Abstract
Disparities in COPD health outcomes have been found with older individuals, men and those of lower socioeconomic status doing worse. We sought to determine if this was due to differences in access to COPD medications. We conducted a retrospective cohort study using population health administrative data from Ontario, Canada, a province with universal prescription drug coverage for older adults. All individuals with COPD aged 67 years and older in 2008 who were not taking inhaled long-acting bronchodilators or inhaled corticosteroids were followed for 2 years. Poisson regression was used to determine the effects of age, sex, and socioeconomic status on the likelihood of initiating one of these medications, after adjusting for potential confounders. Over the study period, 54,050 of 185,698 (29.1%) older individuals with COPD not previously taking any inhaled long-acting bronchodilators or corticosteroids were initiated on one or more of these medications. After adjustment, individuals of low socioeconomic status, measured using neighborhood income level quintiles, were slightly more likely to initiate COPD medications than those of high socioeconomic status (relative risk (RR) 1.05; 95% confidence interval (95% CI) 1.02–1.08). While men received COPD medication at a consistent rate across all age groups, the likelihood that a woman received medication decreased with increasing age. With the exception of older women, there was minimal disparity in prescription for COPD medications. Disparity in health outcomes among Ontario COPD patients is not clearly explained by differences in medication access by socioeconomic status, sex or age.
Introduction
Chronic obstructive pulmonary disease (COPD) affects more than one in four adults, is the third leading cause of death worldwide and is a leading cause of hospitalization (Citation1–3). Across jurisdictions, individuals with COPD of lower socioeconomic status, who are male, and who are older have been shown to have greater mortality, more hospitalizations and/or poorer quality of life even when other factors are adjusted for (Citation4–7). This trend has been documented in Ontario, a Canadian province with a universal health care system intended to prevent health inequities (Citation8).
The cause of worse health outcomes by socioeconomic status, sex, and age is likely multifactorial with differences in smoking patterns, biology, occupational exposures and/or other factors playing a role (Citation4). However, while most of these factors are fixed or difficult to change, one that is modifiable is COPD medication use. COPD medications have been shown to decrease hospitalizations and emergency department visits, improve quality of life, reduce symptoms and possibly reduce mortality (Citation9, Citation10). To realize their benefits, however, patients have to be able to access and use them.
A study conducted in the United States suggested that patients of lower socioeconomic status were less likely to receive COPD medications than their higher socioeconomic status counterparts (Citation11). However, it likely included patients who did not have medication insurance or had co-pays or formulary restrictions that limited their access. Whether such disparity exists among those with prescription medication coverage remains uncertain. It is also unknown, to our knowledge, if disparity in receipt of COPD medications by gender or age exists, although at least one study has found that women are less likely to receive COPD diagnoses—an important precursor to COPD medication prescribing (Citation12).
Recognizing and correcting disparities is important to ensuring fair resource distribution and optimal population health. We conducted the current study to determine if differences in access to COPD medications due to socioeconomic status, gender, and/or age in Ontario, Canada contribute to observed health disparities.
Methods
Study design
We conducted a longitudinal, population-based retrospective cohort study using health administrative data from Ontario to examine the association between socioeconomic status, age, and sex with the initiation of COPD medications. Ontario is a province of Canada with universal health care insurance and a diverse multicultural population of approximately 13 million, including about 1.8 million individuals age 65 years and older. Ethics approval was obtained from the Research Ethics Board of Sunnybrook Health Sciences Centre, Toronto, Canada.
Data sources
The Government of Ontario is the single payer for all medically necessary services for virtually all residents of Ontario across the entire spectrum of providers and hospitals. Details of these services are captured in large health administrative databases that can be linked on an individual level to provide a complete health services profile of each resident. Basic demographic and date of death information are captured in the Registered Persons Database. Prescription medications for all residents 65 years and older are captured in the Ontario Drug Benefit Program database. Hospitalizations, same day surgical procedures and emergency department visits are captured in the Canadian Institute for Health Information Discharge Abstract and National Ambulatory Care Reporting System databases. All health services provided by fee-for-service physicians and “shadow-billings” for physicians paid under alternate payment plans are captured in the Ontario Health Insurance Plan Physician Claims database. Information from all these databases date back to 1991 except for the National Ambulatory Care Reporting system, which dates back to 2002.
Study subjects
The Ontario Chronic Obstructive Pulmonary Disease database is a validated registry of all Ontario residents with COPD generated using the databases described here. It contains information on all individuals aged 35 and older with one or more physician visits and/or one or more hospitalizations for COPD. This COPD case-definition has been previously validated against a clinical reference standard and found to have a positive predictive value of 86% in adults aged 65 years or older (Citation13).
All individuals with COPD aged 67 years and older on the index date of April 1, 2008, who had not received an inhaled long-acting anticholinergic, long-acting beta agonist, or inhaled corticosteroid, but had some form of contact with the health care system, in the previous two years were included and followed until March 31, 2010. Age 67 was chosen to allow for a 2-year look-back for previous medication use. Patients who died or left the province during the follow-up period were excluded.
Exposures
Our primary exposure was socioeconomic status as indicated by neighborhood income level. The 2006 Canadian Census was used to determine the mean household income for each dissemination area of residence within Ontario (Citation14). The dissemination areas were then grouped into quintiles based on their mean household income. We used each patient's postal code, as of their index date, to determine their dissemination area of residence and corresponding neighborhood income quintile (Citation15).
Secondary measures of socioeconomic status were obtained using the 2006 Ontario Marginalization Index (OMI). The OMI uses census data to determine dimensions of social disparity for each dissemination area in Ontario. These dimensions, including material deprivation, residential instability, social dependency, and ethnic concentration, were determined separately and have been previously described elsewhere (Citation16).
We also explored age-sex related differences in the initiation of COPD medications to determine if disparities existed.
Outcome
The primary outcome was receipt of one or more prescriptions for an inhaled long-acting anticholinergic, long-acting beta agonist, or inhaled corticosteroid (including those in combination long-acting beta agonist) during the follow-up period.
Baseline characteristics
Demographic, COPD-related and general care-related characteristics, as well as information on previously prescribed medications and past co-morbidities, were obtained from the health administrative data. A COPD exacerbation was defined as receipt of a short course of an oral steroid or a respiratory antibiotic within 7 days of a physician visit for COPD. Severity of COPD at baseline (index date) was defined as: very severe (≥1 COPD-related hospitalization in the previous 2 years and/or taking long-term oral steroids and/or currently on long term oxygen therapy); severe (≥2 COPD exacerbations and/or ≥1 COPD-related emergency department visit in the previous 2 years); or moderate (≤1 exacerbation during the previous 2 years) (Citation17). A general co-morbidity score was determined for each individual using the John Hopkins Adjusted Clinical Group Case-Mix System and diagnostic information in health services records within two years prior to the index date (Citation18, Citation19).
Analysis
We determined the associations between medication use and socioeconomic status, age, and sex both unadjusted and adjusted for all potential confounding variables (demographics, COPD-related and general care characteristics, co-morbidities, and previous medication use). Since the outcome of interest was common (>25% in the cohort) using standard logistic regression to calculate odds ratios would have overestimated the relative risk of initiating COPD medication between socioeconomic status quintiles (Citation20). Therefore, we used Poisson regression with robust variance estimators, to estimate the relative risk. For all measures of socioeconomic status, those in the highest quintile represented individuals residing in areas with the lowest socioeconomic status.
Postal code information at the index date was missing for 1540 individuals (0.8%) in the cohort. These individuals were retained in the regression models, by including a separate category of ‘missing’ for each socioeconomic status measure, to determine if they differed from individuals with postal code information with respect to their association with the outcome of initiating COPD medication. After statistical adjustment, patients with this missing data did not differ (results not shown) making us confident that there was not significant bias in who was excluded and that we were studying a representative population cohort.
Secondary analyses
To determine if the prescription of one medication drove the observed results, the analyses were repeated twice with receipt of long-acting bronchodilators and then receipt of inhaled corticosteroids as separate outcomes.
To further minimize the potential for unmeasured confounding by studying a group with a more uniform severity of disease, and to examine a subgroup of patients who should have likely been prescribed all three medications of interest, the analysis was repeated in those who were determined to have severe and very severe COPD (Citation17, Citation21).
Statistical analyses were conducted using SAS 9.3 (SAS Institute Inc., Cary, NC) and all tests were two-tailed and used p < 0.05 as the level of statistical significance.
Results
There were 216,846 individuals with COPD 67 years or older who had not filled a prescription for an inhaled long-acting anticholinergic, long-acting beta agonist, or corticosteroid in the two years prior to April 1, 2008. We excluded 28,567 (13.2%) individuals who died and 2,581 (1.2%) individuals who left the province. The final cohort comprised of 185,698 COPD patients, of which 54,050 (29.1%) initiated a COPD medication during the 2-year follow-up. The distribution of COPD patient characteristics across the quintiles of socioeconomic status can be seen in Table (see Appendix Table for a complete list).
Appendix Table 1. Baseline characteristics and outcomes of older individuals with COPD by quintile of neighborhood income
When comparing those with the lowest income (most deprived) to those with the highest income (least deprived), there was a modest but significant increase (relative risk (RR) 1.05; 95% confidence interval (95% CI) 1.02–1.08) in the rate of initiating a COPD medication during follow-up after adjustment for demographics, COPD-related and general care characteristics, co-morbidities, and previous medication use (Table ). When socioeconomic status was measured by material deprivation in the adjusted model, those with the highest material deprivation were more likely (RR 1.03; 95% CI 1.00–1.06) to initiate COPD medication compared to those with the least deprivation; while those with the highest social dependency (RR 0.97; 95% CI 0.94–0.99) and highest ethnic concentration (RR 0.97; 95% CI 0.94–0.99) were less likely to initiate COPD medication compared with those with the lowest social dependency and ethnic concentration, respectively. There was no significant association between initiating COPD medication and residential instability.
In the fully adjusted analysis, the initiation of COPD medication was significantly associated with age group (older COPD patients were less likely to initiate COPD medication compared with younger patients), but not sex (Table ). However, an interaction term between age group and sex was statistically significant (p-value < 0.001). In stratified analyses, the probability of a woman receiving a COPD medication steadily declined as age increased, whereas there was no evidence of an age-related decline among men.
When receipt of a long-acting bronchodilator and receipt of an inhaled corticosteroid were treated as separate outcomes, the adjusted associations with neighborhood income for each were similar in magnitude to the primary analysis, although the statistical significance did vary in some instances (Table ). There were 27,961 COPD patients who had severe or very severe COPD. The associations observed between neighborhood income and initiating COPD medication in this sub-cohort were of similar magnitude compared with all COPD patients, although the estimates were less precise and not statistically significant. The interaction term between age and sex was significant in all of these secondary analyses, with the probability of receiving a COPD medication declining with increasing age among women but remaining stable for men.
Discussion
We conducted a population-based study of all older individuals with physician diagnosed COPD in Ontario, Canada and found that, after statistical adjustment for demographics, COPD-related and general care characteristics, co-morbidities, and previous medication use, women were less likely to receive COPD medications with increasing age, while men received COPD medications at a consistent rate across all age groups. We also found minimal differences in access to COPD medication by socioeconomic status with small observed associations between the lowest and highest quintiles (less than a 5% relative increase or decrease in COPD prescriptions being filled) for all measures after adjustment. These findings were robust when receipt of long-acting bronchodilators and inhaled corticoid steroids were treated as separate outcomes and when only patients with severe COPD were considered.
Our findings are consistent with previous studies that have found suboptimal medication prescribing among older women (Citation22–Citation24). We speculate that this might be because older women, who often live longer than men, may be more likely to be alone and less able to advocate for their optimal care.
Our findings are consistent with one study that found people with diabetes of low socioeconomic status, also insured under Ontario's universal drug programs, were not less likely to fill prescriptions for cardioprotective medications than their high socioeconomic status counterparts (Citation22). They are not, however, consistent with others that demonstrated large differences in cancer screening tests (20% less likely to receive a mammogram) (Citation25); coronary angiography following a myocardial infarction (23% less likely to receive coronary angiography) (Citation26); and inpatient rehabilitation services following a stroke (23% less likely to receive occupational therapy) among Ontario residents living in areas of lowest income compared with residents living in areas of highest income (Citation27). We hypothesize that this discrepancy is because our analyses focused on medications which, compared to these more specialized services, have a lower threshold for access.
Our findings also contrast with a previous United States study that showed that patients of lower socioeconomic status were less likely to receive tiotropium bromide (Citation11). This discrepancy may be due to differences in insurance status as all adults over 65 in Ontario have universal coverage for prescription medications and patients of all ages in the United States do not. Patients in the United States with insurance may also have high medication co-payments and/or drug formulary restrictions that the former do not. Additionally, the difference in findings may be attributable to our study using an ecological measure of socioeconomic status (neighborhood income quintile) and the study from the United States using individual-level measures of education and annual household income less than $20,000 to establish socioeconomic status.
Results from our study suggest that Ontario's Drug Benefit Program is providing relatively equitable access to COPD medications to older patients regardless of socioeconomic status and possibly improved access to men. However, this does not appear to be reducing mortality in the COPD population. In a separate post hoc analysis of all Ontario COPD patients over the age of 67 as of April 1, 2008, we calculated that the 2-year age-sex adjusted mortality rate of the study population was 12.5% higher among those of lower compared to higher socioeconomic status (see Appendix Table ). Also, the age-adjusted mortality rate was 20% significantly higher among men compared to women.
Appendix Table 2. Age and sex adjusted mortality rate of COPD patients aged ≥ 67 years by quintile of neighborhood income
Table 1. Select baseline characteristics and outcomes of older individuals with COPD by quintile of neighborhood income
Table 2. Unadjusted and adjusted rate ratios of initiating COPD medication therapy (long-acting beta-agonists, long-acting anticholinergic, or inhaled corticosteroids) by disparity measure
Table 3. Unadjusted and adjusted rate ratios of initiating COPD medication therapy (long-acting beta-agonists, long-acting anticholinergic, or inhaled corticosteroids) by age and sex
Table 4. Unadjusted and adjusted rate ratios by neighborhood income quintile for secondary analyses
Therefore, access to COPD medications does not seem to explain socioeconomic or sex disparities in COPD mortality—although it might, in part, explain age disparity. Similar results have been reported in the treatment of post-myocardial infarction, where providing patients with full prescription coverage enhanced access to cardioprotective medications compared with usual coverage, but did not reduce subsequent vascular events or need for revascularization (Citation28).
Other factors such as socioeconomic differences in smoking patterns, occupational exposure, or co-morbidity are likely more important contributors (Citation4). Alternatively, it is possible that disparity in outcomes is due to differential access to COPD medication in the many years prior to our study that we do not capture, and the balance in access to COPD medications currently observed may serve to reduce the mortality income gradient in the future. Finally, while access to COPD medication appears fairly equitable by socioeconomic status and increased in men compared to women, adherence to therapy may not be.
Several limitations of our study merit emphasis. First, we used area-level measures of socioeconomic status as individual-level measures were not available. However, this ecological measure has been shown to correlate well with individual-level measures, can account for unmeasurable neighborhood factors that also contribute to socioeconomic status and has proven to be robust in many population based peer review studies (Citation29, Citation30). Second, our study was restricted to older individuals who had prescription medication insurance and its results may not be generalizable to younger COPD patients without insurance. Third, such small relative differences in COPD medication access might be explained by unmeasured confounding, although we adjusted for multiple baseline characteristics, which served as proxies for COPD severity.
Fourth, our cohort of COPD patients was determined using a COPD health administrative case definition rather than clinical measures and misclassification could have caused modest differences in COPD medication access. Finally, our cohort did not include people who died or left the province during the two year follow-up time to ensure that all patients had an equal opportunity to initiate COPD medication, thus its results might not be generalizable to people towards the end of life or who are very mobile.
Conclusion
In summary, we conducted a population-based study of all older patients with physician diagnosed COPD in Ontario and found older women were less likely to initiate COPD medications compared to their older male counterparts and minimal disparity in receipt of COPD medications by socioeconomic status. These findings suggest that Ontario prescription medication insurance is providing relatively equitable access to medications for all COPD patients except older women. They do not, however, help explain why COPD mortality is higher among people of lower socioeconomic status and men in Ontario. Further research should explore disparity in medication access in older women and other modifiable factors (e.g., smoking patterns, occupational exposure, or co-morbidity) that may cause sex and socioeconomic disparity in mortality in the COPD population. It should also evaluate if adherence to COPD medication is consistent across sex and socioeconomic status.
Declaration of Interest Statement
All authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Disclaimer: The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred. ICES had no role in the study design, collection, analysis, or interpretation of the data, writing of the report, or in the decision to submit the report for publication.
Author contributions: All authors contributed to the study concept and design, the analysis and interpretation of the data, and the critical revision of the manuscript for important intellectual content; they have all given their final approval to submit the manuscript for publication.
Study data: The datasets used in this study were linked using unique encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences.
Funding
Dr. Gershon is supported by a Fellowship for Translational Health Research from the Physicians’ Services Incorporated Foundation, Toronto, Ontario and was supported by a New Investigator Award funded by team grant OTG-88591 from the Canadian Institutes of Health Research Institute of Nutrition, Metabolism and Diabetes while working on this study. Dr. To is supported by The Dales Award in Medical Research from the University of Toronto, Toronto, Ontario, Canada. Funding for this project was made available through the Government of 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.
References
- Buist AS, McBurnie MA, Vollmer WM, et al. International variation in the prevalence of COPD (the BOLD Study): a population-based prevalence study. Lancet 2007; 370(9589):741–750.
- Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012; 380(9859):2095–2128.
- Benady S. The Human and Economic Burden of COPD: A Leading Cause of Hospital Admission in Canada. The Canadian Thoracic Society; 2010. Ottawa, Canada.
- Prescott E, Vestbo J. Socioeconomic status and chronic obstructive pulmonary disease. Thorax 1999; 54(8):737–741.
- Gershon AS, Dolmage TE, Stephenson A, Jackson B. Chronic obstructive pulmonary disease and socioeconomic status: a systematic review. COPD 2012; 9(3):216–226.
- Han MK, Postma D, Mannino DM, et al. Gender and chronic obstructive pulmonary disease: why it matters. Am J Respir Crit Care Med 2007; 176(12):1179–1184.
- Nie JX, Wang L, Upshur RE. Mortality of elderly patients in Ontario after hospital admission for chronic obstructive pulmonary disease. Can Respir J 2007; 14(8):485–489.
- Gershon AS, Hwee J, Victor JC Wilton AS, To T. Trends in socioeconomic status–related differences in mortality among people with chronic obstructive pulmonary disease. Annals of the American Thoracic Society 2014, 11(8): 1195–1202.
- Calverley PM, Anderson JA, Celli B, et al. Salmeterol and fluticasone propionate and survival in chronic obstructive pulmonary disease. N Engl J Med 2007; 356(8):775–789.
- Barr RG, Bourbeau J, Camargo CA, Ram FS. Tiotropium for stable chronic obstructive pulmonary disease: A meta-analysis. Thorax 2006; 61(10):854–862.
- Blanc PD, Eisner MD, Yelin EH, et al. Socioeconomic gradients in tiotropium use among adults with COPD. Int J Chron Obstruct Pulmon Dis 2008; 3(3):483–490.
- Chapman KR, Tashkin DP, Pye DJ. Gender bias in the diagnosis of COPD. Chest 2001; 119(6):1691–1695.
- Gershon A, Croxford R, To T, et al. Comparison of inhaled long-acting beta-agonist and anticholinergic effectiveness in older patients with chronic obstructive pulmonary disease: a cohort study. Ann Intern Med 2011; 154(9):583–592.
- PCCF+ Version 4D User's Guide (Geocodes/PCCF). Automated Geographic Coding Based on the Statistics Canada Postal Code Conversion Files, Including Postal Codes to December 2003 (computer program). Ottawa: Statistics Canada; 2004.
- Finkelstein MM. Ecologic proxies for household income: how well do they work for the analysis of health and health care utilization? Canadian J Publ Health 2004; 95(2):90–94.
- Matheson FI, Dunn JR, Smith KL, Moineddin R, Glazier RH. Development of the Canadian Marginalization Index: a new tool for the study of inequality. Can J Publ Health 2012; 103(8 Suppl 2):S12–S16.
- Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global Strategy for the Diagnosis, Management and Prevention of COPD. http://www.goldcopd.org/uploads/users/files/GOLD_Report_2011_Feb21.pdf
- Reid RJ, MacWilliam L, Verhulst L, Roos N, Atkinson M. Performance of the ACG case-mix system in two Canadian provinces. Med Care 2001; 39(1):86–99.
- The Johns Hopkins University ACG Case-Mix System (computer program). Version 8.0. Baltimore, MD: The John Hopkins University; 2006.
- McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003; 157(10):940–943.
- O'Donnell DE, Aaron S, Bourbeau J, et al. Canadian Thoracic Society recommendations for management of chronic obstructive pulmonary disease - 2007 update. Can Respir J 2007; 14 Suppl B:5B–32B.
- Shah BR, Booth GL, Lipscombe LL, Feig DS, Bhattacharyya OK, Bierman AS. Near equality in quality for medication utilization among older adults with diabetes with universal medication insurance in Ontario, Canada. J Eval Clin Pract 2014; 20(2):176–183
- Bierman AS, Pugh MJ, Dhalla I, et al. Sex differences in inappropriate prescribing among elderly veterans. Am J Geriatr Pharmacother 2007; 5(2):147–161.
- Guaraldo L, Cano FG, Damasceno GS, Rozenfeld S. Inappropriate medication use among the elderly: a systematic review of administrative databases. BMC Geriatr 2011; 11:79.
- Borkhoff CM, Saskin R, Rabeneck L, et al. Disparities in Receipt of Screening Tests for Cancer, Diabetes and High Cholesterol in Ontario, Canada: A Population-based Study Using Area-based Methods. Can J Publ Health 2013; 104(4):e284–e290.
- Alter DA, Naylor CD, Austin P, Tu JV. Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction. N Engl J Med 1999; 341(18):1359–1367.
- Kapral MK, Wang H, Mamdani M, Tu JV. Effect of socioeconomic status on treatment and mortality after stroke. Stroke 2002; 33(1):268–273.
- Choudhry NK, Avorn J, Glynn RJ, et al. Full coverage for preventive medications after myocardial infarction. N Engl J Med 2011; 365(22):2088–2097.
- Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Amer J Publ Health 1992; 82(5):703–710.
- Alter DA, Iron K, Austin PC, Naylor CD. Socioeconomic status, service patterns, and perceptions of care among survivors of acute myocardial infarction in Canada. JAMA 2004; 291(9):1100–1107.
Appendices