Abstract
Background:
Early and accurate diagnosis of dementia opens the door to appropriate treatment, support, and counseling. Despite availability of evidence-based guidelines for diagnostic evaluation of dementia, the diagnostic rate in people with dementia is low and the quality of dementia diagnoses is unknown.
Objective:
The overall aim of this register-based study was to analyze the quality of diagnostic evaluation of dementia by assessing nationwide geographical variations in a range of indicators.
Methods:
A register-based cross-sectional study of the entire Danish population aged 65 years or older in 2015 was conducted. The surrogate indicators for diagnostic quality included 1) prevalence rates of dementia diagnoses, 2) incidence rates of dementia diagnoses, 3) age at first diagnosis of dementia, 4) medical specialty responsible for diagnosis, 5) diagnostic rate of dementia subtypes, and 6) use of anti-dementia medication. The indicators were compared across the five Danish regions.
Results:
The national prevalence and incidence of registered dementia diagnoses was 3.0% and 0.5%, respectively. The proportion of patients diagnosed at a dementia specialist department ranged from 60.9% to 90.5% across the five regions, subtype specific diagnosis ranged from 45.3% to 75.5%, and use of anti-dementia medication ranged from 29.2% to 58.3%.
Conclusion:
The observed geographical variations in dementia diagnoses and treatment indicate inequality in the access to appropriate diagnostic evaluation and care for patients with dementia. Our findings call for more awareness of the benefits of timely diagnosis and for improvement in the quality of diagnostic evaluation of dementia.
INTRODUCTION
Worldwide, the prevalence of dementia is expected to approximately double within the next 20 years [1]. Early and accurate diagnosis of dementia paves the way for specific treatment as well as timely support and counseling. National and international evidence-based guidelines on detection, diagnosis, and management of dementia emphasize the importance of a timely diagnosis, including a diagnosis of subtype. Knowledge of the specific diagnosis of dementia subtype such as Alzheimer’s disease, vascular dementia, or frontotemporal dementia is essential for the planning of post-diagnostic management, including specific treatment and care, as well as to identify other treatable conditions [2–4].
Despite availability and dissemination of evidence-based guidelines for diagnostic evaluation in Denmark, dementia is still underdiagnosed in the elderly [5–7], and previous studies have indicated that the quality of dementia diagnostic evaluation was not up to standard [8–10]. Internationally, studies in other countries have consistently found a low diagnostic rate for dementia (39–54.5%) [5, 11]. Furthermore, a German study showed that 46% were registered with unspecified dementia by specialists (neurologists/psychiatrists) [11]. In Denmark, universal health care guarantees free access to public health care services. Diagnostic evaluation of dementia may be initiated by the general practitioner, who will refer to memory clinics based in secondary health care, when a dementia disorder is suspected. Other entry points can be admission to the hospital due to acute illness, being seen in the emergency room or in specialist clinics for other health conditions, where cognitive issue get detected and referred to memory clinics for further investigations.
Danish national registers are unique resources as they provide nationwide coverage and comprehensive information on health care [12]. We hypothesize that there may be a pronounced geographical variation in Danish regions regarding diagnostic rate of dementia and the quality of dementia diagnosis. In a country with universal health care, geographical variation may reflect differences in competencies and attitude toward diagnosing dementia. Knowledge about geographical differences in the Danish population may provide further insight into the extent and nature of variations in clinical practice in the quality of diagnostic evaluation of dementia in other countries with a similar national health care system. Therefore, the overall aim of the present study was to analyze the quality of diagnostic evaluation of dementia in the secondary health care sector, by assessing nationwide geographical variations in a range of predefined indicators for diagnostic quality.
MATERIALS AND METHODS
Study design and ethics
The study was designed as an observational cross-sectional study using data from nationwide Danish registers.
The study was approved by the Danish Data Protection Agency (ID no: 2007-58-0015/30-0667), Statistics Denmark, and the Danish Health and Medicine Authority (ID no: 6-8011-907/1). All data were anonymized and Danish law did not require ethic committee approval or informed patient consent for this type of register-based study.
The registers
All permanent Danish residents are assigned a personal civil registration number at the time of birth or immigration [13], which allows for retrieval of demographic and medical data at an individual level in the nationwide registers [12]. The entire Danish population’s contacts within the secondary health care system have been recorded in the two national hospital registers: The Psychiatric Central Research Register for psychiatric contacts since 1969 and the National Patient Register for somatic contacts since 1977 [14, 15]. Information comprises dates and discharge diagnosis, which are registered according to World Health Organization’s (WHO) International Classification of Diseases (ICD) codes. ICD-8 was used from 1970 to 1993 and ICD-10 from 1994 and onwards.
The Danish National Prescription Register has recorded individual-level data on all prescription drugs sold in pharmacies since 1994 [16]. The anti-dementia drugs available in Denmark are donepezil, rivastigmine, galantamine, and memantine.
Demographical information
Denmark is divided into five regions and subdivided into 98 municipalities. The regions are responsible for the hospitals, the primary care general practitioners (GPs) and specialists. The municipalities are local administrative bodies responsible for primary health care including home care and nursing homes. The Danish Civil Registration System provides information about place of residence, age, sex, and marital status, while Statistics Denmark provides information about living status.
Study population
All Danish residents aged 65 years and older and alive on January 1, 2016 were identified using the Central Population Registry. Dementia cases were identified as those who had been registered with a dementia diagnosis specified as either Alzheimer’s disease, vascular dementia, frontotemporal dementia, other (primary) dementias, or dementia without specification in the National Patient Register or Psychiatric Central Research Register before January 1, 2016, during a hospital-based inpatient admission or at an outpatient visit. Patients diagnosed with dementia in primary care are not included. Patients registered with secondary dementias (a diagnosis of “dementia in other diseases classified elsewhere”) were not included in the study population. The place of residence (region) was identified through the Danish Civil Registration System. Since the validity of the dementia diagnosis has been shown to be low in patients diagnosed earlier than the age of 60 [9], individuals registered with a dementia diagnosis at the age of 60 years old or younger were excluded.
Indicators for diagnostic quality
Patients registered with a dementia diagnosis were classified into prevalent and incident cases. The prevalent cases were individuals who had been registered with a dementia diagnosis any time before January 1, 2016, whereas the incident cases were individuals who had been registered with their first dementia diagnosis from January 1 to December 31, 2015. The following proxy indicators for diagnostic quality were applied: 1) age and sex-standardized prevalence rate of registered dementia diagnoses in the elderly population, 2) age and sex-standardized incidence rate of registered dementia diagnoses in the elderly population, 3) age at first diagnosis of dementia in the incident cases, 4) proportion of incident cases diagnosed at a dementia specialist department, 5) proportion of incident cases with a subtype specific diagnosis, and 6) proportion of incident cases using anti-dementia medication. Patients registered with a dementia diagnosis were classified according to medical specialty of the department at the index contact: departments with dementia specialists (psychiatry, neurology, geriatrics) and other departments (internal medicine, surgery, and other specialties). Using ICD-10 codes for Alzheimer’s disease, vascular dementia, frontotemporal dementia, other (primary) dementias, and dementia without specification, the study population was divided into patients registered with a specific (Alzheimer’s disease, vascular dementia, frontotemporal dementia or other (primary) dementias) versus an unspecific dementia diagnosis (Supplementary Table 1). Users of anti-dementia medication were defined as incident cases redeeming at least one anti-dementia prescription (ATC: N06D) within one year after the date of first dementia diagnosis. This proxy indicator was used as it may reflect the underlying considerations concerning dementia subtype among medical doctors.
An additional analysis was completed to investigate the proportion of patients registered with an unspecific dementia diagnosis from January 1 to December 31, 2014, who had their diagnosis converted to a specific diagnosis within one year after the date of first diagnosis.
Statistical analysis
Normality of continuous variables was graphically determined using histograms. Differences between groups in the five regions were compared using Pearson’s χ2-test or ANOVA, as appropriate. The prevalence and incidence rates of dementia in the regions were adjusted using direct standardization to account for differences in age and sex between geographical areas [17]. A p-value of <0.05 was considered statistically significant. The data analysis was performed using SAS statistical software, version 9.3 (SAS Institute Inc., Cary, NC, USA).
RESULTS
Study population
The study population consisted of 1,079,358 persons aged 65 years and older (Fig. 1). On January 1, 2016, Denmark had 34,040 residents aged 65 years and older registered with a dementia diagnosis. We excluded 2,340 individuals because they had received a diagnosis of dementia before the age of 60 years and excluded another 18 individuals due to missing information on place of residence. This resulted in a study population of 31,672 patients registered with a dementia diagnosis, 5,516 of whom had been diagnosed in 2015. Table 1 presents the characteristics of elderly patients with dementia stratified by region. The proportion of nursing home residents was 48.0% in prevalent cases and 28.4% in incident cases, and highest in the Northern Denmark Region.
Selection of the study population.
Baseline characteristics of prevalent cases with dementia by region in 2015
Numbers are given as percentages and mean (standard deviation).
Baseline characteristics of incident cases with dementia by region in 2015
Numbers are given as percentages and mean (standard deviation).
Indicators for diagnostic quality
Table 2 shows the 2015 indicators for diagnostic quality stratified by region. The age and sex-standardized national prevalence rate of registered dementia diagnoses was 3.0%, ranging from 2.5% to 3.6% in the five regions (p < 0.0001). The age and sex-standardized national incidence of registered dementia diagnoses was 0.5%, ranging across the regions from 0.4% to 0.7% (p < 0.0001). For incident cases diagnosed in 2015, the national mean age at first dementia diagnosis was 81.8 years (SD 7.3) (Table 2). The proportion of patients who were diagnosed at a specialist department varied from 60.9% to 90.5% (p < 0.0001) in 2015. The proportion of patients who were registered with a specific dementia diagnosis at their first diagnosis varied from 45.3% to 75.5% in the regions (p < 0.0001). Additionally, the proportion of patients who were diagnosed with an unspecific dementia diagnosis in 2014 and then received a specific diagnosis within the following 12 months was 3.3% and varied from 2.3% to 4.0% across the five regions (p < 0.0001). The proportion of incident cases who used anti-dementia medication varied from 29.2% to 58.3% in the regions (p < 0.0001).
Indicators for diagnostic quality by region in 2015
1Age and sex-standardized. 2Incident population. 3First diagnosis in 2014. 4Within one year after first dementia diagnosis. 5Data for proportion of dementia patients diagnosed with a specific diagnosis after referral to memory clinics (2016 data from Quality registry) are 94% (total), 96% (North Denmark Region), 96% (Central Denmark Region), 93% (Southern Region of Denmark), 93% (Capital Region of Denmark) and 94% (Zealand Region). Numbers are given as percentages and mean (standard deviation).
DISCUSSION
This study investigated nationwide data to assess the quality of diagnostic evaluation of dementia in 2015. We found significant geographical variations in all predefined indicators for diagnostic quality. These variations may indicate inequality in access to appropriate diagnostic evaluation and post-diagnostic management.
Underdiagnosis of dementia leads to lost opportunities for treatment and counseling and may increase the burden for patients and caregivers. In the present study, the age and sex-standardized prevalence rate of dementia varied from 2.5% to 3.6% in the five regions, while the age and sex-standardized incidence rate of dementia varied from 0.4% to 0.7%. Our findings indicate potential underdiagnosis in certain geographical areas, which may reflect limited access to appropriate diagnostic work-up or inadequate quality of post-diagnostic management in some regions. Based on a Danish population study, the Odense study from 2003 [18], it was estimated that there were approximately 80,739 people aged 65 years and older living with dementia cases in Denmark in 2013 [19], but in our register-based study only 31,672 people aged 65 and older were registered with a dementia diagnosis from secondary health care sector in 2015. Consequently, our findings suggest a low diagnostic rate of dementia at about 30%, although cases diagnosed solely in GP offices are not accounted for in our register-based study. Very fragil patients or patients with advanced dementia may not be referred by their GP for further investigation at a specialist department. This figure corresponds well with an estimate from England that only about one third of persons with dementia were diagnosed [20]. In 2009, the figure has increased after a nationwide action plan. A low diagnostic rate of dementia may reflect a lack of awareness of dementia among health care professionals in hospitals and municipalities. Therefore, our results emphasize the need for action to enhance awareness of dementia to provide equal and high health care quality regardless of geography.
Our register-based data did not contain data on severity of dementia at diagnosis, thus age at diagnosis may serve as a surrogate marker. A previous Danish study based on data from 2003 investigated some of the indicators for diagnostic quality of dementia, without analyzing geographical variations [8]. In the previous study, mean age at the time of first dementia diagnosis was 81.0 years, which is consistent with the finding of 81.8 years in the present study. Similarly, a Swedish study found a mean age at diagnosis of 79.3 years in patients in the primary and secondary health care sector during 2007–2012 [21].
Specialists in geriatrics, neurology and old-age psychiatry are usually considered to be dementia specialists, despite differences in the organization of health care systems across countries [22]. These specialists are highly trained in dementia diagnosis and management in Denmark, therefore they can diagnose dementia with a higher accuracy compared to other medical specialists. A study of dementia care in the US found that neurologists adhered significantly more to evidence-based guidelines for diagnostic evaluation and treatment of dementia compared to GPs and internists, indicating a higher quality in dementia diagnostics and treatment [23]. In 2002, a Danish study reported that 73.6% of dementia diagnoses were made by dementia specialist departments (neurology, psychiatry, and geriatrics) [8], which is compatible with our findings (73.5%). However, it is surprising that this indicator did not improve after more than a decade. Our finding that the proportion of diagnoses registered at dementia specialist departments varied across regions (from 60.9% to 90.5%) may reflect differences in organization of health care and access to diagnostic evaluation.
Diagnostic evaluation of dementia is most often initiated in the primary care setting, where the driving force for a diagnosis is a GP’s suspicion based on patient symptoms or concerns of caregivers [24]. However, several studies have shown that GPs doubt the usefulness of early diagnosis and perceived treatment options as limited [25, 26], which may negatively affect their motivation to initiate a diagnostic evaluation. When GPs suspect dementia, patients may be referred to specialists (memory clinics) for more detailed diagnostic evaluation. In Denmark, memory clinics are organized mainly within the secondary health care system in departments of geriatrics, neurology, or psychiatry. However, patients may also be registered with dementia diagnoses in other hospital departments and without a prior referral for diagnostic evaluation of dementia.
Diagnosis of the subtype of dementia is important given differences in disease course and outcome for different dementia subtypes. Furthermore, without establishing a specific underlying condition, symptoms due to other reversible conditions may be misdiagnosed. Moreover, the strategies for pharmaceutical treatment, care, counseling and support vary according to the specific subtypes. Therefore, a specific diagnosis of dementia can serve as a proxy indicator of good-quality diagnostic evaluation and treatment. A previous Danish study based on data from 2003 found that 44.7% of patients were registered with a subtype specific dementia diagnosis [8], while the proportion was 60.3% in the present study, demonstrating a significant improvement in the last decade. However, the proportion of specific dementia diagnoses is still alarmingly low, especially considering that the geographical differences in diagnostic rates of dementia subtypes ranged from 45.3% to 75.5% across regions. The fact that 3.3% of all patients with unspecific dementia diagnoses in 2014 were rediagnosed with a specific diagnosis within the following 12 months supports the notion that the majority of patients registered with a diagnosis of unspecific dementia were not offered any diagnostic evaluation. Most likely, the high number of unspecific dementia diagnoses may be registered in hospital departments during contacts unrelated to dementia and outside the context of memory clinics. The National Danish Quality Register has documented all diagnostic evaluation performed in memory clinics in 2016, of which 94% were specific subtype diagnoses [27].
The use of anti-dementia medication varied considerably between regions, reflecting the variations in rates of specific diagnoses. Without a specific diagnosis of either Alzheimer’s disease, dementia with Lewy bodies, or Parkinson’s disease with dementia, patients will not have access to appropriate treatment. In Denmark, the diagnosis of a dementia disorder should be established by a specialist in order for the patient to get reimbursed. The proportions of patients who receive anti-dementia medication vary across Europe [28], which may be explained by the differences in prescribing practices and reimbursement policies.
We can only speculate about the reasons for the geographical variations in the quality of diagnostic evaluation of dementia. The barriers to dementia diagnoses and treatment of high quality are manifold. Several factors may play a role, including differences in policy and practice in primary health care for detection of dementia and in access to diagnostic evaluation. Furthermore, differences in knowledge or attitude about the benefits of early and accurate diagnosis of dementia as well as treatment of dementia, may lead to differences in adherence to evidence-based guidelines for diagnostic evaluation of dementia. Finally, lack of access to appropriate post-diagnostic support may reduce diagnostic rates.
Ideally, the quality of the dementia diagnoses registered in the national hospital registers in 2015 should be evaluated through medical chart reviews and clinical examination of the patients. However, such validation would not be feasible in our nationwide study due to limitations in access to patient records. Instead, proxy indicators were used to examine the quality of diagnosing dementia in the secondary health care sector nationwide. The fact that our registers only contain administrative data from ICD-10 coding in the secondary health care sector is a limitation, which may lead to underestimation of the prevalence and incidence rates and to a skewed representation of the quality of diagnostic evaluation. However, the Danish health care system is organized with the primary health care’s GPs as gatekeepers, initiating the diagnostic evaluation and referring to specialists when needed. Another limitation of our study was that it was not possible to identify the subgroup of patients diagnosed in memory clinics, where the quality of diagnostic evaluation presumably would be higher.
One of the strengths of this study is its nationwide population-based design and high reliability of data linked by a unique personal identification number, allowing investigation of all registered dementia diagnoses in the secondary health care sector, thus avoiding problems of selection bias. Data on prescriptions for anti-dementia drugs were complete and included nursing home patients, although we did not have information about indications and whether patients actually consumed their prescribed medicine.
Transparency of geographical data on diagnostic quality, as presented here, may help health care providers initiate appropriate actions as has been seen in the UK dementia strategy [20]. In 2013, the UK government published the Dementia State of the Nation interactive maps, which allowed the public to perceive how local dementia services in their geographical area performed and to view the performance of dementia services across the country, which contributed to improvement in diagnostic rates [29].
However, transparency is not sufficient; other initiatives must also be launched to deliver sustained health care improvements, including education and implementation of evidence-based guidelines and organization of health care, not only to GPs but to secondary health care professionals as well.
The timeliness and accuracy of dementia diagnosis are increasingly relevant given the rapidly aging population, with elderly aged 85 years and older making up the fastest-growing proportion of many nationalities [30]. With age as a major contributor, the incidence of dementia will be expected to make a steep increase, turning poor quality of the diagnostic evaluation and treatment into a greater public health burden over time. Given the present emphasis on improving health care quality, we believe our findings are timely, also in an international context.
Conclusion
The observed geographical variation in registration of dementia diagnoses and diagnostic quality indicators may point to gaps in the access to appropriate diagnostic evaluation and care for patients with dementia. In the next decade, there will be an increasing number of people with dementia and possibly novel diagnostic options and treatments which will place new demands on clinicians such as a timely and specific diagnosis of dementia of high quality. Our findings call for more awareness of the benefits of timely diagnosis and for improvement in the quality of diagnostic evaluation of dementia.
Footnotes
ACKNOWLEDGMENTS
This Danish Dementia Research Centre is supported by the Danish Ministry of Health (file no. 2007-12143-112/59506 and file no. 0901110 /34501). All researchers were independent of the funders. We are grateful for Peter Johannsen, MD, PhD, for valuable support and advise.
