Abstract
Background
For patients with Kawasaki disease (KD), lower socioeconomic status (SES) may adversely affect the timeliness of presentation and initiation of intravenous immune globulin, and coronary artery outcomes. Multipayer systems have been shown to affect health care equity and access to health care negatively. We sought to determine the association of SES with KD outcomes in a single-payer health care system.
Methods
Patients with KD presenting from 2007 to 2017 at a single institution were included. SES data were obtained by matching patient postal code district with data from the 2016 Census Canada.
Results
SES data were linked for 1018 patients. The proportion of households living below the after-tax low-income cutoff in the patient’s postal code district was 13% for not treated, 13% for delayed intravenous immune globulin treatment, and 12% for prompt treatment (P = 0.58). Likewise, the average median annual household income was unrelated to delayed or no treatment. The percentage >15 years of age with advanced education differed between groups at 33%, 29%, and 31% for delayed treatment, prompt treatment, and missed groups, respectively (P = 0.004). SES variables were not significantly different for those with vs without coronary artery aneurysms (max Z-score: >2.5), including the proportion of households living below low-income cutoff (12% vs 13%; P = 0.37), average median annual household income (CAD$81,220 vs $82,055; P = 0.78), and proportion with a university degree (33% vs 31%; P = 0.49), even after adjusting for sex, age, year, and KD type.
Conclusions
Timeliness of treatment for KD and coronary artery outcomes were not associated with SES variables within a single-payer health care system.
Résumé
Contexte
Chez les patients atteints de la maladie de Kawasaki (MK), un statut socioéconomique (SSE) plus difficile pourrait retarder le moment de la première consultation et le début du traitement par immunoglobuline intraveineuse (IgIV) ainsi que peser sur les résultats associés aux artères coronaires. Il a été démontré que les systèmes à payeurs multiples compromettent l’équité en matière de soins de santé et l’accès à ces derniers. Nous avons cherché à déterminer s'il existait un rapport entre le SSE et les résultats associés à la MK au sein d’un système de soins de santé à payeur unique.
Méthodologie
L’étude comprenait des patients atteints de la MK qui se sont présentés à un même établissement entre 2007 et 2017. Les données sur le SSE ont été obtenues en associant le code postal des patients aux données du recensement canadien de 2016.
Résultats
Les données sur le SSE de 1 018 patients ont été répertoriées. La proportion des foyers qui étaient sous le seuil de faible revenu (SFR) après impôt dans la circonscription correspondant à leur code postal était la suivante : 13 % pour les patients non traités, 13 % pour les patients chez qui le traitement par IgIV a été tardif et 12 % pour les patients qui ont rapidement reçu un traitement (p = 0,58). De même, aucune relation n’a été établie entre le revenu annuel médian des ménages et un traitement tardif ou une absence de traitement. Le pourcentage de personnes âgées de plus de 15 ans ayant un niveau de scolarité élevé différait d’un groupe à l’autre, soit respectivement 33 %, 29 % et 31 % pour les groupes à traitement tardif, à traitement rapide et sans traitement (p = 0,004). Les variables en matière de SSE n’étaient pas significativement différentes chez les patients présentant des anévrismes coronariens et chez ceux n’en présentant pas (score z maximal > 2,5), peu importe la proportion des foyers qui étaient sous le SFR après impôt (12 % contre 13 %; p = 0,37), le revenu annuel médian des ménages (81 220 $ CA contre 82 055 $; p = 0,78) ou le taux de diplomation universitaire (33 % contre 31 %; p = 0,49), et ce, même après ajustement en fonction du sexe, de l’âge et du type de MK.
Conclusions
Aucune corrélation n’a été établie entre le SSE et le délai avant l’instauration d’un traitement contre la MK ou les résultats liés aux artères coronaires dans le contexte d’un système de soins de santé à payeur unique.
Kawasaki disease (KD) is the leading cause of paediatric-acquired heart disease in developed countries.1 It is a self-limiting acute vasculitis that can result in coronary artery aneurysms (CAAs) in approximately 25% of untreated cases.1 These CAAs can result in thrombosis or stenosis, which may lead to myocardial infarction and sudden death.2 The timely initiation of treatment with intravenous immune globulin (IVIG) has decreased the incidence of CAAs to 4%.1,3 This drastic decrease in the development of CAAs with treatment has prompted interest in factors that may delay the administration of IVIG.4, 5, 6
Socioeconomic status (SES) is an important determinant of health known to affect outcomes for numerous diseases.6, 7, 8, 9, 10 Different markers have been used to indicate SES, including family income, parental education, rural/urban status, and income disparity.10 Lower SES may adversely affect immune tolerance, timeliness of presentation, initiation of IVIG, and coronary artery outcomes.11, 12 There have been conflicting reports on the effect of SES on KD presentation and outcomes, which may be confounded by the fact that multipayer systems have been shown to negatively affect health care equity and may differentially affect those with lower SES.13 We sought to characterize the association between SES and KD presentation and outcomes in infants and children in Ontario, Canada, within a universal single-payer health care system.
Methods
We performed a retrospective chart review of all patients <18 years of age who presented to the Hospital for Sick Children in Toronto, Canada, from January 1, 2007, to March 31, 2017, who met criteria for KD as defined by the American Heart Association (AHA).1 The Hospital for Sick Children is a large tertiary paediatric referral hospital with a population that reflects those seen in the acute phase of the illness and those referred for consultation. We excluded patients without a Canadian postal code of residence at the time of presentation and those with incomplete clinical data. This study was approved by our institution’s internal research ethics board with waiver of consent.
Patient data were extracted periodically from the medical record and included demographic information, clinical details from the initial hospitalization, and development of complications. Demographic information included age, gender, and forward sortation area at diagnosis. Forward sortation area was defined as a geographical unit based on the Canadian postal code.14 Clinical details from the initial hospitalization included the features of KD at presentation, days of fever before treatment, type and timing of treatment, length of stay, and coronary artery outcomes. Features of KD at presentation were used to classify patients as complete or incomplete KD based on definitions by the AHA.1 Days of fever before initial IVIG treatment were used to group patients into 3 groups: delayed IVIG treatment defined as receiving IVIG after 10 days of fever; prompt IVIG treatment defined as receiving IVIG within 10 days of fever; and missed IVIG treatment. Complications included the development of CAAs and the greatest Z-score of CAAs in any branch as assessed from transthoracic echocardiography.15 CAAs were defined as a coronary artery Z-score ≥2.5 in any branch, and maximal size as small (Z-score: ≥2.5 and <5), medium (Z-score: ≥5 and <10), and large (Z-score ≥10 or absolute dimension >8 mm) as per the AHA criteria.1
Socioeconomic variables were obtained from the 2016 Census Canada through the Canadian Socio-Economic Information Management System. SES variables associated with postal code included the median household income, after-tax low-income cutoff (LICO), percentage with a bachelor’s degree or higher, and rural/urban designation. Median household income represented the median total income of all of a household’s family members before income taxes and deductions during that fiscal year. After-tax LICO was used to define families with “low income,” or below the “poverty line,” and reflects an income of at least 20% greater than average used to obtain necessary goods, clothing, and shelter.16 The percentage of households with a bachelor’s degree or higher is a measure of “educational attainment” and represents the highest successfully completed level of education of a member within a household.17 Rural/urban designation was assigned to each forward sortation area based on the type of mail delivery needed, with rural route drivers used for areas defined as rural. These SES variables by forward sortation area based on postal code were linked to each patient by postal code of residence at the time of the KD episode as a surrogate for an individual patient’s SES data, which were unavailable.
The primary outcome was the time to treatment with IVIG as described above, and the secondary outcome was the development and severity of CAAs.
Data analysis was performed using SAS software, V9.3 (SAS Institute, Cary, NC). Data were described as mean (standard deviation [SD]), median (interquartile range [IQR], values at the 25th and 75th percentile), and frequencies according to the level of measurement of the variable. Associations of outcomes with SES variables were assessed in multivariable linear regression analyses. All regression models were adjusted for age, gender, year, and type of KD. P values < 0.05 were considered statistically significant.
Results
Study population
Baseline characteristics are shown in Table 1. Of 1070 patients admitted over the study period with a diagnosis of KD, postal code information and clinical data were available for 1018 patients (95%), of whom 614 (60%) were male and 365 (36%) had incomplete KD. Patient age ranged from 0.1 to 17 years, with a median of 3.6 years (IQR: 1.8-5.8 years). Of the 1018 patients included, 113 (11%) had delayed treatment, 781 (76%) were treated within 10 days of illness onset, and 124 (12%) did not receive any treatment with the diagnosis being missed. For treated patients, the median number of days of fever before IVIG treatment was 6 days (IQR: 5-8 days). Regarding coronary artery outcomes at maximal involvement, 23 patients (2%) had missing Z-score data. Of the remaining 995 patients, 771 patients (77%) had normal coronary artery dimensions (Z-score: <2), 60 patients (6%) had coronary artery dilation (Z-score: 2-2.5), and 164 patients (16%) had CAAs (Z-score: ≥2.5). Of the 164 patients with CAAs, 97 patients (10%) had small aneurysms (Z-score: 2.5 to <5), 33 patients (3%) had medium aneurysms (Z-score: 5-10), and 34 patients (3%) had large/giant aneurysms (Z-score: >10).
Table 1.
Baseline characteristics
| Variable | Total (%) or median (IQR) |
|---|---|
| Age | 3.6 y (IQR: 1.8-5.8 y) |
| Male | 614 (60) |
| Incomplete KD | 365 (36) |
| Median days to IVIG treatment | 6 d (IQR: 5-8 d) |
| IVIG treatment groups | |
| Delayed treatment (>10 d) | 113 (11) |
| Prompt treatment (<10 d) | 781 (76) |
| Missed treatment | 124 (12) |
| Coronary artery outcomes | |
| Normal coronary arteries | 771 (77) |
| Coronary artery dilation (Z-score: 2 to <2.5) | 60 (6) |
| Small aneurysm (Z-score: 2.5 to <5) | 97 (10) |
| Medium aneurysm (Z-score: 5 to <10) | 33 (3) |
| Large/giant aneurysm (Z-score: ≥10) | 34 (3) |
| Missing | 23 (2) |
| Patient postal code district SES variables | |
| Median proportion of households below the after-tax LICO | 12.1% (IQR: 7.2%-17.0%) |
| Median yearly household income across area medians | $81,014 (IQR: $63,471-$100,291) |
| Median households with a resident who completed a bachelor’s degree or higher | 31.8% (IQR: 23.2%-41.6%) |
| Total rural participants | 32 (3) |
IQR, interquartile range; IVIG, intravenous immune globulin; KD, Kawasaki disease; LICO, low-income cutoff; SES, socioeconomic status.
Regarding the description of patient postal code district SES variables, the median proportion of households below the after-tax LICO was 12.1% (IQR: 7.2%-17.0%) across areas of patient residence, and the median yearly household income across area medians was CAD$81,014 (IQR: $63,471-$100,291). Across areas, a median of 31.8% (IQR: 23.2%-41.6%) of households had a resident who had completed a bachelor’s degree or higher. Only 32 (3%) participants were defined as rural based on their forward sortation area.
Timing of IVIG administration and SES
The timing of IVIG administration was grouped based on those receiving delayed IVIG treatment after 10 days of fever, prompt IVIG treatment within 10 days of fever, and missed KD with no IVIG treatment within the acute phase of illness. Figure 1 illustrates the comparison of the SES variables with the treatment status grouping of patients based on the timing or absence of IVIG administration. The proportion of households living below the after-tax LICO was a mean of 13% (SD: 6%) for those with delayed treatment, 13% (SD: 7%) for those who received prompt treatment, and 12% (SD: 7%) for those not treated (P = 0.58). Similarly, the average median annual household income between groups was $80,214 (SD: $19,103) for those with delayed treatment, $82,437 (SD: $21,892) for those who received prompt treatment, and $81,056 (SD: $21,519) for those not treated (P = 0.52). The average proportion of households with an individual >15 years of age with advanced education between groups, which differed between groups, was 33% (SD: 13%), 29% (SD: 12%), and 31% (SD: 14%) for delayed treatment, prompt treatment, and not treated groups, respectively (P = 0.004). There was an additional weak correlation between the proportion of residences with a bachelor’s degree and median household income (r = 0.25, P < 0.001). When analysing days to IVIG treatment as a continuous variable, greater days from symptom onset to first dose of IVIG was again associated with lower proportion with a bachelor’s degree (r = −0.09, P < 0.009), but not with median household income (r = −0.04, P = 0.23) or proportion living below the after-tax LICO (r = 0.007, P = 0.83). Living in rural postal codes was associated with delayed treatment (16% in the rural group vs 11% in the urban group) and missed treatment (31% in the rural group vs 12% in the urban group, P = 0.002).
Figure 1.
Comparison of socioeconomic variables based on intravenous immune globulin (IVIG) treatment status.
Development of coronary artery aneurysms and SES
The participants were additionally grouped based on whether there was development of CAAs and SES variables compared (Table 2). SES variables were not significantly different for those with vs without coronary artery aneurysms (max Z-score: ≥2.5), including the average proportion of households living below LICO (12% vs 13%; P = 0.12), average median annual household income ($82,553 vs $82,065; P = 0.60), and average proportion with a university degree (33% vs 33%; P = 0.85). In addition, there was no difference in the development of CAAs for those living in rural vs urban areas (16% vs 16%, P = 0.98). Adjusting for sex, age, year, and KD type did not increase associations with SES variables.
Table 2.
Comparison of socioeconomic variables based on the development of coronary artery aneurysms
| With coronary artery aneurysms (max Z-score ≥2.5) | Without coronary artery aneurysms (max Z-score <2.5) | P value | |
|---|---|---|---|
| Proportion of households below the after-tax low-income cutoff (%) | 12 (SD: 7) | 13 (SD: 7) | 0.12 |
| Average median annual household income (CAD$) | 82,553 (SD: 21,030) | 82,065 (SD: 21,780) | 0.60 |
| Proportion of individuals >15 y with an advanced degree (%) | 33 (SD: 14) | 33 (SD: 13) | 0.85 |
SD, standard deviation.
Discussion
Understanding the factors that may contribute to delayed diagnosis and treatment of acute KD may inform the development of targeted public health measures aimed at improving outcomes. SES is an important determinant of health and has been shown to impact presentation to care in many medical conditions.6, 7, 8, 9, 10 In this retrospective cohort study in a single-payer health care system with universal coverage, we noted that SES variables, including median household income and proportion of households living below the after-tax LICO, were not significantly associated with the timeliness of IVIG administration. In addition, the SES variables were not associated with the development of CAAs after adjusting for patient gender, age, year, and type of KD.
The Canadian health care system functions under a single-payer system under the Canada Health Act of 1984. Although often called “universal,” the system is actually a collection of provincially and territorially administered health insurance plans that meet a standard set by the federal government.18 This, in turn, has created a decentralized, publicly funded system covering medical and hospital services funded by taxation at the provincial and federal levels. There certainly are issues with this type of system, namely longer wait times for elective procedures and Canada-specific issues, including health care inequities for underserved populations, particularly the country’s indigenous population.18 These issues are either shared by all single-payer systems or are unique to the Canadian system but require an ongoing study to strive for equal access to care.
The significant reduction in the development of CAAs associated with timely treatment with IVIG highlights the importance of identifying and treating KD promptly.3 Delays in treatment are unfortunately common because of the nonspecific presenting symptoms and a lack of a unifying diagnostic test. Minich et al11 found that across 8 North American centres, 16% of patients were diagnosed after 10 days of fever, with factors associated with late diagnosis including patient age less than 6 months, a diagnosis of incomplete KD, and increased patient distance from a tertiary paediatric hospital. SES information was limited in this study to centres in multipayer health care systems. The SES variables, including percentage of the population with a high school degree or higher, median family income, and percentage of households under the poverty line, were not associated with delayed diagnosis.11 Contrary to this, Dionne et al.12 recently reported that in the setting of a multipayer system, lower SES was associated with a greater likelihood of a delay in treatment with IVIG, prolonged length of hospital stay, and increased risk of CAAs.12 These conflicting reports regarding the association of SES with presentation and outcomes of KD may represent region-specific differences in health insurance coverage that exists in multipayer systems for persons with low SES. Canada’s single-payer system allows the study of SES without this confounding variable.
This study was completed in Ontario, Canada, which had a higher median household income ($74,287 vs $70,336) and slightly higher proportion of household living below the LICO (9.8% vs 9.2%) than the Canadian population as a whole. Our study found that average median household income and proportion of households living below the LICO were not associated with delay or missed IVIG treatment. One factor that may contribute to this finding is the impact of universal health care on willingness and timeliness to presentation. Other potential factors may include regional differences in KD awareness, ethnic distribution, or health beliefs (mistrust of modern medicine), all of which may affect timeliness to presentation or accurate diagnosis of KD.
There was a significant association between lower neighborhood household educational attainment and delayed or missed treatment. This is consistent with previously published literature regarding the association between lower education and poor health. Zajacova and Lawrence17 discussed 4 mediating mechanisms between the two: “economic, health-behavioural, social-psychological, and access to health care.” Whether these factors contribute or are relevant to the KD population is unknown. There is also a potential for confounding in this relationship, especially with median household income, and we did indeed find a weak correlation between household income and education attainment. An additional study in this area is certainly needed.
There was a small proportion of the study population living in designated rural areas. The proportion of the Ontario population that lives in rural postal codes was 13.8% at the time of the 2016 Census19 compared with 3% in our study, which mirrors the finding that there is lower incidence of KD in rural populations.20 Living within a rural designated postal code was associated with delayed or missed IVIG treatment, although was not associated with the development of coronary artery aneurysms. This was based on the small sample of 32 participants living within rural-designated postal codes making inferences on this result difficult.
Despite lower educational attainment and rural status being associated with delayed timing of IVIG treatment, all SES variables were not associated with the development of CAAs.
This study is limited by using area SES data obtained from Census Canada as a surrogate for an individual patient’s SES. The variation of SES within a forward sortation area was not available. In addition, data were obtained from 2016 Census data for all patients over the study period and may not account for changes in neighbourhood SES over time. Information on patient ethnicity is not collected clinically at our centre, so we were not able to adjust for this.
Conclusions
Within a single-payer health care system, SES variables, including median household income and proportion of households living below the after-tax LICO, were not associated with the timeliness of treatment with IVIG or coronary artery outcomes. The association between lower education and treatment timeliness needs further study. This lack of apparent association with SES variables may reflect the positive impact of universal health care as one factor facilitating greater timeliness of presentation to medical services for patients with lower SES and KD.
Acknowledgments
Ethics Statement
This study received ethics approval from the SickKids Research Ethics Board with waiver of consent.
Funding Sources
No funding was received for this study.
Disclosures
BM serves on the editorial board of CJC Pediatric & Congenital Heart Disease. The rest of the authors have no conflicts of interest to disclose.
References
- 1.McCrindle B.W., Rowley A.H., Newburger J.W., et al. Diagnosis, treatment, and long-term management of Kawasaki disease: a scientific statement for health professionals from the American Heart Association. Circulation. 2017;135:e927–e999. doi: 10.1161/CIR.0000000000000484. [DOI] [PubMed] [Google Scholar]
- 2.McCrindle B.W., Manlhiot C., Newburger J.W., et al. Medium-term complications associated with coronary artery aneurysms after Kawasaki disease: a study from the International Kawasaki Disease Registry. J Am Heart Assoc. 2020;9 doi: 10.1161/JAHA.119.016440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.van Stijn D., Korbee J.M., Netea S.A., et al. Treatment and coronary artery aneurysm formation in Kawasaki disease: a per-day risk analysis. J Pediatr. 2022;243:167–172.e1. doi: 10.1016/j.jpeds.2021.12.054. [DOI] [PubMed] [Google Scholar]
- 4.Beiser A.S., Takahashi M., Baker A.L., Sundel R.P., Newburger J.W. A predictive instrument for coronary artery aneurysms in Kawasaki disease. US Multicenter Kawasaki Disease Study Group. Am J Cardiol. 1998;81:1116–1120. doi: 10.1016/s0002-9149(98)00116-7. [DOI] [PubMed] [Google Scholar]
- 5.Kim T., Choi W., Woo C.W., et al. Predictive risk factors for coronary artery abnormalities in Kawasaki disease. Eur J Pediatr. 2007;166:421–425. doi: 10.1007/s00431-006-0251-8. [DOI] [PubMed] [Google Scholar]
- 6.Chen E., Miller G.E. Socioeconomic status and health: mediating and moderating factors. Annu Rev Clin Psychol. 2013;9:723–749. doi: 10.1146/annurev-clinpsy-050212-185634. [DOI] [PubMed] [Google Scholar]
- 7.Pérez-Hernández B., Rubio-Valverde J.R., Nusselder W.J., Mackenbach J.P. Socioeconomic inequalities in disability in Europe: contribution of behavioral, work related and living conditions. Eur J Public Health. 2019;29:640–647. doi: 10.1093/eurpub/ckz009. [DOI] [PubMed] [Google Scholar]
- 8.Wändell P., Carlsson A.C., Gasevic D., Sundquist J., Sundquist K. Neighbourhood socio-economic status and all-cause mortality in adults with atrial fibrillation: a cohort study of patients treated in primary care in Sweden. Int J Cardiol. 2016;202:776–781. doi: 10.1016/j.ijcard.2015.09.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zghebi S.S., Steinke D.T., Carr M.J., et al. Examining trends in type 2 diabetes incidence, prevalence and mortality in the UK between 2004 and 2014. Diabetes Obes Metab. 2017;19:1537–1545. doi: 10.1111/dom.12964. [DOI] [PubMed] [Google Scholar]
- 10.Lord S., Manlhiot C., Tyrrell P.N., et al. Lower socioeconomic status, adiposity and negative health behaviours in youth: a cross-sectional observational study. BMJ Open. 2015;5 doi: 10.1136/bmjopen-2015-008291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Minich L.L., Sleeper L.A., Atz A.M., et al. Delayed diagnosis of Kawasaki disease: what are the risk factors? Pediatrics. 2007;120:e1434–e1440. doi: 10.1542/peds.2007-0815. [DOI] [PubMed] [Google Scholar]
- 12.Dionne A., Bucholz E.M., Gauvreua K., et al. Impact of socioeconomic status on outcomes of patients with Kawasaki disease. J Pediatr. 2019;212:87–92. doi: 10.1016/j.jpeds.2019.05.024. [DOI] [PubMed] [Google Scholar]
- 13.Petrou P., Samoutis G., Lionis C. Single-payer or multipayer health system: a systematic literature review. Public Health. 2018;163:141–152. doi: 10.1016/j.puhe.2018.07.006. [DOI] [PubMed] [Google Scholar]
- 14.Forward sortation area boundary file. Statistics Canada; 2016. 2016. https://www150.statcan.gc.ca/n1/en/catalogue/92-179-X Available at:
- 15.McCrindle B.W., Li J.S., Minich L.L., et al. Coronary artery involvement in children with Kawasaki disease: risk factors from analysis of serial normalized measurements. Circulation. 2007;116:174–179. doi: 10.1161/CIRCULATIONAHA.107.690875. [DOI] [PubMed] [Google Scholar]
- 16.Low income cut-offs for 2008 and low income measures for 2007. Statistics Canada; 2009. Available at: http://www.statcan.gc.ca/pub/75f0002m/2009002/s2-eng.htm. Accessed July 26, 2021.
- 17.Zajacova A., Lawrence E.M. The relationship between education and health: reducing disparities through a contextual approach. Annu Rev Public Health. 2018;39:273–289. doi: 10.1146/annurev-publhealth-031816-044628. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Martin D., Miller A.P., Quesnel-Vallée A., et al. Canada's universal health-care system: achieving its potential. Lancet. 2018;391:1718–1735. doi: 10.1016/S0140-6736(18)30181-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Education reference guide Census of Population 2016. Statistics Canada; 2016. 2016. https://www12.statcan.gc.ca/census-recensement/2016/ref/guides/013/98-500-x2016013-eng.cfm Available at:
- 20.Manlhiot C., Mueller B., O'Shea S., et al. Environmental epidemiology of Kawasaki disease: linking disease etiology, pathogenesis and global distribution. PLoS One. 2018;13 doi: 10.1371/journal.pone.0191087. [DOI] [PMC free article] [PubMed] [Google Scholar]