Self-reported financial barriers to care among patients with cardiovascular-related chronic conditions

Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

David J.T. Campbell, Kathryn King-Shier, Brenda R. Hemmelgarn, Claudia Sanmartin, Paul E. Ronksley, Robert G. Weaver, Marcello Tonelli, Deidre Hennessy and Braden J. Manns

Even in a system with universal health care, financial barriers may reduce access to medical care and can affect health outcomes.  Because many provincial health insurance plans do not cover prescription drugs for all citizens,Note1 some patients may not obtain indicated medications.Note2 Even those with drug coverage may still bear financial burdens such as copayments and deductibles.Note3

Financial barriers may be especially important for people with cardiovascular-related chronic conditions, who typically require ongoing monitoring and long-term use of costly prescription medications.Note4,Note5 People with low socio-economic status Note6,Note7 are particularly vulnerable to experiencing poor health outcomes.Note8,Note9

Earlier research has examined the affordability of health care in Canada, including types of provincial drug coverage,Note10 out-of-pocket expenditures,Note1 and cost-related non-adherence.Note2 Several studies indicate that provision of drug insurance, particularly to people with cardiovascular conditions, increases the use of preventive medications.Note11-14 Others have shown that such insurance is associated with improved clinical outcomes, for instance, with respect to  blood pressureNote14 and cardiac events.Note3 Some studies also suggest that the outcome of insurance provision may be reduced costs in other health care areas, such as emergency department visits and hospitalizations.Note13 However, relationships between different types of financial barriers and care indicators or adverse outcomes in patients with chronic conditions are not clear.

A survey conducted in the four western provinces collected data on self-reported financial barriers to health care among adults who reported that they had been diagnosed with hypertension, diabetes, heart disease and/or stroke.  The goals were to determine the prevalence of various types of self-reported financial barriers, factors associated with experiencing  financial barriers, and implications of these barriers. The present analysis  sought to determine if financial barriers were associated with lower use of guideline-recommended medications, stopping use of one or more recommended medications, and greater out-of-pocket expenditures. The association between self-reported financial barriers and chronic condition-related emergency department visits and hospitalization was also assessed.

Data and methods

Survey design

From February 1 to March 31 2012, Statistics Canada conducted a survey designed by the Interdisciplinary Chronic Disease CollaborationNote15—“Barriers to Care for People with Chronic Health Conditions” (BCPCHC)Note16—using computer-assisted telephone interviews. The survey  collected information about aspects of care and potential barriers, including financial barriers, access barriers, geographical barriers, and health care system-related barriers among people with at least one of four cardiovascular-related chronic conditions. With respondent consent, their BCPCHC  responses were linked to their 2011 Canadian Community Health Survey (CCHS) responses to provide detailed information about hospitalizations and socio-demographic and health characteristics.

The present study was approved by the Conjoint Health Research Ethics Board of the University of Calgary and the Health Research Ethics Board of the University of Alberta. In accordance with Statistics Canada procedures, the survey underwent pilot testing and multiple revisions to ensure that the questionnaire content was relevant to the patient population.Note16

Respondents

The initial sampling frame for the BCPCHC consisted of all 2011 CCHS respondents who: (1) resided in British Columbia, Alberta, Saskatchewan or Manitoba; (2) were aged 40 or older; and (3) self-reported having at least one of heart disease, stroke, diabetes, or hypertension.

Of an initial 4,331 CCHS respondents, 2,582 were considered “in scope” for the BCPCHC; that is, they had not participated in another Statistics Canada CCHS sub-survey in 2011; they confirmed that they had a cardiovascular-related chronic condition; and they agreed to have their BCPCHC results linked to their CCHS responses. From this sampling pool, 2,316 were randomly selected for the BCPCHC. The response rate was 80%, yielding a final study population of 1,849.

Variables

Three types of self-reported financial barriers were examined:  general financial barriers (difficulty paying for services, equipment or medications); financial barriers to accessing medications (not obtaining medications because of cost); and lack of insurance coverage for prescription medications (Appendix).

Respondents were asked if they used statins and acetylsalicylic acid (Aspirin/ASA) regularly in the past month. These drugs are recommended for most patients at high risk of cardiovascular events.Note17,Note18 Compared with ASA (about $50 a year), statins are relatively expensive ($500 to $700 a yearNote19), so patients reporting financial barriers may be less likely to use them on a daily basis. ASA use was selected as a secondary outcome and negative control; it was hypothesized that financial barriers would be less relevant for the use of ASA than for the use of statins. Therefore, people with cardiovascular-related chronic conditions who were not using ASA might have reasons other than cost for non-use.

To identify respondents who, ideally, should be taking these medications, a subset of the total study population at higher cardiovascular risk was defined using an algorithm modified from the Canadian Diabetes Association (CDA) practice guidelinesNote17 and Canadian Hypertension Education Program (CHEP) recommendations.Note18 Although some of the clinical information needed to make this determination (for instance, blood pressure control, family history, symptoms) was lacking, available data were used to generate a high-risk cohort. This was comprised of people who self-reported having heart disease or stroke; those who had both hypertension and diabetes; current smokers; those with hypertension only who were aged 55 or older; and those with diabetes only who were aged 45 or older (men) or 50 or older (women).

Non-adherence to prescribed medication was defined as respondents reporting that they had stopped taking one or more of their medications for at least one week in the past 12 months.

Respondents reported their estimated 12-month out-of-pocket expenses for medications. The estimate was extrapolated from a three-month recall if this was easier for the respondent than recalling the total amount over 12 months.

Respondents were asked if, in the past 12 months, they had visited an emergency department or been hospitalized for care of their cardiovascular-related chronic condition. Their answers were used to derive a binary variable differentiating those who had had an emergency department visit or who had spent at least one night in hospital from those who did neither.

Based on the literature, a list of potential covariates was generated to identify variables that were theoretical confounders (potentially independently associated with both the exposure and the outcome): age, sex, multimorbidity (having more than one of the four selected chronic conditions), having a regular family physician, and selected socio-demographic characteristics. Because multidisciplinary teams have been shown to improve the quality of care for people with chronic conditions,Note20 the models also controlled for involvement of a non-physician health care provider.

Statistical analyses

All analyses were performed using STATA 11.0 (Statacorp, College Station, Texas). Frequency weights were calculated by Statistics Canada to account for non-representative sampling and to reflect the adult population with chronic conditions in the four western provinces.Note21 All percentages and models used these weights, and proportions were stratified by the presence of multimorbidity. Bootstrapping procedures with 500 replications were used to calculate standard errors and confidence intervals around the estimates. As recommended by Statistics Canada, the coefficient of variation was used to determine the reliability of reported percentages.Note22

Log-binomial regression models were used to calculate unadjusted and adjusted prevalence rate ratios (PRR), using a step-wise approach. Initially, bivariate models were generated to test independent associations between the covariate and outcome of interest. Each covariate that was significant in bivariate modelling was then tested in simplified models that included the covariate of interest, the exposure, and the outcome. If the covariate was a potential confounder (change in point estimate of at least 10%), it was included in the full model. The final model was obtained through a process of backwards elimination. Covariates with p-values > 0.10 on the Wald test were removed sequentially; if there was no evidence of confounding (point estimate did not change by at least 10%), the variable was eliminated from the model. Each variable in Table 1 was considered a possible confounder and tested in this way. In addition, joint confounding and effect measure modification were assessed using modelling with interaction terms between lack of drug insurance and province, lack of drug insurance and age, and multimorbidity and age.

Results

Study population

Just over half (52%) of BCPCHC respondents were aged 65 or older; two-thirds (67%) were married; and a large majority (83%) lived in urban areas (Table 1). Respondents who had more than one of the selected cardiovascular-related conditions (multimorbidity) were more likely to be men, to live in lower-income households, and to have relatively low levels of education.

Prevalence of barriers

An estimated 12% of respondents reported general financial barriers to health care; 4% reported financial barriers specifically to accessing medications; and 14% lacked prescription drug insurance  (Table 2).

The prevalence of financial barriers varied by the number of chronic conditions respondents reported. Compared with people who had just one condition, those with at least two were more likely to report general financial barriers (PRR: 2.7; 95% CI: 1.7-4.2) and financial barriers to accessing medications (PRR: 7.6; 95% CI: 2.7-21.7), but they were no more likely to lack drug insurance (PRR: 1.0; 95% CI: 0.7-1.4) (Table 2).

Several patient factors were associated with reporting general financial barriers and lacking drug insurance (Table 3).  Lower income, obesity, and non-White ethnicity were each associated with general financial barriers. Living in a rural area or in a province other than Alberta was associated with not having drug insurance.

Medication use

Almost half (48%) of respondents were using statins (Table 4). The prevalence of statin use was similar among people who did and did not report  general financial barriers (Table 5). However, those who reported difficulty accessing medications because of cost were 50% less likely to be taking statins (adjusted PRR 0.5; 95% CI: 0.3-0.9), compared with those who did not report this difficulty. People without drug insurance were nearly 30% less likely to take statins (adjusted PRR 0.7; 95% CI: 0.6-0.9) than were those with drug insurance.

When the influence of all the variables in the analysis was taken into account, none of the three types of self-reported financial barriers was associated with ASA use (Table 5).

An estimated 13% of respondents had stopped taking one or more prescription medications (Table 4).

A financial obstacle to accessing medications was the only barrier associated with having stopped using prescribed medications for more than a week in the past year (adjusted PRR: 3.5; 95% CI: 1.7-7.3) (Table 5).

Out-of-pocket drug expenditures

Mean annual out-of-pocket drug expenditures amounted to $539 (Table 6), but were significantly higher for people with more than one chronic condition ($714 versus $460). However, this difference reflected the situation among seniors; for people younger than age 65, out-of-pocket drug expenditures did not differ significantly by the number of cardiovascular-related conditions they reported.

Compared with people who reported no barriers, those who reported general financial barriers had significantly higher out-of-pocket expenditures on medications ($466 versus $1,077, p = 0.004) (Table 4). And compared with those who had drug insurance, those who did not reported higher out-of-pocket spending on medications ($480 versus $894, p = 0.003)

Among people reporting financial barriers to medications specifically, the trend was toward higher out-of-pocket expenditures compared with people who did not report such barriers ($900 versus $524). The difference, however, was not statistically significant (p = 0.139), likely because of the small number of respondents identifying this barrier.

Emergency department visits and hospitalization

Overall, 9% of respondents reported an emergency department visit or hospitalization related to their cardiovascular condition in the past year (Table 4). Emergency department visits or hospitalizations were 70% more likely among those who reported a general financial barrier (adjusted PRR: 1.7; 95% CI: 1.0-2.9) than among those who did not (Table 5). However, self-reported financial barriers to medications or lacking drug insurance were not associated with an increased risk of an emergency department visit/hospitalization. Interaction terms between multimorbidity and age, province and age, and province and insurance status were not significant in any of the models.

Discussion

In the four western provinces, self-reported general financial barriers to health care and lacking drug insurance were relatively common among people aged 40 or older with cardiovascular-related chronic conditions (12% to 14%).

A Commonwealth Fund/Health Council of CanadaNote25 survey found a similar percentage of people with chronic conditions reporting financial barriers. The results of the BCPCHC survey provide additional information about health outcomes. Reporting financial barriers seemed to be clinically relevant, as indicated by the 70% increased likelihood of a chronic-condition-related hospitalization or emergency department visit.  This increased risk may be the result of a lack of access to preventive measures, such as monitoring, screening, educational programs and treatments, due to the direct or incidental costs of these services. The findings of the present study are consistent with those of earlier Canadian and American research reporting that ambulatory-care-sensitive hospital admissions are more common among people with limited financial resources.Note26,Note27

Living in a rural location was significantly associated with not having drug insurance. This association may reflect a greater tendency for people in these areas to be self-employed or to work for small businesses that do not provide extended health benefits.

Compared with people who had drug insurance, those who did not were 30% less likely to use statins. People who reported a financial barrier to accessing medications were 50% less likely than those who did not to use statins. These differences may be related to the annual costs of statin use, which range from $500 to $700.Note19 By contrast, use of ASA (a much less expensive medication) did not vary by the perception of financial barriers. The disparity lends further support to the hypothesis that non-use of statins may be related to finances rather than to other unmeasured factors (for instance, an aversion to taking medications), and is similar to what has been noted in other Canadian studies of initiation of statin therapy.Note28

Overall, people without drug insurance had mean annual out-of-pocket medication expenses of $894, compared with $480 for people who had drug insurance. However, individuals who did not have drug insurance were no more likely to report an emergency room visit or hospitalization.

Several characteristics were significantly associated with reporting financial barriers—multimorbidity, non-White ethnicity (including Aboriginal), and an annual household income less than $30,000. Age was not a significant modifier or confounder in any model. This may be due to a dilution of effect, because Alberta is the only province with differential insurance coverage for people aged 65 or older; seniors in the other three western provinces receive no additional insurance benefits. (The provisions of the public drug programs in the four western provinces are outlined in Appendix Text Table A.)

Limitations

This study has several limitations. BCPCHC data were self-reported and subject to the accompanying limitations. Despite adjustment, the potential for residual confounding exists, and differences in outcomes between groups may be related to unmeasured patient differences. For most variables, covariates were stratified into only two groups, and some detail may have been lost.  Moreover, some subgroups were small, thereby limiting the statistical power to detect differences.  Further, the lack of data on blood pressure control and on chronic condition severity may have resulted in misclassification of the need for statins or ASA.  The survey lacks the sensitivity to identify respondents who choose to prioritize medications and spend less on other necessities such as food.Note29 Finally, the survey was administered only to residents of the four western provinces, which potentially limits the generalizability of the findings.

Conclusion

Lack of drug coverage and general perceived cost barriers were reported by more than one in ten adults in western Canada with cardiovascular-related chronic conditions. These barriers were associated with lower use of guideline-recommended medications, an increased likelihood of non-adherence, and an increased likelihood of hospitalizations or emergency department visits. Because those who reported financial barriers comprised a heterogeneous group, and because the relationship between general financial barriers and increased risk of hospitalization is not clear, further research is warranted to better understand the association.

Acknowledgements

This research was supported by an interdisciplinary team grant from Alberta Innovates–Health Solutions, the Interdisciplinary Chronic Disease Collaboration (ICDC), which is funded through the AHFMR Interdisciplinary Team Grants Program. AHFMR is now Alberta Innovates–Health Solutions (AI-HS). David J.T. Campbell is supported by an AI-HS Clinician Fellowship award. Braden J. Manns, Kathryn King-Shier and Brenda R. Hemmelgarn are supported by AI-HS salary awards. Brenda R. Hemmelgarn is also supported by the Roy and Vi Baay Chair in Kidney Research. Paul E. Ronksley is supported by a Frederick Banting and Charles Best Canada Graduate Scholarship from the Canadian Institutes of Health Research. Marcello Tonelli is supported by a Canada Research Chair.  Braden J. Manns, Brenda R. Hemmelgarn and Marcello Tonelli were supported by an alternative funding plan from the Government of Alberta and the Universities of Alberta and Calgary.

Start of text box

Appendix

Barriers to Care for People with Chronic Health Conditions Survey questions

Exposures/Financial barriers

Perceived general financial barrier:

  • “In the past 12 months, how often did you have difficulty paying for services, equipment, medications for chronic conditions?”
    • Yes = always, often, sometimes
    • No = rarely, never

Financial barrier to drugs

  • “In the past 12 months, how often wer e you unable to access medications for your chronic condition due to cost?”
    • Yes =  always, often, sometimes
    • No = rarely, never

Lack of drug insurance

  • “Do you currently have insurance that covers all or part of the cost of prescription medications?”

Quality indicators

Use of statins

  • “In the past month, did you take prescription medication such as Crestor, Lipitor or Zocor to control blood cholesterol?”

Use of acetylsalicylic acid (ASA)

  • “In the past month, did you take aspirin or other ASA (acetylsalicylic acid) medication every day or every second day?”

Adherence to prescribed medication:

  • “Over the past 12 months, have you ever stopped taking one or more of your drugs as prescribed for a week or more?”

Out-of-pocket expenditures

  • “In the past 12 months, what were the out-of-pocket costs for your prescribed medicines, drugs and pharmaceutical products? Estimate the costs incurred by you. Include amounts not covered by insurance, such as exclusions, deductibles and expenses over limits. Exclude payments for which you have been or will be reimbursed. If it is easier for you, estimate your out-of-pocket costs in a 3-month period.”

Chronic-disease-related emergency room visit

  • “How many times have you personally used a hospital emergency department for your condition in the past 12 months?”

Chronic-disease-related hospital admission

  • “In the past 12 months, have you been a patient overnight in a hospital for your condition?”

Covariates

Regular medical doctor

  • “Do you have a regular medical doctor?”

Self-perceived health

  • “In general, would you say your health is . . . ?”
    • At least very good = excellent, very good
    • Less than very good = good, fair, poor

Self-perceived mental health

  • “In general, would you say your mental health is . . . ?”
    • At least very good = excellent, very good
    • Less than very good = good, fair, poor

Inadequate health literacy

  • Imputed variable based on validated 3-item questionnaireNote24

End of text box

Date modified: