The Hand-Foot Skin Reaction and Quality of Life Questionnaire: An Assessment Tool for Oncology

This study describes the development and validation of a brief, patient self-reported questionnaire (the hand-foot skin reaction and quality of life questionnaire) supporting its suitability for use in clinical research to aid in early recognition of symptoms, to evaluate the effectiveness of agents for hand-foot skin reaction (HFSR) or hand-foot syndrome (HFS) treatment within clinical trials, and to evaluate the impact of these treatments on HFS/R-associated patients’ health-related quality of life.

Abstract

Background.

Skin toxicity (hand-foot syndrome/hand-foot skin reaction, HFS/R) related to antineoplastic therapy is a significant issue in oncology practice, with potentially large impacts on health-related quality of life (HRQL).

Materials and Methods.

A patient-reported questionnaire, the hand-foot skin reaction and quality of life (HF-QoL) questionnaire was developed to measure the HFS/R symptoms associated with cancer therapeutic agents and their effect on daily activities. The validity and reliability of the HF-QoL questionnaire was tested in a randomized trial of capecitabine with sorafenib/placebo in 223 patients with locally advanced/metastatic breast cancer. Other measures completed included patient ratings of condition severity, the Functional Assessment of Cancer Therapy-Breast cancer (FACT-B), and the clinician-rated National Cancer Institute-Common Terminology Criteria for Adverse Events (NCI-CTCAE), version 3.0, hand-foot skin reaction grade. The psychometric properties of the HF-QoL tested included structural validity, internal consistency, construct validity, discriminant validity, and responsiveness. Finally, the minimal clinically important difference (MCID) was estimated.

Results.

The HF-QoL instrument comprises a 20-item symptom scale and an 18-item daily activity scale. Each scale demonstrated excellent measurement properties and discriminated between NCI-CTCAE grade and patient-rated condition severity with large effect sizes. The daily activity scale had excellent internal consistency and correlated with the FACT-B and HF-QoL symptom scores. Both HF-QoL scale scores increased linearly with increasing patient-rated condition severity. The MCIDs were estimated as 5 units for daily activities and 8 units for symptoms mean scores.

Conclusion.

The HF-QoL was sensitive to symptoms and HRQL issues associated with HFS/R among participants treated with capecitabine with and without sorafenib. The HF-QoL appears suitable for assessing the HRQL impairment associated with HFS/R to cancer therapies.

Implications for Practice:

Skin toxicity related to anticancer therapies is a significant issue in oncology practice. Several newer agents, as well as older therapies, are associated with the skin toxicity known as hand-foot skin reaction (HFSR) or hand-foot syndrome (HFS). This study describes the development and validation of a brief, patient-reported questionnaire (the hand-foot skin reaction and quality of life questionnaire) supporting its suitability for use in clinical research to aid in early recognition of symptoms, to evaluate the effectiveness of agents for HFS/R treatment within clinical trials, and to evaluate the impact of these treatments on HFS/R-associated patients' health-related quality of life.

Introduction

Skin toxicity related to antineoplastic therapy is a significant issue in oncology practice. Several newer multikinase inhibitors (MKIs), such as sorafenib, sunitinib, axitinib, regorafenib, and BRAF inhibitors (dabrafenib, vemurafenib) are associated with the skin toxicity known as hand-foot skin reaction (HFSR), occurring in 20%–40% of treated patients [1–3]. HFSR is a histopathologically and clinically distinct variant of the more widely known hand-foot syndrome (HFS; palmar-plantar dysesthesia, acral erythema) occurring with older chemotherapeutic agents such as 5-fluorouracil, capecitabine, or doxorubicin [4, 5]. Of these, capecitabine, in particular, is associated with a more intense HFS. Although HFS and HFSR share clinical features such as bilateral occurrence, palmar-plantar localization, tenderness, and pain [6], HFS is more diffuse and HFSR is localized at the pressure points and includes conditions such as inflammatory hyperkeratosis. HFS and HFSR (HFS/R) can occur simultaneously when these older chemotherapeutic agents are used in combination with MKIs, resulting in increased skin toxicity and posing substantial challenges in patient management [6].

Although HFS/R is usually self-limiting and rarely leads to hospitalization or life-threatening manifestations, the symptom burden can result in significant deficits in health-related quality of life (HRQL) [7] and can thus pose potential risk for drug withdrawal, dose reduction, or drug interruption [1, 8, 9]. In treatment regimens with an expected probability of significant HFS/R, recognizing the early symptoms could hasten the use of management strategies [10].

The common practice in oncology clinical studies of monitoring more severe (i.e., grade 2 and 3) skin toxicities [11–15] has provided insufficient information on the extent of HRQL impairment with HFS/R, such as the level and progression of physical discomfort, restricted movement, tiredness, and distress. Patient-reported outcome (PRO) measures record a patient’s global experience with cancer treatment, including both subjectively appraised harms and benefits from treatment [16, 17]. HRQL is a secondary naturalistic goal of treatment in clinical trials and a primary outcome with treatment alternatives effective in prolonging life or controlling disease progression.

In line with the Food and Drug Administration (FDA) recommendations for patient-reported outcome measures [18] and the need for a tool to assess HRQL information pertaining to HFS/R in cancer research settings, we developed and validated a brief, patient self-reported questionnaire (hand-foot skin reaction and quality of life [HF-QoL] questionnaire). Although one self-reported instrument to measure HFS/R exists, the 14-item HFS-14 [19], that instrument solely assesses the functional implications of HFS, including pain, but not the wider symptom burden associated with the condition; the HF-QoL assesses both. The present report describes the scientific methods and key findings of the HF-QoL development and validation study regarding its suitability for use in clinical research with cancer chemotherapeutic agents and targeted agents (sorafenib).

Materials and Methods

HF-QoL Questionnaire Development

As a summary of a more detailed process, the HF-QoL questionnaire was developed in multiple and interactive steps that included input from patients and clinicians. The scientific HFS/R data were systematically reviewed to include an evaluation of the adequacy of existing validated dermatology PRO questionnaires assessing HRQL in HFS/R: the Skindex [20], the Dermatology Quality of Life Index [21, 22], and the Dermatology Specific Quality-of-Life questionnaire [23, 24].

The conceptual framework for the prototype HF-QoL included physical well-being, activities of daily living, role function, social activities, emotional well-being, and mood. Draft items were developed from HFS/R patient and clinician interviews. The prototype HF-QoL was reviewed by 8 clinicians familiar with HFS/R patient outcomes after chemotherapy and targeted therapies. Relevance and face validity of the instrument was then tested via face-to-face semistructured interviews with 7 patients with a recent history of HFS/R after antineoplastic therapy. On institutional review board approval, the final questionnaire was examined via a series of cognitive debriefing interviews with a new sample of 7 patients. These patients had recent or recurrent HFS/R and were undergoing cancer treatment in a comprehensive cancer treatment center; they provided feedback on wording, content, and answer rationale.

The final HF-QoL for validation testing included 48 items grouped conceptually within 2 domains: HFS/R symptoms, comprising 24 items assessing HFS/R signs and symptoms experienced with feet, hands and fingers, and other body areas; and HFS/R daily activities comprising 24 items among subscales for gross physical function, hand-related functioning, social activities, and psychological impact of skin toxicity. Each item was rated on a 5-point scale from 0 (not at all) to 4 (always or extremely) during the previous week. Three additional global items were administered for questionnaire validation and calculation of the minimal clinically important difference (MCID; definition and details provided in Minimal Clinically Important Difference). The three items were the global health appraisal assessed using a feeling thermometer (visual analog scale) rated on a scale of 0 (worst health) to 10 (best health); overall current HFS/R severity rated as 0 (no symptoms at all), 1 (mild), 2 (moderate), and 3 (severe); and changes in HFS/R since completion of the previous questionnaire, rated as 0 (about the same), 1 (better), 2 (worse).

HF-QoL Questionnaire Validation

Study Design and Patients

The item reduction and testing of the validity and reliability of the HF-QOL was performed in SOLTI-0701 (Grupo Español de Estudio, Tratamiento y Otras Estrategias Experimentales en Tumores Sólidos) (European Clinical Trials Database identification no. 2007-000290-32), a large, double-blind, randomized, phase IIb study evaluating the efficacy and safety of sorafenib compared with placebo when administered in combination with capecitabine in patients with locally advanced or metastatic breast cancer. Eligible patients were aged 18 years or older with histologically or cytologically confirmed human epidermal growth factor receptor 2-negative adenocarcinoma of the breast and locally advanced (inoperable) or metastatic disease. The patients were randomly assigned 1:1 to oral capecitabine 1,000 mg/m² twice a day on days 1 to 14 of a 21-day cycle with either oral sorafenib 400 mg twice a day continuously or placebo. The details of this trial have been previously reported [14].

Study Procedures

Patients gave written informed consent, and the study complied with the FDA Good Clinical Practices, Health Protection Branch guidelines, and all applicable ethical, legal, and regulatory requirements. The patients completed the HF-QoL, Functional Assessment of Cancer Therapy-Breast cancer (FACT-B), version 4 [25], and the three global items at each clinic visit. The FACT-B (breast cancer-specific format) assesses 5 dimensions of quality of life: physical, social/family, emotional, and functional well-being among 27 items and a 9-item breast cancer subscale. The FACT-G (FACT-general, nonsite-specific format) is a subset of the FACT-B formed by removing the 9 breast cancer-specific items. The FACT-B was chosen as the HF-QoL comparator, because it contains both breast cancer-specific symptoms and HRQL effects; evidence of the reliability and validity of both the FACT-G and the FACT-B in this patient population have been previously reported [25]. Higher scores on the FACT scales indicate a better quality of life. The HF-QoL and FACT-B were completed at baseline (within 2 weeks of randomization), every 3 (HF-QoL) or 6 (FACT) weeks during the first 24 weeks of the study, and every 9 weeks thereafter until the end of the study or withdrawal.

Clinical evaluations were made after each clinic visit by the investigator using the National Cancer Institute-Common Terminology Criteria for Adverse Events (NCI-CTCAE), version 3.0, grading for HFSR [26] (Table 1). Clinical validity of the HF-QoL was assessed by comparing the HF-QoL scores with the NCI-CTCAE clinical grades.

Table 1.

National Cancer Institute-Common Terminology Criteria for Adverse Events grade

Statistical Analysis

Data analyses were performed using SAS, version 9 (SAS Institute Inc., Cary, NC, http://www.sas.com/en_us/software/analytics/). A psychometric validation sample was created that included data collected from participants with ≥50% of HF-QoL items available at study weeks 3, 6, and 9, this being the timeframe in which skin toxicities are known to become incident. Data are presented as mean ± SD or SE, or n (%). Statistical significance (p < .05) was determined from chi-square tests, t tests, and analyses of variance and covariance, with testing for linear trend, as appropriate for the data distribution. The power of the study was such that a moderate effect size of 0.4 SD could be detected at a 2-sided p < .05 between 2 groups of equal size (n = 100).

HF-QoL Scoring

The HF-QoL symptom and daily activity total scores were transformed to a scale of 0 to 100, based on the sum of each unit-weighted item score divided by the maximum score. Higher scores on the HF-QoL indicate worse quality of life or greater symptom burden.

Conceptual Models

The conceptual model of the HF-QoL daily activity domain was based on items representing an HRQL latent construct in which HFS/R toxicity exerts an overall effect on daily life through symptom burden, affecting the daily activities performed, emotional well-being, and social activities. The final and optimal set of HF-QoL items was identified from an initial set of 24 daily activity and 24 symptoms items based, as appropriate, on factor loadings, interitem correlations, and consistency with the clinician and self-rated HFS/R severity scores. The performance of the HF-QoL daily activity domain was tested using standard psychometric procedures to establish reliability and validity. The HF-QoL symptom cluster items represent an inventory of symptoms and therefore relied on a criterion-referenced measurement approach to assess the measurement properties using the NCI-CTCAE toxicity grade and patient-rated HFS/R condition severity as criteria.

The tests of the daily activity domain included individual item analyses, such as correlations among items, and scale properties, such as internal consistency reliability, construct validity (structural, convergent, known groups), responsiveness/sensitivity to change, and estimation of the MCID. Items were retained based on evidence of measurement congruence or precision on the latter indicators of measurement quality. For the symptoms domain, the items were assessed according to the response distribution and correspondence with patient and clinician global ratings of HFS/R severity and NCI-CTCAE (version 3.0) toxicity grades.

Internal Consistency Reliability

Internal consistency reliability, the extent to which items in a scale are all measuring the same concept [27], was assessed for the HF-QoL daily activity domain using Cronbach’s α, with α ≥0.80 [28] taken to indicate good internal consistency.

Construct Validity

Validity, or the extent to which the instrument measures the concept it is intended to measure [29], was examined by assessing the HF-QoL daily activity scale structure (structural validity) using confirmatory factor analysis (CFA). The SAS system procedure CALIS, version 9.4 (SAS Institute, Inc.) was used to fit the one factor model and the full information maximum likelihood method was used to estimate the parameters under missing data conditions. Selective improvements in the model were achieved by examining misfit in the normalized covariance residuals between the predicted versus observed covariances and the Lagrangian modification statistics from the model. Items were loaded onto one factor per time point, and the factors at each time point were allowed to correlate among each other to handle autocorrelation. Model fit was evaluated using well-known incremental fit indexes such as Bentler’s nonnormed index and the comparative fit index (CFI). It was hypothesized that all HF-QoL daily activity items would load on one factor. A good fit was indicated by a CFI >0.90 and factor loadings of ≥0.40 [30]. For the HFS/R symptom cluster items, a latent, theory-based, conceptual model was not presumed. Instead, a criterion-referenced measurement approach was taken using the NCI-CTCAE toxicity grade as the criterion.

Agreement of the HF-QOL with other measures of similar or overlapping constructs (convergent validity) was assessed by calculating the correlations (Pearson’s r) between the HF-QoL daily activity and FACT-B and FACT-G scores. As a guideline, these correlations should be at least moderate at r ≥ .40 [31].

Known-groups validity was assessed in terms of whether the HF-QoL daily activity and symptom scores distinguished patient- and clinician-rated severity (i.e., patient-rated HSF/R condition severity and clinician-rated NCI-CTCAE grade). The mean values and p values for tests of trend by group level were obtained from mixed model regression with participant as a random effect and week and grade as predictors. Effect sizes were described as Cohen’s F-square statistic with values of 0.02, 0.15, and 0.35 as anchors for small, moderate, and large effects, respectively [32]. It was hypothesized that discrimination would be largest in terms of condition-specific measures. Item performance of the HF-QoL symptom items was based on ability to discriminate between NCI-CTCAE toxicity grades.

Responsiveness

Responsiveness measures the extent that score changes are associated with changes in underlying clinical status (relevant only for the HF-QoL daily activity domain). Adjusted mean values were obtained from mixed models with repeated measures for each patient at weeks 3, 6, and 9, with a change in HFQOL as the dependent variable and changes in grade, previous grade, and week fit to the data. Patient clustering was accounted for using a patient random effect.

Minimal Clinically Important Difference

The MCID, a measure of HF-QOL responsiveness and the smallest group level change or difference indicative of a clinically important improvement or decline, was estimated by calculating the change in slopes for the daily activity and symptoms scores according to a unit change in the NCI-CTCAE toxicity grade over weeks 3–6 and 6–9, combined with a test of linear trend. Changes in NCI-CTCAE grade from weeks 3 to 6 and 6 to 9 thus served as the anchors to examine the same period changes in the HF-QoL scores, adjusting for week and previous HF-QoL score.

Results

Study Population

Between August 2007 and December 2008, 229 female breast cancer patients were assessed at screening. The demographic and clinical characteristics of these patients are listed in Table 2. After excluding 6 patients who provided no HF-QoL data at baseline or at 3 weeks, the final sample comprised 223 patients. The NCI-CTCAE grade for these 223 patients at each assessment is listed in Table 1. At baseline, no evidence was seen of HFS/R (NCI-CTCAE grade 0) in 194 of the 223 patients (87%); for 12%, this had not been assessed or was unknown. By week 9, the proportion of patients with grade 0 toxicity had declined to 34.1%; the proportion with grade 2+ toxicity was 19.2%, and 5.4% had not been assessed (Table 1).

Table 2.

Demographic and clinical characteristics of breast cancer patients assessed at screening (n = 229)

Validity and Reliability of HF-QoL

The CFA of the HF-QoL daily activity items at each study week was consistent with the existence of a single underlying construct. Six items performed suboptimally, however, and were removed, leaving 18 items in the revised final scale (score range, 0–72). Although the factor loadings in the 1-factor model were ≥0.50, the model fit using the CFI statistic ranged from 0.75 to 0.82 across the three time points, less than the ideal 0.90 threshold. Model fit was improved by modifying the structural model to include a first order factor affecting the four daily activity subdomains listed in Table 3, which in turn, individually affected their respective items. Allowing for subscales served to improve the CFI statistic in the new model to 0.91 across all time points.

Table 3.

Hand-foot skin reaction and quality of life instrument with abbreviated item descriptions

For the HFS/R symptom cluster items, 20 items were significantly (p < .05) associated with the NCI-CTCAE grade and constituted the revised final scale (score range, 0–80). The final set of 18 HF-QoL daily activity and 20 symptom items is given in Table 3.

HF-QoL Daily Activity Scale

The item reliability of the HF-QoL daily activity scale assessed using Cronbach’s α coefficient at weeks 3–9 was consistently 0.96. The four subdomains (physical, self-care, social, and psychological) and the total score showed a correlation with the FACT-B and FACT-G total scores in the expected direction (data not shown). For the latter, a representative result was r = −.43 to −.53 (week 6), indicating that the increased quality of life impairment was associated with a higher symptom burden.

The mean HF-QoL daily activity scores stratified by NCI-CTCAE grade (Table 4) show linear associations of increasing scores by toxicity grade (p < .0001). Furthermore, the HF-QoL daily activity scores showed a strong linear trend for increasing score with increasing patient-rated condition severity (data not shown; p < .001). The FACT-B and FACT-G total scores were significantly different across the toxicity grade but did not display an obvious linear pattern in this regard (Table 4).

Table 4.

HF-QoL and FACT scores stratified by NCI-CTCAE toxicity grade (average effect across visits at 3, 6, and 9 weeks)

The mean ± SE changes in daily activity score, rescaled from 0 to 100, by category of grade change combined across weeks 3–9 (i.e., change over weeks 3–6 and 6–9 combined) are listed in Table 5. A decrease in the NCI-CTCAE grade (reduction in toxicity, −1 to −3) during the 9-week period displayed a stepped reduction in HF-QoL deficits; likewise an increase in NCI-CTCAE burden over the same period resulted in an increase in HF-QoL deficits in daily activity scores (overall p = .0024) across the levels of change in the NCI-CTCAE grade. A test of the influence of the previous NCI-CTCAE grade level and specific week (i.e., 3, 6, or 9) showed no significant differences at p < .05 (Wald F-test for interactions with lag grade and week), indicating that neither the week nor the previous NCI-CTCAE grade changed the overall relationship of the changes in NCI-CTCAE grade with the changes in the HF-QoL daily activity scores. The association between grade change and change in HF-QoL daily activity score from weeks 3–9 is shown in Figure 1A.

Table 5.

Adjusted mean ± SE change in HF-QoL daily activity and symptom scale scores stratified by change in NCI-CTCAE toxicity grade for weeks 3–6 and 6–9 combined

Figure 1.

Adjusted hand-foot skin reaction and quality of life (HF-QoL) score changes from weeks 3 to 6 and 6 to 9 combined with the National Cancer Institute-Common Terminology Criteria for Adverse Events toxicity grade change. (A): HF-QoL daily activities. (B): HF-QoL symptoms.

Abbreviation: QOL, quality of life.

HF-QOL Symptom Scale

The correlations of the HF-QoL symptom score with the FACT-B and FACT-G scores were r = −.31 to −.32 (p < .00001) in the expected direction. A significant (p < .001) gradient in adjusted HF-QoL symptom score mean was observed with NCI-CTCAE grades 0–3 (Table 4). Just as was also observed for the daily activity scores, the greater the reduction in the NCI-CTCAE toxicity grade across weeks 3–9, the greater the reduction in HF-QoL symptom score mean (p < .0001) and the greater the increase in NCI-CTCAE grade, the greater the increase in HF-QoL symptom score (Table 5). This result was not modified by week or the previous NCI-CTCAE grade. The association between grade change and change in HF-QoL symptom score from weeks 3–9 is shown in Figure 1B.

Minimal Clinical Important Difference

From the test for linear trend, a unit positive change in toxicity grade (from less to more severe) from weeks 3–6 and 6–9 combined was associated with a mean 5.03 increase (p < .0001) in the HF-QoL daily activity total score and a mean 8.24 increase (p < .0001) increase in the HF-QoL symptom score (scaled from 0 to 100; detailed in Table 5 footnote). Thus, a within-patient change of 5 units or more in the HF-QoL daily activity and 8 units or more in the HF-QoL symptom scores, on a scale from 0 to 100, could be considered meaningful at the individual patient level.

Discussion

The present report describes the development and psychometric properties of a new brief patient-reported instrument, the HF-QoL, which measures the extent of HFS/R symptoms and their effects on daily activities after antineoplastic treatment. Although HFS/R symptoms are common in many oncology treatment settings and graded as toxicity, few patient-reported measures have been developed to assess this treatment burden more accurately. The administration in the present study of the HF-QoL to patients treated with the antineoplastic agent sorafenib, combined with capecitabine, provided evidence of its validity and reliability in this patient population. The HF-QoL daily activity and symptom domains demonstrated sensitivity to small-to-moderate size effects from treatment toxicity and appear to appropriately discern differences in patient-reported outcomes. Both the HF-QOL daily activity and symptom scale scores displayed a gradient of response across an increasing concurrent NCI-CTCAE level and showed responsiveness to changes in the NCI-CTCAE grade over the 9-week period of observation. The estimated MCIDs were approximately 5 units for the daily activity and 9 units for symptom mean, scaled from 0 to 100. A multiple group analysis was conducted using procedure CALIS (SAS Institute, Inc.) comparing parameters between the treatment and control groups when the first factor daily activity model was fit separately to each group. Although significant differences between factor loadings for some items were found, the effect on the composite reliability of the daily activity scale was minor. In the control group, the composite reliability was estimated at 0.90; in the treatment group, the composite reliability was estimated at 0.88.

Finally, because SOLTI-0701 included two treatment groups (sorafenib plus capecitabine vs. placebo plus capecitabine), we examined whether the validity of the HF-QoL could be influenced by the treatment modality, such as through the side effect profile or intensity. Correlations between the FACT-B and FACT-G subscales and the HF-QoL daily activity and symptom scores did not vary by treatment group (p > .05). No differential effects of grade on the daily activity and symptom scores by treatment were found when tested by including interactions between treatment and grade in the regression models (p > .05). Nor did responsiveness to change from grade 0 (week 3) to 1 (week 6) vary significantly by treatment. The absence of a differential was used as one justification for combining the treatment and control groups.

The strengths of the present study included an iterative development process of the HF-QoL, with input from both patients and clinicians with experience of the condition, supporting the instrument’s content validity. The examination of psychometric properties was undertaken within a randomized clinical trial in which an increase in HFS/R occurred after the start of antineoplastic therapy, consistent with the known timelines for clinical expression of skin toxicity after therapy initiation. Clinical validity of the HF-QoL was obtained from its associations with adverse event reporting via the NCI-CTCAE. Although the NCI-CTCAE reports were made from the physician’s perspective and thus might not fully capture the patient’s HRQL symptom burden, statistically significant linear associations were observed between the NCI-CTCAE grade and the HF-QoL daily activity and symptom scores. Although a strong gradient between the HF-QoL scores and the NCI-CTCAE grade was present, in terms of the response to change, it appears that the HF-QoL daily activity scale was more sensitive to a 1-unit decrease in the NCI-CTCAE level than to an increase (worsened toxicity). In this regard, the mean change was −4.55 versus +1.49, respectively. This asymmetry might have resulted because most of the transitions to an increased CTCAE grade were from 0 to 1, signifying a state from no appreciable symptoms to “minimal skin changes or dermatitis (e.g., erythema) without pain,” which might be mild in terms of disruptions to daily functioning. The same change was associated with an increase in the symptom mean score of 7 points, suggesting a more noticeable appearance of skin changes. Structurally, any improvements in CTCAE toxicity during the study period would have had to come from previously higher levels, including “pain interfering with function.” No associations were found between the FACT-B and FACT-G scores and toxicity grade, highlighting the inability of these scales to capture the effects of HFS/R, with the HF-QoL results more clearly differentiating the toxicity grade for skin rash.

However, a substantial amount of HF-QoL score variability was present within the toxicity grades, undermining the notion of a clear action threshold for patient support or dose modification using the toxicity grade alone. Indeed, the PRO measures are valued as an adjunct to physician assessment because of the complexity side effects might have on an individual’s daily life that cannot be “objectively” assessed in toxicity reports. In particular, for some symptoms, such as HFS/R itch or pain, the patient report, by definition, must be the primary assessment source. This demonstrates the value of developing and validating PROs to measure such important toxicities: the instruments can indicate the extent of the problem and highlight the need to develop treatments to address adverse events and help alleviate the suffering brought on by antineoplastic therapies [33]. This is highlighted by a recent study that concluded that the costs of the diagnosis and treatment of dermatologic toxicities associated with antineoplastic therapy are large and generally underdescribed, with the HFS/R associated with sorafenib therapy the most costly dermatologic toxicity [34].

Conclusion

We have described the properties of the HF-QoL, which, by reason of its brevity, should be feasible to administer in a variety of settings to evaluate the effectiveness of agents for HFS/R treatment within clinical trials and the effect of these treatments on HFS/R-associated HRQL. The study has provided an estimate of MCID for both the daily activity and the symptom scales of 5 and 8 points, respectively. These values can be used both to plan sample sizes for future studies and to interpret the individual level changes during treatment.

The possible limitations of the present study included its inclusion only of breast cancer patients, with future studies required to determine the HF-QoL properties for other conditions and therapies. Other studies are also needed to confirm the measurement properties of the instrument.

Author Contributions

Conception/Design: Roger T. Anderson, Karen N. Keating, Fabian Camacho

Collection and/or assembly of data: Roger T. Anderson, Helen A. Doll, Fabian Camacho

Data analysis and interpretation: Roger T. Anderson, Karen N. Keating, Helen A. Doll, Fabian Camacho

Manuscript writing: Roger T. Anderson, Karen N. Keating, Helen A. Doll, Fabian Camacho

Final approval of manuscript: Roger T. Anderson, Karen N. Keating, Helen A. Doll, Fabian Camacho

Disclosures

Helen A. Doll: Bayer (RF); Karen N. Keating: Bayer Healthcare Pharmaceuticals (E). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board