Influence of Socioeconomic Status on Distance Traveled and Care After Stroke
Background and Purpose—Vital to maintaining an efficient delivery of services is an understanding of patient travel patterns during an acute ischemic stroke. Socioeconomic status may influence access to stroke care, including transportation and admission to different facility types.
Methods—We analyzed all acute ischemic stroke admissions between 2003 and 2007 through the Discharge Abstract Database, a national database containing patient-level sociodemographic, diagnostic, procedural, and administrative information across Canada. Socioeconomic status was defined in neighborhood quintiles according to Statistics Canada. Distances between patients and facilities were derived from postal codes. A principal diagnosis of ischemic stroke was identified using the International Classification of Diseases (versions 9 and 10). Analysis of variance and regression analyses were performed with adjustment for demographic characteristics.
Results—Admitted to acute care institutions were 243 410 patients with ischemic stroke. Mean patient age was 72.8 and 49.5% were male; 44.2% traveled beyond their closest center, amounting to an average 7.2 km additional distance traveled. Socioeconomic status quintile had minimal effect on travel patterns, with the lowest socioeconomic status accessing the closest center most frequently (odds ratio, 1.19; 95% confidence interval [CI], 1.13–1.16). Increased utilization of the closest hospital occurred with academic (odds ratio, 6.90; 95% CI, 6.69–7.11) or high-volume (odds ratio, 1.93; 95% CI, 1.88–1.98) facilities. Older patients (β=0.28; 95% CI, 0.27–0.28), expert destination facility (β=0.13; 95% CI, 0.12–0.14), and ambulance use increased travel beyond the closest center.
Conclusions—Patients tend to choose care facilities based on hospital expertise; investment promoting improved regional facilities may be of greatest benefit to patients. Socioeconomic status has little bearing on travel patterns associated with stroke in Canada. These findings may assist in allocating funding to centers and improving patient care.
In the past 2 decades, there has been an unprecedented increase in population age, bringing an increased burden of stroke and the need for efficient high-quality health care delivery.1,2 Escalating levels of organized stroke care have been shown to significantly improve patient outcome.3,4 Unfortunately, these interventions are often limited to larger and more comprehensive stroke centers.5,6 Thus, it is important to identify whether patients' selection of stroke care facilities is based on proximity or level of expertise and resource, and how this travel is affected by socioeconomic status (SES) and ambulance use. A better understanding of patient flow would help determine whether care required matches care provided and how to optimally allocate public resource.
We examined the association between SES and travel patterns during a stroke. Secondarily, we examined the association between patient demographics, ambulance use, hospital expertise, and community type on travel patterns. We hypothesized that patients with low SES receive care at the nearest and less specialized hospital.
Materials and Methods
Study Design and Data Sources
We conducted a retrospective study using the Discharge Abstract Database, a mandatory reporting national database using information on all hospitalizations at acute care institutions across all provinces, except Quebec. From the Discharge Abstract Database, we identified all patient admissions from acute care centers between January 1, 2003 and December 31, 2007, with International Classification of Diseases version 10 diagnosis codes I63.x and I64.x representing “cerebral infarction” and “strokes not specified as hemorrhage or cerebral infarction,” respectively (please see Supplemental Methods for details on data quality; http://stroke.ahajournals.org). Baseline characteristics including demographics, comorbidities, ambulance use, length of stay, and special care unit admission were extracted and are further defined (Supplemental Methods).
Definition of Distances and SES
The forward sorting area (first 3 postal code characters) was used to identify unique census dissemination areas for analysis. Discharge Abstract Database patient and hospital forward sorting areas were linked with Statistics Canada's 2006 Census data. Median household income for each forward sorting area was divided into quintiles and used as a marker of patient SES. “Crow flies” distances were calculated between patient forward sorting areas, discharging hospital, and closest hospital (Supplemental Materials for further details).
The primary outcome measure was travel distance in kilometers. Secondary outcomes include length of hospital stay, admission to specialized care unit, and mortality at discharge.
The Cochran-Armitage test was used to compare categorical variable levels and trends in proportions across SES quintiles. Linear-contrast analysis of variance was used for continuous variable means across quintiles. For those traveling beyond their closest center, median extra distance traveled was determined. Logistic regression was used to examine both factors associated with presentation to the closest stroke center and factors associated with presentation beyond the closest center (http://stroke.ahajournals.org for details).
During the study period (January 1, 2003–December 31, 2007), 243 410 patients presented to participating institutions with a diagnosis of cerebral infarction or stroke (not otherwise specified). Descriptive statistics are summarized in Supplemental Table I. Mean age was 72.8 years and 120 415 (49.5%) patients were male; 145 072 (59.6%) patients arrived by ambulance. Secondary outcome measures are summarized in Supplemental Table II. Mean hospital length of stay was 12.7 days; 44 215 (18.1%) patients were admitted to a special care unit. In-hospital mortality was 41 623 (17.1%).
Half of patients presented to their closest acute care center. A Cochran-Armitage test verified a linearly decreasing trend from lowest to highest SES quintiles (P<0.001; Table 1). The absolute difference between the lowest and highest quintiles was 4.0%. Of patients traveling to more distant centers, the median additional distance traveled was 7.2 km (interquartile range, 2.0–43.7 km) and was similar among all SES strata.
Factors Associated With Presentation to the Closest or More Distant Centers
The most important factors associated with presentation to the closest stroke care center included admission to an academic hospital (odds ratio, 6.90; 95% confidence interval [CI], 6.69–7.11) and high-volume destination hospital (odds ratio, 1.93; 95% CI,1.88–1.98) for those arriving by ambulance. The same factors were found for patients arriving by other means (Table 2).
Of those traveling to more distant centers, the most important factors associated with greater travel distance were older age (β=0.28; 95% CI, 0.27–0.28), high-volume hospital (β=0.13; 95% CI, 0.12–0.14), and academic hospital status (β=0.08; 95% CI, 0.07–0.08; Table 3).
Increased life expectancy, population aging, and limited resources are serious challenges to health care systems. Selecting the correct paradigm of stroke care resource management (centralized versus global investment) is important to efficiently and effectively deliver care. Understanding patient travel patterns and choice of stroke care centers aids in this decision. Any investment should promote equal access to care across all SES strata.
We found that half of patients accessed their closest center and half traveled to more distant centers. All SES quintiles tended to seek specialized stroke care at academic or high-volume centers, both perceived to provide greater quality of care and better outcomes. Rural residents were more likely to choose the closest center, likely because of greater distances to larger centers than for urban residents. These findings were increased further when traveling by ambulance.
Median additional travel for stroke patients receiving care at more distant hospitals was 7.2 km, creating delay in initial assessment and care. SES had minimal effect on additional distance traveled, suggesting stroke patients have equal access to acute care facilities.
Expert hospitals (high-volume and academic) were associated with greater travel distances, likely because of increased complexity of offered services and presumed higher quality of care, encouraging longer travel. Older patients may travel farther because of need for expert (often more distant) care because of a greater number of comorbidities (Supplemental Materials for interpretation of results).
Together, these results suggest the need for investment in greater numbers of regional stroke care centers providing the expertise patients seek with reduced travel distances. Currently, resources are well-distributed to limit major disparities between SES quintiles.
Limitations of the study, use of postal code areas, and SES surrogates are discussed in detail in the supplemental file (Supplemental Methods). Despite these limitations, our study presents a novel analysis looking into the relationship between SES, stroke patient travel patterns, and associated outcomes.
Sources of Funding
Dr. Saposnik was supported by the Clinician Scientist Award from Heart & Stroke Foundation Ontario. Chris Ahuja is supported by the John F Sparks Studenship Award of Queen's University.
The authors thank the Canadian Institute for Health Information (CIHI) for their support through the Graduate Student Data Access Program (GSDAP).
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.635045/-/DC1.
The Stroke Outcomes Research Canada working group (SORCan–www.sorcan.ca).
- Received August 4, 2011.
- Accepted August 18, 2011.
- © 2012 American Heart Association, Inc.
Median age, Canada, 1901–2011. Statistics Canada web site. Available at: http://www12.statcan.ca/English/census01/products/analytic/companion/age/cdamedaget.cfm. Accessed May 10, 2011.
- Saposnik G,
- Fang J,
- O'Donnell M,
- Hachinski V,
- Kapral M,
- Hill MD,
- et al
Stroke Unit Trialists' Collaboration. Organised inpatient (stroke unit) care for stroke. Cochrane Database System Rev. 2007; 4: doi:10.1002/14651858.CD000197.pub2.
- Panella M,
- Brambilla R,
- Marchisio S,
- DiStanislao F