Table 2

Major Methodological Considerations in This Literature

Design FeatureQuestions to Ask
Measurement of staffingDo the staffing measures reflect the type and “dose” of staff actually caring for the patients being studied?

Were the staffing measures collected in a consistent manner (using common definitions) across the organizational units/time periods?
Measurement of outcomesWere outcomes assessed in comparable ways across patients and across settings (units or institutions or time periods)?

Do data sources allow a distinction between complicating conditions present when care was undertaken (which should be considered in the analyses in risk adjustment (below)) from conditions that appeared during care (that are potentially outcomes of nursing care during the hospitalization)?

Were outcomes assessed completely/comprehensively for all patients? What evidence is there regarding the consistency of documentation for the outcomes in question in the data sources?

Does the outcome in question have a plausible association with nursing practice, or is it primarily/entirely associated with factors outside the control of providers?
Risk adjustmentHave the authors conducted fair comparisons between rates of adverse events across hospitals units or time periods by considering potentially important differences in the patients treated across those settings and/or over time?
Data linkageTo what extent do staffing measures represent conditions at the times and places where nursing care affecting the outcomes and measured for this study is given?

Are outcomes attributed to the locations of care where nursing services actually influence the outcome, or do they also reflect the place where detection of the outcome occurs?
Control for confounding factorsHave other aspects of the environments in which patients are cared for that might affect the outcomes been measured and analyzed?

E.g., availability of equipment/supplies, quality of physician care, other types of facility personnel, hospital size, academic affiliation, rural-urban location
Statistical modelingIf the study examines an outcome that is rare in the patient population, has this been considered in any modeling? How is skewness of the data managed?

If the subjects of the study are grouped or nested within larger organizational units (e.g., patients within nursing units within hospitals), has this been handled by the analytic strategy?

Do at least some of the analyses presented depict the complexity of associations between the factors involved through some type of statistical modeling that evaluates impacts of variables simultaneously?

From: Chapter 25, Nurse Staffing and Patient Care Quality and Safety

Cover of Patient Safety and Quality
Patient Safety and Quality: An Evidence-Based Handbook for Nurses.
Hughes RG, editor.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.