Abstract
Purpose
Appropriate use of oral chemotherapy is a challenge for patients and clinicians. The purpose of this study was to analyze cancer patients’ use of oral chemotherapies and identify opportunities to improve adherence.
Methods
We developed a 30-question survey to address frequency and reasons for reducing/skipping doses; sources of information for oral chemotherapy use; perceived importance of food–drug effects; and ease of understanding labeling directions.
Results
Ninety-three patients taking oral chemotherapies with chronic myeloid leukemia, renal cell carcinoma, breast cancer, and colorectal cancer completed the survey. This was a well-educated population with 69% (n = 62) having completed some college; 51% (n = 47) female and 59% (n = 54) older than 50 years of age. Thirty percent of patients reported forgetting to take their oral chemotherapy at least “sometimes”. Younger patients (<50 vs. ≥50, p = 0.002), shorter treatment duration (<6 vs. ≥6 months p = 0.03), or with chronic myeloid leukemia (vs. other diagnoses, p = 0.015) forget to take their oral chemotherapy at higher rates. Twenty-three percent (n = 21) indicated they intentionally skipped their oral chemotherapies and 38% (n = 8) of those did not inform their physicians. Forty-one percent (n = 28) taking drugs with significant food–drug effects did not think about their last meal before taking their oral chemotherapy and 80% (n = 55) did not understand the potential interactions. Additionally, 39% (n = 36/92) never looked at labeling and 15% (n = 14/91) had difficulty understanding label directions.
Conclusion
There are three main barriers associated with appropriate use of oral chemotherapies: misunderstanding about the timing of drug with food; stopping drug without informing physicians; and difficulty understanding labeling directions. A multipronged approach is needed to optimize communication of directions for optimal oral chemotherapy use.
Background
In 2003, only 5% of chemotherapeutic agents were available for oral administration; by 2010, 20–25% of chemotherapy agents in development are oral. Oral antineoplastics (OA) offer convenience to patients, thereby improving quality of life and fostering a sense of independence by increasing patient-involvement in their own care. 1 Despite the benefits, non-adherence with oral chemotherapies (OCs) remains a major challenge to achieve optimal care. Adherence is defined as the extent to which a patient follows directions that a health care practitioner prescribed. 2 While OCs have shown comparable efficacy in terms of response rates and durability, non-adherence has been shown to undermine these outcomes. McCowan et al. reported an association with poorer survival in the 19% of breast cancer patients who were <80% adherent to tamoxifen. 3 Marin et al. showed that 26% of chronic myeloid leukemia patients who were adherent ≤90% of the time to imatinib were less likely to achieve a major or complete molecular response than those who were adherent >90% of the time. 4
The extent of adherence to proper administration of OCs with regard to food is largely unknown. Based on various pharmacokinetic parameters, the FDA’s package labeling carries specific instructions on administration of OCs with regard to food. Additionally, the clinical trials with these agents took into account the effect of food on OCs. Non-adherence to the FDA’s recommendations can lead to increased drug concentration and subsequent toxicity (e.g. if nilotinib is taken with food), gastrointestinal irritation (e.g. if imatinib is taken without food), or decreased absorption and compromised efficacy (e.g. if sorafenib is taken with food). Although there is clear possible clinical sequelae of taking OCs without regard to food, controversy exists over the best strategy to prevent or even utilize drug–food interactions. 5
Although the impact of adherence on disease outcomes is widely accepted, the factors contributing to adherence are multifaceted. Potential barriers to patient adherence include: (a) social and economic; (b) health-care related (inadequate training, ineffective communication); (c) disease-related (severity of symptoms, progression of disease); (d) therapy-related (complex regimens, inconsistent prescription labeling, drug–drug and drug–food interactions, or adverse effects), and (e) patient-related (impaired cognitive ability, forgetfulness, poor health literacy).2,6–8
This study administered a survey to assess several of these factors across a wide array of cancer types. There is a paucity of data regarding patient behavior in this setting and, to our knowledge, this is one of the first studies to report on patient behavior in the context of food–drug interactions (FDIs) and OC. Three issues were evaluated: patients’ description of timing and scheduling of OCs, information sources that patients rely on for proper drug administration, and the barriers patients identify that contribute to non-adherence.
Methods
Population
This was a prospective study of our patient population, which consisted of adult patients (≥18 years old) diagnosed with breast cancer, renal cell carcinoma (RCC), chronic myeloid leukemia (CML), and colorectal cancer (CRC), who were currently receiving OC. From April 2011 to February 2012, patients were identified by electronic searches of the pharmacy information system (PharmNet), and electronic medical records (WebCis), and by contact with the primary oncologists and advanced practice providers (APPs) at the relevant clinics at University of North Carolina Medical Center. Patients who were on hospice, cognitively impaired and requiring assistance with medication administration, or non-English speaking were excluded. Eligible participants were recruited during clinic visits and those interested in participating in the study were given a paper survey to fill out and return before leaving the clinic.
Survey instrument
A 30-question paper survey, adapted from Chew et al.’s health literacy questionnaire, 9 was used (Appendix 1).The survey consisted of four domains: (a) current practices related to patient medication use; (b) issues related to reading and understanding the directions on the medication labels; (c) behavior surrounding OC administration with regard to food; and (d) demographics. Questions were closed-ended, either Likert-type scale or “check all that apply,” with options to write in comments when “other” was the selected response.
Analysis and statistical methods
Descriptive statistics are provided for all survey responses, and comparisons between groups were made using Fisher’s exact test. A multivariable logistic regression modeling were used for the outcome of forgetfulness, and odds ratios (OR) are reported. All analyses were conducted using SAS statistical software version 9.2 (Cary, NC). FDA package insert instructions were the reference for determining “appropriate” timing of OC with regard to food. This study was reviewed and approved by the Institutional Review Board (IRB) at the University of North Carolina Medical Center. Patients were notified that by completing the survey they were giving their consent to participate in the study.
Results
Patient sample
Demographic and clinical characteristics.
Sources of information for proper administration of OA
Sources of information about OC administration.
Reading and understanding the label
Thirty-nine percent (n = 36/92) of patients reported they never look at the labeling information. Although 98% (n = 90/92) of patients reported they have little or no difficulty reading the label, 15% (n = 14/91) of patients reported the label is difficult to understand at least sometimes. Here, reading referred to being able to see the words on the label (font size, boldness, format, etc.) and understanding referred to being able to interpret and follow the directions. Patients who “always/frequently” looked at their labels were less likely to forget to take their OC compared with patients who “sometimes/never” looked at the label (20% vs. 35%, respectively, p = 0.15).
When asked about improvements that could be made to the medication label for easier readability, patients recommended the following: larger font size on the label (46%, n = 29/63), better color coding (16%, n = 10/63), and easier symbols (11%, n = 7/63). Even patients who had no difficulty reading the label preferred larger font size (40%, n = 17/43), better color coding (9%, n = 4/43), and easier symbols (7%, n = 3/43). Patients also had recommendations to help them understand various aspects of the medication label. These included avoiding abbreviations (23%, n = 14/62), easier directions (21%, n = 13/62), and not using stickers (11%, n = 7/62).
Self-reported adherence to proper administration of OA
Forgetfulness
Selected survey responses.
Demographic factors that were associated with increased forgetfulness included being on OCs for greater than 6 months compared with those taking OCs for less than 6 months (p = 0.03); age less than 50 years (p = 0.002); and diagnosis of a malignancy greater than 12 months ago (p = 0.015). When looking at specific cancer types, patients with CML reported higher rates of OC forgetfulness (48%) compared with patients with breast cancer (23%), RCC (18%), and those with GI malignancies (13%) (p = 0.05).
Given that both time on OCs and time from diagnosis were both associated with forgetfulness, a multivariable model was fit to evaluate the association of time from diagnosis, age, and type of cancer with forgetfulness. This exploratory analysis showed that being under 50 and diagnosed over a year ago both remained significantly associated with forgetfulness (OR = 6.4 (95% CI: (2.00, 20.00)) and 12.9 (95% CI: (1.48, 112.15)).
FDIs
Seventy-four percent of patients in our study were taking an OC with significant FDIs. Of these patients, 91% reported they are “very comfortable” taking their OC with regard to food; however, 41% reported “they don’t always think about the last time they ate” in relationship to medication administration. When asked whether or not the OC they were taking had an FDI, 27% gave the wrong answer for the direction of the interaction, 54% did not know and only 20% gave the correct answer. Comparing their responses with the actual timing and scheduling directions indicated in the FDA package insert revealed that 15% of patients were taking their OCs incorrectly.
Barriers to adherence with OA
A variety of factors were reported as reasons why patients were inconsistent with OC administration. Of the patients who reported cutting back on their OC, 45% (9/20) reported that they were instructed to do so by their provider and 35% (7/20) indicated that a side effect of the drug was the primary reason. Of note, 38% (8/21) of patients discontinued therapy due to adverse effects without their doctor’s instructions. The most common types of side-effects experienced by patients who cut back on their OCs included neuropathy (n = 10), diarrhea (n = 5), fatigue (n = 9), nausea (n = 7), and rash (n = 3). Other reasons for inconsistent OC administration included a delay in refilling their OC from their pharmacy (17%, 14/83). The majority of these delays were 1–3 days. Less than 1% of patients reported not taking their OC due to forgetfulness related to travel or emotional distress. Cost-barriers were not a major determinant of patient-reported non-adherence to therapy (n = 3).
Discussion
Adherence to OC continues to be an important area of research. Several authors have reported on adherence rates as it relates to OCs.3,4,8,10,11 Although these trials were able to characterize the rate of non-adherence for individual diseases, there still continues to be a lack of understanding on the patient-reported barriers of adherence in cancer patients. The aim of our study was to characterize these barriers associated with appropriate OC use across a wide spectrum of diseases and demographic backgrounds. This type of study is important to address the need for interventions at a patient, clinician and/or health system level, such as improved patient education and communication by members of the health-care team. Surprisingly, cost was not a major barrier identified in our study likely because most patients had overcome this barrier and were already on therapy. However, cost still remains a major barrier to care and should not be neglected in future considerations of adherence.
Self-reported non-adherence in cancer patients confirmed in a diverse population
Similar to what has been reported in the HIV/AIDS setting, our findings confirm that despite the life-threatening nature of their malignancies, patients are not always adherent with their self-administered OA. 2 We were able to show this in patients across a wide array of malignancies, age groups, and durations of therapy. Two factors that contributed significantly to self-reported non-adherence were younger age (<50 years) and length of diagnosis (>12 months). Younger patients and those who have had their diagnosis ≥12 months may underappreciate the importance of strictly adhering to their OC.
Toxicity is a major barrier to adherence
Timely management of adverse effects is an essential way to prevent non-adherence. It is not known how missed doses due to adverse drug effects impact outcomes for many OCs. A notable finding by Marin et al., who assessed adherence using a medication event monitoring system (MEMS), was that adverse effects to imatinib were predictive of lower rates of adherence and subsequently lower rates of molecular and cytogenetic response. 4 Because patients are taking OCs continuously (sometimes for several years as in CML or breast cancer therapies), patients who experience grade 1 or 2 adverse effects might also be non-adherent. Early management of toxicities is vital to get the desired therapeutic benefit. These interventions and their effect on adherence should be further explored.
Adhering to food–drug instructions is critical to efficacy and/or safety
A novel area of evaluation in our research was patients’ understanding of and adherence to food–drug considerations. Several OCs have important pharmacokinetic changes if the food–drug recommendations by the FDA are not followed. Surprisingly, nearly 40% of patients taking OCs with significant FDIs did not think about the last time they ate. The potential implications of this finding are twofold: patients may not be getting the expected therapeutic outcome and/or they may be experiencing significant toxicity leading to decreased adherence to therapy. In the case of imatinib, not taking it with food could contribute to significant gastrointestinal toxicity, leading to non-adherence, which has been documented in the literature by Marin et al. 4 For other OCs such as lapatinib, nilotinib, pazopanib, and abiraterone, administering with food could lead to significant increases in serum concentrations, leading to toxicity. Because the initial clinical trials were conducted based on the FDA recommended labeling for administration, it is essential that clinicians emphasize this point when counseling their patients about appropriate drug use.
Physicians and drug labels are primary sources of information on proper OC use
Our findings also suggest that improvements to the format and size of instructions on the medication labels will benefit patients in better understanding the instructions. One third of patients reported the medication label was their primary source of information for OC administration with regard to food. In our study, patients’ ability to simply read the labeling information wasn’t always predictive of their ability to understand it. The issue of poor health literacy and its effect on adherence is well known, yet it still prevails as an important issue. 12 Readability of prescription warning labels has also been evaluated with a focus on identifying the required grade-level for most patients’ understanding. Ngoh et al. reported that sixth- or seventh-grade-level and eighth-grade-level readability of labels have a 4.3 times and 12.9 times likelihood of getting misunderstood compared with labels written at a third-grade level, respectively. 13 In addition to the label, appropriate verbal communication by the clinician, additional counseling at the point of dispensation, and ideally, follow-up phone calls should be instituted. 13 Wolf et al. found that although patients could read the label instructions (“take two pills twice daily”), a third of them could not accurately demonstrate understanding (when asked how many pills they would take in one day, they incorrectly answered “two pills” rather than “four pills”). Misunderstandings of labeling instructions can be avoided by using more explicit and simple language to state directions (two tablets in the morning, two tablet in the evening) and organizing the label to reduce distracting elements. 12
Only about one-third of patients mentioned getting their education about OC from a pharmacist. This primarily consisted of a phone conversation from a pharmacist in a mail-order/specialty pharmacy where most of these OC are dispensed. During the study period, clinical pharmacists played very little role in the ambulatory care setting in terms of direct patient education for the OC agents and management of toxicities in our institution. The findings of this study and creation of an institution-based specialty pharmacy have helped justify full time pharmacists in the oncology clinics to ensure the needs of our patients are met with regards to OC agents by providing education, overcoming cost-barriers, managing toxicities, and providing adherence tools.
Next steps: Comprehensive OC program
A multipronged and integrated approach is needed to address the identified barriers to adherence of OCs. A comprehensive OC delivery program needs to be explored to optimize communication of directions for optimal OC use. At the initiation of therapy, patients should be asked to “teach-back” on their OC administration. Because of the continued shortage and time-constraints of oncologists,14,15 clinical pharmacists or other APPs can take the responsibility of patient education on proper OC use. They should emphasize key counseling points with emphasis on OC administration with regard to food. In addition, all toxicities should be addressed since these are common reasons for non-adherence.
Limitations
Limitations of this study include using anonymous surveys, which prevented us from rechecking the reliability of patient responses with their medical records and not directly measuring adherence (MEMs caps, pill counting, etc.). We also do not know what information was communicated to the patient from the health-care team, physician, nurse, or pharmacist. Also, as stated above, there were five versions of the survey with minor changes that were distributed based on feedback from our research team and survey participants. Inclusion of clinical trial patients could also have been a confounder although they only accounted for a small percentage of patients.
Conclusion
We found that there is a significant gap between how we are instructing patients to take their OCs compared with what they reported back. The three main findings were: (a) patients, especially younger patients, are more likely to forget or intentionally cut back their OC without informing their physicians; (b) many patients do not understand the timing of their medications with regard to food; (c) although nearly all patients may look at their medication labels, they are not always able to follow the directions. There needs to be a concerted effort between clinicians, health-systems, and policy makers to ensure optimal management of OC use through the development of structured patient education strategies, validated adherence monitoring tools, and timely management of toxicities.
Footnotes
Acknowledgements
Christine Walko, PharmD, BCOP and Dan Crona, PharmD, PhD.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.