Study setting and population

We used data from a randomized controlled trial comparing negative pressure wound therapy (NPWT) with standard treatment of adults (≥18 years) with acute (≤72 h) extremity blast or gunshot wounds (ClinicalTrials.gov ID: NCT02444598) [11]. Participants were enrolled at two civilian hospitals in Jordan and Iraq from June 9, 2015, to October 24, 2018. At the end of the first surgical procedure, the participants were randomly assigned to NPWT, involving a commercial NPWT device, or standard treatment, involving wound dressings with non-adhesive sterile gauze, according to the International Committee of the Red Cross war surgery protocol [12]. The groups were well balanced in baseline characteristics, and NPWT did not yield superior clinical outcomes compared with standard treatment [13]. For the present study, we included all randomized patients with blast or gunshot wounds who completed both of the study instruments (described below) at baseline, irrespective of treatment modality. Dedicated research nurses collected and entered the data into paper-based case report forms during the study period.

Instruments

The Wound QoL instrument is developed and validated to assess disease-specific HRQoL of patients with hard-to-heal wounds [9]. The instrument consists of 17 items on impairment, which are always assessed retrospectively. The items are assigned to the subscales “body”, “psyche”, and “everyday life”. The subscales have been shown to be valid and responsive [14]. In the present study we used items 1–4 in the “body” subscale, 1) my wound hurt, 2) my wound had a bad smell, 3) there was a disturbing discharge from the wound, and 4) the wound has affected my sleep. The “psyche” and “everyday living” subscales were not deemed relevant for acute wounds. Therefore, these subscales were omitted.

The World Health Organisation 20 question self-reporting questionnaire (SRQ-20) was originally developed as a tool to screen for mental disorder [15], and has been used to assess symptom change over time [16]. The instrument is widely used, has a validated Arabic version [17, 18], and is validated for the use in low- and middle-income country emergency settings [19], as well as for longitudinal assessments [20]. In the present study we used the SRQ-20 questions 1, 2, 3, and 7, which measure somatic symptoms [21].

Reliability

We evaluated the reliability of the “body” subscale of the Wound QoL instrument. To test reliability in regard to internal consistency we estimated the correlation between the four subscale items at baseline by using Cronbach´s alpha coefficient. A Cronbach´s alpha coefficient of 0.7 or higher was considered acceptable [22]. We also tested stability (test-retest reliability) with interclass correlation (ICC) among the participants who had two Wound QoL measurements. An ICC above 70 was considered acceptable [23].

Validity

We assessed validity according to content validity and criterion validity. Content validity refers to the extent to which a measure covers relevant areas of the condition or disease [24]. Criterion validity compares scores from the instrument in focus to an instrument that targets the same phenomena, and is regarded as gold standard [25]. Content validity was assessed by an expert group following discussions on which of the Wound QoL items could be relevant for the patient group under study. Consensus was reached to exclude the “body” subscale’s item number 5 (the treatment of the wound has been a burden) since this item focuses on the treatment of long-lasting or chronic wounds. The remaining four items were considered relevant for the present patient group with acute wounds. Since this discussion, a revised version of Wound QoL has been published, in which item number 5 has been removed from the “body” subscale [26]. We measured criterion validity by comparing the score of the “body” subscale of the Wound QoL instrument to a total score of the SRQ-20 items 1, 2, 3, and 7, using SRQ-20 as gold standard. The correlation between the two instruments was calculated with Spearman´s rho, with scores between 0.30 and 0.49 indicating acceptable correlation [27].

Ceiling and floor effect

Ceiling effect occurs when the scores of an instrument reach the maximum possible score. Floor effect occurs when the scores are equal to the minimum possible score. If ceiling or floor effect occurs in 15 percent or more of the questionnaires, it is considered hard to detect improvement or decline in the rated health status [28].

Statistical analysis

We analysed data using SPSS^® Statistics (IBM Corp., Armonk, New York, USA).

Ethics

This study was part of a larger clinical trial [13], which received ethical approval from the Ethics Review Committee of the Jordan Ministry of Health (MOH REC 150037), the Ethics Review Board of Médecins Sans Frontières (ID 1520), the Research Ethics Committee, Kurdistan Regional Government in Iraq (2:10 6/3/2017), and the Swedish Ethical Review Authority (2019–01975). All participants provided written informed consent.

Reliability

The Cronbach´s alpha coefficient was 0.7. Test-retest analysis was performed on questionnaires from 90 participants with an ICC of 0.47.

Validity

Content validity was discussed within the expert group and the four items constituting the “body” subscale were found to be valid and relevant for the studied patient group. The criterion validity assessments included all 152 participants. Spearman´s rho was 0.20.

Ceiling and floor effect

We found no signs of ceiling effect as no participant received the highest score. Some 19/152 (12.5%) participants had minimum scores, indicating acceptable levels regarding floor effect.

Limitations

Some limitations should be considered. First, the sample size is rather small. However, to prospectively include study participants with acute blast or gunshot wounds is a challenge. Particularly at low-resource or conflict settings. Second, only the “body” subscale of the Wound QoL instrument was used. However, this was the only of the three subscales that was considered relevant. The main limitation is the use of SRQ-20 as gold standard. The SRQ-20 instrument is mainly used as a screening tool for mental disorder. Still, we anticipated that the instrument items measuring somatic symptoms would prove to be transferrable to the physical aspects of HRQoL. The low criterion validity indicates that this might not be the case. Bearing the above mentioned limitations in mind, we still believe that the present study adds important knowledge on the utility and limitations of the SRQ-20 instrument, considering its widespread use.