Development of memory clinics in the Netherlands over the last 20 years

Objectives Memory clinics (MCs) have been established to improve diagnosis and treatment of cognitive disorders, including dementia. The aim of this study was to determine the characteristics and working methods of MCs in the Netherlands in 2016. More insight into different working methods can be used to improve the quality of care in Dutch MCs. Additionally, the findings will be compared with earlier results to investigate the development of MCs since 1998. Methods A survey was sent in 1998, 2004, 2009, and 2017 to all operational Dutch MCs with questions about organization, collaboration, patients, and diagnostic procedures. Results From 1998 to 2016, the number of MCs increased substantially from 12 to 91. The capacity increased from 1560 patients to 24,388. In 1998, most patients received a dementia diagnosis (85%), while in 2016, half of the patients were diagnosed with milder cognitive problems. MCs are more often part of regional care chains and are better embedded within regional care organizations. Diagnostic tools, such as blood tests (97%), neuropsychological assessment (NPA) (95%), and neuroimaging (92%), were used in nearly all MCs. The number of patients in whom these tools were used differed greatly between MCs (NPA: 5%‐100%, neuroimaging: 10%‐100%, and CSF: 0.5%‐80%). There was an increase in the use of NPA, while the use of neuroimaging, CSF, and EEG/ECG decreased by 8% to 15% since 2009. Conclusions Since 1998, MCs have developed substantially and outgrown the primarily research‐based university settings. They are now accepted as regular care facilities for people with cognitive problems.


| INTRODUCTION
Early diagnosis of dementia enables improved understanding of the disease process, provides the opportunity to make decisions concerning the future while cognitive capacities are still relatively intact, 1 and creates a time window to institute early interventions and support for patients and caregivers. 2,3 With no cure available, early diagnosis may also have some drawbacks. An increasingly widespread view acknowledges "timely diagnosis" as a more appropriate concept because it emphasizes a person-centered approach in which the diagnosis is related to the benefit of the patient and not to a disease stage. 4,5 To improve early yet timely diagnosis, the development of specialized multidisciplinary memory clinics (MCs) has been promoted and recommended by national dementia strategies. [6][7][8][9][10][11][12] The first clinics were established during the 1970s in the USA. During this time period, the perspective on the cause of dementia shifted from being an inevitable result of aging to being a disease. 13 The past three decades have shown significant growth in the development of MCs worldwide. The increasing number of MCs has been explained by the increasing prevalence of dementia, licensing of pharmacological treatments, and the improvement in care services. [14][15][16] A comparable increase in MCs can be seen in the Netherlands. To gain more insight into the development and efficacy of these clinics, a first national survey in the Netherlands, the MC Monitor, was published in 1998. 17 20 The aim of this study was to determine the characteristics of MCs in the Netherlands in 2016. Two key topics of the survey were neuropsychological assessment and regional collaboration. More insight into different working methods can be used to improve the quality of care in MCs. In addition, the results were compared with the findings of the previous surveys to investigate the development of MCs in the last 20 years.

| METHODS
To gain more insight into the characteristics of MCs, a semistructured questionnaire was sent out in 1998,2004,2009, and 2017 to all hospital-based MCs in the Netherlands asking for data from the previous year. An MC was defined as a multidisciplinary team with at least two disciplines (at least one medical profession) dedicated to the diagnosis of dementia. All relevant operational clinics were identified using the network of the Alzheimer Center Limburg and through internet searches. Every survey consisted of core items that were repeated in each survey. In addition, relevant items were added over time by an expert group. In addition, the survey was piloted in three academic hospital-based Dutch Alzheimer Centers. In this survey, the following topics were included: organization, collaboration, number of patients, distribution of diagnosis and etiology, referrals, procedures, diagnostic criteria, additional assessments, neuropsychological assessment, treatment, policy, and professionalization. Participants were asked to answer questions by using information derived from official sources, but if this was unavailable, they were allowed to use estimations. In   Table 1). Since 1998, the total number of new patients increased from 1,560 to 24,388. The variation between MCs was large, with the total number of new patients ranging between 30 and 1,000 patients per center. The mean percentage of patients younger than 65 years was 20.6% ± 20.2 (range: 0%-75%). University-based MCs and MCs coordinated by neurology have seen more younger patients on average than non-academic MCs (35% versus 18%, p = 0.02) and MCs coordinated by clinical geriatrics (35% versus 8%, p < .001).

Key points
• Memory clinics in the Netherlands developed considerably in the last 20 years and are now accepted as regular care facilities for people living with cognitive problems and dementia.
• People with cognitive problems attend the memory clinic in an earlier disease stage.
• Memory clinics in the Netherlands are increasingly collaborating with other regional care facilities and have psychosocial interventions more often as part of their treatment options.

| Cognitive screening
In 2016, most MCs used a cognitive screening test during their intake (86%). The MMSE 30 was most frequently used (91%). The outcome of this test was most often used to determine additional diagnostic assessments (78%) or to determine the treatment plan (52%).

| Neuropsychological assessment
The use of an NPA increased from 50% in 1998 to 95% in 2016. The proportion of patients in whom an NPA was performed differed largely between MCs (range: 5%-100%). The reasons for carrying out an NPA were: to support the diagnosis (92%), collect differential diagnostic infor-

| Additional assessment tools
In addition to an NPA, laboratory tests (97%) or neuroimaging studies (92%) were frequently used in MCs. Compared with 2009, the number of MCs using neuroimaging decreased from 100% to 92%, and the number of MCs using EEG, ECG, and CSF decreased from 59% to 45%, 74% to 60%, and 79% to 68%, respectively ( Figure 3). There was a large variation between MCs with respect to the assessment tools used, the per-

| Treatment
From 1998 to 2016, pharmacological treatments were offered in approximately 90% of the MCs (Table 1) 3.10 | Regional collaboration Collaboration with regional healthcare organizations increased from 15% in 1998 to 89% in 2016. In 2016, 78% of all MCs were involved in regional care chains (   neurologists and geriatricians are less frequently involved. 14,15,39,40 The number of psychologists employed in MCs is similar to that of other countries, except for New Zealand, where only 14% of the MCs had a psychologist. 40 These cross-national differences may be related to historical disciplinary developments. NPA, lab tests, and brain imaging are the most frequently used diagnostic assessment tools in MCs. This is comparable with MCs in other countries. 14,16 Since 2009, the use of an NPA has further increased, while EEG, ECG, CSF, and brain imaging tools are used by fewer MCs than in 2009 (a change of 8%-15%). This decrease might be related to the new Dutch multidisciplinary diagnostic guideline in which CSF, for example, is not recommended as a standard routine, and neuroimaging is recommended when the etiologic cause is uncertain. 20 Other speculative reasons might be that new MCs are smaller. The proportion of patients in whom these tools were used, however, did not change (CSF, 2009: 12% and 2016: 15%). The increase in the use of an NPA could be related to the beneficial effect of an NPA on patient outcomes (eg, accuracy of diagnosis). 41,42 Although dementia is still the most common syndrome diagnosis in MCs, diagnoses have shifted towards milder cognitive problems. This is in line with the results from the national English Audit. 35,36 This finding and the increased number of newly referred patients to MCs might be explained by the increased awareness of and attention directed towards dementia and early diagnosis in our society. The proportion of incident cases of dementia diagnosed at an MC has increased 10-fold since 1998 from 6% to 58% (approximately 13,000 patients).
Furthermore, the proportion of patients with cognitive impairment without dementia increased from 10% to 25% (approximately 6,000 patients).
This is in line with the global dementia action plan, which stated that by 2025, 50% of the countries should have diagnosed at least 50% of the incident cases of dementia. 43 The timely diagnosis of dementia is not limited to MCs but is also practiced by general practitioners (GPs) or in community mental health institutions. In the previous survey in 2009, mental health institutions were included and appeared to differ greatly from hospital-based settings (eg, fewer disciplines and diagnostic tools available). They also often did not identify themselves as an MC. Therefore, we have focused on the development of hospital-based MCs in this survey. A timely diagnosis of dementia is being promoted worldwide. [44][45][46] The lack of a disease-modifying treatment calls for a careful consider- best practice between MCs should be shared to improve quality of care. Criticisms of MCs have also been described, such as their role in promoting stigma and issues surrounding over-assessment of patients. 50 In contrast, a European study has shown that MCs facilitate early referrals and to some degree battle against stigma. 53 A strength of this current study is the repeated measurement of a comparable survey over a 20-year period. A high response rate was obtained, and we consequently argue that the study is an adequate representation of the current situation of Dutch MCs. Nonetheless, a few drawbacks should be mentioned. Although we made utmost efforts to include every MC in the Netherlands, we may have missed some newly established MCs. In addition, we did not obtain a response from all identified MCs. The estimated numbers should therefore be carefully interpreted. The nonresponding MCs were all non-academic hospitals, but differed in geographical location, size, and coordinating discipline. Another important point is that the results are mainly based on self-reported estimates rather than objective data from registries.

| CONCLUSION
Since 1998, MCs in the Netherlands have shown substantial development in number, geographical distribution, and total capacity.
MCs are no longer isolated, university-based facilities with a strong focus on scientific research. They are now part of the regular care for the timely diagnosis of dementia and milder cognitive disorders and are integrated into regional care chains. Among MCs, a large diversity in specific working methods and diagnostic tools was identified. This diversity should be the focus of future research to increase transparency of the working methods of individual clinics and to harmonize best practices, which will both improve quality of care in Dutch MCs.