How costly is non-adherence to infliximab in patients with Crohn’s disease?

Non-adherence is the deviation from agreed treatment plans and is common among all fields of medicine1. In chronic diseases such as Inflammatory Bowel Disease, non-adherence leads to adverse patient outcomes including increased flare rates and possibly colorectal cancer1. Current evidence suggests that non-adherence is often multi-factorial, but patients’ beliefs about the necessity for treatment and the fear of possible side-effects have been consistently shown to drive non-adherence2,3. When aiming to assess the true financial burden of non-adherence, the direct costs of additional healthcare usage amongst non-adherent patients must be weighed up against the extra medication costs in adherent patients. Studies addressing this issue need to capture full healthcare usage and costs while also reporting non-adherence accurately with precise medication costs. In patients with mild-to-moderate ulcerative colitis it has been demonstrated that non-adherence to mesalazine leads to increased overall costs4. The small increase in medication cost for adherent patients was far outweighed by vastly increased healthcare usage by non-adherent patients.

In contrast to mesalazine, Infliximab, which is used for moderate-to-severe Crohn’s disease and for the treatment of moderate-to-severe ulcerative colitis, is far more expensive. Findings from the Kane and Shaya4 study on ulcerative colitis can, therefore, not be extrapolated to Crohn’s disease, especially as the study selected patients with mild-to-moderate disease only. When Kane et al.5 examined the costs for Crohn’s disease patients in 2009 they found higher costs for non-adherent patients, but critically the costs of infliximab and its administration were excluded from the analysis. Feagan et al.6 examine in this issue of the Journal of Medical Economics whether non-adherence to Infliximab influences the overall costs, including drug costs in patients with Crohn’s disease. The study demonstrates that intermittently adherent patients had higher direct healthcare costs compared to adherent patients. The absolute difference in overall costs was small, likely secondary in part to the high cost of infliximab. Their efforts to provide evidence in this field of research are laudable, especially as they faced several methodological challenges.

Whether a patient adheres to Infliximab is an important clinical consideration in terms of judging treatment efficacy, the need for escalation of therapy, or even the need for surgery. Given the high cost of infliximab, it should be used continuously as this leads to better clinical results. The efficacy of infliximab can also be influenced by adherence, as treatment gaps lead to sensitization with antibody formation. Traditionally patients are classed as either adherent or non-adherent. In studies examining clinical records and clinical databases, adherence status can be easily determined. Yet, in large insurance claims databases, which are best placed to assess treatment costs, these clinical data are frequently not captured7. Cessation of treatment may reflect a number of events. Firstly the medication may have been stopped by the treating doctor due to side-effects or lack of efficacy. Or, secondly, the patient may have stopped the medication against the doctor’s advice. Only the latter represents non-adherence.

To overcome this issue, Feagan et al.6 have concentrated on patients who were either adherent or intermittently adherent. They define intermittent adherence as patients who continued to receive infliximab, but at longer than the recommended intervals. The concept of differentiating intermittent adherence from full non-adherence for cost analysis purposes is new and the approach is neither validated nor free of faults. Comparing the adherent group to intermittently adherent patients allowed the exclusion of all patients who ceased treatment for clinical reasons. Yet, critically, all patients who ceased treatment because of non-adherence will have been lost to the analysis. Even more importantly, Feagan et al.’s6 definition of intermittent adherence also captures patients who may have not received Infliximab at the recommended intervals due to situations beyond their control. It is standard practice to pause biological agents during infections for examples. This will especially apply to patients with penetrating Crohn’s disease, who may have their infliximab paused for antibiotics, radiological intervention, or surgery. Some patients may have had infusions delayed because of organizational issues within the delivering doctor’s office or hospital. Yet others may have delayed infusions because of holiday or business trips. However, all those will be classified as intermittently adherent. The intermittent adherence definition applied here reflects more whether real world drug delivery in accordance with guidelines has occurred rather than only capturing strictly patient-driven non-adherence.

It remains, therefore, difficult to judge whether patient-driven non-adherence to infliximab has an influence on overall costs. I am somewhat sceptical that Feagan et al.’s6 definition of intermittent adherence will become an agreed way of reporting adherence in future studies given the above limitations. Researchers wanting to use such a measure in future should first aim to validate it. There is no gold standard of measuring adherence1, and adherence to 8-weekly infliximab is difficult to measure. Drug levels are too unreliable, as concentrations can vary significantly between fully adherent patients. Reporting ‘no show’ for appointments by examining clinical records is a sensible way, but is very labour-intensive and often limits the study size that can be practically achieved8. Furthermore, the number of infusions per year can be used for claims databases, but this also does not account for doctor-initiated treatment pauses9. Medicine possession ratios derived from claims databases apply a similar methodology, but are probably less sensitive because of the low frequency of infusions10. Self-report tools of adherence such as MARS are validated for Inflammatory Bowel Disease medication in general, but have not been formally validated for injectable or infusible treatments3.

Despite all the limitations in the current study by Feagan et al.6, there are a number of important messages. Firstly direct costs were considerably lower in the adherent group, a finding also reported by Kane et al.5 in 2009. These cost savings are directly linked to less flares and hospitalizations. We can, therefore, conclude that continuous Infliximab administration at the recommended infusion intervals is associated with better clinical outcomes and reduced direct healthcare costs. Furthermore, the reduced healthcare costs are not outweighed by higher drug-related costs. Whether these effects are truly related to patient-driven non-adherence or to other factors is currently not sufficiently established.

Transparency

Declaration of funding

No funding was received for this manuscript.

Declaration of financial/other relationships

CS has no conflicts of interest of financial disclosures relevant to this manuscript. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.