The committee reviewed the published literature to identify what is known about the relationship between patient safety and health information technology (health IT)-assisted care. In the preliminary analysis, the literature was searched, reviewed, analyzed, sorted into categories (see category definitions below), and summarized into a table. “Health IT-assisted care” means health care and services that incorporate and take advantage of health IT and health information exchanges for the purpose of improving the processes and outcomes of health care services. Health IT-assisted care includes care supported by and involving electronic health records (EHRs), clinical decision support, computerized provider order entry, health information exchange, patient engagement technologies, and other health information technologies used in clinical care.
Extensive search terms were used in four databases—Medline, EMBase, Web of Science, and Cochrane—yielding initial results of 2,868 articles, books, and other literature. A search strategy was developed for each database using terms and Medical Subject Headings focused in subject areas related to patient safety, medical informatics, and other related areas.1
1 Subject areas included quality of health care (accidental harm, adverse events, diagnostic errors, errors of omission or commission, injuries, medication errors, safety, and treatment outcomes) and medical informatics (human factors, informatics, system design, systems analysis, usability, and user–computer interface).
Strategy parameters included limiting the search to human subjects, the English language, and results published between January 2005 and November 2010 because the literature regarding health IT evolves rapidly and continually builds upon itself. Next, hand searches through the references were conducted and relevant studies were included. Primary and secondary research (e.g., meta-analyses, controlled trials) suggested by the public and the committee were also added to the search results. Other literature (e.g., editorials, commentaries) were excluded from the search results.
CATEGORIZATION AND ANALYSIS
Titles and abstracts of the articles produced from the search were rigorously reviewed to determine which studies met the inclusion criteria for study quality and relevance. Pairs of reviewers evaluated titles and abstracts of all studies within each subject area. Each reviewer independently assigned articles to one of three categories with Category 1 being the most relevant and Category 3 being the least relevant (see definitions below).
Category 1: Literature examining how health IT affects patient safety.
1a. Systematic reviews2
1b. Experimental studies
1c. Observational studies
Category 2: Literature describing efforts to improve quality of health care through implementation of one or more of the following components of health IT (e.g., system design, systems analysis, usability, user-computer interface, or human factors).
Category 3: Studies not related to patient safety and health IT.
Reviewers then compared their evaluations, and any disagreement was resolved through discussion. The full texts of articles determined to be Category 1 were retrieved, evaluated, and, if needed, recategorized.
2 A systematic review is defined as a scientific investigation that focuses on a specific question and that uses explicit, preplanned scientific methods to identify, select, assess, and summarize the findings of similar but separate studies. It may or may not include a quantitative synthesis of the results from separate studies (meta-analysis).
The search yielded 128 Category 1 articles. These articles were then placed in two tables—one table summarizing each of the systematic reviews (Table B-1)3 and the other table summarizing each of the experimental and observational studies (Table B-2).4 Within the table, the articles were then organized by the type of health IT component.5
An additional 479 articles were identified as Category 2, which inform broader parameters of patient safety and health IT. These parameters include efforts to improve quality of health care through the implementation of one or more of the following: system design, systems analysis, usability, user-computer interface, or human factors.
The remaining studies were classified as Category 3 studies, which did not meet the inclusion criteria and were not broadly considered in the literature review or table.
Although Category 1 articles published before January 2005 and Category 2 articles were not included in the literature review table, these articles were used throughout the committee’s deliberations as background. Additionally, articles that were not studies or reviews of studies, such as editorials and commentaries, were not included in the literature table but were still considered and used in the analysis of the report.
5 The majority of the studies focused on how individual components of health IT affect patient safety (e.g., alerts, bar-coding, clinical decision support, EHRs, electronic prescribing, patient engagement tools, smartpumps, surveillance tools, and other health IT–assisted care). Therefore, the tables were organized by the components that each article was studying. Articles not focusing on a specific component were placed at the beginning of each table and are labeled as “overview” articles.