Sampling and site access

We used a multi-method case study design (Eisenhardt, 1989; Yin, 1994, 2003). We selected eight case study sites, using the preliminary information that we had gathered in preparation for the proposal. The selection of sites was also informed by meetings with the director and members of the NHS Modernisation Agency team responsible for the treatment centre (TC) programme. The sampling was intended to ensure that the case study sites provided a broad representation of the range of TCs either existing or in development as at March 2003 when the research began.

The selection characteristics that we considered were:

• • geographical (for example, urban/rural);

• • type of host trust;

• • organisational (for example, integral or separate from host trust; NHS ‘star rating’ of host trust; likelihood of foundation hospital status);

• • intended case mix (such as single or multiple specialty, routine or more complex cases);

• • the stage of development (from those that were already open, through to those in the early planning stage);

• • scale (as measured in terms of planned numbers of patients);

• • new/purpose-built or not;

• • degree of private sector involvement;

• • commissioning model (such as whether reliant on multiple or single commissioners).

Ethical approval for the study was sought from a Multi-centre Research Ethics Committee in January 2003, and full approval was granted on 14 April 2003 (MREC/03/07). Management approval for the study from the relevant NHS trust chief executive or TC director in each of the sites was then obtained.

All but one of our original choices agreed to be case study sites (see Table A2.1). The sites were given an information sheet about the study. When they had accepted, a local ‘site representative’ was appointed from among the senior staff associated with the TC. With the help of this person, key initial informants in each site were selected for interview, based on their roles and involvement with the TC.

The sites selected ranged from relatively small initiatives (a single ward, Phase I of Lakenfield) to much larger enterprises including centres that operated essentially as a mini-hospital (Ruckworth and Robbleswade). Some were new purpose-built facilities (Robbleswade, Stanwick, Northendon, Pollhaven and Phase II of Lakenfield), some were refurbishments of facilities within the host organisation – usually a hospital (Brindlesham and St Urban’s) – and one (Ruckworth) was a separate leased building.