Potentially Avoidable Emergency Department Utilization by Persons With Psoriatic Arthritis and Radiographic Axial Spondyloarthritis: A Population-Based Cohort Study

DISCUSSION

We examined utilization and characteristics of ED use by persons with PsA and r-axSpA. We found that half of ED visits were triaged as less urgent or nonurgent, with the most prevalent reasons for presentation being for symptoms that could not be attributed to a specific diagnosis. Injuries and infections accounted for just over 10% of ED visits each, with IA as the most responsible diagnosis, having a significant number of return visits but with admission being relatively rare. Persons from rural settings had nearly twice the mean number of visits per year and were more often triaged as less acute, more frequently returned to ED within 72 hours, and were less frequently hospitalized compared to their urban counterparts. We also identified sex differences in ED use, including female individuals presenting more often during daytime hours compared to male individuals, and a lower frequency of admission for female individuals with PsA relative to male individuals with PsA. These epidemiologic data provide insight into healthcare-seeking approaches among persons with PsA and r-axSpA, particularly with respect to the striking differences in ED utilization patterns between rural and urban settings. Our findings add to the mounting evidence for areas of improvement in the provision of ambulatory care for the timely assessment of less urgent and nonurgent issues, especially with acute care resources being increasingly under strain.⁷

The frequency of ED utilization among both cohorts in our study, which was stable over a decade similar to general ED utilization trends in Canada,⁷ was at least double the annual prevalence in other PsA and r-axSpA cohort reports from the US^8,9,11 and Ontario, Canada.¹⁰ These differences could relate to context-specific factors. For example, there could be increased utilization of acute care services in the universal healthcare model existing in Canada, as it has reduced cost barriers relative to the US. Interprovincial variability of ED utilization within Canada is evident, with a lower overall acuity of ED presentations in Alberta compared to Ontario.¹⁶ The yearly proportion of individuals hospitalized in our study likewise exceeds previous reports from the US,^8,9,11 which has been attributed to a relative lack of acute care beds and longer duration of hospitalization in Canada compared to other economically developed countries.⁷ Causes for the prevalence of acute care use being the most pronounced for individuals with PsA compared to r-axSpA remain to be examined. Other factors driving ED utilization and overcrowding among persons with IA, such as lack of ambulatory care capacity and accessibility,^7,11 would presumably affect individuals in a similar fashion, regardless of PsA or r-axSpA diagnosis.

Between these 2 conditions, we did not observe major differences in key diagnostic categories at ED presentation, such as major adverse CV events, as might be expected with the prevalence of cardiometabolic comorbidity among those with PsA,¹⁷ though elevated CV risk for individuals with r-axSpA is also increasingly recognized.^18,19 Other categories with relevance to disease manifestations and treatment outcomes, such as IA, dermatologic, gastrointestinal, and infectious diagnoses, were similar. Taken together, this trend suggests greater all-cause acute care utilization for persons with PsA vs r-axSpA, which stands in contrast to prior literature finding similar proportions of ED utilization and admission between patients with PsA and r-axSpA in the US.¹¹

Finally, the greater reliance on ED services in rural settings in both cohorts may reflect reduced accessibility to ambulatory care, which previous research has correlated with preventable acute care.^20-22 Alternatively, the prevalence of weekday and daytime presentations to rural EDs may be partly explained by rural primary care physicians (PCPs) seeing patients in those settings while they are simultaneously staffing the ED. The higher rates of return to ED within 72 hours in rural settings may be similarly explained by context-specific factors such as patients intentionally returning when services such as diagnostic imaging become available during daytime weekday hours. Although not statistically significant, we observed a slight trend toward more frequent IA presentations for rural patients in both cohorts, which may signal worse disease outcomes among rural patients, who are known to face significant disparities in access to rheumatology care relative to their urban counterparts.²³

The present study signals gaps in healthcare provision for individuals with rheumatologic concerns including PsA and r-axSpA, with implications for other groups with complex chronic medical conditions navigating the healthcare system. High acute care use often indicates shortcomings in ambulatory care accessibility—in this context reflecting both rheumatology specialty care and primary care—to address nonurgent and disease-specific issues. The shortage of PCPs is well known, with recent Canadian research estimating 1 in 5 people nationally, and a similar portion of those residing in Prairie provinces such as Alberta, as not having a PCP.²⁴ Access barriers to ambulatory rheumatologic care are well established.²⁵ Increased and sustained investments in training and retaining rheumatology and PCPs, improved information technology, and creative solutions for optimizing workflow including payment and care models supporting interprofessional care, are indicated.^7,26 Sex- and gender-specific aspects such as differential access to care, prescribing, and response to therapies are also described in patients with PsA and r-axSpA, among other systemic autoimmune rheumatic diseases, and need to be addressed accordingly in designing accessible models of care. Patient education programs could include information about when and how to seek urgent reassessments for disease flares. Innovative solutions that are proposed include implementation of patient navigation technology and personnel, involvement of allied health and extended role practitioners, telemedicine, and adapting current training programs and practice models to increase the presence of rheumatology expertise in these communities.^23,27

These findings should be considered in the context of several strengths and limitations of our study. A control population was not provided by the data stewards, but comparing the results obtained in the PsA and r-axSpA cohorts relative to the general population would be of interest for future research. The extended run-in period to ascertain incident cases with a decade of follow-up data allows for longitudinal evaluation of ED utilization compared to previous literature, which is largely limited to a constrained 1-year or 2-year sampling. Our study also used a large population-based universal healthcare dataset that captures all acute care interactions and more granular details such as time and day of presentation, triage acuity, and most responsible diagnosis of both initial presentation and 72-hour return to ED, which is lacking in previous similar research. We used validated case definitions to accurately identify PsA and r-axSpA, but the nature of the large dataset nonetheless brought other inherent coding limitations, such as the large proportion of nonspecific diagnoses; the exact effect of this limitation is unclear as these categories could conceivably apply to a range of specific informative ED diagnoses from infection to IA. Other important variables such as comorbidity, pertinent investigations and medications, primary care and rheumatology provider linkage, disease activity, and socioeconomic factors were also not available and would strengthen future research. Last, a cost analysis was not possible with our dataset and would be a valuable aspect for future study as others have shown increased and variable office visit, ED, and hospitalization costs among patients with PsA and r-axSpA.¹¹

In conclusion, in this retrospective study of linked population-based administrative health data in the province of Alberta, Canada, infections and injuries comprised the most frequent responsible diagnoses. Approximately half of persons with PsA and r-axSpA presenting to ED were triaged as less urgent and nonurgent acuity. This was more common in rural areas, as was more frequent re-presentation and hospitalization. These findings highlight areas for improvement in providing urgent and primary care services for persons with PsA and r-axSpA, particularly in rural locations, to optimize utilization of limited healthcare resources.