OBJECTIVES: Retirement villages are semi-closed communities, access usually being gained via village managers. This paper explores issues recruiting a representative resident cohort, as background to a study of residents, to acquire sociodemographic, health and disability data and trial an intervention designed to improve outcomes. METHODS: We planned approaching all Auckland/Waitematā District villages and, via managers, contacting residents ('letter-drop'; 'door-knocks'). In 'small' villages (n ≤ 60 units), we planned contacting all residents, randomly selecting in 'larger' villages. We excluded those with doubtful or absent legal capacity. RESULTS: We approached managers of 53 of 65 villages. Thirty-four permitted recruitment. Some prohibited 'letter-drops' and/or 'door-knocks'. Hence, we recruited volunteers (23 villages) via meetings, posters, newsletters and word-of-mouth, that is representative sampling obtained from 11/34 villages. We recruited 578 residents (median age = 82 years; 420 = female; 217:361 sampled:volunteers), finding differences in baseline parameters of sampled vs. volunteers. CONCLUSION: Due to organisational/managers' policy, and national legislation restrictions, our sample does not represent our intended population well. Researchers should investigate alternative data sources, for example electoral rolls and censuses.
OBJECTIVES: Retirement villages are semi-closed communities, access usually being gained via village managers. This paper explores issues recruiting a representative resident cohort, as background to a study of residents, to acquire sociodemographic, health and disability data and trial an intervention designed to improve outcomes. METHODS: We planned approaching all Auckland/Waitematā District villages and, via managers, contacting residents ('letter-drop'; 'door-knocks'). In 'small' villages (n ≤ 60 units), we planned contacting all residents, randomly selecting in 'larger' villages. We excluded those with doubtful or absent legal capacity. RESULTS: We approached managers of 53 of 65 villages. Thirty-four permitted recruitment. Some prohibited 'letter-drops' and/or 'door-knocks'. Hence, we recruited volunteers (23 villages) via meetings, posters, newsletters and word-of-mouth, that is representative sampling obtained from 11/34 villages. We recruited 578 residents (median age = 82 years; 420 = female; 217:361 sampled:volunteers), finding differences in baseline parameters of sampled vs. volunteers. CONCLUSION: Due to organisational/managers' policy, and national legislation restrictions, our sample does not represent our intended population well. Researchers should investigate alternative data sources, for example electoral rolls and censuses.