The effect of flexible settings on participation rates, data quality, and costs in a longitudinal study of ethnically diverse young adults: the London Determinants of young Adult Social wellbeing and Health (DASH) study

Harding, S., Enayat, Z., Cassidy, A., Read, U., Ferrell, C. and Molaodi, O. (2013) The effect of flexible settings on participation rates, data quality, and costs in a longitudinal study of ethnically diverse young adults: the London Determinants of young Adult Social wellbeing and Health (DASH) study. Lancet, 382(S3), S41. (doi: 10.1016/S0140-6736(13)62466-6)

Abstract

Background: The loss of young and disadvantaged people in cohort studies compromises representativeness and scientific and public value. The Determinants of young Adult Social wellbeing and Health (DASH) study is a longitudinal study of young people in London, mainly from ethnic minorities, and has a social epidemiological focus. The cohort is now aged 21–23 years, an important transition phase to adulthood that influences health and wellbeing in later life. We examine the costs and benefits of different settings on motivation for participation. Methods: 6643 11–13-year-olds from 51 London schools took part in the baseline survey in 2002–03 and 4781 were seen again at age 14–16 years. So far, 81% (5414 of 6643) of the cohort has been traced through friendship networks, social media, and community campaigns. A feasibility study of best practice for full follow-up was done with 400 participants between June, 2012, and August, 2013, the design of which was informed by systematic reviews. A sample of about 50 people per ethnic group (Indian, Pakistani, Bangladeshi, black Africans, black Caribbeans, white British) was chosen to give a reasonable spread by sex and socioeconomic position and across all boroughs and schools. Flexibility in settings was based on number of interview locations available (least, two; intermediate, three; most, six) and appointment times (weekdays 0900–1700 h, other). Interview locations were general practice surgeries and community pharmacies within 2 km of residential postcodes; Muslims living in east London were offered the choice of a Clinical Research Centre close to a large mosque; and central locations (Clinical Research Facility at University College Hospital and King's College London). Participants were given GB£25 shopping vouchers and £30 vouchers if they needed childcare, and reimbursed travel expenses. A main questionnaire covered topics such as socioeconomic circumstances, social support, racism, own and parental health, and health behaviours. Dietary intake over the past 24 h was obtained in a recall questionnaire. Physical measures were done and a blood sample was also taken. The primary outcome was participation rate. Secondary outcomes were cost per participant (CPP) and quality of data. A combination of our own study team and surgery nurses was used. Interviewers were trained for 3 days and compliance to protocol was assessed by regular shadowing and analysis of data by interviewer identification code. Findings: 97% (400 of 413) of those invited took part. Participation increased with flexibility in choice of interview locations (least 15%, 95% CI 11–19; intermediate 27%, 23–32; most 58%, 53–63) and was substantially boosted by weekend and evening appointments (weekday 0900–1700 h 76%, 71–80; other 24%, 17–30). CPP varied by setting from £229 (95% CI 220–239; King's College London, 179 participants seen by study team) to £283 (276–289; Clinical Research Facility, 22 seen by their nurses). Compliance to the protocol and the quality measures were high across settings (p<0•05 for 11 of 13 measures). The CPP for those seen (40 participants) by surgery nurses was intermediate at £240 (95% CI 225–255), but the quality measures were least favourable (eg, consent for bloods 82•5%, 95% CI 70•2–94•8 vs 96•0%, 92•7–99•2 for King's College London). Compliance to protocol was compromised by very busy environments of inner-city surgeries in deprived areas and limited familiarity with population cohorts. Interpretation: Flexibility in interview locations and appointment times boosted participation rates. The use of our own study team rather than surgery nurses was associated with lower CPP and better quality data. Participants were not randomly assigned to a setting, which comprised comparisons of setting-specific participation rates.

• Item Type: Articles
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Read, Dr Ursula and Harding, Professor Seeromanie and Ferrell, Mrs Catherine and Molaodi, Dr Oarabile
• Authors: Harding, S., Enayat, Z., Cassidy, A., Read, U., Ferrell, C., and Molaodi, O.
• College/School: College of Medical Veterinary and Life Sciences > School of Health & Wellbeing > MRC/CSO SPHSU
• Journal Name: Lancet
• Publisher: The Lancet Publishing Group
• ISSN: 0140-6736
• ISSN (Online): 1474-547X