The Wound-QoL instrument. The Wound-QoL instrument was developed by the Institute for Health Services Research in Dermatology and Nursing at University Medical Center, Hamburg, Germany, to fulfill a need for a short, easy-to-use, multidimensional questionnaire measuring HRQOL in patients with hard-to-heal wounds.20 The instrument is a combination of the English CWIS and 2 German wound specific instruments, The Freiburg Life Quality Assessment for Wounds and the Wurzburg Wound Score. Together, the 3 questionnaires comprised a total of 92, 5-point Likert-scaled items from which 17 items were extracted and prospectively analyzed in a virtual validation study. The items were assigned to 3 subscales according to the domains Body (items 1–5), Psyche (items 6–10), and Everyday living (items 11–16). The subscales were summed for an overall global score together with an additional question regarding financial impact. Each item is rated by the respondent on a 5-point Likert-scale where 0 = not at all and 4 = very much; 4 was the worse rated level of HRQOL (see Figure 1).
The psychometric properties (reliability, validity, responsiveness, ceiling and floor effect) then were tested in a validation study performed at 2 different German wound centers with a total sample size of 227 patients with hard-to-heal wounds of various etiologies, showing satisfactory psychometric properties according to internal consistency (Cronbach alpha >0.80), convergent validity (0.5-0.7), and responsiveness (P ≤.001).21 The instrument also was tested for test-retest reliability in a study of 100 patients with hard-to-heal wounds using intraclass correlation (ICC). The result showed excellent reliability (range 0.79-0.86) calculated from the questionnaires, with an interval of 3 to 7 days between first and second completion.22
The translation process. The English version of the Wound-QoL instrument was translated into Swedish according to instructions by the Standard Linguistic Validation Process created by MAPI Research Institute,23 a 7-step process that includes forward and backward translation and pilot testing. The translation process was a collaboration between independent translators and the research group. The re-translated version was sent to the creators of the original version of the instrument, the Institute for Health Services Research in Dermatology and Nursing at University Medical Center, Hamburg, Germany, who evaluated the translation, made adjustments, and gave their approval.
Materials and data collection. Using convenience sampling methods, patients >18 years old with hard-to-heal wounds treated at an outpatient wound clinic in a large city in Sweden were recruited while visiting the clinic for their scheduled wound treatment between August 2015 and July 2016. Patients with less than adequate cognition to complete the questionnaire and patients who had not mastered the Swedish language were excluded.
The patients were asked to complete the Swedish version of the Wound-QoL instrument as well as the generic HRQOL instrument EQ-5D-3L at baseline at study enrollment and after 6 weeks. For the 6-week follow-up, the questionnaires were either distributed at a planned revisit at the wound clinic or sent to the patients with a letter of instruction and a postage-prepaid envelope with instructions to return the forms within 1 week. The patients completed the questionnaires in paper-and-pencil format, then data were transferred into a computer file for analysis. Basic demographic information and wound data, including gender, age, and type and duration of the wound, were obtained from either the patient or his/her medical record in the wound clinic. Patients also were asked in an open-ended question to describe their general experience of completing the Wound-QoL questionnaire; the responses were registered as field notes by the researcher.
All patients gave their verbal informed consent to participate in the study. Patient confidentiality was ensured by coding the questionnaires and records; encoding was handled solely by the researcher in accordance with the Swedish Data Protection Act (SFS,1998:204). The study was performed according to the Helsinki Declaration, and study approval was obtained by the local Ethics Committee (2015/714-31).
Psychometric analysis and statistics. Data from all questionnaires were transferred into a computer program to assess reliability, validity, responsiveness, and ceiling and floor effects. SPSS version 23.0 (IBM, Armonk, NY) was used for all calculations, and all tests were 2-sided with a 95% confidence interval. The results were considered significant at P <.05. Descriptive statistical analysis was performed for presentation of the demographic and wound variables.
Reliability. Reliability was measured with internal consistency to estimate the correlation between the different items for each domain at baseline using Cronbach’s alpha coefficient; Cronbach’s alpha 0.70 or higher was considered acceptable.24 Stability (test-retest reliability) of the instrument over time was measured with an ICC of patients who rated their health as unchanged in the EQ-5D-3L at the 6-week follow-up. An acceptable level of ICC in this study was considered >0.70.25
Validity. Validity in this study was assessed by measuring criterion validity comparing the scores of the Wound-QOL to the gold-standard instrument (EQ-5D-3L), a process similar to that used by the developers of the Wound-QOL instrument.20 All 3 domains of the Wound-QoL instrument together with the global score were compared to the EQ-5D-3L time trade-off (TTO) index and to the EQ-5D-3L VAS (ie, the patient’s self-rated health). The correlations were calculated with Spearman’s rho due to data on an ordinal level and were expected to be negative and moderately strong based on the correlation from the development process.20 A moderately strong correlation was defined as 0.3 to 0.49 according to Cohen.26
Responsiveness. Standardized response mean (SRM) was used for assessing responsiveness, a measurement of the ability of the instrument to detect clinically important changes. The calculation was performed by dividing the difference between baseline and 6-week follow-up in the group assessing their health as improved at the 6-week follow up according to the EQ-5D-3L questionnaire. SRM <0.5 is considered as small ability to detect change, 0.5–0.8 is moderate, and >0.8 is large.27
Ceiling and floor effect. Ceiling and floor effects occur when the scores of an instrument reach the maximum or minimum possible score. If this is noted in 15% or more of the questionnaires, it is hard to detect improvement or decline in the rated health status.28