Introduction

Parkinson’s disease (PD) and atypical parkinsonism (AP) are serious neurodegenerative disorders that affect a significant proportion of the adult population [1–6]. These disorders lead to a deterioration in motor, mental, and functional skills and are associated with significantly increased mortality rates, particularly for patients with AP [7]. These disorders are chronic and progressive, so PD and AP may have serious negative impacts on a patient’s social life [8], family [9, 10], quality of life [11, 12], work [13, 14], and morbidity and mortality [15, 16].

Significant progress has been made in understanding the underlying pathophysiology [17–19], improving diagnostic accuracy [20], and managing these disorders [21–24]. The pathologic disease process is characterized by progressive destruction of multiple brain regions, including the brain stem, the basic forebrain, the extra-pyramidal system, and, in later stages, almost all cerebral regions [25]. Future advances are expected to further our understanding of the brain processes underlying the disease. Great progress has also been made in developing treatment strategies, but, in spite of this, no early intervention is yet possible that reduces the severity of the disease and the associated mortality [26]. Recent studies have identified a reduced quality of life and, thus, a potential socioeconomic impact of the diseases. To date, studies on the economic impact of PD and AP have involved the use of questionnaires with selected patients, the economic burden being estimated from this information [27–32]. The identification or assumptions of costs in these studies have primarily addressed direct costs, whereas indirect costs have not been identified. As such, there has been no specific estimate of disease cost.

In Denmark, it is possible to calculate direct and indirect costs related to diseases because information from public and private hospitals and clinics in the primary and secondary sector (including medication, social, income, and employment data of all patients) is registered in central databases. It is, therefore, possible to estimate the disease effect before, during, and after diagnosis. We conducted the current national population-based study to evaluate the socioeconomic consequences of PD and AP.

Methods

In Denmark, the dates of all hospital contacts are recorded in the National Patient Registry (NPR), which also include primary and secondary diagnoses. The NPR includes administrative information, diagnoses, and diagnostic and treatment procedures using several international classification systems, including the International Classification of Disorders (ICD-10). The NPR is a time-based national database that includes data from all inpatient and outpatient contacts; therefore, the data we used are representative of all patients in Denmark who have received a diagnosis of PD or AP in the primary or secondary sector in both public and private hospitals. As all data are present throughout the observation period we can trace patients retrospectively and prospectively relative to the time of diagnosis. Furthermore, all contacts in the primary sector (general and specialist practice) and the use of medication are coded by The National Health Security and The Danish Medicine Agency, respectively. An overview of the databases is presented in Table 1 There is some underestimation of the patients with parkinsonism, as patients with contact in the primary but not the secondary sector are not coded as having the diagnosis.

Table 1 Sources of information by categories of data (Denmark)
Full size table

The economic consequences of PD and AP were estimated by determining the yearly cost of illness per patient diagnosed with PD (ICD DG209) and AP (ICD DG23.0–9), which includes primarily multiple system atrophy, and comparing that estimate with the cost of healthcare in a matched control group. The health cost was then divided into annual direct and indirect healthcare costs.

Direct costs included costs of hospitalization and outpatient cost weighted by use, according to diagnosis-related groups, and specific outpatient costs—all based on data from the Danish Ministry of Health. The use and costs of drugs were based on data from the National Danish Medicine Agency, which includes the retail price of the drug (and its dispensing costs) multiplied by the number of transactions. The frequencies and costs of consultations with general practitioners and other specialists were based on data from the National Health Security. Social care, nursing home care, cost of transport, etc. are not included in the (direct) cost analyses. The indirect costs, measured from the perspective of societal costs, include those related to a reduced labor supply and to social transfer payments. In Denmark, social transfer payments are income derived from state coffers. These payments include subsistence allowances, pensions, social security, social assistance, public personal support for education, and other payments. Indirect costs were based on income figures from the Coherent Social Statistics database.

Cost-of-illness studies measure the economic burden resulting from disease and illness across a defined population, including direct and indirect costs. Direct costs are the value of resources used in the treatment, care, and rehabilitation of persons with the condition under study. Indirect costs represent the value of economic resources lost because of disease-related work disability and premature mortality. As the patients leave the national data registers at the time of death, the indirect costs estimate reflects only the production loss related to disease-related work disability. It is important to distinguish costs from monetary transfer payments such as disability and welfare payments. Such payments are a transfer of purchasing power to the recipients from the general taxpayer but do not represent net increases in the use of resources and are, therefore, not included in the total cost estimate.

By reviewing the NPR, we identified all patients who received a first diagnosis of PD or AP from 1997–2007. Using data from the Civil Registration System Statistics Denmark (which contains information regarding all social information, income, pensions, etc.), we then randomly selected citizens who were of the same age and sex as the patients. Social compensation was ensured by matching control subjects with patients by civil status and by the area of the country in which they were resident. The ratio of control subjects to patients was 4:1, in order to reduce the variance among controls. Data from patients and matched control subjects that could not be identified in the Coherent Social Statistics database were excluded from the sample. More than 99% of the observations in the two groups were successfully matched. Patients and matched control subjects were followed from the year of diagnosis until 2007. If a patient or control case was not present in the CPR register on the 1st of January each year, they were not included in the dataset for that year. Patients who were not in the CPR register were typically dead, in prison, or had emigrated to another country. Costs were measured on a yearly basis and adjusted to 2005 prices using the health sector price index for health sector costs; the general price index was applied to nonmedical costs. All costs were measured in DKK and converted into Euros (€1: DKK 7.45).

The study was approved by the Danish Data Protection Agency. Data were handled in a manner that did not reveal the identity of any patient or control subject, so neither individual consent nor ethical approval was required. Statistical analysis was performed using SAS 9.1.3 (SAS, Inc., Cary, NC). Statistical significance of the cost estimates was determined for a nonparametric bootstrap analysis [33]. Since some of the samples were small or their distributions were non-normal or both, we use bootstrap to estimate the p values to attain more accurate results, than would be possible using the t test without bootstrapping. We used 100 iterations, which is sufficient for these populations.

For the large samples, the normality assumption will hold due to the central limit theorem, and the p results before and after bootstrap will be the same.

Results

In total, 13,400 patients with PD and 647 patients with AP were identified and subsequently matched with 53,600 and 2,588 control subjects, respectively. The age distribution of the patients and control subjects is shown in Table 2 As expected, the initial diagnosis of PD or AP was made in relatively old patients.

Table 2 Age and sex of patients with Parkinson’s disease or atypical parkinsonism in the year of diagnosis
Full size table

Direct costs: outpatient clinic, hospital, primary care, and drug costs

More patients than control subjects were treated in outpatient clinics, were hospitalized, and had contact with the primary care system. Compared with control subjects, more patients were taking medication and more received public support to pay for their medications (Table 3).

Table 3 Percentage of patients and control receiving health services, their source of income, unemployment, transfer income, and pension or retirement for patients with Parkinson’s disease and atypical parkinsonism
Full size table

Indirect cost, social costs, employment rate, and income

More patients than control subjects received social services. Conversely, fewer patients with PD or AP than control subjects received income from employment (Table 3). Patients with PD or AP had lower employment rates and significantly higher social transfer rates than did control subjects. The employment rate for patients steadily decreased over time after they received a diagnosis of PD or AP, compared with the same period for control subjects (Table 4). A corresponding increase in social transfer expenses took place, whereas the rate of retirement was similar in both groups (not shown).

Table 4 Sources of information and the average annual health cost (in Euros) per person–year by cost categories in Denmark
Full size table

Total health costs per year

The sources of information and the average annual healthcare cost per person–year by cost categories for patients with PD or AP in Denmark, compared with age- and sex-matched control subjects, are presented in Table 4. The sum of direct net healthcare costs (general practitioner services, hospital services, and medication) and indirect costs (loss of labor market income) were €6,500 for patients with PD and €9,294 for patients with AP. Social transfer payments were all significantly higher in patients with PD or AP compared with control subjects.

Influence of age and sex on the employment and direct and indirect costs

The employment rate before and after a diagnosis has been established is shown in Fig 1 for PD and AP patients. This shows that the employment rate is affected up to 8 years before a diagnosis is assigned in PD and AP patients. Consequently, the employment, total direct and indirect costs are affected before and after the diagnosis is established, peaking at the time of diagnosis (Fig. 1, PD and AP, lower trace). This peak in costs is primarily due to increased direct costs.

Fig. 1
figure 1

The proportion of patients with a diagnosis of Parkinson’s disease (PD) or atypical parkinsonism (AP) and respective control subjects who were employed before diagnosis, at diagnosis, and in follow-up years (upper trace). Average direct and indirect costs before, at, and after the establishment of a diagnosis of PD or AP. Eight years before and 8 years after diagnosis versus 8 years before diagnosis and any length of time after. The overall difference is significant (p < 0.001)

Full size image

The relationships of age and sex with direct and indirect costs for both men and women are shown in Fig. 2 for PD. Patients with AP show a similar pattern, which is not illustrated here. Age and sex have pronounced effects on direct costs especially with respect to higher expenses due to hospitalization and medication among younger patients. The indirect costs were significantly higher among the patients with PD/AP for all ages. The patients receive increased pension payments and have lower income levels, especially among the younger and middle-aged patients.

Fig. 2
figure 2figure 2

Direct and indirect cost by 5-year decades for patients with a diagnosis of Parkinson’s disease (PD) or atypical parkinsonism (AP). Data are presented as cost per patient. The data for females (a) and males (b), respectively, are shown. All differences are highly significant (p < 0.001)

Full size image

Discussion

Patients with PD and AP had significantly higher rates of contact with all sectors of the healthcare system: general practice, outpatient clinics, and in-hospital services. Patients had higher rates of medication use and of publicly supported payment for medication. The total expenses were higher in the group of patients with PD or AP than in age- and sex-matched control subjects. Patients with PD and AP also had lower employment rates and significantly more often received welfare payments; employed patients had lower incomes, compared with employed control subjects. The study demonstrates that PD and AP have social and economic consequences several years before their diagnosis is established.

The differences between patients with PD or AP and control subjects were considerable: medication costs were more than three times higher for patients with PD or AP, hospital costs were 3 times more, total healthcare costs were more than double, and the employment rate was more than 30% lower than for control subjects. Employed patients earned only two-thirds of the income of employed control subjects. Patients with PD or AP had much higher social expenditures than did control subjects. These costs are present for all ages, although they tend to be greater among the youngest patients due to increased direct costs and loss of income.

Previous studies regarding the burden of parkinsonism have focused on the direct costs, e.g., hospital services and use of treatment procedures, such as medication, surgical intervention, physiotherapy, etc. Parkinsonism significantly reduces the quality of life of patients, family members, and caregivers [12, 34–38] Several studies have documented the increased direct costs of PD using different models, e.g., quality-of-life estimates [27, 30, 39] or model estimates [40, 41], whereas others have evaluated the effect of pharmacological treatment [29, 31, 42–46] or subthalamic stimulation [28, 47–50]. Still other studies have addressed management issues related to the care of patients with AP and PD, including the effect on caregivers and families [36, 51–56] and the influence on economic and other costs. However, most of these studies have examined the economic effects of PD and AP in a limited fashion, primarily focusing on direct costs, such as reduction in the use of selected types of medication or whether subthalamic stimulation reduces the need for medication. PD and AP also have significant consequences related to work capabilities, pensions, transfer income, and other public support even though these diseases have a late onset [14, 57]; however, the importance of these factors has not previously been estimated.

The impact of PD and AP on work capabilities is considerable. As can be seen in Figs. 1 and 2, even at the time of diagnosis, patients already have significant reductions in their levels of employment. Thus, employment rate is affected before a diagnosis of PD or AP is made, with the rate falling even further after the diagnosis is confirmed. This finding has also been shown in other chronic and progressive disorders, such as narcolepsy [58]. PD and AP patients have a significant influence on social and health cost variables a long time (up to 8 years) before the diagnosis is established. The slight increase in employment rate in AP patients (Fig. 2) is due to the selection mechanism. Several factors may explain this: (1) patients may have (pre-motor?) symptoms long before PD or AP is diagnosed; (2) patients may not be able to seek a pension before they receive a diagnosis, and so remain employed before their diagnosis is made; and, (3) patients may only seek professional help when they have reached the point at which their social lives are impaired and their symptoms lead to exhaustion. Also direct costs due to hospital and health care contacts were affected up to 8 years prior to the diagnosis of PD or AP. This finding fits well with the hypothesis that the patients present symptoms and disease manifestations prior to the motor manifestations, e.g., hyposmia, urological, depressive, REM behavior disorder, which consequently leads to increased health care contacts [25]. A potential consequence of this finding is that although PD and AP may present further disease progression, the effect of the disease is already significant before and at the time of diagnosis. Disease modification intervention may affect motor skills and quality of life, but it is limited in terms of its social consequences. This may explain, in part, why many of the current treatment modalities for PD and AP have only limited effects on direct and indirect costs associated with the diseases. Consequently, if disease modification is to have an effect, earlier disease identification is important, i.e., during the pre-motor phase of disease development.

PD and AP patients tend to have one or several health contacts with the secondary sector, but patients with early or minor symptoms and contacts with the primary sector may be underestimated, as contacts in the primary sector are registered but not with a diagnosis. In general, the database represents an almost complete national patient sample. This is possible because all Danes are registered using social security codes, and data concerning health, medication, social, and employment are recorded. Therefore, all contacts with the primary and secondary sector, including diagnostic and treatment procedures, were included in these costs. Furthermore, all indirect costs, including transfer payment costs, and income levels were included in the analysis. We did not include criteria for any other verification of the diagnoses of PD or AP, nor did we subdivide by AP type. AP represents a wide range of diagnoses with a poor prognosis. This group of disease represent a higher mortality rates, but we did not include loss of labor as this effect is smaller in the age range and difficult to estimate precisely. It is important to stress that the control group is not defined as a group of healthy subjects; they were selected on the basis of age, sex, geography, and social factors. To be included as control subjects, they could not have a diagnosis of PD or AP, but they may very well have had other disorders—the costs of which would have been included in the figures presented in Table 3. The differences observed between patients and control subjects, thus, represent those between people with PD or AP and a random population-based sample.

Consequently, the study presents the objectively determined direct and indirect costs of PD and AP that were evaluated in a national sample, compared with a control group with a similar age, sex, and social composition, over a 10-year period.

In conclusion, AP and PD lead to significantly higher health-related and social transfer costs, and lower levels of employment and income. These effects are present up to 8 years prior to the diagnosis of motor disorders indicating the significant impact of pre-motor symptoms on health and work capabilities. Appropriate treatment must be provided not only to improve quality of life, but also to assist patients in their ability to continue taking part in their family and professional lives. Additional research is needed in early disease identification, disease management, and the effects of PD and AP on quality of life, socioeconomic factors, work capabilities, and healthcare needs so that these costs for patients and society can be reduced.