Harnessing routinely collected data for the evaluation of early years interventions: insights from a scoping review...

Routinely collected data (RCD) describes data about individuals that is documented routinely in practice. Evaluation of early years interventions is imperative for improving child health and development, and utilisation of RCD for research could overcome limitations with traditional study designs, such as participant burden, attrition and disappointment bias. A systematic understanding of if (and how) RCD is being used to evaluate early years interventions is lacking.

To scope the literature on how RCD is being used to evaluate the effectiveness of early years interventions being delivered in the UK.

The study protocol was registered online (osf.io/cug36). Included studies were interventions for expectant parents and parents of children aged 0-5 years, with a quantitative measure obtained from RCD as an outcome. The study scope was limited to the UK, from January 2000 to November 2024. Completed studies and study protocols were eligible, encompassing grey literature and peer reviewed sources. Medline, Psycinfo and Embase databases were searched via Ovid.

32 unique studies published 2009-2024 had used or were planning to use RCD as an outcome in an evaluation of an early years intervention. Most measured more than one outcome, and the most common were birth outcomes and child education. Many studies noted limitations of RCD, particularly being limited by measures available, and fewer noted strengths.

Whilst RCD expands evaluation opportunities for measuring multiple relevant parent and child outcomes simultaneously, researchers are limited by what measures are available in RCD. We make specific recommendations to improve the use of RCD in early years evaluations. With careful consideration, RCD can provide useful insights regarding the impacts of early years interventions. This scoping review reveals an emerging methodology that may have increasing importance for evaluating early years policies and interventions, but may currently be constrained by data availability and quality.