times were 1.8-2.6 times lower than crude incidences. These findings indicate limitations of incidence methods for reporting substantial patient-reported symptoms because a considerable proportion of patients with symptoms do not experience them persistently over time, as they may fluctuate or get successfully treated. In contrast, the LAPERS method for longitudinal analysis identifies patients with clinically relevant, long-lasting symptoms.

Systematic review of analytical methods applied to longitudinal studies of malaria

BackgroundModelling risk of malaria in longitudinal studies is common, because individuals are at risk for repeated infections over time. Malaria infections result in acquired immunity to clinical malaria disease. Prospective cohorts are an ideal design to relate the historical exposure to infection and development of clinical malaria over time, and analysis methods should consider the longitudinal nature of the data. Models must take into account the acquisition of immunity to disease that increases with each infection and the heterogeneous exposure to bites from infected Anopheles mosquitoes. Methods that fail to capture these important factors in malaria risk will not accurately model risk of malaria infection or disease.MethodsStatistical methods applied to prospective cohort studies of clinical malaria or Plasmodium falciparum infection and disease were reviewed to assess trends in usage of the appropriate statistical methods. The study was designed to test the hypothesis that studies often fail to use appropriate statistical methods but that this would improve with the recent increase in accessibility to and expertise in longitudinal data analysis.ResultsOf 197 articles reviewed, the most commonly reported methods included contingency tables which comprised Pearson Chi-square, Fisher exact and McNemar’s tests (n = 102, 51.8%), Student’s t-tests (n = 82, 41.6%), followed by Cox models (n = 62, 31.5%) and Kaplan–Meier estimators (n = 59, 30.0%). The longitudinal analysis methods generalized estimating equations and mixed-effects models were reported in 41 (20.8%) and 24 (12.2%) articles, respectively, and increased in use over time. A positive trend in choice of more appropriate analytical methods was identified over time.ConclusionsDespite similar study designs across the reports, the statistical methods varied substantially and often represented overly simplistic models of risk. The results underscore the need for more effort to be channelled towards adopting standardized longitudinal methods to analyse prospective cohort studies of malaria infection and disease.

Effectiveness of energy renovations: a reassessment based on actual consumption savings

Energy renovations offer unique opportunities to increase the energy efficiency of the built environment and for the existing housing stock, they are the most important solution. Usually, energy savings are based on modelling calculations. However, recent research has shown that the predicted energy consumption differs largely from the actual consumption. In this paper, the effectiveness of energy measures is re-assessed based on actual consumption data. We use a monitoring system, which contains information about the energy performance of around 60% of the Dutch non-profit housing sector (circa 1.2 million dwellings). We connect the data from this monitoring system to actual energy consumption data from Statistics Netherlands on a dwelling level. Using longitudinal analysis methods, from 2010 to 2014, we are able to identify the energy efficiency improvements of the stock and determine the effectiveness of different measures in terms of actual energy savings. The results reveal the actual energy savings of different efficiency measures and highlight the significance of the actual energy consumption when a renovation is planned or realized.

Longitudinal Modeling Approaches to Assess the Association Between Changes in 2 Clinical Outcome Assessments.

Understanding how one clinical outcome assessment (COA) (eg, a patient-reported outcome [PRO]) relates to a second COA (eg, a clinician-reported outcome [ClinRO]) may provide insights into disease burden or treatment efficacy. We aimed to briefly review commonly used cross-sectional methods to evaluate the association between a PRO and a ClinRO and to demonstrate the advantages of longitudinal modeling approaches, particularly a joint mixed model for repeated measures (MMRM), to evaluate this association. We generated an example longitudinal data set that included a PRO measured on an 11-point numeric rating scale and a binary ClinRO. The association between change in PRO score and ClinRO response at each time point was examined using 2 cross-sectional analyses: point biserial correlation and logistic regression. We conducted longitudinal analyses of the association between the 2 COAs across time points using MMRM and joint MMRM approaches. Point-biserial correlation and logistic regression analyses correctly captured the "built in" associations between the 2 COAs that strengthened over time, but each association was applicable only for a single time point. The MMRM approach provided correlations over time but only for a single outcome variable. The joint MMRM approach modeled the relationship between both outcome variables simultaneously, allowing for evaluation of the correlations both within and between the variables over time. Each analysis demonstrated the relationship between PRO score changes and ClinRO response. Longitudinal analysis methods, particularly the joint MMRM, allow for a more thorough examination of the correlations among the 2 outcomes than cross-sectional analysis methods.

Effectiveness of energy renovations: a reassessment based on actual consumption savings

Energy renovations offer unique opportunities to increase the energy efficiency of the built environment and for the existing housing stock; they are the most important solution. Usually, energy savings are based on modeling calculations. However, recent research has shown that the predicted energy consumption differs largely from the actual consumption. In this paper, the effectiveness of energy measures is re-assessed based on actual consumption data. We use a monitoring system, which contains information about the energy performance of around 60% of the Dutch non-profit housing sector (circa 1.2 million dwellings). We connect the data from this monitoring system to actual energy consumption data from Statistics Netherlands on a dwelling level. Using longitudinal analysis methods, from 2010 to 2014, we are able to identify the energy efficiency improvements of the stock and determine the effectiveness of different measures in terms of actual energy savings. The results reveal the actual energy savings of different efficiency measures and highlight the significance of the actual energy consumption when a renovation is planned or realized.

Effectiveness of energy renovations: a reassessment based on actual consumption savings

Energy renovations offer unique opportunities to increase the energy efficiency of the built environment and for the existing housing stock, they are the most important solution. Usually, energy savings are based on modelling calculations. However, recent research has shown that the predicted energy consumption differs largely from the actual consumption. In this paper, the effectiveness of energy measures is re-assessed based on actual consumption data. We use a monitoring system, which contains information about the energy performance of around 60% of the Dutch non-profit housing sector (circa 1.2 million dwellings). We connect the data from this monitoring system to actual energy consumption data from Statistics Netherlands on a dwelling level. Using longitudinal analysis methods, from 2010 to 2014, we are able to identify the energy efficiency improvements of the stock and determine the effectiveness of different measures in terms of actual energy savings. The results reveal the actual energy savings of different efficiency measures and highlight the significance of the actual energy consumption when a renovation is planned or realized.

Gene-by-Psychosocial Factor Interactions Influence Diastolic Blood Pressure in European and African Ancestry Populations: Meta-Analysis of Four Cohort Studies.

Inter-individual variability in blood pressure (BP) is influenced by both genetic and non-genetic factors including socioeconomic and psychosocial stressors. A deeper understanding of the gene-by-socioeconomic/psychosocial factor interactions on BP may help to identify individuals that are genetically susceptible to high BP in specific social contexts. In this study, we used a genomic region-based method for longitudinal analysis, Longitudinal Gene-Environment-Wide Interaction Studies (LGEWIS), to evaluate the effects of interactions between known socioeconomic/psychosocial and genetic risk factors on systolic and diastolic BP in four large epidemiologic cohorts of European and/or African ancestry. After correction for multiple testing, two interactions were significantly associated with diastolic BP. In European ancestry participants, outward/trait anger score had a significant interaction with the C10orf107 genomic region (p = 0.0019). In African ancestry participants, depressive symptom score had a significant interaction with the HFE genomic region (p = 0.0048). This study provides a foundation for using genomic region-based longitudinal analysis to identify subgroups of the population that may be at greater risk of elevated BP due to the combined influence of genetic and socioeconomic/psychosocial risk factors.

Validation of two portable bioelectrical impedance analyses for the assessment of body composition in school age children.

BackgroundBioelectrical impedance analysis (BIA) is a convenient and child-friendly method for longitudinal analysis of changes in body composition. However, most validation studies of BIA have been performed on adult Caucasians. The present cross-sectional study investigated the validity of two portable BIA devices, the Inbody 230 (BIA8MF) and the Tanita BC-418 (BIA8SF), in healthy Taiwanese children.MethodsChildren aged 7–12 years (72 boys and 78 girls) were recruited. Body composition was measured by the BIA8SF and the BIA8MF. Dual X-ray absorptiometry (DXA) was used as the reference method.ResultsThere were strong linear correlations in body composition measurements between the BIA8SF and DXA and between the BIA8MF and DXA. Both BIAs underestimated fat mass (FM) and percentage body fat (%BF) relative to DXA in both genders The degree of agreement in lean body mass (LBM), FM, and %BF estimates was higher between BIA8MF and DXA than between BIA8SF and DXA. The Lin’s concordance correlation coefficient (ρc) for LBM8MF met the criteria of substantial to perfect agreement whereas the ρc for FM8MF met the criteria of fair to substantial agreement. Bland-Altman analysis showed a clinically acceptable agreement between LBM measures by BIA8MF and DXA. The limit of agreement in %BF estimation by BIA and DXA were wide and the errors were clinically important. For the estimation of ALM, BIA8SF and BIA8MF both provided poor accuracy.ConclusionsFor all children, LBM measures were precise and accurate using the BIA8MF whereas clinically significant errors occurred in FM and %BF estimates. Both BIAs underestimated FM and %BF in children. Thus, the body composition results obtained using the inbuilt equations of the BIA8SF and BIA8MF should be interpreted with caution, and high quality validation studies for specific subgroups of children are required prior to field research.

Impact of Health System Inputs on Health Outcome: A Multilevel Longitudinal Analysis of Botswana National Antiretroviral Program (2002-2013).

ObjectiveTo measure the association between the number of doctors, nurses and hospital beds per 10,000 people and individual HIV-infected patient outcomes in Botswana.DesignAnalysis of routinely collected longitudinal data from 97,627 patients who received ART through the Botswana National HIV/AIDS Treatment Program across all 24 health districts from 2002 to 2013. Doctors, nurses, and hospital bed density data at district-level were collected from various sources.MethodsA multilevel, longitudinal analysis method was used to analyze the data at both patient- and district-level simultaneously to measure the impact of the health system input at district-level on probability of death or loss-to-follow-up (LTFU) at the individual level. A marginal structural model was used to account for LTFU over time.ResultsIncreasing doctor density from one doctor to two doctors per 10,000 population decreased the predicted probability of death for each patient by 27%. Nurse density changes from 20 nurses to 25 nurses decreased the predicted probability of death by 28%. Nine percent decrease was noted in predicted mortality of an individual in the Masa program for every five hospital bed density increase.ConclusionConsiderable variation was observed in doctors, nurses, and hospital bed density across health districts. Predictive margins of mortality and LTFU were inversely correlated with doctor, nurse and hospital bed density. The doctor density had much greater impact than nurse or bed density on mortality or LTFU of individual patients. While long-term investment in training more healthcare professionals should be made, redistribution of available doctors and nurses can be a feasible solution in the short term.

Temporally Constrained Group Sparse Learning for Longitudinal Data Analysis in Alzheimer's Disease.

Sparse learning has been widely investigated for analysis of brain images to assist the diagnosis of Alzheimer's disease and its prodromal stage, i.e., mild cognitive impairment. However, most existing sparse learning-based studies only adopt cross-sectional analysis methods, where the sparse model is learned using data from a single time-point. Actually, multiple time-points of data are often available in brain imaging applications, which can be used in some longitudinal analysis methods to better uncover the disease progression patterns. Accordingly, in this paper, we propose a novel temporallyconstrained group sparse learning method aiming for longitudinal analysis with multiple time-points of data. Specifically, we learn a sparse linear regression model by using the imaging data from multiple time-points, where a group regularization term is first employed to group the weights for the same brain region across different time-points together. Furthermore, to reflect the smooth changes between data derived from adjacent time-points, we incorporate two smoothness regularization terms into the objective function, i.e., one fused smoothness term thatrequires that the differences between two successive weight vectors from adjacent time-points should be small, and another output smoothness term thatrequires the differences between outputs of two successive models from adjacent time-points should also be small. We develop an efficient optimization algorithm to solve the proposed objective function. Experimental results on ADNI database demonstrate that, compared with conventional sparse learning-based methods, our proposed method can achieve improved regression performance and also help in discovering disease-related biomarkers.