Measuring the Symptom Burden of Lung Cancer: The Validity and Utility of the Lung Cancer Module of the M. D. Anderson Symptom Inventory

A study was conducted to establish the psychometric properties of a module of the M. D. Anderson Symptom Inventory developed specifically for patients with lung cancer (MDASI-LC). The MDASI-LC includes the 13 core MDASI symptoms of cancer and three lung cancer–specific items: coughing, constipation, and sore throat. The MDASI-LC module is a valid, reliable, and sensitive instrument for assessing the severity of symptoms of lung cancer and their interference in patients' daily functioning.

Abstract

We conducted a study to establish the psychometric properties of a module of the M. D. Anderson Symptom Inventory (MDASI) developed specifically for patients with lung cancer (MDASI-LC). The MDASI measures 13 common “core” symptoms of cancer and its treatment. The MDASI-LC includes the 13 core MDASI symptom items and three lung cancer–specific items: coughing, constipation, and sore throat. MDASI-LC items were administered to three cohorts of patients with lung cancer undergoing either chemotherapy or chemoradiotherapy. Known-group validity and criterion (concurrent) validity of the MDASI-LC were evaluated using the Eastern Cooperative Oncology Group performance status and the 12-item Short-Form Health Survey. The internal consistency and test-retest reliability of the module were adequate, with Cronbach coefficient α-values of 0.83 or higher for all module items and subscales. The sensitivity of the MDASI-LC to changes in patient performance status (disease progression) and to continuing cancer treatment (effects of treatment) was established. Cognitive debriefing of a subset of participants provided evidence for content validity and indicated that the MDASI core items and three additional lung cancer–specific items were clear, relevant to patients, and easy to understand; only two patients suggested additional symptom items. As expected, the item “sore throat” was sensitive only for patients receiving chemoradiotherapy. The MDASI-LC is a valid, reliable, and sensitive symptom-assessment instrument whose use can enhance clinical studies of symptom status in patients with lung cancer and epidemiological and prevalence studies of symptom severity across various cancer types.

Introduction

Patients with cancer experience multiple symptoms simultaneously. Both the cancer itself and its treatment contribute to these symptoms, creating a burden that greatly distresses patients and impairs their functional capacity [1]. Lung cancer is notorious for the prevalence and severity of the symptoms it produces. For patients with advanced lung cancer, symptom alleviation is often one of several primary aims of continued treatment; however, once the disease reaches its end stages, symptom control becomes a critical component of disease management. Therefore, measures that reflect the symptom status of patients with lung cancer are necessary to improve care [2], facilitate individual treatment decisions [3], and evaluate the efficacy of emerging cancer treatments [4]; they are also useful for predicting survival [5–7].

Symptom assessment requires psychometrically validated tools that are easy to use and quick to administer. Several instruments have been developed for use with lung cancer patients—e.g., the Functional Assessment of Cancer Therapy for lung (FACT-Lung) [8] and the European Organisation for the Research and Treatment in Cancer lung module (EORTC QLQ-LC13) [9]. However, these instruments were designed within the conceptual framework of health-related quality of life, and although many quality-of-life measures include symptom items, they are not necessarily symptom instruments per se. Therefore, the recent systematic review of cancer symptom-assessment instruments by Kirkova et al. [10] did not include either the FACT-Lung [8] or the EORTC QLQ-LC13 [9]. Another instrument developed specifically for lung cancer, the Lung Cancer Symptom Scale (LCSS) [11], was not intended for use in studies that include patients with other cancers.

The M. D. Anderson Symptom Inventory (MDASI) [12] was designed to assess the severity of common cancer-related and treatment-related symptoms that may better reflect the symptom experience of the cancer population. Kirkova et al. [10] identified several additional advantages of the MDASI over other measures. (1) The MDASI's 13 “core” symptoms are experienced by most cancer patients, suggesting that the MDASI is comprehensive, yet brief enough to avoid being a burden to answer. (2) The MDASI assesses not only the intensity of cancer-related symptoms but also the level of symptom interference with daily functioning. (3) The instrument's 0–10 numerical scale response-option format is readily understood even by less-educated patients, easy to translate into other languages, and readily adaptable for telephone, computer, and other electronic forms of administration.

Symptoms specific to a particular cancer, treatment method, or treatment site can be added to the core MDASI. These “MDASI modules” include the 13 symptom and 6 interference items of the core MDASI, augmented by additional disease-specific or treatment-specific symptom items. MDASI modules have been developed for patients with brain tumors [13], thyroid cancer [14], head and neck cancer [15], gastrointestinal cancer [16], and treatment-related heart failure [17]. The number of additional module-specific symptom items is minimized to keep the MDASI concise and easy to use in clinical and clinical-research settings and to facilitate repeated measurement.

The goal of our study was to develop, cognitively debrief, demonstrate the sensitivity of, and establish the validity of a MDASI module specific to lung cancer (MDASI-LC; see supplemental online Figure 1). This module asks patients to describe symptoms related to cancer in general and to rate symptoms related specifically to lung cancer and its treatment.

Materials and Methods

Study Participants

Three patient cohorts were used to evaluate the psychometric properties of the new MDASI-LC module. Each cohort rated both the 13 core MDASI items and the 3 lung cancer–specific symptoms (coughing, constipation, sore throat), as described below. This study was approved by the Institutional Review Board of The University of Texas MD Anderson Cancer Center in Houston, Texas.

Cohort 1 comprised patients with advanced lung cancer being treated with chemotherapy. The lung cancer–specific symptom items rated by these patients were coughing and constipation; the group was not asked about sore throat because no radiotherapy-related esophagitis was expected. Data collected before chemotherapy began (baseline) and at the end of two cycles of chemotherapy were used in our overall analysis and to test the MDASI-LC's sensitivity to change in performance status.

Cohort 2 comprised patients with locally advanced lung cancer who were undergoing chemoradiotherapy. The lung cancer–specific symptoms these patients rated included coughing and sore throat. This group was not asked about constipation because very few patients were expected to be taking strong opioids to manage disease-related pain. Symptom data at baseline (before treatment began) and at the end of six weeks of therapy were used in our overall analysis and to provide evidence of the sensitivity of the MDASI-LC to treatment effects. Data from cohorts 1 and 2 were combined for our analysis of the MDASI-LC's measurement characteristics, including tests of validity and estimates of reliability.

Cohort 3 consisted of patients with early-stage or advanced-stage lung cancer who were receiving chemoradiotherapy. This cohort served several purposes: (1) to cognitively debrief patients with lung cancer about the appropriateness of the MDASI core and module-specific items as a reflection of the symptoms they experienced; (2) to probe the need to include additional symptoms to better reflect their experience; (3) to inquire about the appropriateness of the wording of the items; (4) to assay the difficulty patients had in responding to the scale; and (5) to estimate the short-term (day 1 and day 2) reliability of the measure.

Patients were recruited from the Departments of Medical Oncology and Radiation Oncology at MD Anderson. To be eligible for this study, patients were required to be at least 18 years old, speak English, have a pathological diagnosis of lung cancer, and be scheduled for a new chemotherapy cycle (cohort 1) or chemoradiotherapy (cohorts 2 and 3) when enrolled. Patients were excluded if clinical research staff felt that they did not understand the intent of the study or could not complete the assessment measures. All patients provided written consent to participate.

Data-Collection Methods and Measures

At the time of patient enrollment, research staff asked study participants to complete self-administered questionnaires, answered questions, and assisted with completion of survey forms as needed.

General Survey Questionnaire

Patient demographic information (e.g., sex, age, marital status, education level, and employment status) was collected during the initial clinic visit using a general survey questionnaire.

Clinician Checklist

A study-specific clinician checklist was used to collect medical background information from hospital records, including treatment, presence of metastases, and cancer diagnosis, location, and staging.

Eastern Cooperative Oncology Group performance status (ECOG PS [18]) was used to estimate disease severity. ECOG PS is a physician-rated measure of functional ability, ranging from 0 (fully active, able to carry on all predisease performance without restriction) to 4 (completely disabled, cannot perform self care, totally confined to bed or chair).

The MDASI

The MDASI asks patients to rate the severity of 13 disease-related and treatment-related symptoms during the past 24 hours [12]. Each symptom (pain, fatigue, nausea, disturbed sleep, emotional distress, shortness of breath, difficulty remembering, lack of appetite, drowsiness, dry mouth, sadness, vomiting, and numbness or tingling) is rated on an 11-point scale ranging from 0 (not present) to 10 (as bad as you can imagine).

Patients also rate the degree to which symptoms interfered with various aspects of life during the past 24 hours. Each interference item (general activity, mood, normal work [including both work outside the home and housework], relations with other people, walking ability, and enjoyment of life) is rated on an 11-point scale ranging from 0 (did not interfere) to 10 (interfered completely). The interference factor can be decomposed into (1) an activity-related interference dimension consisting of the items normal work, general activity, and walking ability and (2) a mood-related interference dimension composed of the item's relations with people, enjoyment of life, and mood [19].

MDASI-LC.

To create the MDASI-LC, we added three symptoms—coughing, constipation, and sore throat—to the core MDASI. Selection of these lung cancer–specific symptom items was based on literature review and clinician input. Coughing was included in all module versions tested (i.e., for all three cohorts). Constipation, a side effect commonly associated with opioid use, was included only for those patients who were expected to be prescribed strong opioids (cohorts 1 and 3). Sore throat was included only for patients receiving chemoradiotherapy (cohorts 2 and 3).

Scoring the MDASI-LC.

The ratings in the MDASI-LC module can be averaged into several subscale scores: mean severity (13 core symptom items plus the lung cancer–specific items), mean core (13 core symptom items only), and mean interference (interference items only). The interference items can further be broken down into mean activity-related interference (work, general activity, and walking ability) and mean mood-related interference (relations with people, enjoyment of life, and mood).

A more sensitive characterization of symptoms for a given cohort may use a subset of the most severe symptoms reported by that group. Symptom items may be used individually or in subsets without summary scoring if specified a priori. Specific symptom items can also be used based on the expected (i.e., prespecified) outcome. For example, we hypothesized that lung cancer patients undergoing chemoradiotherapy would experience worsening sore throat.

MDASI module development is incremental, and we expected that the three additional lung cancer–specific symptom items would make sense only when combined with the 13 core symptom items. For example, the lung cancer–specific item of coughing and the core symptom item of shortness of breath are frequent, and often severe, among lung cancer patients. This line of reasoning further indicates that lung cancer–specific symptom items should be used in concert with core symptom items. Module items that are not relevant to a particular disease or treatment (e.g., patients not receiving radiation therapy would not be expected to have sore throat) may be dropped from the module to minimize floor effect. All 19 core symptom and interference items must be retained.

Twelve-Item Short-Form Health Survey

To evaluate the validity of the MDASI-LC in comparison with an established instrument, we used the Medical Outcomes Study 12-item short-form questionnaire (SF-12) [20]. The SF-12 is a valid and reliable measure of generic health status and quality of life that was designed for use with a broad range of populations with chronic disease. We chose the SF-12 to examine the relationship between impairment of health status and reported interference from symptoms, recognizing that it is a general health status measure. Perceived health status should correlate more with symptom interference than with symptom severity. Thus, we administered the SF-12 to cohort 1, hypothesizing that the MDASI-LC's mood-related interference subscale would be more correlated with the SF-12 mental component score, and the activity-related interference subscale would be more correlated with the SF-12 physical component score. We further hypothesized that the mean MDASI-LC severity and core subscale scores would correlate moderately with SF-12 physical and mental component scores.

Cognitive Debriefing

Cohort 3 underwent cognitive debriefing of the MDASI-LC. Cognitive debriefing is a critical component of instrument development that is used to assess the relevance of instrument items, the ease of responding to the items, and the comprehensibility and clarity of the items. We asked patients whether they were comfortable answering the questions and whether they had suggestions for making the items more comfortable to answer. They were asked whether any of the symptom items seemed repetitive and which, if any, redundant items might be deleted. They were also queried on whether the scoring system (0–10 numeric scale) was easy to use and understand and whether they were comfortable using it. Finally, patients were asked to verify the relevance of the lung cancer–specific items and whether any other important symptoms had not been included in the questionnaire. This process ensured that the MDASI-LC adequately captured salient lung cancer–specific symptoms, whether caused by the disease itself or by treatment, that were not already included in the core MDASI.

Statistical Analysis

All statistical analyses were conducted using Statistical Package for the Social Sciences (SPSS) software version 16.0 [21]. Correlations, means, standard deviations (SDs), ranges, and 95% confidence limits (CLs) were computed for all symptoms and subscales. Statistical significance was set using a two-tailed α-level of 0.05.

Reliability of the MDASI-LC

Reliability refers to the extent to which the items in a scale are measuring the same concept. Cronbach coefficient α-values were computed to estimate the internal consistency reliability of the three MDASI-LC subscales: the core subscale (13 MDASI symptom items), the severity subscale (core plus lung-specific items), and the interference subscale (six interference items). The criterion for good internal consistency (reliability) requires a Cronbach α-value of 0.70 or higher [22]. We used cohort 3 data from assessments made on consecutive days to evaluate test-retest reliability, calculated using intraclass correlations for the three MDASI-LC subscales.

Validity of the MDASI-LC

Content Validity.

Content validity assesses whether the items in a questionnaire adequately represent the construct of interest. Cognitive debriefing results provide evidence for the relevance of the symptom items for the patient's disease and treatment conditions. Responses are elicited via patient interview.

Criterion Validity.

Criterion, or concurrent, validity refers to the extent to which an instrument is related to another instrument that measures a similar, but not the same, concept. To show concurrent validity, we correlated MDASI subscale scores and items with the SF-12 physical and mental component scores.

Construct Validity.

Construct validation requires demonstrating that the instrument measures the underlying construct it is intended to measure. Various methods of establishing construct validity can be used, such as differentiation between groups (known-group validity), factor analysis, and multitrait–multimethod matrix. For this paper, data from cohort 1 and independent-sample t tests were used to demonstrate known-group validity, which refers to the extent to which an instrument can distinguish between groups known to be clinically different.

Sensitivity of the MDASI-LC

Sensitivity is defined as the ability of the instrument to detect change in outcome using the instrument's subscales or items when such change is expected. We conducted several tests to assess the MDASI-LC's sensitivity to changes in performance status and to the insult of treatment. Effect sizes were calculated to estimate the magnitude of differences in subscale scores and items [23, 24].

First, we evaluated whether the MDASI-LC module could detect a worsening of symptoms among patients with deteriorating performance status. Using our cohort 1 data, we examined whether the MDASI-LC could detect changes in patients whose performance status changed during the course of their disease—particularly, whether symptom severity increased for patients whose performance status deteriorated over time. We also expected patients whose performance status improved to show improved symptom scores. Change scores for MDASI-LC subscales and individual items were computed and were considered to be clinically meaningful at 0.5 SD or higher, the level often used in distribution-based methods of determining meaningful differences [25]. Second, we evaluated whether the MDASI-LC could detect a worsening of symptoms among patients with continued, significant insult due to treatment. We tested the cohort 2 data to demonstrate worsening symptoms during ∼6 weeks of treatment, beginning at the start of chemoradiotherapy (baseline). We used paired t tests to examine increases in the three subscales and in the individual items during treatment. We hypothesized that pain, fatigue, nausea, lack of appetite, drowsiness, vomiting, and sore throat would worsen over time in this group of patients.

Results

Demographic and Clinical Characteristics of Study Cohorts

Demographic and clinical characteristics of the three patient cohorts are summarized in Table 1. Cohort 1 comprised 177 patients with advanced lung cancer, cohort 2 comprised 62 patients with locally advanced lung cancer, and cohort 3 comprised 20 patients with early-stage or advanced-stage lung cancer. Cohort 1 received chemotherapy only; cohorts 2 and 3 underwent concurrent chemotherapy and radiotherapy.

Table 1.

Demographic and clinical characteristics of study cohorts (n = 259)

Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory.

Overall, the average age was 61 years and the mean education level was 13 years. There were more men than women, and the sample was predominantly non-Hispanic Caucasian. Most patients had good performance status. Approximately 18% had stage I or II disease.

Results of Cognitive Debriefing

The 20 patients who underwent cognitive debriefing completed the MDASI-LC in ∼2 minutes, on average. All 20 participants reported that the questionnaire was easy to complete, easy to understand, and not repetitive. They were very comfortable answering the questions and had no problems with the understandability, readability, or number of questions asked. Two patients (10% of the cognitive debriefing cohort) suggested diarrhea as a possible additional symptom. No other additional lung cancer or treatment symptoms were suggested in open interviews with patients. All patients found the 0–10 numeric scale easy to use and understand and were comfortable using it.

Pretreatment Symptom Severity

For cohort 1, mean pretreatment subscale scores for the core, severity, and interference subscales were 2.50, 2.94, and 2.98, respectively. The most severe core symptoms reported were fatigue, pain, disturbed sleep, and drowsiness; nausea and vomiting were the least severe core symptoms reported. Coughing was the most severe lung cancer–specific item reported. Our previous studies have shown that a rating of 5 or greater (on a 0–10 numeric rating scale) for pain [19] and for fatigue [26] indicates a moderate to severe symptom that significantly impairs daily functioning. Using this cutpoint, we found that 40% and 46% of patients reported moderate to severe pain and fatigue, respectively. Applying the same cutpoint to other symptoms reported by this cohort, we found that 36%, 32%, and 30% of patients were experiencing moderate to severe disturbed sleep, drowsiness, and coughing, respectively. In contrast, 10% and 6% of patients reported moderate to severe nausea and vomiting, respectively.

For Cohort 2, mean pretreatment subscale scores for the core, severity, and interference subscales were 1.69, 1.64, and 2.24, respectively. Fatigue, drowsiness, pain, and lack of appetite were the most severe core symptoms with 34%, 15%, 21%, and 17% of patients, respectively, reporting them to be moderate to severe. Sore throat was moderate to severe before treatment for 2% of this cohort.

Symptom Severity During Treatment

For cohort 1, mean scores for the core, severity, and interference subscales during treatment were 2.54, 2.61, and 3.11, respectively. The percentages of cohort 1 patients reporting moderate to severe symptom levels were 44% for fatigue, 32% for pain, 32% for drowsiness, 30% for lack of appetite, and 24% for disturbed sleep. Coughing continued to be the most severe lung cancer–specific item, with 24% of the cohort reporting moderate to severe levels.

For cohort 2, mean scores for the core, severity, and interference subscales after 6 weeks of chemoradiotherapy were 2.45, 2.51, and 3.24, respectively. Fatigue, pain, drowsiness, and lack of appetite were at least 1.0 point higher than their baseline values. The mean value for sore throat during treatment was at least 2.5 points higher than at baseline. The percentages of patients reporting moderate to severe symptoms levels were 33% for fatigue, 16% for pain, 19% for drowsiness, 23% for lack of appetite, and 9% for sore throat.

Validation of the MDASI-LC

Reliability

Internal Consistency.

The MDASI-LC subscales showed good internal consistency reliability (Table 2). Pretreatment Cronbach coefficient α-values were 0.89 for the core subscale, at least 0.88 for the severity subscale, and 0.90 for the interference subscale. All subscale scores were 0.91 and higher during treatment.

Table 2.

Internal consistency reliability of the MDASI-LC before and during treatment

^aBefore patients started chemotherapy and/or radiotherapy.

^bAfter 2 cycles of chemotherapy and/or 6 weeks of radiotherapy.

Abbreviation: MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory.

Test-Retest Reliability.

The intraclass correlations of the MDASI core, severity, and interference subscales administered one day apart (n = 20) were also indicative of good test-retest reliability. All values were at least 0.83 (Table 3).

Table 3.

Test-retest reliability of the MDASI-LC administered 1 day apart in a subset of patients with lung cancer (cohort 3, n = 20)

Abbreviation: MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory.

Validity

Content Validity.

Results of cognitive debriefing provided evidence of content validity. Only one additional symptom, diarrhea, was suggested by two patients.

Criterion (Concurrent) Validity.

Our analysis of the concurrent validity of MDASI-LC items using data from cohort 1 (n = 177) showed that combinations of items we assessed were correlated with both the physical and mental components of SF-12 scores (p < .001 for all comparisons) (Table 4). As expected, the activity-related items in the interference subscale of the MDASI-LC module were more strongly correlated with the physical than mental component score of the SF-12 instrument. Conversely, the mood-related items in the interference subscale were more strongly correlated with the SF-12 mental component score.

Table 4.

Criterion/concurrent validity of MDASI-LC module items compared with SF-12 scores (cohort 1, n = 177)

^aMean core, average of 13 core symptoms items; mean lung-specific, average of two lung-specific items; mean severity, average of 13 core and two lung-specific items (from cohort 1).

^bStatistically significant at p < .001.

Abbreviations: MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory; SF-12, Medical Outcomes Study 12-item short-form health questionnaire.

Construct (Known-Group) Validity.

Known-group validity comparisons were made for the MDASI-LC subscales relative to ECOG PS scores, which were available for 232 patients in cohorts 1 and 2. The MDASI-LC discriminated between patients with good versus poor performance status: patients with good ECOG PS had significantly lower scores for all three subscales than did patients with poor ECOG PS (all p < .05) (Table 5). Similar results were seen for lung cancer–specific symptoms (p < .05). Effect-size differences were 0.65 and higher, indicating medium to large effect sizes [23, 24].

Table 5.

Known-group validity: Comparison of MDASI-LC symptom and interference subscale scores and ECOG PS scores (cohorts 1 and 2, n = 232 patients for whom ECOG PS data were available)

^aStatistically significant at p < .05.

Abbreviations: DIFF, mean difference; ECOG PS, Eastern Cooperative Oncology Group performance status; LCL, lower 95% confidence limit for the group mean; MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory; UCL, upper 95% confidence limit for the group mean.

Sensitivity

Sensitivity to Change in Performance Status.

Using data from the 73 of 177 patients in cohort 1 (advanced-stage lung cancer treated with chemotherapy) who had ECOG PS data available, we assessed whether the MDASI-LC module could detect symptom changes when performance status changed during the course of treatment. We found that the MDASI core and MDASI core plus coughing and constipation subscales were correlated with change in ECOG PS (Table 6). Change scores between patients whose ECOG PS worsened over time and patients whose ECOG PS remained the same or improved were statistically significant for both subscales. Individual items from the MDASI core, such as pain, fatigue, distress, lack of appetite, sadness, and difficulty remembering, were also correlated with change in ECOG PS. These differences were clinically meaningful as reflected by effect sizes of 0.5 SD and higher.

Table 6.

Sensitivity of the MDASI-LC to change in performance status (cohort 1, n = 75 patients for whom ECOG PS data were available)

^aChange scores were calculated such that positive values indicate higher symptom severity at follow-up.

^bStatistically significant at p < .05.

Abbreviation: MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory.

Sensitivity to the Impact of Cancer Therapy.

Statistically significant increases from baseline in MDASI-LC subscales and items were observed in cohort 2 (n = 62) during chemoradiotherapy (Table 7). The MDASI-LC subscale consisting of core symptom items plus the lung cancer–specific symptom items of coughing and sore throat had a larger effect size than the MDASI core subscale. As we expected, the item “sore throat” worsened significantly during treatment. The differences we observed were clinically important, as indicated by the magnitude of the effect sizes (>0.5 SD) for these symptom items.

Table 7.

Sensitivity of MDASI-LC items to symptom changes at end of treatment (cohort 2, n = 62)

^aStatistically significant at p < .05.

Abbreviation: MDASI-LC, lung cancer module of the M. D. Anderson Symptom Inventory.

Discussion

In this study we tested a lung cancer–specific module of the MDASI in three cohorts encompassing patients with early-stage or advanced-stage lung cancer as well as patients who received chemotherapy alone or combined with radiotherapy. The results provide strong psychometric evidence for the use of the MDASI-LC. The module's severity and interference subscales exhibited high test-retest reliability on consecutive days and acceptable internal consistency reliability. The MDASI-LC subscales were sensitive to changes in performance status (related to disease) and to the impact of therapy, as evidenced by significant correlations between MDASI-LC ratings and both ECOG PS and treatment effects over time. In addition, cognitive debriefing results showed that most participants found that the questionnaire represented the most important symptoms related to lung cancer and was easy to use and understand, thus addressing guidance from the FDA on the development of patient-reported outcomes to be used in support of labeling claims [27].

MDASI modules offer advantages over other symptom-assessment measures. First, data collected with MDASI modules (as opposed to disease-specific symptom scales) can be used to compare symptom prevalence and severity across cancer types, which is necessary for epidemiological studies and clinical trials that may include patients with different types of cancers. By rank ordering the severity of core symptom items from one type of cancer to another, researchers can identify most of the symptoms that are consistently burdensome for patients with cancer and thus can compare “symptom burden” across cancers. For example, the five most severe symptoms reported by our study participants—fatigue, shortness of breath, disturbed sleep, pain, and drowsiness—are core MDASI items that are also used in modules for other cancer types [13–15, 17]. This adaptability may not be possible for an instrument designed exclusively for patients with lung cancer.

Second, because validation is costly and requires time and effort, each validation of a MDASI module provides incremental evidence for the validity, reliability, and sensitivity of the original instrument, which contains a set of “legacy” (core) items that are subjected to cognitive debriefing by increasing numbers of patients and are repeatedly tested for sensitivity. The core items can be included in new modules with fewer of the expected psychometric steps used in instrument development, with the exception of the demonstration that the core items are relevant and sensitive to change in the target patient group.

Finally, the MDASI is available in several linguistically and psychometrically validated foreign language versions [28–34]. A study of symptom ratings made by patients with cancer in four countries using four different language versions of the MDASI [35] provided evidence that there are minimal variations in MDASI symptom ratings due to culture and language.

The ability of a patient-reported outcomes instrument to detect change was one of the criteria set forth in the FDA's guidance on the use of patient-reported outcomes in labeling claims [27]. In particular, the regulatory agency is interested to see whether changes in the scores are related to changes in the patient's state. We have shown here that the MDASI-LC subscales are sensitive to changes in performance status (related to disease) and to the impact of therapy. Furthermore, a recent study demonstrated that symptom items from the MDASI, such as coughing, fatigue, and shortness of breath, were predictive of survival in lung cancer [5]. Another study showed that fatigue, distress, and sadness were significant predictors of how symptoms interfered with daily functioning [36]. These studies further establish the ability of the MDASI symptom items to detect change in the patient's state.

Our study had limitations. First, the lung cancer–specific symptom items were not the same in each of the three patient cohorts. Instead, we used symptom sets that we refer to as an “experimental pool” of lung cancer–specific items in which cognitive debriefing of patients was favorable toward the current symptom set. Two patients in the cognitive debriefing cohort suggested the addition of diarrhea to the MDASI-LC. If diarrhea caused by newer targeted-therapy agents becomes an increasing problem in this population, it may be reasonable to include it in a future revision. Second, most of our study participants were drawn from a comprehensive cancer center and therefore may or may not be representative of lung cancer patients in general.

Our study also had strengths. We used three cohorts of patients with early-stage or advanced lung cancer and receiving chemotherapy alone or in combination with radiotherapy, thus expanding the generalizability of the psychometric properties of the MDASI-LC module. Finally, unlike other lung cancer assessment tools such as the FACT-Lung and EORTC-QLQ-LC13, the MDASI-LC is based on the concept of symptom burden rather than health-related quality of life. Measures of symptom burden such as the MDASI-LC may be sufficient to let stakeholders such as patients, clinicians, and regulators make informed decisions about evaluating new cancer therapies [1].

Conclusion

The MDASI-LC module is a valid, reliable, and sensitive instrument for assessing the severity of symptoms of lung cancer and their interference in patients' daily functioning.

Author Contributions

Conception/Design: Tito R. Mendoza, Xin Shelley Wang, Charles Lu, Guadalupe R. Palos, Zhongxing Liao, Charles S. Cleeland

Provision of study material or patients: Xin Shelley Wang, Charles Lu, Zhongxing Liao, Charles S. Cleeland

Collection and/or assembly of data: Tito R. Mendoza, Xin Shelley Wang, Guadalupe R. Palos, Gary M. Mobley, Shitij Kapoor

Data analysis and interpretation: Tito R. Mendoza, Xin Shelley Wang, Gary M. Mobley, Shitij Kapoor, Charles S. Cleeland

Manuscript writing: Tito R. Mendoza, Xin Shelley Wang, Charles S. Cleeland

Final approval of manuscript: Tito R. Mendoza, Xin Shelley Wang, Charles S. Cleeland