Abstract
Disability weights are critical to assess the cost-effectiveness of dengue control interventions, but existing estimates are few and variable. To generate robust estimates, we combined a meta-analysis and original data from Puerto Rico for acute episodes, and a model for chronic (persistent) dengue, which leads to depression, fatigue and other symptoms. The meta-analysis searched major data bases for empirical studies through 2016 with “dengue” combined with “quality of life,” “disability,” or similar terms in 4 languages. The Puerto Rico study was based on detailed retrospective interviews with 101 (69 hospitalized and 32 ambulatory) patients with laboratory-confirmed dengue across the island. To compute the disability-adjusted life year (DALY) burden for acute episodes, we fit the extent of disability by day of illness (where 0 denoted no disability and 1 disability equivalent to death) to reported data and computed the area under this disability curve. The model for chronic dengue was based on a prevalence regression (C Tiga et al, AJTMH, 2016) and the disability weight for “major depressive disorder, mild episode” (J Salomon et al, Lancet, 2015). The meta-analysis found 5 published studies. Results are presented as means ± standard errors. The Puerto Rico study found: durations (in days) 15.31±0.99 (all cases), 12.13±1.26 (ambulatory cases), 16.80±1.30 (hospitalized cases); DALY per episode 0.0174±0.0012 (all cases), 0.0150±0.0017 (ambulatory cases), 0.0185±0.0016 (hospitalized cases); disability weights per day 0.4105±0.0145 (all cases), 0.4292±0.0275 (ambulatory cases), 0.4018±0.0170 (hospitalized cases). The model for persistent dengue found 0.092 person years and a DALY burden of 0.0133±0.0026 per dengue episode. Combining all studies, ambulatory and hospitalized episodes have average DALY burdens per episode of 0.0116±0.0026 and 0.0148±0.0053, respectively (acute phase) and 0.0249 ±0.0037 and 0.0281±0.0034, respectively (both phases). These results are within the broad range of previous estimates, but offer more systematic estimates of burden for cost-effectiveness analyses and burden of disease estimates.