Living with NPWT

• Smell is dose-responsive and consistent across three of four scales. Faint, moderate and strong odour give Life impact −3.45, −7.71, −11.32; Psychological −5.40, −7.20, −11.50; Social −6.58, −9.36, −11.36. All significant.
• Drainage is significant on three scales: Life impact −5.65, Psychological −3.43, Social −5.50.
• Any comorbidity other than diabetes or peripheral vascular disease is significant on all four scales.
• Diabetes is significant on Life impact only (−3.90).
• Five or more wounds is significant on Life impact (−6.78) and Social (−9.33).
• The Social scale has a 24.4% ceiling effect (306 of 1,256), which limits its ability to detect improvement.

CAT: OCEBM 4 · This is a cross-sectional survey and the authors state plainly in their Limitations that 'the results of our research can only imply associations'. Three features break any causal reading of the vacuum coefficient and they should be recorded in this order. First, no adjustment for wound severity. The models adjust for wound size in square centimetres, duration in seven bands, aetiology in fourteen categories, number, location, drainage and smell, plus diabetes, peripheral vascular disease and other comorbidity - but for NO depth, NO exposed bone or tendon, NO Wagner, University of Texas or WIfI grade, NO infection status, NO ankle-brachial index or perfusion measure, NO inpatient status, NO pain and NO healing trajectory. Sicker wounds get NPWT, and that residual severity is what the coefficient is most likely carrying. Wound size is a two-dimensional area proxy with a median of 4 square centimetres and a range to 1,282, entered untransformed and linearly. Second, the exposure window does not overlap the outcome window. NPWT is measured as 'use of suction device within the past 3-6 months'; all four scales have a ONE-WEEK recall. The authors concede it - 'it would have been preferable if participants were asked if they were treated with NPWT in the past week' - so temporal order is not established. Third, recruitment stream is neither a covariate nor a stratifying variable, despite an 862 to 420 split with materially different wound states: drainage was present in 72.5% of the clinic stream against 48.6% of the online stream, recoverable by subtraction from the paper's own tables; wound characteristics were clinician-corroborated in three of four clinic countries and ENTIRELY SELF-REPORTED in the online stream and in the Danish clinic subgroup; and no country variable appears in any model either. Add that the life impact coefficient sits below the instrument's own measurement error - the paper quotes a smallest detectable change of 2.5 points on the Psychological scale, and the Life impact vacuum interval of -7.33 to -0.24 includes values far smaller than that; only the Social interval lies wholly beyond it - and that it is one of 173 unadjusted multivariable coefficient tests at p < 0.05, where roughly 8 or 9 false positives are expected by chance. The sleep findings are on different ground. Sleep interference is the dominant term in every model, with standardised betas of -0.30 to -0.57 and coefficients of -23.64 on Life impact and -29.55 on Social for the most disturbed category; These are the only coefficients in the paper whose entire confidence interval lies beyond the smallest detectable change. One caveat the paper does not name: the Sleep-scale model regresses a five-item scale about trouble falling and staying asleep on a single item asking how many nights the wound interfered with sleep, so its standardised beta of -0.57 is largely the same question answered twice.

Figures: Figures checked

Group means on a 1 to 4 scale, higher better. No p-value, confidence interval or standard deviation is given for any individual item.

The pattern is coherent and is the citable content: the machine is judged competent at its job and acceptable-looking, and worst on being physically active, sleeping, noise and enjoying life. Item 1 is the only item where the reconstruction group scored higher.

The free-text confirms it. Benefits: "I felt more supported with this device than in a bra."; "The dressing made me feel at ease and relieved my anxiety about how I would look after the surgery."; "It was so good that I forgot that I was attached to it."; "I walked 1 mile in the evening on the same day as my surgery." Areas for improvement: "The power unit was too heavy to carry around."; "The instructions are unclear regarding how to charge the device."; "The noise from the device prevented sleep."; "The tubing often got tangled with drain tubing."; "I was scared about getting the device wet in the shower."

CAT: OCEBM 4 · Start LOW (single-arm observational). -1 risk of bias: No control group, no complication endpoint, no baseline measure so nothing is a change score, no blinding and no adjustment; Direction of data collection is stated both ways - 'retrospectively' in the Key Findings box against a fixed prospective administration point and consent 'before data collection' in the Methods; patients are described as 'consecutively undergoing' with no screening log, no eligibility criteria, no exclusions and no denominator of patients approached; and no recruitment dates appear anywhere, no named ethics committee, no approval number, no protocol and no registration. -1 imprecision: n = 25 with a subgroup of 12, no sample size justification, and no statistical methods section at all - no test named, no software, no significance threshold, no handling of missing data, and no standard deviations, confidence intervals or n values for any individual item. The sm-versus-ibr gap is real arithmetic on the authors' own summary statistics - 74.8 plus-or-minus 19.9 against 53.9 plus-or-minus 9.6, difference 20.9 points, Welch t = 3.384, p = 0.0034, 95% CI 7.90 to 33.90, Hedges g 1.28 (0.41 to 2.14) - but it is confounded by a 22-YEAR median age gap (74 against 52) and by drains being present only in the reconstruction arm, neither adjusted, and it has no published minimal important difference against which to interpret it. Net VERY LOW for any comparative reading. The item-level domain pattern is on better ground: it is descriptive, it is internally coherent, it matches the free-text comments, and it corroborates independently what Miyanaga's Japanese inpatients and Simonsen's 1,282-patient dataset both found.

Figures: Figures checked

The abstract elevates it to one of three headline problems. In the Results it appears twice.

Under Independence: "The patients felt safe and confident with the dressing at home. Some experienced a lack of information of how to handle the PICO dressing." Quotation: "I didn't get any information but I still felt confident coming home."

Under Awareness: "Some participants felt they constantly had to make sure that the green light indicator was signaling or if there were any indication of malfunctioning. These participants considered that the information about the indicators at discharge was poor, if any." Quotation: "It wasn't very reliable. You had to keep an eye on it all the time".

And under Problem-Solving: "Some participants never found the reason for the beeps and light flashes and seemed to accept this without any further consideration."

The gap sits at three specific points, and none of them is in the operative consent: what the indicator light and the alarm mean and what each requires the patient to do; whether the device may be showered with; and what to do if the pump is dropped, beeps or appears to stop, and whom to call. Patients filled all three gaps themselves — removing and reinserting batteries, cycling the pump off and on, or substituting cloth washing for showering — without contacting the service. So the absence of calls to the ward cannot be read as evidence that the information was adequate.

CAT: Qualitative - GRADE-CERQual applied. Low to Moderate confidence for the INFORMATION-GAP finding; Very low for the four-patient preference finding · GRADE-CERQual assesses four components. Methodological limitations: moderate concerns. The study is nested in a well-conducted randomised trial with ethics approval and consent, two coders working independently then jointly, and all-author review to consensus. But the stated qualitative tradition CONTRADICTS ITSELF - 'qualitative content analysis' in the abstract, 'Braun & Clarke thematic analysis' in the Methods - and saturation is claimed while citing a method that expressly rejects saturation. There is no reflexivity statement: the only sentence touching the interviewer's position is 'The interviewer had been involved in the care of only one participant', and nothing in the text discloses that the interviewer is a vascular surgeon and a named co-investigator on the parent trial testing the device. No member checking, no transcript return, no audit trail, no coding tree, no data-availability statement and no agreement statistic. Language is never stated and quotations were 'translated by the authors' with no back-translation. SRQR is claimed with no checklist supplied and COREQ, the instrument specific to interview studies, is not used. RELEVANCE: HIGH for a South African vascular practice - elective infra-inguinal revascularisation with an inguinal incision, 7 days at 80 mmHg, discharged wearing the device, which is exactly the clinical situation the wiki's groin pages describe. COHERENCE: MODERATE for the information-gap finding, which appears in three separate sub-themes with consistent quotations; LOW for the preference finding, where a single quotation is offered and it is about perceived HEALING rather than stated preference. Adequacy of data: limited - 15 participants, median interview 10.53 minutes with the shortest at 6, only 8 of 15 participants quoted with one participant supplying three of the ten quotations, and frequency reported only as 'some', 'many' and 'more than one' so no theme can be weighted. Net: low to moderate confidence for the information gap; VERY LOW for the preference finding, where the elicitation question is not printed, appended or deposited anywhere, the four patients could not be blinded because one groin visibly carried a pump, the interviewer was on the trial team, and the denominator is four of an unknown number since the paper never states how many of the 15 had bilateral incisions.

Figures: Figures checked

CAT: Qualitative - GRADE-CERQual applied. Moderate confidence for the PAIN and SLEEP findings; Low for everything logistical, on transferability grounds · GRADE-CERQual, four components. Methodological limitations: moderate to serious concerns. Strengths: an interview guide reviewed by qualitative researchers, wound nursing specialists and plastic surgeons; digital recording with verbatim transcription; two coders with all-author validation; member checking claimed. Weaknesses: a design tradition mismatch, since a content-analysis design is justified by citing a PHENOMENOLOGY methods paper; Independent double coding is not claimed, with no disagreement rule, no agreement statistic, no software and no audit trail; saturation asserted with no matrix and no stopping rule; COREQ claimed with no checklist and numerous items unreported; and no reflexivity statement of any kind, which matters because the interviewer is a faculty member of the treating institution, two co-authors are in that hospital's plastic surgery department, and recruitment was performed by the patients' own ward nurses - courtesy bias therefore runs towards understatement of harms and overstatement of gratitude, and no step is reported to counter it. Translation is a further limitation: the study was conducted in Japanese and published in English by an UNNAMED translator with NO BACK-TRANSLATION, and coding appears to have occurred in Japanese with translation applied afterwards, so English readers cannot audit the code-to-quotation link; the published English is in places ungrammatical, which is a signal of an unverified translation step. RELEVANCE: MODERATE for pain and sleep, which are device-intrinsic; LOW for everything logistical, because every participant was an inpatient with mains power, an IV stand, nursing staff doing canister and dressing changes, and ward mates within earshot - and the authors state that NPWT remains an inpatient-only treatment in Japan with insurance requiring hospitalisation for the duration. The burdens described (unplugging the cord for the toilet, running an IV-stand tyre over the tube, worrying about ward mates) are WARD burdens, not home burdens, and the community burdens that dominate Western practice - dressing supply, self-management, district-nurse access, transport, returning to work carrying a pump - are structurally absent from this dataset. COHERENCE: GOOD; seven themes with 19 sub-themes and a quotation for each. ADEQUACY: LIMITED - 17 participants at wildly unequal exposure (treatment days 5 to 89, median 17) and the paper nowhere states at what point in the course each interview took place, nor whether 'treatment days' is duration at interview or total; denominators are supplied for three themes and not for the other four. NET: MODERATE for pain and sleep; LOW for the logistical themes on transferability.

Figures: Figures checked

• Shared theme: "threat to normality", with three categories — impact on daily life, manageability, powerlessness.
• Moist-dressing patients: pain dominated, especially during and after changes ("hurts like hell 1 hour after the change").
• NPWT patients: pain almost absent from the diaries; the main concern was the machine — alarms felt like a threat to safety; fear it would switch off if squeezed or fell; fear of pulling the tube off during sleep; some slept only on their back.

CAT: Qualitative study — GRADE-CERQual not formally applied; confidence moderate for the NPWT themes (clear method, saturation reported, small diary sample, single centre). · A small, well-described qualitative study. Its finding is specific and useful: patients on standard dressings wrote mostly about pain at dressing changes, while patients on NPWT barely mentioned pain but were preoccupied with the machine — alarms, fear of the tube pinching or the pump falling at night, sleeping only on their back. Both groups shared a theme of 'threat to normality'. It supports patient education and alarm-response instructions for anyone sent home on NPWT.

Figures: Figures checked