Validity and test retest reliability of the vascular quality of life Questionnaire-6: a short form of a disease-specific health-related quality of life instrument for patients with peripheral arterial disease

Abstract

Background: Many existing patient-reported outcome measures are extensive regarding both patient burden and administration, and in terms of analysing and reporting results. The VascuQoL-6 (VQ6) - a short version of the original Vascular Quality of Life Questionnaire (VascuQoL), a disease-specific instrument for peripheral arterial disease - was recently developed. However, the VQ6 has not yet been empirical tested with regard to content validity, construct validity and test retest reliability. Our aim was, therefore, to explore both the validity and the reliability of the VQ-6 in a target population with established peripheral arterial disease.

Methods: Two hundred patients treated at two vascular centres were consecutively recruited for the survey. Administered questionnaires included VQ6 and the Short Form Health Survey-36 (SF-36). Out of the 200 patients, 150 also received a second VQ6 questionnaire for a test-retest assessment. Further, a purposive sample of 22 patients consented to participate in cognitive interviews. All included patients suffer from peripheral arterial disease. The questionnaire data was tested by both Rasch analysis and traditional psychometric methods, while the cognitive interviews were analysed descriptively.

Results: The validity and reliability of the VQ6, as tested in a target population without the surrounding 19 items from the original VascuQoL, was high, in general, and a good fit to the Rasch model was observed. Further, an excellent internal consistency and significant correlations between comparable dimensions in SF-36 were demonstrated. In the test-retest analysis, the percentage agreement was somewhat poor (<70%) in the six items. However, no systematic disagreements between the two assessments were seen in any of the six items, and the test-retest assessment for the VQ6 sum score showed an acceptable intraclass correlation coefficient (0.86). Finally, all items in the VQ6 were considered as both understandable and relevant by the interviewed patients.

Conclusions: The VQ6 has acceptable to good psychometric properties with regard to data quality, scale assumptions, targeting, validity and reliability. Further, VQ6 seems to be easy to use and comprehend within the target population of patients with PAD.

Keywords: Cognitive interviews; Health-related quality of life; Peripheral arterial disease; Rasch analysis; VascuQoL; VascuQoL-6.

Conflict of interest statement

Ethics approval and consent to participate

All patients were informed that participation was voluntary, and written informed consent forms for participation were collected from each patient. The study was approved by the Regional Ethical Review Board at Lund University (no 315/2008 and no 750/2013).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Fig. 1

Category thresholds between the four response options in category probability curves for the six items in VQ6

Fig. 2

Person-item map for the six items in VQ6 showing the ability of the patients to the left and the difficulty of the items (VQ1-VQ6) to the right of a linear scale (from −6 to +6 logits). The measure (logit) shows the logarithm of the odds ratio for being able to perform an item activity successfully. The distance between a respondent and an item shows the probability of the respondents to perform successfully the item activity. Preferable, person and item should center on the same mean value. M, S and T represent mean, 1SD and 2SD, respectively. Each # is 2 persons. The instrument is well targeted because the item means and person means were separated by only 0.48 logits

Fig. 3

The contingency table (a) (T1 vertical = first measurement, T2 horizontal = second measurement) and the ROC curve (b) show no systematic disagree in concentration between the two assessments of item 4. ROC, relative operating characteristic

Fig. 4

The contingency table (a) (T1 vertical = first measurement, T2 horizontal = second measurement) and the ROC curve (b) show no systematic disagree in concentration between the two assessments of item 3. ROC, relative operating characteristic

Fig. 5

The Bland-Altman plot display the differences plotted against the averages of the two measurements of VQ6 sum score. The limits of agreement are defined as the mean difference ± 1.96SD of differences

Fig. 6

Cognitive interviews with patients (n = 22)