The burden of headache disorders in North India: methodology, and validation of a Hindi version of the HARDSHIP questionnaire, for a community-based survey in Delhi and national capital territory region

The study adopted LTB’s standardized protocol for population-based studies in a cross-sectional survey using multistage random sampling conducted in urban Delhi and a surrounding rural area. Trained interviewers visited households unannounced, randomly selected one adult member from each and applied the Hindi version of HARDSHIP in face-to-face interviews. The most bothersome headache reported by participants was classified algorithmically into headache on ≥ 15 days/month (H15 +), migraine (including definite and probable) or tension-type headache (including definite and probable). These diagnoses were mutually exclusive. All participants diagnosed with H15 + and a 10% subsample of all others were additionally assessed by headache specialists and classified as above. We estimated the sensitivity and specificity of HARDSHIP diagnoses by comparison with the specialists’ diagnoses.

From 3,040 eligible households, 2,066 participants were interviewed. The participating proportions were 98.3% in rural areas but 52.9% in urban Delhi. In the validation subsample of 291 participants (149 rural, 142 urban), 61 did not report any headache (seven of those assessed by HARDSHIP, eight by headache specialists and 46 by both) [kappa = 0.83; 95% CI: 0.74-0.91]. In the remaining 230 participants who reported headache in the preceding year, sensitivity, specificity and kappa with (95% CI) were 0.73 (0.65-0.79), 0.80 (0.67-0.90) and 0.43 (0.34-0.58) for migraine; 0.71 (0.56-0.83), 0.80 (0.730.85) and 0.43 (0.37-0.62) for TTH and 0.75 (0.47-0.94), 0.93 (0.89-0.96) and 0.46 (0.34-0.58) for H15 + respectively.

This study validates the Hindi version of HARDSHIP, finding its performance similar to those of other versions. It can be used to conduct population surveys in other Hindi-speaking regions of India.