Comparison of imputation methods for handling missing covariate data when fitting a Cox proportional hazards model: a resampling study

Maternal early life and prenatal stress in relation to birth outcomes in Argentinian mothers

Environmental influences before and during pregnancy significantly impact offspring development. This study investigates open research questions regarding the associations between maternal early life stress (ELS), prenatal psychosocial stress, prenatal hair cortisol (HC), and birth outcomes in Argentinian women. Data on ELS, prenatal life events, HC (two samples representing first and second half of pregnancy), and birth outcomes were collected from middle-class Argentinian women (N = 69) upon delivery. Linear mixed models indicated that HC increased from the first half to the second half of pregnancy with considerable variability in the starting values and slopes between individuals. Mothers who experienced more ELS, were taller, or more educated, tended to show lower increases in HC. Older age was positively related to HC increases. Our data did not suggest an interaction between ELS and prenatal life events in relation to HC. We found that the change in HC was most likely negatively associated with birth weight. Our data are most compatible with either a weak or the absence of an association between ELS or prenatal life events and absolute values of HC. Mothers with stronger increases in hair cortisol tended to have newborns with slightly lower birth weight. Hence, ELS and birthweight may either have been related to changes in cortisol exposure during pregnancy or to factors that influence accumulation or retention of cortisol in hair.

Skills training followed by either EMDR or narrative therapy for posttraumatic stress disorder in adult survivors of childhood abuse: a randomized controlled trial

Background: Individuals suffering from PTSD following childhood abuse represent a large subgroup of patients attending mental health services. The aim of phase-based treatment is to tailor treatment to the specific needs to childhood abuse survivors with PTSD with a Skills Training in Affective and Interpersonal Regulation (STAIR) phase, in which emotion dysregulation and interpersonal problems are targeted, and a trauma-focused phase.Objective: The purpose of this study was to compare STAIR + Eye Movement Desensitization and Reprocessing (EMDR) vs. STAIR + Narrative Therapy (NT) as treatments for PTSD following childhood-onset trauma in a routine clinical setting.Method: Sixty-eight adults were randomly assigned to STAIR/EMDR (8 STAIR-sessions followed by 12 EMDR-sessions) or STAIR/NT (8 STAIR-sessions followed by 12 NT-sessions). Assessments took place at pre-treatment, after each treatment phase and at 3 and 12 months post-intervention follow-up. Primary outcomes were interviewer-rated and self-reported symptom levels of PTSD. Secondary outcomes included symptom levels of depression and disturbances in emotion regulation and interpersonal skills.Results: Multilevel analyses in the intent-to-treat sample indicated that patients in both treatments improved substantially on PTSD symptom severity (CAPS: d = 0.81 to 1.29; PDS: d = 1.68 to 2.15), as well as on symptom levels of depression, anxiety, emotion regulation, dissociation and interpersonal skills. Effects increased or were maintained until 12-month follow-up. At mid-treatment, after STAIR, patients in both treatments improved moderately on PTSD symptom severity (PDS: d = 1.68 to 2.15), as well as on symptom levels of depression (BDI: d = .32 to .31). Symptoms of anxiety, emotion dysregulation, interpersonal problems and dissociation were not decreased after STAIR. There were no significant differences between the two conditions on any outcome.Conclusion: PTSD in adult survivors of childhood interpersonal trauma can effectively be treated by phase-based interventions using either EMDR or NT in the trauma-processing phase.Trial registration: ClinicalTrials.gov identifier: NCT01443182..

Chronotype and emotional/behavioral problems mediate the association between leisure screen time and academic achievement in children

Leisure screen time is associated with poor academic achievement; however, the mechanism underlying this relationship is unclear. Chronotypes and emotional/behavioral problems may be linked to this association. This study aimed to examine the associations between leisure screen time, chronotype, emotional/behavioral problems, and academic achievement using mediation analysis. A total of 113 children aged 9-12 years participated in this study. All participants were assessed for leisure screen time, chronotype, emotional/behavioral problems, and academic achievement. Leisure screen time was evaluated using a self-reported questionnaire. Chronotypes were measured using The Japanese Children's Chronotype Questionnaire, and the morningness/eveningness (M/E) score was calculated. Emotional/behavioral problems were assessed using The Japanese Strengths and Difficulties Questionnaire, and the total difficulties score (TDS) was calculated. Academic achievement was assessed by the homeroom teacher for each of the seven school subjects. Partial correlation analysis adjusted for grade, sex, and sleep duration indicated that leisure screen time was associated with M/E scores and academic achievement (p < 0.05). There was a positive association between M/E score and TDS (p < 0.05) and a negative association between TDS and academic achievement (p < 0.05). A mediation analysis adjusted for grade, sex, and sleep duration was performed. There was a significant total effect of leisure screen time on academic achievement (p < 0.05). Additionally, the M/E score and TDS significantly mediated the association between leisure screen time and academic achievement (p < 0.05). Our findings suggest that the serial path between chronotype and emotional/behavioral problems weakly but significantly mediates the association of leisure screen time with academic achievement.

Risk perception as a motivational resource during the COVID-19 pandemic: the role of vaccination status and emerging variants

BACKGROUND

People's perceived risk of being infected and having severe illness was conceived as a motivational source of adherence to behavioral measures during the COVID-19 crisis.

METHODS

We used online self-reported data, spanning 20 months of the COVID-19 crisis in Belgium (n = 221,791; 34.4% vaccinated; July 2020 - March 2022) to study the association between risk perception and motivation.

RESULTS

Both perceived infection probability and severity fluctuated across time as a function of the characteristics of emerging variants, with unvaccinated persons perceiving decreasingly less risk compared to vaccinated ones. Perceived severity (and not perceived probability) was the most critical predictor of autonomous motivation for adherence to health-protective measures, a pattern observed at both the between-day and between-person level among both vaccinated and unvaccinated individuals. An integrated process model further indicated that on days with higher hospitalization load, participants reported being more adherent because risk severity and autonomous motivation for adherence were more elevated on these days.

CONCLUSIONS

These findings suggest that risk severity served as a critical and dynamic resource for adherence to behavioral measures because it fostered greater autonomous regulation.

The prognostic value of a geriatric risk score for older patients undergoing emergency surgery of colorectal cancer: A retrospective cohort study

INTRODUCTION

Emergency surgery of colorectal cancer is associated with high mortality rates in older patients. We investigated whether information on four geriatric domains has prognostic value for 30-day mortality and postoperative morbidity including severe complications.

MATERIALS AND METHODS

All consecutive patients aged 70 years or older who underwent emergency colorectal cancer surgery in six Dutch hospitals (2014-2017) were studied. Presence of geriatric risk factors was scored prior to surgery as either 0 (risk absent) or 1 (risk present) in each of four geriatric domains and summed up to calculate a sumscore with a value between 0 and 4. In addition, we separately investigated the use of a mobility aid. Primary outcome was 30-day mortality. Secondary outcomes were any postoperative complications and severe complications. Multivariable logistic regression model was used to evaluate the sumscore and outcomes.

RESULTS

Two hundred seven patients were included. Median age was 79.4 years. One hundred seventy-five patients (76%) presented with obstruction, 22 (11%) with a perforation, and 17 (8%) with severe anemia. Mortality rates were 2.9%, 13.6%, and 29.6% for patients with a sumscore of 0, 1-2, and 3-4 respectively, with odds ratio (OR) 4.8 [95% confidence interval (CI) 1.03-22.95] and OR 10.6 [95% CI 1.99-56.34] for a sumscore of 1-2 and 3-4 respectively. Use of a mobility aid was associated with increased mortality OR 8.0 [95% CI 2.74-23.43] and severe complications OR 2.31 [95% CI 1.17-4.55].

DISCUSSION

This geriatric sumscore and the use of a mobility aid have strong association with 30-day mortality after emergency surgery of colorectal cancer. This could provide better insight into surgical risk and help select high-risk patients for alternative strategies.

Comment on Oberman & Vink: Should we fix or simulate the complete data in simulation studies evaluating missing data methods?

For simulation studies that evaluate methods of handling missing data, we argue that generating partially observed data by fixing the complete data and repeatedly simulating the missingness indicators is a superficially attractive idea but only rarely appropriate to use.

Comparative analysis of anticholinergic burden scales to explain iatrogenic cognitive impairment and self-reported side effects in the euthymic phase of bipolar disorders: Results from the FACE-BD cohort

Bipolar disorders (BD) are characterized by cognitive impairment during the euthymic phase, to which treatments can contribute. The anticholinergic properties of medications, i.e., the ability of a treatment to inhibit cholinergic receptors, are associated with cognitive impairment in elderly patients and people with schizophrenia but this association has not been well characterized in individuals with remitted BD. Moreover, the validity of only one anticholinergic burden scale designed to assess the anticholinergic load of medications has been tested in BD. In a literature review, we identified 31 existing scales. We first measured the associations between 27 out of the 31 scales and objective cognitive impairment in bivariable regressions. We then adjusted the bivariable models with covariates: the scales significantly associated with cognitive impairment in bivariable and multiple logistic regressions were defined as having good concurrent validity to assess cognitive impairment. In a sample of 2,031 individuals with euthymic BD evaluated with a neuropsychological battery, two scales had good concurrent validity to assess cognitive impairment, whereas chlorpromazine equivalents, lorazepam equivalents, the number of antipsychotics, or the number of treatments had not. Finally, similar analyses with subjective anticholinergic side-effects as outcome variables reported 14 scales with good concurrent validity to assess self-reported peripheral anticholinergic side-effects and 13 to assess self-reported central anticholinergic side-effects. Thus, we identified valid scales to monitor the anticholinergic burden in BD, which may be useful in estimating iatrogenic cognitive impairment in studies investigating cognition in BD.

Dealing with missing data in laboratory test results used as a baseline covariate: results of multi-hospital cohort studies utilizing a database system contributing to MID-NET® in Japan

BACKGROUND

To evaluate missing data methods applied to laboratory test results used for confounding adjustment, utilizing data from 10 MID-NET®-collaborative hospitals.

METHODS

Using two scenarios, five methods dealing with missing laboratory test results were applied, including three missing data methods (single regression imputation (SRI), multiple imputation (MI), and inverse probability weighted (IPW) method). We compared the point estimates of adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) between the five methods. Hospital variability in missing data was considered using the hospital-specific approach and overall approach. Confounding adjustment methods were propensity score (PS) weighting, PS matching, and regression adjustment.

RESULTS

In Scenario 1, the risk of diabetes due to second-generation antipsychotics was compared with that due to first-generation antipsychotics. The aHR adjusted by PS weighting using SRI, MI, and IPW by the hospital-specific-approach was 0.61 [95%CI, 0.39-0.96], 0.63 [95%CI, 0.42-0.93], and 0.76 [95%CI, 0.46-1.25], respectively. In Scenario 2, the risk of liver injuries due to rosuvastatin was compared with that due to atorvastatin. Although PS matching largely contributed to differences in aHRs between methods, PS weighting provided no substantial difference in point estimates of aHRs between SRI and MI, similar to Scenario 1. The results of SRI and MI in both scenarios showed no considerable changes, even upon changing the approaches considering hospital variations.

CONCLUSIONS

SRI and MI provide similar point estimates of aHR. Two approaches considering hospital variations did not markedly affect the results. Adjustment by PS matching should be used carefully.

Development of the PREDS score to predict in-hospital mortality of patients with Ebola virus disease under advanced supportive care: Results from the EVISTA cohort in the Democratic Republic of the Congo

BACKGROUND

As mortality remains high for patients with Ebola virus disease (EVD) despite new treatment options, the ability to level up the provided supportive care and to predict the risk of death is of major importance. This analysis of the EVISTA cohort aims to describe advanced supportive care provided to EVD patients in the Democratic Republic of the Congo (DRC) and to develop a simple risk score for predicting in-hospital death, called PREDS.

METHODS

In this prospective cohort (NCT04815175), patients were recruited during the 10^th EVD outbreak in the DRC across three Ebola Treatment Centers (ETCs). Demographic, clinical, biological, virological and treatment data were collected. We evaluated factors known to affect the risk of in-hospital death and applied univariate and multivariate Cox proportional-hazards analyses to derive the risk score in a training dataset. We validated the score in an internal-validation dataset, applying C-statistics as a measure of discrimination.

FINDINGS

Between August 1^st 2018 and December 31^th 2019, 711 patients were enrolled in the study. Regarding supportive care, patients received vasopressive drug (n = 111), blood transfusion (n = 101), oxygen therapy (n = 250) and cardio-pulmonary ultrasound (n = 15). Overall, 323 (45%) patients died before day 28. Six independent prognostic factors were identified (ALT, creatinine, modified NEWS2 score, viral load, age and symptom duration). The final score range from 0 to 13 points, with a good concordance (C = 86.24%) and calibration with the Hosmer-Lemeshow test (p = 0.12).

INTERPRETATION

The implementation of advanced supportive care is possible for EVD patients in emergency settings. PREDS is a simple, accurate tool that could help in orienting early advanced care for at-risk patients after external validation.

FUNDING

This study was funded by ALIMA.

Multiple Imputation to Salvage Partial Respondents: Analysis of the Forgotten Joint Score-12 After Total Hip Arthroplasty

BACKGROUND

Missing responses are common when Asian patients complete the Forgotten Joint Score-12 (FJS-12), which is widely used to evaluate total hip arthroplasty (THA). We aimed to provide orthopaedic researchers with a solution for handling missing values in such patient-reported outcome measures (PROMs).

METHODS

Patients who had undergone primary THA between 1998 and 2016 (n = 1,021) were investigated in 2020. The FJS-12 and 9 other PROMs, including questions related to Asian lifestyle activities, were administered. Risk factors for missing FJS-12 items were investigated. Partial respondents were matched with complete respondents; then, in each pair, the items not completed by the partial respondent were deleted from the responses of the complete respondent. Predictive mean matching (PMM) was performed in an attempt to recover the deleted items, using 65 sets of imputation models. After the missing values had been imputed, we explored patient characteristics that affected the FJS-12, using data from all complete and partial respondents.

RESULTS

A total of 652 patients responded to the survey (393 complete and 193 partial respondents). Partial respondents were older, more often female, and less active. Older respondents were more likely to skip items involving the bed, while those who reported a better ability to sit in the seiza style (traditional Japanese floor sitting) were more likely to skip items about chair sitting. The imputed FJS-12 value exhibited excellent reliability (intraclass correlation coefficient for agreement with the true scores, 0.985). FJS-12 values of complete respondents were significantly higher than those of respondents with 4 to 11 missing items (51.6 versus 32.8, p < 0.001). Older age was associated with higher FJS-12 values, which was revealed only via analysis of the multiply imputed data sets (p < 0.001).

CONCLUSIONS

Analysis of only complete FJS-12 responses after THA resulted in a nonresponse bias, preferentially excluding older, female, and less active individuals and those with a traditional floor living style. Multiple imputation could provide a solution to scoring and analyzing PROMs with missing responses by permitting the inclusion of partial respondents.

LEVEL OF EVIDENCE

Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

Predicting in-hospital mortality for MIMIC-III patients: A nomogram combined with SOFA score

Predicting the mortality of patients provides a reference for doctors to judge their physical condition. This study aimed to construct a nomogram to improve the prediction accuracy of patients' mortality. Patients with severe diseases were screened from the Medical Information Mart for Intensive Care (MIMIC) III database; 70% of patients were randomly selected as the training set for the model establishment, while 30% were used as the test set. The least absolute shrinkage and selection operator (LASSO) regression method was used to filter variables and select predictors. A multivariable logistic regression fit was used to determine the association between in-hospital mortality and risk factors and to construct a nomogram. A total of 9276 patients were included. The area under the curve (AUC) for the clinical nomogram based on risk factors selected by LASSO and multivariable logistic regressions were 0.849 (95% confidence interval [CI]: 0.835-0.863) and 0.821 (95% CI: 0.795-0.846) in the training and test sets, respectively. Therefore, this nomogram might help predict the in-hospital mortality of patients admitted to the intensive care unit (ICU).

Development and validation of prognostic models for anal cancer outcomes using distributed learning: protocol for the international multi-centre atomCAT2 study

BACKGROUND

Anal cancer is a rare cancer with rising incidence. Despite the relatively good outcomes conferred by state-of-the-art chemoradiotherapy, further improving disease control and reducing toxicity has proven challenging. Developing and validating prognostic models using routinely collected data may provide new insights for treatment development and selection. However, due to the rarity of the cancer, it can be difficult to obtain sufficient data, especially from single centres, to develop and validate robust models. Moreover, multi-centre model development is hampered by ethical barriers and data protection regulations that often limit accessibility to patient data. Distributed (or federated) learning allows models to be developed using data from multiple centres without any individual-level patient data leaving the originating centre, therefore preserving patient data privacy. This work builds on the proof-of-concept three-centre atomCAT1 study and describes the protocol for the multi-centre atomCAT2 study, which aims to develop and validate robust prognostic models for three clinically important outcomes in anal cancer following chemoradiotherapy.

METHODS

This is a retrospective multi-centre cohort study, investigating overall survival, locoregional control and freedom from distant metastasis after primary chemoradiotherapy for anal squamous cell carcinoma. Patient data will be extracted and organised at each participating radiotherapy centre (n = 18). Candidate prognostic factors have been identified through literature review and expert opinion. Summary statistics will be calculated and exchanged between centres prior to modelling. The primary analysis will involve developing and validating Cox proportional hazards models across centres for each outcome through distributed learning. Outcomes at specific timepoints of interest and factor effect estimates will be reported, allowing for outcome prediction for future patients.

DISCUSSION

The atomCAT2 study will analyse one of the largest available cross-institutional cohorts of patients with anal cancer treated with chemoradiotherapy. The analysis aims to provide information on current international clinical practice outcomes and may aid the personalisation and design of future anal cancer clinical trials through contributing to a better understanding of patient risk stratification.

Fibroblast Growth Factor 23 and Risk of New Onset Heart Failure With Preserved or Reduced Ejection Fraction: The PREVEND Study

Background

The role of fibroblast growth factor 23 (FGF23) in the development of new‐onset heart failure (HF) with reduced (HFrEF) or preserved ejection fraction (HFpEF) in the general population is unknown. Therefore, we set out to investigate associations of C‐terminal FGF23 with development of new‐onset HF and, more specifically, with HFrEF or HFpEF in a large, prospective, population‐based cohort.

Methods and Results

We studied 6830 participants (aged 53.8±12.1 years; 49.7% men; estimated glomerular filtration rate, 93.1±15.7 mL/min per 1.73 m²) in the community‐based PREVEND (Prevention of Renal and Vascular End‐Stage Disease) study who were free of HF at baseline. Cross‐sectional multivariable linear regression analysis showed that ferritin (standardized β, −0.24; P<0.001) and estimated glomerular filtration rate (standardized β, −0.13; P<0.001) were the strongest independent correlates of FGF23. Multivariable Cox proportional hazard regression was used to study the association between baseline FGF23 and incident HF, HFrEF (ejection fraction ≤40%) or HFpEF (ejection fraction ≥50%). After median follow‐up of 7.4 [IQR 6.9–7.9] years, 227 individuals (3.3%) developed new‐onset HF, of whom 132 had HFrEF and 88 had HFpEF. A higher FGF23 level was associated with an increased risk of incident HF (fully adjusted hazard ratio, 1.29 [95% CI, 1.06–1.57]) and with an increased risk of incident HFrEF (fully adjusted hazard ratio, 1.31 [95% CI, 1.01–1.69]). The association between FGF23 and incident HFpEF lost statistical significance after multivariable adjustment (hazard ratio, 1.22 [95% CI, 0.87–1.71]).

Conclusions

Higher FGF23 is independently associated with new‐onset HFrEF in analyses fully adjusted for cardiovascular risk factors and other potential confounders. The association between FGF23 and incident HFpEF lost statistical significance upon multivariable adjustment.

Multiple imputation for cause-specific Cox models: Assessing methods for estimation and prediction

In studies analyzing competing time-to-event outcomes, interest often lies in both estimating the effects of baseline covariates on the cause-specific hazards and predicting cumulative incidence functions. When missing values occur in these baseline covariates, they may be discarded as part of a complete-case analysis or multiply imputed. In the latter case, the imputations may be performed either compatibly with a substantive model pre-specified as a cause-specific Cox model [substantive model compatible fully conditional specification (SMC-FCS)], or approximately so [multivariate imputation by chained equations (MICE)]. In a large simulation study, we assessed the performance of these three different methods in terms of estimating cause-specific regression coefficients and predicting cumulative incidence functions. Concerning regression coefficients, results provide further support for use of SMC-FCS over MICE, particularly when covariate effects are large and the baseline hazards of the competing events are substantially different. Complete-case analysis also shows adequate performance in settings where missingness is not outcome dependent. With regard to cumulative incidence prediction, SMC-FCS and MICE are performed more similarly, as also evidenced in the illustrative analysis of competing outcomes following a hematopoietic stem cell transplantation. The findings are discussed alongside recommendations for practising statisticians.

Assessing proximate intermediates between ambient temperature, hospital admissions, and mortality in hemodialysis patients

BACKGROUND

Typical thermoregulatory responses to elevated temperatures among healthy individuals include reduced blood pressure and perspiration. Individuals with end-stage kidney disease (ESKD) are susceptible to systemic fluctuations caused by ambient temperature changes that may increase morbidity and mortality. We investigated whether pre-dialysis systolic blood pressure (preSBP) and interdialytic weight gain (IDWG) can independently mediate the association between ambient temperature, all-cause hospital admissions (ACHA), and all-cause mortality (ACM).

METHODS

The study population consisted of ESKD patients receiving hemodialysis treatments at Fresenius Medical Care facilities in Philadelphia County, PA, from 2011 to 2019 (n = 1981). Within a time-to-event framework, we estimated the association between daily maximum dry-bulb temperature (TMAX) and, as separate models, ACHA and ACM during warmer calendar months. Clinically measured preSBP and IDWG responses to temperature increases were estimated using linear mixed effect models. We employed the difference (c-c') method to decompose total effect models for ACHA and ACM using preSBP and IDWG as time-dependent mediators. Covariate adjustments for exposure-mediator and total and direct effect models include age, race, ethnicity, blood pressure medication use, treatment location, preSBP, and IDWG. We considered lags up to two days for exposure and 1-day lag for mediator variables (Lag 2-Lag 1) to assure temporality between exposure-outcome models. Sensitivity analyses for 2-day (Lag 2-only) and 1-day (Lag 1-only) lag structures were also conducted.

RESULTS

Based on Lag 2- Lag 1 temporal ordering, 1 °C increase in daily TMAX was associated with increased hazard of ACHA by 1.4% (adjusted hazard ratio (HR), 1.014; 95% confidence interval, 1.007-1.021) and ACM 7.5% (adjusted HR, 1.075, 1.050-1.100). Short-term lag exposures to 1 °C increase in temperature predicted mean reductions in IDWG and preSBP by 0.013-0.015% and 0.168-0.229 mmHg, respectively. Mediation analysis for ACHA identified significant indirect effects for all three studied pathways (preSBP, IDWG, and preSBP + IDWG) and significant indirect effects for IDWG and conjoined preSBP + IDWG pathways for ACM. Of note, only 1.03% of the association between temperature and ACM was mediated through preSBP. The mechanistic path for IDWG, independent of preSBP, demonstrated inconsistent mediation and, consequently, potential suppression effects in ACHA (-15.5%) and ACM (-6.3%) based on combined pathway models. Proportion mediated estimates from preSBP + IDWG pathways achieved 2.2% and 0.3% in combined pathway analysis for ACHA and ACM outcomes, respectively. Lag 2 discrete-time ACM mediation models exhibited consistent mediation for all three pathways suggesting that 2-day lag in IDWG and preSBP responses can explain 2.11% and 4.41% of total effect association between temperature and mortality, respectively.

CONCLUSION

We corroborated the previously reported association between ambient temperature, ACHA and ACM. Our results foster the understanding of potential physiological linkages that may explain or suppress temperature-driven hospital admissions and mortality risks. Of note, concomitant changes in preSBP and IDWG may have little intermediary effect when analyzed in combined pathway models. These findings advance our assessment of candidate interventions to reduce the impact of outdoor temperature change on ESKD patients.

Overall survival in the real-world and clinical trials: a case study validating external controls in advanced melanoma

Aims: We assessed the suitability of real-world data (RWD) as an external control for analysis of overall survival (OS) compared with clinical trial data (CTD) in advanced melanoma. Methods: OS among adults receiving ipilimumab for advanced melanoma was compared between trials (CTD group) and the Flatiron Health database (RWD group) using Cox models. Adjusted analyses accounted for differences in baseline factors; missing data were addressed through multiple imputation. Results: After adjusting for baseline factors and accounting for missingness, OS was similar in the CTD (n = 241) versus RWD groups (n = 816) (hazard ratio: 0.98; 95% CI: 0.75-1.26). Conclusion: Flatiron Health data is suitable to construct external control groups for OS in advanced melanoma trials after adjusting for baseline factors and missing data.

Peripheral immune cell reactivity and neural response to reward in patients with depression and anhedonia

Increased levels of peripheral cytokines have been previously associated with depression in preclinical and clinical research. Although the precise nature of peripheral immune dysfunction in depression remains unclear, evidence from animal studies points towards a dysregulated response of peripheral leukocytes as a risk factor for stress susceptibility. This study examined dynamic release of inflammatory blood factors from peripheral blood mononuclear cells (PBMC) in depressed patients and associations with neural and behavioral measures of reward processing. Thirty unmedicated patients meeting criteria for unipolar depressive disorder and 21 healthy control volunteers were enrolled. PBMCs were isolated from whole blood and stimulated ex vivo with lipopolysaccharide (LPS). Olink multiplex assay was used to analyze a large panel of inflammatory proteins. Participants completed functional magnetic resonance imaging with an incentive flanker task to probe neural responses to reward anticipation, as well as clinical measures of anhedonia and pleasure including the Temporal Experience of Pleasure Scale (TEPS) and the Snaith-Hamilton Pleasure Scale (SHAPS). LPS stimulation revealed larger increases in immune factors in depressed compared to healthy subjects using an aggregate immune score (t₄₉= 2.83, p = 0.007). Higher peripheral immune score was associated with reduced neural responses to reward anticipation within the ventral striatum (VS) (r = -0.39, p = 0.01), and with reduced anticipation of pleasure as measured with the TEPS anticipatory sub-score (r = -0.318, p = 0.023). Our study provides new evidence suggesting that dynamic hyper-reactivity of peripheral leukocytes in depressed patients is associated with blunted activation of the brain reward system and lower subjective anticipation of pleasure.

Multiple imputation to balance unbalanced designs for two-way analysis of variance

A balanced ANOVA design provides an unambiguous interpretation of the F-tests, and has more power than an unbalanced design. In earlier literature, multiple imputation was proposed to create balance in unbalanced designs, as an alternative to Type-III sum of squares. In the current simulation study we studied four pooled statistics for multiple imputation, namely D₀, D₁, D₂, and D₃ in unbalanced data, and compared them with Type-III sum of squares. Statistics D₁ and D₂ generally performed best regarding Type-I error rates, and had power rates closest to that of Type-III sum of squares. Additionally, for the interaction, D₁ produced power rates higher than Type-III sum of squares. For multiply imputed datasets D₁ and D₂ may be the best methods for pooling the results in multiply imputed datasets, and for unbalanced data, D₁ might be a good alternative to Type-III sum of squares regarding the interaction.

Handling Complex Missing Data Using Random Forest Approach for an Air Quality Monitoring Dataset: A Case Study of Kuwait Environmental Data (2012 to 2018)

In environmental research, missing data are often a challenge for statistical modeling. This paper addressed some advanced techniques to deal with missing values in a data set measuring air quality using a multiple imputation (MI) approach. MCAR, MAR, and NMAR missing data techniques are applied to the data set. Five missing data levels are considered: 5%, 10%, 20%, 30%, and 40%. The imputation method used in this paper is an iterative imputation method, missForest, which is related to the random forest approach. Air quality data sets were gathered from five monitoring stations in Kuwait, aggregated to a daily basis. Logarithm transformation was carried out for all pollutant data, in order to normalize their distributions and to minimize skewness. We found high levels of missing values for NO2 (18.4%), CO (18.5%), PM10 (57.4%), SO2 (19.0%), and O3 (18.2%) data. Climatological data (i.e., air temperature, relative humidity, wind direction, and wind speed) were used as control variables for better estimation. The results show that the MAR technique had the lowest RMSE and MAE. We conclude that MI using the missForest approach has a high level of accuracy in estimating missing values. MissForest had the lowest imputation error (RMSE and MAE) among the other imputation methods and, thus, can be considered to be appropriate for analyzing air quality data.

Adolescent Transitions in Self-Management Strategies and Young Adult Alcohol Use

Individuals use a variety of strategies to manage their thoughts, emotions, and behaviors across the lifespan. In this study, we used latent class analysis to derive distinct subtypes of self-management skills in early adolescence and latent transition analysis to examine whether movement between different classes was associated with later young adult alcohol use. Assessments of behavioral self-control, affective self-regulation, and cognitive self-reinforcement were obtained in the seventh and 10th grades from students participating in two independent drug prevention trials (control group participants only, N = 3,939). Assessment of alcohol use was obtained when participants were young adults (23–26). A model distinguishing four subtypes of self-management skills fit best for both the seventh and 10th grades. While findings indicated modest stability in class structure over time, maintaining class membership characterized by high cognitive self-reinforcement and high affective self-regulation was consistently protective in terms of young adult alcohol use relative to movement from this to other classes. Transitions in class membership involving an expansion of self-management strategies were protective and associated with lower levels of young adult alcohol use and transitions involving a contraction of self-management strategies associated with higher young adult alcohol use. This study illustrates the important use of person-centered techniques to exemplify how typologies of self-management during adolescence can play a protective role in young adult alcohol use.

A tool to predict progression of non-alcoholic fatty liver disease in severely obese patients

BACKGROUND & AIMS

Severely obese patients are a growing population at risk of non-alcoholic fatty liver disease (NAFLD). Considering the increasing burden, a predictive tool of NAFLD progression would be of interest. Our objective was to provide a tool allowing general practitioners to identify and refer the patients most at risk, and specialists to estimate disease progression and adapt the therapeutic strategy.

METHODS

This predictive tool is based on a Markov model simulating steatosis, fibrosis and non-alcoholic steatohepatitis (NASH) evolution. This model was developped from data of 1801 severely obese, bariatric surgery candidates, with histological assessment, integrating duration of exposure to risk factors. It is then able to predict current disease severity in the absence of assessment, and future cirrhosis risk based on current stage.

RESULTS

The model quantifies the impact of sex, body-mass index at 20, diabetes, age of overweight onset, on progression. For example, for 40-year-old severely obese patients seen by the general practitioners: (a) non-diabetic woman overweight at 20, and (b) diabetic man overweight at 10, without disease assessment, the model predicts their current risk to have NASH or F3-F4: for (a) 5.7% and 0.6%, for (b) 16.1% and 10.0% respectively. If those patients have been diagnosed F2 by the specialist, the model predicts the 5-year cirrhosis risk: 1.8% in the absence of NASH and 6.0% in its presence for (a), 10.3% and 26.7% respectively, for (b).

CONCLUSIONS

This model provides a decision-making tool to predict the risk of liver disease that could help manage severely obese patients.

Much ado about nothing: Multiple imputation to balance unbalanced designs for two-way analysis of variance

In earlier literature, multiple imputation was proposed to create balance in unbalanced designs, as an alternative to Type III sum of squares in two-way ANOVA. In the current simulation study we studied four pooled statistics for multiple imputation, namely D₀, D₁, D₂, and D₃ in unbalanced data, and compared these statistics with Type III sum of squares. Statistics D₀ and D₂ generally performed best regarding Type-I error rates, and had power rates closest to that of Type III sum of squares. However, none of the statistics produced power rates higher than Type III sum of squares. The results lead to the conclusion that for multiply imputed datasets D₀ and D₂ may be the best methods for pooling the results of multiparameter estimates in multiply imputed datasets, and that for unbalanced data, Type III sum of square is to be preferred over using multiple imputation in obtaining ANOVA results.

Effect of High-Dose vs Standard-Dose Vitamin D3 Supplementation on Body Composition among Patients with Advanced or Metastatic Colorectal Cancer: A Randomized Trial

Simple Summary

Skeletal muscle and adipose tissue express the vitamin D receptor and may be a mechanism through which vitamin D supplementation slows cancer progression and reduces cancer death. It is unknown if high-dose vitamin D₃ impacts skeletal muscle and adipose tissue, as compared with standard-dose vitamin D₃, in patients with advanced or metastatic colorectal cancer. In this exploratory analysis of a phase II randomized trial, high-dose vitamin D₃ did not lead to changes of body weight, body mass index, muscle area, muscle attenuation, visceral adipose tissue area, or subcutaneous adipose tissue area, as compared with standard-dose vitamin D₃. High-dose vitamin D₃ did not change body composition in patients receiving chemotherapy for advanced or metastatic colorectal cancer.

Abstract

Skeletal muscle and adipose tissue express the vitamin D receptor and may be a mechanism through which vitamin D supplementation slows cancer progression and reduces cancer death. In this exploratory analysis of a double-blind, multicenter, randomized phase II clinical trial, 105 patients with advanced or metastatic colorectal cancer who were receiving chemotherapy were randomized to either high-dose vitamin D₃ (4000 IU) or standard-dose (400 IU) vitamin D₃. Body composition was measured with abdominal computed tomography at enrollment (baseline) and after cycle 8 of chemotherapy (16 weeks). As compared with standard-dose vitamin D₃, high-dose vitamin D₃ did not significantly change body weight [−0.7 kg; (95% CI: −3.5, 2.0)], body mass index [−0.2 kg/m²; (95% CI: −1.2, 0.7)], muscle area [−1.7 cm²; (95% CI: −9.6, 6.3)], muscle attenuation [−0.4 HU; (95% CI: −4.2, 3.2)], visceral adipose tissue area [−7.5 cm²; (95% CI: −24.5, 9.6)], or subcutaneous adipose tissue area [−8.3 cm²; (95% CI: −35.5, 18.9)] over the first 8 cycles of chemotherapy. Among patients with advanced or metastatic colorectal cancer, the addition of high-dose vitamin D₃, vs standard-dose vitamin D₃, to standard chemotherapy did not result in any changes in body composition.

SpaTemHTP: A Data Analysis Pipeline for Efficient Processing and Utilization of Temporal High-Throughput Phenotyping Data

The rapid development of phenotyping technologies over the last years gave the opportunity to study plant development over time. The treatment of the massive amount of data collected by high-throughput phenotyping (HTP) platforms is however an important challenge for the plant science community. An important issue is to accurately estimate, over time, the genotypic component of plant phenotype. In outdoor and field-based HTP platforms, phenotype measurements can be substantially affected by data-generation inaccuracies or failures, leading to erroneous or missing data. To solve that problem, we developed an analytical pipeline composed of three modules: detection of outliers, imputation of missing values, and mixed-model genotype adjusted means computation with spatial adjustment. The pipeline was tested on three different traits (3D leaf area, projected leaf area, and plant height), in two crops (chickpea, sorghum), measured during two seasons. Using real-data analyses and simulations, we showed that the sequential application of the three pipeline steps was particularly useful to estimate smooth genotype growth curves from raw data containing a large amount of noise, a situation that is potentially frequent in data generated on outdoor HTP platforms. The procedure we propose can handle up to 50% of missing values. It is also robust to data contamination rates between 20 and 30% of the data. The pipeline was further extended to model the genotype time series data. A change-point analysis allowed the determination of growth phases and the optimal timing where genotypic differences were the largest. The estimated genotypic values were used to cluster the genotypes during the optimal growth phase. Through a two-way analysis of variance (ANOVA), clusters were found to be consistently defined throughout the growth duration. Therefore, we could show, on a wide range of scenarios, that the pipeline facilitated efficient extraction of useful information from outdoor HTP platform data. High-quality plant growth time series data is also provided to support breeding decisions. The R code of the pipeline is available at https://github.com/ICRISAT-GEMS/SpaTemHTP.

Various Types of Perceived Job Discrimination and Sleep Health Among Working Women: Findings From the Sister Study

Job discrimination, a social stressor, may lead to sleep health disparities among workers; yet, limited research has examined this relationship and specific sources of job discrimination. We used a US sample of working women (n = 26,085), participants in the Sister Study (2008-2016), to examine the associations of perceived job discrimination due to sex, race, age, health conditions, and/or sexual orientation with sleep health. Cross-sectionally, linear or logistic regression models revealed that each source of job discrimination was independently associated with different sleep problems after controlling for other sources of job discrimination. Longitudinally, among participants without short sleep (<7 hours/night) at time 1 (2012-2014), age-specific job discrimination was associated with 21% increased odds of new-onset short sleep (odds ratio = 1.21, 95% confidence interval: 1.03, 1.43) at time 2 (2014-2016). Among those without insomnia symptoms at time 1, race-specific job discrimination was associated with 37% increased odds of new-onset insomnia symptoms (odds ratio = 1.37, 95% confidence interval: 1.07, 1.75) at time 2. Sex- and health-specific job discrimination also predicted new-onset sleepiness. There were dose-response relationships such that a greater number of sources of job discrimination (≥3) was associated with greater odds of prevalent and incident sleep problems. Perceived job discrimination may contribute to working women's poor sleep health over time, raising concerns about sleep health disparities emanating from the workplace.

Image Based Data Mining Using Per-voxel Cox Regression

Image Based Data Mining (IBDM) is a novel analysis technique allowing the interrogation of large amounts of routine radiotherapy data. Using this technique, unexpected correlations have been identified between dose close to the prostate and biochemical relapse, and between dose to the base of the heart and survival in lung cancer. However, most analyses to date have considered only dose when identifying a region of interest, with confounding variables accounted for post-hoc, most often using a multivariate Cox regression. In this work, we introduce a novel method to account for confounding variables directly in the analysis, by performing a Cox regression in every voxel of the dose distribution, and apply it in the analysis of a large cohort of lung cancer patients. Our method produces three-dimensional maps of hazard for clinical variables, accounting for dose at each spatial location in the patient. Results confirm that a region of interest exists in the base of the heart where those patients with poor performance status (PS), PS > 1, have a stronger adverse reaction to incidental dose, but that the effect changes when considering other clinical variables, with patient age becoming dominant. Analyses such as this will help shape future clinical trials in which hypotheses generated by the analysis will be tested.

[Predictive Mean Matching as an alternative imputation method to hot deck in Vigitel]

This study aimed to describe the estimated means for weight, height, and body mass index (BMI) according to two imputation methods, using data from Vigitel (Risk and Protective Factors Surveillance System for Chronic Non-Communicable Diseases Through Telephone Interview). This was a cross-sectional study that used secondary data from the Vigitel survey from 2006 to 2017. The two imputation methods used in the study were hot deck and Predictive Mean Matching (PMM). The weight and height variables imputed by hot deck were provided by Vigitel. Two models were conducted with PMM: (i) explanatory variables - city, sex, age in years, race/color, and schooling; (ii) explanatory variables - city, sex, and age in years. Weight and height were the outcome variables in the two models. PMM combines linear regression and random selection of the value for imputation. Linear prediction is used as a measure of distance between the missing value and the possible donors, thereby creating the virtual space with the candidate cases for yielding the value for imputation. One of the candidates from the pool is randomly selected, and its value is assigned to the missing unit. BMI was calculated by dividing weight in kilograms by height squared. The result shows the means and standard deviations for weight, height, and BMI according to imputation method and year. The estimates used the survey module from Stata, which considers the sampling effects. The mean values for weight, height, and BMI estimated by hot deck and PMM were similar. The results with the Vigitel data suggest the applicability of PMM to the set of health surveys.

Does entry to center-based childcare affect gut microbial colonization in young infants?

Entry to center-based childcare (CC) at three months of life can be an important challenge for infants as it includes major stressors such as long maternal separations and frequently changing caregivers. Stress and the new environment may in turn alter the composition of the gut microbiota with possible implications for future health outcomes. As part of an ongoing longitudinal study, we investigated whether CC, as compared to being cared for by the parents at home, alters the composition of the gut microbiota, while accounting for known covariates of the infant gut microbiota. Stool samples of infants who entered CC (n = 49) and control infants (n = 49) were obtained before and four weeks after CC entrance. Using Redundancy analysis, Random Forests and Bayesian linear models we found that infant gut microbiota was not affected in a uniform way by entry to CC. In line with the literature, breastfeeding, birth mode, age, and the presence of siblings were shown to significantly impact the microbial composition.

Comparative Assessment of Predictive Performance of PRECISE-DAPT, CRUSADE, and ACUITY Scores in Risk Stratifying 30-Day Bleeding Events

Background  The utility of the PRECISE-DAPT score in predicting short-term major bleeding, either alone, or in comparison with the CRUSADE and ACUITY scores, has not been investigated. This analysis compared the predictive performances of the three bleeding scores in stratifying the risk of 30-day major bleeding postpercutaneous coronary intervention in patients with dual-antiplatelet therapy.

Methods  In this post hoc subanalysis of the GLOBAL LEADERS trial, the primary safety objective (bleeding according to the Bleeding Academic Research Consortium [BARC] criteria [type 3 or 5]) was assessed at 30 days according to the three scores in the overall population, and in patients with acute (ACS) and chronic coronary syndrome (CCS).

Results  In a total of 15,968 patients, we calculated all three scores in 14,709 (92.1%). Irrespective of clinical presentation, the PRECISE-DAPT ( c -statistics: 0.648, 0.653, and 0.641, respectively), CRUSADE ( c -statistics: 0.641, 0.639, and 0.644, respectively), and ACUITY ( c -statistics: 0.633, 0.638, and 0.623, respectively) scores were no significant between-score differences in discriminatory performance for BARC 3 or 5 bleeding up to 30 days, and similarly the PRECISE-DAPT score had a comparable discriminative capacity according to the integrated discrimination improvement when compared with the other scores. In ACS, the CRUSADE score had a poor calibration ability (Hosmer–Lemeshow goodness-of-fit [GOF] chi-square = 15.561, p  = 0.049), whereas in CCS, the PRECISE-DAPT score had poor calibration (GOF chi-square = 15.758, p  = 0.046).

Conclusion  The PRECISE-DAPT score might be clinically useful in the overall population and ACS patients for the prediction of short-term major bleeding considering its discriminative and calibration abilities.

School Climate as a Universal Intervention to Prevent Substance Use Initiation in Early Adolescence: A Longitudinal Study

Initiation of substance use often starts during adolescence, with tobacco and alcohol use frequently preceding the use of marijuana and other illicit drugs. Studies suggest that a positive school climate may prevent substance use while promoting healthy student behaviors. The purpose of this study was to determine the longitudinal associations between school climate and substance use initiation in a group of middle school students. Parallel latent growth curve modeling was used to examine changes among study variables longitudinally using a sample of 2,097 sixth-, seventh-, and eighth-grade students across 16 regional schools located in three counties in West Virginia. Results suggest that a positive school climate may prevent substance use initiation (β = -0.07 to -0.25, p < .01). However, perceptions of school climate decreased on their own over time (β = -0.28 to -0.66, p < .01). Furthermore, substance use initiation also increased as students grew older (β = 0.96 to 0.99, p < .01) and reduced the effects of school climate longitudinally (β = -0.07 to -0.24, p < .01). Early substance use initiation may be a warning sign of other underlying student issues and requires additional school support to foster student success. Findings suggest that a positive school climate may delay substance use initiation and promote school success. School climate may, therefore, be useful as an intervention to support school-based health promotion.

Standardized Regression Coefficients and Newly Proposed Estimators for [Formula: see text] in Multiply Imputed Data

Whenever statistical analyses are applied to multiply imputed datasets, specific formulas are needed to combine the results into one overall analysis, also called combination rules. In the context of regression analysis, combination rules for the unstandardized regression coefficients, the t-tests of the regression coefficients, and the F-tests for testing [Formula: see text] for significance have long been established. However, there is still no general agreement on how to combine the point estimators of [Formula: see text] in multiple regression applied to multiply imputed datasets. Additionally, no combination rules for standardized regression coefficients and their confidence intervals seem to have been developed at all. In the current article, two sets of combination rules for the standardized regression coefficients and their confidence intervals are proposed, and their statistical properties are discussed. Additionally, two improved point estimators of [Formula: see text] in multiply imputed data are proposed, which in their computation use the pooled standardized regression coefficients. Simulations show that the proposed pooled standardized coefficients produce only small bias and that their 95% confidence intervals produce coverage close to the theoretical 95%. Furthermore, the simulations show that the newly proposed pooled estimates for [Formula: see text] are less biased than two earlier proposed pooled estimates.

School Climate as an Intervention to Reduce Academic Failure and Educate the Whole Child: A Longitudinal Study

BACKGROUND

Preventing student academic failure is crucial to student health and life success. Previous studies suggest a positive school climate may reduce students' risk for academic failure and contribute to academic success. The purpose of this study was to determine the longitudinal associations between school climate and academic grades in a group of middle school students who transition into high school.

METHODS

Parallel latent growth curve modeling was used to examine changes among study variables longitudinally using a sample of 2604 in 6th, 7th, and 8th-grade students across 16 regional schools located in 3 counties in West Virginia.

RESULTS

Students with higher perceptions of a positive school climate exhibited sustained or improved academic achievement over time (β = 0.22 to 0.30, p < .01). Higher positive perceptions of school climate appear to sustain students who earn As/Bs (β = 0.20 to 0.27, p < .01) and strengthen students who earn Cs/Ds/Fs (β = -0.16 to -0.46, p < .05).

CONCLUSIONS

Positive student perceptions of school climate may sustain high academic performance while strengthening students who earn Cs/Ds/Fs. School climate may be useful as an intervention to support school-based health promotion to reduce the achievement gap in the United States.

Approaches for missing covariate data in logistic regression with MNAR sensitivity analyses

Data with missing covariate values but fully observed binary outcomes are an important subset of the missing data challenge. Common approaches are complete case analysis (CCA) and multiple imputation (MI). While CCA relies on missing completely at random (MCAR), MI usually relies on a missing at random (MAR) assumption to produce unbiased results. For MI involving logistic regression models, it is also important to consider several missing not at random (MNAR) conditions under which CCA is asymptotically unbiased and, as we show, MI is also valid in some cases. We use a data application and simulation study to compare the performance of several machine learning and parametric MI methods under a fully conditional specification framework (MI-FCS). Our simulation includes five scenarios involving MCAR, MAR, and MNAR under predictable and nonpredictable conditions, where "predictable" indicates missingness is not associated with the outcome. We build on previous results in the literature to show MI and CCA can both produce unbiased results under more conditions than some analysts may realize. When both approaches were valid, we found that MI-FCS was at least as good as CCA in terms of estimated bias and coverage, and was superior when missingness involved a categorical covariate. We also demonstrate how MNAR sensitivity analysis can build confidence that unbiased results were obtained, including under MNAR-predictable, when CCA and MI are both valid. Since the missingness mechanism cannot be identified from observed data, investigators should compare results from MI and CCA when both are plausibly valid, followed by MNAR sensitivity analysis.

Recovery of health-related quality of life after burn injuries: An individual participant data meta-analysis

Background

A prominent outcome measure within burn care is health related quality of life (HRQL). Until now, no model for long-term recovery of HRQL exists for adult burn patients which requires large samples with repeated measurements. Re-use and the combination of existing data is a way to achieve larger data samples that enable the estimation of long-term recovery models. The aim of this secondary data analysis was to assess the recovery of HRQL after a burn injury over time.

Methods and findings

Data from ten European studies on generic HRQL assessed in adult burn patients (either with the EQ-5D or SF-36) from five different countries were merged into one dataset. SF-36 outcomes were transformed into EQ-5D outcomes. A 24-month recovery of HRQL (EQ-5D utility) was modeled using a linear mixed-effects model and adjusted for important patient and burn characteristics. Subgroups of patients with mild and intermediate burns (≤20% total body surface area (TBSA) burned) and with major burns (>20% TBSA burned) were compared. The combined database included 1687 patients with a mean age of 43 (SD 15) years and a median %TBSA burned of 9% (IQR 4–18). There was large improvement in HRQL up to six months after burns, and HRQL remained relatively stable afterwards (studied up to 24 months post burn). However, the estimated EQ-5D utility scores remained below the norm scores of the general population. In this large sample, females, patients with a long hospital stay and patients with major burns had a delayed and worse recovery. The proportion of patients that reported problems for the EQ-5D dimensions ranged from 100% (pain/discomfort at baseline in patients with major burns) to 10% (self-care ≥3 months after injury in patients with mild and intermediate burns). After 24 months, both subgroups of burn patients did not reach the level of the general population in the dimensions pain/discomfort and anxiety/depression, and patients with major burns in the dimension usual activities. A main limitation of the study includes that the variables in the model were limited to age, gender, %TBSA, LOS and time since burn as these were the only variables available in all datasets.

Conclusions

The 24-month recovery model can be used in clinical practice to inform patients on expected HRQL outcomes and provide clinicians insights into the expected recovery of HRQL. In this way, a delayed recovery can be recognized in an early stage and timely interventions can be started in order to improve patient outcomes. However, external validation of the developed model is needed before implementation into clinical practice. Furthermore, our study showed the benefit of secondary data usage within the field of burns.

Predicting Prostate Cancer Death with Different Pretreatment Risk Stratification Tools: A Head-to-head Comparison in a Nationwide Cohort Study

BACKGROUND

Numerous pretreatment risk classification tools are available for prostate cancer. Which tool is best in predicting prostate cancer death is unclear.

OBJECTIVE

To systematically compare the prognostic performance of the most commonly used pretreatment risk stratification tools for prostate cancer.

DESIGN, SETTING, AND PARTICIPANTS

A nationwide cohort study was conducted, including 154 811 men in Prostate Cancer data Base Sweden (PCBaSe) 4.0 diagnosed with nonmetastatic prostate cancer during 1998-2016 and followed through 2016.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

We compared the D'Amico, National Institute for Health and Care Excellence (NICE), European Association of Urology (EAU), Genito-Urinary Radiation Oncologists of Canada (GUROC), American Urological Association (AUA), National Comprehensive Cancer Network (NCCN), and Cambridge Prognostic Groups (CPG) risk group systems; the Cancer of the Prostate Risk Assessment (CAPRA) score; and the Memorial Sloan Kettering Cancer Center (MSKCC) nomogram in predicting prostate cancer death by estimating the concordance index (C-index) and the observed versus predicted cumulative incidences at different follow-up times.

RESULTS AND LIMITATIONS

A total of 139 515 men were included in the main analysis, of whom 15 961 died from prostate cancer during follow-up. The C-index at 10 yr of follow-up ranged from 0.73 (95% confidence interval [CI]: 0.72-0.73) to 0.81 (95% CI: 0.80-0.81) across the compared tools. The MSKCC nomogram (C-index: 0.81, 95% CI: 0.80-0.81), CAPRA score (C-index: 0.80, 95% CI: 0.79-0.81), and CPG system (C-index: 0.78, 95% CI: 0.78-0.79) performed the best. The order of performance between the tools remained in analyses stratified by primary treatment and year of diagnosis. The predicted cumulative incidences were close to the observed ones, with some underestimation at 5 yr. It is a limitation that the study was conducted solely in a Swedish setting (ie, case mix).

CONCLUSIONS

The MSKCC nomogram, CAPRA score, and CPG risk grouping system performed better in discriminating prostate cancer death than the D'Amico and D'Amico-derived systems (NICE, GUROC, EAU, AUA, and NCCN). Use of these tools may improve clinical decision making.

PATIENT SUMMARY

There are numerous pretreatment risk classification tools that can aid treatment decision for prostate cancer. We systematically compared the prognostic performance of the most commonly used tools in a large cohort of Swedish men with prostate cancer. The Memorial Sloan Kettering Cancer Center nomogram, Cancer of the Prostate Risk Assessment score, and Cambridge Prognostic Groups performed best in predicting prostate cancer death. The use of these tools may improve treatment decisions.

Methodological standards for the development and evaluation of clinical prediction rules: a review of the literature

Clinical prediction rules (CPRs) that predict the absolute risk of a clinical condition or future outcome for individual patients are abundant in the medical literature; however, systematic reviews have demonstrated shortcomings in the methodological quality and reporting of prediction studies. To maximise the potential and clinical usefulness of CPRs, they must be rigorously developed and validated, and their impact on clinical practice and patient outcomes must be evaluated. This review aims to present a comprehensive overview of the stages involved in the development, validation and evaluation of CPRs, and to describe in detail the methodological standards required at each stage, illustrated with examples where appropriate. Important features of the study design, statistical analysis, modelling strategy, data collection, performance assessment, CPR presentation and reporting are discussed, in addition to other, often overlooked aspects such as the acceptability, cost-effectiveness and longer-term implementation of CPRs, and their comparison with clinical judgement. Although the development and evaluation of a robust, clinically useful CPR is anything but straightforward, adherence to the plethora of methodological standards, recommendations and frameworks at each stage will assist in the development of a rigorous CPR that has the potential to contribute usefully to clinical practice and decision-making and have a positive impact on patient care.

A systematic review of methodological quality of model development studies predicting prognostic outcome for resectable pancreatic cancer

OBJECTIVES

To assess the methodological quality of prognostic model development studies pertaining to post resection prognosis of pancreatic ductal adenocarcinoma (PDAC).

DESIGN/SETTING

A narrative systematic review of international peer reviewed journals DATA SOURCE: Searches were conducted of: MEDLINE, Embase, PubMed, Cochrane database and Google Scholar for predictive modelling studies applied to the outcome of prognosis for patients with PDAC post resection. Predictive modelling studies in this context included prediction model development studies with and without external validation and external validation studies with model updating. Data was extracted following the Checklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies (CHARMS) checklist.

PRIMARY AND SECONDARY OUTCOME MEASURES

Primary outcomes were all components of the CHARMS checklist. Secondary outcomes included frequency of variables included across predictive models.

RESULTS

263 studies underwent full text review. 15 studies met the inclusion criteria. 3 studies underwent external validation. Multivariable Cox proportional hazard regression was the most commonly employed modelling method (n=13). 10 studies were based on single centre databases. Five used prospective databases, seven used retrospective databases and three used cancer data registry. The mean number of candidate predictors was 19.47 (range 7 to 50). The most commonly included variables were tumour grade (n=9), age (n=8), tumour stage (n=7) and tumour size (n=5). Mean sample size was 1367 (range 50 to 6400). 5 studies reached statistical power. None of the studies reported blinding of outcome measurement for predictor values. The most common form of presentation was nomograms (n=5) and prognostic scores (n=5) followed by prognostic calculators (n=3) and prognostic index (n=2).

CONCLUSIONS

Areas for improvement in future predictive model development have been highlighted relating to: general aspects of model development and reporting, applicability of models and sources of bias.

TRIAL REGISTRATION NUMBER

CRD42018105942.

Patient-reported scar quality of adults after burn injuries: A five-year multicenter follow-up study

Scar formation is an important adverse consequence of burns. How patients appraise their scar quality is often studied shortly after sustaining the injury, but information in the long‐term is scarce. Our aim was, therefore, to evaluate long‐term patient‐reported quality of burn scars. Adults with a burn center admission of ≥1 day between August 2011 and September 2012 were invited to complete a questionnaire on long‐term consequences of burns. We enriched this sample with patients with severe burns (>20% total body surface area [TBSA] burned or TBSA full thickness >5%) treated between January 2010 and March 2013. Self‐reported scar quality was assessed with the Patient Scale of the Patient and Observer Scar Assessment Scale (POSAS). Patients completed this scale for their—in their opinion—most severe scar ≥5 years after burns. This study included 251 patients with a mean %TBSA burned of 10%. The vast majority (91.4%) reported at least minor differences with normal skin (POSAS item score ≥2) on one or more scar characteristics and 78.9% of the patients’ overall opinion was that their scar deviated from normal skin. Patients with severe burns had higher POSAS scores, representing worse scar quality, than patients with mild/intermediate burns, except for color, which was high in both groups. A longer hospital stay predicted reduced scar quality (both mean POSAS and mean overall opinion of the scar) in multivariate analyses. In addition, female gender was also associated with a poorer overall opinion of the scar. In conclusion, this study provides new insights in long‐term scar quality. Scars differed from normal skin in a large part of the burn population more than 5 years after burns, especially in those with severe burns. Female gender is associated with a poorer patients’ overall opinion of their scar, which may be an indication of gender differences in perception of scar quality after burns.

Codability of industry and occupation information from cancer registry records: Differences by patient demographics, casefinding source, payor, and cancer type

INTRODUCTION

Industry and occupation (I&O) information collected by cancer registries is useful for assessing associations among jobs and malignancies. However, systematic differences in I&O availability can bias findings.

METHODS

Codability by patient demographics, payor, identifying (casefinding) source, and cancer site was assessed using I&O text from first primaries diagnosed 2011-2012 and reported to California Cancer Registry. I&O were coded to a U.S. Census code or classified as blank/inadequate/unknown, retired, or not working for pay.

RESULTS

Industry was codable for 37% of cases; 50% had "unknown" and 9% "retired" instead of usual industry. Cases initially reported by hospitals, covered by preferred providers, or with known occupational etiology had highest codable industry; cases from private pathology laboratories, with Medicaid, or diagnosed in outpatient settings had least. Occupation results were similar.

CONCLUSIONS

Recording usual I&O for retirees and improving linkages for reporting entities without patient access would improve I&O codability and research validity.

Joint modeling of multivariate hearing thresholds measured longitudinally at multiple frequencies

Pure-tone thresholds are used to estimate hearing acuity and, when measured longitudinally, can characterize age-related changes in hearing. Measured at multiple-frequencies, multiple-irregular time points, for right and left ears, these longitudinal studies of age-related hearing loss produce data of inherent complexity due to: 1) multivariate outcomes at different frequencies; 2) longitudinal measurements taken at subject-specific time intervals; and 3) inter-ear correlations due to clustering and nesting. To address limitations in existing methods, we propose a multivariate generalized linear mixed model(mGLMM) and assess its performance. We demonstrate its application using a unique dataset from a cohort study of age-related hearing loss.

Shed urinary ALCAM is an independent prognostic biomarker of three-year overall survival after cystectomy in patients with bladder cancer

Proteins involved in tumor cell migration can potentially serve as markers of invasive disease. Activated Leukocyte Cell Adhesion Molecule (ALCAM) promotes adhesion, while shedding of its extracellular domain is associated with migration. We hypothesized that shed ALCAM in biofluids could be predictive of progressive disease. ALCAM expression in tumor (n = 198) and shedding in biofluids (n = 120) were measured in two separate VUMC bladder cancer cystectomy cohorts by immunofluorescence and enzyme-linked immunosorbent assay, respectively. The primary outcome measure was accuracy of predicting 3-year overall survival (OS) with shed ALCAM compared to standard clinical indicators alone, assessed by multivariable Cox regression and concordance-indices. Validation was performed by internal bootstrap, a cohort from a second institution (n = 64), and treatment of missing data with multiple-imputation. While ALCAM mRNA expression was unchanged, histological detection of ALCAM decreased with increasing stage (P = 0.004). Importantly, urine ALCAM was elevated 17.0-fold (P < 0.0001) above non-cancer controls, correlated positively with tumor stage (P = 0.018), was an independent predictor of OS after adjusting for age, tumor stage, lymph-node status, and hematuria (HR, 1.46; 95% CI, 1.03–2.06; P = 0.002), and improved prediction of OS by 3.3% (concordance-index, 78.5% vs. 75.2%). Urine ALCAM remained an independent predictor of OS after accounting for treatment with Bacillus Calmette-Guerin, carcinoma in situ, lymph-node dissection, lymphovascular invasion, urine creatinine, and adjuvant chemotherapy (HR, 1.10; 95% CI, 1.02–1.19; P = 0.011). In conclusion, shed ALCAM may be a novel prognostic biomarker in bladder cancer, although prospective validation studies are warranted. These findings demonstrate that markers reporting on cell motility can act as prognostic indicators.

Identification of patients with stable chest pain deriving minimal value from coronary computed tomography angiography: An external validation of the PROMISE minimal-risk tool

BACKGROUND

The PROspective Multicenter Imaging Study for Evaluation of chest pain (PROMISE) minimal-risk tool was recently developed to identify patients with suspected stable angina at very low risk of coronary artery disease (CAD) and clinical events. We assessed the external validity of this tool within the context of the Scottish Computed Tomography of the HEART (SCOT-HEART) multicenter randomised controlled trial of patients with suspected stable angina due to coronary disease.

METHODS

The minimal-risk tool was applied to 1764 patients with complete imaging and follow-up data. External validity was compared with the guideline-endorsed CAD Consortium (CADC) risk score and determined through tests of model discrimination and calibration.

RESULTS

A total of 531 (30.1%, mean age 52.4years, female 62.0%) patients were classified as minimal-risk. Compared to the remainder of the validation cohort, this group had lower estimated pre-test probability of coronary disease according to the CADC model (30.0% vs 47.0%, p<0.001). The PROMISE minimal-risk tool improved discrimination compared with the CADC model (c-statistic 0.785 vs 0.730, p<0.001) and was improved further following re-estimation of covariate coefficients (c-statistic 0.805, p<0.001). Model calibration was initially poor (χ2 197.6, Hosmer-Lemeshow [HL] p<0.001), with significant overestimation of probability of minimal risk, but improved significantly following revision of the PROMISE minimal-risk intercept and covariate coefficients (χ2 5.6, HL p=0.692).

CONCLUSION AND RELEVANCE

Despite overestimating the probability of minimal-risk, the PROMISE minimal-risk tool outperforms the CADC model with regards to prognostic discrimination in patients with suspected stable angina, and may assist clinicians in decisions regarding non-invasive testing.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT01149590.

Multiple Score Comparison: a network meta-analysis approach to comparison and external validation of prognostic scores

BACKGROUND

Prediction models and prognostic scores have been increasingly popular in both clinical practice and clinical research settings, for example to aid in risk-based decision making or control for confounding. In many medical fields, a large number of prognostic scores are available, but practitioners may find it difficult to choose between them due to lack of external validation as well as lack of comparisons between them.

METHODS

Borrowing methodology from network meta-analysis, we describe an approach to Multiple Score Comparison meta-analysis (MSC) which permits concurrent external validation and comparisons of prognostic scores using individual patient data (IPD) arising from a large-scale international collaboration. We describe the challenges in adapting network meta-analysis to the MSC setting, for instance the need to explicitly include correlations between the scores on a cohort level, and how to deal with many multi-score studies. We propose first using IPD to make cohort-level aggregate discrimination or calibration scores, comparing all to a common comparator. Then, standard network meta-analysis techniques can be applied, taking care to consider correlation structures in cohorts with multiple scores. Transitivity, consistency and heterogeneity are also examined.

RESULTS

We provide a clinical application, comparing prognostic scores for 3-year mortality in patients with chronic obstructive pulmonary disease using data from a large-scale collaborative initiative. We focus on the discriminative properties of the prognostic scores. Our results show clear differences in performance, with ADO and eBODE showing higher discrimination with respect to mortality than other considered scores. The assumptions of transitivity and local and global consistency were not violated. Heterogeneity was small.

CONCLUSIONS

We applied a network meta-analytic methodology to externally validate and concurrently compare the prognostic properties of clinical scores. Our large-scale external validation indicates that the scores with the best discriminative properties to predict 3 year mortality in patients with COPD are ADO and eBODE.

A Concept-Wide Association Study of Clinical Notes to Discover New Predictors of Kidney Failure

BACKGROUND AND OBJECTIVES

Identifying predictors of kidney disease progression is critical toward the development of strategies to prevent kidney failure. Clinical notes provide a unique opportunity for big data approaches to identify novel risk factors for disease.

DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS

We used natural language processing tools to extract concepts from the preceding year's clinical notes among patients newly referred to a tertiary care center's outpatient nephrology clinics and retrospectively evaluated these concepts as predictors for the subsequent development of ESRD using proportional subdistribution hazards (competing risk) regression. The primary outcome was time to ESRD, accounting for a competing risk of death. We identified predictors from univariate and multivariate (adjusting for Tangri linear predictor) models using a 5% threshold for false discovery rate (q value <0.05). We included all patients seen by an adult outpatient nephrologist between January 1, 2004 and June 18, 2014 and excluded patients seen only by transplant nephrology, with preexisting ESRD, with fewer than five clinical notes, with no follow-up, or with no baseline creatinine values.

RESULTS

Among the 4013 patients selected in the final study cohort, we identified 960 concepts in the unadjusted analysis and 885 concepts in the adjusted analysis. Novel predictors identified included high-dose ascorbic acid (adjusted hazard ratio, 5.48; 95% confidence interval, 2.80 to 10.70; q<0.001) and fast food (adjusted hazard ratio, 4.34; 95% confidence interval, 2.55 to 7.40; q<0.001).

CONCLUSIONS

Novel predictors of human disease may be identified using an unbiased approach to analyze text from the electronic health record.

Competing risk of death and end-stage renal disease in incident chronic kidney disease (stages 3 to 5): the EPIRAN community-based study

BACKGROUND

Although chronic kidney disease (CKD) affects a growing number of people, epidemiologic data on incident CKD in the general population are scarce. Screening strategies to increase early CKD detection have been developed.

METHODS

From a community-based sample of 4,409 individuals residing in a well-defined geographical area, we determined the number of patients having a first serum creatinine value ≥1.7 mg/dL and present for at least 3 months that allowed us to calculate an annual incidence rate of CKD (stages 3 to 5). CKD (stages 3 to 5) was defined by estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2. We also described the primary care, outcomes and risk factors associated with outcomes using competing risks analyses for these CKD patients.

RESULTS

A total of 631 incident CKD patients (stages 3 to 5) were followed-up until the occurrence of death and dialysis initiation for more than 3 years. The annual incidence rate of CKD (stages 3 to 5) was estimated at 977.7 per million inhabitants. Analyses were performed on 514 patients with available medical data. During the study, 155 patients (30.2 %) were referred to a nephrologist, 193 (37.5 %) died and 58 (11.3 %) reached end-stage renal disease and initiated dialysis. A total of 139 patients (27.6 %) had a fast decline of their renal function, 92 (18.3 %) a moderate decline and the 272 remaining patients had a physiological decline (21.1 %) or a small improvement of their renal function (33.0 %). Predictors of death found in both Cox and Fine-Gray multivariable regression models included age at diagnosis, anemia, active neoplasia and chronic heart failure, but not a low glomerular filtration rate (GFR). Age at diagnosis, anemia and a low GFR were independently associated with dialysis initiation in Cox model, but anemia was not found to be a risk factor for dialysis initiation in Fine-Gray model.

CONCLUSIONS

This large cohort study provided useful epidemiological data on incident CKD (stages 3 to 5) and stressed the need to improve the hands-on implementation of clinical practice guidelines for the evaluation and the management of CKD in primary care.