Cohort study of 8,000 adults aged 50-64 years who were recruited from GP practices across England that are part of the Clinical Practice Research Database (CPRD). At baseline, people agreed to complete questionnaires about their work and health, and determinants of these, annually for 3 years and gave permission for their responses to be linked to their anonymised records in CPRD so that we have relatively low-cost access to information about diagnoses, consultations in primary and secondary care and medications prescribed. Data collection is on-going and will run initially until 2018. The need to keep Britain’s ageing population economically active has prompted government policies aimed at extending working lives. However, working beyond the traditional retirement age may not be feasible for those with major health problems of ageing, and depending on occupational and personal circumstances (e.g. savings, retirement intentions, domestic responsibilities, whether work is arduous, rewarding), might be either good or bad for health. We will recruit 8,000 50-64 year-olds from 24 practices contributing to the Clinical Practice Research Datalink (CPRD, formerly GPRD). Participants will complete questionnaires about their work and home circumstances, initially over a 3-year follow-up, and with their permission, we will access their health data via the CPRD. The inter-relation of changes in employment (with reasons) and changes in health (e.g. major new illnesses, new treatments, mortality) will be examined. CPRD linkage offers major advantages, notably cost-effective capture of frequent, detailed, objective health data. We will be able to examine the impact of health on work at older ages (e.g. how often arthritis suffers have to quit work) and of work on health (e.g. whether mental health is helped or worsened by deferring retirement, and in what circumstances). Findings should inform government policy and improve the design of work for older people.

Data have been collected via postal questionnaire. Patients are registered with 24 general practices currently participating in CPRD data collection. The location of these practices depends on the willingness of GPs to support the study. Ideally, however, there will be purposive quota sampling aimed at ensuring a reasonable geographical spread e.g. at least 25% of total list size from each of the South, Midlands and North. We consider this sensible as unemployment rates and patterns of illness behaviour and consulting will vary geographically. However, there is no requirement that the distribution of respondents’ occupations should be nationally representative. Subjects received a questionnaire at baseline and a briefer questionnaire at follow-up (annually for 3 years initially). The age limits have been chosen on the expectation that some subjects will already be retired at baseline, some will retire at follow-up with others deferring retirement, and many will develop retirement plans. No other inclusion or exclusion criteria are envisaged - all patients in the relevant age band from each participating practice were invited to participate for the duration of follow-up. The questionnaire at baseline asked about: 1) current work status; 2) among those in paid work – main occupation, length of service, pattern of work (e.g. salaried vs. piece work, permanent vs. temporary), employer’s size, physical and psychosocial working conditions (e.g. kneeling, climbing, digging, heavy lifting, standing, shift and night work, work demands and support), job satisfaction, conflicts at work, job security, income protection in illness; self-reported ability to cope with work demands); 3) financial status (e.g. contribution to total household income, housing tenure, affordability of consumer durables, pension provision; 4) attitudes to work and retirement – how long the person would like to work, how long they need to work, their intended retirement age; 5) demographic, social and anthropometric data – education and qualifications, marital status, dependants and caring commitments, household composition, height and weight; 6) leisure and social activities; 7) smoking and alcohol history; 8) selected health items: sickness absence in past 12 months; regional pains in past 12 months; Self-Rated Health (SRH); abridged Sleep Problems Scale; Brief Symptom Inventory (BSI) somatising scale; Center for Epidemiologic Studies Depression Scale (CES-D);and the Warwick-Edinburgh Mental Well-being Scale (MWBS); brief items on items on frailty (Fried frailty index) and cognition. The follow-up questionnaire assessed changes from baseline in: 1) job circumstances, with reasons (job loss, new job, job modification, for health-related or other reasons); 2) health (e.g. hospital referrals, new diagnoses, new treatments, new workplace injuries, changes in SRH, BSI, CES-D, MWBS, frailty, cognition); and 3) attitudes towards retirement (including those modified by spouse’s health and employment). The CPRD provides the main source of information on health at baseline and over follow-up.