Introduction

Cancer chemotherapy is associated with a predictable spectrum of dose-related toxic effects including nausea and vomiting, which may be present for up to 5–7 days after dosing. Symptoms that occur within 24 h after the initiation of chemotherapy are traditionally defined as acute; those that occur after 24 h are defined as delayed [10, 11].

Cisplatin is one of the most emetogenic chemotherapeutic agents. High-dose cisplatin (>50 mg/m2) predictably evokes acute vomiting in approximately 100% of patients and delayed vomiting in approximately 70%–90% of patients in the absence of preventive therapy. Cisplatin has been used as the benchmark chemotherapy in the evaluation of the efficacy of antiemetic therapies.

While current antiemetic treatments have resulted in much improved control of chemotherapy-induced nausea and vomiting (CINV) especially during the acute phase, many cancer patients still experience these complications [14], either because of failure of antiemetic treatments or because of failure to prescribe appropriate antiemetic therapy. CINV, especially during the delayed phase, can lead to significant problems for patients such as an inability to perform daily activities and difficulties with eating and drinking. In some cases, patients experiencing severe CINV are unable to eat during the week following treatment, may require intravenous volume replacement therapy or delay future treatment [14]. Patients consistently rank nausea and vomiting as one of the most distressing side effects of cancer chemotherapy [9, 12].

At the time of this study, only one patient-completed instrument, the Functional Living Index—Emesis (FLIE), existed to directly assess the impact of CINV on patients' daily lives. Other methods and instruments have been used to assess the overall impact of chemotherapy on patients' well-being or to assess treatment burden, but these methods are not specific to nausea and vomiting [2, 3]. As originally developed by Lindley et al., the FLIE was used with a 3-day recall period [13]. The purpose of this analysis was to assess the measurement characteristics of the FLIE instrument using a longer recall period of 5 days in order to quantify the impact of CINV throughout the acute and delayed phases following chemotherapy initiation. The measurement characteristics of the 5-day recall version were also compared to those for the original 3-day version.

Patients and methods

The development of the FLIE has been described previously by Lindley et al. [13]. The FLIE instrument was modeled after the Functional Living Index—Cancer, a patient-completed multidimensional quality-of-life evaluative instrument [15]. The FLIE is a validated nausea and vomiting-specific patient-reported outcome instrument comprising two domains (Vomiting and Nausea) with nine identical items in each domain. The first item in each domain asks the patient to rate how much nausea (vomiting) he/she has experienced over the past 5 days. The remaining eight items assess the impact of nausea (vomiting) on the following aspects of a patient's daily life: ability to enjoy meals/liquids, ability to prepare meals/do household tasks, ability to perform daily functions, ability to perform usual recreation/leisure activities, willingness to spend time with family and friends, extent to which the side effect has caused personal hardship and hardship on others.

Each item is answered using a 100-mm (1 to 7 points) visual analog scale (VAS) with anchors corresponding to "none"/"not at all" and "a great deal" and tick-marks dividing the scale into six equal categories (Fig. 1). Items within the domain are weighted equally and summed to create the domain score. The two domain scores are then summed to create a total score. Higher scores are more favorable and indicate greater ability to maintain daily life. The endpoint, "no impact on daily life", is operationally defined as an average item score of >6 on the seven-point scale. This cut-off was chosen based on face validity. A patient whose average item response is >6 has explicitly chosen the highest category anchored by "none" or "not at all" to describe the level of impact on daily life.

Fig. 1.
figure 1

Example of a FLIE item with definition of "No Impact on Daily Life

Full size image

The modified version with 5-day recall was used to assess the overall impact of CINV on patients' daily lives following their first round of highly emetogenic chemotherapy. Data from a large, multinational, double-blind, randomized, parallel-group, controlled, phase IIb clinical trial of a novel oral NK-1 receptor antagonist (MK-0869, aprepitant; 3-[[(2R,3S)-3-(p-fluorophenyl)-2-[[(R)-methyl-3,5-bis(trifluoromethyl)benzyl]oxy]morpholino]methyl]-3-1,2,4-triazolin-5-one tachykinin receptor antagonist) were used for these analyses. Patients of at least 18 years of age, scheduled to receive cisplatin ≥70 mg/m2 for the first time, were enrolled in the study. Patients were excluded if they were mentally incapacitated, used any illicit drugs including marijuana or excessive alcohol, were to receive stem cell rescue therapy, had received an investigational drug within the last 4 weeks, were to receive multiple-day chemotherapy with cisplatin, had vomited within the 24 hours prior to the start of the study, were to receive therapy to the abdomen or pelvis, had a symptomatic primary or metastatic CNS malignancy, or the study medication was contraindicated due to the patient's current medical status or concomitant medications. Table 1 outlines the treatment regimens for patients included in this analysis.

Table 1. Treatment regimens
Full size table

Data were pooled across treatment groups for all analyses. The number of emetic episodes, nausea ratings on a 100-mm VAS, and use of rescue medications were recorded by the patients in a 5-day daily diary. The patients also completed the 5-day recall version of the FLIE questionnaire on day 1, for training purposes only, and again on day 6 during their first cycle of chemotherapy.

Measurement characteristics

Reliability

Two aspects of instrument reliability are routinely assessed as part of questionnaire development. The first, also termed reproducibility or test-retest reliability, is assessed through repeated administrations of the instrument to patients with stable disease. The second, internal consistency or scale reliability, refers to the degree to which questions within a domain measure the same concept. Due to the fluctuation in frequency and severity of CINV during the first cycle of chemotherapy, it was not possible to assess reproducibility. Internal consistency of the FLIE was assessed by Cronbach's alpha [7]. Alpha values above 0.75 indicate excellent internal consistency, although values above 0.95 may imply redundancy [17].

Validity

Two types of instrument validity, construct and convergent, were assessed. Correlations between items and domain scores using Spearman's correlation were used to assess construct validity, the relative appropriateness of item groupings within each domain. Acceptable construct validity is demonstrated by item-domain correlations stronger within domains than across domains. The degree to which a measure reflects the disease severity (convergent validity), was explored by assessing the correlation between FLIE total and domain scores and independent clinical endpoints such as number of emetic episodes, nausea ratings, and use of rescue medications. FLIE total and domain scores were expected to be moderately correlated to the severity of nausea, frequency of vomiting episodes and use of rescue medications. Additionally, the proportion of patients reporting "no impact on daily life" (average domain item score of >6 on the seven-point scale) on the Total score was compared (Pearson chi-square) among patients with and without Total Control. Total control was defined as no vomiting, no use of rescue medication and no nausea (<5 mm on the VAS) during the 5 days after chemotherapy.

Missing data

Finally, as part of the assessment of this new version, the frequency of missing data for individual questions or sections of the questionnaire was evaluated.

Results

The patient demographics and clinical characteristics are presented in Table 2. Patients included in this analysis were primarily Caucasian (73.8%) and male (62.8%) with a mean age of 57 years. Of the full clinical trial sample completing cycle 1, 93% of the patients (187 of 202) completed the questionnaire. Of those completed, 4 were excluded from the analysis because they were completed outside the allowable time-frame (i.e., completed after day 8), leaving 183 patients for this analysis.

Table 2. Baseline patient demographics and clinical characteristics
Full size table

Reliability

Internal consistency assessed by Cronbach's alpha was 0.79 for both the Nausea and Vomiting domains. According to the guidelines established by Streiner [17], the internal consistency of the original 3-day version (alpha values above 0.9) [13] would appear to indicate some level of redundancy; however, both domains in this modified version were found to have excellent internal consistency.

Validity

Table 3 presents the item-domain correlation coefficients for each item in each domain. Acceptable construct validity was observed with item-domain correlations stronger within domains (r=0.74 to 0.97) than across domains (r=0.52 to 0.76).

Table 3. Item-domain correlations for construct validity
Full size table

The correlations presented in Table 4 support the convergent validity of each domain. The correlations were moderately strong and negative indicating that patients experiencing less nausea and vomiting were more likely to report favorable outcomes as assessed by the FLIE questionnaire. These results are consistent with those of the original version with 3-day recall wherein correlations of −0.65 and −0.68 were observed between FLIE scores and patient reports of nausea and vomiting, respectively. Likewise, the percentage of patients reporting no impact on daily life as assessed by the FLIE 5 days after chemotherapy was significantly smaller among those experiencing either nausea or vomiting or both compared to those with no nausea or vomiting (Table 5).

Table 4. Correlations between domain scores and clinical endpoints for concurrent validity (n=183)
Full size table
Table 5. Proportion of patients reporting "no impact on daily life" among patients with and without CINV (n=182)
Full size table

Figures 2 and 3 present scatter plots of the FLIE Nausea and Vomiting domain scores by average nausea VAS score, and average number of emetic episodes, respectively, during the 5 days after chemotherapy. Patients reporting more severe and frequent nausea and vomiting reported more impact of their CINV on daily life compared to those with less severe symptoms. Finally, Fig. 4 presents the cumulative distribution of FLIE Total scores by level of CINV control. The proportion of patients with an average item score >6 (i.e., "no impact on daily life") in the Total Control group was significantly greater than among those without Total Control (96.6% vs 52.1%; P<0.01). Fewer than 5% of the patients who reported no nausea, no vomiting and no use of rescue medication had average item FLIE scores ≤6 versus nearly 50% of those who reported some nausea, vomiting or rescue medication use over the 5 days.

Fig. 2.
figure 2

Distribution of FLIE scores by nausea VAS score: less nausea associated with higher average FLIE Nausea Domain item score

Full size image
Fig. 3.
figure 3

Distribution of FLIE scores by vomiting episodes: less vomiting associated with higher average FLIE Vomiting Domain item score

Full size image
Fig. 4.
figure 4

Cumulative distribution of FLIE Total scores by CINV control (n=183)

Full size image

Missing data

Finally, the analysis of missing data showed excellent compliance to the questionnaire instructions. Two of the 183 patients who completed the questionnaire according to the protocol missed two and three questions on the FLIE, respectively. According to the scoring guidelines, a domain score may be computed if at least five items in each domain are completed. Among the study sample, all patients had the required number to calculate both domain scores.

Discussion

Though enumeration of vomiting episodes and ratings of nausea are useful for evaluating the clinical efficacy of antiemetics, these measures do not assess the full impact of CINV on the daily life of patients (e.g., daily functioning, appetite, family life, etc.) and thus are not capable of demonstrating the broader impact of treatment. The value of including a patient-reported outcome that assesses the impact of CINV on patients' everyday lives in antiemetic clinical trials is to supplement the information obtained from clinical endpoints. The factors most important to patients may have direct impact on their future health-care decisions and willingness to continue treatment.

The FLIE questionnaire is one of the only published validated nausea- and vomiting-specific instrument currently available to assess the impact of CINV on a patient's ability to maintain daily life, and has been used in clinical trials of marketed antiemetics such as ondansetron [5, 6, 8, 16]. Although the Functional Assessment of Cancer Therapy—General (FACT-G) [4] and the European Organization for Research and Treatment Center Quality of Life Questionnaire (EORTC QLQ-C30) [1] include items that assess the amount of vomiting or nausea, they do not assess the impact of CINV on patient functioning.

The FLIE questionnaire, previously validated by Lindley et al., was shown to be a sensitive and acceptable measure for use within 3 days following chemotherapy initiation [13]. While the FLIE was developed to specifically assess the impact of nausea and vomiting on patient daily function, comparisons with other more general measures such as the Functional Living Index—Cancer support the content validity of the FLIE as a patient-reported outcome measure [15].

The internal consistency, construct and convergent validity of the modified version with 5-day recall as reported here are satisfactory for both the domains and the questionnaire as a whole. The results of convergent validity and internal consistency for this 5-day version are consistent with those previously reported for the original 3-day version [13]. Additionally, the 5-day version of the questionnaire was able to discriminate among patients with varying disease severity. By modifying the recall period to 5 days after receiving chemotherapy, we were able to quantify the impact of CINV on patients' daily lives throughout both the acute and delayed phases.

A valid instrument is essential for understanding results and drawing conclusions about test medications within a clinical trial setting. Likewise, an instrument such as the FLIE may be useful in clinical practice to better understand the effect of treatment on patients' everyday lives. The results from this analysis indicated the 5-day recall version of the FLIE was an appropriate instrument for assessing the impact of CINV on patients' daily lives during both the acute and delayed phases following highly emetogenic chemotherapy including cisplatin. Furthermore, the modified instrument was able to quantify the difference in impact of CINV between patients with and without controlled CINV. The FLIE data clearly indicated that uncontrolled CINV among patients receiving highly emetogenic chemotherapy had a significant negative impact on patients' daily lives compared to those whose side effects were controlled.

Due to fluctuations in severity and frequency of CINV, reproducibility was not assessed. An assessment of change in FLIE scores over time was not appropriate since this was a prevention trial among chemotherapy-naive patients during their first cycle of chemotherapy. Comparisons with other more general patient-reported outcomes were not made for two reasons: (1) the modification of the recall period was not expected to affect the content validity of the questionnaire, and (2) the patient burden associated with self-completed questionnaires was too great to justify the inclusion of another questionnaire. Future studies using the 5-day recall version of the FLIE should include assessment of the instrument's responsiveness to changes in symptom severity over multiple chemotherapy cycles. Further study of the questionnaire among patients receiving non-cisplatin-based chemotherapies is also warranted.

The 5-day recall version of the FLIE was shown to be an acceptable and reliable patient-reported outcome instrument suitable for use in cancer clinical trials to assess the impact of CINV on patients' daily lives over the acute and delayed phases. The FLIE may also be useful in everyday clinical practice to better understand patient outcomes and the impact of current antiemetic regimens following highly emetogenic chemotherapy.