Handbook of data analysis of electronic health records (EHR) using SAS software / Behrouz Ehsani-Moghaddam.

Format:

Book

Author/Creator:

Ehsani-Moghaddam, Behrouz, author.

Series:

Health Care in Transition

Language:

English

Subjects (All):

Medical records--Data processing--Handbooks, manuals, etc.

Medical records.

Physical Description:

1 online resource (182 pages)

Edition:

1st ed.

Place of Publication:

New York : Nova Science Publishers, [2023]

Summary:

"Electronic Health Records (EHR) are longitudinal data that are stored in a database that captures current and new patients at different points in time. Since EHR data come from multiple different vendors and open-source products, they can be messy, inconsistent, and often need to be harmonized and reformatted properly before they can provide real-world insights about patients using statistical techniques. This book goes beyond the general data manipulation, viewing the data analysis issues in a wider and more practical context. It covers all major steps of analysis of EHR data, even those instructions that cannot be taught in any classroom. The reader can have hands-on experience using the codes that are provided in the book and by utilizing the accompanying data that are available for free. This book is not restricted to one specific discipline but rather will be of interest to scientists working in any area where analyzing electronic public health data using SAS program is necessary. The material is aimed at the reader who are already familiar with applied statistics at an undergraduate level or higher"-- Provided by publisher.

Contents:

Intro

Contents

Preface

Part I: Basic Information about EHR Data and Data Manipulation

Chapter 1

EHR Observations

Abstract

1.1. EHR Definition

1.2. Structure of EHR Data

1.3. The Structure of a Typical EHR Data Quality and Security

1.3.1. EHR Data Quality

1.3.1.1. Accuracy and Precision

1.3.1.2. Coherence

1.3.1.3. Completeness and Comprehensiveness

1.3.1.4. Consistency

1.3.1.5. Data Cleaning

1.3.1.6. Randomness

1.3.1.7. Timeliness

1.3.1.8. Uniqueness

1.3.2. EHR Security

1.4. Types of Variables

1.5. Record Linkage

References

Chapter 2

Data Transfer from Database Management Systems to SAS

2.1. SQL Server

2.1.1. How to Access a SQL Server Database from SAS

Program 2.1. Creation of Libraries to Extract SQL Tables

2.1.2. Importing Individual EHR Tables

Program 2.2. Creation of a Library to Extract SQL Tables

Program 2.3. Importing Individual EHR Tables

2.2. Python

Program 2.4. Creating CSV File from Vaccine Data in Python

Program 2.5. Importing CSV File Created in Python into SAS

2.3. Oracle Healthcare Repository

Program 2.6. Importing Oracle Tables into SAS

Chapter 3

Creating Temporary and Permanent Data Sets

Introduction

3.1. Making Temporary and Permanent Data Sets

Program 3.1: Example of Creating a Data Set by Using DATA Step and INPUT Statement

Program 3.2: Making Data Sets by FILENAME, LIBNAME or Another Data Set

3.2. Exploring Your Data Set

3.2.1. SAS Explorer

3.2.2. CONTENTS Procedure

Program 3.3: Contents View Using PROC CONTENTS and PROC DATASETS

3.2.3. PRINT Procedure

Program 3.4: Data Exploration Using PROC PRINT

3.2.4. FREQ Procedure

Program 3.5: Data Exploration Using PROC FREQ

3.2.5. MEANS/SUMMARY Procedure.

Program 3.6: Data Exploration Using PROC MEANS and PROC SUMMARY

3.2.6. UNIVARIATE Procedure

Program 3.7: Data Exploration Using PROC UNIVARIATE

3.3. Creating a Subset of Data

3.3.1. KEEP and DROP Statements

Program 3.7: Subsetting Data Using KEEP and DROP

3.3.2. Subsetting Data with PROC SQL

Program 3.8: Subsetting Data Using PROC SQL

3.3.3. Subsetting Using IF/WHERE/IF…THEN DELETE

Program 3.9: Subsetting Data Using IF/WHERE/IF… THEN DELETE

3.3.4. Subsetting Data Using CONTAINS, LIKE, FIND, INDEX and SUBSTR Functions

Program 3.10: Subsetting Data Using CONTAINS, LIKE, FIND, INDEX and SUBSTR Functions

3.4. Exporting Data from SAS to Other Programs

Program 3.11: Exporting SAS Data Sets to Other Programs

Chapter 4

Retrieving Patient Information

4.1. Combining EHR Data Sets

4.1.1. Concatenation

Program 4.1: Example of Creating Data Set by Using a DATA Step and SET Statement

Program 4.2: Example of Creating Data Set Using BY and Interleaving Method

4.1.2. Match-Merging

Program 4.3: Example of Creating Data Sets Using Match-Merging Technique

4.2. Creating New Variables

4.2.1. Creating New Variables Using LENGTH or ATTRIB Statements

Program 4.4: Example of Creating Variable Using LENGTH or ATTRIB Statements

4.2.2. Creating New Variables from Existing Variables

Program 4.5: Example of Creating Variables Using Existing Variables

4.3. Removing Duplicate or Unnecessary Records

4.3.1. Example 1

Program 4.6: Example of Removing Duplicate or Unwanted Observations

4.3.2. Example 2

Program 4.7: Example of Removing Duplicate Disease Data

4.3.3. Example 3

Program 4.8: Example of Removing Unwanted Repeated Weight Values

4.4. Changing Date Format

Program 4.9: Changing Date Format

4.5. Estimation of Patient Age

Program 4.10: Estimation of Patient Age.

4.6. Conversion of Variable Type to Numeric/Character

Program 4.11: Conversion of Variable Type to Numeric/Character

4.7. Importance of Patient Encounter Date

4.7.1. Investigation on the Completeness of Encounter Date

Program 4.12: Finding Patients without Encounter Date

4.7.2. Making an Inclusive Encounter Date

Program 4.13: Making an Inclusive Encounter Date

Part II: Analysis of Longitudinal EHR Data

Chapter 5

Data Extraction from Text and Analysis: Adverse Events Following Immunization

5.1. Objectives

5.2. Methodology

5.2.1. Changing Lower Case to Upper Case and Vice Versa

Program 5.1: Changing Character Variable Data from Lowercase to Uppercase or Vice Versa

5.2.2. Parsing the Character String

5.2.2.1. Step 1

5.2.2.2. Step 2

Program 5.2: Conversion of Text from Adverse Effects of Vaccination to Numeric Variables

5.3. Analysis and Results

Program 5.3: Frequency and Distribution of the Most Common Adverse Effects of Vaccination

Program 5.4: The Risk of Adverse Effects Related to the Age and Sex of the Vaccine Recipient

Chapter 6

Prevalence Estimation for Acute Diseases (A Cross-Sectional Cohort Study)

6.1. Objectives

6.2. Methodology

6.2.1. Case Definition

6.2.2. Crude Prevalence Estimation

6.2.3. Age-Sex Adjustment (Standardization)

6.3. Analysis and Results

6.3.1. Objective 1: Crude Prevalence Estimation

6.3.1.1. Step 1

Program 6.1: Creating Subset Data (Denominator) Using Match-Merge Technique

6.3.1.2. Step 2

Program 6.2: Adding Billing, Encounter_Diagnosis and Health_Condition Data to Patient_encounter Data Set

6.3.1.3. Step 3

Program 6.3: Creating a HZ Variable and Estimation of Patient Age at Onset of the Disease

6.3.1.4. Step 4.

Program 6.4: Removing Duplicate HZ Observations for Each Patient

6.3.1.5. Step 5

Program 6.5: Descriptive Statistics and Prevalence Estimation for HZ by Sex and Age

6.3.2. Objective 2: Age-Sex Standardization

Program 6.6: Descriptive Statistics and Prevalence Estimation for HZ by Sex and Age

Chapter 7

Prevalence Estimation for Chronic Diseases

7.1. Objectives

7.2. Methodology

7.2.1. Case Definition

7.2.2. Crude Prevalence Estimation

7.3. Analysis and Results

7.3.1. Objective 1: Crude Prevalence Estimation

7.3.1.1. Step 1

Program 7.1: Creating a Subset Data Using Match-Merge Technique for Patients who Had at Least One Visit from Jan. 1st, 2019 to Dec. 31st 2020

7.3.1.2. Step 2

Program 7.2: Creation of COPD Data Set and Exclusion of Asthma Patients

7.3.1.3. Step 3

Program 7.3: Creation of Medication Data Set for COPD Patients

7.3.1.4. Step 4

Program 7.4: Estimation of Patient Age at Onset of COPD Disease

7.3.1.5. Step 5

Program 7.5: Descriptive Statistics for Age and Prevalence Estimation for COPD Patients

7.3.2. Objective 2: Number of Healthcare Visits

Program 7.6: Number of Visit Estimation for COPD Patients

Chapter 8

Disease Case Validation

8.1. Objectives

8.2. Methodology

8.2.1. Precision Metrics

8.3. Analysis and Results

8.3.1. Objective: Cross-validation for COPD Classified Cases

Program 8.1: Cross-validation of COPD Data with Gold Standard

Program 8.2: False Positive and False Negative Probabilities Resulting from Cross-validation of COPD Data with Gold Standard

8.3.2. Sample Size Calculation

Program 8.3: Sample Size for Two Groups of COPD Patients Using the ANOVA Method

Chapter 9

Multiple Logistic Regression

9.1. Objectives

9.2. Methodology.

9.2.1. Data Set

9.3. Analysis and Results

9.3.1. MLR Model Fit Using PROC HPGENSELECT

Program 9.1: Variable Selection and Model Fitting for Infant Mortality Data Using PROC HPGENSELECT

Program 9.2: Variable Selection and Model Fitting for Infant Mortality Data Using LASSO Technique

Program 9.3: Merging Birthwgt and Out Data Sets for Estimationof Probability of a Specific Observation

Chapter 10

Machine Learning for Medical Diagnoses

10.1. Objectives

10.2. Methodology

10.2.1. Data Set

10.2.2. Creating Project, Library, and Data Source

10.2.2.1. Creating Project and Diagram

10.2.2.2. Saving the Created SAS Table

10.2.2.3. Data Source

10.2.3. Creating a Flow Diagram

10.2.4. Data Exploration

10.2.5. Imputation and Transformation

10.2.6. Variable Selection

10.2.7. Multicollinearity

10.2.8. Data Partitioning

10.2.9. Building and Assessing Models

10.2.10. Model Comparison

10.2.11. Scoring a Data Set Using the Selected Model

10.2.12. Estimation of Precision Metrics

Program 10.1: Estimation of Precision Metrics for ANN Model

Conclusion

Index

About the Author

Blank Page.

Notes:

Includes bibliographical references and index.

Description based on print version record.

Other Format:

Print version: Ehsani-Moghaddam, Behrouz Handbook of Data Analysis of Electronic Health Records (EHR) Using SAS Software

ISBN:

9798886974379