HURT (Headache Under-Response to Treatment) questionnaire in the management of primary headache disorders: reliability, validity and clinical utility of the Arabic version

Background: To support better headache management in primary care, the Global Campaign against Headache developed an 8-question outcome measure, the Headache Under-Response to Treatment (HURT) questionnaire. HURT was designed by an expert consensus group with patient-input. It assesses the need for and response to treatment, and provides guidance on actions to optimize therapy. It has proven content validity. We aim to evaluate the Arabic version of HURT for clinical utility in primary care in Saudi Arabia. Methods: HURT was translated according to the Global Campaign's translation protocol. We assessed test-retest reliability in consecutive patients of four primary-care centres, who completed HURT at two visits 4-6 weeks apart while receiving usual care. We then provided training in headache management to the GPs practising in these centres, which were randomized in pairs to control (standard care) or intervention (care guided by implementation of HURT). We assessed responsiveness of HURT to clinical change by comparing base-line responses to HURT questions 1-6 with those at follow up. We assessed clinical utility by comparing outcomes between control and intervention pairs after 3 months, using locally-developed 5-point verbal-rating scales: the patient-satisfaction scale (PSS) and doctor-satisfaction scale (DSS). Results: For test-retest reliability in 40 patients, intra-class correlation coefficients were 0.66-0.78 for questions 1-4 and 0.90-0.93 for questions 5-7 (all P <= 0.001). For the dichotomous response to question 8, Kappa coefficient = 1 (P < 0.0001). Internal consistency was good (Cronbach's alpha = 0.74). In 342 patients, HURT signalled clinical improvement over 3 months through statistically significant changes in responses to questions 1-6. PSS scores were higher among those in whom HURT recorded improvement, and also higher among those with less severe headache at baseline. Patients treated with guidance from HURT (n = 207) were more satisfied than controls (n = 135), but this did not quite reach statistical significance (P = 0.06). Conclusion: The Arabic HURT Questionnaire is reliable and responsive to clinical change in Arabic-speaking headache patients in primary care. HURT showed clinical utility in this first assessment, conducted in parallel with studies elsewhere in other languages, but this needs further study. Other Arabic instruments are not available as standards for comparison.