Mariken Stegmann

Institutional Review Board of the University Medical Centre Groningen reviewed the  protocol and informed consent was obtained from all interviewees. We report this  research based on the Consolidated Criteria for Reporting Qualitative Research. 22 Data collection of medical files  Medical hospital files were selected by purposive sampling. We aimed to include patients  of different age, gender and tumour stage who were treated by oncology specialists and  GPs of different experience levels, ages and genders.  We extracted all referral letters from medical files, including those sent by mail or fax, and  those sent digitally using ZorgDomein (Care Domain). 23  The ZorgDomein system provides  GPs with a predefined format that automatically imports some key information (e.g.  prescriptions or past history) from the GP’s file into the referral letter. This system has  become widely used in the Netherlands over the last decade. The precise format differs  slightly by the reason for referral and by department, as all formats are composed at a  regional level by collaboration between specialists and GPs. 23 Next, we extracted the specialist letters written from the time of referral to the end of the  initial phase of treatment (which was defined as the first three months after diagnosis and  the start of any other treatment thereafter, if described in the initial treatment plan).  Notes were also extracted regarding outpatient visits, allied health professional (e.g.  psychotherapists, social workers and dieticians) visits, telephone calls and multidisciplinary  meetings. In the multidisciplinary meetings, all patients diagnosed with a specific type of  cancer were discussed by specialists of different relevant departments (e.g. surgeons,  medical oncology specialists, radiologists, pathologists, pulmonologists and radiation  oncology specialists), but without input from GPs. However, sometimes reports of these  meetings were sent to GPs. We excluded documents used to exchange information  between in‐house healthcare providers.  Data collection of interviews  To confirm our data, healthcare providers were recruited for the interviews by purposive  sampling. We aimed to include GPs and oncology specialists from different departments,  age and gender. Discussed topics were the content and quality of any referral letters,  specialist letters, other ways of communication and non‐shared in‐ formation. All  interviews were recorded, five were fully transcribed, and for the others, all relevant  quotations were written down.  Data analysis  Descriptive statistical analyses are given about the number of pages of correspondence and  the time interval between consultation and correspondence. To illustrate the findings of  the medical files and interviews, quotations (Q) are presented and identified by a  department and a document number.  During a pilot study of five files of patients with breast cancer, a preliminary code list was  produced, and checked by two other researchers, which was used when analysing the  patient files. Relevant text fragments were identified and coded, before the coded text  fragments were analysed thematically, all super‐ vised by experienced researchers. 24–26 68 Chapter 6