Trends in incidence, prevalence, and mortality of neuromuscular disease in Ontario, Canada: A population-based retrospective cohort study (2003-2014)

Background Population trends of disease prevalence and incidence over time measure burden of disease and inform healthcare planning. Neuromuscular disorders (NMD) affect muscle and nerve function with varying degrees of severity and disease progression. Objective Using health administrative databases we described trends in incidence, prevalence, and mortality of adults and children with NMD. We also explored place of death and use of palliative care. Methods Population-based (Ontario, Canada) cohort study (2003 to 2014) of adults and children with NMD identified using International Classification of Disease and health insurance billing codes within administrative health databases. Results Adult disease prevalence increased on average per year by 8% (95% confidence interval (CI) 6% to 10%, P <.001), with the largest increase in adults18-39 years. Childhood disease prevalence increased by 10% (95% CI 8% to 11%, P <.0001) per year, with the largest increase in children 0 to 5 years. Prevalence increased across all diagnoses except amyotrophic lateral sclerosis and spinal muscular atrophy for adults and all diagnoses for children. Adult incidence decreased by 3% (95% CI -4% to -2%, P <.0001) but incidence remained stable in children. Death occurred in 34,336 (18.5%) adults; 21,236 (61.8%) of whom received palliative care. Death occurred in 1,009 (5.6%) children; 507 (50.2%) of whom received palliative care. Mortality decreased over time in adults (odds ratio (OR) 0.86, 95% CI 0.86–0.87, P <.0001) and children (OR 0.79, 95% CI 0.76–0.82, P <.0001). Use of palliative care over time increased for adults (OR 1.18, 95% CI 1.09 to 1.28, P <.0001) and children (OR 1.22, 95% CI 1.20 to 1.23, P <.0001). Conclusions In both adults and children, NMD prevalence is rising and mortality rates are declining. In adults incidence is decreasing while in children it remains stable. This confirms on a population-based level the increased survival of children and adults with NMD.

Introduction Neuromuscular disorders (NMD) vary in severity and disease progression [1,2]. These disorders result in longstanding functional deficits that result in substantial utilization of healthcare resources in addition to the emotional, financial and social burden to those affected individuals and their families. Some forms of NMD are hereditary and manifest at birth or early childhood whereas the incidence of other forms increases with age [3]. On an international level, an aging population means the prevalence of NMD is likely to increase.
Although some estimates of NMD prevalence exist [4,5], case ascertainments methods are highly variable and frequently biased. These include chart reviews, surveys, family histories, interviews with relatives and patient registries [4]. Few studies are conducted using population based health administration databases containing all publicly funded healthcare encounters for a region or an entire country. Population based estimates of disease prevalence and trends over time inform an understanding of the natural history as well as the impact of newer management interventions on the epidemiology of disease measure burden of disease and inform healthcare planning [6].
Understanding the magnitude of NMD at a population level is important to understand the impact on the healthcare system and the healthcare burden of individuals to inform healthcare policy. Therefore, we conducted a population based cohort study using health administrative databases to determine the prevalence, incidence, and mortality for adults and children with NMD. Our secondary objectives were to describe prevalence and incidence for various types of NMD, location of death, and use of palliative care services.

Study design
We conducted a retrospective population-based study using health administrative databases for the province of Ontario, Canada held at the Institute for Clinical Evaluative Sciences (ICES) using unique encoded identifiers to link databases. These databases contain anonymized data for all residents of the province of Ontario, population approximately 13 million [7]. In Canada, universal public health insurance funded through general taxation covers cost of all medically necessary care.

Population and data sources
Form the Ontario Registered Persons Database (RPDB)we identified adults aged 18 to 105 and children aged 0 to <18 years with NMD using hospital diagnostic codes (International Classification of Disease (ICD)-9, ICD-10) or physician billing (Ontario Health Insurance Plan (OHIP)) codes a priori considered specific to NMD within health administration databases. We verified physician billing codes for NMD with neurologists working with research team members. We sought the first instance (from fiscal year 2003 onwards) of these codes considered the 'most responsible' or secondary contributing diagnoses indicating hospital admission with a NMD diagnosis in the Discharge Abstract Database. This database includes all hospitalizations including demographic, procedural data (up to 24 comorbidities and procedures), and in-hospital death [8]. We also examined hospital diagnostic codes in the National Ambulatory Care Reporting System containing all emergency department visits and physician billings codes in the OHIP database [9] that contains all physician billings including procedures ( Fig  1). To establish death outside of hospital we linked to the RPDB. To determine use of palliative care we sought validated palliative care codes [10] in the the hospitalization, ambulatory care and physician databases as well as the Continuing Care Reporting System, the Complex Continuing Care database, and the Home Care Database.
For individuals with the first occurrence of a NMD diagnostic code found using a physician billing code, we sought to confirm NMD diagnosis by looking backwards and forwards in time (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) in the hospitalization database for a NMD-related hospital admission identified through an ICD code. For individuals identified using the physician billing code 349 which is used as a non-specific catch all code for NMD, we again looked backwards and forwards in the hospitalization (1998-2014) and ambulatory care (2000-2014) databases and sought ICD codes and physician billings from a neurologist or for an electromyogram to increase our certainty the patient had NMD.

Outcomes
We established annual prevalence and annual incidence considering all adults and all children with a NMD diagnosis, and separately for those adults and children identified through hospitalization or ED presentation during that year. We also established annual prevalence and incidence for each NMD disease category. However, due to inability to classify disease type further with identification of NMD using physician billing we restricted our estimates of prevalence and incidence to those individuals with case ascertainment through hospital and ED diagnostic (ICD) codes. To establish incidence, we used a five-year look back window to confirm the individual had not received any health services for NMD previously and calculated annual incidence from 1 st April 2008 to 31 st March 2015. Individuals with prior health services utilization for NMD in these years were counted as prevalent but not incident. We established allcause mortality rates, identified death location (in or outside a hospital), and determined use of palliative care.

Ethical considerations
We conducted our study according to a pre-specified protocol approved by the Research Ethics Board at Sunnybrook Health Sciences Centre, Toronto, Ontario and according to privacy regulations of the Institute for Clinical Evaluative Sciences.

Statistical analyses
We estimated annual prevalence by dividing the number of affected individuals alive at the end of each fiscal year by the census population estimated in the corresponding year from the RPDB. To estimate annual incidence we divided the number of new cases by the number of individuals at risk for NMD, considered as all adults resident in the province of Ontario without NMD, also from the RPDB estimated census population. Using patient level data we examined linear trends of prevalence and incidence (overall and by age and sex categories) over time using linear regression models with year as an independent variable [15]. We examined Trends in incidence, prevalence, and mortality of neuromuscular disease trends in mortality and proportion of adults and children receiving palliative care using logistic regression. We conducted analyses in SAS Enterprise Guide 7.1 (SAS Institute Inc., Cary, NC, USA).

Prevalence
For adults, the crude prevalence of NMD considering only those identified through an ED presentation or hospital admission averaged over the 12-year period (Fiscal years 2003-2014) was 72.6/100,000 adults whereas it was 1,344/100,000 adults when including those identified through physician billing (See Table A in S1 File for total eligible patient population and NMD prevalent cases). Prevalence of patients identified through ED presentation or hospital admission increased on average per year by 8% (95% CI 6% to 10%, P <.0001 for linear trend). Prevalence increased across all diagnoses except ALS and SMA (Table 1). Although prevalence was highest in adults aged � 65 years and increased for all adult age categories, the greatest average increase was in younger adults: 18-39 years 11% per year (95% CI 10% to 11%) P <.0001; 40-64 8% per year (95% CI 7% to 9%) P <.0001 and in adults aged �65 years 4% per year (95% CI 2% to 5%, P <.0001) (Fig 2). Prevalence increased by 8% (95% CI 7% to 8%) in both males and females ( Table B in S1 File).
For children, incidence of NMD over the 7-year period (fiscal years 2008-2014) was 4.7/ 100,000 children considering only those identified through ED presentation or hospital admission and 56/100,000 including children identified through physician billing. Incidence remained relatively static overall (2%, 95% CI -7% to 2%, P = 0.169 linear trend) and across all age categories (Fig 3). Incidence decreased over time for multiple sclerosis, neuropathies, and myasthenia gravis ( Table 2). Incidence decreased by 7% (95% CI -9% to -4%) in males but remained unchanged in females (Table C in S1 File).

Mortality and use of palliative care
Of adults with NMD identified through hospitalization or ED presentation, 34,336 (18.5%) died over the 12-year period (  Trends in incidence, prevalence, and mortality of neuromuscular disease

Discussion
In this retrospective population based study of adults and children with NMD, we found prevalence of those requiring ED presentation or hospital admission to be increasing over time among both adults and children. Prevalence in adults is highest in the oldest age category but increased the most in the youngest age category, suggesting increased survival of children with NMD into adulthood. Overall incidence is decreasing in adults but remains static in children.
Diseases for which the incidence remained unchanged in adults include neuropathies, muscular dystrophy, metabolic diseases, ALS, spina bifida and myasthenia gravis. Mortality was modest overall and has decreased over time. Two thirds of adults and 50% of children received palliative care. Receipt of palliative care has increased over time for adults and children. Trends in incidence, prevalence, and mortality of neuromuscular disease Although there have been studies reporting point prevalence or birth prevalence of various forms of NMD [16,17], none have been population-based studies in which trends in prevalence and incidence of NMD were ascertained. Prevalence is dependent on incidence and disease duration, the latter being related to disease resolution or death. Given the declining incidence in adults and static incidence in children and given that most forms of NMD do not resolve, the increasing prevalence observed can be attributed to increased survival. This is supported by the decline in crude mortality rates over the 12-year study period for both adults and children and the increased prevalence in both the youngest adult and child age categories.
The reduction in mortality found in our study is consistent with reports that identified an increased survival in various NMD populations including Duchenne Muscular Dystrophy [18], Guillain Barré syndrome [19], and ALS [20]. Reasons for this reduction are likely multifactorial. For those individuals in whom the disease progresses to respiratory failure improvements to monitoring of respiratory complications, newer technologies for lung hygiene [18]  Trends in incidence, prevalence, and mortality of neuromuscular disease and nocturnal ventilation [21] are likely contributors in addition to more timely access to specialist services, including multi-disciplinary and inter-professional specialist clinics [22]. Almost two thirds of adults and half the children that died received palliative care, with adult use increasing over time. Although the Canadian healthcare system offers advanced palliative care [23] most is delivered within hospitals rather than the community, with substantial variation in access to palliative care regardless of diagnosis related to geographic location [24]. Barriers to the utilization of palliative care for children and adolescents with NMD include poor family awareness of palliative care services and what they provide [25], service availability and cost, age-appropriateness of services, perceptions of palliative care and perceived benefit [26]. Trends in incidence, prevalence, and mortality of neuromuscular disease Comparison with international reports of prevalence and incidence rates is challenging due to differences in case ascertainment rates and denominator. We found the prevalence of muscular dystrophy to be 8.6/100,000 adults and 5.6/100,000 children in 2014 which is lower than other international reports and likely due to restricting our estimates to those identified through hospital admission or ED presentation and therefore likely to have more severe disease. Global estimates of the prevalence of combined muscular dystrophies derived from systematic reviews of existing studies, estimate the pooled prevalence to range from 16.1/100,000 [4] to 25.1/100,000 people [16]. We observes a slightly lower prevalence (2.1/100,000 in 2014) and incidence (0.5/100,000) of ALS compared with European ALS registers (prevalence 7-9/ 100,000, incidence 2.6/100,000) [27]. Our observations were more in keeping with recent US estimates of ALS prevalence (4 5.0/100,000) [5].
Study limitations include the possibility of misclassification, either under or over ascertainment of NMD due to the absence of validated algorithms for case ascertainment. We restricted reporting of disease specific prevalence and incidence to those individuals identified through hospital diagnostic codes and therefore these data likely underestimate the true population rates that include mild or early disease. However, we are able to provide population-based trends in prevalence, incidence, and mortality of adults and children with NMD requiring emergent care. Coding within the health administrative database prevented more disease specific descriptions of prevalence and incidence i.e., Duchenne muscular dystrophy, the most common muscular dystrophy. We were also unable to distinguish between acute and chronic disease. Finally, delivery of palliative care may not always be coded in the health administrative databases, as some physicians may be unaware of palliative care billing codes.

Conclusion
In this population-based study of adults and children with NMD, we observed a rising prevalence of NMD, particularly in younger adults and children, whereas mortality is declining. The incidence of NMD is declining in adults with no change in children. These findings confirm on a population-based level the increased survival of children and adults with NMD, and will inform healthcare planning for NMD in the future. Although a reasonable proportion of adults and children received palliative care before death, we recommend further improvements in delivery and uptake of palliative care.
Supporting information S1 File. Cohort creation codes and additional study data. (DOCX)