× You are not currently logged in. To receive all the benefits our site has to offer, we encourage you to log in now.

Demystifying Patient Data: Using Medical Records in Healthcare Design Research

Patient-level medical records data has been a largely untapped resource in healthcare design research, despite its potential to help our field generate stronger evidence of the impact of design on patient outcomes. This webinar will highlight how patient data can be used in research, discuss the process of requesting and acquiring patient data, and explain how to conduct analyses to test design hypotheses.
Patient outcomes, such as length of stay and occurrence of adverse events, are of upmost importance in healthcare settings. Most hospital design studies have focused on nursing staff, and those that have investigated impact on patients have tended to use aggregated HCAHPS scores that do not get to the most important aspects of care at an individual patient or encounter level. Hospital medical records can bridge that gap and provide a rich data source for evaluation of actual patient outcomes.
Presenters will discuss how to engage healthcare organizations in design research using patient data, while complying with the HIPAA Privacy Rule. Topics include definitions of key terms, data use agreements, coordinating a data request and defining the parameters of a limited data set. Assessment of risk and IRB review will also be covered. Research coordination and logistical issues from a recent hospital study will be used to illustrate the process of working with clients in a real-world setting. Then, using length of stay (LOS) as an outcome of interest, presenters will walk through the steps of an analysis comparing patient data before and after a design intervention.
By applying an appropriate statistical model and controlling for patient characteristics and condition, such as demographics, socioeconomic status, acuity, and comorbidities, any remaining difference in LOS can be at least partially attributed to the difference in care environments. Presenters will explain how sample sizes are determined to detect a significant effect, the steps taken to designate independent variables to include for the best model fit, and the assumptions needed to form a valid statistical analysis. Data from a recent hospital study will be used to demonstrate the analysis and subsequent findings.