Qualitative interviews when developing the haemophilia caregiver impact measure© (HCI) documented the importance of capturing the positive aspects of caregiving, not just the negative. The present study thus investigates the construct underlying the positive emotions HCI subscale and tests models proposing a more comprehensive way of thinking about this construct. Secondary analysis was implemented on longitudinal web-based survey data (n = 323) from haemophilia A or haemophilia B caregivers. Person-reported outcomes (PROs) included the HCI, the PROMIS-10 and Ryff psychological well-being subscales. Predictors included caregiver demographics; patient haemophilia characteristics; exercise; adherence; and quality-of-life (QOL) appraisal processes as measured by the brief appraisal inventory (BAI) which yields composite scores assessing awareness of challenges, fulfillment and growth, stay positive, social comparison and interpersonal problem-solving. Second-order factor analysis, structural equation modelling and residual modelling were implemented. A structural equation model fit the data well that contained bifactor representation of well-being with a general factor comprised of environmental mastery, positive relations with others, physical functioning and emotional functioning. Positive emotions was modelled as a component of well-being, with a unique component ('Alchemy') characterized by its associations with stay positive, and awareness of challenges appraisals, and difficulty paying bills. Alchemy had positive linear relationships with the first two, and a positive quadratic relationship with difficulty paying bills. Adopting positive-focused ways of thinking about one's life limitations may transform the negatives of haemophilia caregiving into something positive. Such cognitive habits reflect an awareness and acceptance of the limitations imposed by haemophilia caregiving. © 2019 John Wiley & Sons Ltd.