Secondary Logo

Psychometric Testing of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire–Chemotherapy-Induced Peripheral Neuropathy 20-Item Scale Using Pooled Chemotherapy-Induced Peripheral Neuropathy Outcome Measures Standardization and Alliance for Clinical Trials in Oncology A151408 Study Data

Smith, Ellen M. Lavoie, PhD, APRN, AOCN, FAAN; Banerjee, Tanima, MS; Yang, James J., PhD; Bridges, Celia M., BA, BSN, RN; Alberti, Paola, MD, PhD(c); Sloan, Jeff A., PhD; Loprinzi, Charles, MD

doi: 10.1097/NCC.0000000000000596
ARTICLES
Free

Background No criterion-standard patient-reported outcome measure of chemotherapy-induced peripheral neuropathy (CIPN) exists.

Objectives The aims of this study were to reevaluate the sensitivity, reliability, and validity of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire–CIPN (QLQ-CIPN20) measure and suggest possible revisions that could strengthen it.

Methods Cross-sectional QLQ-CIPN20 data from 8 European countries (n = 271) were pooled with data from 4 North American multisite CIPN intervention trials (n = 884). The combined sample (N = 1155) included patients with varied cancer diagnoses who had received neurotoxic chemotherapy. Item score ranges, Cronbach's α, and exploratory factor analysis were used to evaluate sensitivity, internal consistency, and structural validity.

Results Individual item mean scores ranged from 1.21 to 2.34 (SD range, 0.55–1.17). All item scores encompassed the entire 1 to 4 range. We recommend that 4 items be removed because of low item-item score correlations (r < 0.30). On the basis of the remaining 16 items, 88% of the variance was explained by 2 factors whose Cronbach's α coefficients were .90 and .85. However, items lacked conceptual alignment with previously published factor structures.

Conclusion Using a large, diverse sample of European and North American participants, the reduced 16-item QLQ-CIPN20 is sensitive and internally consistent. However, factor analysis results revealed an unstable factor structure.

Implications for Practice The use of a reliable, valid, and sensitive criterion-standard QLQ-CIPN20 variant in clinical practice settings could improve function, quality of life, and CIPN symptom control by facilitating patient reporting and thereby clinician awareness of this underrecognized consequence of cancer therapy.

Author Affiliations: University of Michigan School of Nursing (Drs Smith and Yang and Ms Bridges); Institute for Healthcare Policy & Innovation, University of Michigan (Ms Banerjee), Ann Arbor; University of Milano-Bicocca (Dr Alberti), Italy; Alliance Statistics and Data Center (Dr Sloan), and Mayo Clinic (Dr Loprinzi), Rochester, Minnesota.

Research reported in this publication was supported by the National Cancer Institute of the National Institute of Health under award numbers UG1CA189823 (Alliance for Clinical Trials in Oncology NCORP Grant), U10CA180790, and R03 CA186183-02. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors have no conflicts of interest to disclose.

Correspondence: Ellen M. Lavoie Smith, PhD, University of Michigan School of Nursing, 400 N Ingalls, Ann Arbor, MI 648109 (ellenls@med.umich.edu).

Accepted for publication January 25, 2018.

Chemotherapeutic agents are a mainstay of current cancer treatment. Many are neurotoxic,1 causing chemotherapy-induced peripheral neuropathy (CIPN) in almost all patients who receive them.2 Chemotherapy-induced peripheral neuropathy incidence and severity vary depending on comorbidities and the specific neurotoxic chemotherapeutic agent, dosage, and treatment regimen.2–5 Numbness, tingling, and neuropathic pain in the extremities are common CIPN symptoms.2–4,6–8 Less common but nonetheless disturbing symptoms may include muscle weakness, sensory ataxia that is manifested by impaired balance, and autonomic nervous system dysfunction such as constipation, orthostatic hypotension, urinary retention, and erectile dysfunction.2,4,6–9 Any of these symptoms may require chemotherapy dose reductions and/or delays, which might compromise cancer treatment efficacy.4 Symptoms can become chronic, leading to functional disability and diminished quality of life (QOL).10–12

Although several measurement tools for assessing CIPN are available, there is no consensus about which measure to use; furthermore, most have not undergone rigorous psychometric testing via assessment of multiple psychometric properties.13,14 The most commonly used CIPN outcome measures are toxicity grading scales such as the National Cancer Institute Common Terminology Criteria for Adverse Events13,14; however, they have limited sensitivity13–17 and may miss improvements in CIPN resulting from a truly effective experimental treatment. In support of this point, a 2011 Cochrane review provides evidence that suboptimal CIPN measurement is a barrier to discovering new treatment.18 The authors reviewed 16 randomized placebo-controlled studies and concluded that there is no proven way to prevent CIPN.18 Although there are several possible explanations for the failure of these studies to identify effective interventions (eg, small sample sizes, possible subtherapeutic drug intervention dosing, our general lack of knowledge regarding CIPN mechanisms), the authors concluded that, in 12 of the 16 studies, CIPN measurement limitations contributed to an inability to accurately quantify the true impact of the interventions.18 Furthermore, Gewandter and colleagues19 reported in their 2017 review article that, of 7 published intervention trials for painful CIPN, only one used an appropriate primary outcome measure; this was the only trial of the seven reviewed to reveal an effective CIPN treatment.20 These studies provide evidence that researchers are using neither a consistent CIPN measure across trials, which limits cross-comparisons, nor the best and most appropriate CIPN measures. As a result, clinicians may be withholding potentially helpful treatments, based in part on flawed or inappropriate CIPN outcome measurement, which has limited our ability to accurately detect effects in intervention studies. The current study specifically addresses this problem by providing psychometric evidence that can inform the selection of CIPN measures for use in future intervention trials.

Several CIPN patient-reported outcome (PRO) measures have been tested in cancer populations and could be selected as a criterion standard measurement tool for future studies. Extensive psychometric testing has revealed strong psychometric properties in 2 such measures: the Functional Assessment of Cancer Therapy/Gynaecologic Oncology Group–Neurotoxicity scale and the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire–CIPN 20-item scale (EORTC QLQ-CIPN20).21–31 The EORTC QLQ-CIPN20 was chosen for further testing in the current study for 2 reasons: (1) we had access to QLQ-CIPN20 data from several large, multisite studies that were conducted in North America and Europe, which allowed psychometric testing in a more diverse sample than had been possible before, and (2) the QLQ-CIPN20 is in contention as a potential criterion-standard measure based on 9 studies that have provided evidence supporting multiple types of reliability and validity: internal consistency and test-retest reliability; sensitivity; content, structural, convergent, and discriminant validity; paper/electronic mode equivalence; and responsiveness.24–32 Although compelling empirical evidence suggests that the QLQ-CIPN20 does have strong psychometric properties, there is conflicting evidence regarding the instrument's structural validity/factor structure. Smith and colleagues29 conducted a secondary data analysis to test the instrument's internal consistency reliability, construct and convergent validity, and responsiveness to change over time. Results supported the reliability and validity of the QLQ-CIPN20, and sensory and motor scales exhibited moderate-high responsiveness to change. However, factor analysis results indicated that, instead of the original 3 QLQ-CIPN20 subscales (sensory, motor, and autonomic), a 2-factor structure of upper and lower extremity subscales may be more reliable and valid. In addition, low item-item correlations (r < 0.30) were found between all items comprising the autonomic scale, as well as the hearing loss item. The analyses suggested eliminating the 4 problematic items that compromised validity and devoting further attention to the resulting 16-item version of the instrument. In a subsequent study, Kieffer and colleagues26 were unable to confirm a factor structure and therefore recommended scoring the instrument as an additive checklist. Thus, the purpose of this current secondary data analysis of the pooled CIPN Outcome Measures Standardisation (CI-PeriNomS) study and the Alliance for Clinical Trials in Oncology (Alliance) study A151408 data was to retest the QLQ-CIPN's sensitivity, reliability, and structural validity using data obtained from a large international sample and to suggest possible revisions that could strengthen the instrument for use as a future criterion-standard measure.

Back to Top | Article Outline

Theoretical Framework

The EORTC QLQ-CIPN20 was designed to supplement the EORTC QLQ-30 QOL instrument.24 Consistent with Ferrell and colleagues'33 QOL model, the QLQ-CIPN20 measures symptoms of CIPN known to influence QOL in cancer survivors.20,34,35 The general factor model and classical measurement theory guided the methods that were used to test the QLQ-CIPN20's construct validity.36 On the basis of this theory, related scale items will cluster to form multiple factors (eg, sensory, motor, or autonomic nerve function), and each factor represents a dimension of the larger construct or latent variable (CIPN).37

The operational definitions of sensory, motor, and autonomic neuropathy used in this study are described hereinafter. Sensory neuropathy is demonstrated by new patient-reported sensations of numbness, tingling, or neuropathic pain in bilateral upper and lower extremities after the administration of neurotoxic chemotherapy. Sensory neuropathy is also evidenced by deficits in functions that require normal sensation in the toes/feet and fingers/hands. For example, impaired balance can be caused by foot numbness that impairs individuals' ability to detect the location of their feet on the floor. Motor neuropathy is evidenced by new-onset cramps in the hands and/or feet after neurotoxic chemotherapy treatment and by functional deficits due to weakness in the fingers/hands (eg, problems holding objects or opening jars) and toes/feet/legs (eg, difficulty walking or climbing stairs). However, ascribing exclusively either a sensory or motor neuropathy designation to functional deficits may not always be possible: in some cases, both sensory and motor peripheral nerve dysfunctions are involved (eg, inability to hold a pen can be caused by a combination of numbness and weakness in the fingers/hands). Finally, autonomic neuropathy can be demonstrated by problems that arise after neurotoxic chemotherapy treatment: new onset of erectile dysfunction, blurred vision, or dizziness due to orthostatic hypertension.

Back to Top | Article Outline

Methods

A secondary data analysis was conducted using pooled QLQ-CIPN20 data from a large sample (N = 1155) of European and North American patients. The data had been previously obtained from 271 patients from 25 sites in 8 countries in Europe (CI-PeriNomS)27 and 884 patients treated in 4 North American multisite CIPN treatment and prevention trials (N06CA, N08C1, N08CA, and N08CB) conducted by the North Central Cancer Treatment Group (NCCTG),8,38–40 which is now part of the Alliance, a large North American multisite cancer-focused research network. Institutional review boards/ethics committees at participating sites approved all studies. All study participants provided a signed, dated informed consent document before participation in the original studies.

The cross-sectional CI-PeriNomS study was designed to assess outcome measures for CIPN measurement and to test their validity and reliability. Participants were recruited at 25 participating academic and medical sites in Europe (Italy, 11; the Netherlands, 5; Spain, France, and Germany, 2 each; and the United Kingdom, Greece, and Austria, 1 each). The inclusion and exclusion criteria provided control for confounding variables that could compromise the ability to evaluate the QLQ-CIPN20's validity and reliability as a PRO measure of CIPN. Key inclusion criteria were (1) new-onset CIPN after standard chemotherapy, (2) aged 18 years or older, (3) clinically stable at least 1 month before screening for enrollment, and (4) Karnofsky performance score of 70 or greater. Key exclusion criteria were (1) active underlying malignancy and poor prognosis, (2) chemotherapy planned during study period, (3) conditions that would complicate assessment (eg, diabetes, alcohol abuse, neurological conditions), (4) severe depression, and (5) known peripheral neuropathy from other illness or medications.27 The QLQ-CIPN20 was translated and back-translated into several European languages using the EORTC's standardized procedures.41 European patients completed the QLQ-CIPN20 version in their native languages: either in the original English or in the Italian, Spanish, Greek, Dutch, German, or French translations. Participants of the CI-PeriNomS completed the QLQ-CIPN20 at 2 time points: (1) the first research visit and (2) a second visit 2 to 3 weeks later.

Three of the 4 North American NCCTG studies were designed to test a CIPN intervention.38,40,42 N06CA was a randomized, double-blind, placebo-controlled trial evaluating a topical agent for established CIPN.40 N08CA and N08CB were randomized, double-blind, placebo-controlled trials designed to evaluate neuroprotective agents for the prevention of paclitaxel/carboplatin-42 and oxaliplatin-induced38 neuropathy, respectively. Of note, none of the interventions tested in any of these trials was efficacious. The fourth study, N08C1, was a descriptive, longitudinal study that explored CIPN incidence and severity over time in patients receiving neurotoxic chemotherapy.8 Participants in these 4 studies were recruited from 125 academic and community sites located throughout the United States and Canada. Key inclusion criteria were age of 18 years or older, life expectancy of 4 to 6 months or greater (1 and 3 studies, respectively), and ECOG performance status of 0 to 1 (3 studies). Individuals with peripheral neuropathy from other causes and those receiving treatment of CIPN (3 studies) were excluded.29 North American study participants for the NCCTG completed the QLQ-CIPN20 at various time points.

The combined sample, from diverse geographic backgrounds, included patients aged 18 years or older with varied cancer diagnoses who were treated with all types of neurotoxic chemotherapy. All participants whose data were pooled for the analysis had either developed CIPN by the end of the parent study (ie, N08CA, N08C1, and N08CB) or entered the parent study with preexisting CIPN (ie, CI-PeriNomS and N06CA). Baseline data—of subjects who had already received neurotoxic chemotherapy and had preexisting CIPN—from the CI-PeriNomS and North American NCCTG N06CA trials and the final (12-week) time point data from the other 3 North American NCCTG studies—whose subjects were scheduled to receive neurotoxic chemotherapy—were used for this secondary analysis, because these were the time points for each cohort when CIPN was the most severe.

The QLQ-CIPN20 assesses sensory (9 items), motor (8 items), and autonomic (3 items) symptoms.24 Individuals indicate the degree to which they have experienced these symptoms during the previous week, using a 4-point Likert scale (1, not at all; 2, a little; 3, quite a bit; 4, very much). Raw scale scores range from 9 to 36 for sensory and from 8 to 32 for motor symptoms. Autonomic raw scores range from 2 to 8 for women and from 3 to 12 for men, because of the additional erectile function item.24 All scores are converted linearly to a 0 to 100 scale, and higher scores indicate a higher symptom burden.

Back to Top | Article Outline

Analyses

Data from the NCCTG and CI-PeriNomS studies were pooled (N = 1155). Subscale scores were converted to 100-point scales by averaging the items and multiplying by 25. The autonomic male subscale was constructed using items 16, 17, and 20; the female version does not include item 20 regarding erectile dysfunction. Sensitivity was determined based on whether the single item scores encompassed the full score range and whether responses clustered in low or high categories that reflected floor or ceiling effects. Internal consistency reliability was assessed using Cronbach's α and item-item correlations.43 Strong psychometric properties were indicated by α coefficients of .80 or greater and an item-item correlation range from 0.30 to 0.70; correlations greater than 0.70 suggest item redundancy. Complete pairwise item responses were used to estimate item-item associations. Structural validity of the QLQ-CIPN20 was initially assessed by confirmatory factor analysis (CFA) using structural equation modeling. To assess the CFA model fit, we evaluated several fit indices (χ2, root mean square error of approximation [RMSEA], and the comparative fit index). The RMSEA was considered the primary measure of fit, with values of 0.05 or less indicating a good fit and values greater than 0.08 indicating a mediocre fit.44 Per our previous exploratory factor analysis research, we hypothesized that the QLQ-CIPN20 CFA would reveal 2 distinct factors corresponding to upper and lower extremity CIPN.29 However, on the basis of published standards for acceptable fit indices,44,45 results of the current CFA did not support a 2-factor upper and lower extremity structure. Therefore, an exploratory factor analysis was conducted using established factor analysis guidelines.37,46 We removed 4 items with low item-item correlations (<0.30) before conducting the exploratory factor analyses. Bartlett's test of sphericity and Kaiser-Meyer-Olkin measures were used to evaluate item associations. The number of factors was determined using a scree plot, eigenvalues, and the percentage of the variance explained by the factors.

Back to Top | Article Outline

Results

Sample Characteristics

Table 1 illustrates the samples' demographic characteristics. The pooled sample (N = 1155) included patients with varied cancer diagnoses (ie, colon, breast, lung, myeloma, lymphoma, ovarian) who had received neurotoxic chemotherapy treatments. The mean age of the pooled sample was 59.4 years (range, 23–100 years), and most were female (64.9%). Because data regarding the frequencies of the varied diagnoses, neurotoxic treatments, and racial backgrounds were not available from all 5 studies, they are not reported.

Table 1

Table 1

Back to Top | Article Outline

Sensitivity

Table 2 illustrates QLQ-CIPN20 individual item and subscale scores for the pooled sample and the frequencies of floor and ceiling responses (minimum and maximum scores). On the basis of the original 3-factor structure, the mean pooled sensory, motor, and autonomic-male and autonomic-female subscale scores were 45.0 (SD, 16.4), 38.3 (SD, 14.3), 39.5 (SD, 14.1), and 36.2 (SD, 14.4), respectively.

Table 2

Table 2

Tingling was the most severe CIPN symptom (mean, 2.30–2.34; SD, 1.0–1.1). The lowest-scoring item was “Did you have difficulty using the pedals?” (mean, 1.2; SD, 0.5). Because all items encompassed the entire score range (1–4), no floor or ceiling effects were found. The Figure illustrates the frequency of response options for each item. Less than 5% of the sample reported severe (score, 4) cramps, inability to distinguish between hot and cold, difficulty holding a pen, foot drop, or leg weakness. Although uncommon, more patients reported severe shooting or burning pain in the toes or feet (7.2%) than in the fingers or hands (4.2%). Very few participants reported severe dizziness (1.5%), blurred vision (1%), difficulty hearing (0.9%), or difficulty using pedals (1.4%).

Figure

Figure

Back to Top | Article Outline

Internal Consistency Reliability

The Cronbach's α coefficients for the 3 subscales—sensory, motor, and autonomic female/male—of the original QLQ-CIPN20 instrument were .87, .83, and .62/.39, respectively. The low autonomic subscale α coefficients for men and women suggest that the items in this subscale are not internally consistent. Item-item, item-subscale, and item-total correlations are provided in Table 3. Consistent with our previous findings,29 the item-item correlation coefficients for items 16 to 20 (all autonomic subscale items and the hearing loss and pedal items) were generally less than 0.30, suggesting that these items may be unreliable indicators of peripheral neuropathy. The pedal item (#19) was retained because the research team hypothesized that participants who were concerned about potentially losing their driving privileges due to CIPN may have answered “not at all,” and this might explain the low item mean and standard deviation. However, because of the low α coefficients and item-item correlations, the autonomic and hearing loss items (items 16–18 and 20) were deleted before performing the factor analysis, resulting in a 16-item measure. Deletion of items 16, 17, and 20 is justified given that dizziness, blurred vision, and erectile dysfunction are caused by medications and other comorbid illnesses and cannot be solely attributed to CIPN. In addition, the hearing loss item (#18) was deleted because, although hearing loss can be caused by cisplatin treatment, it is commonly associated with other etiologies such as aging and noise exposure.

Table 3

Table 3

Items that are too highly correlated with other items also can be problematic. The item-item correlations between items 2 (tingling in the toes or feet) and 4 (numbness in the toes or feet) (r = .80) and items 1 (tingling in the fingers or hands) and 3 (numbness in the toes or feet) (r = 0.74) were greater than 0.70, suggesting item redundancy or that patients were confused about the difference between numbness and tingling. Nonetheless, because numbness and tingling are common clinical manifestations of CIPN, items 1 to 4 were retained. The Cronbach's α coefficients that emerged when testing the 16-item 2-factor scale were .90 and .85, which are higher than those of the original 20-item measure.

Back to Top | Article Outline

Structural Validity

Structural validity was assessed by examining the factor structure of the QLQ-CIPN20, beginning with a CFA. The data used in this analysis were from participants with no missing values (n = 946). χ2 describes similarity of the observed and expected matrices. Acceptable model fit is indicated by a χ2 probability greater than or equal to 0.05. For the CFA model, the χ2 value was 2345.79 and P < .0001. Root mean square error of approximation indicates the amount of unexplained variance or residual. The 0.1496 RMSEA value was larger than the criterion of 0.05 or less for good fit. In addition, the comparative fit index value (0.76) did not meet the criterion (0.90 or larger) for acceptable model fit.44 Therefore, several fit statistics indicated that the CFA model fit was unacceptable.

Because of this indication of an unacceptable model fit, 4 items with low item-item correlations were removed before conducting an exploratory factor analysis of the resulting 16-item version. A 2-factor solution was the best based on eigenvalues of at least 1.00, factor loadings of 0.40 or greater, or an explained 5% or greater of the variance in scores. A scree plot was also examined. The eigenvalues for factors 1 and 2 were 6.85 and 0.95, respectively. A moderate factor-to-factor correlation (r = 0.62) supports that oblimin (promax) rotation was an appropriate rotation approach. Bartlett's test indicated a factorable correlation matrix (χ2 = 8734.33, P < .0001), and the Kaiser-Meyer-Olkin measure of sampling adequacy was adequate at 0.88.37

Factors 1 and 2 explained 88% of the cumulative variance. Factor loading from the rotated factor-loading pattern matrix is reported in Table 4. Factor 1, which explained 77% of the variance, contains 7 items addressing numbness and tingling in the upper and lower extremities and difficulty manipulating small objects, opening jars, and holding a pen. Factor 2, which explained 11% of the variance, is composed of 9 items that assess cramps, burning pain, hot/cold temperature sensation, foot drop, and difficulty walking, climbing stars, and using the pedals of a car. The “pedals” item factor coefficient was 0.35, less than the 0.40 cut-point for inclusion in the factor.

Table 4

Table 4

Back to Top | Article Outline

Discussion/Conclusion

This study is one of the first known psychometric analyses of the QLQ-CIPN20 using a large, geographically diverse, multilingual sample from North America and Europe. The results suggest that the 16-item QLQ-CIPN20 is internally consistent and sensitive when used in patients in this population. Sensory neuropathy was the most severe type of neuropathy reported by the pooled sample.

Consistent with our previous findings,29 several items that were originally designed to evaluate autonomic neuropathy were not internally reliable, specifically those evaluating dizziness, blurred vision, and erectile dysfunction. These clinical problems, which can be due to a variety of confounding factors, are difficult to attribute solely to CIPN. The same is true for the hearing loss item; although this item should be retained when using the QLQ-CIPN20 to assess patients receiving cisplatin, it is important to recognize that hearing loss due to other causes will confound the reliability and validity of this item as a measure of CIPN-associated hearing loss.

After excluding those 4 items from the analysis, the remaining 16 items comprising the 2 factors/subscales were highly internally consistent based on improved α coefficients of 0.80 or greater and explained a high percentage of the variance in CIPN scores. This suggests that deleting these items does not compromise the instrument's validity and internal consistency reliability. Another benefit of item deletion is that the shorter instrument is more parsimonious and less burdensome for patients to complete.

We also provide evidence that the 16-item QLQ-CIPN20 possesses structural validity based on factor loadings and fit indices. However, similar to Kieffer et al's26 findings, the items did not cluster together based on either of the 2 previously reported conceptualizations: the 2-factor upper and lower extremity CIPN29 and the 3-factor sensory, motor, or autonomic neuropathy.24 Yet this lack of conceptual alignment may not pose a significant concern because some items may quantify uniquely important CIPN-related symptoms or functional deficits that do not necessarily cluster into distinct categories or latent variables. Moreover, CIPN signs and symptoms may arise from both sensory and motor nerve pathology, which invalidates the relevance of separate sensory and motor subscales. Thus, the QLQ-CIPN20's structural validity may be less important than other types of validity.

More specifically, 1 cogent explanation for the unstable factor structure may be linked to how participants interpreted the meaning of the QLQ-CIPN20 items and whether these interpretations are consistent with what clinicians and researchers expect. As 1 example, we know that clinicians can easily explain the difference between the sensations of numbness and tingling, but these differences may have been difficult for participants to discern, which could explain why the numbness and tingling items were highly correlated.

Moreover, social desirability can compromise instrument validity. Social desirability will influence item validity when patients provide answers that are socially desirable but not necessarily truthful. Regarding the current analysis, patients may not have provided valid responses to the driving (pedals) item because those who feared the loss of their driving privileges may not have answered this question honestly. A dishonest response is an invalid response.

These concerns about content validity and social desirability are validated by our subsequent research. After completing this secondary data analysis, the first in a series of three, we conducted a second study using cognitive interviews with 25 patients with CIPN to evaluate the content validity of the 16 items.32 We learned that patients can be confused about the difference between numbness and tingling, and this confusion likely explains the high item-item correlations among these items in the current study. In response to this discovery, we recommended the addition of “loss of feeling” parenthetically after the word “numbness” to clarify that the absence of sensation defines numbness but not tingling.32 Our results also provide a viable explanation for the low item-item correlations among the “difficulty using the pedals when driving a car” item and other scale items. Participants reported that others might not answer this question honestly because of concerns that their answers could result in the loss of driving privileges.32

To our knowledge, the previously described qualitative study is the only published research to rigorously evaluate the QLQ-CIPN20's content validity. Although the 20-item parent measure has been extensively tested, the shortened and slightly revised version needs further testing before it can be recommended as a criterion-standard PRO CIPN measure. Because the item revisions we made to improve clarity may have strengthened the QLQ-CIPN20's psychometric properties, we tested this revised tool via a prospective, longitudinal, case-control study involving 120 patients with cancer who were receiving a variety of neurotoxic agents and 30 healthy controls. Consistent with the findings of the current study and those of Kieffer et al,26 the results did not support the instrument's structural validity. The conceptual overlap among some signs and symptoms of sensory and motor neuropathy provides further support that structural validity may not be an important indicator of the psychometric strength of a PRO CIPN measure. The results of this last phase in our series of 3 studies suggest that a CIPN15 total summed score has strong psychometric properties: internal consistency and stability reliability, concurrent and contrasting group validity, sensitivity, and responsiveness to change.47

This study has several limitations. In addition to issues related to content validity, cultural and language differences may have influenced the factor structure findings. The importance and appropriateness of reporting pain may have varied significantly within the pooled sample, and hesitancy to report pain could have influenced the pain item scores. Furthermore, although the EORTC-CIPN20 was carefully forward- and back-translated, language variations across 8 European countries and North America may partially account for the unstable factor structure. In addition, the number of subjects using the English-language version was approximately 3 times the number of those who used all other language versions combined, which may have skewed results toward the English version; however, we did not stratify our analysis by language because of the limitation that some small subsample sizes in the CI-PeriNomS group would compromise the validity of the findings.

In summary, this article presents the results of phase 1 of a 3-phase study to extensively evaluate the EORTC's copyrighted QLQ-CIPN20 and to recommend revisions to further improve the tool. Our results suggest that the reduced 16-item tool is more internally consistent and parsimonious than the full 20-item version. Moreover, the reduced version is sensitive when used by patients with CIPN caused by varied classes of neurotoxic drugs. Although factor analysis results revealed an unstable factor structure, this is likely not a major concern based on subsequent testing of a summed score. Findings of the current study informed phase 2 in which we conducted cognitive interviews to test the content validity of the best items emerging from phase 1.32 Subsequently, via a prospective case-control study47 in a third phase, we tested the reliability, validity, sensitivity, and responsiveness to change of a revised QLQ-CIPN20 (the CIPN15) that emerged from phase 2. Thus, the results of the current study informed two others. When considering the new evidence from this and our other 2 studies,32,47 the QLQ-CIPN20 is now the most extensively studied CIPN PRO measure available for use in research and clinical settings. The tool has been tested in populations that have received many different classes of neurotoxic drugs and in those from geographically and linguistically diverse backgrounds in North America and Europe. Although instruments are never fully validated, the reduced version emerging from this current research and, perhaps, a further refined version based on the 2 subsequent phases should be considered for use in future CIPN intervention trials based on an extensive body of empirical evidence supporting the QLQ-CIPN's strong psychometric properties.

Very few effective treatments for CIPN are known, in part because our ability to discover them hinges on having a reliable and valid outcome measure for use in intervention research. Consistent use of a strong PRO measure, such as the suggested revision of the QLQ-CIPN instrument, will improve measurement validity in subsequent intervention trials and thereby aid in the discovery of effective treatments. The outcome of this psychometric study has taken us 1 step closer to identifying a parsimonious and psychometrically sound measure that researchers can agree to use in future intervention trials, a measure that will be more likely to uncover effective treatments for patients suffering with CIPN. The necessary ongoing research to test modified versions of the QLQ-CIPN instrument could be facilitated via secondary research aims in future intervention studies that use the instrument.

Back to Top | Article Outline

ACKNOWLEDGMENTS

The authors acknowledge Nusara Prasertsri, PhD, RN, and Warunee Phligbua, PhD, RN, for their international perspectives during data interpretation. They also acknowledge Dr David Cornblath for his role in facilitating collaborations with the European CI-PeriNomS investigators. The CI-PeriNomS and Alliance for Clinical Trials in Oncology (A151408) groups provided the data for this research but did not participate in the data analysis described in this article.

Back to Top | Article Outline

References

1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin. 2013;63(1):11–30. Accessed January 24, 2013.
2. Argyriou AA, Bruna J, Marmiroli P, Cavaletti G. Chemotherapy-induced peripheral neurotoxicity (CIPN): an update. Crit Rev Oncol Hematol. 2012;82(1):51–77.
3. Windebank AJ, Grisold W. Chemotherapy-induced neuropathy. J Peripher Nerv Syst. 2008;13:27–46.
4. Hausheer FH, Schilsky RL, Bain S, Berghorn EJ, Lieberman F. Diagnosis, management, and evaluation of chemotherapy-induced peripheral neuropathy. Semin Oncol. 2006;33(1):15–49.
5. Stubblefield MD, Burstein HJ, Burton AW, et al. NCCN task force report: management of neuropathy in cancer. J Natl Compr Canc Ne. 2009;7(suppl 5):S1–S26.
6. Argyriou AA, Zolota V, Kyriakopoulou O, Kalofonos HP. Toxic peripheral neuropathy associated with commonly used chemotherapeutic agents. J BUON. 2010;15(3):435–446.
7. Paice JA. Clinical challenges: chemotherapy-induced peripheral neuropathy. Semin Oncol Nurs. 2009;25(2 suppl 1):S8–S19.
8. Loprinzi CL, Reeves BN, Dakhil SR, et al. Natural history of paclitaxel-associated acute pain syndrome: prospective cohort study NCCTG N08C1. J Clin Oncol. 2011;29(11):1472–1478.
9. Visovsky C, Daly BJ. Clinical evaluation and patterns of chemotherapy-induced peripheral neuropathy. J Am Acad Nurse Pract. 2004;16(8):353–359.
10. Wampler MA, Miaskowski C, Hamel K, Byl N, Rugo H, Topp LS. The modified Total Neuropathy Score: a clinically feasible and valid measure of taxane-induced peripheral neuropathy in women with breast cancer. Support Oncol. 2006;4(8):W9–W16.
11. Bakitas MA. Background noise: the experience of chemotherapy-induced peripheral neuropathy. Nurs Res. 2007;56(5):323–331.
12. Smith EM, Bakitas MA, Homel P, et al. Preliminary assessment of a neuropathic pain treatment and referral algorithm for patients with cancer. J Pain Symptom Manage. 2011;42(6):822–838.
13. Cavaletti G, Frigeni B, Lanzani F, et al. Chemotherapy-induced peripheral neurotoxicity assessment: a critical revision of the currently available tools. Eur J Cancer. 2010;46(3):479–494.
14. Griffith K, Merkies IS, Hill E, Cornblath D. Measures of chemotherapy-induced peripheral neuropathy: a systematic review of psychometric properties. J Peripher Nerv Syst. 2010;15(4):314–325.
15. Postma TJ, Heimans JJ, Muller MJ, Ossenkoppele GJ, Vermorken JB, Aaronson NK. Pitfalls in grading severity of chemotherapy-induced peripheral neuropathy. Ann Oncol. 1998;9(7):739–744.
16. Frigeni B, Piatti M, Lanzani F, et al. Chemotherapy-induced peripheral neurotoxicity can be misdiagnosed by the National Cancer Institute Common Toxicity scale. J Peripher Nerv Syst. 2011;16(3):228–236.
17. Lavoie Smith EM, Cohen JA, Pett MA, Beck SL. The validity of neuropathy and neuropathic pain measures in patients with cancer receiving taxanes and platinums. Oncol Nurs Forum. 2011;38(2):133–142.
18. Albers JW, Chaudhry V, Cavaletti G, Donehower RC. Interventions for preventing neuropathy caused by cisplatin and related compounds. Cochrane Database Syst Rev. 2011;(2):CD005228.
19. Gewandter JS, Dworkin RH, Finnerup NB, Mohile NA. Painful chemotherapy-induced peripheral neuropathy: lack of treatment efficacy or the wrong clinical trial methodology? Pain. 2017;158(1):30–33.
20. Smith EM, Pang H, Cirrincione C, et al. Effect of duloxetine on pain, function, and quality of life among patients with chemotherapy-induced painful peripheral neuropathy: a randomized clinical trial. JAMA. 2013;309(13):1359–1367.
21. Calhoun EA, Welshman EE, Chang CH, et al. Psychometric evaluation of the functional Assessment of Cancer Therapy/Gynecologic Oncology Group-Neurotoxicity (Fact/GOG-Ntx) questionnaire for patients receiving systemic chemotherapy. Int J Gynecol Cancer. 2003;13(6):741–748.
22. Cella D, Peterman A, Hudgens S, Webster K, Socinski MA. Measuring the side effects of taxane therapy in oncology: the Functional Assessment of Cancer Therapy-taxane (FACT-taxane). Cancer. 2003;98(4):822–831.
23. Kopec JA, Land SR, Cecchini RS, et al. Validation of a self-reported neurotoxicity scale in patients with operable colon cancer receiving oxaliplatin. Support Oncology. 2006;4(8):W1–W7.
24. Postma TJ, Aaronson NK, Heimans JJ, et al. The development of an EORTC quality of life questionnaire to assess chemotherapy-induced peripheral neuropathy: the QLQ-CIPN20. Eur J Cancer. 2005;41(8):1135–1139.
25. Alberti P, Rossi E, Cornblath DR, et al. Physician-assessed and patient-reported outcome measures in chemotherapy-induced sensory peripheral neurotoxicity: two sides of the same coin. Ann Oncol. 2014;25(1):257–264.
26. Kieffer JM, Postma TJ, van de Poll-Franse L, et al. Evaluation of the psychometric properties of the EORTC chemotherapy-induced peripheral neuropathy questionnaire (QLQ-CIPN20). Qual Life Res. 2017;26(11):2999–3010.
27. Cavaletti G, Cornblath DR, Merkies IS, et al. The chemotherapy-induced peripheral neuropathy outcome measures standardization study: from consensus to the first validity and reliability findings. Ann Oncol. 2013;24(2):454–462.
28. Knoerl R, Gray E, Stricker C, et al. Electronic versus paper-pencil methods for assessing chemotherapy-induced peripheral neuropathy. Support Care Cancer. 2017;25(11):3437–3446.
29. Lavoie Smith EM, Barton DL, Qin R, Steen PD, Aaronson NK, Loprinzi CL. Assessing patient-reported peripheral neuropathy: the reliability and validity of the European Organization for Research and Treatment of Cancer QLQ-CIPN20 questionnaire. Qual Life Res. 2013;22(10):2787–2799.
30. Velasco R, Postma TJ, Aaronson N, Simo M, Bruna J. Preliminary validation of the EORTC Chemotherapy-Induced Peripheral Neuropathy Quality of Life Questionnaire (QLQ-CIPN20) Spanish version in a series of multiple myeloma patients treated with bortezomib: P.022*. Neuro-oncology. 2010;12(3 suppl 3):26.
31. Kim HY, Kang JH, Youn HJ, et al. Reliability and validity of the Korean version of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire to assess chemotherapy-induced peripheral neuropathy. J Korean Acad Nurs. 2014;44(6):735–742.
32. Smith EM, Haupt R, Kelly J, et al. The content validity of a chemotherapy-induced peripheral neuropathy patient-reported outcome measure. Oncol Nurs Forum. 2017;44(5):580–588.
33. Ferrell BR, Dow KH, Leigh S, Ly J, Gulasekaram P. Quality of life in long-term cancer survivors. Oncol Nurs Forum. 1995;22(6):915–922.
34. Mols F, Beijers T, Lemmens V, van den Hurk CJ, Vreugdenhil G, van de Poll-Franse LV. Chemotherapy-induced neuropathy and its association with quality of life among 2- to 11-year colorectal cancer survivors: results from the population-based PROFILES registry. J Clin Oncol. 2013;31(21):2699–2707.
35. Tofthagen C, Donovan KA, Morgan MA, Shibata D, Yeh Y. Oxaliplatin-induced peripheral neuropathy's effects on health-related quality of life of colorectal cancer survivors. Support Care Cancer. 2013;21(12):3307–3313.
36. DeVellis RF. Scale Development: Theory and Applications. Vol. 26. 2nd ed. Thousand Oaks, CA: Sage Publications; 2003:171.
37. Pett MA, Lackey NR, Sullivan JL. Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research. Thousand Oaks, CA: Sage Publications; 2003:348.
38. Loprinzi CL, Qin R, Dakhil SR, et al. Phase III randomized, placebo-controlled, double-blind study of intravenous calcium and magnesium to prevent oxaliplatin-induced sensory neurotoxicity (N08CB/alliance). J Clin Oncol. 2013;32(10):997–1005.
39. Reeves BN, Dakhil SR, Sloan JA, et al. Further data supporting that paclitaxel-associated acute pain syndrome is associated with development of peripheral neuropathy: North Central Cancer Treatment Group trial N08C1. Cancer. 2012;118(20):5171–5178.
40. Barton DL, Wos EJ, Qin R, et al. A double-blind, placebo-controlled trial of a topical treatment for chemotherapy-induced peripheral neuropathy: NCCTG trial N06CA. Support Care Cancer. 2011;19(6):833–841.
41. Kuliś D, Bottomley A, Velikova G, Greimel E, Koller M. EORTC translation manual. http://groups.eortc.be/qol/sites/default/files/img/newsletter/translation_manual_2017.pdf. Updated 2017. Accessed May 30, 2017.
42. Leal AD, Qin R, Atherton PJ, et al. North Central Cancer Treatment Group/Alliance trial N08CA—the use of glutathione for prevention of paclitaxel/carboplatin-induced peripheral neuropathy: a phase 3 randomized, double-blind, placebo-controlled study. Cancer. 2014;120(12):1890–1897.
43. DeVon HA, Block ME, Moyle-Wright P, et al. A psychometric toolbox for testing validity and reliability. J Nurs Scholarsh. 2007;39(2):155–164.
44. Hu LL. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Model. 1999;6(1):1–55.
45. Clayton MF, Pett MA. Modeling relationships in clinical research using path analysis part II: evaluating the model. J Spec Pediatr Nurs. 2011;16(1):75–79.
46. Fayers PM. Quality-of-life measurement in clinical trials—the impact of causal variables. J Biopharm Stat. 2004;14(1):155–176.
47. Smith EML, Knoerl R, Yang JJ, Kanzawa-Lee G, Lee D, Bridges CM. In Search of a Gold Standard Patient-Reported Outcome Measure for Use in Chemotherapy- Induced Peripheral Neuropathy Clinical Trials. Cancer Control. 2018;25(1):1073274818756608.
Keywords:

Chemotherapy-induced peripheral neuropathy (CIPN); QLQ-CIPN20; Reliability; Sensitivity; Validity

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved