OBJECTIVES: Even in a publicly funded health care system, access to care may be related to socioeconomic status (SES). For children, delays in surgical procedures can have profound functional, social, and psychological effects with lifelong impact. The purpose of this study was to determine whether SES was related to meeting surgical wait time access targets for children. We also assessed the effects of gender, age, and distance to hospital on meeting access targets.
METHODS: Patient addresses, referral wait times, and surgical wait times were obtained for 39 287 surgical procedures between 2005 and 2011 at the Hospital for Sick Children. Using census data, we derived household income quintile, distance to hospital, and indices of social and material deprivation. These indices were correlated with the percentage of children meeting clinic referral wait time targets and receiving surgery within the Pediatric Canadian Access Targets for Surgery.
RESULTS: Across all SES quintiles, 33% of children exceeded their referral wait time targets, and 28% of children exceeded their surgical wait time targets. Indices of material or social deprivation and age did not correlate with the time from referral to clinic consultation (P = .54, .40, and .58, respectively). Gender was statistically significant (P < .001), but the difference was small (odds ratio = 0.87 for girls). Distance was also statistically significant (P = .005), and these differences translate into clinically meaningful differences in meeting wait time targets. Regarding completion of surgical procedures, material deprivation, distance, and gender did not correlate with longer wait times for surgery (P = .44, .09, .59, respectively). Social deprivation was statistically significant (P = .02) but not clinically significant. Increasing patient age was significantly associated with increased proportion of out-of-window wait times (P < .001). SES did not affect the timeliness of completion of surgery even when the urgency of the surgery (priority level based on diagnosis) was considered.
CONCLUSIONS: SES does not predict the timeliness of delivery for pediatric surgical services.
- wait times
- socioeconomic status
- pediatric surgery
- equal access health care
- access targets
- health care delivery
- publicly funded health care
- CI —
- confidence interval
- GEE —
- generalized estimating equation
- OOW —
- out of window
- OR —
- odds ratio
- P-CATS —
- Paediatric Canadian Access Targets for Surgery
- SES —
- socioeconomic status
- W1 —
- Wait 1 (the time between referral from primary care physician to the time of surgical consultation)
- W2 —
- Wait 2 (the time between surgical consultation and completion of surgery)
What’s Known on This Subject:
Socioeconomic status (SES) often influences timeliness of health care delivery, even in publicly funded systems. Children need prompt surgical care for a variety of time-sensitive developmental conditions, and children of lower SES may be especially vulnerable to delays in surgery.
What This Study Adds:
It is unknown whether a publicly funded system’s ability to provide timely pediatric surgical care is related to SES. In 39 327 consecutive surgical cases, we demonstrate that SES need not influence timeliness of surgical care in a publicly funded system.
Access is a key component of high-quality care.1 Excessive wait times, especially for surgery, are often cited as a weakness of publicly funded health care systems.2 Previous studies have suggested that even in a system with universal access, socioeconomic status (SES) may affect access to care.3 Such barriers to care might further disadvantage a vulnerable segment of the population already at risk for elevated morbidity and mortality.4–6 Lower SES has been associated with lower cancer survival rates,7 less access to cardiac procedures, greater mortality after stroke or myocardial infarction,6,8 and less access to same-day surgery.9
When children need specialized surgery, barriers to care and longer wait times have an even more profound impact. Pediatric surgical interventions are seldom discretionary.10 Timeliness of care, especially for developmental conditions, is critical, because excessive wait times adversely affect the outcomes after surgery.11 These wait times increase the risk of children needing additional surgical procedures.12,13 Furthermore, prompt access to surgical care can decrease rates of serious adverse outcomes14 and prevent unnecessary emergency department visits.14 Delays in surgical care may put the child at greater risk. The sequelae of these delays can also have profound social and psychosocial effects on these children, especially at key stages of development, while placing immense pressure on families.15 Despite the known influence of SES on adult health outcomes, there is very little research in children. Children may be particularly vulnerable to deprivation, so physicians, policymakers, and the public need to determine whether SES has an influence on access to care in a publicly funded system. The Paediatric Canadian Access Targets for Surgery (P-CATS) is a 7-stage priority classification system developed to establish wait time targets for pediatric patients to receive clinic consultation (Wait 1 [W1]) or surgical procedures (Wait 2 [W2]). The P-CATS was developed in 2005 and revised in 2008 by 11 expert panels across 11 surgical subspecialties using a nominal group technique to build consensus on acceptable surgical wait times. Clinicians used available evidence and expert opinion to establish the maximum period that children could wait for surgery. Access targets were assigned a priority classification level consistent between all subspecialties, resulting in an associated wait time target for each surgical procedure. These priority-based access targets were linked to 867 identified diagnoses in 11 pediatric surgical subspecialties.10,16 Unfortunately, 30% of children in Canada do not meet their P-CATS access targets for surgery.10 The purpose of this study was to determine whether SES was related to meeting referral and surgical wait time access targets for children. Secondary objectives included assessing for impact of patient gender, age, and distance to the health care center.
We retrospectively studied all surgical clinic referrals and surgical procedures occurring at the Hospital for Sick Children. The institutional ethics board waived the need to obtain informed consent from each person in the study because sensitive personal data were not collected. Furthermore, we ensured that no identifiable information was associated with the data set. The data set did not contain any diagnoses, conditions, or surgical treatments performed.
We analyzed 2 wait time periods associated with surgical interventions. First, we considered the time elapsed between referral from the primary care provider and the completion of the surgical consultation (W1) and second, the time between the decision to proceed with surgery and completion of surgery (W2). All referral and surgical wait times at the Hospital for Sick Children were collected in 2 databases. These databases also contain the P-CATS priority rating of each patient and condition and whether individual patients waited outside their prescribed wait time window. The outcome variable used in the analysis was whether the patient was treated within P-CATS targets (in window) or not (out of window [OOW]). This variable was established by using the P-CATS guidelines from 864 established targets across 11 surgical specialties based on expert consensus.16
All patients receiving scheduled surgery at the Hospital for Sick Children between January 2005 and December 2011 were included in the study. No patients were excluded. This population included 29 375 referrals, which resulted in 39 402 consecutive, unique surgical procedures. Some patients received multiple referrals to different surgical specialists. For each referral and surgical procedure, both the postal code and whether that patient was treated inside or outside the prescribed P-CATS wait time target were captured and linked to the data set. Patients’ actual wait times were then compared with their target wait times to determine the percentage of patients receiving surgery within the access targets.
Postal codes of individual patients were used to determine each patient’s estimated household income and their Indices of Material and Social Deprivation using 2006 Canadian Census data. Although this approach does not provide the individual household SES, the regional SES may be even more important than individual family or patient SES. The Indices of Material and Social Deprivation were previously developed for the Canadian Ministry of Health5 based on the work of Townsend and colleagues17–20 and were calculated by using postal code data for each unique census dissemination area (the smallest geographic unit for which the census releases information; each contains 400–700 people). The Index of Social Deprivation describes the percentage of adults living alone; adults who are separated, divorced, or widowed; and families headed by a single parent. The Index of Material Deprivation describes the percentage of people lacking a high school diploma, unemployment rates, and mean income. Each index was standardized to yield a composite score, expressed as a number either <0 or >0. Higher values indicate higher levels of deprivation. For example, a value of +1 would indicate that that person is 1 SD more deprived than a person at 0 or the mean.5,21 Distance from the hospital was calculated by using the postal code of the permanent address of each patient. Gender and age of the patient at the time of surgical intervention were also collected.
Data from household income and the Indices of Material and Social Deprivation were divided into quintiles, and the proportions of patients out of target were then determined. A χ2 analysis was performed to detect any statistical differences in meeting wait time targets between groups. Both the referral data and the surgical data contained repeated measures (multiple referrals or surgeries of the same patient). Because of this clustering, a generalized estimating equation (GEE) for repeated measures was also performed for W1 and W2.
For W1 the GEE model was corrected for gender, age, social and material deprivation indices, and distance from the hospital. For the W2 the same variables were used. Income was not included in the analysis because it constitutes one-third of the material deprivation; therefore, its inclusion would confound analysis. A priori, α was set to .05.
Complete wait time data and postal code information were obtained for 39 402 surgical procedures at the Hospital for Sick Children between January 2005 and December 2011. Among clinic consultations, 71 (0.2%) of 29 375 referrals were missing income data, and 43 (0.2%) were missing data needed to calculate material and social deprivation indices. For surgical wait times, income information was incomplete for 115 (0.3%) patients, and information used to develop material and social deprivation indices was incomplete for 129 (0.3%) patients. Some of the data missing were the result of incomplete census data sets (eg, patients living in newly added postal codes would lack complete data). Cases lacking complete data were excluded.
Surgical clinic consultations (W1) were completed OOW in 33% of all patients. There were no differences in access (Table 1) between quintiles when we compared household Indices of Material and Social Deprivation for W1 (P = .68, .46, respectively). Quintile 1 represents the lowest SES level, and 5 represents the highest. Patients had multiple referrals such that of the 29 375 referrals, ∼15 000 were unique. To address this clustering, indices of Material and Social Deprivation were evaluated by using GEE analysis. There was no significant difference with respect to material deprivation, social deprivation, or age (P = .54, .40, .58, respectively). Distance from the referral clinic was statistically significant (P = .005). Although gender was statistically significant, this difference was not clinically significant, yielding an odds ratio (OR) for girls of only 0.87 (95% confidence interval [CI], 0.83–0.92). However, the effect of distance was both statistically and clinically significant because the chance of being OOW increased by 18% for every 150 km between the family’s home and the clinic.
Surgical procedures (W2) were completed OOW in 28% of all patients. There were no differences (Table 1) between quintiles when we compared Social Deprivation for W2 (P = .18). Indices of Material Deprivation alone correlated with greater W2 (P = .04). However, the absolute difference in OOW rates between Material Deprivation quintile 1 (29.3%) and quintile 5 (27.5%) was only 1.8%. In some cases, for each referral a patient may have received multiple surgeries. Of the 39 402 surgeries, ∼29 000 were unique. As a result, this data set also contained clustering. Using GEE repeated-measures modeling to account for clustering, we evaluated material and social deprivation, distance, gender, age of patients, and priority ranking of surgery. Although there were statistical differences between proportions OOW between surgical priorities (Table 2), the interaction between material and social deprivation in the model was insignificant (P = .28 and .15, respectively. This indicates that SES does not play a role within priority rankings and was removed from the GEE model for the additional analysis. Material deprivation, gender, and distance were not significant (P = .44, .49, .09). Indices of Social Deprivation alone correlated with greater W2 (P = .02) under GEE analysis. However, the differences in ORs were found between quintiles 1 and 2 and between 1 and 3, not between 1 and 4 or 1 and 5 (Table 3). This indicates that there is no difference between the most affluent and least affluent groups, representing little clinical significance. These small differences are shown in Fig 1, which illustrates the proportion of patients OOW by surgical service and social deprivation quintile. Age of the patient was also significant. The parameter estimate (Fig 2) for age indicates that for every 5-year increase in age, a child is 9% more likely to be OOW, with 95% CI between 6% and 11%. The priority of surgery was also found to be significant (P < .001; Table 2), indicating that the more urgent the surgery is, the more likely a patient is to be OOW. When the effect of SES on wait times within individual priority rankings was analyzed, there were no significant differences. Even the most urgent cases are not influenced by the patient’s demographic factors. Adjusted and unadjusted CIs, ORs, and P values are shown in Table 3. There were no differences between adjusted and unadjusted analyses. Adjusted and adjusted data for W1 and for univariate analysis were also insignificant.
Wait times were also evaluated individually within all 10 surgical subspecialties. Material and Social Deprivation indices yielded no significant differences for all 10 surgical divisions. However, there was a significant difference (P < .001) in the proportions of OOW cases between surgical specialties (Fig 1). These proportions range from ∼15% for cardiac surgery to ∼40% for ophthalmology.
SES had no clinically significant effect on either clinic referral wait times (W1) or surgical wait times (W2). Although the Index of Material Deprivation showed a statistically significant difference between quintile groups for W2 in unadjusted analysis, when we used GEE repeated-measures modeling to account for clustering and corrected for the other variables, there was no significance associated with material deprivation (P = .44). The statistical power of this study is very high, enabling detection of small differences between quintiles; however, these differences in unadjusted analyses are probably not clinically significant and, more importantly in multivariable analyses, were not statistically significant. In the evaluation of W2 using GEE repeated measures, age of patient was found to be significant along with social deprivation (P = .005) and priority level (P < .001). As previously stated, the parameter estimated for age indicates that for every 5-year increase in age, a child is ∼9% more likely to be OOW, with 95% CI between 6% and 11% (Fig 2). The significance of this finding is easily explained. As described previously, timely intervention for young children is paramount for reducing morbidity and achieving positive outcomes. Therefore, younger children are probably prioritized because of these concerns. The significance of social deprivation was demonstrated when we compared the most affluent group (quintile 1) with quintiles 2 and 3. There were no differences between these quintile groups and the least affluent quintiles 4 and 5, indicating that this difference is not clinically significant. The chance of a patient’s wait time being OOW increased with higher priority (Table 2). However, in the lowest priority (priority VI, <1 year), OOW proportions dropped to ∼3%. It is easier to complete a given intervention within the wait time window if there is more time to organize the surgery. Wait time targets in priority I are within 24 hours. This is a difficult target to reach, and there is less time to meet it. Furthermore, with already full wait lists and without widely understood triage criteria, first-come first-served scheduling results in patients not meeting access targets. There was no influence of SES within priority rankings. The differences in OOW proportions between surgical divisions warrant investigation, and these differences may be related to resource allocation within surgical departments.
Analysis of the referral wait time data indicates that gender and distance to the clinic are statistically significant factors. The wait time differences attributable to gender are small, with an OR for girls of 0.87 (95% CI, 0.83–0.92). This is not a clinically significant difference. Differences in distance were a clinically significant finding that correlates with an increase in odds of being OOW of 18% for every 150 km away from the clinic. It may be more acceptable to miss a surgical consultation appointment than it is to miss a scheduled surgery, which may push patients out of their prescribed wait time to see a surgeon. Furthermore, both planned and purposeful trips are made around consults, and referring physicians and patients often avoid referrals until absolutely necessary. In some cases this leads to exceeding W1.
Few studies have investigated the effects of SES on surgical wait times, and none have assessed pediatric surgical wait times. The literature shows that SES often predicts health outcomes. SES was found to have no influence on wait times for elective surgery,22 trauma center performance,21 major joint arthroplasty19,23 or breast cancer care.24 However, SES has been correlated with cancer survival,7 access to invasive cardiac procedures, and mortality after stroke or myocardial infarction.6,8
A recent report from 20 hospitals found that greater household income was related to greater access to same-day surgery.9 Discrepancies in the effects of SES on access to care from different medical disciplines and medical centers are difficult to reconcile but are probably related to a lack of standardized access targets and variable methods of SES measurement. In contrast, our study used the pediatric wait time access targets that are currently used as a national standard. We have also analyzed outcomes across all surgical disciplines, and we have also assessed both referral wait times (W1) and surgical wait times (W2). Furthermore, we were the first to comprehensively assess access for children. Finally, we used 3 established measures of SES.
Our results are important because they demonstrate that access to specialized surgical care for children in Ontario is equal across all socioeconomic levels, not only when assessed as household income, unemployment, and education but also when assessed as influences such as single-parent households captured in the Social Deprivation Index. Furthermore, this study shows that the effectiveness of standardized assignment of surgical wait times by P-CATS can be monitored in a real-world setting. Also, our study suggests that the present system serves the children of Ontario reasonably well and that this approach may be applied to the care of other vulnerable populations in North America and elsewhere.
This study has potential limitations. First, we used measures of SES linked to postal codes rather than individual household income. However, there is a large body of literature surrounding the development and use of methods using regional data for analysis of large populations. This method was pioneered in the United Kingdom and has become the gold standard for evaluating SES in large populations.25 Regardless, any proper classification of this kind depends on the data from which it is constructed. In our case the data were collected directly from the Canadian Census data, and these methods were adapted and validated for use by Health Canada, were published as “A Deprivation Index for Health Planning in Canada”5 and can be accessed through the Public Health Agency of Canada. These standards on which SES of large populations can be quantified have been used to analyze trauma center performance, elective surgery wait times, preterm birth outcomes, and physician resource utilization.3,5,21,22,24 Although SES of an area is not always associated with the income of an individual, these SES scores reflect the socioeconomic characteristics of localities. A person’s environment may even be a better indicator than individual income because it effectively captures the person’s environment, an important component of SES. Finally, it is important to note that in our study individual incomes or numerical quantification of material deprivation were not directly used. Patients were simply allocated within quintiles to indicate where in the socioeconomic spectrum their demographics fit. Conclusions were then drawn through statistical comparison of these quintiles, which were then compared against each other. Because there were no differences between groups, it is valid to state that SES does not affect wait times for pediatric patients.
Second, wait time targets were established based on a mixture of expert consensus and published clinical evidence. The resulting wait time targets (P-CATS) are nationally accepted targets formulated by >100 clinicians. Also, when P-CATS were evaluated empirically, exceeding the target for surgery resulted in worse outcomes and higher incidence of complications. Third, this study used proportion of patients exceeding access targets based on standardized condition-dependent wait times as an indicator of wait time success or failure. However, this method does not assess how many days OOW a given patient waited. Finally, although our study found little relationship between SES and wait times, it remains unknown whether SES was associated with the quality of surgical outcomes. Additional research will be needed to determine whether SES affects patient health and satisfaction after surgical intervention.
In summary, we found little relationship between SES and access to pediatric surgical care. Our results are important and reassuring given the risks and consequences of delays in surgery for crucially time-sensitive pediatric surgeries during the developmental years.11,12
We thank Alexander Mosoiu, Ruth Croxford, and Dan Napier for their support of this project.
- Accepted May 7, 2014.
- Address correspondence to Gregory H. Borschel, MD, Department of Surgery, Division of Plastic and Reconstructive Surgery, The Hospital for Sick Children and University of Toronto, 555 University Avenue, Toronto, Ontario, Canada, M5G 1X8. E-mail:
Mr Szynkaruk participated in the study design, collected and analyzed the data, and wrote the first draft of the manuscript; Mr Stephens participated in the study design, analyzed the data, and helped edit the manuscript; Drs Borschel and Wright participated in the study design, collected and analyzed the data, and critically reviewed the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Funding for this work was provided by the RB Salter Chair in Pediatric Surgery.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Swift EK
- Kapral MK,
- Wang H,
- Mamdani M,
- Tu JV
- ↵Murphy KGR, Glazier R, Wang X, Holton E, Fazli G, Ho M. Hospital care for all: an equity report on differences in household income among patients at Toronto Central Local Health Integration Network (TC LHIN) hospitals, 2008–2010. 2012. Available at: http://www.ices.on.ca/Publications/Atlases-and-Reports/2012/Hospital-Care-For-All
- Wright JG,
- Menaker RJ,
- Canadian Paediatric Surgical Wait Times Study Group
- Lawrence JT,
- Argawal N,
- Ganley TJ
- Ahn H,
- Kreder H,
- Mahomed N,
- Beaton D,
- Wright JG
- Jarman B
- Jarman B,
- Townsend P,
- Carstairs V
- Carstairs V,
- Morris R
- Shortt SE,
- Shaw RA
- Carstairs VMR
- Copyright © 2014 by the American Academy of Pediatrics