Estimating the Frequency of Inpatient Adverse Events Using a 2-Step Retrospective Chart Review: Initial Observational Cohort Study

TL;DR

This study tested two methods to detect inpatient adverse events in a hospital in a lower- to middle-income country, finding one method more effective.

Contribution

The study adapts and evaluates a methodology for detecting adverse events in resource-limited healthcare settings.

Findings

Tool 1B identified 30 adverse events per 100 admissions, compared to 1 event using tool 1A.

60% of the adverse events identified by tool 1B were considered preventable.

Tool 1B was found to be more sensitive and suitable for use in lower- to middle-income countries.

Abstract

Lower- and middle-income countries bear a disproportionate share of the global burden of adverse events in health care. Despite this, patient safety research is predominantly conducted in high-income countries with well-developed health care systems, resulting in evidence and methodologies that have limited applicability in resource-constrained settings.

This pilot study primarily aimed to identify the most suitable methodology for a full-scale study to detect inpatient adverse events at a tertiary care hospital in a lower- to middle-income country. Second, we aimed to use our experience with this study to further adapt the selected methodology to our setting.

This external pilot study used a 2-step retrospective chart review methodology. Two separate screening tools, tools 1A and 1B, were compared to identify which performed better in our setting. We reviewed the medical records of patients who were discharged between January 1, 2019, and December 31, 2019, from a tertiary care hospital in a lower- to middle-income country in South Asia. The main outcome of interest was the rate of adverse events among hospitalized patients reported as the total number of adverse events experienced per 100 admissions.

A total of 100 medical records were screened using tool 1A, with the mean patient age being 39.2 (SD 27.7) years and the mean length of stay being 3.3 (SD 2.8) days. Only 1 adverse event was identified using tool 1A, resulting in an adverse event rate of 1 event per 100 hospital admissions. Tool 1B was also used to screen a total of 100 medical records. The mean patient age was found to be 39.8 (SD 28.4) years, with the mean length of stay being 3.5 (SD 3.4) days. A total of 30 adverse events were identified across 22 patient files, with 18 (60%) considered preventable, resulting in an adverse event rate of 30 events per 100 hospital admissions.

This study demonstrates that tool 1B, adapted from the Global Trigger Tool for Measuring Adverse Events, represents an appropriate and sensitive methodology to identify adverse events among hospitalized patients in a lower- to middle-income country. Furthermore, the findings and experiences of this study were used to improve the design and procedures of our research methodology before implementation in a full-scale study.

RR2-10.1136/bmjopen-2023-076971