DNP Scholarly Project · Results & Discussion

Section 4: what each arm has to do

Both regulatory gates are clear and the observation and chart-audit arms have usable data. But every number currently written in Section 4 is mock, and the live data does not say what the draft says it says.

Prepared August 26, 2026 Observation n = 93 Retrospective n = 68 Education n = 0

Read this before writing anything

The ambient temperature association runs backward

Warmer rooms are associated with colder patients at emergence: Pearson r = −0.221, slope −0.042 °C per °F, p ≈ 0.03. The current draft claims +0.11 °C per °F. Ambient explains under 5% of the variance in emergence core temperature.

The most likely reason is confounding by indication: providers turn the room up for cases they already expect to run cold. All three 76 °F cases were general surgery emerging at 35.7, 35.4, and 35.9. The room was likely warmed because the patient was cold, which reverses the causal arrow.

No arm may claim that warmer rooms produce warmer patients. What survives, and is genuinely the project's contribution, is that ambient OR temperature sits below every published minimum in 79.6% of cases and appears nowhere in the anesthesia record.

What all three of you must do

These apply to every arm. Settle them together before anyone drafts, so the three papers do not contradict each other in front of the same committee.

Replace every number.Nothing currently in any Section 4 came from live data. Treat the existing files as structural templates only.
Say Evidence-Based Practice, not research.Dr. Breazeale's July 17 determination classified the project as EBP and requires that any publication describe it that way. Audit every use of "study," "research," and "subjects."
Include a threshold sensitivity analysis.A ±0.1 °C shift in the cut point swings incidence by roughly 9 points in both arms, and that is exactly the Philips monitor's stated accuracy. Roughly one case in six cannot be classified either way.
Report core temperature as continuous, with the 36.0 °C rate as secondary.This single change dissolves three separate dead ends: the one severe case, the absent warming comparison group, and the low event count that made the logistic regression untenable.

Ameer — Observation arm

Ameer Shaw

Prospective observation 7 cases short

n = 93 of ~100 target · hypothermia 29.0% · mean emergence 36.28 °C · ambient mean 66.2 °F

Rewrite the ambient section around the negative association.Report r = −0.221 honestly, then explain confounding by indication. This is your largest single rewrite and it inverts the draft's central claim.
Lead with the finding that survives.74 of 93 cases (79.6%) ran below the AORN 68 °F minimum, and 82.8% fell outside the ASHRAE 170 range. Ambient was undocumented in all 93. That is the contribution.
Delete the quasi-separation narrative.The draft describes 57 unwarmed cases forcing an undefined coefficient. You have 2 unwarmed cases. The problem it solves no longer exists.
Cut "active warming is necessary but not sufficient."91 of 93 cases were warmed. There is no comparison group, so the claim cannot be supported from this dataset.
Move the Chen SSI inference into limitations.OR 2.12 is the below-35 °C estimate and you have exactly one case below 35 °C. Report severe hypothermia as n = 1 and stop there.
Flip the benchmark comparison.The draft says your rate is twice the published 39–44%. It is 29.0%, which is below it. Explain why: you measured at emergence only, not at any point intraoperatively.
Fix three miscoded rows before analyzing.6/7 15:17 recorded 36.2 °C flagged hypothermic; both 6/19 rows recorded 35.4 and 35.9 flagged normothermic. Temperature is authoritative.
Soften the precision sentence in Section 3.It currently says ±0.1 °C is "well within the precision required" to classify against 36.0 °C. Your own threshold analysis shows that is false for one case in six.

Keagen — Retrospective arm

Keagen Ferrell-Clegg

Retrospective chart audit Target met

n = 68 of ~60 target · emergence hypothermia 23.5% · induction hypothermia 19/68

Make the trajectory the story, not the endpoint.You have preoperative, induction, and emergence temperatures on every case. Emergence normothermia alone is an insensitive endpoint, which the team established back in June. The three-point trajectory is what only your arm can show.
Report the eleven flip cases.Eleven cases were hypothermic at induction but normothermic at emergence. One landed on exactly 36.0 and was counted as a recovery. Two more cleared by 0.2 °C or less.
Reclassify the seven cases without forced air.Six are cardiac cases documented as using an underbody water warmer. That is a difference in warming modality, not an absence of warming. Reporting them as unwarmed would be wrong.
Run the threshold analysis on your own arm.42.6% of your cases sit at or below 36.2 °C, and a ±0.1 °C shift moves your rate from 20.6% to 29.4%.
State how you handled missing preoperative values.Five cases are recorded ND for preoperative temperature. Name the convention and apply it consistently.

Ryan — Education arm

Ryan Curley

Provider education, pre/post Blocked — no data

n = 0 responses · form live since Aug 21 · recruitment email not yet sent

Send the recruitment email. Nothing else can start.The Google Form has been live since August 21 with zero responses because no invitations have gone out. Every task below depends on this one.
Verify how the form writes pre and post to the sheet.Because pre-survey, PowerPoint, and post-survey are sequential sections of one submission, per-respondent pairing may be achievable in a single response row. Confirm this before committing to a paired t-test. The anonymity design has no pairing code, so if it does not pair, switch to an unpaired comparison.
Use KR-20, not Cronbach's alpha.Your items are dichotomous. This has been flagged since June and is still uncorrected.
Fix the cover letter typo.Step 1 reads "Complete an 10-item." Small, but it is in an IRB-approved participant-facing document.
Write the analysis plan now, against the empty structure.You can draft everything except the numbers while recruitment runs. Do not wait for responses to start Section 4.

Mock versus real

What the observation Section 4 currently claims, against what the data shows. Struck values appear in the draft today.

Section_4_Results_Discussion_OBSERVATION.docx
Metric	In the draft	Actual
Cases	500	93
Emergence hypothermia	85.2%	29.0%
Severe hypothermia	~20%	1 case
Mean emergence core	35.41 °C	36.28 °C
Ambient mean	62.6 °F	66.2 °F
Ambient range	58.3–67.8 °F	60–76 °F
Active warming used	88.6%	97.8%
Unwarmed cases	57	2
Ambient–core slope	+0.11 °C/°F	−0.042 °C/°F
Variance explained	R² = 0.52	R² = 0.049

The paragraph to build around

Drafted for the observation arm. Keagen can mirror the structure with his own figures.

Emergence hypothermia occurred in 29.0% of cases at the ASPAN 36.0 °C threshold. However, threshold sensitivity analysis demonstrates the estimate is unstable at that cut point, ranging from 24.7% to 34.4% across a ±0.1 °C shift, which is precisely the stated accuracy of the monitoring equipment. Nearly half of all cases (47.3%) fell at or below 36.2 °C. Dichotomous classification at 36.0 °C therefore understates the thermal burden observed, and continuous core temperature is reported as the primary outcome. Proposed opening, Results

All figures computed directly from the live Google Forms response sheets on August 26, 2026. Full tables, the threshold curve, and the flagged document edits are recorded in MEMORY.md in the DNP Project folder.