Percentage of cancer patients dying in their usual place of residence (at home or in residential care). Place of death: Dying at home and in residential care: derived from the reimbursement of nomenclature code Dying in the hospital: derived from the hospitalisation information as obtained from the Intermutualistic Agency (IMA) hospitalisation database. Of note, this database as such also describes hospitalisations based on administrative reimbursement data through an internally validated algorithm. Hospitalisations for one day only were not taken into account as hospitalisations

To determine the place where the patient died, the algorithm of the Pallcare project was used. The following order was used to determine the place of death: Date of death = end date/date of medical act ppp End date in hospital on day of death End date in residential care on day of death Home nursing at day of death

Date of death – 1 = end date/date of medical act End date in residential care one day before day of death Home nursing one day before day of death End date in hospital one day before death

Date of death – 2 or 3 = end date/date medical act End date in residential care two or three days before death Home nursing two or three days before death Home nursing during last week before death.

Number of patients who died at home or in residential care.

Number of patients diagnosed with cancer that died within the studied time period.

Patients with more than one invasive tumour (until 2015) Patients without a Belgian residence Patients without national social security number Patients for whom no Intermutualistic Agency (IMA) data of the year of death were available (=3.9%) Patients with an uncertain date of diagnosis Age at diagnosis <18 years

Other restrictions: Maximum 3 years of follow-up was applied. Analyses are limited to patients who died before January 1st of the third year following the year of incidence. For example: patients diagnosed in 2006 were included in case they died in 2006, 2007 or 2008; patients diagnosed in 2007 were included in case they died in 2007, 2008 or 2009, etc. These cohorts were defined in order to assure that for every diagnosis, the same follow-up period could be taken into account. By applying this definition, a bias induced by varying length of the follow-up period could be avoided.

None

Belgian Cancer Registry (BCR) linked with Intermutualistic Agency (IMA) data. The Belgian Cancer Registry is an exhaustive national register of cancer cases. The IMA dataset includes all information as collected by the seven existing health insurance agencies in Belgium. This database covers data on reimbursed medication and reimbursement for the use of health services, collected in the framework of the compulsory Belgian health care and benefits insurance program.

Public reporting.

kce.fgov.be/sites/default/files/2023-06/KCE370PerformanceRevisionframeworkconceptualReport.pdf

www.healthybelgium.be/metadata/hspa/eol4.pdf