Combining structured and unstructured data for predictive models: a deep learning approach

Zhang, Dongdong; Yin, Changchang; Zeng, Jucheng; Yuan, Xiaohui; Zhang, Ping

doi:10.1186/s12911-020-01297-6

Cited by 142 publications

(95 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In health care, DL models have also been used for predicting disease onset, prognosis, and health outcomes. [8][9][10] For example, Rajkomar and colleagues (2018) demonstrated that the long short-term memory (LSTM) DL algorithm was superior to traditional non-DL predictive models for predicting health-care outcomes in the general inpatient population. 11 In the area of HF, LSTM has been utilized to predict disease onset with promising performance.…”

Section: Introductionmentioning

confidence: 99%

Using Deep Learning to Identify High-Risk Patients with Heart Failure with Reduced Ejection Fraction

Wang

Chen

Tan

et al. 2021

JHEOR

View full text Add to dashboard Cite

Background: Deep Learning (DL) has not been well-established as a method to identify high-risk patients among patients with heart failure (HF). Objectives: This study aimed to use DL models to predict hospitalizations, worsening HF events, and 30-day and 90-day readmissions in patients with heart failure with reduced ejection fraction (HFrEF). Methods: We analyzed the data of adult HFrEF patients from the IBM® MarketScan® Commercial and Medicare Supplement databases between January 1, 2015 and December 31, 2017. A sequential model architecture based on bi-directional long short-term memory (Bi-LSTM) layers was utilized. For DL models to predict HF hospitalizations and worsening HF events, we utilized two study designs: with and without a buffer window. For comparison, we also tested multiple traditional machine learning models including logistic regression, random forest, and eXtreme Gradient Boosting (XGBoost). Model performance was assessed by area under the curve (AUC) values, precision, and recall on an independent testing dataset. Results: A total of 47 498 HFrEF patients were included; 9427 with at least one HF hospitalization. The best AUCs of DL models without a buffer window in predicting HF hospitalizations and worsening HF events in the total patient cohort were 0.977 and 0.972; with a 7-day buffer window the best AUCs were 0.573 and 0.608, respectively. The best AUCs in predicting 30- and 90-day readmissions in all adult patients were 0.597 and 0.614, respectively. An AUC of 0.861 was attained for prediction of 90-day readmission in patients aged 18-64. For all outcomes assessed, the DL approach outperformed traditional machine learning models. Discussion: The DL approach can automate feature engineering during the model learning, which can increase the clinical applicability and lead to comparable or better model performance. However, the lack of granular clinical data, and sample size and imbalance issues may have limited the model’s performance. Conclusions: A DL approach using Bi-LSTM was shown to be a feasible and useful tool to predict HF-related outcomes. This study can help inform the future development and deployment of predictive tools to identify high-risk HFrEF patients and ultimately facilitate targeted interventions in clinical practice.

show abstract

Section: Introductionmentioning

confidence: 99%

Using Deep Learning to Identify High-Risk Patients with Heart Failure with Reduced Ejection Fraction

Wang

Chen

Tan

et al. 2021

JHEOR

View full text Add to dashboard Cite

show abstract

“…Davoodi et al [29] and Hoogendoorn et al [24] predicted after 24 h and within a median of 72 h, respectively. Studies by Tang et al [33], Caicedo-Torres et al [36], Sha et al [38], and Zhang et al [41] predicted in-hospital mortality irrespective of the admission or discharge time.…”

Section: Mortality Predictionmentioning

confidence: 96%

“…In addition to the most commonly used data elements, other clinical information such as medications, intake/output variables, risk scores, and comorbidities were also utilized. Weissman et al [31] and Zhang et al [41] used clinical variables from both structured and unstructured data types for mortality prediction.…”

Section: Mortality Predictionmentioning

confidence: 99%

Application of Machine Learning in Intensive Care Unit (ICU) Settings Using MIMIC Dataset: Systematic Review

Syed

Sexton

et al. 2021

Informatics

View full text Add to dashboard Cite

Modern Intensive Care Units (ICUs) provide continuous monitoring of critically ill patients susceptible to many complications affecting morbidity and mortality. ICU settings require a high staff-to-patient ratio and generates a sheer volume of data. For clinicians, the real-time interpretation of data and decision-making is a challenging task. Machine Learning (ML) techniques in ICUs are making headway in the early detection of high-risk events due to increased processing power and freely available datasets such as the Medical Information Mart for Intensive Care (MIMIC). We conducted a systematic literature review to evaluate the effectiveness of applying ML in the ICU settings using the MIMIC dataset. A total of 322 articles were reviewed and a quantitative descriptive analysis was performed on 61 qualified articles that applied ML techniques in ICU settings using MIMIC data. We assembled the qualified articles to provide insights into the areas of application, clinical variables used, and treatment outcomes that can pave the way for further adoption of this promising technology and possible use in routine clinical decision-making. The lessons learned from our review can provide guidance to researchers on application of ML techniques to increase their rate of adoption in healthcare.

show abstract

“…Data from the general population and subpopulations are yet to be effectively integrated with algorithms that contribute to a hybrid model of care as part of a general guide [5]. Machine learning and deep learning have until recently only been fed structured data in medical informatics, neglecting invaluable information from unstructured clinical notes [52]. Unstructured data from medical records combined with deep learning tools hold promise in identifying the most suitable candidates for a study on youth depression, but a reliable recommendation system has not been established [53].…”

Section: Digital Mental Health Implementation In a Hybrid Model Of Carementioning

confidence: 99%

Digital Mental Health Challenges and the Horizon Ahead for Solutions

Balcombe¹,

Leo²

2021

JMIR Ment Health

View full text Add to dashboard Cite

The demand outstripping supply of mental health resources during the COVID-19 pandemic presents opportunities for digital technology tools to fill this new gap and, in the process, demonstrate capabilities to increase their effectiveness and efficiency. However, technology-enabled services have faced challenges in being sustainably implemented despite showing promising outcomes in efficacy trials since the early 2000s. The ongoing failure of these implementations has been addressed in reconceptualized models and frameworks, along with various efforts to branch out among disparate developers and clinical researchers to provide them with a key for furthering evaluative research. However, the limitations of traditional research methods in dealing with the complexities of mental health care warrant a diversified approach. The crux of the challenges of digital mental health implementation is the efficacy and evaluation of existing studies. Web-based interventions are increasingly used during the pandemic, allowing for affordable access to psychological therapies. However, a lagging infrastructure and skill base has limited the application of digital solutions in mental health care. Methodologies need to be converged owing to the rapid development of digital technologies that have outpaced the evaluation of rigorous digital mental health interventions and strategies to prevent mental illness. The functions and implications of human-computer interaction require a better understanding to overcome engagement barriers, especially with predictive technologies. Explainable artificial intelligence is being incorporated into digital mental health implementation to obtain positive and responsible outcomes. Investment in digital platforms and associated apps for real-time screening, tracking, and treatment offer the promise of cost-effectiveness in vulnerable populations. Although machine learning has been limited by study conduct and reporting methods, the increasing use of unstructured data has strengthened its potential. Early evidence suggests that the advantages outweigh the disadvantages of incrementing such technology. The limitations of an evidence-based approach require better integration of decision support tools to guide policymakers with digital mental health implementation. There is a complex range of issues with effectiveness, equity, access, and ethics (eg, privacy, confidentiality, fairness, transparency, reproducibility, and accountability), which warrant resolution. Evidence-informed policies, development of eminent digital products and services, and skills to use and maintain these solutions are required. Studies need to focus on developing digital platforms with explainable artificial intelligence–based apps to enhance resilience and guide the treatment decisions of mental health practitioners. Investments in digital mental health should ensure their safety and workability. End users should encourage the use of innovative methods to encourage developers to effectively evaluate their products and services and to render them a worthwhile investment. Technology-enabled services in a hybrid model of care are most likely to be effective (eg, specialists using these services among vulnerable, at-risk populations but not severe cases of mental ill health).

show abstract

Combining structured and unstructured data for predictive models: a deep learning approach

Cited by 142 publications

References 22 publications

Using Deep Learning to Identify High-Risk Patients with Heart Failure with Reduced Ejection Fraction

Using Deep Learning to Identify High-Risk Patients with Heart Failure with Reduced Ejection Fraction

Application of Machine Learning in Intensive Care Unit (ICU) Settings Using MIMIC Dataset: Systematic Review

Digital Mental Health Challenges and the Horizon Ahead for Solutions

Contact Info

Product

Resources

About