Background Physical activity (PA) reduces the risk of at least 20 medical conditions, including coronary artery disease, stroke, cancer and depression. Physical inactivity accounts for over 3 million preventable deaths per year worldwide. Despite clear evidence of the benefits, PA intervention studies typically have low recruitment (6–35%) and investigators often struggle to recruit representative participants. Research suggests that there are significant health differences between those who respond to invitations to participate and those who do not. Understanding the reasons for non-participation is essential in order to redress this potential health inequality.

Methods Just over 11,000 patients aged 45–75 years from seven south-west London general practices were invited to take part in a randomised controlled PA trial, which included a 3-month long intervention. Approximately 10% were recruited to the trial. We telephone interviewed a purposive (age, gender, ethnicity and practice) sample of 30 people who declined participation but consented to be contacted again. We used a topic guide, developed from a preceding physical activity trial. Interviews were audio-recorded, transcribed verbatim and thematically analysed.

Results Interviewees were supportive of walking as a form of exercise suitable for most people, the importance of our research and of primary care as an appropriate location for such interventions. Key ‘internal’ or ‘personal’ barriers to participation in the trial were the perception of being too physically active to participate; medical problems; and a lack of interest in increasing PA. Key ‘external’ barriers included work; travel from home; and other commitments. Less frequently cited reasons were generally ‘trial-related’ and included a reluctance to be randomised; the length of the intervention; wearing the pedometer; a preference for a different physical activity; and perceiving the trial literature to be aimed at a different group, for example older people.

Conclusion The commonest reason for non-participation was a perception of being already sufficiently active, suggesting that, for some individuals, programmes aimed at maintaining rather than increasing PA levels may be better received. However, it is established that most people over-estimate their activity levels, so education regarding objectively adequate levels of PA is important. Recruitment might be improved by adaptations to allow people to continue the programme while travelling, shorter programmes and tailoring activity to allow for medical problems or PA preferences. Taking care to ensure trial literature is succinct and inclusive is also important. Despite declining, interviewees were generally supportive of primary care as an appropriate location for such interventions.