Estimating Health State Utilities in Hemophagocytic Lymphohistiocytosis

Abstract Background Hemophagocyti.c lymphohistiocytosis (HLH) is a rare and severe disorder characterized by abnormal activation of the immune system. Primary HLH causes prolonged fever, spleen and liver enlargement, and organ dysfunction, usually in infancy and early childhood and is fatal if left untreated. As effective treatment options emerge, such as emapalumab-lzsg, Health Technology Assessment bodies around the world will assess them in terms of cost-effectiveness. This study was designed to estimate quality of life weights (utilities) for such analyses. Methods Vignettes were developed describing HLH treatment related health states. Health states included active HLH, HLH plus neurological symptoms, receiving chemotherapy, undergoing stem cell transplant (SCT), graft versus host disease (GVHD), cure and end of life care. The vignettes were based on information from in depth interviews with clinical specialists; and qualitative research with four parents of children with primary HLH aged between 1 and 18 years old. The vignettes were then assessed in time trade off (TTO) interviews with members of the UK general public in one on one face to face interviews with trained, experienced interviewers. Preference data were analysed using the generalised estimating equations framework. Results Detailed qualitative data captured the substantial burden of this disease for young children. One hundred participants completed the TTO interviews. The utility score for Active HLH was estimated as 0.32 (95% CI, 0.24 to 0.39). Values for other states were HLH plus neurological symptoms (0.27, 95%CI 0.18–0.35), receiving chemotherapy (0.26, 95%CI 0.17–0.34), undergoing SCT (0.18, 95%CI 0.07–0.28), GVHD (0.07, 95%CI -0.04-0.17), cure (0.72, 95%CI 0.67–0.77) and end of life care (− 0.17, 95%CI -0.27- -0.07). Conclusions This study has estimated utility weights for seven different HLH related states which are based on detailed input from carers and physicians and have good face validity. There are few other options for collecting these data in an ultra-rare setting.