Secondary Analysis of Electronic Health Records.

• Author: MIT Critical Data, M. I. T. Critical.
• Place / Publishing House: Cham : : Springer International Publishing AG,, 2016.; ©2016.
• Year of Publication: 2016
• Edition: 1st ed.
• Language: English
• Physical Description: 1 online resource (435 pages)

Table of Contents

Intro -- Preface -- MIT Critical Data -- Contents -- Setting the Stage: Rationale Behind and Challenges to Health Data Analysis -- Introduction -- 1 Objectives of the Secondary Analysis of Electronic Health Record Data -- 1.1 Introduction -- 1.2 Current Research Climate -- 1.3 Power of the Electronic Health Record -- 1.4 Pitfalls and Challenges -- 1.5 Conclusion -- References -- 2 Review of Clinical Databases -- 2.1 Introduction -- 2.2 Background -- 2.3 The Medical Information Mart for Intensive Care (MIMIC) Database -- 2.3.1 Included Variables -- 2.3.2 Access and Interface -- 2.4 PCORnet -- 2.4.1 Included Variables -- 2.4.2 Access and Interface -- 2.5 Open NHS -- 2.5.1 Included Variables -- 2.5.2 Access and Interface -- 2.6 Other Ongoing Research -- 2.6.1 eICU-Philips -- 2.6.2 VistA -- 2.6.3 NSQUIP -- References -- 3 Challenges and Opportunities in Secondary Analyses of Electronic Health Record Data -- 3.1 Introduction -- 3.2 Challenges in Secondary Analysis of Electronic Health Records Data -- 3.3 Opportunities in Secondary Analysis of Electronic Health Records Data -- 3.4 Secondary EHR Analyses as Alternatives to Randomized Controlled Clinical Trials -- 3.5 Demonstrating the Power of Secondary EHR Analysis: Examples in Pharmacovigilance and Clinical Care -- 3.6 A New Paradigm for Supporting Evidence-Based Practice and Ethical Considerations -- References -- 4 Pulling It All Together: Envisioning a Data-Driven, Ideal Care System -- 4.1 Use Case Examples Based on Unavoidable Medical Heterogeneity -- 4.2 Clinical Workflow, Documentation, and Decisions -- 4.3 Levels of Precision and Personalization -- 4.4 Coordination, Communication, and Guidance Through the Clinical Labyrinth -- 4.5 Safety and Quality in an ICS -- 4.6 Conclusion -- References -- 5 The Story of MIMIC -- 5.1 The Vision -- 5.2 Data Acquisition -- 5.2.1 Clinical Data.

5.2.2 Physiological Data -- 5.2.3 Death Data -- 5.3 Data Merger and Organization -- 5.4 Data Sharing -- 5.5 Updating -- 5.6 Support -- 5.7 Lessons Learned -- 5.8 Future Directions -- Acknowledgments -- References -- 6 Integrating Non-clinical Data with EHRs -- 6.1 Introduction -- 6.2 Non-clinical Factors and Determinants of Health -- 6.3 Increasing Data Availability -- 6.4 Integration, Application and Calibration -- 6.5 A Well-Connected Empowerment -- 6.6 Conclusion -- References -- 7 Using EHR to Conduct Outcome and Health Services Research -- 7.1 Introduction -- 7.2 The Rise of EHRs in Health Services Research -- 7.2.1 The EHR in Outcomes and Observational Studies -- 7.2.2 The EHR as Tool to Facilitate Patient Enrollment in Prospective Trials -- 7.2.3 The EHR as Tool to Study and Improve Patient Outcomes -- 7.3 How to Avoid Common Pitfalls When Using EHR to Do Health Services Research -- 7.3.1 Step 1: Recognize the Fallibility of the EHR -- 7.3.2 Step 2: Understand Confounding, Bias, and Missing Data When Using the EHR for Research -- 7.4 Future Directions for the EHR and Health Services Research -- 7.4.1 Ensuring Adequate Patient Privacy Protection -- 7.5 Multidimensional Collaborations -- 7.6 Conclusion -- References -- 8 Residual Confounding Lurking in Big Data: A Source of Error -- 8.1 Introduction -- 8.2 Confounding Variables in Big Data -- 8.2.1 The Obesity Paradox -- 8.2.2 Selection Bias -- 8.2.3 Uncertain Pathophysiology -- 8.3 Conclusion -- References -- A Cookbook: From Research Question Formulation to Validation of Findings -- 9 Formulating the Research Question -- 9.1 Introduction -- 9.2 The Clinical Scenario: Impact of Indwelling Arterial Catheters -- 9.3 Turning Clinical Questions into Research Questions -- 9.3.1 Study Sample -- 9.3.2 Exposure -- 9.3.3 Outcome -- 9.4 Matching Study Design to the Research Question.

9.5 Types of Observational Research -- 9.6 Choosing the Right Database -- 9.7 Putting It Together -- References -- 10 Defining the Patient Cohort -- 10.1 Introduction -- 10.2 PART 1-Theoretical Concepts -- 10.2.1 Exposure and Outcome of Interest -- 10.2.2 Comparison Group -- 10.2.3 Building the Study Cohort -- 10.2.4 Hidden Exposures -- 10.2.5 Data Visualization -- 10.2.6 Study Cohort Fidelity -- 10.3 PART 2-Case Study: Cohort Selection -- References -- 11 Data Preparation -- 11.1 Introduction -- 11.2 Part 1-Theoretical Concepts -- 11.2.1 Categories of Hospital Data -- 11.2.2 Context and Collaboration -- 11.2.3 Quantitative and Qualitative Data -- 11.2.4 Data Files and Databases -- 11.2.5 Reproducibility -- 11.3 Part 2-Practical Examples of Data Preparation -- 11.3.1 MIMIC Tables -- 11.3.2 SQL Basics -- 11.3.3 Joins -- 11.3.4 Ranking Across Rows Using a Window Function -- 11.3.5 Making Queries More Manageable Using WITH -- References -- 12 Data Pre-processing -- 12.1 Introduction -- 12.2 Part 1-Theoretical Concepts -- 12.2.1 Data Cleaning -- 12.2.2 Data Integration -- 12.2.3 Data Transformation -- 12.2.4 Data Reduction -- 12.3 PART 2-Examples of Data Pre-processing in R -- 12.3.1 R-The Basics -- 12.3.2 Data Integration -- 12.3.3 Data Transformation -- 12.3.4 Data Reduction -- 12.4 Conclusion -- References -- 13 Missing Data -- 13.1 Introduction -- 13.2 Part 1-Theoretical Concepts -- 13.2.1 Types of Missingness -- 13.2.2 Proportion of Missing Data -- 13.2.3 Dealing with Missing Data -- Available-Case Analysis -- Weighting-Case Analysis -- Mean and Median -- Linear Interpolation -- Hot Deck and Cold Deck -- Last Observation Carried Forward -- Linear Regression -- Stochastic Regression -- Multiple-Value Imputation -- K-Nearest Neighbors -- 13.2.4 Choice of the Best Imputation Method -- 13.3 Part 2-Case Study.

13.3.1 Proportion of Missing Data and Possible Reasons for Missingness -- 13.3.2 Univariate Missingness Analysis -- Linear Regression Imputation -- Stochastic Linear Regression Imputation -- 13.3.3 Evaluating the Performance of Imputation Methods on Mortality Prediction -- 13.4 Conclusion -- References -- 14 Noise Versus Outliers -- 14.1 Introduction -- 14.2 Part 1-Theoretical Concepts -- 14.3 Statistical Methods -- 14.3.1 Tukey's Method -- 14.3.2 Z-Score -- 14.3.3 Modified Z-Score -- 14.3.4 Interquartile Range with Log-Normal Distribution -- 14.3.5 Ordinary and Studentized Residuals -- 14.3.6 Cook's Distance -- 14.3.7 Mahalanobis Distance -- 14.4 Proximity Based Models -- 14.4.1 k-Means -- 14.4.2 k-Medoids -- 14.4.3 Criteria for Outlier Detection -- 14.5 Supervised Outlier Detection -- 14.6 Outlier Analysis Using Expert Knowledge -- 14.7 Case Study: Identification of Outliers in the Indwelling Arterial Catheter (IAC) Study -- 14.8 Expert Knowledge Analysis -- 14.9 Univariate Analysis -- 14.10 Multivariable Analysis -- 14.11 Classification of Mortality in IAC and Non-IAC Patients -- 14.12 Conclusions and Summary -- Code Appendix -- References -- 15 Exploratory Data Analysis -- 15.1 Introduction -- 15.2 Part 1-Theoretical Concepts -- 15.2.1 Suggested EDA Techniques -- 15.2.2 Non-graphical EDA -- 15.2.3 Graphical EDA -- 15.3 Part 2-Case Study -- 15.3.1 Non-graphical EDA -- 15.3.2 Graphical EDA -- 15.4 Conclusion -- Code Appendix -- References -- 16 Data Analysis -- 16.1 Introduction to Data Analysis -- 16.1.1 Introduction -- 16.1.2 Identifying Data Types and Study Objectives -- 16.1.3 Case Study Data -- 16.2 Linear Regression -- 16.2.1 Section Goals -- 16.2.2 Introduction -- 16.2.3 Model Selection -- 16.2.4 Reporting and Interpreting Linear Regression -- 16.2.5 Caveats and Conclusions -- 16.3 Logistic Regression -- 16.3.1 Section Goals.

16.3.2 Introduction -- 16.3.3 2 × 2 Tables -- 16.3.4 Introducing Logistic Regression -- 16.3.5 Hypothesis Testing and Model Selection -- 16.3.6 Confidence Intervals -- 16.3.7 Prediction -- 16.3.8 Presenting and Interpreting Logistic Regression Analysis -- 16.3.9 Caveats and Conclusions -- 16.4 Survival Analysis -- 16.4.1 Section Goals -- 16.4.2 Introduction -- 16.4.3 Kaplan-Meier Survival Curves -- 16.4.4 Cox Proportional Hazards Models -- 16.4.5 Caveats and Conclusions -- 16.5 Case Study and Summary -- 16.5.1 Section Goals -- 16.5.2 Introduction -- 16.5.3 Logistic Regression Analysis -- 16.5.4 Conclusion and Summary -- References -- 17 Sensitivity Analysis and Model Validation -- 17.1 Introduction -- 17.2 Part 1-Theoretical Concepts -- 17.2.1 Bias and Variance -- 17.2.2 Common Evaluation Tools -- 17.2.3 Sensitivity Analysis -- 17.2.4 Validation -- 17.3 Case Study: Examples of Validation and Sensitivity Analysis -- 17.3.1 Analysis 1: Varying the Inclusion Criteria of Time to Mechanical Ventilation -- 17.3.2 Analysis 2: Changing the Caliper Level for Propensity Matching -- 17.3.3 Analysis 3: Hosmer-Lemeshow Test -- 17.3.4 Implications for a 'Failing' Model -- 17.4 Conclusion -- Code Appendix -- References -- Case Studies Using MIMIC -- Introduction -- 18 Trend Analysis: Evolution of Tidal Volume Over Time for Patients Receiving Invasive Mechanical Ventilation -- 18.1 Introduction -- 18.2 Study Dataset -- 18.3 Study Pre-processing -- 18.4 Study Methods -- 18.5 Study Analysis -- 18.6 Study Conclusions -- 18.7 Next Steps -- 18.8 Connections -- Code Appendix -- References -- 19 Instrumental Variable Analysis of Electronic Health Records -- 19.1 Introduction -- 19.2 Methods -- 19.2.1 Dataset -- 19.2.2 Methodology -- 19.2.3 Pre-processing -- 19.3 Results -- 19.4 Next Steps -- 19.5 Conclusions -- Code Appendix -- References.

20 Mortality Prediction in the ICU Based on MIMIC-II Results from the Super ICU Learner Algorithm (SICULA) Project.