Search:
Be Well CYP and young people’s wellbeing: encouraging findings, important caveats
How can we better support young people’s mental health and wellbeing? Social prescribing connects young people with support beyond clinical care, such as community activities and practical help. At #BeeWell, we worked with Big Life and PBE to explore how wellbeing scores among Be Well CYP participants compared with those of a matched comparison group drawn from our survey of secondary-school pupils.
As part of the wider Be Well service, Big Life delivered Be Well CYP, a specialist pathway offering one-to-one wellbeing coaching for young people aged 13 to 25 across Manchester. Coaching began with a person-centred assessment of each young person’s interests, circumstances and goals. Coaches then tailored support to those goals, whether that meant strategies for managing emotions and everyday worries, help finding local activities, or connections to practical support. Sessions took place in community settings, including community centres and outdoor spaces, and young people received between one and 14 sessions during a period of support.
Our analysis found an encouraging association: Be Well CYP participants had higher wellbeing scores than the comparison group after we took account of selected differences between them. However, this does not establish that the programme caused the difference. Read on to find out why:
What did we look at?
We compared wellbeing scores recorded after participation in Be Well CYP with scores from young people in the #BeeWell survey of secondary-school pupils.
Both datasets used the Short Warwick–Edinburgh Mental Wellbeing Scale, or SWEMWBS: a short questionnaire measuring self-reported wellbeing. Higher scores indicate better wellbeing.
The main analysis compared scores between groups, rather than changes over time. A separate exploratory analysis of changes over time involved only a small number of participants and was inconclusive.
How did we make the comparison?
People who join a programme may differ from those who do not. Simply comparing their average scores could therefore be misleading.
We used exact matching to create groups with the same recorded gender, ethnicity, sexuality, cohort year and level of neighbourhood deprivation. Deprivation was measured using Index of Multiple Deprivation deciles, which describe relative disadvantage in an area.
Of the original 322 participants, 232 had the information needed for the main analysis. We found matches for 229 participants, using 34,537 #BeeWell survey observations.
We then used a statistical model, weighted using the matching results, to take account of age and the matching characteristics. The model allowed the association with participation to vary across these characteristics and estimated an average difference in wellbeing scores for the matched participants.
This addressed selected recorded differences—not every factor that might affect wellbeing.
What did we find?
The analysis estimated that Be Well CYP participants scored about 2.4 points higher on the wellbeing questionnaire than the comparison group after adjustment.
This is an estimated difference between groups, not a measured improvement of 2.4 points for each participant. It does not establish how much of the difference, if any, was caused by Be Well CYP. Nor does this analysis alone tell us how meaningful that difference was in young people’s everyday lives.
Several alternative approaches also estimated a positive difference, although they did not resolve the age mismatch or other limitations of the comparison data.
Why is the age difference important?
Participants were, on average, about 19 years old, while the comparison young people were about 13. Requiring matches of exactly the same recorded age left only eight participants with matches. We therefore included age in the statistical model instead.
This retained more participants, but did not solve the mismatch. Because there were few similar-aged comparison pupils, the estimated difference depended partly on extending the modelled relationship between age and wellbeing beyond the ages well represented in the comparison group.
This is called extrapolation. If wellbeing changes with age differently from the pattern assumed by the model, the estimated difference could be misleading. Statistical adjustment cannot substitute for comparing similar-aged young people.
What else could affect the result?
Other differences could matter—for example, wellbeing before joining the programme, reasons for referral or support received elsewhere. Matching cannot account for differences that are not adequately measured and included.
Missing information also restricted the sample. Participants with complete records may differ from those excluded, so the findings should not automatically be applied to everyone supported by Be Well CYP, or to young people across the UK.
Service records and surveys collected information in different ways. Ages and wellbeing scores may not have been recorded at equivalent points in time; matching on cohort year alone did not guarantee that measurement dates aligned.
What should happen next?
The finding warrants further investigation, but is not proof of programme effectiveness. A stronger evaluation would:
Our conclusion is one of cautious encouragement. The analysis found higher wellbeing scores among Be Well CYP participants, but better comparison data and a stronger evaluation are needed to understand why that difference arose.
The accompanying PBE report presents a separate value-for-money assessment, using the estimated wellbeing difference alongside additional assumptions about how much of the difference was caused by the programme, how long any benefit lasted, and its monetary value.
