Premium
Objective Assessment and Thematic Categorization of Patient‐audible Information in an Emergency Department
Author(s) -
Zhang Xiao C.,
Kobayashi Leo,
Berger Markus,
Reddy Pranav M.,
Chheng Darin B.,
Gorham Sara A.,
Pathania Shivany,
Stern Sarah P.,
Icaza Milson Elio,
Jay Gregory D.,
Baruch Jay M.
Publication year - 2015
Publication title -
academic emergency medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.221
H-Index - 124
eISSN - 1553-2712
pISSN - 1069-6563
DOI - 10.1111/acem.12762
Subject(s) - medicine , categorization , emergency department , thematic map , medical emergency , thematic analysis , nursing , artificial intelligence , qualitative research , cartography , social science , sociology , computer science , geography
Objectives The objective was to assess and categorize the understandable components of patient‐audible information (e.g., provider conversations) in emergency department ( ED ) care areas and to initiate a baseline ED soundscape assessment. Methods Investigators at an academic referral hospital accessed 21 deidentified transcripts of recordings made with binaural in‐ear microphones in patient rooms ( n = 10) and spaces adjacent to nurses' stations ( n = 11), during ED staff sign‐outs as part of an approved quality management process. Transcribed materials were classified by speaker (health care provider, patient/family/friend, or unknown). Using qualitative analysis software and predefined thematic categories, two investigators then independently coded each transcript by word, phrase, clause, and/or sentence for general content, patient information, and HIPAA ‐defined patient identifiers. Scheduled reviews were used to resolve any data coding discrepancies. Results Patient room recordings featured a median of 11 (interquartile range [ IQR ] = 2 to 33) understandable words per minute (wpm) over 16.2 ( IQR = 15.1 to 18.4) minutes; nurses' station recordings featured 74 ( IQR = 47 to 109) understandable wpm over 17.0 ( IQR = 15.4 to 20.3) minutes. Transcript content from patient room recordings was categorized as follows: clinical, 44.8% ( IQR = 17.7% to 62.2%); nonclinical, 0.0% ( IQR = 0.0% to 0.0%); inappropriate (provider), 0.0% ( IQR = 0.0% to 0.0%); and unknown, 6.0% ( IQR = 1.7% to 58.2%). Transcript content from nurses' stations was categorized as follows: clinical, 86.0% ( IQR = 68.7% to 94.7%); nonclinical, 1.2% ( IQR = 0.0% to 19.5%); inappropriate (provider), 0.1% ( IQR = 0.0% to 2.3%); and unknown, 1.3% ( IQR = 0.0% to 7.1%). Limited patient information was audible on patient room recordings. Audible patient information at nurses' stations was coded as follows (median words per sign‐out sample): general patient history, 116 ( IQR = 19 to 206); social history, 12 ( IQR = 4 to 19); physical examination, 39 ( IQR = 19 to 56); imaging results, 0 ( IQR = 0 to 21); laboratory results, 7 ( IQR = 0 to 22); other results, 0 ( IQR = 0 to 3); medical decision‐making, 39 ( IQR = 10 to 69); management (general), 118 ( IQR = 79 to 235); pain management, 4 ( IQR = 0 to 53); and disposition, 42 ( IQR = 22 to 60). Medians of 0 ( IQR = 0 to 0) and 3 ( IQR = 1 to 4) patient name identifiers were audible on in‐room and nurses' station sign‐out recordings, respectively. Conclusions Sound recordings in an ED setting captured audible and understandable provider discussions that included confidential, protected health information and discernible quantities of nonclinical content.