No abstract

BackgroundThe use of full-population databases is under-explored to study the use, quality and costs of end-of-life care. Using the case of Belgium, we explored: (1) which full-population databases provide valid information about end-of-life care, (2) what procedures are there to use these databases, and (3) what is needed to integrate separate databases.MethodsTechnical and privacy-related aspects of linking and accessing Belgian administrative databases and disease registries were assessed in cooperation with the database administrators and privacy commission bodies. For all relevant databases, we followed procedures in cooperation with database administrators to link the databases and to access the data.ResultsWe identified several databases as fitting for end-of-life care research in Belgium: the InterMutualistic Agency's national registry of health care claims data, the Belgian Cancer Registry including data on incidence of cancer, and databases administrated by Statistics Belgium including data from the death certificate database, the socio-economic survey and fiscal data. To obtain access to the data, approval was required from all database administrators, supervisory bodies and two separate national privacy bodies. Two Trusted Third Parties linked the databases via a deterministic matching procedure using multiple encrypted social security numbers.ConclusionIn this article we describe how various routinely collected population-level databases and disease registries can be accessed and linked to study patterns in the use, quality and costs of end-of-life care in the full population and in specific diagnostic groups.Electronic supplementary materialThe online version of this article (doi:10.1186/s12904-016-0159-7) contains supplementary material, which is available to authorized users.

The number of packages is a more appropriate measure than the number of DDDs when assessing outpatient antibiotic use over time and the impact of awareness campaigns in countries dispensing 'complete packages'. We recommended the use of different complementary measures or caution when interpreting trends based only on DDDs.

End-of-life cancer care has been criticized as frequently inappropriate and aggressive. Providing appropriate care to people with cancer is a public health priority. Quality indicators are considered a valid way to evaluate the appropriateness of end-of-life cancer care within a health care system. We conducted a population-level retrospective observational study of all cancer decedents in Belgium in 2012 to assess end-of-life care and risk factors for exposure to care. We linked eight full-population databases on health care use, cancer diagnoses, and demographic and socioeconomic variables. We used analysis of variance to examine factors associated with exposure to appropriate or inappropriate end-of-life cancer care. Of the 26,464 people in Belgium who died from cancer in 2012, 47 percent received specialist palliative care, and 30 percent died at home. In the last thirty days of life, 17 percent received chemotherapy, and 66 percent received diagnostic testing. For 17 percent, palliative care was initiated only in the last fourteen days of life. Our results suggest a need to focus policy on reducing aggressive and inappropriate care at the end of life and an opportunity to increase the proportion of people who receive specialist palliative care and die at home.

Our study found indications of inappropriate end-of-life care in people with COPD, such as high percentages of diagnostic testing and hospital admissions and low proportions receiving specialized palliative care. Risk-adjusted variation between regions was high for several QIs, indicating the usefulness of relative performance standards to improve quality of end-of-life COPD care.