Sinnott S, Buckley C, O’Riordan D, Bradley C, Whelton H. The effect of co-payments for prescriptions on adherence to prescription medicines in publicly insured populations; a systematic review and meta-analysis. PLoS One. 28 May 2013;8 (5):e64914.

Lesen E, Sundell KA, Carlsten A, Mardby A, Jonsson AK. Is the level of patient co-payment for medicines associated with refill adherence in Sweden? Eur J Public Health. Published online 23 Jun 2013.

Background

According to the Canadian Institute for Health Information, spending on prescription drugs in Canada grew from $8.0 billion to $19.0 billion between 1998 and 2007, more than doubling in under a decade, with an average annual growth rate of 10%[1]. Other developed countries have experienced similar spending growth.

To contain expenditures, public and private insurers often resort to consumer cost-sharing, frequently in the form the form of co-payments.

Some patients are likely to find co-payments onerous, such as those on public assistance, pensions, or other fixed incomes, threatening their ability to pay for prescribed medications. Non-adherence to medication for chronic diseases, such as hypertension, diabetes, and mental disorders, results in morbidity, mortality, and additional health-care costs.

In this e-round, we review two studies on co-payments, one in the US, and another in Sweden.

Sinnot et al undertook a systematic review and meta-analysis to determine how medication adherence is affected by a new or increased co-payment among patients covered by public insurance. The studies generally used one of four methods to measure adherence (i.e. proportion of days covered; medication possession ratio; daily defined dose; ReComp algorithm).

Lesen et al examined the effect of co-payments on adherence to medications for 2210 Swedish patients with epilepsy. In the jurisdiction studied, patients had a tiered co-payment based on the price of medication, but with an annual limit; once the maximum accumulated co-payment was reached, the co-payment was eliminated. The researchers used a continuous measure of medication acquisition (CMA) based on the supply of medication units obtained per unit of time relative to the prescribed amount. Non-adherence was defined as a CMA less than 80%. Covariates included demographics, family income, and other personal health-care expenditures.

Findings

Of 22 included studies in the Sinnot et al study meta-analysis, 7 studies, representing 199,996 Veterans Administration or Medicare patients, were eligible for the meta-analysis. They typically examined adherence to treatments for chronic diseases such as hypertension, diabetes, depression and hyperlipidemia. The odds ratio for non-adherence was 1.11 (95% CI, 1.09-1.14) indicating an 11% increased risk of non-adherence when co-payments ranged from $5 to $70. All studies in Sinnot et al were of weak to moderate quality, and typically suffered from significant differences between the study and comparison groups, and limited follow-up data. Some studies failed to control for possible confounders. The funnel plot for included studies was asymmetrical suggesting publication bias (i.e. the possibility that studies showing no effect of co-payments remain unpublished, thus overestimating the true effect).

In the Lesen et al study, associations between patients’ co-payment for all medicines and refill adherence were assessed with multilevel mixed-effects linear regression, accounting for clustering within patients. Exemption from co-payment, compared to co-payments of 25% or more, was associated with greater adherence. In a mixed-model analysis of covariance (ANCOVA), a co-payment of 50% or less was associated with 2-4% greater adherence than a 100% co-payment. Non-adherence was also associated with younger age, and being on social or unemployment benefits, or in the lowest income quintile. Non-adherence may have been underestimated because data on patients who never obtained the first prescription were not captured.

Comment

Qualitative reviews of the evidence on co-payments have previously shown that they decrease access to care, and adherence to interventions. The Sinnot et al systematic review and meta-analysis is the first quantitative analysis providing a summary odds ratio. The Lesen et al study reports on a Swedish population with generous health care coverage and an overall high level of adherence, but still shows that even in this context, adherence to medications is sensitive to co-payment.

As Canada wrestles with how best to finance health care, these two studies provide evidence that individual co-payments do not provide a health risk-free or equitable solution. These studies also provide evidence that as we explore financing options for a universal Pharmacare program, we should consider the health consequences of drug co-payments, especially for the increasing number of Canadians reliant on pharmaceutical interventions.