Direct Medical Costs of Spinal Muscular Atrophy in the Catalonia Region: A Population-Based Analysis

Spinal muscular atrophy (SMA) is a rare disorder, estimated to affect 1 per 10,000 live births. Patients affected with SMA often require intensive, chronic healthcare, which represents great social and economic costs. This study aimed to evaluate the direct medical costs of SMA, from the National Health System perspective in Catalonia, and provide regional data for the development of optimal disease management protocols and resource allocation decisions at the regional level. A retrospective, population-based study was designed based on admission records from primary care centres, hospitals and specialised care settings (inpatient and outpatient care), emergency services and extended care facilities obtained from a regional governmental claims database. A total of 396 patients met the inclusion criteria. Annual direct medical costs summed €58,606 per patient, taking into account the use of healthcare resources at all levels of care and excluding the cost of prescription medication. Specialised care represented 81% of the expenses that were mostly associated with respiratory manifestations of SMA. In the year 2016, 71.26% of patients with SMA had four or more systems affected by a chronic condition, versus 23.50% in the general population, which had an impact on healthcare use. Inpatient extended care and the increased presence of multimorbid chronic conditions in patients with SMA must be taken into account in order to develop multidisciplinary treatment protocols that reflect the complexity of SMA. Forthcoming resource allocation decisions should reflect the intensive use of specialised care registered in patients with SMA.


Introduction
Chronic and genetic conditions such as spinal muscular atrophy (SMA) pose a challenge for families and the healthcare system in terms of economic costs [1]. In the US, approximately 78% of health spending was estimated to be destined to treat patients with chronic conditions, a sum that grows exponentially as the number of chronic diseases per patient increases [2]. In Spain, a similar percentage is estimated, considering that these conditions are responsible for 80% and 60% of primary and specialised care consultations, respectively [3].
SMA is a rare disorder, estimated to affect 1 per 10,000 live births when considering all SMA types together [4]. However, the different degrees of motor function impediment derived from the degeneration of the alpha motor neurons of the spinal cord define extremely high healthcare costs to be assumed by patients, families and healthcare systems [1,5]. Previous evaluations in Spain were based on questionnaires completed by caregivers, in which their sociodemographic characteristics and use of healthcare and non-healthcare resources were measured [6]. Such studies remark the socioeconomic costs of the disease, yet regional data, including a detailed evaluation of direct medical costs, is still needed in Spain. The chronicity associated with the disease and its multisystemic symptomatology require complex protocols that may increase treatment costs greatly [7]. Thus, the economic burden associated with SMA is estimated to be significant within the healthcare system due to patients' need for extended care, long-term ventilation, or nutritional support [8,9]. Additionally, the number of multimorbidities in these patients is relatively high, including scoliosis, joint contractures, fatigue, and other orthopaedic problems [10,11], while healthcare systems are rarely prepared to handle multimorbidity in favour of individual diseases, with patients visiting a large number of different medical specialists every year [12,13]. Cross-sectional studies have shown that a high multimorbidity number relates to poorer quality of life and more intensive use of healthcare resources than in patients without multimorbidity, which ultimately correlates with higher medical costs [14]. Contrarily, systems with successful integrated multimorbidity care are typically found in countries in which primary care takes a central role [15].
The National Healthcare System in Spain is predominantly publicly funded, with a strong primary care system focused on improving health system coordination and addressing chronic conditions [16]. Health competences are transferred to the Autonomous Communities, whereas the National Healthcare System remains responsible for coordination and certain nationwide policies [17]. In the Spanish region of Catalonia, the regional government has developed a specialised programme that gathers details associated with patient admission records, going back to the year 2007. This data analysis programme for research and innovation in health (Public Data Analysis for Health Research and Innovation Program [PADRIS]) aims to provide a tool for the evaluation of disease management, assessing not only healthcare use but also patient characteristics in terms of multimorbidity [18]. This way, the regional government aims to adapt the healthcare system to the population needs via the development of new protocols that reflect disease complexity, with a special focus on multidisciplinary care. In addition, the PADRIS programme aims to assist regional health departments that require data acquired at the regional level for reimbursement and resource allocation decisions. This real-world evidence provides the necessary data for health policy decisions in public health and revision of the established protocols and guidelines [19].
In this context, the objective of this study was to estimate the direct medical costs that are derived from SMA in the Catalonia region by means of the PADRIS database, as well as to assess the presence of multimorbid conditions among these patients.

Data Extraction
Anonymised admission records corresponding to patients diagnosed with SMA types I, II, III, and IV, as well as progressive SMA, in the Catalonia region (7.5 million inhabitants) between 1 January 2014 and 31 December 2017 were extracted from the PADRIS database under the supervision of the Health Evaluation and Quality Agency of Catalonia (AQuAS). Within the database, data were codified via the International Statistical Classification of Diseases and Related Health Problems, Ninth Revision, Clinical Modification (ICD-9-CM) [20]. Thus, to identify patients with SMA, the following ICD-9-CM codes were used: 335.0, SMA-I or Werdnig-Hoffmann; 335.19, SMA in adults (SMA-IV) and other, including SMA-II; 335.11, SMA-III or Kugelberg-Welander; 335.21, pure progressive SMA specified as such.
This study did not involve human participants, data were previously anonymised, and there was no access to identifying information, therefore patient consent was not required in accordance with the Declaration of Helsinki and Spanish legislation (Law 14/2007, 3 July, on biomedical research, Spain). The University of Barcelona's Bioethics Commission approved the research plan before data extraction.

Study Variables
The files included admission data from primary care centres, hospitals and specialised care settings (inpatient, outpatient care and any physiotherapy sessions taking place therein), emergency services and extended care facilities, together with information regarding pharmaceutical consumption and laboratory analyses.
Separately, the database classified patients into multimorbidity groups, performed automatically according to the number of systems affected by chronic conditions, using a morbidity measurement developed by the Catalan Health Institute (ICS). This measurement of morbidity uses Adjusted Morbidity Groups (Grupos de Morbilidad Ajustados, GMA) as an alternative to the Clinical Risk Group (CRG) and classifies patients into seven groups: healthy population, pregnancy, acute pathology, chronic disease in one system, chronic disease in two or three systems, chronic disease in four or more systems, and neoplasia; each one of these groups (excluding the healthy population) is divided into five levels of complexity that is determined by their use of resources, risk of hospitalisation and mortality [21]. Therefore, in this method, higher GMA levels correlate with more intensive use of healthcare resources and thus major medical costs [21]. Comorbidity indicators were extracted for the year 2016.

Data Analysis
The patient population was categorised into SMA types by analysing the first admission registered per patient due to SMA during the study period; the complete admission data were used for the remaining analyses. The direct medical costs of SMA between 2014 and 2017 were calculated based on the regional admission and medical procedures tariffs list determined by the Catalan government for the year 2013, which was the list of reference within the study period [22,23].
The distinct centres were considered independently. In all cases, a first portion of direct medical costs was approximated based on the mean cost per admission determined by the Catalan government, which considers the expenses associated with personnel, nutrition (diet) and the average cost of general medical equipment [22]. The costs associated with specific procedures and healthcare (inpatient and outpatient) required by patients with SMA, treatment (examination, medication and surgery) and specialised equipment were calculated separately (taking into account that a patient with SMA might require more complex or specialised care) by multiplying the cost per procedure or per diagnosis as determined by the diagnosis-related group (DRG) by the total number of procedures or diagnoses registered [22]. In addition, the following parameters were fixed in the analysis: the cost of specialised inpatient admissions was based on the average cost per day for hospitalisations of 6 or more days; the costs incurred by extended care facilities were given, considering patients requiring level B support (out of three levels: medium complexity, pathology and dependency) [22,23].
Data presentation is mainly descriptive. Statistical analyses were performed using Microsoft Excel© Professional Plus 2010 (Microsoft Corporation, Redmond, WA, USA).

Results
In total, 26,888 admission files were obtained from the distinct healthcare settings, corresponding to 396 patients admitted with SMA in the Catalonia region (7.5 million inhabitants) between 2014 and 2017. The patient population comprised 170 (42.9%) patients with progressive SMA, 114 (28.7%) patients with SMA-III, 39 (9.8%) patients with SMA-II or SMA-IV, 22 (5.5%) patients with SMA-I, and 51 (13.0%) unclassified patients. All patients were attended to in primary and specialised care centres and emergency rooms at some point during the study period, while only 286 patients were attended to in extended care facilities between 2014 and 2016. The majority of admissions, i.e. 22,936 (85.30%), were registered in primary care centres, 830 admissions (3.09%) were registered in specialised care centres, emergency services registered 2836 admissions (10.55%), and extended care facilities registered 286 admissions (1.06%), with no trends observed during the study period (Table 1). When an inpatient, specialised care admissions registered a mean length of stay of 8 days. The number of admissions corresponding to male patients was slightly above 50% in all settings.
The average direct annual medical cost associated with SMA in the Catalonia region was €22,838,916, corresponding to €58,606 per patient. Hospitalisation costs represented 73% of expenses and were principally justified due to patients' extended need for mechanical ventilation (71.64%); physiotherapy was also accounted for in this section (Table 2). Costs originated in emergency rooms were principally derived from admissions due to acute and chronic respiratory failure, obstructive chronic bronchitis, essential hypertension and chronic kidney disease.
Laboratory tests in patients with SMA did not represent significant sanitary expenses since analyses revealed a predominance of routine blood tests. Over the study period, 92,073 tests were registered, involving 337 patients and corresponding principally to complete blood count studies and measures of glucose, creatinine, cholesterol, potassium, sodium and various antibodies. Similar outcomes were obtained from the pharmacy dispensation analysis; external  (Table 3). Overall, 6.68% of patients were exempted from contributions due to old age or long-term unemployment. Additionally, multimorbidity was assessed based on GMA values for the year 2016. The data obtained corresponded to 348 patients (Table 4). According to this classification (the year 2016), 71.26% of patients diagnosed with SMA had four or more systems affected by a chronic condition, while this value was 23.50% in the general population (Fig. 1). Instead, the general population included a higher

Discussion
Understanding healthcare usage patterns will be crucial to establishing protocols that take into account the needs of chronically ill patients, especially patients affected by more than one chronic condition. In addition, the revision of protocols and guidelines generally requires real-world evidence [19]. This study aimed to evaluate, in detail, the direct medical costs that derive from SMA in the Catalonia region, while assessing the presence of multimorbid conditions in these patients and how they correlate with disease burden, vital in a region that relies on regionally acquired data for reimbursement and resource allocation plans [24]. Multimorbidity is considered to be one of the determining challenges for healthcare management in the coming generations [25]. It has been proven that as the number of diagnosed chronic conditions increases, healthcare expenses grow exponentially, as does the number of admissions per patient [2,26]. Herein, 71.26% of patients included in the study had four or more systems affected by a chronic condition in 2016, whereas in the general population, this value was 23.50% for the same year [27]. A previous analysis in the general population showed that increased complexity correlated with 2-11 times more emergency room admissions and up to an 11-fold increase in pharmaceutical costs [21]. Similar outcomes can be assumed in patients diagnosed with SMA, hence the importance of complexity and multimorbidity in upcoming protocol revision and resource allocation.
In this study, costs resulting from hospitalisations were responsible for the majority of expenses, although only 830 admissions were registered and were mostly due to complications derived from respiratory difficulties. The role of specialised care in resource expenditure is considerable. As a consequence, the goal of healthcare providers is to increase the efficiency of specialised care while promoting the use of primary care. Some evidence suggests that broadening the care offered in primary care centres can reduce medical costs; however, definitive data are hardly obtainable since more specialised primary care may lead to the early detection of conditions that require unavoidable hospitalisation [28,29]. The opinion of clinicians and experts should be considered together with this data in order to promote health measures and for protocol modification.
The overall cost of SMA care has been estimated on multiple occasions, however huge discrepancies exist between data obtained from different regions and datasets. In Spain, a  [6]. Similarly, in Germany, the direct medical cost of SMA was established at €14,342 per patient in 2016, giving great weight to hospital inpatient care [30]. These estimations are significantly lower than the costs predicted in the present study, where the annual healthcare costs associated with patients with SMA were estimated to reach €58,606 per patient. These discrepancies may be highly associated with the methodology. Herein, specialised care, and in particular mechanical ventilation and other procedures related to respiratory failure, were estimated separately and accounted for most of the expenses. Comparably, in 2018, the cost of a single hospitalisation for patients with SMA-I in the US was approximately €45,000, billing mostly for nutritional and ventilator support, attributable to the complexity and duration of the care provided [9]. Also in the US, estimations based on US military healthcare system data approximated a cost of €42,300 per year of direct medical costs for patients with SMA [31].
In contrast, costs estimated for the general population are considerably lower. The direct public healthcare cost per person in Spain was approximately €1594 during 2017, with 63.2% of total costs dedicated to specialised care [32].
Finally, data in Table 3 show that the healthcare system covers the majority of pharmaceutical costs. For 64.89% of patients, a minimum of 90% of costs are covered, having to contribute solely to 10% of the pharmaceutical costs. Current expenses are not significant, although patients' contributions may become relevant as new treatments arise and are released into the market [33]. One of the novel treatments for SMA is nusinersen, a modified antisense oligonucleotide for treating SMA caused by mutations in chromosome 5q [34]. The approval of nusinersen in Spain was subsequent to this study and its use will presumably increase the overall medical costs covered by the Spanish National Healthcare System [35].
A series of limitations influenced the results of this study. The ICD-9-CM codification used in the database impedes the analysis per SMA type, given the lack of specific codes for SMA-II and SMA-IV. The use of a fixed cost per admission plus a variable portion corresponding to medical procedures may lead to an overestimation of the costs associated with hospital bed, nutrition and personnel. On the other hand, rehabilitation taking place in external private centres was not included.
This study offers an evaluation of the direct medical costs of SMA in Catalonia, aiming to assist regional resource allocation decisions. Nonetheless, complementary research analysing informal care costs, indirect medical costs and nonmedical costs will be required to quantify the total burden of this condition.

Conclusions
SMA represents a significant burden for the health care system, i.e. €58,606 per patient, primarily associated with hospitalisations and the care of patients with respiratory failure. Patients require extended inpatient care, which determines intensive use of resources and can be associated with the high presence of multimorbid conditions; 71.26% of patients had four or more systems affected by a chronic condition. Healthcare systems need to adapt to patients' needs as the prevalence of multimorbidity increases, and to promote multidisciplinary care.

Conflicts of interest Josep Darbà declares that no conflicts of interest exist.
Ethical standards This project was approved by the University of Barcelona's Bioethics Commission. Data were anonymised prior to extraction in accordance with the Declaration of Helsinki and Spanish legislation (Law 14/2007, 3 July, on biomedical research, Spain). Patient consent was not required for this study.
Funding The author received no specific funding for this work.