Abstract
Background: Out-of-pocket drug costs lead many Canadians to engage in cost-related nonadherence to prescription medications, but our understanding of other consequences such as borrowing money remains incomplete. In this descriptive study, we sought to quantify the frequency of borrowing to pay for prescription drugs in Canada and characteristics of Canadians who borrowed money for this purpose.
Methods: In partnership with Statistics Canada, we designed and administered a cross-sectional rapid-response module in the Canadian Community Health Survey administered by telephone to Canadians aged 12 years or more between January and June 2016. We restricted our analyses to participants who responded to the question regarding borrowing money to pay for prescription drugs and used logistic regression to identify characteristics associated with borrowing.
Results: A total of 28 091 Canadians responded to the survey (overall response rate 61.8%). The weighted proportion of respondents who reported having borrowed money to pay for prescription drugs in the previous year was 2.5% (95% confidence interval 2.2%–2.8%), an estimated 731 000 Canadians. The odds of borrowing were higher among younger adults, people in poor health and people lacking prescription drug insurance. Other factors associated with increased adjusted odds of borrowing were having 2 or more chronic conditions, low household income and higher out-of-pocket prescription drug costs.
Interpretation: Many Canadians reported borrowing money to pay for out-of-pocket prescription drug costs, and borrowing was more prevalent among already vulnerable groups that also report other compensatory behaviours to address challenges in paying for prescription drugs. Future research should investigate policy responses intended to increase equity in access to prescription drugs.
All Canadian residents are insured for medically necessary hospital care and physician services, without out-of-pocket charges at the point of care. However, this insurance excludes prescription drugs used outside of hospitals. Although many Canadians have some form of prescription drug insurance through work-related benefits or public programs, others lack any drug coverage.1 In addition, Canadians with prescription drug insurance still often have to bear some or all of the costs of their drugs owing to insurance plan deductibles, copayments and gaps in insurance whereby some drugs are not covered.1,2 The out-of-pocket pharmaceutical costs borne by uninsured or underinsured Canadians can be substantial and tend to disproportionately affect potentially vulnerable populations,3–6 including children, older people, ethnic minorities and those who are socioeconomically disadvantaged.7,8 The consequences of high out-of-pocket costs vary. Patients have reported engaging in compensatory behaviours including cost-related nonadherence to prescription medications,6,9,10 which affects about 8% of Canadians with a drug prescription.11–14 Patients also make trade-offs against spending in other areas of the household budget to be able to afford prescription drugs.6,9,15–17
Patients have also reported borrowing money, including increasing credit card debt, to compensate for high drug costs and borrowing money from family and friends as a way to cope with high health care costs.10,18–21 However, most of this evidence is from the United States, which has a markedly different health insurance system and different levels of out-of-pocket drug costs from Canada. Furthermore, few studies distinguish between borrowing to pay for medication costs specifically and health care costs in general. In addition, existing studies have tended to focus on patients with specific high-cost illnesses rather than the general population.18,20–22 In sum, little is known about borrowing to finance prescription medicine in Canada. Therefore, we sought in this descriptive study to quantify the frequency of borrowing to pay for prescription drugs in Canada and to identify predictors of such activity.
Methods
Data sources
The data for this study came from a cross-sectional rapid-response module in the Canadian Community Health Survey administered by telephone to Canadians aged 12 years or more between January and June 2016. In brief, our team collaborated with Statistics Canada to design a module of questions exploring the consequences of out-of-pocket drug charges for patients. The questions in the module drew on validated questions in the literature, previous iterations of the Canadian Community Health Survey and previous qualitative work by our team.5,6,9,10,12,17,23,24 The questions were vetted and underwent qualitative testing by Statistics Canada in both French and English. The module25 was administered by Statistics Canada in accordance with Canadian Community Health Survey guidelines.26 Respondents were asked a range of sociodemographic and health questions, including whether they had borrowed money to pay for prescription medicines: “In the last 12 months, have you or anyone else in the household ever had to borrow money to pay for your prescriptions?” Following rigorous pilot testing conducted by Statistics Canada, the phrasing of this question specifically excluded examples of kinds of borrowing (e.g., from friends or family, on a credit card) to minimize confusion for the respondent. We restricted our analyses in this study to participants who responded to this question.
Statistical analysis
We calculated the total number of respondents and nationally representative estimates of the weighted proportions of the population reporting having to borrow money to pay for prescription drugs. We used multivariate logistic regression to investigate the factors associated with reporting having borrowed money to pay for prescription drugs. We included variables previously shown to be associated with difficulty paying for medications: sex, age, self-reported health status, number of chronic conditions (including arthritis, chronic obstructive pulmonary disease, diabetes, cancer, heart disease, high blood pressure and mood disorders), ethnicity, household income, education and prescription drug insurance status.5,6
We used step-wise multiple imputation methods to fill in missing data: we first imputed values for the variable missing the most data and then used the imputed values to the next highest, and so on until all missing variables had been imputed.27,28 Once the imputation for each variable was complete, we recombined the data sets to incorporate the adjustments to variance. To incorporate the complex sampling design of the Canadian Community Health Survey into our population estimates, we used survey weights provided by Statistics Canada and bootstrapping to calculate confidence intervals (CIs).29
Ethics approval
This study was approved by the University of British Columbia Behavioural Research Ethics Board.
Results
Descriptive characteristics
Responses were collected from 28 091 people, with a combined Canada-wide response rate of 61.8%. Of the respondents, 572 (2.0% of the total sample) were excluded because they responded “Don’t know” or refused to answer the question on borrowing, leaving a sample of 27 519 respondents. Data on 1 or more variables were missing for 1390 respondents (5.0%) across 4 variables (self-reported health status, out of pocket drug costs, education and prescription drug insurance), with a maximum of 2.4% of the total sample for any single variable (out-of-pocket drug costs). Table 1 presents weighted proportions of the total population: 50.7% of the respondents were female, 48.6% were less than age 45 years, 11.1% reported having fair or poor health, 20.3% had an annual household income of less than $40 000, and 20.3% reported having no prescription drug insurance.
Borrowing to finance out-of-pocket prescription drug costs
The weighted proportion of respondents who reported having borrowed money to pay for prescription medications in the previous year was 2.5% (95% CI 2.2%–2.8%) (Table 1). At the population level, this is equivalent to an estimated 731 000 Canadians (95% CI 639 000–824 000). Compared to respondents who did not report borrowing money to pay for prescription medications, those who reported borrowing tended to be younger and in poorer health, to have more chronic conditions, and to report government or no prescription drug insurance.
Figure 1 provides an illustration of some of the differences in rates of borrowing money to pay for prescription drugs across different groups in our sample. Respondents aged 19–34 years had the highest rate of borrowing (3.7%, 95% CI 2.9–4.4) (Figure 1, A). Respondents with a household income of less than $20 000 a year had more than twice the rate of borrowing (6.6%, 95% CI 5.0–8.3) than those earning $40 000 a year or more (Figure 1, B). Respondents who lacked any kind of prescription drug insurance had the highest rate of borrowing (4.6%, 95% CI 3.5–5.7), and those with employer-based drug insurance had the lowest rate of borrowing (1.3%, 95% CI 1.0–1.6) (Figure 1, C). As annual out-of-pocket costs on prescription drugs increased, respondents were more likely to report having borrowed money to pay for prescription drugs ($501–$1000: 9.5%, 95% CI 6.9–12.0; > $1000: 11.5%, 95% CI 8.6–14.3) (Figure 1, D).
Our multivariate logistic regression model showed that younger age was associated with higher odds of borrowing to pay for prescription medications. After we controlled for other factors (including insurance status and size of out-of-pocket costs), respondents aged 19–34 had more than 3.5 times the odds of borrowing (adjusted odds ratio [OR] 3.7, 95% CI 2.3–5.7) compared to those aged 45–54 (Table 2). Respondents aged more than 54 had less than half the odds of borrowing compared to those aged 45–54. Having poor self-reported health status was associated with substantially higher odds of borrowing (adjusted OR 7.7, 95% CI 3.7–15.9) compared to having excellent health. Insurance coverage was also important: respondents who reported having either government drug insurance or no drug insurance had twice the odds of borrowing to pay out-of-pocket costs for prescriptions (adjusted OR 2.0, 95% CI 1.4–2.9) compared to those with employer-sponsored drug insurance. Other factors associated with increased adjusted odds of borrowing were having 2 or more chronic conditions, having a household income less than $40 000 and spending more money out of pocket on prescription drugs.
Extent of out-of-pocket costs among borrowers
Among those who reported having borrowed money to pay for prescription drugs in the previous year (n = 6798), the largest group (an estimated 247 397 [33.8%], 95% CI 28.1–39.5) borrowed money to pay for comparatively low drug costs, $200 or less. Another estimated 195 859 (26.8%) (95% CI 21.2–32.4) borrowed money to pay out-of-pocket drug costs of $201–$500. Finally, an estimated 143 601 (19.6%) (95% CI 14.7–24.6) and 113 303 (15.5%) (95% CI 11.8–19.2) of those who reported borrowing money did so for out-of-pocket drug costs of $501–$1000 and more than $1000, respectively.
Interpretation
The out-of-pocket costs associated with having to pay for prescription drugs have important implications for Canadians. We found that 2.5% of Canadians, or an estimated 731 000 people, borrowed money to pay for the out-of-pocket costs of their prescription drugs in the previous year. This represents another form of compensatory behaviour to deal with drug charges on top of the substantial rates of cost-related nonadherence to prescription medications and tradeoffs with other expenditures patients in Canada report.11 Furthermore, we found that borrowing was most common among groups who also display these other compensatory behaviours, including younger people, those with lower self-reported health, those with government or no drug insurance, and those with lower household income.11 Notably, we found that borrowing to pay for prescription drugs occurred at all levels of out-of-pocket costs, and over 60% of borrowing reported by our respondents in the previous year occurred for out-of-pocket costs of $500 or less.
Our findings are consistent with other research on cost-related nonadherence and associated compensatory behaviours. 6,9,11 Studies from the US have shown cost-related nonadherence to be associated with food insecurity and cutting back on necessities.10,16 This is sobering in light of the fact that most drugs for which patients need to make these tradeoffs are relatively inexpensive.10,11 Our findings suggest that the consequences of high drug costs are more extensive than just reducing adherence to medicines; there are likely impacts on other aspects of patients’ quality of life.
There is an opportunity for health care providers to intervene and initiate conversations with patients to help support those who are at greatest risk of cost-related nonadherence and the attendant compensatory behaviours. Prior research suggests that patients who have high levels of trust in their health care providers tend to be at lower risk for cost-related nonadherence.30 Although many physicians feel they are not well positioned to manage conversations about medication costs, many health care providers also believe it is their responsibility to engage patients in these conversations.31,32 Prescribers can improve medication affordability for patients by staying up to date on drug costs, prescribing the most cost-effective alternative, frequently reviewing medication regimens for opportunities to deprescribe and prescribing generic drugs.31,33
Multiple policy interventions can be used to address the negative impacts on patients who are struggling to pay for prescription drugs.34 For example, some provinces have recently implemented policy changes to help address cost-related nonadherence: Ontario implemented a pharmacare program for youth under the age of 25 who lack private prescription drug insurance, and British Columbia has reduced or eliminated public drug insurance deductibles for families earning less than $45 000 per year.36,36 Future studies should investigate the impact of such changes on both cost-related nonadherence to prescription medications and borrowing behaviours.
Limitations
Like other survey-based study designs, our results are based on patient self-report, which is potentially susceptible to recall bias and social desirability bias. If recall bias had an effect on our findings, this likely would have resulted in conservative estimates.37 We were unable to ask more specific questions about the kind of borrowing patients engaged in (e.g., from family and friends, a financial institution, or a payday loan or cash advance lender, against a mortgage or home, on a credit card) and were unable to assess how respondents interpreted the question or to gauge the amount of money that might have been borrowed.
Conclusion
Our findings show that many Canadians have borrowed money to pay for out-of-pocket prescription drug costs. Borrowing occurred for relatively inexpensive drugs as well as more costly ones and disproportionately affected vulnerable groups such as those who have low income, those with poor self-reported health status and those who lack prescription drug insurance. In Canada, some provinces are already taking steps to implement policy changes to help these more vulnerable groups address cost-related nonadherence to prescription medications and other associated behaviours. Future research should investigate the impacts of such changes that could increase equity in access to prescription drugs.
Footnotes
Competing interests: Michael Law has consulted for Health Canada and has acted as an expert witness for the Attorney General of Canada. Anne Holbrook is a therapeutics expert and drug policy consultant for the Ontario public drug programs, the Canadian Agency for Drugs and Technologies in Health and Hamilton hospitals. No other competing interests were declared.
This article has been peer reviewed.
Contributors: Ashra Kolhatkar drafted the manuscript, and Lucy Cheng, Steven Morgan, Laurie Goldsmith, Irfan Dhalla, Anne Holbrook and Michael Law critically revised the manuscript for important intellectual content. All of the authors contributed substantially to the study conception and design, and analysis and interpretation of the data, approved the final version to be published and agreed to be accountable for all aspects of the work.
Funding: This analysis was funded by Operating Grant MOP-126020 (principal investigator: Michael Law) and Foundation Scheme Grant FDN-148412 (principal investigator: Michael Law) from the Canadian Institutes of Health Research. Michael Law received salary support through a Canada Research Chair in Access to Medicines and a Michael Smith Foundation for Health Research Scholar Award.
Disclaimer: Any opinions or conclusions expressed in this publication do not necessarily represent the opinions or conclusions of Health Quality Ontario. No endorsement is intended or should be inferred. This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this article are those of the authors and are independent from the funding sources. No endorsement by ICES or the MOHLTC is intended or should be inferred.
Supplemental information: For reviewer comments and the original submission of this manuscript, please see www.cmajopen.ca/content/6/4/E544/suppl/DC1.
References
- Copyright 2018, Joule Inc. or its licensors