Dr D. A. Revicki, Center for Health Outcomes Research, MEDTAP International, 7101 Wisconsin Ave., Suite 600, Bethesda, MD 20814, USA. E-mail: Revicki@medtap.com
Background : Patient-based symptom assessments are necessary to evaluate the effectiveness of medical treatments for gastroparesis.
Aim : To summarize the development and measurement qualities of the Gastroparesis Cardinal Symptom Index (GCSI), a new measure of gastroparesis-related symptoms.
Methods : The GCSI was based on reviews of the medical literature, clinician interviews and patient focus groups. The measurement qualities (i.e. reliability, validity) of the GCSI were examined in 169 gastroparesis patients. Patients were recruited from seven clinical centres in the USA to participate in this observational study. Patients completed the GCSI, SF-36 Health Survey and disability day questions at a baseline visit and again after 8 weeks. Clinicians independently rated the severity of the patients' symptoms, and both clinicians and patients rated the change in gastroparesis-related symptoms over the 8-week study.
Results: The GCSI consists of three sub-scales: post-prandial fullness/early satiety, nausea/vomiting and bloating. The internal consistency reliability was 0.84 and the test–re-test reliability was 0.76 for the GCSI total score. Significant relationships were observed between the clinician-assessed symptom severity and the GCSI total score, and significant associations were found between the GCSI scores and SF-36 physical and mental component summary scores and restricted activity and bed disability days. Patients with greater symptom severity, as rated by clinicians, reported greater symptom severity on the GCSI. The GCSI total scores were responsive to changes in overall gastroparesis symptoms as assessed by clinicians (P = 0.0002) and patients (P = 0.002).
Conclusion: The findings of this study indicate that the GCSI is a reliable and valid instrument for measuring the symptom severity in patients with gastroparesis.
Gastroparesis is a chronic gastrointestinal disorder defined by delayed gastric emptying of food solids with or without liquids. Gastroparesis is suspected on the basis of clinical symptoms, such as post-prandial fullness, early satiety, bloating, nausea and vomiting, and can be confirmed by a significant delayed gastric emptying test.1,2 Symptoms suggestive of gastroparesis are present in approximately 7–15% of the population.3 The aetiology of gastroparesis is variable, with about one-third of patients having idiopathic gastroparesis, 24% with gastroparesis due to diabetes mellitus and 19% with a post-surgical aetiology.4 The prevalence of gastroparesis is increased in diabetes, and 30–50% of patients with type 1 or type 2 diabetes have been reported to have gastroparesis.5
Symptom severity and disease-specific health-related quality of life outcomes are important for the evaluation of the effectiveness of treatments for gastrointestinal disorders.6–9 Patient-reported symptom severity and health-related quality of life measures are important for the evaluation of functional gastrointestinal disorders, as they are the only measures that directly reflect the patients' experience of symptom severity, functioning and well-being. Clinicians and investigators rely on patient reports of their symptoms to manage functional gastrointestinal disorders and to monitor the effectiveness of treatment. Although clinical tests for gastroparesis are useful for diagnosis, they are of limited value for the evaluation of treatment outcomes as they are invasive, time consuming and may not always be closely correlated with patient symptom reports.
There is a need to develop psychometrically sound, patient-rated symptom measures to evaluate the effectiveness of therapies for gastroparesis. These measures, if properly developed, will prove useful as clinical end-points in randomized clinical trials of new medical treatments. Several symptom scales have been developed to evaluate outcomes in gastroparesis.10–12 Only limited information is available on the psychometric qualities of these instruments,9 and most rely on clinician assessments of symptoms elicited from patients. Previous clinical trials comparing treatments for gastroparesis have employed clinical symptom scales assembled to cover the key symptoms associated with gastroparesis.4,11–16 Frequently, these clinical assessment measures have not been evaluated for reliability and validity; thus, it is unknown whether they can be justified as clinical end-points.
This study evaluated the psychometric characteristics of a recently developed patient-rated symptom severity scale, the Gastroparesis Cardinal Symptom Index (GCSI), for use in gastroparesis clinical studies. The GCSI was developed as part of a larger patient outcomes project for the development of the Patient Assessment of Upper Gastrointestinal Disorders-Symptom Severity Index (PAGI-SYM).17,18 The current study involved 169 patients diagnosed with gastroparesis recruited from seven clinical centres, and examined the reliability and validity of the GCSI.
A sample of 169 patients diagnosed with gastroparesis was identified and recruited from seven university-based clinical centres in the USA. The clinical centres involved in the study included Johns Hopkins University (n = 61), University of Kansas (n = 35), Cleveland Clinic Foundation (n = 29), Joslin Diabetes Center (n = 21), University of North Carolina (n = 12), University of Missouri (n = 7) and Washington Hospital Center/Medstar Research Center (n = 4). Eligible study subjects were identified from the existing patient population receiving medical care within each participating clinical centre by clinical site co-ordinators based on the inclusion/exclusion criteria. Subjects were eligible if they had a documented diagnosis of gastroparesis with symptoms for at least 6 months, were aged 18 years or over, were able to read and understand English and were willing to give consent to participate in the study. Patients were excluded if they had a history of gastric surgery or cancer of the gastrointestinal tract or a history of psychiatric disorders or cognitive impairment that would interfere with participation in the study. Over 75% of the eligible patients who were contacted agreed to participate in the study. The primary reason (19%) for non-participation was travel inconvenience associated with attending the 8-week study visit. The patients' current treatment regimens were continued as clinically prescribed and were not modified for participation in this study, as it was non-interventional; physicians determined the treatment for the clinical management of the patients' gastroparesis symptoms.
Diagnosis of gastroparesis
The diagnosis of gastroparesis was based on established clinical criteria2 and gastric emptying tests performed within each clinical centre. Documentation of the diagnosis of gastroparesis in the medical record was required (e.g. gastroparesis, gastric motility disorder, delayed gastric emptying). For 66.3% of the sample, there was direct evidence of delayed gastric emptying tests performed within the participating clinical centre. For 41 patients (24.3%), the gastric emptying tests were performed by the referring clinical centre and there was evidence within the medical record of abnormal gastric emptying test results. For seven patients (4.1%), the diagnosis of gastroparesis was based on endoscopy, symptoms and clinical history and, for the remaining nine patients (5.3%), gastroparesis was secondary to diabetes and no gastric emptying tests were performed. Therefore, evidence of delayed gastric emptying was documented in 90% of the sample patients. The remaining patients had evidence in their medical records confirming a diagnosis of gastroparesis. Subjects with recent gastric emptying times were compared with those without, and no significant differences in GCSI scores were found between the two groups.
Development of the GCSI
The GCSI is a subset of items from the longer PAGI-SYM instrument, which was developed to assess the severity of relevant gastrointestinal symptoms associated with upper gastrointestinal disorders.17,18 The development of the PAGI-SYM instrument followed the standard instrument development process: content and items were based on reviews of the medical literature,9 examination of previously developed gastrointestinal symptom scales and interviews with clinicians and patient focus groups. The GCSI consists of three sub-scales of the PAGI-SYM instrument, selected to measure important symptoms related to gastroparesis: nausea/vomiting (three items), post-prandial fullness/early satiety (four items) and bloating (two items). The nine-item GCSI is reproduced in the Appendix. The rationale for the selection of these three sets of gastrointestinal symptoms was based on reviews of the gastroparesis-related medical literature and the recommendations of an expert panel of gastroenterologists. A six-point Likert response scale, ranging from 0 (none) to 5 (very severe), with a 2-week recall period, was used to rate the severity of each symptom. The GCSI total score was constructed as the average of the three symptom sub-scales. GCSI total scores were in the range 0–5, with higher scores reflecting greater symptom severity.
Clinical and other patient outcome measures
Information was collected on the demographic and clinical characteristics of the patients at the first visit (baseline). In addition, data were collected on disability days and health-related quality of life using the SF-36 Health Survey19 at baseline and at the 8-week follow-up assessment. Clinical symptoms were assessed by physicians and an Overall Treatment Evaluation (OTE) rating scale20,21 was completed by physicians and patients at 8 weeks. To evaluate the test–re-test reliability, a randomly selected sub-sample of 30 patients completed the GCSI and OTE at 2 weeks.
Clinical characteristics. Investigators assessed gastroparesis-related symptoms, including nausea, abdominal distension, bloating, early satiety and vomiting, at baseline and week 8. These symptoms were rated on a four-point severity scale ranging from 0 (none) to 3 (severe), and have been used in previous clinical trials in gastroparesis.12,13,15 This clinician-assessed symptom rating scale was used as an independent clinician-based validation measure.
The change in overall gastroparesis-related symptom status was measured using the OTE.20,21 The first question required the physician to indicate whether the patient's gastroparesis-related symptoms had improved, remained the same or worsened since the last evaluation. If the patient's symptoms had improved, the physician rated the degree of improvement on a seven-point scale from ‘Almost the same, hardly better at all’ (1) to ‘A very great deal better’ (7). If the patient's symptoms had worsened, the physician rated the degree of worsening on a seven-point scale from ‘Almost the same, hardly worse at all’ (− 1) to ‘A very great deal worse’ (− 7). The OTE has been used in numerous clinical studies and therapeutic areas.20–24 Patients also rated the change in their gastroparesis-related symptoms using the same OTE questions. The OTE was used to categorize subjects at 2 and 8 weeks on the basis of their overall gastroparesis-related symptoms (improved, stable or worse).
Health-related quality of life. The health-related quality of life was measured using the physical component summary (PCS) and mental component summary (MCS) scores based on the SF-36 Health Survey. The SF-36 is a generic health status instrument developed for application in primary care and chronic disease populations.19,25 The PCS and MCS scores are transformed, on the basis of normative US general population data, to have a mean of 50 and standard deviation of 10; higher scores indicate better health-related quality of life. The SF-36 summary scores have excellent reliability and validity in US general and chronic disease populations,25 including patients with gastroparesis.4,9,12,13,26
Disability days. Study patients completed questions on bed disability days and restricted activity days over the past 3 months. These questions have been used extensively in epidemiological studies.27,28
A psychometric analysis was designed to evaluate the reliability, validity and responsiveness of the GCSI. Internal consistency reliability was assessed using Cronbach's alpha.29,30 Subjects were classified as improved, stable or worse with regard to their gastroparesis-related symptoms using the OTE at 2 and 8 weeks; a clinically stable status was defined as a patient-rated OTE score between − 1 and 1. Test–re-test reliability was evaluated in the sub-sample of clinically stable subjects interviewed at baseline and week 2 using intra-class correlation coefficients.
Construct validity was evaluated by examining the relationship between the GCSI scores and the gastroparesis-related symptoms rated by clinicians, disability days and SF-36 summary scores. Pearson correlation coefficients were used to examine the magnitude and direction of these relationships. Study subjects were categorized into four groups on the basis of clinician-rated gastroparesis symptom severity (none, mild, moderate, severe). Analysis of covariance (ancova) models, adjusting for age and gender, were used to compare mean GCSI scores by gastroparesis severity group. It was hypothesized that higher (more severe) GCSI scores would be associated with clinician symptom ratings, lower PCS and MCS scores and a greater number of disability days. Responsiveness to changes in clinical status was evaluated using the clinician-rated and patient-rated OTE scales and observed changes in GCSI scores. Patients were classified as improved, stable or worse with regard to their gastroparesis-related symptoms based on the OTE at 8 weeks. ancova models, adjusting for gender and age, were used to compare the mean GCSI scores by a change in clinical status.
A total of 169 patients diagnosed with gastroparesis participated in this study. The patient characteristics are summarized in Table 1. Seventy-seven per cent were women, 49% were diabetics and the average age was 45.7 years (range, 18–78 years). The mean baseline PCS score was 33.9 (s.d., 10.6) in the gastroparesis sample, 16.1 points lower than the US general population norm (mean, 50.0; s.d., 10.0). The mean baseline MCS score was 40.8 (s.d., 11.4), 9.2 points lower than the US general population norm (mean, 50.0; s.d., 10.0).
Table 1. Demographic characteristics of gastroparesis subjects (n = 169)
Age (years), mean (s.d.)
Living situation (%)
Educational level (%)
Work status (%)
Working (full- or part-time)
Student (full- or part-time)
High blood pressure
The mean GCSI total score at baseline was 2.56 (s.d., 1.05) and, after 8 weeks, 2.26 (s.d., 1.1). At baseline, the mean GCSI post-prandial fullness/early satiety score was 3.01 (s.d., 1.24), bloating score 2.69 (s.d., 1.51) and nausea/vomiting score 1.88 (s.d., 1.50). The GCSI total and sub-scale scores were in the range 0–5, demonstrating that the full range of possible scores were expressed in the study sample.
The internal consistency reliability for the GCSI total score was 0.84 at baseline and 0.89 at the 8-week follow-up visit. Internal consistency reliabilities for the three GCSI sub-scales ranged from 0.83 to 0.85 (see Table 2). The test–re-test reliability was 0.76 for the GCSI and ranged from 0.68 to 0.81 for the three sub-scales. In stable subjects, the GCSI total score varied by only 0.02 points over 2 weeks.
Table 2. Internal consistency reliability and 2-week reproducibility of the Gastroparesis Cardinal Symptom Index (GCSI) total and sub-scale scores
The construct validity of the GCSI total scores was evaluated by determining the associations with clinician-rated symptoms, SF-36 summary scores and disability days. Table 3 summarizes the ancovas comparing the mean GCSI total scores by the physician-rated severity of gastroparesis symptoms. The mean GCSI scores varied significantly by severity of bloating (P < 0.0001), abdominal distension (P < 0.0001), early satiety (P < 0.0001), nausea (P < 0.0001) and vomiting (P < 0.0001). Consistently, across all symptoms measured, there were higher (more severe) mean GCSI scores associated with worse clinician ratings of symptom severity (see Table 4).
Table 3. Mean Gastroparesis Cardinal Symptom Index (GCSI) total scores (standard deviation in parentheses) by clinician-rated symptoms
All listed contrasts are statistically significant at P < 0.05.
PCS and MCS scores and bed disability and restricted activity days were correlated with GCSI scores at both baseline and 8 weeks (see Table 4). The mean GCSI total scores varied significantly by the frequency of bed disability days reported during the previous 3 months (P < 0.0001; see Figure 1). Gastroparesis patients reporting greater than or equal to 14 bed disability days had a mean GCSI score of 3.2, compared with a mean GCSI score of 2.0 for patients with no reported bed disability days (P < 0.0001), and means ranging from 2.6 to 2.7 for those with more than one but less than 14 bed disability days. The mean GCSI scores varied significantly by the frequency of restricted activity days (P < 0.0002; see Figure 2). The mean GCSI score was 1.9 for patients reporting no restricted activity days; this was 0.4 points lower (P < 0.05) than the mean GCSI score for those reporting 1–4 restricted activity days, 0.6 points lower (P < 0.005) than for those reporting 5–13 restricted activity days and 1.2 points lower (P < 0.0001) than for those reporting ≥ 14 restricted activity days.
The responsiveness to changes in the clinical symptom status was evaluated by assessing the relationship between the clinician and patient ratings of changes in overall gastrointestinal symptoms during the period between baseline and the 8-week follow-up assessment. The mean baseline to 8-week change in the GCSI total score by the change in clinical status, based on clinician ratings, is summarized in Figure 3. There were statistically significant differences in GCSI change scores between the improved and stable/worsened group (P < 0.0001). The improved group demonstrated a 0.75 point decrease in the mean GCSI score, whereas the stable/worsened group showed a change of − 0.06 points. Statistically significant differences in GCSI change scores were also found between the patient-rated clinical status groups (P = 0.0004) (see Figure 4). The improved group demonstrated a 0.74 point decrease in the mean GCSI total score, whereas the stable/worsened group showed a change of − 0.12 points.
There are no patient-reported symptom assessment instruments with good measurement qualities available for use in clinical trials and other studies in gastroparesis. Most gastroparesis clinical studies rely on patient reports of symptoms, either based on physician interviews or patient-completed questionnaires. We developed a patient-based symptom instrument, the GCSI, to assess the severity of gastroparesis-related symptoms for clinical trials. The GCSI total score was based on a combination of nausea/vomiting, post-prandial fullness/early satiety and bloating sub-scales. Although bloating is not commonly included as a symptom of gastroparesis, it is consistent with previous studies. Based on a large US-based population survey, bloating was grouped as a symptom of delayed gastric emptying or gastroparesis.31 Within the current study, the factor analysis results indicated that bloating was an important symptom in this gastroparesis population.
Based on the results of this study, the GCSI is a reliable measure of symptom severity, and we have documented evidence supporting the construct validity and responsiveness of the GCSI in gastroparesis patients. The reliability of the GCSI is acceptable for group level comparisons.29,30 The findings suggest that the GCSI total scores are internally consistent and have excellent reproducibility.
The results show that the mean GCSI score varies significantly by clinician-based assessment of the gastroparesis-related symptoms. Patients rated by clinicians as having more severe symptoms are more likely to report greater symptom severity on the GCSI. Consistent and systematic differences in the mean GCSI score were observed across clinician-rated symptoms differentiating mild, moderate and severe symptom severity.
Patients who are rated as improved over 8 weeks, by either physicians or the patients themselves, demonstrate larger decreases in the mean GCSI score compared with patients who remain unchanged or who have worsened. A decrease (indicating an improvement in symptom severity) of 0.75 points in the mean GCSI total score was seen in patients assessed by physicians as improving in clinical status. These findings were consistently observed for the clinician- and patient-based assessments of the change in overall gastroparesis status.
Patients reporting greater gastroparesis-related symptom severity also report greater numbers of bed disability and restricted activity days. The findings demonstrate that, as the mean GCSI score increases, a greater number of disability days are reported. A similar relationship is observed between the GCSI score and the number of restricted activity days: as the mean GCSI score increases, indicating greater symptom severity, an increasing frequency of restricted activity days is reported.
Statistically significant correlations were also observed between the GCSI scores and the summary scores from a generic health status instrument, the SF-36 Health Survey. There were statistically significant correlations between the GCSI and the PCS and MCS scores, but the largest correlations were seen for the measures of physical functioning. As expected, there were lower, but still significant, correlations between the GCSI and the measures of psychological functioning and well-being. The results indicate that, as the gastroparesis-related symptom severity increases, there is greater impairment in patient functioning and well-being.
Gastroparesis patients in this study reported significantly more impairment of measures of functioning and well-being compared with the US general population. Previous studies have noted decrements in health-related quality of life amongst patients with gastroparesis.4,12,13 The SF-36 summary scores observed in this gastroparesis patient sample were similar to those seen in a study of diabetes-related gastroparesis.13
There are several limitations that should be considered when reviewing these results. First, the findings may only be generalizable to Caucasian females, as few non-Caucasians and males were included in the sample. This is not surprising given that the majority of gastroparesis patients are female. We have no reason to believe that the results of this study are not generalizable to other demographic groups, but there is little literature available on gastroparesis in non-Caucasian populations.
Second, it is possible that social desirability or anxiety symptoms may have influenced the patient-reported scores on the GCSI. However, the findings suggest a significant association between patient-reported and clinician-reported symptoms, and therefore it is believed that these other factors may have had a minimal impact on the patient response to the GCSI items.
Finally, the concurrent measurement of gastric emptying times was not required to document the diagnosis of gastroparesis as part of this instrument development and validation study. The clinical centres participating in the study were very experienced in the diagnosis and treatment of gastroparesis and documentation was required from the medical records; therefore, we believe that these patients were correctly diagnosed with gastroparesis. We were unable to evaluate the relationship between objective gastric emptying test results and GCSI scores in this study; future research should be completed to determine the strength of this relationship. In addition, future research is needed to evaluate the impact of prokinetic therapy on GCSI scores in gastroparesis patients. Given the evidence on the sensitivity to change, based on the study results, it is very likely that the GCSI will be responsive to changes in clinical status over time and to the effectiveness of treatment for gastroparesis. Clearly, clinical trials are needed to demonstrate the responsiveness of this symptom outcome measure in patients with gastroparesis.
In conclusion, the GCSI is a relatively brief symptom severity instrument for gastroparesis with demonstrated reliability and evidence supporting construct validity and responsiveness to change in clinical status. The psychometric evidence on the GCSI meets the recommended guidelines for sufficient support for substantiating claims of clinical benefit for European regulatory agencies32 and the US Food and Drug Administration.33,34 Additional research is needed to confirm these psychometric findings in samples of patients with gastroparesis from other countries. Owing to the systematic attention given to the identification of important gastrointestinal symptoms, careful instrument development and language translation and cultural adaptation in the PAGI-SYM project,17 it is probable that the psychometric qualities summarized in this study will be confirmed. The GCSI may be a useful clinical end-point for clinical trials comparing medical therapies for gastroparesis, and may also be useful in monitoring patient outcomes and quality of care in medical practice.
We would like to thank all participating investigators, study co-ordinators and patients for their participation, and Jordana Schmier for her tireless effort on behalf of the study. Funding for this study was provided by Janssen Pharmaceutica N.V.
Appendix: gastroparesis cardinal symptom index
This questionnaire asks you about the severity of symptoms you may have related to your gastrointestinal problem. There are no right or wrong answers. Please answer each question as accurately as possible.
For each symptom, please circle the number that best describes how severe the symptom has been during the past 2 weeks. If you have not experienced this symptom, circle 0. If the symptom has been very mild, circle 1. If the symptom has been mild, circle 2. If it has been moderate, circle 3. If it has been severe, circle 4. If it has been very severe, circle 5. Please be sure to answer every question.