Table 1 Background information about the ARCH Corpus database [27]

This Corpus houses a digitally stored collection of patient / practitioner consultation data that includes video-recorded consultations and verbatim transcripts.

Included data has been collected since 2003 as part of funded studies.

These studies include the:

 - Interaction Study (IS) exploring clinical decision-making when rationing is explicit;

 - Tracking Study (TS) exploring communication processes throughout a single complete episode of care of patients referred from primary to secondary care;

 - Diabetes Study (DS) tracking the contact of newly diagnosed patients with Type 2 diabetes with healthcare professionals over a six-month period.

The combined dataset for the IS and TS comprised an unselected sample of 183 routine consultations involving 15 general practitioners recorded in 21 participating general practices in the Wellington region of New Zealand between 2004 and 2007. The data for the DS was purposively collected in Wellington and Auckland; for this study, 36 patients newly diagnosed with type 2 diabetes were recruited prospectively via 21 general practices (6 of which had participated in the previous studies) and tracked for a period of 6 months between 2008 and 2011. Their consultations with general practitioners and nurses at the general practice and related consultations with allied health professional staff were video-recorded.

All consultations were made in the course of ‘practice as usual’ and are therefore typical of routine interactions occurring in New Zealand healthcare.

Written consent was sought from participants in these original studies to use their data for secondary analysis.

At the time of sampling the ARCH Corpus included 418 video-recorded consultations recorded between 2003 and 2011, 337 of which were eligible for inclusion. These comprised 247 individual patients, 30 general practitioners, 31 nurses and 15 other practitioners.