Evidence suggests that engagement in care (EIC) may be worse in young people with perinatal HIV (PHIV) than in adults or children living with HIV. However, there is no consensus on how best to measure EIC; and most studies use a simplistic definition based on number of clinic visits attended per year and examine limited predictors of EIC. In this thesis, I took an existing EIC algorithm for adults living with HIV in England, and adapted it to young people with PHIV in the Adolescent and Adults Living with Perinatal HIV cohort (AALPHI), using data from 2013-2015. A wide range of potential predictors of EIC from the AALPHI dataset were explored in logistic regression models (allowing for clustered data). Predictors of EIC identified in the quantitative analysis were then explored in focus groups with young people with PHIV to help contextualise the findings and to explore if they resonated with the experiences of young people themselves. Of 3,585 months of total follow-up in 306 young people, 3,126 (87%) person-months were classified as engaged in care. Multivariable predictors associated with poorer odds of being engaged in care were: baseline viral load >50c/mL vs. viral load ≤50c/mL; Asian/mixed ethnicity vs. black ethnicity; ever self-harmed vs. not; self-assessed adherence as bad/not so good/not on ART vs. good/excellent. Findings from the focus groups support and expand the quantitative results. Young people described actively choosing when to and when not to attend clinic depending on what they thought their viral load was or to hide non-adherence and self-harm. My adapted algorithm provides a more sensitive method to measure EIC in young people with PHIV. Identifying which young people are less likely to engage in care may allow targeted interventions to support young people to attend clinic and optimise their health outcomes. Findings from the focus groups provide a much broader understanding of the social meaning of EIC.