Creating High-Quality Synthetic Health Data: Framework for Model Development and Validation

BACKGROUND

Electronic health records are a valuable source of patient information that must be properly deidentified before being shared with researchers. This process requires expertise and time. In addition, synthetic data have considerably reduced the restrictions on the use and sharing of real data, allowing researchers to access it more rapidly with far fewer privacy constraints. Therefore, there has been a growing interest in establishing a method to generate synthetic data that protects patients' privacy while properly reflecting the data.

OBJECTIVE

This study aims to develop and validate a model that generates valuable synthetic longitudinal health data while protecting the privacy of the patients whose data are collected.

METHODS

We investigated the best model for generating synthetic health data, with a focus on longitudinal observations. We developed a generative model that relies on the generalized canonical polyadic (GCP) tensor decomposition. This model also involves sampling from a latent factor matrix of GCP decomposition, which contains patient factors, using sequential decision trees, copula, and Hamiltonian Monte Carlo methods. We applied the proposed model to samples from the MIMIC-III (version 1.4) data set. Numerous analyses and experiments were conducted with different data structures and scenarios. We assessed the similarity between our synthetic data and the real data by conducting utility assessments. These assessments evaluate the structure and general patterns present in the data, such as dependency structure, descriptive statistics, and marginal distributions. Regarding privacy disclosure, our model preserves privacy by preventing the direct sharing of patient information and eliminating the one-to-one link between the observed and model tensor records. This was achieved by simulating and modeling a latent factor matrix of GCP decomposition associated with patients.

RESULTS

The findings show that our model is a promising method for generating synthetic longitudinal health data that is similar enough to real data. It can preserve the utility and privacy of the original data while also handling various data structures and scenarios. In certain experiments, all simulation methods used in the model produced the same high level of performance. Our model is also capable of addressing the challenge of sampling patients from electronic health records. This means that we can simulate a variety of patients in the synthetic data set, which may differ in number from the patients in the original data.

CONCLUSIONS

We have presented a generative model for producing synthetic longitudinal health data. The model is formulated by applying the GCP tensor decomposition. We have provided 3 approaches for the synthesis and simulation of a latent factor matrix following the process of factorization. In brief, we have reduced the challenge of synthesizing massive longitudinal health data to synthesizing a nonlongitudinal and significantly smaller data set.

Novel Approach for Detecting Respiratory Syncytial Virus in Pediatric Patients Using Machine Learning Models Based on Patient-Reported Symptoms: Model Development and Validation Study

BACKGROUND

Respiratory syncytial virus (RSV) affects children, causing serious infections, particularly in high-risk groups. Given the seasonality of RSV and the importance of rapid isolation of infected individuals, there is an urgent need for more efficient diagnostic methods to expedite this process.

OBJECTIVE

This study aimed to investigate the performance of a machine learning model that leverages the temporal diversity of symptom onset for detecting RSV infections and elucidate its discriminatory ability.

METHODS

The study was conducted in pediatric and emergency outpatient settings in Japan. We developed a detection model that remotely confirms RSV infection based on patient-reported symptom information obtained using a structured electronic template incorporating the differential points of skilled pediatricians. An extreme gradient boosting-based machine learning model was developed using the data of 4174 patients aged ≤24 months who underwent RSV rapid antigen testing. These patients visited either the pediatric or emergency department of Yokohama City Municipal Hospital between January 1, 2009, and December 31, 2015. The primary outcome was the diagnostic accuracy of the machine learning model for RSV infection, as determined by rapid antigen testing, measured using the area under the receiver operating characteristic curve. The clinical efficacy was evaluated by calculating the discriminative performance based on the number of days elapsed since the onset of the first symptom and exclusion rates based on thresholds of reasonable sensitivity and specificity.

RESULTS

Our model demonstrated an area under the receiver operating characteristic curve of 0.811 (95% CI 0.784-0.833) with good calibration and 0.746 (95% CI 0.694-0.794) for patients within 3 days of onset. It accurately captured the temporal evolution of symptoms; based on adjusted thresholds equivalent to those of a rapid antigen test, our model predicted that 6.9% (95% CI 5.4%-8.5%) of patients in the entire cohort would be positive and 68.7% (95% CI 65.4%-71.9%) would be negative. Our model could eliminate the need for additional testing in approximately three-quarters of all patients.

CONCLUSIONS

Our model may facilitate the immediate detection of RSV infection in outpatient settings and, potentially, in home environments. This approach could streamline the diagnostic process, reduce discomfort caused by invasive tests in children, and allow rapid implementation of appropriate treatments and isolation at home. The findings underscore the potential of machine learning in augmenting clinical decision-making in the early detection of RSV infection.

Development and validation of machine learning models to predict frailty risk for elderly

AIMS

Early identification and intervention of the frailty of the elderly will help lighten the burden of social medical care and improve the quality of life of the elderly. Therefore, we used machine learning (ML) algorithm to develop models to predict frailty risk in the elderly.

DESIGN

A prospective cohort study.

METHODS

We collected data on 6997 elderly people from Chinese Longitudinal Healthy Longevity Study wave 6-7 surveys (2011-2012, 2014). After the baseline survey in 1998 (wave 1), the project conducted follow-up surveys (wave 2-8) in 2000-2018. The osteoporotic fractures index was used to assess frailty. Four ML algorithms (random forest [RF], support vector machine, XGBoost and logistic regression [LR]) were used to develop models to identify the risk factors of frailty and predict the risk of frailty. Different ML models were used for the prediction of frailty risk in the elderly and frailty risk was trained on a cohort of 4385 elderly people with frailty (split into a training cohort [75%] and internal validation cohort [25%]). The best-performing model for each study outcome was tested in an external validation cohort of 6997 elderly people with frailty pooled from the surveys (wave 6-7). Model performance was assessed by receiver operating curve and F2-score.

RESULTS

Among the four ML models, the F2-score values were similar (0.91 vs. 0.91 vs. 0.88 vs. 0.90), and the area under the curve (AUC) values of RF model was the highest (0.75), followed by LR model (0.74). In the final two models, the AUC values of RF and LR model were similar (0.77 vs. 0.76) and their accuracy was identical (87.4% vs. 87.4%).

CONCLUSION

Our study developed a preliminary prediction model based on two different ML approaches to help predict frailty risk in the elderly.

IMPACT

The presented models from this study can be used to inform healthcare providers to predict the frailty probability among older adults and maybe help guide the development of effective frailty risk management interventions.

IMPLICATIONS FOR THE PROFESSION AND/OR PATIENT CARE

Detecting frailty at an early stage and implementing timely targeted interventions may help to improve the allocation of health care resources and to reduce frailty-related burden. Identifying risk factors for frailty could be beneficial to provide tailored and personalized care intervention for older adults to more accurately prevent or improve their frail conditions so as to improve their quality of life.

REPORTING METHOD

The study has adhered to STROBE guidelines.

PATIENT OR PUBLIC CONTRIBUTION

No patient or public contribution.

Development and validation of a clinical model to predict preconception risk of gestational diabetes mellitus in nulliparous women: A retrospective cohort study

OBJECTIVE

To develop and validate a model to predict the preconception risk of gestational diabetes mellitus (GDM) in nulliparous women.

METHODS

This was a retrospective cohort study. A total of 1565 women in early pregnancy who underwent preconception health examinations in the Women and Children's Hospital of Chongqing Medical University between January 2020 and June 2021 were invited to participate in a questionnaire survey. Logistic regression analysis was performed to determine the preconception risk factors for GDM. These factors were used to construct a model to predict GDM risk in nulliparous women. Then, the model was used to assess the preconception risk of GDM in 1060 nulliparous women.

RESULTS

Independent preconception risk factors for GDM included the following: age 35 years or greater, diastolic blood pressure 80 mm Hg or greater, fasting plasma glucose 5.1 mmol/L or greater, body mass index (BMI, calculated as weight in kilograms divided by the square of height in meters) 24 or greater, weight gain 10 kg or greater in the year before pregnancy, age of menarche 15 years or greater, three or more previous pregnancies, daily staple food intake 300 g or greater, fondness for sweets, and family history of diabetes. BMI less than 18.5, daily physical activity duration 1 h or greater, and high-intensity physical activity were protective factors. These factors were used to construct a model to predict GDM risk in nulliparous women, and the incidence of GDM significantly increased as the risk score increased. The area under the curve of the prediction model was 0.82 (95% confidence interval 0.80-0.85).

CONCLUSION

The preconception GDM risk prediction model demonstrated good predictive efficacy and can be used to identify populations at high risk of GDM before pregnancy, which provides the possibility for preconception intervention.

Guidelines to support newly qualified professional nurses for effective clinical practice

BACKGROUND

 Transitioning to a professional role is difficult for newly qualified professional nurses. Given the challenges that these nurses experience during the transition to practice, support is essential for them to become efficient, safe, confident, and competent in their professional roles.

OBJECTIVES

 The purpose of this study was to explore the transition experiences of newly qualified professional nurses to develop a preceptorship model.

METHOD

 This study employed a qualitative approach to purposively collect data. Concept analyses were conducted applying the steps suggested by Walker and Avant, and the related concepts were classified utilising the survey list of Dickoff, James and Wiedenbach's practice theory.

RESULTS

 A preceptorship model for the facilitation of guidance and support in the clinical area for newly qualified professional nurses was developed. The model consists of six components, namely, the clinical environment, the operational manager and preceptor, the newly qualified professional nurse, the preceptorship, the assessment of learning, and the outcome.

CONCLUSION

 The study revealed that newly qualified professional nurses face many transition challenges when entering clinical practice. They are thrown far in, experience a reality shock, and are not ready to start performing their professional role. The participants agreed that guidance and support are needed for their independent practice role.Contribution: The preceptorship model for newly qualified professional nurses would be necessary for the transition period within hospitals. This preceptorship model may be implemented by nursing education institutions as part of their curriculum to prepare pre-qualifying students for the professional role.

Comparing Predictive Performance of Time Invariant and Time Variant Clinical Prediction Models in Cardiac Surgery

Clinical prediction models are increasingly used across healthcare to support clinical decision making. Existing methods and models are time-invariant and thus ignore the changes in populations and healthcare practice that occur over time. We aimed to compare the performance of time-invariant with time-variant models in UK National Adult Cardiac Surgery Audit data from Manchester University NHS Foundation Trust between 2009 and 2019. Data from 2009-2011 were used for initial model fitting, and data from 2012-2019 for validation and updating. We fitted four models to the data: a time-invariant logistic regression model (not updated), a logistic model which was updated every year and validated it in each subsequent year, a logistic regression model where the intercept is a function of calendar time (not updated), and a continually updating Bayesian logistic model which was updated with each new observation and continuously validated. We report predictive performance over the complete validation cohort and for each year in the validation data. Over the complete validation data, the Bayesian model had the best predictive performance.

Unraveling neural pathways of political engagement: bridging neuromarketing and political science for understanding voter behavior and political leader perception

Introduction

Political neuromarketing is an emerging interdisciplinary field integrating marketing, neuroscience, and psychology to decipher voter behavior and political leader perception. This interdisciplinary field offers novel techniques to understand complex phenomena such as voter engagement, political leadership, and party branding.

Methods

This study aims to understand the neural activation patterns of voters when they are exposed to political leaders using functional near-infrared spectroscopy (fNIRS) and machine learning methods. We recruited participants and recorded their brain activity using fNIRS when they were exposed to images of different political leaders.

Results

This neuroimaging method (fNIRS) reveals brain regions central to brand perception, including the dorsolateral prefrontal cortex (dlPFC), the dorsomedial prefrontal cortex (dmPFC), and the ventromedial prefrontal cortex (vmPFC). Machine learning methods were used to predict the participants' perceptions of leaders based on their brain activity. The study has identified the brain regions that are involved in processing political stimuli and making judgments about political leaders. Within this study, the best-performing machine learning model, LightGBM, achieved a highest accuracy score of 0.78, underscoring its efficacy in predicting voters' perceptions of political leaders based on the brain activity of the former.

Discussion

The findings from this study provide new insights into the neural basis of political decision-making and the development of effective political marketing campaigns while bridging neuromarketing, political science, and machine learning, in turn enabling predictive insights into voter preferences and behavior.

Developing the theoretical model of Chinese physical education teachers' health communication competence: based on grounded theory

Background

Physical education teachers' health communication competence is a key factor in health promotion. Although health communication is a multidisciplinary field, medical practitioners are the primary focus of health communication research, whereas physical education teachers are marginalized. Therefore, this study proposes a theoretical model of health communication competence for physical education teachers.

Methods

This qualitative research utilized interviews as the primary data collection method. Purposeful sampling was employed to select participants, including university teachers, primary and secondary school teachers, and health education professionals from diverse regions of China. A total of 31 participants were interviewed through two focus groups (N = 15) and individual semi-structured interviews (N = 16). Grounded theory was used to analyze and code the collected interview materials.

Results

The health communication competence of physical education teachers consisted of three main categories, 10 subcategories, 30 concepts, and 240 statement labels. The three main categories were as follows: (i) foundations of health communication knowledge and skills (this category encompassed three subcategories, namely sport and health knowledge reserve, health beliefs, and health behaviors); (ii) health communication perception competence (this category included two subcategories, namely health risk and crisis perception competence and communication audience perception competence); and (iii) practical competence of health communication (this category consisted of five subcategories, namely language expression competence, organizational and design competence, utilization of new media tools competence, communication content selection and processing competence, and professional skills).

Conclusion

The theoretical model of health communication competence in this study provides a foundation for the involvement of physical education teachers in health communication work. It can serve as a reference for the development of both pre-service health education courses and in-service training systems for physical education teachers. Future research can expand the sample size and geographic coverage to further validate the applicability of the findings. Additionally, investigating the factors influencing the formation of the identified competencies is recommended.

The Development and Implementation of a Model to Facilitate Self-Awareness of Professionalism for Enrolled Nurses

Within the South African context, the nursing profession comprises different nursing cadres. The enrolled nurse is considered a sub-category of nursing and therefore does not carry the title of 'professional' as in a professional nurse. The purpose of the study was to develop, describe, implement, and evaluate a model for the facilitation of self-awareness for the professionalism of enrolled nurses at a specific nursing agency in Gauteng. A theory generating, qualitative, exploratory, descriptive and contextual design was used and was conducted following Chinn and Kramer's four stages of model development. The model can benefit nursing education because it relates to an essential aspect of growth and maturity in one's career. Ultimately, the facilitation of self-awareness for professionalism can advance one's career, or the lack of self-awareness may impede one's career. Developing, describing, implementing and evaluating this model to facilitate self-awareness for the professionalism of enrolled nurses at a specific nursing agency in Gauteng provides an original contribution to the theory in nursing professionalism and ethos. This model can be utilised as a tool to facilitate self-awareness for the professionalism of enrolled nurses at a nursing agency.

Noninvasive Staging of Lymph Node Status in Breast Cancer Using Machine Learning: External Validation and Further Model Development

BACKGROUND

Most patients diagnosed with breast cancer present with a node-negative disease. Sentinel lymph node biopsy (SLNB) is routinely used for axillary staging, leaving patients with healthy axillary lymph nodes without therapeutic effects but at risk of morbidities from the intervention. Numerous studies have developed nodal status prediction models for noninvasive axillary staging using postoperative data or imaging features that are not part of the diagnostic workup. Lymphovascular invasion (LVI) is a top-ranked predictor of nodal metastasis; however, its preoperative assessment is challenging.

OBJECTIVE

This paper aimed to externally validate a multilayer perceptron (MLP) model for noninvasive lymph node staging (NILS) in a large population-based cohort (n=18,633) and develop a new MLP in the same cohort. Data were extracted from the Swedish National Quality Register for Breast Cancer (NKBC, 2014-2017), comprising only routinely and preoperatively available documented clinicopathological variables. A secondary aim was to develop and validate an LVI MLP for imputation of missing LVI status to increase the preoperative feasibility of the original NILS model.

METHODS

Three nonoverlapping cohorts were used for model development and validation. A total of 4 MLPs for nodal status and 1 LVI MLP were developed using 11 to 12 routinely available predictors. Three nodal status models were used to account for the different availabilities of LVI status in the cohorts and external validation in NKBC. The fourth nodal status model was developed for 80% (14,906/18,663) of NKBC cases and validated in the remaining 20% (3727/18,663). Three alternatives for imputation of LVI status were compared. The discriminatory capacity was evaluated using the validation area under the receiver operating characteristics curve (AUC) in 3 of the nodal status models. The clinical feasibility of the models was evaluated using calibration and decision curve analyses.

RESULTS

External validation of the original NILS model was performed in NKBC (AUC 0.699, 95% CI 0.690-0.708) with good calibration and the potential of sparing 16% of patients with node-negative disease from SLNB. The LVI model was externally validated (AUC 0.747, 95% CI 0.694-0.799) with good calibration but did not improve the discriminatory performance of the nodal status models. A new nodal status model was developed in NKBC without information on LVI (AUC 0.709, 95% CI: 0.688-0.729), with excellent calibration in the holdout internal validation cohort, resulting in the potential omission of 24% of patients from unnecessary SLNBs.

CONCLUSIONS

The NILS model was externally validated in NKBC, where the imputation of LVI status did not improve the model's discriminatory performance. A new nodal status model demonstrated the feasibility of using register data comprising only the variables available in the preoperative setting for NILS using machine learning. Future steps include ongoing preoperative validation of the NILS model and extending the model with, for example, mammography images.

Methodologic Issues Specific to Prediction Model Development and Evaluation

Developing and evaluating statistical prediction models is challenging, and many pitfalls can arise. This article identifies what the authors believe are some common methodologic concerns that may be encountered. We describe each problem and make suggestions regarding how to address them. The hope is that this article will result in higher-quality publications of statistical prediction models.

Immune-related gene-based model predicts the survival of colorectal carcinoma and reflected various biological statuses

Bakcground: Prognosis of colorectal cancer (CRC) varies due to complex genetic-microenviromental interactions, and multiple gene-based prognostic models have been highlighted. Material and Method: In this work, the immune-related genes' expression-based model was developed and the scores of each sample were calculated. The correlation between the model and clinical information, immune infiltration, drug response and biological pathways were analyzed. Results: The high-score samples have a significantly longer survival (overall survival and progression-free survival) period than those with a low score, which was validated across seven datasets containing 1,325 samples (GSE17536 (N = 115), GSE17537 (N = 55), GSE33113 (N = 90), GSE37892 (N = 130), GSE38832 (N = 74), GSE39582 (N = 481), and TCGA (N = 380)). The score is significantly associated with clinical indicators, including age and stage, and further associated with PD-1/PD-L1 gene expression. Furthermore, high-score samples have significantly higher APC and a lower MUC5B mutation rate. The high-score samples show more immune infiltration (including CD4+ and CD8+ T cells, M1/M2 macrophages, and NK cells). Enriched pathway analyses showed that cancer-related pathways, including immune-related pathways, were significantly activated in high-score samples and that some drugs have significantly lower IC50 values than those with low score. Conclusion: The model developed based on immune-related genes is robust and reflected various statuses of CRC and may be a potential clinical indicator.

A Natural Language Processing Model for COVID-19 Detection Based on Dutch General Practice Electronic Health Records by Using Bidirectional Encoder Representations From Transformers: Development and Validation Study

BACKGROUND

Natural language processing (NLP) models such as bidirectional encoder representations from transformers (BERT) hold promise in revolutionizing disease identification from electronic health records (EHRs) by potentially enhancing efficiency and accuracy. However, their practical application in practice settings demands a comprehensive and multidisciplinary approach to development and validation. The COVID-19 pandemic highlighted challenges in disease identification due to limited testing availability and challenges in handling unstructured data. In the Netherlands, where general practitioners (GPs) serve as the first point of contact for health care, EHRs generated by these primary care providers contain a wealth of potentially valuable information. Nonetheless, the unstructured nature of free-text entries in EHRs poses challenges in identifying trends, detecting disease outbreaks, or accurately pinpointing COVID-19 cases.

OBJECTIVE

This study aims to develop and validate a BERT model for detecting COVID-19 consultations in general practice EHRs in the Netherlands.

METHODS

The BERT model was initially pretrained on Dutch language data and fine-tuned using a comprehensive EHR data set comprising confirmed COVID-19 GP consultations and non-COVID-19-related consultations. The data set was partitioned into a training and development set, and the model's performance was evaluated on an independent test set that served as the primary measure of its effectiveness in COVID-19 detection. To validate the final model, its performance was assessed through 3 approaches. First, external validation was applied on an EHR data set from a different geographic region in the Netherlands. Second, validation was conducted using results of polymerase chain reaction (PCR) test data obtained from municipal health services. Lastly, correlation between predicted outcomes and COVID-19-related hospitalizations in the Netherlands was assessed, encompassing the period around the outbreak of the pandemic in the Netherlands, that is, the period before widespread testing.

RESULTS

The model development used 300,359 GP consultations. We developed a highly accurate model for COVID-19 consultations (accuracy 0.97, F1-score 0.90, precision 0.85, recall 0.85, specificity 0.99). External validations showed comparable high performance. Validation on PCR test data showed high recall but low precision and specificity. Validation using hospital data showed significant correlation between COVID-19 predictions of the model and COVID-19-related hospitalizations (F1-score 96.8; P<.001; R2=0.69). Most importantly, the model was able to predict COVID-19 cases weeks before the first confirmed case in the Netherlands.

CONCLUSIONS

The developed BERT model was able to accurately identify COVID-19 cases among GP consultations even preceding confirmed cases. The validated efficacy of our BERT model highlights the potential of NLP models to identify disease outbreaks early, exemplifying the power of multidisciplinary efforts in harnessing technology for disease identification. Moreover, the implications of this study extend beyond COVID-19 and offer a blueprint for the early recognition of various illnesses, revealing that such models could revolutionize disease surveillance.

Development and validation of a model for measuring alcohol consumption from transdermal alcohol content data among college students

BACKGROUND AND AIMS

Transdermal alcohol content (TAC) data collected by wearable alcohol monitors could potentially contribute to alcohol research, but raw data from the devices are challenging to interpret. We aimed to develop and validate a model using TAC data to detect alcohol drinking.

DESIGN

We used a model development and validation study design.

SETTING

Indiana, USA PARTICIPANTS: In March to April 2021, we enrolled 84 college students who reported drinking at least once a week (median age = 20 years, 73% white, 70% female). We observed participants' alcohol drinking behavior for 1 week.

MEASUREMENTS

Participants wore BACtrack Skyn monitors (TAC data), provided self-reported drinking start times in real time (smartphone app) and completed daily surveys about their prior day of drinking. We developed a model using signal filtering, peak detection algorithm, regression and hyperparameter optimization. The input was TAC and outputs were alcohol drinking frequency, start time and magnitude. We validated the model using daily surveys (internal validation) and data collected from college students in 2019 (external validation).

FINDINGS

Participants (N = 84) self-reported 213 drinking events. Monitors collected 10 915 hours of TAC. In internal validation, the model had a sensitivity of 70.9% (95% CI = 64.1%-77.0%) and a specificity of 73.9% (68.9%-78.5%) in detecting drinking events. The median absolute time difference between self-reported and model-detected drinking start times was 59 min. Mean absolute error (MAE) for the reported and detected number of drinks was 2.8 drinks. In an exploratory external validation among five participants, number of drinking events, sensitivity, specificity, median time difference and MAE were 15%, 67%, 100%, 45 minutes and 0.9 drinks, respectively. Our model's output was correlated with breath alcohol concentration data (Spearman's correlation [95% CI] = 0.88 [0.77, 0.94]).

CONCLUSION

This study, the largest of its kind to date, developed and validated a model for detecting alcohol drinking using transdermal alcohol content data collected with a new generation of alcohol monitors. The model and its source code are available as Supporting Information (https://osf.io/xngbk).

The bystander intervention for problematic alcohol use model (BIPAUM)

Objective: The study aimed to identify phases of bystander intervention (BI) for problematic alcohol use (PAU) among college students. Participants: Twenty focus groups and nine interviews were conducted. Methods: Transcripts were thematically analyzed. Results: The phases of the Bystander Intervention for Problematic Alcohol Use Model (BIPAUM) include: (1) plan in advance, (2) notice and interpret a sign, (3) decide (i.e., assume responsibility, assess support/feasibility to intervene, and identify intervention strategy), (4) intervene, and (5) assess outcomes. Assessing outcomes loops to influence future behavior and each phase is influenced by barriers and facilitators. Conclusions: These unique phases should be considered when designing and evaluating intervention programs for PAU to meet students' needs and better reduce PAU. Future research should empirically test the BIPAUM. The results of the current study demonstrate a promising opportunity for applying BI to PAU, with the goal of reducing risky drinking among college students.

Preclinical testing of antimicrobials for cystic fibrosis lung infections: current needs and future priorities

A workshop was held by the PIPE-CF strategic research centre to consider preclinical testing of antimicrobials for cystic fibrosis (CF). The workshop brought together groups of people from the CF community to discuss current challenges and identify priorities when developing CF therapeutics. This paper summarizes the key points from the workshop from the different sessions, including talks given by presenters on the day and round table discussions. Currently, it is felt that there is a large disconnect throughout the community, with communication between patients, clinicians and researchers being the main issue. This leads to little consideration being given to factors such as treatment regimes, routes of administration and side effects when developing new therapies, that could alter the day-to-day lifestyles of people living with CF. Translation of numerical data that are obtained in the laboratory to successful outcomes of clinical trials is also a key challenge facing researchers today. Laboratory assays in preclinical testing involve basing results on bacterial clearance and decrease in viable cells, when these are not factors that are considered when determining the success of a treatment in the clinic. However, there are several models currently in development that seek to tackle some of these issues, such as the organ-on-a-chip technology and adaptation of a hollow-fibre model, as well as the development of media that aim to mimic the niche environments of a CF respiratory tract. It is hoped that by summarizing these opinions and discussing current research, the communication gap between groups can begin to close.

Corrigendum: Estimation of subjective quality of life in schizophrenic patients using speech features

[This corrects the article DOI: 10.3389/fresc.2023.1121034.].

Agile Mentorship: A Longitudinal Exploratory Analysis

Effective mentorship relationships increase mentee academic success and satisfaction. However, existing mentorship models are limited by miscommunication, undefined roles, and mismatched goals. The agile mentorship process aims to address these limitations by leveraging insights from agile science and the existing evidence on effective mentorship models to support effective mentoring relationships in healthcare environments. To illustrate the agile mentorship process and the growth of a mentored clinician-scientist (H.L., first author), we describe the model and share qualitative findings generated from the independent analysis of 18 months of mentee reflections. In two iterative cycles, reflections (n = 56) were analyzed using exploratory content and relational analysis. Coauthors C.S. and B.T. employed inductive and deductive coding approaches to explore the data using an ontological lens. We discuss and share quotes representing the identified four main themes. Identification of shortcomings, adaptive perspective, managing relationships, and personal growth. In addition, personal growth had three subthemes: Awareness, continual reflection, and toolkit development. In summary, the reflections of one mentee within the agile mentorship process illustrated the growth process which occurred within an effective mentorship relationship. The agile mentorship process is a scalable and sustainable framework that is adaptable to various career development processes. Further evaluation is needed to understand the longitudinal impact of the model on mentee performance and satisfaction.

Developing a Model for Integrating of Tuberculosis, Human Immunodeficiency Virus and Primary Healthcare Services in Oliver Reginald (O.R) Tambo District, Eastern Cape, South Africa

Despite the policy, frameworks for integration exist; integration of TB and HIV services is far from ideal in many resource-limited countries, including South Africa. Few studies have examined the advantages and disadvantages of integrated TB and HIV care in public health facilities, and even fewer have proposed conceptual models for proven integration. This study aims to fill this vacuum by describing the development of a paradigm for integrating TB, HIV, and patient services in a single facility and highlights the importance of TB-HIV services for greater accessibility under one roof. Development of the proposed model occurred in several phases that included assessment of the existing integration model for TB-HIV and synthesis of quantitative and qualitative data from the study sites, which were selected public health facilities in rural and peri-urban areas in the Oliver Reginald (O.R.) Tambo District Municipality in the Eastern Cape, South Africa. Secondary data on clinical outcomes from 2009-2013 TB-HIV were obtained from various sources for the quantitative analysis of Part 1. Qualitative data included focus group discussions with patients and healthcare workers, which were analyzed thematically in Parts 2 and 3. The development of a potentially better model and the validation of this model shows that the district health system was strengthened by the guiding principles of the model, which placed a strong emphasis on inputs, processes, outcomes, and integration effects. The model is adaptable to different healthcare delivery systems but requires the support of patients, providers (professionals and institutions), payers, and policymakers to be successful.

Multiparametric Quantitative Imaging in Risk Prediction: Recommendations for Data Acquisition, Technical Performance Assessment, and Model Development and Validation

Combinations of multiple quantitative imaging biomarkers (QIBs) are often able to predict the likelihood of an event of interest such as death or disease recurrence more effectively than single imaging measurements can alone. The development of such multiparametric quantitative imaging and evaluation of its fitness of use differs from the analogous processes for individual QIBs in several key aspects. A computational procedure to combine the QIB values into a model output must be specified. The output must also be reproducible and be shown to have reasonably strong ability to predict the risk of an event of interest. Attention must be paid to statistical issues not often encountered in the single QIB scenario, including overfitting and bias in the estimates of model performance. This is the fourth in a five-part series on statistical methodology for assessing the technical performance of multiparametric quantitative imaging. Considerations for data acquisition are discussed and recommendations from the literature on methodology to construct and evaluate QIB-based models for risk prediction are summarized. The findings in the literature upon which these recommendations are based are demonstrated through simulation studies. The concepts in this manuscript are applied to a real-life example involving prediction of major adverse cardiac events using automated plaque analysis.

AAPM task group report 273: Recommendations on best practices for AI and machine learning for computer-aided diagnosis in medical imaging

Rapid advances in artificial intelligence (AI) and machine learning, and specifically in deep learning (DL) techniques, have enabled broad application of these methods in health care. The promise of the DL approach has spurred further interest in computer-aided diagnosis (CAD) development and applications using both "traditional" machine learning methods and newer DL-based methods. We use the term CAD-AI to refer to this expanded clinical decision support environment that uses traditional and DL-based AI methods. Numerous studies have been published to date on the development of machine learning tools for computer-aided, or AI-assisted, clinical tasks. However, most of these machine learning models are not ready for clinical deployment. It is of paramount importance to ensure that a clinical decision support tool undergoes proper training and rigorous validation of its generalizability and robustness before adoption for patient care in the clinic. To address these important issues, the American Association of Physicists in Medicine (AAPM) Computer-Aided Image Analysis Subcommittee (CADSC) is charged, in part, to develop recommendations on practices and standards for the development and performance assessment of computer-aided decision support systems. The committee has previously published two opinion papers on the evaluation of CAD systems and issues associated with user training and quality assurance of these systems in the clinic. With machine learning techniques continuing to evolve and CAD applications expanding to new stages of the patient care process, the current task group report considers the broader issues common to the development of most, if not all, CAD-AI applications and their translation from the bench to the clinic. The goal is to bring attention to the proper training and validation of machine learning algorithms that may improve their generalizability and reliability and accelerate the adoption of CAD-AI systems for clinical decision support.

The development and validation of screening tools for semi-automated surveillance of surgical site infection following various surgeries

Background

Surveillance of surgical site infections (SSIs) is essential for better prevention. We developed a screening method for SSIs in adults.

Methods

The training dataset included data from patients who underwent orthopedic surgeries (N = 1,090), colorectal surgeries (N = 817), and abdominal hysterectomies (N = 523) during 2015-2018. The gold standard for the validation of the screening tool was the presence of SSI as determined by a trained infection control practitioner, via manual full medical record review, using the US Center for Disease Control and Prevention criteria. Using multivariable regression models, we identified the correlates of SSI. Patients who had at least one of these correlates were classified as likely to having SSI and those who did not have any of the correlates were classified as unlikely to have SSI. We calculated the sensitivity and specificity of this tool compared to the gold standard and applied the tool to a validation dataset (N = 1,310, years 2019-2020).

Results

SSI was diagnosed by an infection control specialist in 8.2, 5.2, and 31.2% of the patients in the training dataset who underwent hysterectomies, orthopedic surgeries and colorectal surgeries, respectively, vs. 6.2, 6.6, and 25.5%, respectively, in the validation dataset. The correlates of SSI after abdominal hysterectomy were prolonged hospitalization, ordering wound or blood culture, emergency room visit and reoperation; in orthopedic surgery, emergency room visit, wound culture, reoperation, and documentation of SSI, and in colorectal surgeries prolonged hospitalization, readmission, and ordering wound or blood cultures. Area under the curve was >90%. The sensitivity and specificity (95% CI) of the screening tool were 98% (88-100) and 58% (53-62), for abdominal hysterectomy, 91% (81-96) and 82% (80-84) in orthopedic surgeries and 96% (90-98) and 62% (58-66) in colorectal surgeries. The corresponding values for the validation dataset were 89% (67-97) and 75% (69-80) in abdominal hysterectomy; 85% (72-93) and 83% (80-86) in orthopedic surgeries and 98% (93-99) and 59% (53-64) in colorectal surgeries. The number of files needed to be fully reviewed declined by 61-66.

Conclusion

The presented semi-automated simple screening tool for SSI surveillance had good sensitivity and specificity and it has great potential of reducing workload and improving SSI surveillance.

Digital health and capability maturity models-a critical thematic review and conceptual synthesis of the literature

OBJECTIVE

A literature review of capability maturity models (MMs) to inform the conceptualization, development, implementation, evaluation, and mainstreaming of MMs in digital health (DH).

METHODS

Electronic databases were searched using "digital health," "maturity models," and related terms based on the Digital Health Profile and Maturity Assessment Toolkit Maturity Model (DHPMAT-MM). Covidence was used to screen, identify, capture, and achieve consensus on data extracted by the authors. Descriptive statistics were generated. A thematic analysis and conceptual synthesis were conducted.

FINDINGS

Diverse domain-specific MMs and model development, implementation, and evaluation methods were found. The spread and pattern of different MMs verified the essential DH foundations and five maturity stages of the DHPMAT-MM. An unanticipated finding was the existence of a new category of community-facing MMs. Common characteristics included:1. A dynamic lifecycle approach to digital capability maturity, which is:a. responsive to environmental changes and may improve or worsen over time;b. accumulative, incorporating the attributes of the preceding stage; andc. sequential, where no maturity stage must be skipped.2. Sociotechnical quality improvement of the DH ecosystem and MM, which includes:a. investing in the organization's human, hardware, and software resources andb. a need to engage and improve the DH competencies of citizens.

CONCLUSIONS

The diversity in MMs and variability in methods and content can create cognitive dissonance. A metamodel like the DHPMAT-MM can logically unify the many domain-specific MMs and guide the overall implementation and evaluation of DH ecosystems and MMs over the maturity lifecycle.

Prospective validation of a dynamic prognostic model for identifying COVID-19 patients at high risk of rapid deterioration

BACKGROUND

We sought to develop and prospectively validate a dynamic model that incorporates changes in biomarkers to predict rapid clinical deterioration in patients hospitalized for COVID-19.

METHODS

We established a retrospective cohort of hospitalized patients aged ≥18 years with laboratory-confirmed COVID-19 using electronic health records (EHR) from a large integrated care delivery network in Massachusetts including >40 facilities from March to November 2020. A total of 71 factors, including time-varying vital signs and laboratory findings during hospitalization were screened. We used elastic net regression and tree-based scan statistics for variable selection to predict rapid deterioration, defined as progression by two levels of a published severity scale in the next 24 h. The development cohort included the first 70% of patients identified chronologically in calendar time; the latter 30% served as the validation cohort. A cut-off point was estimated to alert clinicians of high risk of imminent clinical deterioration.

RESULTS

Overall, 3706 patients (2587 in the development and 1119 in the validation cohort) met the eligibility criteria with a median of 6 days of follow-up. Twenty-four variables were selected in the final model, including 16 dynamic changes of laboratory results or vital signs. Area under the ROC curve was 0.81 (95% CI, 0.79-0.82) in the development set and 0.74 (95% CI, 0.71-0.78) in the validation set. The model was well calibrated (slope = 0.84 and intercept = -0.07 on the calibration plot in the validation set). The estimated cut-off point, with a positive predictive value of 83%, was 0.78.

CONCLUSIONS

Our prospectively validated dynamic prognostic model demonstrated temporal generalizability in a rapidly evolving pandemic and can be used to inform day-to-day treatment and resource allocation decisions based on dynamic changes in biophysiological factors.

Stochastic, individual animal systems simulation model of beef cow-calf production: development and validation

A stochastic, individual animal systems simulation model describing U.S. beef cow-calf production was developed and parameterized to match typical U.S. Angus genetics under cow-calf production conditions in the Kansas Flint Hills. Model simulation results were compared to available actual, multivariate U.S. cow-calf production data reported according to beef cow-calf standardized performance analysis (SPA) methodology through North Dakota State University's CHAPS program to assess model validity. Individual animal nutrition, reproduction, growth, and health characteristics, as well as production state are determined on a daily time step. Any number of days can be simulated. These capabilities allow for decision analysis and assessment of long-run outcomes of various genetic, management, and economic scenarios regarding multiple metrics simultaneously. Parameterizing the model to match Kansas Flint Hills production conditions for the years 1995 through 2018, 32 different genetic combinations for mature cow weight and peak lactation potential were simulated with 100 iterations each. Sire mature cow weight genetics ranged from 454 to 771 kg in 45 to 46 kg increments. Sire peak lactation genetics were considered at 6.8, 9, 11.3, and 13.6 kg/d for all eight mature cow weights. Utilizing model results for the years 2000 to 2018, raw model results were assessed against actual historical cow-calf production data. Exploratory factor analysis was applied to interpret the underlying factor scores of model output relative to actual cow-calf production data. Comparing modeled herd output with CHAPS herd data, median average calf weaning age, average cow age, percent pregnant per cow exposed, and percent calf mortality per calf born of model output was 3.4 d greater, 0.2 yr greater, 1 percentage point less, and 1.7 percentage points greater, respectively. Subtracting the median CHAPS pre-weaning average daily gain from the median modeled pre-weaning average daily gain for each of the eight respective mature cow weight genetics categories, and then calculating the median of the eight values, the median difference was -0.21 kg/d. Performing the same calculation for birth weight and adjusted 205 d weaning weight, the modeled data was 4.9 and 48.6 kg lighter than the CHAPS data, respectively. Management and genetic details underlying the CHAPS data were unknown.

Modeling the association between socioeconomic features and risk of flood damage: A local-scale case study in Sri Lanka

Floods cause severe damage to people as well as to properties. The same flood can cause different levels of damage to different households, but investigations into floods tend to be conducted on regional and national scales, thereby missing these local variations. It is therefore necessary to understand individual experiences of flood damage to implement effective flood management strategies on a local scale. The main objectives of this study were to develop a model that represents the relationship between socioeconomic conditions and flood damage at a local scale, and to understand the socioeconomic factors most closely tied to flood damage. The analysis is novel in that it considers not only the impact of flood characteristics, but also the impact of social, economic, and geographic factors on flood damage. This analysis derives from a quantitative modeling approach based on community responses, with the responses obtained through questionnaire surveys that consider four consecutive floods of differing severity. Path analysis was used to develop a model to represent the relationships between these factors. A randomly selected sample of 150 data points was used for model development, and nine random samples of 150 data points were used to validate the model. Results suggest that poor households, located in vulnerable, low-lying areas near rivers, suffer the most from being exposed to frequent, severe floods. Further, the results show that the socioeconomic factors with the most significant bearing on flood damage are per capita income and geographic location of the household. The results can be represented as a cycle, showing that social, economic, geographic, and flood characteristics are interrelated in ways that influence flood damage. This empirical analysis highlights a need for local-scale flood damage assessments, as offered in this article but seldom seen in other relevant literature. Our assessment was achieved by analyzing the impact of socioeconomic and geographic conditions and considering the relationship between flood characteristics and flood damage.

A model to assess the impact of digital technologies on the health-related quality of life

OBJECTIVES

Health-related quality of life (HRQoL) is a vital instrument to account for individuals' well-being in various settings. However, no model of HRQoL allows for examining the effect of digital technology on HRQoL. Therefore, we extend an established HRQoL model by adding a digital technology-related construct. We refer to this extension as the technology-affected health-related quality of life (TA-HRQoL).

METHODS

We investigate the extended TA-HRQoL model through a survey. In the survey, we exemplify the use of digital technology through a device for self-managing bladder dysfunction. Hence, we explore whether the model extension proposed is valid and how determinants of the HRQoL affect patients with bladder dysfunction.

RESULTS

The results indicate that the use of digital technology improves the HRQoL. In our exemplary use scenario, the digital technology decreases bladder-related functional impairments and increases well-being and life satisfaction directly.

CONCLUSIONS

Our study may provide evidence for the influence of digital technologies on the HRQoL, thus supporting our model extension. We consider our proposed TA-HRQoL model as valid and as useful to account for the influence of digital technology on an individual's HRQoL. With the TA-HRQoL model, the impact of a digital technology on an individual's HRQoL can be assessed.

Training Ultrasound Image Classification Deep-Learning Algorithms for Pneumothorax Detection Using a Synthetic Tissue Phantom Apparatus

Ultrasound (US) imaging is a critical tool in emergency and military medicine because of its portability and immediate nature. However, proper image interpretation requires skill, limiting its utility in remote applications for conditions such as pneumothorax (PTX) which requires rapid intervention. Artificial intelligence has the potential to automate ultrasound image analysis for various pathophysiological conditions. Training models require large data sets and a means of troubleshooting in real-time for ultrasound integration deployment, and they also require large animal models or clinical testing. Here, we detail the development of a dynamic synthetic tissue phantom model for PTX and its use in training image classification algorithms. The model comprises a synthetic gelatin phantom cast in a custom 3D-printed rib mold and a lung mimicking phantom. When compared to PTX images acquired in swine, images from the phantom were similar in both PTX negative and positive mimicking scenarios. We then used a deep learning image classification algorithm, which we previously developed for shrapnel detection, to accurately predict the presence of PTX in swine images by only training on phantom image sets, highlighting the utility for a tissue phantom for AI applications.

The Advantage of Case-Tailored Information Metrics for the Development of Predictive Models, Calculated Profit in Credit Scoring

This paper compares model development strategies based on different performance metrics. The study was conducted in the area of credit risk modeling with the usage of diverse metrics, including general-purpose Area Under the ROC curve (AUC), problem-dedicated Expected Maximum Profit (EMP) and the novel case-tailored Calculated Profit (CP). The metrics were used to optimize competitive credit risk scoring models based on two predictive algorithms that are widely used in the financial industry: Logistic Regression and extreme gradient boosting machine (XGBoost). A dataset provided by the American Fannie Mae agency was utilized to conduct the study. In addition to the baseline study, the paper also includes a stability analysis. In each case examined the proposed CP metric that allowed us to achieve the most profitable loan portfolio.

Simulating indoor inorganic aerosols of outdoor origin with the inorganic aerosol thermodynamic equilibrium model ISORROPIA

Outdoor aerosols can transform and have their composition altered upon transport indoors. Herein, IMAGES, a platform that simulates indoor organic aerosol with the 2-dimensional volatility basis set (2D-VBS), was extended to incorporate the inorganic aerosol thermodynamic equilibrium model, ISORROPIA. The model performance was evaluated by comparing aerosol component predictions to indoor measurements from an aerosol mass spectrometer taken during the summer and winter seasons. Since ammonia was not measured in the validation dataset, outdoor ammonia was estimated from aerosol measurements using a novel pH-based algorithm, while nitric acid was held constant. Modeled indoor ammonia sources included temperature-based occupant and surface emissions. Sensitivity to the nitric acid indoor surface deposition rate β g , HNO 3 , g was explored by varying it in model runs, which did not affect modeled sulfate due to its non-volatile nature, though the fitting of a filter efficiency was required for good correlations of modeled sulfate with measurements in both seasons. Modeled summertime nitrate well-matched measured observations when β g , HNO 3 , g = 2.75 h - 1 , but wintertime comparisons were poor, possibly due to missing thermodynamic processes within the heating, ventilating, and air-conditioning (HVAC) system. Ammonium was consistently overpredicted, potentially due to neglecting thirdhand smoke impacts observed in the field campaign, as well as HVAC impacts.

A Model for Estimating Biological Age From Physiological Biomarkers of Healthy Aging: Cross-sectional Study

BACKGROUND

Individual differences in the rate of aging and susceptibility to disease are not accounted for by chronological age alone. These individual differences are better explained by biological age, which may be estimated by biomarker prediction models. In the light of the aging demographics of the global population and the increase in lifestyle-related morbidities, it is interesting to invent a new biological age model to be used for health promotion.

OBJECTIVE

This study aims to develop a model that estimates biological age based on physiological biomarkers of healthy aging.

METHODS

Carefully selected physiological variables from a healthy study population of 100 women and men were used as biomarkers to establish an estimate of biological age. Principal component analysis was applied to the biomarkers and the first principal component was used to define the algorithm estimating biological age.

RESULTS

The first principal component accounted for 31% in women and 25% in men of the total variance in the biological age model combining mean arterial pressure, glycated hemoglobin, waist circumference, forced expiratory volume in 1 second, maximal oxygen consumption, adiponectin, high-density lipoprotein, total cholesterol, and soluble urokinase-type plasminogen activator receptor. The correlation between the corrected biological age and chronological age was r=0.86 (P<.001) and r=0.81 (P<.001) for women and men, respectively, and the agreement was high and unbiased. No difference was found between mean chronological age and mean biological age, and the slope of the regression line was near 1 for both sexes.

CONCLUSIONS

Estimating biological age from these 9 biomarkers of aging can be used to assess general health compared with the healthy aging trajectory. This may be useful to evaluate health interventions and as an aid to enhance awareness of individual health risks and behavior when deviating from this trajectory.

TRIAL REGISTRATION

ClinicalTrials.gov NCT03680768; https://clinicaltrials.gov/ct2/show/NCT03680768.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

RR2-10.2196/19209.

Development of the PhysioVessel: a customizable platform for simulating physiological fluid resuscitation

Uncontrolled hemorrhage is a leading cause of death in trauma situations. Developing solutions to automate hemorrhagic shock resuscitation may improve the outcomes for trauma patients. However, testing and development of automated solutions to address critical care interventions, oftentimes require extensive large animal studies for even initial troubleshooting. The use of accurate laboratory or in-silico models may provide a way to reduce the need for large animal datasets. Here, a tabletop model, for use in the development of fluid resuscitation with physiologically relevant pressure-volume responsiveness for high throughput testing, is presented. The design approach shown can be applied to any pressure-volume dataset through a process of curve-fitting, 3D modeling, and fabrication of a fluid reservoir shaped to the precise curve fit. Two case studies are presented here based on different resuscitation fluids: whole blood and crystalloid resuscitation. Both scenarios were derived from data acquired during porcine hemorrhage studies, used a pressure-volume curve to design and fabricate a 3D model, and evaluated to show that the test platform mimics the physiological data. The vessels produced based on data collected from pigs infused with whole blood and crystalloid were able to reproduce normalized pressure-volume curves within one standard deviation of the porcine data with mean residual differences of 0.018 and 0.016, respectively. This design process is useful for developing closed-loop algorithms for resuscitation and can simplify initial testing of technologies for this life-saving medical intervention.

Multi-Level Sensing Technologies in Landslide Research-Hrvatska Kostajnica Case Study, Croatia

In March 2018, a landslide in Hrvatska Kostajnica completely destroyed multiple households. The damage was extensive, and lives were endangered. The question remains: Can it happen again? To enhance the knowledge and understanding of the soil and rock behaviour before, during, and after this geo-hazard event, multi-level sensing technologies in landslide research were applied. Day after the event field mapping and unmanned aerial vehicle (UAV) data were collected with the inspection of available orthophoto and "geo" data. For the landslide, a new geological column was developed with mineralogical and geochemical analyses. The application of differential interferometric synthetic aperture radar (DInSAR) for detecting ground surface displacement was undertaken in order to determine pre-failure behaviour and to give indications about post-failure deformations. In 2020, electrical resistivity tomography (ERT) in the landslide body was undertaken to determine the depth of the landslide surface, and in 2021 ERT measurements in the vicinity of the landslide area were performed to obtain undisturbed material properties. Moreover, in 2021, detailed light detection and ranging (LIDAR) data were acquired for the area. All these different level data sets are being analyzed in order to develop a reliable landslide model as a first step towards answering the aforementioned question. Based on applied multi-level sensing technologies and acquired data, the landslide model is taking shape. However, further detailed research is still recommended.

A model to understand HIV-related stigma and the psychosocial well-being of children orphaned by AIDS: a theory generative approach

HIV-related stigma has negatively impacted the psychosocial well-being of children who have been orphaned by AIDS-related causes. Response to reducing stigma and ensuring child well-being is hindered by the limited understanding of HIV-related stigma and how it affects the psychosocial well-being of children. Due to the call for a comprehensive understanding of HIV-related stigma, this study aimed to develop a model to understand the manner in which HIV-related stigma affects the psychosocial well-being of children orphaned by AIDS. The study implemented a mixed method, exploratory, sequential design within a theory generative approach that included concept development, statement development, model description, and model evaluation. The developed model indicated that HIV-related stigma is embedded in social interaction and mediated by children orphaned by AIDS response to stigma. HIV-related stigma and maladaptive coping strategies collectively affect several domains of child psychosocial well-being and elevate psychosocial distress. This is the first model to provide a child-centred understanding of HIV-related stigma and its consequences for psychosocial well-being. The model may be used to guide future research and inform the development of appropriate interventions.

A Deep Learning Approach to Refine the Identification of High-Quality Clinical Research Articles From the Biomedical Literature: Protocol for Algorithm Development and Validation

Background

A barrier to practicing evidence-based medicine is the rapidly increasing body of biomedical literature. Use of method terms to limit the search can help reduce the burden of screening articles for clinical relevance; however, such terms are limited by their partial dependence on indexing terms and usually produce low precision, especially when high sensitivity is required. Machine learning has been applied to the identification of high-quality literature with the potential to achieve high precision without sacrificing sensitivity. The use of artificial intelligence has shown promise to improve the efficiency of identifying sound evidence.

Objective

The primary objective of this research is to derive and validate deep learning machine models using iterations of Bidirectional Encoder Representations from Transformers (BERT) to retrieve high-quality, high-relevance evidence for clinical consideration from the biomedical literature.

Methods

Using the HuggingFace Transformers library, we will experiment with variations of BERT models, including BERT, BioBERT, BlueBERT, and PubMedBERT, to determine which have the best performance in article identification based on quality criteria. Our experiments will utilize a large data set of over 150,000 PubMed citations from 2012 to 2020 that have been manually labeled based on their methodological rigor for clinical use. We will evaluate and report on the performance of the classifiers in categorizing articles based on their likelihood of meeting quality criteria. We will report fine-tuning hyperparameters for each model, as well as their performance metrics, including recall (sensitivity), specificity, precision, accuracy, F-score, the number of articles that need to be read before finding one that is positive (meets criteria), and classification probability scores.

Results

Initial model development is underway, with further development planned for early 2022. Performance testing is expected to star in February 2022. Results will be published in 2022.

Conclusions

The experiments will aim to improve the precision of retrieving high-quality articles by applying a machine learning classifier to PubMed searching.

International Registered Report Identifier (IRRID)

DERR1-10.2196/29398

Machine Learning-Based Prediction of COVID-19 Mortality With Limited Attributes to Expedite Patient Prognosis and Triage: Retrospective Observational Study

Background

The onset and development of the COVID-19 pandemic have placed pressure on hospital resources and staff worldwide. The integration of more streamlined predictive modeling in prognosis and triage–related decision-making can partly ease this pressure.

Objective

The objective of this study is to assess the performance impact of dimensionality reduction on COVID-19 mortality prediction models, demonstrating the high impact of a limited number of features to limit the need for complex variable gathering before reaching meaningful risk labelling in clinical settings.

Methods

Standard machine learning classifiers were employed to predict an outcome of either death or recovery using 25 patient-level variables, spanning symptoms, comorbidities, and demographic information, from a geographically diverse sample representing 17 countries. The effects of feature reduction on the data were tested by running classifiers on a high-quality data set of 212 patients with populated entries for all 25 available features. The full data set was compared to two reduced variations with 7 features and 1 feature, respectively, extracted using univariate mutual information and chi-square testing. Classifier performance on each data set was then assessed on the basis of accuracy, sensitivity, specificity, and received operating characteristic–derived area under the curve metrics to quantify benefit or loss from reduction.

Results

The performance of the classifiers on the 212-patient sample resulted in strong mortality detection, with the highest performing model achieving specificity of 90.7% (95% CI 89.1%-92.3%) and sensitivity of 92.0% (95% CI 91.0%-92.9%). Dimensionality reduction provided strong benefits for performance. The baseline accuracy of a random forest classifier increased from 89.2% (95% CI 88.0%-90.4%) to 92.5% (95% CI 91.9%-93.0%) when training on 7 chi-square–extracted features and to 90.8% (95% CI 89.8%-91.7%) when training on 7 mutual information–extracted features. Reduction impact on a separate logistic classifier was mixed; however, when present, losses were marginal compared to the extent of feature reduction, altogether showing that reduction either improves performance or can reduce the variable-sourcing burden at hospital admission with little performance loss. Extreme feature reduction to a single most salient feature, often age, demonstrated large standalone explanatory power, with the best-performing model achieving an accuracy of 81.6% (95% CI 81.1%-82.1%); this demonstrates the relatively marginal improvement that additional variables bring to the tested models.

Conclusions

Predictive statistical models have promising performance in early prediction of death among patients with COVID-19. Strong dimensionality reduction was shown to further improve baseline performance on selected classifiers and only marginally reduce it in others, highlighting the importance of feature reduction in future model construction and the feasibility of deprioritizing large, hard-to-source, and nonessential feature sets in real world settings.

An Open-Globe Porcine Injury Platform for Assessing Therapeutics and Characterizing Biological Effects

Open-globe injuries can result in permanent vision loss, partly due to extended delays between injury and medical intervention. Even with early intervention, the management of open-globe injuries remains a challenge for ophthalmologists, mostly due to inadequate or suboptimal current therapies. To aid in the development of novel therapeutics and track toxicological and pathophysiological changes, this article details an open-globe injury platform capable of inducing injuries in enucleated porcine eyes. The injury platform relies on a high-speed solenoid device to mimic explosive injury scenarios, allowing for large, complex injury shapes and sizes that are often observed in casualties and are more difficult to treat. The system can be implemented with precise computer control of the injury mechanism to allow for more complex setups. Also, the system can make use of real-time intraocular pressure measurement to track changes during injury induction and to assess therapeutic efficacy for restoring intraocular pressure and the integrity of the eye. These protocols will assist with implementation of the injury model in prospective laboratories seeking to develop therapeutics or studying biological changes that occur from this type of traumatic injury. Published 2020. U.S. Government. Basic Protocol 1: Preparing gelatin molds and porcine eye tissue Basic Protocol 2: Creating an open-globe injury using a solenoid device Alternate Protocol 1: Constructing a computer-controlled system for open-globe injury Alternate Protocol 2: Constructing a pressure measurement system for tracking intraocular pressure Support Protocol 1: Assessing ocular compliance in porcine eyes Support Protocol 2: Assessing outflow rate from the anterior chamber Support Protocol 3: Assessing burst pressure in porcine eyes.

A Binary Logistic Regression Model as a Tool to Predict Craft Beer Susceptibility to Microbial Spoilage

Beer spoilage caused by microorganisms, which is a major concern for brewers, produces undesirable aromas and flavors in the final product and substantial financial losses. To address this problem, brewers need easy-to-apply tools that inform them of beer susceptibility to the microbial spoilage. In this study, a growth/no growth (G/NG) binary logistic regression model to predict this susceptibility was developed. Values of beer physicochemical parameters such as pH, alcohol content (% ABV), bitterness units (IBU), and yeast-fermentable extract (% YFE) obtained from the analysis of twenty commercially available craft beers were used to prepare 22 adjusted beers at different levels of each parameter studied. These preparations were assigned as a first group of samples, while 17 commercially available beers samples as a second group. The results of G/NG from both groups, after artificially inoculating with one wild yeast and different lactic acid bacteria (LAB) previously adapted to grow in a beer-type beverage, were used to design the model. The developed G/NG model correctly classified 276 of 331 analyzed cases and its predictive ability was 100% in external validation. This G/NG model has good sensitivity and goodness of fit (87% and 83.4%, respectively) and provides the potential to predict craft beer susceptibility to microbial spoilage.

Development and Implementation of Couple-Based Collaborative Management Model of Type 2 Diabetes Mellitus for Community-Dwelling Chinese Older Adults: A Pilot Randomized Trial

Background: To mobilize family's positive involvement in improving and sustaining self-management activities of older adults with diabetes, we developed a couple-based collaborative management model (CCMM) for community-dwelling older Chinese. Methods: The model was developed stepwise through applying theoretical models, interviewing older couples and community healthcare workers, as well as incorporating expert reviews. A 3-month pilot study was conducted to test the model's feasibility and its treatment effects by linear regression on 18 pairs of older couples aged 60 years+, who were equally divided into a couple-based intervention arm and a patient-only control arm. Results: The developed CCMM covered four theory-driven intervention modules: dyadic assessment, dyadic education, dyadic behavior-change training, and dyadic monitoring. Each module was delivered by community healthcare workers and targeted at older couples as the management units. Based on interviews with older couples and healthcare workers, 4 weekly education and training group sessions and 2-month weekly behavior change booster calls were designed to address older adults' main management barriers. These modules and session contents were evaluated as essential and relevant by the expert panel. Furthermore, the CCMM showed good feasibility and acceptability in the pilot, with non-significant yet more positive changes in physiological outcomes of diabetic participants and couples' well-being and exercise levels of these in the intervention arm than their controlled counterparts. Conclusion: We systematically developed a couple-based collaborative management model of diabetes, which was well-received by healthcare practitioners and highly feasible among older Chinese couples living in the community. The model's treatment effects need to be verified in fully powered randomized controlled trials. Clinical Trial Registration: http://www.chictr.org.cn/showproj.aspx?proj=42964, identifier: ChiCTR1900027137.

Development and Proof of Concept of a Low-Cost Ultrasound Training Model for Diagnosis of Giant Cell Arteritis Using 3D Printing

Objectives: Currently, ultrasound (US) is widely used for the diagnosis of giant cell arteritis (GCA). Our aim was to develop a low-cost US training model for diagnosis of GCA of the temporal and axillary artery using a modern 3D printing system. Methods: We designed an US training model, which enables measurement of the intima-media thickness (IMT) of temporal and axillary arteries using Autodesk Fusion360. This model was printed using a modern 3D printer (Formlabs Form3) and embedded in ballistic gelatine. The ultrasound images including measurement of the IMT by ultrasound specialists in GCA were compared to ultrasound images in acute GCA and healthy subjects. Results: Our ultrasound training model of the axillary and temporal artery displayed a very similar ultrasound morphology compared to real US images and fulfilled the OMERACT ultrasound definitions of normal and pathological temporal and axillary arteries in GCA. The IMT measurements were in line with published cut-off values for normal and pathological IMT values in GCA and healthy individuals. When testing the models on blinded US specialists in GCA, they were identified correctly in all test rounds with an intra-class coefficient of 0.99. Conclusion: The production of low-cost ultrasound training models of normal and pathological temporal and axillary arteries in GCA, which fulfil the OMERACT ultrasound definitions and adhere to the published IMT cut-off values in GCA, is feasible. Ultrasound specialists identified each respective model correctly in every case.

Developing clinical prediction models for nonrecovery in older patients seeking care for back pain: the back complaints in the elders prospective cohort study

ABSTRACT

Back pain is a leading cause of disability worldwide and is common in older adults. No clinical prediction models for poor long-term outcomes have been developed in older patients with back pain. This study aimed to develop and internally validate 3 clinical prediction models for nonrecovery in this population. A prospective cohort study in general practice was conducted (Back Complaints in the Elders, Netherlands), including 675 patients >55 years with a new episode of care for back pain. Three definitions of nonrecovery were used combining 6-month and 12-month follow-up data: (1) persistent back pain, (2) persistent disability, and (3) perceived nonrecovery. Sample size calculation resulted in a maximum of 14 candidate predictors that were selected from back pain prognostic literature and clinical experience. Multivariable logistic regression was used to develop the models (backward selection procedure). Models' performance was evaluated with explained variance (Nagelkerke's R2), calibration (Hosmer-Lemeshow test), and discrimination (area under the curve [AUC]) measures. The models were internally validated in 250 bootstrapped samples to correct for overoptimism. All 3 models displayed good overall performance during development and internal validation (ie, R2 > 30%; AUC > 0.77). The model predicting persistent disability performed best, showing good calibration, discrimination (AUC 0.86, 95% confidence interval 0.83-0.89; optimism-adjusted AUC 0.85), and explained variance (R2 49%, optimism-adjusted R2 46%). Common predictors in all models were: age, chronic duration, disability, a recent back pain episode, and patients' recovery expectations. Spinal morning stiffness and pain during spinal rotation were included in 2 of 3 models. These models should be externally validated before being used in a clinical primary care setting.

Aminoglycoside-induced Hearing Loss Among Patients Being Treated for Drug-resistant Tuberculosis in South Africa: A Prediction Model

BACKGROUND

Individuals treated for drug-resistant tuberculosis (DR-TB) with aminoglycosides (AGs) in resource-limited settings often experience permanent hearing loss, yet there is no practical method to identify those at higher risk. We sought to develop a clinical prediction model of AG-induced hearing loss among patients initiating DR-TB treatment in South Africa.

METHODS

Using nested, prospective data from a cohort of 379 South African adults being treated for confirmed DR-TB with AG-based regimens we developed the prediction model using multiple logistic regression. Predictors were collected from clinical, audiological, and laboratory evaluations conducted at the initiation of DR-TB treatment. The outcome of AG-induced hearing loss was identified from audiometric and clinical evaluation by a worsened hearing threshold compared with baseline during the 6-month intensive phase.

RESULTS

Sixty-three percent of participants (n = 238) developed any level of hearing loss. The model predicting hearing loss at frequencies from 250 to 8000 Hz included weekly AG dose, human immunodeficiency virus status with CD4 count, age, serum albumin, body mass index, and pre-existing hearing loss. This model demonstrated reasonable discrimination (area under the receiver operating characteristic curve [AUC] = 0.71) and calibration (χ2[8] = 6.10, P = .636). Using a cutoff of 80% predicted probability of hearing loss, the positive predictive value of this model was 83% and negative predictive value was 40%. Model discrimination was similar for ultrahigh-frequency hearing loss (frequencies >9000 Hz; AUC = 0.81) but weaker for clinically determined hearing loss (AUC = 0.60).

CONCLUSIONS

This model may identify patients with DR-TB who are at highest risk of developing AG-induced ototoxicity and may help prioritize patients for AG-sparing regimens in clinical settings where access is limited.

Simulating forest resilience: A review

AIM

Simulation models are important tools for quantifying the resilience (i.e., persistence under changed environmental conditions) of forest ecosystems to global change. We synthesized the modelling literature on forest resilience, summarizing common models and applications in resilience research, and scrutinizing the implementation of important resilience mechanisms in these models. Models applied to assess resilience are highly diverse, and our goal was to assess how well they account for important resilience mechanisms identified in experimental and empirical research.

LOCATION

Global.

TIME PERIOD

1994 to 2019.

MAJOR TAXA STUDIED

Trees.

METHODS

We reviewed the forest resilience literature using online databases, selecting 119 simulation modelling studies for further analysis. We identified a set of resilience mechanisms from the general resilience literature and analysed models for their representation of these mechanisms. Analyses were grouped by investigated drivers (resilience to what) and responses (resilience of what), as well as by the type of model being used.

RESULTS

Models used to study forest resilience varied widely, from analytical approaches to complex landscape simulators. The most commonly addressed questions were associated with resilience of forest cover to fire. Important resilience mechanisms pertaining to regeneration, soil processes, and disturbance legacies were explicitly simulated in only 34 to 46% of the model applications.

MAIN CONCLUSIONS

We found a large gap between processes identified as underpinning forest resilience in the theoretical and empirical literature, and those represented in models used to assess forest resilience. Contemporary forest models developed for other goals may be poorly suited for studying forest resilience during an era of accelerating change. Our results highlight the need for a new wave of model development to enhance understanding of and management for resilient forests.

Deep Learning Benchmarks on L1000 Gene Expression Data

Gene expression data can offer deep, physiological insights beyond the static coding of the genome alone. We believe that realizing this potential requires specialized, high-capacity machine learning methods capable of using underlying biological structure, but the development of such models is hampered by the lack of published benchmark tasks and well characterized baselines. In this work, we establish such benchmarks and baselines by profiling many classifiers against biologically motivated tasks on two curated views of a large, public gene expression dataset (the LINCS corpus) and one privately produced dataset. We provide these two curated views of the public LINCS dataset and our benchmark tasks to enable direct comparisons to future methodological work and help spur deep learning method development on this modality. In addition to profiling a battery of traditional classifiers, including linear models, random forests, decision trees, K nearest neighbor (KNN) classifiers, and feed-forward artificial neural networks (FF-ANNs), we also test a method novel to this data modality: graph convolugtional neural networks (GCNNs), which allow us to incorporate prior biological domain knowledge. We find that GCNNs can be highly performant, with large datasets, whereas FF-ANNs consistently perform well. Non-neural classifiers are dominated by linear models and KNN classifiers.

A Novel Prognostic Scoring System Integrating Gene Expressions and Clinicopathological Characteristics to Predict Very Early Relapse in Node-Negative Estrogen Receptor-Positive/HER2-Negative Breast Cancer

Background: Despite low aggressiveness in tumor biology and high responsiveness to endocrine therapy, subgroups of patients with estrogen receptor-positive/HER2-negative (ER+/HER2-) breast cancer relapse early in the first two years after initiation of endocrine therapy, indicating potential endocrine resistance. Accordingly, we attempted to establish a scoring system to inform the first-2-year prognosis (F2P Score). Methods: Patients with node-negative ER+/HER2- breast cancer and complete data of gene expressions in a 21-gene panel were retrospectively retrieved from Shanghai Jiao Tong University Breast Cancer Database (SJTU-BCDB). The F2P Score was established based on the clinical and genomic variables associated with the first-2-year relapse after shrinkage correction and validated using the bootstrap resampling method. Model performance was quantified by Harrell's concordance-index (C-index) and Bayesian information criteria (BIC). Results: The F2P Score was established by integrating the clinical (age and tumor size) and genomic (ESR1, PGR, BCL2, CD68, GSTM1, and BAG1) variables with a C-index of 0.71 and BIC of 397.46. Bootstrap C-index was 0.72 (95% CI, 0.62-0.81) and BIC was 396.75 (95% CI, 252.37-541.13). A higher score indicated an increased likelihood of a first-2-year relapse, when used as continuous (HR, 2.94; 95% CI, 1.87-4.61) or categorical (HR, 3.68; 95% CI, 1.70-8.00) predictors in multivariate analysis. Both continuous and categorical F2P Score also remained prognostic for overall survival and other endpoints. No significant interaction was observed between the F2P Score and treatment subgroups. Additionally, the F2P Score outperformed the IHC4, clinical treatment score and 21-gene test in predicting first-2-year relapse. Conclusion: The F2P Score reported herein, integrating the clinicopathological and genomic variables, may inform prognosis and endocrine responsiveness. After the benefits and risks have been considered, treatment escalation may be an alternative strategy for patients with a higher score.

Data-driven Evolution Equation Reconstruction for Parameter-Dependent Nonlinear Dynamical Systems

When studying observations of chemical reaction dynamics, closed form equations based on a putative mechanism may not be available. Yet when sufficient data from experimental observations can be obtained, even without knowing what exactly the physical meaning of the parameter settings or recorded variables are, data-driven methods can be used to construct minimal (and in a sense, robust) realizations of the system. The approach attempts, in a sense, to circumvent physical understanding, by building intrinsic "information geometries" of the observed data, and thus enabling prediction without physical/chemical knowledge. Here we use such an approach to obtain evolution equations for a data-driven realization of the original system - in effect, allowing prediction based on the informed interrogation of the agnostically organized observation database. We illustrate the approach on observations of (a) the normal form for the cusp singularity, (b) a cusp singularity for the nonisothermal CSTR, and (c) a random invertible transformation of the nonisothermal CSTR, showing that one can predict even when the observables are not "simply explainable" physical quantities. We discuss current limitations and possible extensions of the procedure.

Developments in the MPI-M Earth System Model version 1.2 (MPI-ESM1.2) and Its Response to Increasing CO2

A new release of the Max Planck Institute for Meteorology Earth System Model version 1.2 (MPI-ESM1.2) is presented. The development focused on correcting errors in and improving the physical processes representation, as well as improving the computational performance, versatility, and overall user friendliness. In addition to new radiation and aerosol parameterizations of the atmosphere, several relatively large, but partly compensating, coding errors in the model's cloud, convection, and turbulence parameterizations were corrected. The representation of land processes was refined by introducing a multilayer soil hydrology scheme, extending the land biogeochemistry to include the nitrogen cycle, replacing the soil and litter decomposition model and improving the representation of wildfires. The ocean biogeochemistry now represents cyanobacteria prognostically in order to capture the response of nitrogen fixation to changing climate conditions and further includes improved detritus settling and numerous other refinements. As something new, in addition to limiting drift and minimizing certain biases, the instrumental record warming was explicitly taken into account during the tuning process. To this end, a very high climate sensitivity of around 7 K caused by low-level clouds in the tropics as found in an intermediate model version was addressed, as it was not deemed possible to match observed warming otherwise. As a result, the model has a climate sensitivity to a doubling of CO2 over preindustrial conditions of 2.77 K, maintaining the previously identified highly nonlinear global mean response to increasing CO2 forcing, which nonetheless can be represented by a simple two-layer model.

ChemSuite: A package for chemoinformatics calculations and machine learning

Prediction of biological and toxicological properties of small molecules using in silico approaches has become a wide practice in pharmaceutical research to lessen the cost and enhance productivity. The development of a tool "ChemSuite," a stand-alone application for chemoinformatics calculations and machine-learning model development, is reported. Availability of multi-functional features makes it widely acceptable in various fields. Force field such as UFF is incorporated in tool for optimization of molecules. Packages like RDKit, PyDPI and PaDEL help to calculate 1D, 2D and 3D descriptors and more than 10 types of fingerprints. MinMax Scaler and Z-Score algorithms are available to normalize descriptor values. Varied descriptor selection and machine-learning algorithms are available for model development. It allows the user to add their own algorithm or extend the software for various scientific purposes. It is free, open source and has user-friendly graphical interface, and it can work on all major platforms.

Population Pharmacokinetic Analyses and Model Validation of Tolvaptan in Subjects With Autosomal Dominant Polycystic Kidney Disease

Tolvaptan is the first approved drug treatment to slow kidney function decline in adults at risk of rapidly progressing autosomal dominant polycystic kidney disease (ADPKD). The objective is to develop (1091 subjects, 7335 observations) and validate (678 subjects, 3012 observations) a population pharmacokinetic model to describe tolvaptan pharmacokinetics in ADPKD subjects. The final model was evaluated with a bootstrapping method. The final model was internally and externally evaluated using visual predictive checks (VPC). Pharmacokinetics was best described by a 1‐compartmental model with 0‐order absorption, nonlinear relative bioavailability (F1), and first‐order elimination. Accounting for changes in F1 significantly improved the model: as the dose increased from 15 mg to 120 mg, F1 decreased by 36%. Population estimates for clearance/F (CL/F), volume of distribution/F (Vd/F), duration of absorption (D1), the highest dose at which F1 is lowest, and the amount of dose at which F1 is 50% were 12.6 L·h^‐1, 110 L, 0.58 hour, 182 mg, and 166 mg, respectively. The interindividual variability was 64% in CL/F, 70% in Vd/F, and 238% in D1. Residual variability was described by a combined‐error model. The VPC (500 data sets simulated) showed that 76% to 92% of the observed data fell within the 90% prediction intervals. The model stability assessed by a 1000‐run bootstrap analysis showed that the mean parameter estimates of data were within 10% of those obtained with the final model. The developed model is robust and stable. Internal and external validation confirmed the model ability to describe the data optimally.