If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to: Laura C. Rosella, PhD, MHSc, Dalla Lana School of Public Health, University of Toronto, 155 College St, Health Sciences Building 6th Floor, Toronto, Ontario, Canada M5T 3M7
Public Health OntarioInstitute for Clinical Evaluative SciencesDalla Lana School of Public HealthInstitute of Health Management Policy and Evaluation, University of Toronto
Institute for Clinical Evaluative SciencesInstitute of Health Management Policy and Evaluation, University of TorontoToronto Rehabilitation Institute, Toronto, Ontario, Canada
Healthcare spending occurs disproportionately among a very small portion of the population. Research on these high-cost users (HCUs) of health care has been overwhelmingly cross-sectional in nature and limited to the few sociodemographic and clinical characteristics available in health administrative databases. This study is the first to bridge this knowledge gap by applying a population health lens to HCUs. We investigate associations between a broad range of SES characteristics and future HCUs.
Methods
A cohort of adults from two cycles of large, nationally representative health surveys conducted in 2003 and 2005 was linked to population-based health administrative databases from a universal healthcare plan for Ontario, Canada. Comprehensive person-centered estimates of annual healthcare spending were calculated for the subsequent 5 years following interview. Baseline HCUs (top 5%) were excluded and healthcare spending for non-HCUs was analyzed. Adjusted for predisposition and need factors, the odds of future HCU status (over 5 years) were estimated according to various individual, household, and neighborhood SES factors. Analyses were conducted in 2014.
Results
Low income (personal and household); less than post-secondary education; and living in high-dependency neighborhoods greatly increased the odds of future HCUs. After adjustment, future HCU status was most strongly associated with food insecurity, personal income, and non-homeownership. Living in highly deprived or low ethnic concentration neighborhoods also increased the odds of becoming an HCU.
Conclusions
Findings suggest that addressing social determinants of health, such as food and housing security, may be important components of interventions aiming to improve health outcomes and reduce costs.
Introduction
Health services utilization and subsequent healthcare spending occurs disproportionately among a very small portion of the population. This phenomenon, commonly referred to as the “high-cost users” (HCUs) of health care, is seen across healthcare sectors and health systems globally.
Wodchis WP. The concentration of health care spending: little ado (yet) about much (money). Paper presented at: Canadian Association for Health Services and Policy Research (CASHPR) 2012 Conference; May 30, 2012; Montreal. www.cahspr.ca/web/uploads/presentations/C6.1_Walter_P._Wodchis.pdf.
have shown that since at least the 1970s, the top 5% of healthcare users in the U.S. have consistently accounted for more than 50% of expenditures, whereas the bottom 50% incur less than 5% of costs. In Canada, the numbers are very much the same.
Wodchis WP. The concentration of health care spending: little ado (yet) about much (money). Paper presented at: Canadian Association for Health Services and Policy Research (CASHPR) 2012 Conference; May 30, 2012; Montreal. www.cahspr.ca/web/uploads/presentations/C6.1_Walter_P._Wodchis.pdf.
During the 2007–2008 fiscal year, the top 5% of healthcare users in Canada’s most populous province, Ontario, accounted for two thirds of total of healthcare expenditures; a mere 1% was spent on the bottom 50%.
Wodchis WP. The concentration of health care spending: little ado (yet) about much (money). Paper presented at: Canadian Association for Health Services and Policy Research (CASHPR) 2012 Conference; May 30, 2012; Montreal. www.cahspr.ca/web/uploads/presentations/C6.1_Walter_P._Wodchis.pdf.
Given increasing concern over the sustainability of the healthcare system, quality of care, and patient outcomes, there has been a renewed interest in HCUs in recent years.
Efforts to address high-cost use have primarily been limited to case management of high-risk patients, such as the elderly and those with multi-morbidity.
observed a distinct relationship between SES and frequent users of primary care in the 1994 National Population Health Survey. It is reasonable to expect that the SES–health gradient fundamental to so many inequities in health outcomes would extend to HCUs.
To date, HCU research has been overwhelmingly cross-sectional and clinically focused, limited to the few sociodemographic characteristics available in administrative databases or relying on existing ecological measures. This approach has provided limited opportunities to study the drivers of high-cost use, especially those that may be intervened upon in the community before costs accrue.
These considerations have important implications for how the link between SES and high-cost use is understood and for policies and programs targeting HCUs.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots.”.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots.”.
This study is the first to bridge this knowledge gap by investigating a broad range of individual, household, and neighborhood SES characteristics and high-cost use among a large, longitudinal, population-based cohort of non-HCU adults in Ontario, Canada.
Methods
Data Sources
Participants from two cycles of the Canadian Community Health Survey (CCHS), Cycle 2.1 (2003–2004) and Cycle 3.1 (2005–2006), were linked to population-based health administrative databases for Ontario, Canada. The CCHS is a cross-sectional survey administered by Statistics Canada, representative of 98% of the Canadian population aged ≥12 years living in private dwellings.
All permanent residents of Ontario are covered by a single-payer insurance system referred to as the Ontario Health Insurance Plan (OHIP) and all related healthcare encounters are recorded in health administrative databases. Healthcare spending was calculated for all key sources of healthcare expenditure, including hospital admissions, same day surgery, emergency department visits, physician payments, rehabilitation, complex continuing care, and prescriptions filled for individuals eligible for the Ontario Drug Benefit (seniors, individuals living in long-term care or special care homes, residents receiving social assistance, and those with high relative drug costs). Healthcare spending was calculated using a person-centered methodology developed for Ontario administrative data, which is valid for costing from 2003 onward.
Wodchis WP, Bushmeneva K, Nikitovic M, McKillop I. Guidelines on Person-level Costing Using Administrative Databases in Ontario. Working Paper Series. Vol. 1. Toronto: Health System Performance Research Network; 2013.
Annual per-person costs were calculated for each of the 5 years following CCHS interview, and individuals were ranked according to percentiles of cost within each CCHS cohort; HCUs were defined as those who ranked in the top 5% according to total annual spending.
Participants were excluded if they could not be successfully linked to administrative data or were OHIP-ineligible for the entire observation window. Given differences in utilization patterns and profiles between adults and children, the sample was restricted to adults (aged ≥18 years).
Prior healthcare utilization was drawn from administrative data captured in the 2 years prior, and Aggregated Diagnosis Groups (ADGs) scores, a measure of comorbidity, were calculated.
Unlike other marginalization indices, ON-Marg is multifaceted and measures four area-level dimensions of marginalization: residential instability, material deprivation, ethnic concentration, and dependency.
All other variables were captured from CCHS interview.
Household food security status is a derived variable, where a household is considered food insecure if any of the following conditions were met in the 12 months prior: (1) A member of the household worried there would not be enough food to eat because of lack of money; (2) a member’s food intake was reduced as a result of there being not enough food to eat; or (3) the desired variety or quality of foods was not eaten because of lack of money.
In Cycle 2.1, food security was defined dichotomously (secure versus insecure); however, in Cycle 3.1, the definition was expanded to look at levels of food security: food secure versus food insecure without hunger, with moderate hunger, and with severe hunger. The CCHS uses a validated household food security model, which has been described in detail elsewhere.
The conceptual model of this study and variable selection was guided by the Andersen–Newman Framework of the individual determinants of health services utilization, a theoretic framework that states that the amount of health services a person uses is dependent on the predisposing factors of demographics (e.g., age, sex, marital status); social structure (e.g., education, ethnicity, residential mobility); and values and beliefs (e.g., attitude toward health services) and enabling factors: family (e.g., income, access to regular source of health care) and community (e.g., urban/rural); and illness level both perceived (e.g., general health, diagnoses) and evaluated (e.g., diagnoses).
This framework of health services utilization was selected given its relevance to a population health, and more specifically a social determinants of health, approach to health services research.
Statistical Analysis
Baseline HCU, defined as the top 5% users in the year following interview, were excluded in order to investigate the upstream determinants of being an HCU, that is, the factors associated with future high-cost use among a cohort currently not HCUs. Annual spending for non-HCUs was tracked for each of the 5 years following interview. The outcome of interest was ever being an HCU in at least one of the 5 years following interview (ever-HCUs); never-HCUs were in the bottom 95% of users in all 5 years. Unadjusted, age-adjusted, ADG-adjusted, and age sex ADG–adjusted logistic regression models were used to quantify the associations between individual, household, and neighborhood SES measures and the odds of becoming an HCU in any of the 5 years following survey.
Food security, an optional CCHS module, was asked only in certain health regions within Ontario during Cycle 2.1; however, all regions opted into this module during Cycle 3.1. A sensitivity analysis was conducted using Cycle 3.1, which would be more representative of the Ontario population and less prone to bias than the food security responses collected from Cycle 2.1, to examine any significant changes in associations between food security and ever-HCUs.
Bootstrap sampling weights provided by Statistics Canada were applied, using balanced repeated replication, to all analyses to adjust for the complex survey design of the CCHS and to produce population-based estimates.
Weighted 95% confidence limits were calculated for all estimates.
All statistical analyses were performed in 2014 using SAS, version 9.3. The study design received ethical approval from the Ethics Review Boards of Public Health Ontario and Sunnybrook Health Sciences Center.
Results
The final sample consisted of 55,734 Ontario adults who were successfully linked to administrative data and were not HCUs (i.e., not in top 5%) at baseline. In the 5 years following interview, 16.3% of the cohort were HCUs in at least 1 year (95% CI=15.8%, 16.7%). Whereas 32.8% of ever-HCUs were HCUs in multiple years, only 2.4% of ever-HCUs persisted over all 5 years.
Respondents who became HCUs tended to be older, white, and female; have lower household income; and have less than post-secondary education (Table 1). They also more commonly lived in areas of higher dependency or lower ethnic concentration (Figure 1). In the unadjusted models (Table 2), HCUs were most strongly associated with low income and low household education. Strong gradients of association were noted for lower household income, higher area dependency, and lower ethnic concentration. After adjusting for age, these associations were attenuated, with the exception of homeownership and food security, which were strengthened (OR=1.55, 95% CI=1.33, 1.82). A similar attenuation was noted in the ADG-adjusted models. Clear gradients were again noted for lower household income, higher area dependence, and lower ethnic concentration.
Weighted using bootstrap weights provided by Statistics Canada to provide population estimates. Percentages represent “percent responded” unless otherwise stated.
of SES Characteristics According to 5-Year Utilization Trajectories Within a Cohort of Adult Ontarians
Overall % (95% CI)
HCU (Top 5%) in the 5 years following CCHS interview
Food security: an optional CCHS module that was only asked in certain health regions in Cycle 2.1, and included for all Ontario respondents in Cycle 3.1.
The Ontario Marginalization (ON-Marg) Index has been described in detail elsewhere.
Dependency index
Quintile 1 (lowest)
20.0 (19.3, 20.7)
14.9 (13.8, 16.0)
21.0 (20.2, 21.8)
Quintile 2
22.7 (22.0, 23.5)
20.7 (19.4, 22.0)
23.1 (22.4, 23.9)
Quintile 3
21.2 (20.5, 22.0)
20.3 (18.9, 21.6)
21.4 (20.7, 22.2)
Quintile 4
17.4 (16.8, 18.1)
19.1 (17.8, 20.4)
17.1 (16.4, 17.8)
Quintile 5 (highest)
17.5 (16.8, 18.1)
24.0 (22.8, 25.2)
16.3 (15,7, 16.9)
Material deprivation
Quintile 1 (lowest)
23.2 (22.5, 23.9)
20.7 (19.4, 22.0)
23.7 (22.9, 24.4)
Quintile 2
23.3 (22.6, 24.0)
22.2 (21.0, 23.4)
23.5 (22.8, 24.3)
Quintile 3
21.0 (20.3, 21.6)
22.0 (20.7, 23.4)
20.7 (20.0, 21.5)
Quintile 4
17.9 (17.3, 18.6)
19.5 (18.3, 20.8)
17.6 (16.9, 18.3)
Quintile 5 (highest)
13.6 (13.0, 14.1)
14.5 (13.3, 15.6)
13.4 (12.8, 14.0)
Residential instability
Quintile 1 (lowest)
26.0 (25.3, 26.7)
20.6 (19.3, 21.9)
27.0 (26.2, 27.8)
Quintile 2
20.8 (20.2, 21.5)
21.2 (19.9, 22.5)
20.8 (20.1, 21.5)
Quintile 3
16.6 (16.1, 17.2)
18.2 (17.1, 19.4)
16.3 (15.7, 17.0)
Quintile 4
18.9 (18.2, 19.6)
19.8 (18.5, 21.0)
18.7 (18.0, 19.4)
Quintile 5 (highest)
16.6 (16.1, 17.2)
19.1 (18.0, 20.3)
16.1 (15.6, 16.7)
Ethnic concentration
Quintile 1 (lowest)
12.8 (12.4, 13.1)
15.7 (14.9, 16.6)
12.2 (11.8, 12.6)
Quintile 2
16.9 (16.4, 17.4)
18.6 (17.6, 19.6)
16.6 (16.1, 17.1)
Quintile 3
18.5 (17.9, 19.1)
19.2 (18.1, 20.4)
18.4 (17.7, 19.0)
Quintile 4
20.2 (19.5, 20.9)
20.1 (18.7, 21.4)
20.3 (19.5, 21.0)
Quintile 5 (highest)
30.5 (29.8, 31.2)
25.3 (23.7, 26.9)
31.5 (30.7, 32.3)
HCU, high cost users; CCHS, Canadian Community Health Survey.
a Weighted using bootstrap weights provided by Statistics Canada to provide population estimates. Percentages represent “percent responded” unless otherwise stated.
b Other: divorced, separated, widowed, or single.
c Food security: an optional CCHS module that was only asked in certain health regions in Cycle 2.1, and included for all Ontario respondents in Cycle 3.1.
d Respondent or member of their household owns the home.
e The Ontario Marginalization (ON-Marg) Index has been described in detail elsewhere.
OR calculated using binomial logistic regression; ref group: never-HCUs. Weighted using bootstrap weights provided by Statistics Canada to provide population-based estimates.
of ever being an HCU in the 5 years post-interview
ADG, Aggregated Diagnosis Group; HCU, high cost user.
a OR calculated using binomial logistic regression; ref group: never-HCUs. Weighted using bootstrap weights provided by Statistics Canada to provide population-based estimates.
b Johns Hopkins’ ADG score, a clinical measure of morbidity and predictor of mortality.
c Adjusted for age, sex, and ADG score.
d Other: separated, divorced, widowed, or single; married includes common-law.
e Respondent or other member of household owns family dwelling.
f The Ontario Marginalization (ON-Marg) Index has been described in detail elsewhere.
In the fully adjusted model, becoming an HCU was most associated with food insecurity. Compared to individuals from food-secure households, the odds of becoming an HCU within the next 5 years were 46% greater for those living with food insecurity (95% CI=1.24, 1.71). Clear gradients were noted for higher residential instability and lower household income (only significant for low income); ethnic concentration was no longer associated with becoming an HCU.
Restricting the sample to Cycle 3.1, where the optional CCHS food security module was asked of all Ontario respondents, did not notably alter the direction or significance of the model results. In fact, the original estimates were slightly more conservative than those produced using only Cycle 3.1; the most notable change occurred for the fully adjusted estimate (5.2% increase).
Discussion
This study provides a longitudinal perspective on high-cost use trajectories that is lacking from the majority of HCU research. Compared to the cross-sectional perspective, high-cost use outcomes appear to be more transient and occur more diffusely when looking over multiple years. Additionally, this study offers a detailed characterization of the multiple SES factors associated with high-cost healthcare utilization in the context of a universal healthcare system. The results of this study support the idea of SES as a multidimensional concept operating at various levels.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots.”.
The strength and significance of these associations varied across SES measures, aggregation level, and after adjusting for confounders. The multidimensional nature of SES raises important policy questions because some factors may be more amenable to intervention than others.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots.”.
Similarly, some interventions may be more appropriate for individuals, whereas others may have greater impact at the household or community level. This also suggests that caution should be used when applying ecological measures to individual outcomes to avoid ecological fallacies. Further, SES factors not traditionally included in health services research, such as housing and food security, had important associations with high-cost use, even after adjusting for clinical and sociodemographic factors. These results support emerging evidence from programs addressing health outcomes through social interventions, such as those taking “housing first” and “hot spotting” approaches.
Clear SES–high-cost use gradients were also noted, demonstrating the importance of looking across the SES continuum, as opposed to simply high versus low SES. Notably, a clear protective gradient was associated with areas of higher ethnic concentration, which may suggest that a healthy immigrant effect may be important, or the finding may be a result of differences in culture and beliefs surrounding health services use or barriers to accessing services.
Overall, the SES–high-cost use relationships observed in this study remained even after adjusting for predisposition and illness level, suggesting that these differences are not merely due to differences in health status or healthcare-seeking behavior but are the result of SES disparities and access inequities.
This study is the first to bridge this knowledge gap by applying a population health lens to comprehensively investigate the issue of HCUs according to multiple measures of SES, demographics, and health status using a population-based source of administrative data linked to health survey data. To date, high-cost use research has lacked the ability to investigate the multidimensional nature of SES owing to the scope of data available in administrative databases. As a result, many previous studies have captured SES using only income or education, or have relied on area-level SES measures.
Despite this limitation, several studies have found an association between SES and high health services use. For example, a 2009 study by Lemstra and colleagues
found that low-income residents of Saskatoon, Saskatchewan, had overall healthcare costs 35% higher than high-income residents. The link between SES and high cost or frequent healthcare use has been corroborated in other Canadian provinces, within the U.S., and other countries.
Wodchis WP. The concentration of health care spending: little ado (yet) about much (money). Paper presented at: Canadian Association for Health Services and Policy Research (CASHPR) 2012 Conference; May 30, 2012; Montreal. www.cahspr.ca/web/uploads/presentations/C6.1_Walter_P._Wodchis.pdf.
Additionally, low SES has also been linked to increased risk of preventable hospitalizations, higher rates of hospitalization and longer stays if admitted, frequent emergency department use, and poor continuity of care.
More generally, it is known that SES affects health outcomes, and subsequently healthcare use and costs, through a diverse array of mechanisms. Factors besides income and education, such as food security, housing, and neighborhood, can affect health.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots.”.
stresses food security as an important public health issue, describing food insecurity as a deprivation of a basic human need, which has lasting effects on an individual’s health over the life course.
The results of this study suggest that multiple SES factors may have important roles in a patient’s progression to high-cost and frequent health services use. However, the health-related impacts of SES separate from those of mediating factors, such as risky lifestyle behaviors, are complex, and longitudinal associations with high-cost use trajectories over longer time frames are not yet clear, making it difficult to determine optimal policies and interventions for prevention.
These relationships should receive consideration when interpreting high-cost use research or designing related interventions.
Limitations
This study is strengthened by the novel use of a large, nationally representative survey linked to a comprehensive and longitudinal source of administrative data, allowing us to create a population-based sample of non-HCUs and more broadly characterize the upstream social determinants of becoming an HCU. Further, this study is the first to track transitions into high-cost use states over multiple years, and one of the few to define HCUs according to comprehensive estimates of total healthcare spending (i.e., across all key service sectors as opposed to one sector). However, there are some limitations that must be mentioned. Foremost, the CCHS sampling frame is limited to Canadians living in private dwellings. Despite being representative of approximately 98% of the population, this excludes Canadians residing in institutions, on Aboriginal reserves, and full-time members of the Canadian Forces, in addition to those living in certain remote areas.
Therefore, Ontarians residing in long-term care facilities, mental health institutions, or hospitals at the time of interview are excluded from these analyses. Importantly, HCUs were excluded at baseline and any subsequent transfers into these facilities would be captured through the databases in this study. Thus, this would not be a considerable limitation to this study design. Similarly, homeless individuals and First Nations people living on reserve are not represented by the CCHS.
These individuals, who are likely at higher risk of becoming HCUs given the link with SES, are important areas for future high-cost use research. Further, investigating high-cost use trajectories over longer time frames (i.e., decades), high-cost use persistence, and high-cost use transitions may provide additional insights into the relationship between SES and high-cost use outcomes.
There were limits in the variables available for inclusion and how they may be categorized. For example, investigating levels of food insecurity, compared to a dichotomous indicator, may provide additional insights or have important policy implications. Similarly, the ability to adjust for “values and beliefs” was limited, outside of what can be inferred through prior utilization history (ADGs). Furthermore, because the CCHS collects self-reported data, there is the potential for social desirability bias.
There is also the potential that respondents were previously HCUs or already on this trajectory and the nature of their prior health may have influenced some factors (e.g., low income as a result of not being able to work because of prolonged illness). However, baseline HCUs were excluded and, with the exception of income, many of the SES variables investigated here are typically stable over time.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Although comprehensive, health expenditures were limited to only those covered by Ontario’s universal health insurance plan. Except for eligible members of the adult population (e.g., those aged >65 years, receiving government assistance, or with specific diseases), OHIP coverage excludes items such as prescription drug costs, dental care, and allied health services.
Wodchis WP, Bushmeneva K, Nikitovic M, McKillop I. Guidelines on Person-level Costing Using Administrative Databases in Ontario. Working Paper Series. Vol. 1. Toronto: Health System Performance Research Network; 2013.
Compared with costs associated with hospital care and physician services, these represent a relatively smaller proportion of spending. It should be noted that although universal coverage has eliminated many barriers to accessing health care, some obstacles remain, notably geography. For instance, remote rural areas are known to have poorer access to health services than major urban centers.
Indeed, a moderate, albeit statistically significant, association between rural residence and becoming an HCU was observed in this study, even after adjusting for age, sex, and ADG.
Conclusions
HCUs are often framed by cost savings and sustainability concerns, but the underlying issue is also one of health disparities and social inequities.
Solar O, Irwin A. A conceptual framework for action on the social determinants of health. Social Determinants of Health Discussion Paper 2 (Policy and Practice). Geneva: WHO, 2010.
This study suggests that future high-cost healthcare use is strongly associated with multiple dimensions of SES, including income, education, homeownership, food security, and neighborhood marginalization. The root causes of high-cost and frequent healthcare use are entrenched in SES and are often overlooked in health services research. A multitude of factors affecting the SES–health gradient lie outside of the healthcare system
; in order to effect change before patients become HCUs or begin down that trajectory of use, an upstream population-based lens must be taken to this problem. Indeed, health disparities and SES inequities are at the core of public health, and collaborative, intersectorial approaches will allow us to address high-cost use from within and outside the healthcare system by aligning public health and healthcare goals. Understanding high-cost use from a broader perspective, including a comprehensive understanding of SES, is paramount for informing policies and interventions aiming to mitigate high-cost use events and achieving the common goal of improved population health.
Acknowledgments
This project was funded by the Project Initiation Fund, an annual competitive grant provided internally through Public Health Ontario. This study was also supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). These data sets were linked using unique encoded identifiers and analyzed at the ICES. The opinions, results, and conclusions reported in this paper are those of the authors and are independent of the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred.
LR and VG conceived of the original study. LR provided epidemiologic expertise, guided the study design and analysis, interpreted the results, and critically revised the manuscript. TF helped guide the study design and analysis, interpreted the results, and drafted the manuscript. AC carried out all analyses and critically revised the final manuscript. JP and AP provided contextual insights, helped interpret the results, contributed social determinants of health expertise to discussions, and critically revised the manuscript. WP, VG, and HM contributed expert advice and provided feedback on the study design, interpretation, and the final manuscript. All authors read and approved the final manuscript.
No financial disclosures were reported by the authors of this paper.
References
Berk M.L.
Monheit A.
The concentration of health expenditures: an update.
Wodchis WP. The concentration of health care spending: little ado (yet) about much (money). Paper presented at: Canadian Association for Health Services and Policy Research (CASHPR) 2012 Conference; May 30, 2012; Montreal. www.cahspr.ca/web/uploads/presentations/C6.1_Walter_P._Wodchis.pdf.
Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic status and health: dimensions and mechanisms. Cambridge, MA: National Bureau of Economic Research (NBER), 2008. NBER Working Paper No.: 14333. www.nber.org/papers/w14333.pdf.
Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots.”.
Wodchis WP, Bushmeneva K, Nikitovic M, McKillop I. Guidelines on Person-level Costing Using Administrative Databases in Ontario. Working Paper Series. Vol. 1. Toronto: Health System Performance Research Network; 2013.
Solar O, Irwin A. A conceptual framework for action on the social determinants of health. Social Determinants of Health Discussion Paper 2 (Policy and Practice). Geneva: WHO, 2010.
00082-3&id=gr1.jpg){kind=link}