The COPD Assessment Test (CAT): Short- and Medium-term Response to Pulmonary Rehabilitation

Abstract

Background: The COPD Assessment Test (CAT) is a recently introduced instrument to assess health-related quality of life in COPD. We aimed to evaluate the longitudinal change in CAT following Pulmonary Rehabilitation (PR), and test the relationship between CAT and CRQ-Self Report (SR) over time. We hypothesised that the CAT would show similar responsiveness to PR as the CRQ-SR both in the short and medium-term. Methods: 118 COPD patients completed an eight-week outpatient multidisciplinary PR programme. CAT, CRQ-SR and the incremental shuttle walk (ISW) were measured prior to starting PR (T1), completion of PR (T2) and 6 months after completion of PR (T3). Results: There was a significant improvement in CAT, CRQ-SR and ISW immediately following PR (p < 0.001). Although there was decline between T2 and T3, CAT, CRQ-SR and ISW remained significantly better at T3 compared with T1 (ANOVA p < 0.001). Both between T1-T2 and between T2-T3, change in CAT correlated significantly with change in CRQ (both r  = -0.44 and p < 0.001). The slope of the relationship between CAT change and CRQ-SR change at T1-T2 and T2-T3 was not significantly different (ANCOVA: intercept p  =  0.79, interaction effect p  =  0.95). Conclusions: In COPD, the CAT score is immediately responsive to PR and remains improved at 6 months. There is no significant difference in the short and medium term changes in the CAT and CRQ-SR following PR. We propose that for most clinical indications for assessing health-related quality of life in COPD, the CAT is a robust and practical alternative to longer-established instruments such as the CRQ-SR.

Keywords :

• COPD
• Pulmonary Rehabilitation
• Quality of Life
• Health Status
• COPD Assessment Test (CAT)
• Chronic Respiratory Disease Questionnaire

Introduction

The COPD Assessment Test (CAT) was recently introduced as a simple, short, patient-completed instrument to assess health-related quality of life and symptom burden in patients with chronic obstructive pulmonary disease (COPD) (1). It comprises 8 questions, each scored between 0–5, giving a total score out of 40. Scores of 0–10, 11–20, 21–30 and 31–40 represent mild, moderate, severe or very severe clinical impact. The CAT has good internal consistency and test-retest reliability (1), correlates strongly with the COPD-specific version of the St. George's Respiratory Questionnaire (SGRQ) and is able to distinguish between stable patients and those undergoing an exacerbation (1, 2).

Recently, we demonstrated the short-term responsiveness of the CAT to pulmonary rehabilitation (PR) (3), and showed that changes in CAT with PR correlated significantly with changes in other well-established health status questionnaires, such as the Chronic Respiratory Questionnaire (CRQ) and SGRQ (3). Although the CRQ, SGRQ and other outcome measures such as the incremental shuttle walk (ISW) are responsive to PR, improvements decrease with time and usually return to baseline by 12 months (4).

There are currently no medium- to long-term longitudinal data on the CAT, particularly in the context of response to PR. Although the CAT has distinct advantages over other health-related quality of life questionnaires in terms of simplicity, speed of completion and scoring (5), further data are required to assess the medium and long term properties of the instrument before widespread adoption. The aim of the current study was to evaluate the longitudinal change in CAT following PR, and test the relationship between CAT and CRQ over time. We hypothesised that the CAT would show similar responsiveness to PR as the CRQ-SR both in the short and medium-term.

Methods

187 consecutive COPD patients referred to the Harefield Pulmonary Rehabilitation (PR) service between January and October 2010 participated in the study. The study was approved by the Riverside Research Ethics Committee. We prospectively measured the COPD Assessment Test (CAT) (1), the self-report Chronic Respiratory Questionnaire (CRQ-SR) (6, 7) and the incremental shuttle walk (ISW) (8) at enrolment prior to starting PR (T1), completion of PR (T2) and six months after completion of PR (T3). At the time of measurement, participants were blinded to their previous results.

The PR programme was an eight-week multidisciplinary outpatient exercise and education programme comprising two supervised and at least one additional home session per week. The exercise training was individualised and included a mixture of aerobic and strength training. Initial walking exercise prescription was based on the outcome of the ISW and speed set at 80% of predicted peak oxygen consumption (9), whilst initial cycling prescription was based on symptom scores. Workloads and duration of exercise were increased through the programme as tolerated. The education classes covered issues including exercise, medication use, diet, coping strategies, increasing physical activity and recognising and managing infections, with a particular emphasis on self-management. On completion of the PR course, patients were offered individualised exercise advice and goal-setting, but no formal supervised training sessions.

To test our hypothesis, we examined whether the relationships between change in CAT and change in CRQ-SR between T2 and T3 and between T1 and T2 were significantly different. We also tested whether a significant difference existed between change in CAT, CRQ-SR and ISW at completion of PR (T2) and 6 months post-PR (T3).

Statistical analyses and graphical presentations were performed using GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA) or PASW version 18 (IBM, USA). Paired t-tests or non-parametric equivalent (Wilcoxon signed rank tests) and repeated measures ANOVA with post-hoc multiple comparison (Tukey's HSD) were used for comparison of outcome measures between time points. Relationships between change in CAT and other outcome measures were calculated using Pearson's correlation coefficient. ANCOVA (interaction effect) was used to compare the slopes of the relationship between change in CAT and CRQ-SR at T1-T2 and T2-T3. Data are presented as mean (standard deviation) or median (interquartile range) if non-parametric unless otherwise stated.

Results

Of the 187 patients, 35 did not begin PR, and a further 30 dropped out from PR and did not return for assessment at T2. 4 patients did not return for assessment at T3. Mean baseline CAT score for these 69 patients was 23.3 and not significantly different from the participants completing the study (p  =  0.49). The baseline characteristics of the remaining 118 patients are presented in Table 1. Four patients did not complete an ISW at T3 due to recent lower limb musculoskeletal injury (n  =  2) or joint surgery (n  =  2).

Table 1.  Baseline characteristics

Age (years)	73 (8)
FEV1%	47 (33, 67)
Gender (M:F)	60:58
Body Mass Index (kg.m-2)	28.5 (23.9, 32.0)
ISW (metres)	130 (50, 250)
CRQ Dyspnoea	13 (8, 16)
CRQ Fatigue	13 (10, 17)
CRQ Emotion	31 (25, 38)
CRQ Mastery	20 (14, 23)
CRQ Total	75 (58, 91)
CAT score	22.1 (7.5)

Data expressed as mean (standard deviation) or median (25th, 75th centile).

As expected, there was a significant improvement in CAT, CRQ-SR, all domains of CRQ-SR and ISW immediately following PR (Table 2; fall in CAT and increase in CRQ-SR and ISW represents improvement). Although there was a deterioration in most measures between T2 and T3 (Table 3), repeated measures ANOVA demonstrated that CAT, CRQ-SR and ISW at T3 remained significantly better than baseline (p < 0.001). Univariate correlations between T1-T2 and T2-T3 changes in CAT and changes in CRQ-SR and ISW during the same time period are shown in Table 4.

Table 2.  Response to PR (T2-T1)

	Pre-Pr (T1)	Post-PR (T2)	p-Value
CAT	22.1 (7.5)	19.2 (7.6)	<0.001
CRQ Dyspnoea	13 (8, 16)	16 (11, 22)	<0.001
CRQ Fatigue	13 (10, 17)	16.5 (5.7)	<0.001
CRQ Emotion	31 (25, 38)	34.7 (8.7)	<0.001
CRQ Mastery	20 (14, 23)	22 (16, 25)	<0.001
CRQ Total	75 (58, 91)	88.7 (24.1)	<0.001
ISW, metres	130 (50, 250)	215 (100, 320)	<0.001

Data presented as mean (standard deviation) or median (interquartile range).

Table 3.  Medium-term response to PR (T3-T2)

	Post-PR (T2)	6 months post-PR (T3)	p-Value
CAT	19.2 (7.6)	20.7 (7.6)	0.008
CRQ Dyspnoea	16 (11, 22)	15.5 (6.9)	0.02
CRQ Fatigue	16.5 (5.7)	14.7 (5.0)	<0.001
CRQ Emotion	34.7 (8.7)	32.9 (8.6)	0.001
CRQ Mastery	22 (16, 25)	20 (15, 25)	0.44
CRQ Total	88.7 (24.1)	86 (63, 98)	<0.001
ISW, metres	215 (100, 320)	170 (68, 300)	<0.001

Data presented as mean (standard deviation) or median (interquartile range).

Table 4.  Relationship between delta CAT, delta CRQ and delta ISW between T2-T1 and T3-T2

T2-T1 ΔCAT	r	p-value
ΔCRQ Total	-0.44	<0.001
ΔISW	-0.24	0.01
T3-T2 ΔCAT	r	p-value
ΔCRQ Total	-0.44	<0.001
ΔISW	-0.10	0.31

N = 118 except for ΔISW (114 observations) between T3-T2.

CRQ, self-report Chronic Respiratory Questionnaire; ISW, incremental shuttle walk.

CAT, CRQ-SR (total and individual domain scores), and ISW at T1, T2 and T3 are plotted in Figures 1 and 2, showing graphically that CAT responds in a similar way to CRQ-SR. A formal test of the relationship between CAT and CRQ-SR between T1-T2 and T2-T3 was performed using ANCOVA; the slope and intercept of the relationship between CAT change and CRQ-SR change over these two periods were not significantly different (ANCOVA: intercept p  =  0.79, slope p  =  0.95; Figure 3).

Figure 1.  Mean (95% Confidence Intervals) COPD Assessment test (CAT) scores and median (interquartile range) Chronic Respiratory Disease Questionnaire (CRQ) total and domain scores plotted pre-, immediately post- and 6 months after pulmonary rehabilitation.

Figure 2.  Mean (95% Confidence Intervals) COPD Assessment test (CAT) scores and median (interquartile range) incremental shuttle walk plotted pre-, immediately post- and 6 months after pulmonary rehabilitation.

Figure 3.  Delta CAT versus delta CRQ between T1 and T2 and between T2 and T3. Filled circles represent T1-T2 and open circles represent T2-T3. Solid line represents T1-T2 regression line and dashed line represents T2-T3 regression line. ANCOVA: intercept p = 0.79, slope p = 0.95.

Discussion

The main finding of the present study is that after initial short term response to PR, the CAT declined with time, and was similar to the deterioration in health status recorded by the well-established CRQ-SR (both total and individual domain scores). It is important to note that the relationship between changes in CAT and CRQ-SR were closely matched whether measuring initial response to rehabilitation or the subsequent partial loss. This provides evidence that the CAT score has similar longitudinal properties to the CRQ-SR, and is therefore a practical alternative in evaluating longitudinal change in health status in COPD patients, especially after PR.

Study limitations

This study is the first to demonstrate that medium term longitudinal changes in health status with PR, as well as the immediate responsiveness, can be captured by the CAT. COPD patients were unselected, participating in the clinical PR setting, with data collected prospectively. 118 of the 187 patients (63%) initially referred for PR participated in the final data analysis, reflecting the well-recognised issues of uptake, adherence and completion of PR outside of randomised clinical trials (10).

Despite this, the sample size was relatively large; for example, greater than seen in the intervention arms of major PR randomised controlled trials (4) (11). We acknowledge that data was collected from only one centre. However, the same centre recently participated in a multi-centre study comprising seven PR programmes in the UK, and had equivalent PR results to the other centres (3). As such, the results from this study are likely to be generalizable to routine clinical practice.

Significance of findings

The CRQ-SR and SGRQ are the two best-established respiratory specific health related quality of life questionnaires and the most commonly used in the PR setting. Both have proposed thresholds for clinical significance, and the long term response to PR has been previously described (4,11). In contrast to the CAT, which consists of only 8 questions, the CRQ-SR and the SGRQ consists of 20 and 50 questions respectively, with further time and effort required for the health professional to score the questionnaire.

A recent study in the PR setting showed that the CAT was considerably quicker to complete than the SGRQ (average time 107 seconds versus 578 seconds)(5). Given that staffing time is a major component of the cost for PR, adoption of a simpler and quicker health status questionnaire could lead to a more efficient delivery of the service and subsequent cost savings. This is particularly important given the known resource limitations associated with PR (12).

The CAT score appears to be a suitable candidate instrument—it is significantly less time-consuming than CRQ or SGRQ to complete or score (usually less than 2 minutes)(5), and has been shown previously (and in this study) that it is responsive to PR (3, 13). Improvements in CAT with PR correlate significantly with improvements in CRQ-SR and SGRQ (3, 13). Therefore, for most clinical indications associated with PR in COPD patients (for example, assessing response to intervention, or benchmarking a service), the CAT is a practical alternative to the CRQ-SR or SGRQ.

Another candidate questionnaire is the Clinical COPD Questionnaire (CCQ) (14), which consists of 10 questions and has recently been shown to be responsive to PR in severe COPD patients (5). However, although considerably shorter than the SGRQ, the CCQ does take longer to complete than the CAT (5).

There are some weaknesses associated with the CAT. First, it has only been validated in patients with COPD and there is increasing evidence that PR is useful for patients with chronic respiratory disorders other than COPD. Second, unlike the CRQ-SR and SGRQ, the CAT does not provide multi-dimensional domain scores. Although of interest to the patient and practitioner, from a practical point of view this information is rarely used to change the intervention delivered and is not needed for most clinical purposes. There is no evidence to suggest that knowledge of these domains leads to improvements in patient outcomes or care.

Conclusion

In COPD patients, the CAT score is immediately responsive to PR and remains improved at 6 months after PR. There is no significant difference in the short and medium term changes in the CAT and CRQ-SR following PR. We propose that for most routine clinical indications for assessing health-related quality of life in COPD patients, the CAT is a robust and practical alternative to longer-established instruments such as the CRQ-SR.

Source of Support

WD-CM is supported by a National Institute for Health Research Clinician Scientist Award and a Medical Research Council (UK) New Investigator Research Grant. JLC and SSCK are supported by the Medical Research Council. This project was undertaken at the NIHR Respiratory Biomedical Research Unit at the Royal Brompton and Harefield NHS Foundation Trust and Imperial College London; MIP's salary is part funded by the Biomedical Research Unit. The views expressed in this publication are those of the authors and not necessarily those of the NHS, The National Institute for Health Research nor the Department of Health. PWJ has received an institutional unrestricted research grant from Glaxo Smith Kline (GSK).

Declaration of Interests

The authors alone are responsible for the content and writing of the paper.