The long-term effects of pulmonary rehabilitation in patients with asthma and chronic obstructive...

transcript

REVIEW ARTICLE

The Long-Term Effects of Pulmonary Rehabilitation in Patients With Asthma and Chronic Obstructive Pulmonary Disease: A Research Synthesis Walter Cambach, MSc, Robert C. Wagenaar, PhD, Tim W Koelman, PT, A.R.J. Ton van Keimpema, PhD, Han C.G. Kemp& Phb

ABSTRACT. Cambach W, Wagenaar RC, Koelman TW, van Keimpema ARJ, Kemper HCG. The long-term effects of pulmonary rehabilitation in patients with asthma and chronic obstructive pulmonary disease: a research synthesis. Arch Phys Med Rehabil 1999;80:103-11.

Objective: To present a critical review and meta-analysis of studies evaluating the long-term effects of pulmonary rehabilitation in patients with asthma and chronic obstructive pulmonary disease (COPD).

Data Sources: A database of articles published over the last 45 years, compiled by using medical subject heading key words pulmonary, obstructive, rehabilitation, and exercise. Articles not written in English, Dutch, or German and abstracts were excluded.

Study Selection: Selected studies (1) evaluated the effects of pulmonary rehabilitation, (2) included patients with asthma or COPD older than 18 years, (3) evaluated outcome measures of exercise capacity or health related quality of life (HRQL), and (4) included a control condition lacking exercise training.

Data Extraction: Independent extraction by two reviewers. Data Synthesis: For each outcome, summary effects were

computed by pooling standardized mean differences as well as raw mean differences. Significant improvements were found for all outcomes (p < ,001). Sensitivity analyses for methodological quality of the selected studies did not change summary effect sizes. Effect sizes were significantly heterogeneous for the outcome endurance time (p < .OOOl). Pooling raw mean differences revealed overall effects in 6-minute walking distance (49 2 26m) and all 4 dimensions of the chronic respiratory questionnaire (range, 0.5 + 0.3 to 0.8 i 0.3 points), indicating substantial improvements in these outcomes. Signifi- cant summary effect sizes were found up to 9 months after finishing rehabilitation for maximal exercise capacity (p < ,003) and 6-minute walking distance (p < .005).

Conclusions: Patients with asthma and COPD benefit from pulmonary rehabilitation.

0 1999 by the American Congress of Rehabilitation Medi- cine and the American Academy of Physical Medicine and Rehabilitation

From the Research Center Primary Secondary Health Care, University Hospital Vrije Universiteit, Amsterdam (Mr. Cambach): the Department of Physiotherapy, University Hospital Vrije Universiteit and the Research Institute for Fundamental and Clinical Human Movement Sciences, Amsterdam (Dr. Wagenaar, Mr. Koelman); the Lung and Asthma Centmm, Medical Centmm De Klokkenberg, Breda (Dr. van Keimpema); and the Institute for Research in Extramural Medicine, Medical Faculty, Vrije Universiteit, Amsterdam (Dr. Kemper), The Netherlands.

Submitted for publication April 2, 1998. Accepted in revised form July 14, 1998. No commercial party having a direct financial interest in the results of the research

supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated.

Reprint requests to W. Cambach, Department of Physiotherapy, University Hospital Vrije Universiteit, PO Box 7057, 1007 MB Amsterdam, The Netherlands.

0 1999 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation

0003.9993/99/8001-4948$3.00/O

F ROM THE 1960s ONWARDS, inspired by a number of historical studies on the effects of pulmonary rehabilita-

tion,‘s2 a variety of rehabilitation programs have been devel- oped for patients with asthma and chronic obstructive pulmonary disease (COPD). Because many patients with asthma and COPD may be physically inactive and deconditioned from their impaired lung function and their dyspnea (as well as their fear of dyspnea), the major objectives of such programs are to improve exercise capacity3-5 and health-related quality of life (HRQL) 4.6

In several reviews that have summarized pulmonary rehabilitation programs, it has been concluded that these programs result in improvements in exercise capacity3*4,7-g and HRQL.3 Recent developments in pulmonary rehabilitation concern the implementation of programs in home-based settingslO,ll and physiotherapy practice-based settings.r2 In all of these programs, patients were supervised by general practitioners, nurses, and physiotherapists.

Lacasse and colleaguest3 published a first meta-analysis on the effects of pulmonary rehabilitation, and provided, on the basis of the results of 14 randomized control trials, firm evidence for the acute effects of pulmonary rehabilitation compared with conventional treatment in patients with COPD in terms of maximal exercise capacity, functional exercise capacity, and HRQL. Still, many questions concerning the effects of pulmonary rehabilitation remain unanswered.

In an attempt to address some of these questions, the purpose of the present study was to conduct a thorough critical review and a meta-analysis of studies evaluating the efficacy of pulmonary rehabilitation in patients with asthma or COPD. The meta-analysis sought to investigate the influence of the method- ologies of the selected studies on outcomes in terms of maximal exercise capacity, functional exercise capacity, and HRQL. The effects of rehabilitation programs that included exercise training were compared with the effects of either conventional treatment or an alternative treatment lacking exercise training. It was questioned whether or not the heterogeneity in effect sizes for functional exercise capacity reported by Lacasse and colleaguesr3 was caused by pooling the results of submaximal cycling tests and walk tests. Because more studies were included in the present study than in the one of Lacasse and colleagues,t3 first the summary effect sizes were calculated and expressed in the metric of the most commonly used measure. Summary effect sizes were also calculated for the long-term effects of pulmonary rehabilitation.

METHODS

Literature Search The database of articles published over the last 45 years was

compiled (1) by means of Index Medicus and MEDLINE using medical subject heading key words pulmonary, obstructive, rehabilitation, and exercise, (2) by routine searching of Current Contents, and (3) by looking for cross-references in studies and

Arch Phys Med Rehabil Vol 80, January 1999

104 LONG-TERM EFFECTS OF PULMONARY REHABILITATION, Cambach

review articles. Articles not written in English, Dutch, or German and abstracts were excluded.

Study Selection Intervention studies were included that (1) examined the

effects of pulmonary rehabilitation programs, (2) concerned patients with asthma and/or COPD older than 18 years, (3) evaluated the outcome measures exercise capacity and/or HRQL, (4) included at least one treatment group receiving exercise training (defined as including at least one of the following training modalities: walking, cycling, stepping exercises, or stair climbing) and at least one control group lacking exercise training, and (5) reported sufficient data to calculate an effect size. These criteria were applied on all studies by two reviewers (WC and ARJK). When disagreement persisted, a third reviewer (RCW) made the final decision.

included randomization procedures with adequate concealment or studies that included double-blinding revealed significantly smaller treatment effects than studies in which these criteria were not satisfied.14 The remaining criteria stemmed from existing scales and checklists. 15x16 It was anticipated that none of the selected studies would match criterion 6 because patients cannot be blinded with respect to the exercise interventions.

After consensus meetings, all controlled studies were evaluated by two reviewers (WC, TWK) to assess methodologic quality. The criteria were scored on a dichotomic scale, that is, studies in which a criterion was matched were scored “1” and studies in which a criterion was not matched, insufficiently matched, irrelevant, or not clear were scored “0.” When disagreement persisted, a third reviewer (RCW) made the final decision.

Meta-Analysis Critical Review

The quality of each study was assessed by applying a checklist containing 20 criteria (table 1). In this way, potential confounders for the effect sizes of individual studies could be traced.

Criteria 1, 4, and 6 were formulated because studies that

Overall effect scores. Most studies reported sufficient statistics to calculate the standardized mean difference, that is, the difference between changes in the treatment group and changes in the control group divided by the pooled standard deviation (SD) of the baseline outcome measure in the treatment and control groups. When in one study two or more rehabilitation

Table 1: The 20 Criteria Applied in Assessing the Methodologic Quality of the Randomized and Nonrandomized Control Trials

Methodologic Quality of Studies Scoring

Randomization procedure is adequate if: 1. Concealed allocation is applied (eg, sealed envelopes);

2. Random sequence generation is applied (eg, random-number table, coin tossing, computer random numbers). Matching procedure is adequate if:

3. Patients are matched according to each of the criteria age, FEV,, and exercise capacity (eg, maximal workload, walking distance).

Blinding procedure is adequate if:

4. Those performing the measurements are kept unaware of the groups to which the patients have been assigned;

5. Those performing the statistical analysis are kept unaware of the groups to which the patients have been assigned; 6. Patients are kept unaware with respect to the condition to which they have been assigned to and/or to the purpose of the

study. Description of dropouts and analysis of data are adequate if:

7. The number of dropouts are described for each group separately;

8. The number of intercurrent dropouts are described for the treatment group; 9. Additional analysis accounting for patients dropping out after randomization (eg, intention-to-treat analysis, worst-case

analysis) has been applied.

Measurement instruments are adequate if: 10. Statistically significant test-retest correlation coefficients and/or intra, interobserver reliability of instrument(s) measuring

exercise capacity have been reported by the authors or have been established in studies cited by the authors; 11. The instrument(s) measuring exercise capacity are compared statistically with other instrument(s) measuring the same

modality by the authors or in studies cited by the authors; 12. Statistically significant test-retest correlation coefficients and/or intra, inter-observer reliability of instrument(s) measuring

quality of life have been reported by the authors or have been established in studies cited by the authors;

13. The instrument(s) measuring quality of life are compared statistically with other instrument(s) measuring the same modality by the authors or in studies cited by the authors.

Control of cointervention(s) are adequate if:

14. Cointerventions leading to systematic differences between groups are avoided; 15. Adjunctive (medical) treatments (ie, oxygen therapy, drug therapy, training regimen) are reported for each group sepa-

rately. Comparability of baseline patient characteristics are adequate if:

16. Age, FEV,, and outcomes of exercise capacity and quality of life are comparable between experimental and control

group(s). Description of the training regimen (dose) is adequate if the:

17. Intensity of each of the training modalities applied during the exercise training sessions has been predetermined in terms

of (percentage) of heart rate or (percentage) of maximum workload; 18. Duration of each of the training modalities applied during the exercise training sessions has been predetermined; 19. Frequency of each of the training modalities applied during the exercise training sessions has been predetermined; 20. Actual intensity and/or duration of exercise training sessions has (or have) been reported.

O/l O/l

011 O/l

O/l O/l

011 O/l O/l O/l

LONG-TERM EFFECTS OF PULMONARY REHABILITATION, Cambach 105

programs (eg, a program including patient education, walking, and cycling, and a program including patient education and inspiratory muscle training) were compared with conventional treatment, the rehabilitation program that approached the previously mentioned definition of exercise training (ie, training that included at least one of the following: walking, cycling, stepping exercises, or stair climbing) most adequately, ie, the program including patient education, walking, and cycling, was chosen.

Unbiased estimates of effect sizes were combined to compute a weighted summary effect size according to the fixed effects model. The methods used to calculate unbiased effect sizes, to combine effect sizes, and to test for statistical significance and homogeneity were described by Hedges and Olkin.17 For each outcome the effect size and the 95% confidence interval (CI) around the mean weighted summary effect size were reported in SD units.

Outcomes of interest included maximal exercise capacity (maximal workload or power during cycling, maximal speed and slope during walking on a treadmill), endurance time (the time that a submaximal workload could be tolerated during cycling on an ergometer or walking on a treadmill), walking distance (the distance walked during either 4, 6, or 12 minutes), and instruments measuring HRQL (disease-specific and generic questionnaires). Effect sizes of each of the outcome measures were separately aggregated.

Additional analyses were conducted for outcome measures expressed as raw mean differences in the metric of the most commonly used measure, that is, for symptom-limited cycle tests expressed in W, for 4-, 6-, and 1Zminute walk tests in meters, and for each of the 4 dimensions of the chronic respiratory questionnaire (CRQ), ie, dyspnea, fatigue, emotion, and mastery (defined as the extent to which the subject feels in control of, or able to cope with, the illness) on a 7-point scale.18 The methods for combining raw mean differences and for testing homogeneity across studies are described by Cooper and Hedges.lg Because endurance time was assessed during walking and cycling at different submaximal loads (range 60% to 95% Wmax) we did not pool the raw mean differences from these tests.

Sensitivity analysis. Subgroup analyses were applied when significant heterogeneity was found across a set of effect sizes. Anumber of potential sources of heterogeneity were specified a priori to examine whether or not elements of study design (randomized control trial versus nonrandomized control trial), patient characteristics (COPD versus asthma and COPD), forced expiratory volume in one second (FEVl) (lower than 50th percentile [P5a] versus larger than PRO), setting of the rehabilitation program (home-based or physiotherapy practice- based programs versus outpatient-based or inpatient-based programs), and type of experimental control (conventional treatment versus concurrent intervention) were related to varia- tion in effect sizes between studies.

The influence of the methodological quality of individual studies on summary effect size was analyzed by using a weighing factor for each study. This weighing factor was calculated by dividing the quality score of each study by the maximum feasible score. In this way, studies with relatively high methodological quality scores were given more weight than studies with relatively low quality scores.

With respect to HRQL, it was, in our opinion, incorrect to pool effect sizes obtained from reliable, valid, and sensitive disease-specific questionnaires (ie, CRQ18 baseline/transition dyspnea index [BDI/TDI] ,*O and medicopsychological questionnaire for lung patients [MPQL]*i) with effect sizes obtained

from generic questionnaires (quality of well-being scale [QWB],22 Bandura scale23), disease-specific questionnaires (shortness of breath questionnaire [SBQ]24), and self- constructed questionnaires, that lacked reliability, validity, or sensitivity. In addition, the CRQ was the most frequently applied measure for HRQL, while out of the other reliable, valid, and sensitive instruments, one study applied exclusively the MPQL4 and one study exclusively the BDI/TD125 (table 2). Therefore, outcome measures of HRQL were initially derived from studies using the dimensions of the CRQ, which were then compared with a composite effect size based on all questionnaires (eg, MPQL, QWB, self-constructed questionnaires). This composite effect size was calculated according to the procedure described earlier.17

In one study, lo the results of two groups receiving rehabilitation in two different settings were compared with a common control group receiving no treatment. In this case, the group that received community-based rehabilitation and the group that received outpatient hospital-based rehabilitation were considered as two separate studies. The same routine was used with respect to another studyll in which the results of two groups receiving physiotherapy maintenance sessions (once a week, once a month) were compared with a common control group receiving no maintenance sessions. As effect sizes under these circumstances are statistically dependent, the influence of successively leaving out the results of each treatment group on summary effect size and 95% CI was analyzed. Finally, in case a significant heterogeneity was found for a particular outcome measure, results of fixed effects analysis were compared with results of random effects analysis.

RESULTS A total of 79 studies were identified as potentially eligible.

After a more detailed review, 42 studies lacking a control condition were excluded. (A list of the uncontrolled studies can be obtained from the authors.) Excluded for further analysis were 6 controlled studies in which both the experimental and control group received exercise training,26-31 and 5 controlled studies in which the rehabilitation program either included no physical training at alp2 or only weightlifting training,33,34 upper arm training,35 and isolated conditioning of peripheral skeletal muscles groups. 36 Furthermore, 4 controlled studies were excluded in which the outcomes were not reported in terms of exercise capacity or HRQL,37-40 and 3 controlled studies in which insufficient data were reported to compute any estimate of effect size.41-43 Finally, one controlled study was excluded because most of the participating patients suffered from coalworkers’ pneumoconiosis.44 The reviewers (WC, ARJK) agreed to include 18 articles4,10,11J2,25~45-57 in the meta-analysis (K: .9358) (table 2).

Critical Review The results of the methodologic quality score of the 18

selected studies are presented in table 3. There was agreement between the two reviewers (WC, TWK) on 332 of the 360 criteria scored (K: 0.8). The methodologic quality score varied from 11%47 to 63%12 of the maximum feasible score of 19 points. The lack of an adequate randomization procedure (12 of 14 studies), blinding procedure (all studies), analysis accounting for patients dropping out after allocation (17 of 18 studies), and description of the actual intensity and/or duration of the exercise training sessions (14 of 18 studies) were most pro- nounced. In addition, in all 18 studies one or more reliable and valid instruments measuring exercise capacity were lacking.

Table 2: Summary of the Relevant Characteristics of the Rehabilitation Programs

Reference

-Sample Size* -Allocation -Setting

-Program Duration -Average FEV* -Components: -Session Frequency -Average Age* Experimental Group -Duration of Exercise -Mode of Exercise During Program/During

Iyrs) -Components: Training Session Assessment -Patients Control Group -Exercise Intensity+ -QOUADL Instrument

McGavin et al,4s 1977

Sinclair et al,46 1980

Pardy et al,47 1981

Reid et al,48 1984

Jones et al,49 1985

Ries et al,51 1986

Busch et aLso 1988

Dekhuijzen

et al,52 1990

Gosselink et al,53 1990

Lake et al,” 1990

Weiner et al,55 1992

Cox et al,4 1993

Goldstein et al,ss 1994

Reardon et al,25 1994

Ries et al,57 1995

-12112

-Random -Home-based

-17/l 6

-Quasi -Outpatient

-Quasi

-Outpatient

-Random -Outpatient

-8/6 -Random -Home-based

-715 -Random

-Outpatient

-Random

-Home-based

-20/20

-Quasi

-Outpatient

-16/I 5

-Random -Outpatient

-6/7 -Random

-Outpatient

-12/I 2

-Random

-Outpatient

-44143

-Quasi -Outpatient

-40140

-Random -Inpatient

-IO/IO -Random -Outpatient

-57162 -Random -Outpatient

-1 .OL/1.2L

-61157 COPD

-1.1L/1.0L

-66164

-31%/25% pred

-67158 -COPD -l.lL/l.lL

-68/61 -COPD

-0.8L/O.7L -64163

-0.9Ul .OL -67162

-0.8L/O.8L -66165

-1.5U1.4L

-58160

-1.5L/2.2L -57158

-Asthma and COPD

-0.7L/l .OL -72166

-33%/39% pred

-64162 -COPD

-2.3U2.4L -43144

-Asthma and COPD

-35%/35% pred -66165

-COPD -35%/33% -66166

-1.2L/1.2L

-62/64 -COPD

-ET -No intervention

-No intervention

-ET, BE, education, PS

-ET, BE

-No intervention

-ET, BE, education

-ET, BE, education, RE*

-No intervention

-ET, education, RE,

PS -No intervention

-No intervention

-ET, BE, education, RE, PS

-No intervention

-Education

-12 weeks -Daily

-At least 2 minutes

-Not specified -10 months

-Daily -At least 15 minutes

-Not specified -2 months

-Daily -Not specified

-Not specified -6 weeks

-5 times a week -30 minutes

-75% Wmax”

-10 weeks -Daily

-Not specified -Not specified

-6 weeks -Daily

-At least 15 minutes

-Not specified -18 weeks -Daily

-0.5-18 minutes -Not specified

-10 weeks

-5 days a week -Not specified -Not specified

-12 weeks

-2 times a week -At least 30 minutes

-6O%-75% Wmax -8 weeks

-3 times a week -At least 20 minutes

-Not specified -6 months -3 times a week

-At least 30 minutes -Up to 50% Wmax

-12 weeks -Daily

-At least 30 minutes -Not specified

-24 weeks -At least 3 times a week -At least 30 minutes

-Not specified -6 weeks -2 times a week

-At least 30 minutes -7O%-85% HRmaxli -8 weeks

-12 sessions -Up to 30 minutes -Not specified

-Stairs/l2MD, SLCT -Well-being

-12MD, stairs/l2MD -Well-being

-Cycling, treadmill, stairs/l2MD, SLCT,

-Treadmill/SLTT, STT -

-Walking, stepping, upper limb/l2MD, SLCT, SCT

-Lubin test, Affectometer test

-Treadmill, walking/l2MD, SLCT, STT -

-Walking or stairs, upper limb/SLCT, SLST

-CRQ (dyspnea-score)

-Not specified/l2MD, SLCT

-ADL score

-Cycling, rowing, stairs/6MD, SLCT, SCT -CRQ, MPQL

-Walking, arm cycling/GMD, SLAT, SLCT

-Bandura scale

-Cycling, rowing, upper limb/l2MD,

SLCT, SCT -

-Cycling, walking, 12MD, swimming/ 12MD, SLCT

-Treadmill, walking upper limb/GMD, SLCT, SCT

CRQ, BDl/TDl

-Cycling, treadmill, stairs, upper limb/ SLTT

-BDI/TDI

-Walking, treadmill/SLTT, STT -QWB, SBQ

Arch Phys Med Rehabil Vol80, January 1999

Table 2: Summary of the Relevant Characteristics of the Rehabilitation Programs (Cont’d)

Reference

Strijbos

et al,l0 1996

Wijkstra

et al,‘l 1996

Cambach

et al,lz 1997

-Program Duration -Average FEV* -Components: -Session Frequency

-Sample Size* -Average Age* Experimental Group -Duration of Exercise -Mode of Exercise During Program/During -Allocation (vs) -Components: Training Session Assessment -Setting -Patients Control Group -Exercise Intensity+ -QOL/ADL Instrument

-15/15 -1 .OL/l .OL -ET, BE, education, -12 weeks -Walking, stairs, cycling/4MD, SLCT

-Random -61163 RE -Daily -Well-being -Outpatient -COPD -No intervention -At least 15 minutes

-70% Wmax -28/l 5 -1.2L/1.2L -ET, BE, RE, IMT, -12 weeks -Cycling, upper limb/GMD, SLCT

-Random -64162 education -Daily -CRQ

-Home-based COPD -No intervention -At least 30 min

-6O%-70% Wmax -37129 -2.3Ll2.2L -ET, BE, education, -12 weeks -Cycling, rowing, stepping/GMD, SCT

-Random -49155 RE* -3 times a week -CRQ -Physiotherapy -Asthma and -No intervention -At least 30 minutes

practice- COPD -6O%-70% Wmax based

Abbreviations: FEV 1, forced expiratory volume in 1 second; QOL, quality of life; ADL, activities of daily living; ET, exercise training; IMT, inspiratory muscle training; BE, breathing exercises; PS, psychological support; VMT, ventilatory muscle training; RE, relaxation exercises; 12MD, IZ-minute walking test; SLCT, symptom limited cycle test; SCT, submaximal cycle test; STT, submaximal treadmill test; CRQ, chronic respiratory disease questionnaire; 6MD, 6-minute walking test; MPQL, medicopsychological questionnaire for lung patients; SLAT, symptom limited arm test; BDI/TDI, baseline/transitional dyspnea index; SLTT, symptom limited treadmill test; QWB, Quality of Well-Being Scale; SBQ, shortness of breath questionnaire; 4MD, 4-minute walking test. * Rehabilitation group/control group. + Specified either in terms of percentage of heart rate or percentage of maximum workload, or “not specified.” * Randomized cross-over design (only first phase of the trial is considered). * Percentage of maximal workload achieved during a symptom limited exercise test. 11 Percentage of maximal heart rate achieved during a symptom limited exercise test.

Meta-Analysis Exercise capacity. Analyses for maximal exercise capac-

ity, endurance time, and walking distance showed statistically significant summary effect sizes (p < .OOOl), indicating improvements in these outcomes in favor of programs including exercise training (table 4). The overall heterogeneity statistics were not significant for maximal exercise capacity and walking distance (p > .16), implying that effect sizes were homogeneous across studies.

A significant heterogeneity of effect sizes was found for

endurance time (p < .OOOl). The subgroup analyses presented in table 5 indicate that the FEVt explains some of the heterogeneity in endurance time (ie, no significant heterogeneity in effect size was found across the four studies that included patients with a FEVt lower than Pss and the three studies that included patients with a FEVl larger than P5a [p = . IO]).

Sensitivity analyses for methodological quality showed identical summary effect sizes and CIs for maximal exercise capacity and waking distance. The latter analyses slightly changed the summary effect size for endurance time (ie, 1.3; CI

Table 3: Methodologic Quality of the Studies

Allocation

Authors/ Procedure

Criteria

McGavin et al,45 1977

Sinclair et aI/” 1980 Pardy et al,47 1981 Reid et al,48 1984

Jones et aL4$ 1985

Ries et al,51 1966 Busch et al,50 1988 Dekhuijzen et al,52 1990

Gosselink et al,53 1990 Lake et al,54 1990

Weiner et al,55 1992 Cox et a l,4 1993

Goldstein et al,56 1994 Reardon et al,25 1994 Ries et al,a7 1995

Strijbos et al,lo 1996 Wijkstra et al,‘l 1996 Cambach et al,lz 1997

Blinding Procedure

Dropouts Reliability and ITT and Validity

7 8 9 10 11 12 13

Confoundina

14 15 16

Description of Program

17 18 19 20 Total Score

- - 1 - - 0 0 0 -

0 0 - 0 0 - - - 0

0 0 - 0 0 -

0 0 - - - 0

0 0 - 0 0 - 1 1 -

Total score: 2 2 1

0 0 0 0 0 0 0 0 0

0 0 0 1 0 0 0 0 0

0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

1 0 0 0 0 0

0 0 0 0 0 0 0 0 0

1 1 0 0 0 0 0

1 1 0 0 0 0 0 10000--

- oooo--

1000000 1 0 0 0 0 - -

1000011 - 000000

1000011 1 0 0 0 0 0 0

- oooo--

1 0 0 0 0 0 0 1000011

- 000011 1 0 0 0 0 0 0 1 0 0 0 0 0 0

1 1 0 0 0 1 1 1 1 1 0 0 1 1

1 0 0 1 0 1 1 0 0

1 0 0 1 1 1

1 0 1 1 1 1 1 1 0

1 0 1 1 1 1

1 1 0 1 1 1 1 0 0

14 4 1 0 0 6 6 18 6 9

0 1 1 0

0 1 1 0 0 0 0 1 1110

0 1 1 0

1110 1 0 1 1 0 0 1 0

1110 1110

1110 0 1 1 0 0 1 1 0

1 0 1 0 0 1 1 0

0 0 1 0 1110 1 1 1 1

9 13 17 3

Table 4: Overall Effects of Pulmonary Rehabilitation

Outcome Measure N Effect* 95% Cl pValue+ aTotal*

Exercise capacity Maximal exercise capacity 12 0.4 0.2-0.6 <.OOOl NS; .87 Endurance time 7 1.2 0.9-1.5 <.OOOl <.OOOl Walking distance 15 0.5 0.3-0.7 c.0001 NS;.16

HRQL Dyspnea 5 0.7 0.4-1.0 <.OOOl NS; .41 Total score 7 0.6 0.4-0.8 <.OOOl NS; .30

Fatigue 4 0.6 0.3-0.9 .OOOl NS; .73 Emotion 4 0.5 0.2-0.7 .OOl NS; .66 Mastery 4 0.6 0.3-0.9 <.OOOl NS; .96 Total score 11 0.6 0.5-0.7 <.OOOl NS; .I4

Dyspnea, fatigue, emotion, and mastery scores based on the CRQ. Total scores based on all questionnaires. Abbreviations: N, number of studies; QTotal, total heterogeneity. * Standard deviation units. + Significance indicates rejection of the null hypothesis that the summary effect equals 0. * Significance indicates rejection of the hypothesis of homogeneity.

1.1 to 1.6). Sensitivity analyses for the multiple comparisons in the studies by Strijbos and colleagueslo and Wijkstra and colleagues’l did not change the summary effect sizes and CIs. Table 6 demonstrates for endurance time the much wider CI resulting from random effects analysis in comparison with fixed effects analysis, but the summary effect size remained statistically significant (p = .02).

Additional analyses of the raw mean differences revealed significant improvements for maximal exercise capacity and walking distance (p < .OOOl; table 7). For the 4- and 6minute walk tests no significant heterogeneity was found for the mean differences across studies (p > .91).

Analysis of the long-term effects of rehabilitation was possible on the basis of studies using maximal exercise tests and walk tests (table 8). The results indicate significant improvements up to 9 months after rehabilitation in both outcome measures in favor of the patients who received the program with exercise training (p < .OOS).

HRQL. Significant summary effect sizes were found for the dimensions dyspnea, fatigue, emotion, and mastery (p < .OOl), indicating improvements in these dimensions in favor of patients receiving the program that included exercise training. The summary effect sizes were not significantly heterogeneous (p > .41).

Sensitivity analyses for methodological quality revealed identical summary effect sizes for all four dimensions of the

Table 6: Comparison Between the Results of Fixed Effects Analysis Versus Random Effects Analysis for Endurance Time

Effect Variance Standard P N Size* Component Error 95% Cl Value+

Endurance time fixed effects model 71.2 - 0.1 1.0-I .5 <.OOOl

Endurance time random effects model 7 1.0 1.7 0.3 0.04-2.1 .02

Abbreviation: N, number of studies. * Effect sizes in SD units. + Significance indicates rejection of the null hypothesis that the summary effect size equals 0.

CRQ. The summary effect size and CI of the five studies using the dyspnea dimension of the CRQ, and the composite effect size and confidence interval of the seven studies using either the CRQ or one of the other questionnaires measuring dyspnea were comparable (table 4). Also, the summary effect size and CI of the four studies reporting the dimensions fatigue, emotion, and mastery of the CRQ, and the composite effect size and CI of the 11 studies using either the CRQ or one of the other questionnaires measuring HRQL, were comparable (table 4).

Additional analyses of the raw mean differences demonstrated significant improvements for dyspnea, fatigue, emotion, and mastery (p < .0005; table 7). For all of the dimensions of the CRQ the mean differences were not significantly heterogeneous across studies (p > .49).

DISCUSSION The present research synthesis corroborates the findings of

the meta-analysis by Lacasse and colleagues13 as well as those of a large number of reviews3,5s7-9 that the acute effects of pulmonary rehabilitation programs comprising exercise training involve improvements in exercise capacity and HRQL.

The results indicate that, with the exception of endurance time, the outcomes were not significantly heterogeneous with respect to study design, patient characteristics, organizational setting, and type of experimental control. With respect to study design, Kwakkel and colleagues5g also found no significant differences between randomized control trials and nonrandomized control trials in a comparable research synthesis on the effects of intensity of rehabilitation after stroke. Concerning patient characteristics, one recent study indicated that both patients with asthma and COPD may benefit from rehabilitation in terms of exercise capacity and HRQL.i2 Another study

Table 5: Results of Subgroup Analysis

Endurance Time RCT NCT COPD

Asthma and

FEV, i50th

Percentile

FEV, 250th

Percentile Community-

Based Inpatient/

Outpatient

No Intervention

in Control Group

Intervention in Control

N 6 1 5 2 4 3 1 6 4 3 Effect size* 1.4 - 1.711 1.0 1.8 0.9 1.4 2.2 1 .o 1.7 0.8 p Value+ <.OOOl .003 <.OOOl <.OOOl <.OOOl <.OOOl <.OOOl <.OOOl <.OOOl <.OOOl Q-within* .Ol <.OOOl <.OOOl <.OOOl <.OOOl Q-between’ <.OOOl ,009 NS; .I0 .OOl .0008

Abbreviations: RCT, randomized control trial; NCT, nonrandomized control trial; N, number of studies; Q-within, within-group heterogeneity; Q-between, between-group heterogeneity; NS, not significant. * Effect sizes in SD units. + Significance indicates rejection of the null hypothesis that the summary effect size equals 0. * Significance indicates rejection of the hypothesis of homogeneity of effect sizes within groups. § Significance indicates rejection of the hypothesis of my homogeneity of effect sizes between groups. 11 Significant difference between effect sizes as measured by analyzing contrasts among mean effects; p 4.05.

Arch Phys Med Rehabil Vol80, January 1999

Table 7: Overall Effects of Pulmonary Rehabilitation

Outcome Measure N Effect* 95% Cl pValue+ Q Total*

Exercise capacity Maximal exercise capacity 8 26W 23-29 <.OOOl .05

Endurance time - - - - -

Walking distance 4-minute

6-minute 12-minute

HRQL Dyspnea

Fatigue Emotion

Mastery

2 30m 20-40 1.0001 NS; .93

5 49m 23-75 .OOOl NS; .91 8 133m 128-138 <.OOOl <.OOOl

5 0.8 0.5-1.1 <.OOOl NS; .49

4 0.7 0.4-1.0 <.OOOl NS; .57 4 0.5 0.2-0.8 .0005 NS; .70

4 0.8 0.5-1.1 <.OOOl NS; 66

Dyspnea, fatigue, emotion, and mastery scores based on the CRQ. Abbreviations: N, number of studies; Q total, total heterogeneity; NS, not significant. * Raw mean differences. + Significance indicates rejection of the null hypothesis that the summary effect equals 0. * Significance indicates rejection of the hypothesis of homogeneity.

demonstrated equal improvements in exercise capacity and HRQL directly after rehabilitation in a home-based setting and an outpatient clinical setting, implying that programs run in these settings may yield similar results.1°

It may be hypothesized on the basis of the results of the present research synthesis that conventional care is as effective as education alone or inspiratory muscle training (IMT) alone (ie, effect sizes were homogeneous with respect to type of experimental control). The results of two meta-analyses on the effects of education (including elements such as coping with stress, breathing techniques, and use of medication)60 and IMT6i support this hypothesis. However, in the meta-analysis by Devine and Pearcy60 a maximum of only three relevant studies focused on the effect of education alone. The latter studies lacked a complete description of the experimental treatment. Furthermore, Smith and colleague@ made no distinc- tion between, on the one hand, studies on the effects of IMT added to conventional care, and, on the other, IMT added to exercise training. Therefore, it can be concluded that the effect of education alone or combined with exercise training remains unclear, and that the evidence for the additional benefits of IMT over exercise training alone is equivocal.62

The methodological quality of the majority of the primary studies included in this research synthesis showed some major shortcomings, which might have biased the outcomes. For example, it was observed that only two studies met the criteria for an adequate randomization procedure, none met those for blinding, and only one applied an alternative analysis accounting for dropouts after allocation. Furthermore, it should be noted that with the exception of the tests measuring walking distance, the validity and reliability of instruments measuring

Table 8: Overall Long-Term Effects of Pulmonary Rehabilitation

Outcome Measure N Effect” 95% Cl pValue+ Q Total”

Exercise capacity

Maximal exercise capacity 6 0.3 0.1-0.5 <.003 NS; .93

Walking distance 5 0.4 0.1-0.7 <.005 NS; .26

Abbreviations: N, number of studies: Q Total, total heterogeneity; NS, not significant. * SD units. + Significance indicates rejection of the null hypothesis that the summary effect equals 0. * Significance indicates rejection of the hypothesis of homogeneity.

exercise capacity still have not been demonstrated.12 However, sensitivity analyses for methodologic quality of studies revealed no marked influence on maximal exercise capacity, functional exercise capacity, and HRQL.

Lacasse and colleagues13 reported significant heterogeneity for outcomes of functional exercise capacity that included outcomes of submaximal cycle tests and walk tests. In the present meta-analysis, separate analyses of the results of submaximal tests (endurance time) and walk tests (walking distance) demonstrated that heterogeneity in the effect sizes of the individual studies was caused by the heterogeneity in effect sizes reported for submaximal tests (p < .OOOl), and not for walk tests (p = .16). Some of the heterogeneity in the effect sizes of the submaximal tests (endurance time) could be explained by means of the FEVt of patients.

For most outcome measures it was possible to compute raw mean differences in the metric of the most commonly used measure. The improvement of 26 2 3W in maximal exercise capacity must be interpreted with caution because the results in maximal exercise capacity were significantly heterogeneous across studies (p = .05; table 7). The study by Cox and colleagues,4 reporting the greatest improvement in maximal exercise capacity (27W) with the largest sample size (83 patients), had a large influence on the overall mean effect. The patients who participated in this outpatient rehabilitation program were relatively young (mean age: 44 years), and suffered from relatively mild asthma and COPD (mean FEVt, 2.4L). Leaving out the latter study in the analysis revealed a more conservative improvement of 11 f 8W (heterogeneity test: p = .72), approaching the overall mean effect of SW (CI 3 to 17) estimated by Lacasse.13

The improvement in 6-minute walking distance of 49 + 26m demonstrated in the present meta-analysis approached the estimated minimal clinical important difference of 50m (CI 37 to 71).13 Lacasse, l3 however, demonstrated a larger summary effect size in 6-minute walking distance (56m), and, in particular, a wider confidence interval (CI 28 to 93). This difference in findings between the present meta-analysis and the one of Lacasse13 may be explained by differences in the methods used to calculate the overall mean effect in terms of the original metric or by differences in the number and characteristics of the selected studies (9 of the 14 studies that were selected in Lacasse’s meta-analysis were included in the present meta- analysis).

The improvements in all 4 dimensions of the CRQ found in the present meta-analysis matched the estimated minimal clinical important change in score of 20.5 points per item (CI 0.4 to 0.7) in HRQL.63 The findings of the present study support those of Lacasse,13 with the exception of the higher overall mean value and the wider CI for dyspnea reported in Lacasse’s study (ie, 1.0 points, CI 0.6 to 1.5). Again, this finding may be related to the previously mentioned differences between the two meta-analyses.

The present meta-analysis reports significant improvements in maximal exercise capacity and walking distance up to 9 months after rehabilitation in favor of the patients who received the program including exercise training. In some of the studies, the patients received a supervised maintenance exercise program after rehabilitation,*1,57 while in other studies such maintenance programs were absent4,10 The homogeneity across the study results suggests that maintenance sessions after rehabilitation may not be crucial. It is possible, however, that most of the patients in the studies that lacked maintenance sessions4J0 might have been prone to adapt an active way of life. This may be explained by the fact that the patients in the

latter studies were either relatively young (mean age: 44 years)4 or received therapy sessions in their locality.1°

Publication bias, indicating that studies reporting statistically significant results tend to be published more frequently than studies without such significant results, may interfere with the summary effect size estimates found. Additional analysis,19 however, revealed that on average 35 unpublished studies (range: 4 studies for maximal exercise capacity to 114 studies for endurance time) with an average effect of zero would be needed to reverse the findings of the present study, which provides reasonable evidence that publication bias was not a major threat.

In summary, the present research synthesis demonstrated statistically significant improvements in terms of exercise capacity and HRQL in patients with asthma and COPD. Sensitivity analyses for methodological quality of the selected studies did not change summary effect sizes. Effect sizes were significantly heterogeneous for the outcome measure endurance time, but not for walking distance. Significant improvements were found up to 9 months after rehabilitation for maximal exercise capacity and walking distance.

Future research on the relative efficacy of components of the rehabilitation program, the appropriate setting of rehabilitation, and the minimally required intensity and frequency of maintenance training sessions for patients with asthma or COPD is necessary. Furthermore, questions concerning the prognostic value of (potential) determinants of the outcome of rehabilitation in patients with asthma and COPD are still open to further investigation. Finally, the development of reliable, valid, and sensitive estimates of clinically relevant improvements may further substantiate whether or not improvements in exercise capacity and HRQL are really meaningful for individual patients.64

Acknowledgment: The authors thank Gert Kwakkel for his contribution to the data analysis and the authors of the original articles included in this research synthesis who provided additional data on our request.

References Barach AL, Bickerman HA, Beck G. Advances in treatment of nontuberculous pulmonary disease. Bull N Y Acad Med 1952;28: 353-84. Petty TL. Pulmonary rehabilitation in perspective: historical roots, present status, and future projections. Thorax 1993;48:855-62. Ries AL. Position paper of the American Association of Cardiovas- cular and Pulmonary Rehabilitation. Scientific basis of pulmonary rehabilitation. .I Cardiopulm Rehabil 1990;10:418-41. Cox NJ, Hendricks JC, Binkhorst RA, Van Herwaarden CLA. A pulmonary rehabilitation program for patients with asthma and mild chronic obstructive pulmonary disease (COPD). Lung 1993; 171:235-44. Casaburi R. Exercise training in chronic obstructive pulmonary disease. In: Casaburi R, Petty TL, editors. Principles and practice of pulmonary rehabilitation. Philadelphia: WB Saunders; 1993. p. 204-24. Curtis JR, Deyo RA, Hudson LD. Health-related quality of life among patients with chronic obstructive pulmonary disease. Thorax 1994;49:162-70. Lertzman MM, Chemiack RM. State of the art. Rehabilitation of patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 1976;114:1145-65. Bohning W, Golinske M, Wettengel R. Rehabilitation bei atemweg- serkrankungen. Internist 1985;26:228-32. Celli BR. Pulmonary rehabilitation in patients with COPD. Am J Respir Crit Care Med 1995;152:861-4. Strijbos JH, Postma DS, Van Altena R, Gimeno F, Koeter GH. A comparison between an outpatient hospital-based pulmonary rehabilitation program and a home-care pulmonary rehabilitation

program in patients with COPD. A follow-up of 18 months. Chest 1996;109:366-72. Wijkstra PJ, Ten Verger? EM, Van Altena R, Otten V, Kraan J, Postma DS, et al. Long term effects of rehabilitation at home on quality of life and exercise tolerance in patients with chronic obstructive pulmonary disease. Thorax 1996;50:824-8. Cambach W, Chadwick-Straver RVM, Wagenaar RC, Van Keim- pema ARJ, Kemper HCG. The effects of a community-based pulmonary rehabilitation programme on exercise tolerance and quality of life: A randomized controlled trial. Em Resp J 1997;lO: 104-13. Lacasse Y, Wong E, Guyatt GH, King D, Cook DJ, Goldstein RS. Meta-analysis of respiratory rehabilitation in chronic obstructive pulmonary disease. Lancet 1996;348: 1115-9. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995;273: 408-12. Chalmers TC, Smith H, Blackbum B, Silverman B, Schroeder B, Reitman D, et al. A method for assessing the quality of a randomized control trial. Control Clin Trials 198 1;2:3 l-49. DerSimonian R, Charette LJ, McPeek B, Mosteller F. Reporting on methods in clinical trials. N Engl J Med 1982;306:1332-7. Hedges LV, Olkin I, editors. Statistical methods for meta-analysis. New York: Academic Press; 1985. Guyatt GH, Belman LB, Townsend M, Plugsley SO, Chambers LW. A measure of quality of life for clinical trials in chronic lung disease. Thorax 1987;42:773-8. Cooper H, Hedges LV, editors. The handbook of research synthesis. New York: Russel Sage; 1994. Mahler DA, Weinberg DH, Wells CK, Feinstein AR. The measurement of dyspnea: contents, interobserver agreement, and physiologic correlates of two new clinical indexes. Chest 1984;85: 751-8. Cox NJM, Hendricks JCM, Nijkhuisen R, Binkhorst RA, Van Herwaarden CA. Usefulness of a medicopsychological questionnaire for lung patients. Int J Rehabil Res 1991;14:267-72. Kaplan RM, Atkins CJ, Timms R. Validity of a quality of well-being scale as an outcome measure in chronic obstructive pulmonary disease. J Chron Dis 1984;37:85-95. Bandura A, Adam NE. Analysis of self efficacy theory of behavioral change. Cogn Ther Res 1977;1:287-308. Archibald CJ, Guidotti TL. Degree of objectively measured impairment and perceived shortness of breath with activities of daily living in patients with chronic obstructive pulmonary disease. Can J Rehabil 1987;1:45-54. Reardon J, Awad E, Normandin E, Vale F, Clark B, Zuwallack RL. The effect of comprehensive outpatient pulmonary rehabilitation on dyspnea. Chesti994;105: 1046-52. - - Hale T. Sorieas J. Hamlev EJ. Effectiveness of an exercise regime

1 vu ,

on the rehabilitation of chronic obstructive lung disease patTents using heart-rate as the parameter. Br J Sports Med 1976;10:71-5. Casaburi R, Patessio A, Ioli F, Zanaboni S, Donner CF, Wasserman K. Reductions in exercise lactic acidosis and ventilation as a result of exercise training in patients with obstructive lung disease. Am Rev Respir Dis 1991;143:9-18. Ries AL, Ellis B, Hawkins RWH. Upper-extremity exercise training in chronic obstructive pulmonary disease. Chest 1988;93: 688-92. Martinez FJ, Vogel PD, DuPont DN, Stanopoulus I, Gray A, Beamis JF. Supported arm exercise vs unsupported arm exercise in the rehabilitation of patients with severe chronic airflow obstruction. Chest 1993;103:1397-1402. Wanke Th, Formanek D, Lahrmann H, Brath H, Wild M, Wagner C, et al. Effects of combined inspiratory muscle and cycle ergometer training on exercise performance in patients with COPD. Eur Respir J 1994;7:2205-11. Vallet G, Ahmdid S, Serres I, Fabre C, Bourgouin D, Desplan J, et al. Comparison of two training programs in chronic airway limitation patients: standardized versus individualized protocols. Eur Respir J 1997;10:114-22. Ambrosino N, Paggiaro PL, Macchi M, Filieri M, Toma G, Aghini Lombardi F, et al. A study of short-term effect of rehabilitative

therapy in chronic obstructive pulmonary disease. Respiration 1981;41:40-4. O’Hara WJ, Lasachuk KE, Matheson PC, Renahan MC, Schlotter DG, Lilker ES. Weight training and backpacking in chronic pulmonary disease. Respir Care 1984;29: 1202-10. Simpson K, Killian K, McCartney N, Stubbing DG, Jones NL. Randomized controlled trial of weightlifting exercise in patients with chronic airflow limitation. Thorax 1992;47:70-5. Bauldoff GS, Hoffman LA, Sciurba F, Zullo TG. Home-based, upper-arm exercise training for patients with chronic obstructive pulmonary disease. Heart Lung 1996;25:288-94. Clark CJ, Cochrane L, Mackay E. Low intensity peripheral muscle conditioning improves exercise tolerance and breathlessness in COPD. Eur‘Resp J 1996;9:2590-6. Haas A. Cardon H. Rehabilitation in chronic obstructive oulmo- nary disease. Med Clin North Am 1969;53:593-606. A Sergysels R, De Coster A, Degre S, Denolin H. Functional evaluation of a physical rehabilitation program including breathing exercises and bicycle training in chronic obstructive lung disease. Respiration 1979;38:105-11. Flicker M, Wanke T, Zwick H. Trainingsergebnisse am Beispiel von drei Patientengruppen. Prax Klin Pneumol 1988;42:628-33. Cochrane LM, CJ Clark. Benefits and problems of a physical training programm for asthmatic patients. Thorax 1990;45:345-51. Chester EH, Belman MJ, Bahler RC, Baum GL, Schey G, Buch P. Multidisciplinary treatment of chronic pulmonary insufficienty. 3. The effect of physical training on cardiopulmonary performance in patients with COPD. Chest 1977;72:695-702. Booker HA. Exercise training and breathing control in patients with chronic airflow limitation. Physiotherapy 1984;70:258-60. Berrv MJ. Adair NE. Sevenskv KS. Ouinbv A. Lever HM. Insp&ato& muscle training and- whole-Gody -rec&ditioning in chronic obstructive pulmonary disease. A controlled randomized trial. Am J Respir Crit Care Med 1996;153:1812-6. Cockcroft AE, Saunders MJ, Berry G. Randomized controlled trial of rehabilitation in chronic respiratory disability. Thorax 1981;36: _ _ 200-3. McGavin CR, Gupta SP, Lloyd EL, McHardy JR. Physical rehabilitation for chronic bronchitic: results of a controlled trial of exercise in the home. Thorax 1977;32:307-11. Sinclair DJM, Ingram CG. Controlled trial of supervised exercise training in chronic bronchitis. BMJ 1980;23:519-20. Pardv RL. Rivington RN. Desoas PJ. Macklem PT. Insuiratorv mu&e training compare2 wi& physiotherapy in pat&s wi& chronic airflow limitation. Am Rev Respir Dis 1981;123:421-5. Reid WD, Warren CPW. Ventilatory muscle strength and endurance training in eldery subjects and patients with chronic airflow limitation: A pilot study. Physiother Can 1984;36:305-11. Jones DT, Thomson RJ, Sears MR. Physical exercise and resistive

breathing training in severe chronic airways obstruction--are they effective? Eur J Respir Dis 1985;67:159-66.

50. Busch AJ, McClements JD. Effects of a supervised home exercise program on patients with severe chronic obstructive pulmonary disease. Phys Ther 1988;68:469-74.

51. Ries AL, Moser KM. Comparison of isocapnic hyperventilation and walking exercise training at home in pulmonary rehabilitation. Chest 1986;90:285-9.

52. Dekhuijzen PNR, Folgering ThM, Van Herwaarden CLA. Target- flow inspiratory muscle training at home and during pulmonary rehabilitation in COPD patients with a ventilatory limitation during exercise. Lung 1990;168 Suppl:502-8.

53. Gosselink HAAM, Wagenaar RC, Van Keimpema ARJ, Chadwick- Straver RVM. Het effect van een reactiveringsprogramma bij patiEnten met CARA. Ned T Fysiotherapie 199O;lOO: 193-9.

54. Lake FR, Henderson K, Briffa T, Openshaw J, Musk AW. Upper-limb and lower-limb exercise training in patients with chronic airflow obstruction. Chest 1990;97:1077-82.

55. Weiner P, Azgad Y, Ganam R. Inspiratory muscle training combined with general exercise reconditioning in patients with COPD. Chest 1992;102:1351-6.

56. Goldstein RS, Gort EH, Stubbing D, Avendano MA, Guyatt GH. Randomised controlled trial of respiratory rehabilitation. Lancet 1994;344:1394-7.

57. Ries AL, Kaplan RM, Limberg TM, Prewitt LM. Effects of pulmonary rehabilitation on physiologic and psychosocial outcomes in patients with chronic obstructive pulmonary disease. Ann Intern Med 1995;122:823-32.

58. Cohen J. Weighted kappa: Nominal scale agreement with provi- sion for scaled disagreement or partial credit. Psycho1 Bull 1968;70:213-20.

59. Kwakkel G, Wagenaar RC, Koelman TW, Lankhorst GJ, Koetsier JC. Effects of intensity of rehabilitation after stroke. A research synthesis. Stroke 1997;28:1550-6.

60. Devine EC, Pearcy J. Meta-analysis of the effects of psychoeduca- tional care in adults with chronic obstructive pulmonary disease. Patient education and counseling. 1996;29:167-78.

61. Smith K, Cook D, Guyatt GH, Madhavan J, Oxman AD. Respira- tory muscle training in chronic airflow limitation: A meta-analysis. Am Rev Respir Dis 1992;145:533-9.

62. Lacasse Y, Guyatt GH, Goldstein RS. The components of a respiratory rehabilitation program A systematic overview. Chest 1997;111:1077-88.

63. Jaeschke R, Singer J, Guyatt GH. Measurement of health status: ascertaining the minimum clinically important difference. Control ClinTrials 1989;10:407-15.

64. Wright JG. The minimal important difference: Who’s to say what is important? J Clin Epidemiol 1996;11:1221-2.

The long-term effects of pulmonary rehabilitation in patients with asthma and chronic obstructive...

Documents