Survival prediction models: an introduction to discrete-time modeling

Prioritising deteriorating patients using time-to-event analysis: prediction model development and internal-external validation

BACKGROUND

Binary classification models are frequently used to predict clinical deterioration, however they ignore information on the timing of events. An alternative is to apply time-to-event models, augmenting clinical workflows by ranking patients by predicted risks. This study examines how and why time-to-event modelling of vital signs data can help prioritise deterioration assessments using lift curves, and develops a prediction model to stratify acute care inpatients by risk of clinical deterioration.

METHODS

We developed and validated a Cox regression for time to in-hospital mortality. The model used time-varying covariates to estimate the risk of clinical deterioration. Adult inpatient medical records from 5 Australian hospitals between 1 January 2019 and 31 December 2020 were used for model development and validation. Model discrimination and calibration were assessed using internal-external cross validation. A discrete-time logistic regression model predicting death within 24 h with the same covariates was used as a comparator to the Cox regression model to estimate differences in predictive performance between the binary and time-to-event outcome modelling approaches.

RESULTS

Our data contained 150,342 admissions and 1016 deaths. Model discrimination was higher for Cox regression than for discrete-time logistic regression, with cross-validated AUCs of 0.96 and 0.93, respectively, for mortality predictions within 24 h, declining to 0.93 and 0.88, respectively, for mortality predictions within 1 week. Calibration plots showed that calibration varied by hospital, but this can be mitigated by ranking patients by predicted risks.

CONCLUSION

Time-varying covariate Cox models can be powerful tools for triaging patients, which may lead to more efficient and effective care in time-poor environments when the times between observations are highly variable.

18F-Fluorodeoxyglucose Positron Emission Tomography-Based Risk Score Model for Prediction of Five-Year Survival Outcome after Curative Resection of Non-Small-Cell Lung Cancer

The aim of our retrospective study is to develop and assess an imaging-based model utilizing 18F-FDG PET parameters for predicting the five-year survival in non-small-cell lung cancer (NSCLC) patients after curative surgery. A total of 361 NSCLC patients who underwent curative surgery were assigned to the training set (n = 253) and the test set (n = 108). The LASSO regression model was used to construct a PET-based risk score for predicting five-year survival. A hybrid model that combined the PET-based risk score and clinical variables was developed using multivariate logistic regression analysis. The predictive performance was determined by the area under the curve (AUC). The individual features with the best predictive performances were co-occurrence_contrast (AUC = 0.675) and SUL peak (AUC = 0.671). The PET-based risk score was identified as an independent predictor after adjusting for clinical variables (OR 5.231, 95% CI 1.987-6.932; p = 0.009). The hybrid model, which integrated clinical variables, significantly outperformed the PET-based risk score alone in predictive accuracy (AUC = 0.771 vs. 0.696, p = 0.022), a finding that was consistent in the test set. The PET-based risk score, especially when integrated with clinical variables, demonstrates good predictive ability for five-year survival in NSCLC patients following curative surgery.

Effects of long-term exposure to outdoor air pollution on COVID-19 incidence: A population-based cohort study accounting for SARS-CoV-2 exposure levels in the Netherlands

Several studies have linked air pollution to COVID-19 morbidity and severity. However, these studies do not account for exposure levels to SARS-CoV-2, nor for different sources of air pollution. We analyzed individual-level data for 8.3 million adults in the Netherlands to assess associations between long-term exposure to ambient air pollution and SARS-CoV-2 infection (i.e., positive test) and COVID-19 hospitalisation risks, accounting for spatiotemporal variation in SARS-CoV-2 exposure levels during the first two major epidemic waves (February 2020-February 2021). We estimated average annual concentrations of PM10, PM2.5 and NO2 at residential addresses, overall and by PM source (road traffic, industry, livestock, other agricultural sources, foreign sources, other Dutch sources), at 1 × 1 km resolution, and weekly SARS-CoV-2 exposure at municipal level. Using generalized additive models, we performed interval-censored survival analyses to assess associations between individuals' average exposure to PM10, PM2.5 and NO2 in the three years before the pandemic (2017-2019) and COVID-19-outcomes, adjusting for SARS-CoV-2 exposure, individual and area-specific confounders. In single-pollutant models, per interquartile (IQR) increase in exposure, PM10 was associated with 7% increased infection risk and 16% increased hospitalisation risk, PM2.5 with 8% increased infection risk and 18% increased hospitalisation risk, and NO2 with 3% increased infection risk and 11% increased hospitalisation risk. Bi-pollutant models suggested that effects were mainly driven by PM. Associations for PM were confirmed when stratifying by urbanization degree, epidemic wave and testing policy. All emission sources of PM, except industry, showed adverse effects on both outcomes. Livestock showed the most detrimental effects per unit exposure, whereas road traffic affected severity (hospitalisation) more than infection risk. This study shows that long-term exposure to air pollution increases both SARS-CoV-2 infection and COVID-19 hospitalisation risks, even after controlling for SARS-CoV-2 exposure levels, and that PM may have differential effects on these COVID-19 outcomes depending on the emission source.

Reviving Decades-Old Wisdom: Longitudinal Analysis of Renin-Angiotensin System Inhibitors and Its Effects on Acute Ischemic Stroke to Improve Outcomes

BACKGROUND

While renin-angiotensin system (RAS) inhibitors have a longstanding history in blood pressure control, their suitability as first-line in-patient treatment may be limited due to prolonged half-life and kidney failure concerns.

METHODS

Using a cohort design, we assessed the impact of RAS inhibitors, either alone or in combination with beta-blockers, on mortality, while exploring interactions, including those related to end-stage renal disease and serum creatinine levels. Eligible subjects were Acute Ischemic Stroke (AIS) patients aged 18 or older with specific subtypes who received in-patient antihypertensive treatment. The primary outcome was mortality rates. Statistical analyses included cross-sectional and longitudinal approaches, employing generalized linear models, G-computation, and discrete-time survival analysis over a 20-day follow-up period.

RESULTS

In our study of 3,058 AIS patients, those using RAS inhibitors had significantly lower in-hospital mortality (2.2%) compared to non-users (12.1%), resulting in a relative risk (RR) of 0.18 (95% CI: 0.12-0.26). Further analysis using G-computation revealed a marked reduction in mortality risk associated with RAS inhibitors (0.0281 vs. 0.0913, risk difference [RD] of 6.31% or 0.0631, 95% CI: 0.046-0.079). Subgroup analysis demonstrated notable benefits, with individuals having creatinine levels below and above 1.3 mg/dl exhibiting statistically significant RD (RD -0.0510 vs. -0.0895), and a significant difference in paired comparison (-0.0385 or 3.85%, CI 0.023-0.054). Additionally, longitudinal analysis confirmed a consistent daily reduction of 0.93% in mortality risk associated with the intake of RAS inhibitors.

CONCLUSIONS

RAS inhibitors are associated with a significant reduction in in-hospital mortality in AIS patients, suggesting potential clinical benefits in improving patient outcomes.

Association between particulate air pollution and hypertensive disorders in pregnancy: A retrospective cohort study

BACKGROUND

Epidemiological findings regarding the association of particulate matter ≤2.5 μm (PM2.5) exposure with hypertensive disorders in pregnancy (HDP) are inconsistent; evidence for HDP risk related to PM2.5 components, mixture effects, and windows of susceptibility is limited. We aimed to investigate the relationships between HDP and exposure to PM2.5 during pregnancy.

METHODS AND FINDINGS

A large retrospective cohort study was conducted among mothers with singleton pregnancies in Kaiser Permanente Southern California from 2008 to 2017. HDP were defined by International Classification of Diseases-9/10 (ICD-9/10) diagnostic codes and were classified into 2 subcategories based on the severity of HDP: gestational hypertension (GH) and preeclampsia and eclampsia (PE-E). Monthly averages of PM2.5 total mass and its constituents (i.e., sulfate, nitrate, ammonium, organic matter, and black carbon) were estimated using outputs from a fine-resolution geoscience-derived model. Multilevel Cox proportional hazard models were used to fit single-pollutant models; quantile g-computation approach was applied to estimate the joint effect of PM2.5 constituents. The distributed lag model was applied to estimate the association between monthly PM2.5 exposure and HDP risk. This study included 386,361 participants (30.3 ± 6.1 years) with 4.8% (17,977/373,905) GH and 5.0% (19,381/386,361) PE-E cases, respectively. In single-pollutant models, we observed increased relative risks for PE-E associated with exposures to PM2.5 total mass [adjusted hazard ratio (HR) per interquartile range: 1.07, 95% confidence interval (CI) [1.04, 1.10] p < 0.001], black carbon [HR = 1.12 (95% CI [1.08, 1.16] p < 0.001)] and organic matter [HR = 1.06 (95% CI [1.03, 1.09] p < 0.001)], but not for GH. The population attributable fraction for PE-E corresponding to the standards of the US Environmental Protection Agency (9 μg/m3) was 6.37%. In multi-pollutant models, the PM2.5 mixture was associated with an increased relative risk of PE-E ([HR = 1.05 (95% CI [1.03, 1.07] p < 0.001)], simultaneous increase in PM2.5 constituents of interest by a quartile) and PM2.5 black carbon gave the greatest contribution of the overall mixture effects (71%) among all individual constituents. The susceptible window is the late first trimester and second trimester. Furthermore, the risks of PE-E associated with PM2.5 exposure were significantly higher among Hispanic and African American mothers and mothers who live in low- to middle-income neighborhoods (p < 0.05 for Cochran's Q test). Study limitations include potential exposure misclassification solely based on residential outdoor air pollution, misclassification of disease status defined by ICD codes, the date of diagnosis not reflecting the actual time of onset, and lack of information on potential covariates and unmeasured factors for HDP.

CONCLUSIONS

Our findings add to the literature on associations between air pollution exposure and HDP. To our knowledge, this is the first study reporting that specific air pollution components, mixture effects, and susceptible windows of PM2.5 may affect GH and PE-E differently.

Development and Validation of a Risk Score Predicting Death Without Transplant in Adult Heart Transplant Candidates

Importance

The US heart allocation system prioritizes medically urgent candidates with a high risk of dying without transplant. The current therapy-based 6-status system is susceptible to manipulation and has limited rank ordering ability.

Objective

To develop and validate a candidate risk score that incorporates current clinical, laboratory, and hemodynamic data.

Design, Setting, and Participants

A registry-based observational study of adult heart transplant candidates (aged ≥18 years) from the US heart allocation system listed between January 1, 2019, and December 31, 2022, split by center into training (70%) and test (30%) datasets. Adult candidates were listed between January 1, 2019, and December 31, 2022.

Main Outcomes and Measures

A US candidate risk score (US-CRS) model was developed by adding a predefined set of predictors to the current French Candidate Risk Score (French-CRS) model. Sensitivity analyses were performed, which included intra-aortic balloon pumps (IABP) and percutaneous ventricular assist devices (VAD) in the definition of short-term mechanical circulatory support (MCS) for the US-CRS. Performance of the US-CRS model, French-CRS model, and 6-status model in the test dataset was evaluated by time-dependent area under the receiver operating characteristic curve (AUC) for death without transplant within 6 weeks and overall survival concordance (c-index) with integrated AUC.

Results

A total of 16 905 adult heart transplant candidates were listed (mean [SD] age, 53 [13] years; 73% male; 58% White); 796 patients (4.7%) died without a transplant. The final US-CRS contained time-varying short-term MCS (ventricular assist-extracorporeal membrane oxygenation or temporary surgical VAD), the log of bilirubin, estimated glomerular filtration rate, the log of B-type natriuretic peptide, albumin, sodium, and durable left ventricular assist device. In the test dataset, the AUC for death within 6 weeks of listing for the US-CRS model was 0.79 (95% CI, 0.75-0.83), for the French-CRS model was 0.72 (95% CI, 0.67-0.76), and 6-status model was 0.68 (95% CI, 0.62-0.73). Overall c-index for the US-CRS model was 0.76 (95% CI, 0.73-0.80), for the French-CRS model was 0.69 (95% CI, 0.65-0.73), and 6-status model was 0.67 (95% CI, 0.63-0.71). Classifying IABP and percutaneous VAD as short-term MCS reduced the effect size by 54%.

Conclusions and Relevance

In this registry-based study of US heart transplant candidates, a continuous multivariable allocation score outperformed the 6-status system in rank ordering heart transplant candidates by medical urgency and may be useful for the medical urgency component of heart allocation.

Predicting recurrent interventions after radiocephalic arteriovenous fistula creation with machine learning and the PREDICT-AVF web app

OBJECTIVE

Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines discourage ongoing access salvage attempts after two interventions prior to successful use or more than three interventions per year overall. The goal was to develop a tool for prediction of radiocephalic arteriovenous fistula (AVF) intervention requirements to help guide shared decision-making about access appropriateness.

METHODS

Prospective cohort study of 914 adult patients in the United States and Canada undergoing radiocephalic AVF creation at one of the 39 centers participating in the PATENCY-1 or -2 trials. Clinical data, including demographics, comorbidities, access history, anatomic features, and post-operative ultrasound measurements at 4-6 and 12 weeks were used to predict recurrent interventions required at 1 year postoperatively. Cox proportional hazards, random survival forest, pooled logistic, and elastic net recurrent event survival prediction models were built using a combination of baseline characteristics and post-operative ultrasound measurements. A web application was created, which generates patient-specific predictions contextualized with the KDOQI guidelines.

RESULTS

Patients underwent an estimated 1.04 (95% CI 0.94-1.13) interventions in the first year. Mean (SD) age was 57 (13) years; 22% were female. Radiocephalic AVFs were created at the snuffbox (2%), wrist (74%), or proximal forearm (24%). Using baseline characteristics, the random survival forest model performed best, with an area under the receiver operating characteristic curve (AUROC) of 0.75 (95% CI 0.67-0.82) at 1 year. The addition of ultrasound information to baseline characteristics did not substantially improve performance; however, Cox models using either 4-6- or 12-week post-operative ultrasound information alone had the best discrimination performance, with AUROCs of 0.77 (0.70-0.85) and 0.76 (0.70-0.83) at 1 year. The interactive web application is deployed at https://predict-avf.com.

CONCLUSIONS

The PREDICT-AVF web application can guide patient counseling and guideline-concordant shared decision-making as part of a patient-centered end-stage kidney disease life plan.

Applying the concept of niche breadth to understand urban tree mortality in the UK

Accelerated climate change has raised concerns about heightened vulnerability of urban trees, spurring the need to reevaluate their suitability. The urgency has also driven the widespread application of climatic niche-based models. In particular, the concept of niche breadth (NB), the range of environmental conditions that species can tolerate, is commonly estimated based on species occurrence data over the selected geographic range to predict species response to changing conditions. However, in urban environments where many species are cultivated out of the NB of their natural distributions, additional empirical evidence beyond presence and absence is needed not only to test the true tolerance limits but also to evaluate species' adaptive capacity to future climate. In this research, mortality trends of Acer and Quercus species spanning a 21-year period (2000-2021) from tree inventories of three major UK botanic gardens - the Royal Botanic Gardens, Kew (KEW), Westonbirt, the National Arboretum (WESB), and the Royal Botanic Garden Edinburgh (RBGE) - were analyzed in relation to their NB under long-term drought stress. As a result, Acer species were more responsive to drought and heat stress. For Acer, positioning below the lower limits of the precipitation of warmest quarter led to an increase in the probability of annual mortality by 1.2 and 1.3 % at KEW and RBGE respectively. In addition, the mean cumulative mortality rate increased corresponding to an increase in the number of niche positions below the lower limits of the selected bioclimatic variables. On the other hand, Quercus species in general exhibited comparable resilience regardless of their niche positions. Moreover, Mediterranean oaks were most tolerant, with cumulative mortality rates that were lower than those of native oaks in the UK. These findings further highlight the importance of incorporating ecological performance and recognizing species-specific adaptive strategies in climatic niche modeling.

Multitask Learning with Convolutional Neural Networks and Vision Transformers Can Improve Outcome Prediction for Head and Neck Cancer Patients

Neural-network-based outcome predictions may enable further treatment personalization of patients with head and neck cancer. The development of neural networks can prove challenging when a limited number of cases is available. Therefore, we investigated whether multitask learning strategies, implemented through the simultaneous optimization of two distinct outcome objectives (multi-outcome) and combined with a tumor segmentation task, can lead to improved performance of convolutional neural networks (CNNs) and vision transformers (ViTs). Model training was conducted on two distinct multicenter datasets for the endpoints loco-regional control (LRC) and progression-free survival (PFS), respectively. The first dataset consisted of pre-treatment computed tomography (CT) imaging for 290 patients and the second dataset contained combined positron emission tomography (PET)/CT data of 224 patients. Discriminative performance was assessed by the concordance index (C-index). Risk stratification was evaluated using log-rank tests. Across both datasets, CNN and ViT model ensembles achieved similar results. Multitask approaches showed favorable performance in most investigations. Multi-outcome CNN models trained with segmentation loss were identified as the optimal strategy across cohorts. On the PET/CT dataset, an ensemble of multi-outcome CNNs trained with segmentation loss achieved the best discrimination (C-index: 0.29, 95% confidence interval (CI): 0.22-0.36) and successfully stratified patients into groups with low and high risk of disease progression (p=0.003). On the CT dataset, ensembles of multi-outcome CNNs and of single-outcome ViTs trained with segmentation loss performed best (C-index: 0.26 and 0.26, CI: 0.18-0.34 and 0.18-0.35, respectively), both with significant risk stratification for LRC in independent validation (p=0.002 and p=0.011). Further validation of the developed multitask-learning models is planned based on a prospective validation study, which has recently completed recruitment.

Association of Antepartum and Postpartum Air Pollution Exposure With Postpartum Depression in Southern California

Importance

Women are especially vulnerable to mental health matters post partum because of biological, emotional, and social changes during this period. However, epidemiologic evidence of an association between air pollution exposure and postpartum depression (PPD) is limited.

Objective

To examine the associations between antepartum and postpartum maternal air pollution exposure and PPD.

Design, Setting, and Participants

This retrospective cohort study used data from Kaiser Permanente Southern California (KPSC) electronic health records and included women who had singleton live births at KPSC facilities between January 1, 2008, and December 31, 2016. Data were analyzed between January 1 and May 10, 2023.

Exposures

Ambient air pollution exposures were assessed based on maternal residential addresses using monthly averages of particulate matter less than or equal to 2.5 μm (PM2.5), particulate matter less than or equal to 10 μm (PM10), nitrogen dioxide (NO2), and ozone (O3) from spatial interpolation of monitoring station measurements. Constituents of PM2.5 (sulfate, nitrate, ammonium, organic matter, and black carbon) were obtained from fine-resolution geoscience-derived models based on satellite, ground-based monitor, and chemical transport modeling data.

Main Outcomes and Measures

Participants with an Edinburgh Postnatal Depression Scale score of 10 or higher during the 6 months after giving birth were referred to a clinical interview for further assessment and diagnosis. Ascertainment of PPD was defined using a combination of diagnostic codes and prescription medications.

Results

The study included 340 679 participants (mean [SD] age, 30.05 [5.81] years), with 25 674 having PPD (7.54%). Increased risks for PPD were observed to be associated with per-IQR increases in antepartum and postpartum exposures to O3 (adjusted odds ratio [AOR], 1.09; 95% CI, 1.06-1.12), PM10 (AOR, 1.02; 95% CI, 1.00-1.04), and PM2.5 (AOR, 1.02; 95% CI, 1. 00-1.03) but not with NO2; PPD risks were mainly associated with PM2.5 organic matter and black carbon. Overall, a higher risk of PPD was associated with O3 during the entire pregnancy and postpartum periods and with PM exposure during the late pregnancy and postpartum periods.

Conclusions and Relevance

The study findings suggest that long-term exposure to antepartum and postpartum air pollution was associated with higher PPD risks. Identifying the modifiable environmental risk factors and developing interventions are important public health issues to improve maternal mental health and alleviate the disease burden of PPD.

Analysis of Heat Exposure During Pregnancy and Severe Maternal Morbidity

Importance

The rate of severe maternal morbidity (SMM) is continuously increasing in the US. Evidence regarding the associations of climate-related exposure, such as environmental heat, with SMM is lacking.

Objective

To examine associations between long- and short-term maternal heat exposure and SMM.

Design, Setting, and Participants

This retrospective population-based epidemiological cohort study took place at a large integrated health care organization, Kaiser Permanente Southern California, between January 1, 2008, and December 31, 2018. Data were analyzed from February to April 2023. Singleton pregnancies with data on SMM diagnosis status were included.

Exposures

Moderate, high, and extreme heat days, defined as daily maximum temperatures exceeding the 75th, 90th, and 95th percentiles of the time series data from May through September 2007 to 2018 in Southern California, respectively. Long-term exposures were measured by the proportions of different heat days during pregnancy and by trimester. Short-term exposures were represented by binary variables of heatwaves with 9 different definitions (combining percentile thresholds with 3 durations; ie, ≥2, ≥3, and ≥4 consecutive days) during the last gestational week.

Main Outcomes and Measures

The primary outcome was SMM during delivery hospitalization, measured by 20 subconditions excluding blood transfusion. Discrete-time logistic regression was used to estimate associations with long- and short-term heat exposure. Effect modification by maternal characteristics and green space exposure was examined using interaction terms.

Results

There were 3446 SMM cases (0.9%) among 403 602 pregnancies (mean [SD] age, 30.3 [5.7] years). Significant associations were observed with long-term heat exposure during pregnancy and during the third trimester. High exposure (≥80th percentile of the proportions) to extreme heat days during pregnancy and during the third trimester were associated with a 27% (95% CI, 17%-37%; P < .001) and 28% (95% CI, 17%-41%; P < .001) increase in risk of SMM, respectively. Elevated SMM risks were significantly associated with short-term heatwave exposure under all heatwave definitions. The magnitude of associations generally increased from the least severe (HWD1: daily maximum temperature >75th percentile lasting for ≥2 days; odds ratio [OR], 1.32; 95% CI, 1.17-1.48; P < .001) to the most severe heatwave exposure (HWD9: daily maximum temperature >95th percentile lasting for ≥4 days; OR, 2.39; 95% CI, 1.62-3.54; P < .001). Greater associations were observed among mothers with lower educational attainment (OR for high exposure to extreme heat days during pregnancy, 1.43; 95% CI, 1.26-1.63; P < .001) or whose pregnancies started in the cold season (November through April; OR, 1.37; 95% CI, 1.24-1.53; P < .001).

Conclusions and Relevance

In this retrospective cohort study, long- and short-term heat exposure during pregnancy was associated with higher risk of SMM. These results might have important implications for SMM prevention, particularly in a changing climate.

Predicting incident dementia in cerebral small vessel disease: comparison of machine learning and traditional statistical models

Background

Cerebral small vessel disease (SVD) contributes to 45% of dementia cases worldwide, yet we lack a reliable model for predicting dementia in SVD. Past attempts largely relied on traditional statistical approaches. Here, we investigated whether machine learning (ML) methods improved prediction of incident dementia in SVD from baseline SVD-related features over traditional statistical methods.

Methods

We included three cohorts with varying SVD severity (RUN DMC, n = 503; SCANS, n = 121; HARMONISATION, n = 265). Baseline demographics, vascular risk factors, cognitive scores, and magnetic resonance imaging (MRI) features of SVD were used for prediction. We conducted both survival analysis and classification analysis predicting 3-year dementia risk. For each analysis, several ML methods were evaluated against standard Cox or logistic regression. Finally, we compared the feature importance ranked by different models.

Results

We included 789 participants without missing data in the survival analysis, amongst whom 108 (13.7%) developed dementia during a median follow-up of 5.4 years. Excluding those censored before three years, we included 750 participants in the classification analysis, amongst whom 48 (6.4%) developed dementia by year 3. Comparing statistical and ML models, only regularised Cox/logistic regression outperformed their statistical counterparts overall, but not significantly so in survival analysis. Baseline cognition was highly predictive, and global cognition was the most important feature.

Conclusions

When using baseline SVD-related features to predict dementia in SVD, the ML survival or classification models we evaluated brought little improvement over traditional statistical approaches. The benefits of ML should be evaluated with caution, especially given limited sample size and features.

Findings From the World Mental Health Surveys of Civil Violence Exposure and Its Association With Subsequent Onset and Persistence of Mental Disorders

Importance

Understanding the association of civil violence with mental disorders is important for developing effective postconflict recovery policies.

Objective

To estimate the association between exposure to civil violence and the subsequent onset and persistence of common mental disorders (in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition [DSM-IV]) in representative surveys of civilians from countries that have experienced civil violence since World War II.

Design, Setting, and Participants

This study used data from cross-sectional World Health Organization World Mental Health (WMH) surveys administered to households between February 5, 2001, and January 5, 2022, in 7 countries that experienced periods of civil violence after World War II (Argentina, Colombia, Lebanon, Nigeria, Northern Ireland, Peru, and South Africa). Data from respondents in other WMH surveys who immigrated from countries with civil violence in Africa and Latin America were also included. Representative samples comprised adults (aged ≥18 years) from eligible countries. Data analysis was performed from February 10 to 13, 2023.

Exposures

Exposure was defined as a self-report of having been a civilian in a war zone or region of terror. Related stressors (being displaced, witnessing atrocities, or being a combatant) were also assessed. Exposures occurred a median of 21 (IQR, 12-30) years before the interview.

Main Outcomes and Measures

The main outcome was the retrospectively reported lifetime prevalence and 12-month persistence (estimated by calculating 12-month prevalence among lifetime cases) of DSM-IV anxiety, mood, and externalizing (alcohol use, illicit drug use, or intermittent explosive) disorders.

Results

This study included 18 212 respondents from 7 countries. Of these individuals, 2096 reported that they were exposed to civil violence (56.5% were men; median age, 40 [IQR, 30-52] years) and 16 116 were not exposed (45.2% were men; median age, 35 [IQR, 26-48] years). Respondents who reported being exposed to civil violence had a significantly elevated onset risk of anxiety (risk ratio [RR], 1.8 [95% CI, 1.5-2.1]), mood (RR, 1.5 [95% CI, 1.3-1.7]), and externalizing (RR, 1.6 [95% CI, 1.3-1.9]) disorders. Combatants additionally had a significantly elevated onset risk of anxiety disorders (RR, 2.0 [95% CI, 1.3-3.1]) and refugees had an increased onset risk of mood (RR, 1.5 [95% CI, 1.1-2.0]) and externalizing (RR, 1.6 [95% CI, 1.0-2.4]) disorders. Elevated disorder onset risks persisted for more than 2 decades if conflicts persisted but not after either termination of hostilities or emigration. Persistence (ie, 12-month prevalence among respondents with lifetime prevalence of the disorder), in comparison, was generally not associated with exposure.

Conclusions

In this survey study of exposure to civil violence, exposure was associated with an elevated risk of mental disorders among civilians for many years after initial exposure. These findings suggest that policy makers should recognize these associations when projecting future mental disorder treatment needs in countries experiencing civil violence and among affected migrants.

Survival analysis 101: an easy start guide to analysing time-to-event data

Survival analysis, also called time-to-event analysis, is a common approach to handling event data in cardiovascular nursing and health-related research. Survival analysis is used to describe, explain, and/or predict the occurrence and timing of events. There is a specific language used and methods designed to handle the unique nature of event data. In this methods paper, we provide an 'easy start guide' to using survival analysis by (i) providing a step-by-step guide and (ii) applying the steps with example data. Specifically, we analyse cardiovascular event data over 6 months in a sample of patients with heart failure.

Investigating the performance of machine learning algorithms in predicting the survival of COVID-19 patients: A cross section study of Iran

Background and Aims

Like early diagnosis, predicting the survival of patients with Coronavirus Disease 2019 (COVID-19) is of great importance. Survival prediction models help doctors be more cautious to treat the patients who are at high risk of dying because of medical conditions. This study aims to predict the survival of hospitalized patients with COVID-19 by comparing the accuracy of machine learning (ML) models.

Methods

It is a cross-sectional study which was performed in 2022 in Fasa city in Iran country. The research data set was extracted from the period February 18, 2020 to February 10, 2021, and contains 2442 hospitalized patients' records with 84 features. A comparison was made between the efficiency of five ML algorithms to predict survival, includes Naive Bayes (NB), K-nearest neighbors (KNN), random forest (RF), decision tree (DT), and multilayer perceptron (MLP). Modeling steps were done with Python language in the Anaconda Navigator 3 environment.

Results

Our findings show that NB algorithm had better performance than others with accuracy, precision, recall, F-score, and area under receiver operating characteristic curve of 97%, 96%, 96%, 96%, and 97%, respectively. Based on the analysis of factors affecting survival, heart disease, pulmonary diseases and blood related disease were the most important disease related to death.

Conclusion

The development of software systems based on NB will be effective to predict the survival of COVID-19 patients.

Application of 18 F-fluorodeoxyglucose PET/CT radiomic features and machine learning to predict early recurrence of non-small cell lung cancer after curative-intent therapy

OBJECTIVE

To predict the recurrence of non-small cell lung cancer (NSCLC) within 2 years after curative-intent treatment using a machine-learning approach with PET/CT-based radiomics.

PATIENTS AND METHODS

A total of 77 NSCLC patients who underwent pretreatment 18 F-fluorodeoxyglucose PET/CT were retrospectively analyzed. Five clinical features (age, sex, tumor stage, tumor histology, and smoking status) and 48 radiomic features extracted from primary tumors on PET were used for binary classifications. These were ranked, and a subset of useful features was selected based on Gini coefficient scores in terms of associations with relapsed status. Areas under the receiver operating characteristics curves (AUC) were yielded by six machine-learning algorithms (support vector machine, random forest, neural network, naive Bayes, logistic regression, and gradient boosting). Model performances were compared and validated via random sampling.

RESULTS

A PET/CT-based radiomic model was developed and validated for predicting the recurrence of NSCLC during the first 2 years after curation. The most important features were SD and variance of standardized uptake value, followed by low-intensity short-zone emphasis and high-intensity zone emphasis. The naive Bayes model with the 15 best-ranked features displayed the best performance (AUC: 0.816). Prediction models using the five best PET-derived features outperformed those using five clinical variables.

CONCLUSION

The machine learning model using PET-derived radiomic features showed good performance for predicting the recurrence of NSCLC during the first 2 years after a curative intent therapy. PET/CT-based radiomic features may help clinicians improve the risk stratification of relapsed NSCLC.