The number of older people in the Netherlands with a migration background is increasing, especially in the three biggest cities. Older non-western men and women have poorer physical and psychosocial health than Dutch peers in a comparable socioeconomic situation, and these differences are expected to be more distinct in frail older people. Both dementia and diabetes are one of the most frequent problems of old age, especially among the frail elderly, and these conditions are interrelated. The prevalence of diabetes and dementia will increase substantially in the coming years. With the ageing population, the relative increase of older migrants and the relatively unfavourable health of this group, the amount of primary care that is needed for this group is likely to increase. However little is known about the care-specific aspects of these two conditions in people with a non-western migrant background.
This study aims to identify areas of improvement for appropriate health care use and the quality of care for older people with dementia and/or diabetes of non-western migrants in a mixed-methods approach. First, we shall compare the health care use and quality of care between non-western migrants and people of Dutch descent in various health care registries. We shall also compare people living in a big city and more rural regions.
For the quantitative analysis we will use four extensive databases which contain routine health care (administrative) information. For that, the PHARMO database, the DEPEND database, the database of the Diabetes Care System in West Friesland and the database of the Academic GP network of Amsterdam UMC will be linked to each other were applicable, and to the database of StatLine of Statistics Netherlands (CBS) to obtain information regarding ethnicity, social economic status and whether the patients are living in a more rural, or urban setting.
We shall compare non-western migrants and people of Dutch descent of age 65 and older who have been coded to have dementia or diabetes by their GP using the international classification of primary care (ICPC: P70 or T92). The primary outcomes are a set of health care use parameters and quality indicators (QIs) that we selected from the available literature and that are available in the registries. As general chronic health care use parameters we will use the number of primary out-of-hours care visits, the number of acute hospitalisation, time from diagnosis to institutionalisation, time from diagnosis to death and the proportion of patients using antipsychotics. Main QI ’s for general elderly care will be: the proportion of people with potentially inappropriate prescribing according to the Fick criteria, time from diagnosis to death, use of antipsychotics, the proportion of people who do not receive a yearly influenza vaccination and the Herfindahl-Hirschman Index as a measure for continuity of care. Main QI’s for dementia care will be: the proportion of patient taking neuroleptic medication without ICPC codes that might explain the use of neuroleptic drugs. Also, some of the previously mentioned health care indicators will be analysed as proxies for quality of dementia care: the number of primary out-of-hours care visits, number of acute hospitalisations. Foremost QI’s for diabetes are the proportion of people with annually measured blood pressure, the proportion of people with an annually registered foot examination, the proportion of people with an annually measured kidney function and the proportion of people that have a retinal examination.
We will provide descriptive statistics for each database. We will compare differences for people between non-western migrant people and native Dutch people with dementia and diabetes using different statistical methods depending on the type of outcome measures. Linear regression or linear mixed models or if needed Poisson regression for the continuous or count variables; survival analysis (with competing-risk analysis if needed) for time to event parameters and logistic regression or multilevel models for dichotomous measures. Potential confounders, including age, number of chronic diseases, sex, and household income, will be analysed and added to the models when appropriate. Subgroup analyses for specific migration backgrounds will be conducted if the sample size is sufficient.
Secondly in the qualitative part of this study, we shall formulate improvement suggestions for general practice based on the discussion of the results of the first part of our study with both primary care professionals (GP’s and practise nurses). We shall do this also with a panel of experts with lived migrant experiences who represent the patient groups aiming for diversity in both ethnicity and gender of the panel members.
We aim to disseminate the insights and results of this project widely to GPs, GP trainees, patients organisations, migrant organisations and healthcare policymakers.