NOTE-1073: Jensen Hughes — Post-incident analysis of the Brookhollow Apartment fire

Contribution

A consultant’s post-incident analysis of a fatal apartment fire in Wichita, covering Sedgwick County Emergency Communications (SCEC), the Wichita Fire Department (WFD), Sedgwick County EMS and the Wichita Police Department, from the first 911 call (3:58 a.m.) to the last engine leaving (11:05 a.m.). It reconstructs the timeline from 911 audio, six radio channels, CAD records and interviews, compares each agency with its own policies and with APCO, NENA and NFPA standards, and makes 30 numbered findings with recommendations. Its distinctive result is that most failures were departures from procedures the agencies already had.

Key insight

The written system was sound and the practised one was not. SCEC’s procedures were “thorough, well-organized, and up-to-date” (p. 31) and the WFD Operations Manual “thorough, well-organized, and up-to-date relative to most current practices” (p. 47); yet the dispatchers on duty could not make the new CAD recommend a second alarm, the TAC channel was unmonitored by an informal agreement contrary to SCEC policy, command was established late and did not control the fireground, and the mayday was handled against the OM’s own mayday section. WFD firefighters, asked what they could have done better, said “there was no room for improvement” (p. C-5).

Assumptions

• The reviewers relied on data the agencies provided, and say some was withheld: the WFD Fire Investigation Unit kept back its documentation because its investigation was open, so the fire’s origin, development and the victim’s last minutes could not be fully reconstructed.
• All times are “approximations” aligned across 911 calls, radio and CAD; the radio traffic in Appendix A is paraphrased, “not verbatim”.
• The yardsticks are each agency’s own policies, APCO 1.113.2, NENA, and NFPA 1710, 1561, 1407, 1521, 921 and 1033.
• Names of individuals other than Ms. Bedeski are withheld.

Key results

• Overall (pp. 7, 81–82). “we found no single point of failure by any of the agencies involved that could be solely attributed to the fatality”. “Jensen Hughes identified a number of systemic shortcomings that, when combined, contributed to an unorganized rescue effort.” She stopped speaking 48 seconds into her call, before WFD was dispatched, and “Jensen Hughes cannot conclude that the shortcomings identified would have absolutely changed the ultimate outcome of this fatal incident” (p. 8).
• Dispatch. Her call was handled within APCO benchmarks (Table 1). The call taker likely could not hear her unit number, but “did not use the instant recall feature to replay the audio” (p. 8); recorded audio is clearer than what a dispatcher hears, and dual headsets impair comprehension. The second alarm, requested twice, was dispatched about eighteen minutes later because dispatchers and the supervisor could not get the new CAD to recommend units; afterwards “multiple dispatchers on multiple shifts were not exactly able to recall how to dispatch a second alarm” (p. 42), though the academy covers it. The “man trapped in 302” message was broadcast on Ops but not on TAC as the Fire OM requires.
• Fireground. WFD arrived within NFPA 1710 times and had 41 personnel on scene before the second alarm, more than the standard requires. Three attack lines ran from one engine’s tank before a water supply was secured, and the engine ran out. A “Fire Attack” mode was declared, but crews worked in an undeclared “Fast Attack/Rescue” mode with no Rapid Intervention Crew. Six units arrived before command was established by the seventh. Crews searched adjacent buildings before Building 300; “a significant number of additional units arrived and freelanced without clear coordination”, which “may be an accepted cultural norm in the WFD” (p. 59). Radio discipline was poor; officers were “often unsure of the department policies” on channels, and some “acknowledged they were not following them constantly” (p. 60).
• The mayday. About seven minutes after arrival part of the stair roof collapsed on the attack crews (Engine 15 and Squad 15) and a mayday was called; command went defensive and ordered everyone out. Everyone was accounted for in about six and a half minutes and no firefighter was reported injured. The report calls a mayday “an emergency incident occurring within the larger context of an emergency incident” and requires “a dual focus: addressing the immediate emergency while maintaining situational awareness of the ongoing incident” (p. 61). WFD policy already said “During a mayday, no fire ground assignment should be abandoned” (p. 54). But “search and rescue of Building 300 was not reprioritized and reinitiated for over eleven (11) minutes after the Mayday was called, despite the Mayday being under control within a few minutes of its occurrence” (p. 55). Policy was departed from in channel discipline, the RIC Group Leader assignment, progress reports, the alarm upgrade, and the end: “nor did Command announce when the Mayday was complete per WFD policy” (p. 63). Most firefighters recalled the academy as their last formal multi-company mayday training.
• After the mayday. A primary search of Unit 306 found Ms. Bedeski in the bathtub about 3.5 minutes after search was resumed. No triage area was set up; EMS, not dispatched to fires unless someone is reported trapped, had to find patients.
• Investigation. The Fire Investigation Unit’s work did not follow NFPA 921 or its own manual; it had not issued an origin-and-cause report; the fire’s cause was “Undetermined” and still “Under Investigation”; and the FIU did not take part in the department’s After-Action Review.
• Underlying challenges (pp. 10, 81). Years of “Tensions and animosity” between SCEC and agencies it serves; “the lack of nearby comparable advancing agencies fosters complacency”; and “The fire department appears to lack an emphasis on self-reflection and accountability within its culture.”
• Other factors. Smoking materials discarded near the stair; the stair’s combustible construction; wind; a smoke alarm with a low battery weeks before; the door of Unit 306 possibly open; and her retreat to a windowless bathroom.

Claims

id	claim	strength	support
C1	No single agency failure caused the death; combined shortcomings produced an unorganized rescue, and whether they changed the outcome cannot be concluded	moderate	timeline reconstruction; her voice stopped before dispatch (pp. 7–8, 81–82)
C2	The dispatcher probably could not hear the unit number, and did not use instant recall that might have recovered it	moderate	interviews, audio review, console testing of headset audio (pp. 40–44)
C3	The second alarm was delayed about eighteen minutes because staff could not operate a CAD feature they had been trained on	strong	CAD records, radio, interviews, SCEC’s own post-incident check (pp. 41–42)
C4	Incident command did not establish effective command and control; companies freelanced	moderate to strong	radio traffic and interviews; freelancing as a norm is flagged as needing other incidents to confirm (pp. 58–60)
C5	After the mayday, search of the involved building was not resumed for over eleven minutes, contrary to WFD’s own mayday policy	strong	radio timeline (Appendix A, Tables 6, 10) and policy text (pp. 54–55, 63)
C6	The FIU investigation did not follow NFPA 921 or its own manual	moderate	documents provided; the FIU withheld others (pp. 65–75)
C7	WFD culture lacks self-reflection and accountability; SCEC–agency relations have been marked by years of distrust	weak to moderate	interviews, including WFD members’ “no room for improvement” (pp. 10, C-2–C-5)

Concepts

• Mayday: a firefighter in life-threatening distress; here an “incident within an incident” (p. 63).
• Fire Attack and Fast Attack/Rescue modes: WFD operational modes; the second relaxes firefighter safety requirements for time-critical rescues and is meant to be brief.
• IRIC / RIC: initial and full Rapid Intervention Crews, standing by to rescue firefighters.
• Freelancing: companies acting without reporting for assignment.
• TAC and Ops channels: the fireground tactical channel, where command talks to companies, and the operations channel to dispatch.
• Fire scene benchmarks: required radio reports to dispatch (water supply, primary search, under control).

Connections

The review was recommended unanimously by the Emergency Communications Advisory Board on 5 December 2023 (Appendix B). Its standards are NFPA’s and APCO’s; it does not cite NIMS by name, though its command analysis follows the Incident Command System’s terms. The county’s dashboard on the same page tracks the 40 recommendation items (the 30 findings, some with several recommendations); fetched on 2026-10-22, it showed 37 marked “Implemented” and 3 “In Progress”.

Bearing on the record

• CASE-135. Mostly supported, with corrections.
  • The quoted findings (“contributed to an unorganized rescue effort”, “cannot conclude that the shortcomings identified would have absolutely changed the ultimate outcome”, the tensions with emergency communications, the lack of “self-reflection and accountability”) are in the report verbatim.
  • The exchange’s details marked unverified are now checkable. The report does call a mayday an “incident within an incident” and “an emergency incident occurring within the larger context of an emergency incident”; the exchange’s “an emergency arising within another emergency” is a paraphrase, not a quotation. The mayday and the original incident were not managed together: search was not resumed for over eleven minutes. “Command did not … announce when the Mayday was complete” is in the report. The later mayday procedure is not in the report (it predates it), but the county’s dashboard records that by 24 October 2025 WFD adopted a General Orders chapter including “COMMAND PROCEDURES - MAYDAY”, exercised in multi-agency training.
  • “Four firefighters caught in a stairwell collapse” is not in the report. It says portions of the stair roof “collapsed on Engine 15 and Squad 15” (staffed by three and two), and that “There were no reported firefighters injured”. The number four comes from press coverage and should be marked so, or dropped.
  • The report’s title page reads “Post-Incident Analysis” over “Independent Review of the Brookhollow Apartment Fire”. It gives the address as 8165 E. Central in the body and 8650 W. Central in the cover letter, and dates the FIU’s arrival “October 13, 2024” once; these are the report’s errors.
• CLAIM-027 and QUESTION-078 (which regime is effective; acting as one without synchronizing). The case’s reading, two regimes demanding authority at once, is the report’s own: “A Mayday requires a dual focus” and “should interrupt the primary incident operations as little as possible”. What the report adds is that the composition failed by default to suspension: command suspended search entirely and did not restore it for eleven minutes after the mayday was under control.
• TERM-017 and “release”. The case notes that TERM-017 names suspension and not release. The report supports the gap concretely: the mayday regime was never formally ended (“nor did Command announce when the Mayday was complete”), and the suspended regime was resumed late. Release has its own procedure in WFD policy, and it was the step skipped.
• CLAIM-422 (act, procedure, maintained capacity). This report is a stronger case for the claim than the case presently makes it. The procedures existed before the fire: WFD’s mayday section, its rule that second engines secure water, SCEC’s TAC monitoring rule and its academy training on second alarms. The capacities were not maintained: dispatchers on several shifts could not recall the CAD step; the TAC rule had been set aside by informal agreement; multi-company mayday training had lapsed since the academy. So the record of a procedure was not evidence of the capacity even before any correction was made. The county’s dashboard is the claim’s third object in miniature: it records 37 of 40 recommendations “Implemented”, including, for Recommendation 16, updates that began “The policy is under review”. An implemented recommendation is a procedure adopted; whether the capacity is maintained is not something the dashboard measures.
• CASE-135 and A23’s caution that a review “may recommend reform without demonstrating that the institution actually changed”. Supported in spirit, and the report itself recommends the measurement that would tell: WFD “should examine a sample of recent fire incidents and consider whether Incident Command exercised effective command and control” (Finding 13).
• CLAIM-416. WFD met every NFPA 1710 response-time benchmark and staffed above the standard, and the rescue was still unorganized: a test passed (arrival and staffing) is not the capacity the test is meant to indicate.
• NIMS (LIT-1365). The freelancing and self-dispatch the report describes (Engine 4, Rescue 2 and Engine 10 self-dispatched; the FIU self-dispatched) depart from the NIMS principle that resources not requested “should refrain from spontaneous deployment” (NOTE-1072). The report does not make this comparison.
• No THEORY filed: one incident, analysed against its own policies.
• No instruction for machine-learning practice.

Limitations

• One incident, analysed by consultants for the commissioning governments; the fire department’s investigators withheld material.
• Times are approximate and radio quotations paraphrased.
• Cultural findings rest on interviews, and the report itself says freelancing as a norm would need other incidents to confirm.
• It does not establish what would have saved Ms. Bedeski, and says so.

Open questions

• Did the recommended monthly incident reviews (adopted, per the county’s dashboard, in 2025) change command performance at later fires? That is the maintained-capacity question CLAIM-422 asks, and only later incidents can answer it.
• Was the eleven-minute pause a failure of training, of the mayday procedure’s wording, or of command attention? The report recommends training without separating them.