Family Practice Vol. 19, No. 6, 665-674
© Oxford University Press 2002
A randomized controlled trial of group aerobic exercise in primary care patients with persistent, unexplained physical symptoms
a Departments of Psychiatry,
b Primary Care and
d Clinical Psychology, University of Liverpool and
c Department of Physiotherapy, Keele University, UK.
Sarah Peters, Department of Psychiatry, University of Liverpool, Royal Liverpool University Hospital, Liverpool L69 3GA, UK; E-mail: speters{at}liv.ac.uk
Peters S, Stanley I, Rose M, Kaney S and Salmon P. A randomized controlled trial of group aerobic exercise in primary care patients with persistent, unexplained physical symptoms. Family Practice 2002; 19: 665–674.
Received 4 December 2001; Revised 29 April 2002; Accepted 16 July 2002.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
Background. The management of persistent, unexplained physical symptoms is challenging and often unsatisfactory for patients and doctors. Aerobic exercise training has benefited patients referred to secondary care with symptoms of chronic fatigue and fibromyalgia. It is not known if this approach is either possible or beneficial for patients with the broader range of persistent, unexplained symptoms found in primary care.
Objectives. To examine the feasibility and effects of aerobic exercise training in primary care patients with unexplained physical symptoms persisting more than 12 months.
Methods. Randomized comparison (n = 228) of aerobic exercise with stretching as control among patients recruited from primary care. Training comprised 20, one-hour, sessions led by NHS physiotherapists. Adherence to training was recorded along with two groups of outcome measures: (i) documented symptoms and health care use, monitored from six months before to six months after training; and (ii) self-reported measures including emotional state and perceived disability, assessed before, during and six months after training.
Results. Exercise training proved feasible: more than 70% of referred patients attended for assessment and were randomized to aerobic or control exercise; 78% of eligible patients attended the first session; and median attendance was 11 sessions for both programmes. Primary care consultations and prescriptions were significantly reduced in the 6 months after training; extent of reduction was related to attendance at training sessions, irrespective of type. Whilst self-reported measures improved similarly during both training programmes, improvements were unrelated to level of attendance.
Conclusion. For primary care patients with persistent, unexplained physical symptoms willing to be involved in exercise training, aerobic exercise offers no benefits over non-aerobic exercise. Whilst the observed reduction in primary health care use following exercise training is potentially of practical importance in a group of patients characterized by high consultation rates, improvement in patients’ subjective state was not clearly attributable to exercise training.
Keywords. Exercise, persistent unexplained physical symptoms, primary care, randomized controlled trial.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
Persistent, physical symptoms that defy explanation in terms of physical pathology have been recognized in medical practice for several centuries.1 In succeeding generations the labels attached to such presentations have changed with fashionable medical views of their nature: hysteria,2 neurasthenia3 and, currently, somatization disorders4,5 reflecting the varying importance attached to physical, social and psychological causative factors.
Because they borrow the language of medicine to construct and articulate their illness,4,6 patients with persistent, unexplained physical symptoms (PUPS) present a formidable management challenge. They are frequently, but not invariably, anxious or depressed7–9 but effective treatment is often rendered problematic by patients’ denial of emotional problems. To the dismay of their GPs,10 typically, they consult frequently,11 seek and obtain above average rates of specialist advice but remain dissatisfied with their care.6 The outcome of existing medical management of PUPS is unsatisfactory. Although most new presentations of unexplained abdominal symptoms in primary care have resolved by one year,12 abridged somatization (4 and 6 current symptoms in males and females, respectively) persists in around half of patients.13 Long-term follow-up of neurological and motor symptoms referred to hospital reveals continuing physical disability and psychological problems.14,15
This perplexing situation has directed attention to alternative approaches to management.16–18 Physical exercise is known to influence emotional state and well-being19 and controlled trials in Chronic Fatigue Syndrome (CFS)20 and fibromyalgia21 have demonstrated its effectiveness in secondary care. However, it is not known if exercise training is feasible or effective for patients with PUPS found in primary care.11
By contrast with the relatively selected and symptomatically homogeneous groups that are found in secondary care clinics, the heterogeneity of PUPS patients in primary care, and their scepticism about medical explanation and intervention6 are likely to present considerable challenges to engagement. Therefore, in the present study, we examined (i) the proportion of such patients that can be recruited into group exercise training programmes led by NHS physiotherapists; (ii) the level of adherence to exercise training that can be achieved; and (iii) whether aerobic exercise training offers any benefit to patients, as judged by self-reported measures and documented health care use. Stretching training was chosen as a control for group activity, contact with therapist and mastery of physical skills.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
Subjects
Following approval of relevant ethical committees, GPs in Liverpool and the surrounding area were invited to refer into the study patients with physical symptoms that had persisted for at least 12 months and remained unexplained following specialist investigation. In all, 323 patients were referred: 15 were excluded because of hypertension and/or ischaemic heart disease (n = 9), asthma (n = 4) or history of psychosis (n = 2). No patient regularly exercised more than 3 times weekly. Eighty patients failed to attend for interview or declined to take part. The remaining 228 (74% of those eligible) attended for initial assessment, agreed to participate in the study and had normal ECGs at rest and upon exercise. Following initial assessment, consecutive patients were randomly allocated to either an aerobic exercise or a stretching training programme. Patients were not told the content of the programme they had been allocated to until their first session. It was not feasible to blind therapists or research assistant to randomization. However, throughout the study, patients remained blind to the hypothesis and the content of the alternative exercise programme. Recruitment, randomization and induction into training are summarized in Figure 1
.
|
Intervention
Both exercise programmes involved attendance at a university gymnasium for one-hour sessions twice weekly. Patients exercised in groups of ten and remained in the same group for the ten week duration of training. Each session of training was led by one of two NHS physiotherapists, who alternated supervision of aerobic and stretching groups. As part of their training, patients were taught to use a portable digital pulse monitor (Polar, Finland) and to monitor their radial pulse. In both training programmes, distinctive explanatory models were presented to patients to link the content of training to potential amelioration of their symptoms. Exercise was presented as ‘a way of keeping the body conditioned so better able to cope with the strains put upon it by everyday life’. Furthermore, during exercise sessions, therapists adapted explanations to fit specific presenting complaints and patients’ beliefs, for example acknowledging that in chronic fatigue, energy is perceived as limited. The programmes are described in detail elsewhere.22
Intensity of training.. In aerobic training, effort was titrated to maintain patients’ heart rate at 60–65% of age-adjusted estimated maximum, a level which has previously been shown to be optimal for achieving psychological benefits in a general population sample selected for high anxiety.23 Each session began with 10 minutes warm-up before a circuit of specific exercises for 40 minutes, concluding with a 10 minute cool-down.
Stretching training was designed to maintain heart rate below 50% of the estimated age-adjusted maximum. Relaxation exercises for 10 minutes began and completed the session. During the 40 minutes of stretching exercises, involving all major joint complexes, patients were asked to maintain muscle tension at a level that provoked a feeling of ‘stretch’.
To maximize compliance with the prescribed level of exercise, patients were prompted to monitor their pulse at regular intervals during each session. Following the 3rd, 12th and 17th sessions, individual feedback of pulse readings was provided using a graphic display on a laptop computer.
Homework.. Patients in each programme were asked to carry out specific exercises between sessions for at least 20 minutes three times weekly, guided by written and pictorial information, and to record exercises undertaken in a structured homework diary. At the end of every session, homework diaries were reviewed by the physiotherapist with individual patients, and targets for the following week were agreed between them. At the end of training, patients received advice on continuing a level and type of exercise consistent with their training programme and with individuals’ circumstances including local access to sporting and other facilities.
Assessment
Social characteristics..
Socio-demographic data gathered at recruitment comprised age, gender and self-reported employment status (paid employment, retired, homemaker, student, unemployed or not working because of illness). Postcodes were used to allocate patients to deciles of a deprivation index for Merseyside.24
Clinical data.. For each patient referred to the study, the GP was asked to indicate the ‘predominant complaint/ symptom’. These were grouped into four categories: fatigue and tiredness; emotional problems; musculoskeletal pain; and other. In addition, clinical data were systematically transcribed from GP records onto a coding sheet covering two periods (six months immediately before the start of exercise training and six months after completion of training). Information included: the number of GP consultations; number, and number of types, of prescriptions; number, and number of types, of contact with secondary care specialities; and whether or not the patient was newly referred to secondary care during this period. In addition, all symptoms noted in GP records during the two periods were extracted. Rather than adopt an existing classification of morbidity we endeavoured to preserve the language used by GPs to enter symptoms. This led to a framework of 48 categories. For common symptoms a number of categories were closely related but not necessarily identical; for example, we preserved ‘tiredness’, ‘ME’, ‘weak/washed out’ and ‘post-viral syndrome’ as distinct symptom categories. However, in order to keep the total number of categories within reasonable bounds, some infrequent descriptive symptoms were consolidated within a single category (e.g. ‘knee pain’, ‘leg pain’, ‘ankle pain’ and ‘calf pain’ were coded as ‘lower limb pain’).25 For the period of training, only numbers of consultations, prescriptions and secondary care contacts were recorded. Each was doubled for comparability with the 6 month values.
Self-report.. After the first training session, patients were asked to rate the extent to which they expected the training programme to benefit their health on a 5-point Likert scale (‘not at all’, ‘a little’, ‘somewhat’, ‘quite a lot’, ‘very much so’, scored 0–4).
Other self-report data were gathered at recruitment (T1), at the start (T2), mid-point (T3) and completion (T4) of training, and six months after training ended (T5). Measurements assessed emotional state (Hospital Anxiety and Depression Scale, HAD)26 and somatization (Modified Somatic Perception Questionnaire, MSPQ; and the somatization subscale of the Hopkins Symptom Checklist, HSCL).27,28 A rating was made by patients of the extent to which their main problem interfered with their life, using a 5-point Likert scale (‘not at all’, ‘a little’, ‘somewhat’, ‘quite a lot’, ‘very much’, scored 0–4). Perceived disability was assessed by the Health Outcomes Questionnaire: Short Form 36 (SF–36) at recruitment (T1), after training (T4) and at six months follow-up (T5).29
Adherence to training.. For each patient we recorded the number of training sessions attended, content of homework diary (number of prescribed tasks completed and number of non-prescribed aerobic episodes reported), and five pulse rates taken at regular intervals from 15–50 minutes after the start of the 3rd, 12th and 17th session. From the pulse data, a mean rate was calculated for each patient and compared with either the prescribed minimum (aerobic) or maximum (stretching and aerobic) rates. From previous evidence of the effects of exercising at the level intended in this study, aerobic fitness was not expected to increase.19 Nevertheless, as a precaution, a physiological index of fitness (VO2max [ml/min/kg body weight]) was estimated by the 3-step ergometer test30 at recruitment (T1), after training (T4) and at six months follow-up (T5).
On the basis of their long-term disability, adverse mental state and disaffection with medical care,6,18 we anticipated that patients would find difficulty in achieving high levels of attendance at training sessions, thus reducing opportunities for ongoing assessment. To minimize loss of data, non-attenders at training sessions were sent questionnaires at T3, T4 and T5. At follow-up (T5) we aimed to maximize completion of the Hospital Anxiety and Depression Scales (HADS) and patients who did not attend at this time were not only sent questionnaires by mail but also reminded up to four times by post or telephone to complete and return the HADS.
|
Statistical analysis |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
Strategy
Randomization and representativeness of data providers.. Randomization was validated by comparing groups on T1 values of each variable. Representativeness of those who attended the first exercise session were assessed by comparing attenders and non-attenders on T1 values of each variable. For each subsequent assessment time, those who attended and provided questionnaire data were compared to those who did not, using data from T2. Records of attendance at exercise sessions were tabulated as indices of adherence to training; their relationship was examined with variables measured at T1 and T2.
Changes over time and effects of training.. The outcome variables were: HADS depression and anxiety; SF–36 scales; somatization scales (HSCL, MSPQ); number of GP consultations; number, and number of types, of prescription; number of GP recorded symptoms; number, and number of types, of secondary care contacts; and whether or not the patient was newly referred to secondary care.
‘Partial intention to treat’ analyses were carried out in which all patients who started training and provided data were included. The power of repeated-measures analysis across all time-points would have been compromised by the pattern of missing data. Therefore, successive two-way ANOVA compared the groups across pairs of time-points. First, comparisons of data from T1 and T2 were used to assess baseline changes for variables measured at both times. Then, data from T2 and T5 were used to assess changes from baseline that persisted to follow-up. In addition, data from T2 (or, for variables not measured at this time, from T1) were compared to T3 data in order to detect changes apparent by mid-training, and to T4 data to detect changes apparent at the end of training. We also examined whether changes over time depended on attendance at training sessions and whether this relationship depended on the type of training programme (aerobic or stretching). Finally, to examine whether and how the different variables were interrelated, correlations between them were examined at T2 and T5.
Methods.. Relationships between categorical variables were examined by chi-square tests. Relationships between continuous variables were tested by product-moment or rank-order correlations for normally distributed and other variables, respectively. Normally distributed variables were compared between groups by t-tests; others were compared by Mann-Whitney U tests. Changes over time in normally distributed variables were assessed by two-way ANOVA, after which significant interactions were explored by post hoc t-tests. Changes in non-normally distributed continuous variables were assessed in the following way. First, change over time across both groups was tested by the Wilcoxon test. Then differences between groups on each occasion were compared by Mann-Whitney U tests. For binary variables, changes over time were assessed by the McNemar test; then differences between groups were assessed by chi-square tests.
Hierarchical multiple linear regression analysis was used to examine the relationship between predictor variables and outcome in the two programmes. Binary variables were analysed by multiple logistic regression. For each analysis, the value at six month follow-up was regressed on predictor variables entered in the order: baseline value (T2 or, for variables not measured at T2, T1); total number of sessions attended; training group; interaction of group with total sessions attended.
Analyses were by SPSS 9.0 and Genstat 5. The criterion for significance was P < 0.01.
|
Results |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
Sample characteristics
Of the 228 patients who were eligible for randomization, 107 (47%) were male and 121 (53%) female. Their mean age was 44 years (range 9–73). At recruitment, 46 (20.2%) were in paid employment, 20 (8.8%) classified themselves as unemployed, 16 (7%) as retired and 24 (10.5%) were classified as ‘other’ (e.g. homemaker, student). However, the largest group of patients were not working because of illness (n = 111, 48.7%). A further 11 patients (4.8%) had never been in paid employment. On the basis of their most recent employment, 94 (41.2% of the total sample) categorized themselves as manual workers, 35 (15.4%) as clerical or shop workers, 32 (14%) as technical or secretarial and 53 (23.2%) as managerial or professional. The remaining three patients were in full-time education (1.3%). Eighty-two patients (36%) reported being in receipt of sickness or disability allowance from their employer (23) or the state (59) because of their illness. A further 55 (24.1%) received state benefit because of low income. However, numbers in each decile of deprivation for Merseyside did not differ significantly from expected values (
2 = 11.41, d.f. = 9, P > 0.10). GPs nominated the main problem for 205 patients and, as expected, fatigue (n = 70), musculoskeletal problems (n = 52), and emotional distress (n = 30) were common categories (the remaining 53 patients being identified as having other problems). Patients randomized to aerobic and stretching groups did not differ on any demographic, clinical or psychological variable.
From the 228 patients who were randomized, 177 (78%) attended the first exercise session and completed questionnaires at this time. They did not differ on any variable from those who did not attend. Subsequent analyses are therefore based on the sample of 177 (Fig. 1
). Presenting complaints were provided by GPs for 175 (98.9%) of these patients and GP records were successfully searched for 155 (88%). Data on the HADS were provided by 136 patients (77% of those who started training) at T5, 112 (63%) at T3 and 115 (65%) at T4. There were only two isolated differences between providers and non-providers of these data on any occasion. Providers at T4 were more deprived (U = 2173, d.f. = 163, P < 0.01) and had better SF–36 physical functioning (t = 2.73, d.f. = 172, P < 0.01) at T1.
Data distribution and re-coding
All continuous variables were normally distributed except: deprivation index, self-rating of interference with life, two SF–36 scales (mental and physical role limitation), attendance at the training programme, number and types of secondary care contact. The SF–36 scales were re-coded so that all non-zero scores were given a value of one. Numbers of secondary care contacts were analysed by non-parametric statistics but were re-coded, for the purposes of logistic regression analysis only, so that all non-zero scores took a value of one.
Expectations and adherence
Patients who attended the first session of each programme expected similar levels of benefit: aerobic mean = 2.56 (SD = 0.95); stretching mean = 2.35 (SD = 0.98). The median number of sessions attended was 11 with 95 patients (53%) attending 10 or more sessions and 61 (34%) attending 15 or more sessions. There were no significant differences in level of attendance between aerobic and stretching training. The number of sessions attended was unrelated to socio-demographic characteristics, main symptom, initial expectation of benefit and baseline values of self-report and health care use variables.
Only 88 (50%) of the 177 who began training completed a homework diary (42 and 46 in the stretching and aerobic programmes, respectively). Analysis of the diaries revealed similar levels of completion of prescribed tasks in both programmes. Mean pulse rate was significantly higher during aerobic training than stretching: 111.1 (SD = 18.0) versus 86.2 (SD = 13.2); t = 7.60, d.f. = 90, P < 0.001. As intended, the majority of patients’ pulse rates (61%) were below the maximum prescribed level in the stretching programme and above the minimum prescribed level in the aerobic programme (66%). A proportion (39%) of patients’ pulse rates in the aerobic programme was above the prescribed range.
As anticipated, VO2 max remained unchanged at T5 and did not differ between groups (Table 1).
|
Changes over time
Changes over time, analysed for the 177 who began training, are presented in Figure 2
for principal variables, and means of measurement for aerobic and stretching groups at each time point are shown in Table 1.
|
Clinical data.. Health care use declined at six month follow-up and also, where measured, by the end of training. Specifically, the number of symptoms recorded by the GP declined from T2 to T5 (F = 100.04, d.f. = 1,153, P < 0.001). The number of GP consultations declined from T2 to T5 (F = 50.69, d.f. = 1,153, P < 0.001) and from T2 to T4 (F = 106.16, d.f. = 1,153, P < 0.001). The number of prescriptions declined from T2 to T5 (F = 11.95, d.f. = 1,153, P < 0.001) and from T2 to T4 (F = 39.82, d.f. = 1,153, P < 0.001); the number of types of prescription also declined from T2 to T5 (F = 13.63, d.f. = 1,153, P < 0.001). The number of secondary care contacts declined from T2 to T5 (Wilcoxon Z = –4.20, P < 0.001) and from T2 to T4 (Wilcoxon Z = –5.52, P < 0.001). The number of types of secondary care contacts also declined from T2 to T5 (Wilcoxon Z = –4.38, P < 0.001). The number of patients who were newly referred to secondary care declined from T2 to T5 (McNemar,
2 = 9.80, P < 0.01).
Self-report.
- Baseline stability. Between T1 and T2, there were no changes in somatization (MSPQ, HSCL) or depression and no interactions of time with type of allocated training programme. The only significant change was that anxiety declined (F = 8.15, d.f. = 1,175, P < 0.01).
- Psychological state and physical symptoms. In general, patients improved by T5. Improvement in emotional state, but not physical symptoms, was apparent from mid-way through training (T3). Depression declined from T2 to T5 (F = 14.02, d.f. = 1,134, P < 0.001); the decline was significant by T3 (F = 7.89, d.f. = 1,110, P < 0.01) and at T4 (F = 21.24, d.f. = 1,113, P < 0.001). Anxiety also declined from time T2 to T5 (F = 9.84, d.f. = 1,134 P < 0.01); the decline was significant by T3 (F = 12.91, d.f. = 1,110, P < 0.001) and at T4 (F = 15.18, d.f. = 1,113, P < 0.001). Somatization (HSCL) declined from T2 to T5 (F = 9.75, d.f. = 1,129, P < 0.01) but not by T3 or T4. Similarly, MSPQ scores declined from T2 to T5 (F = 6.97, d.f. = 1,129, P < 0.01) but not by T3 or T4.
- Disability. There was partial improvement after training. Specifically, there was a highly significant decline in self-rated interference with life from T2 to T5 (Wilcoxon Z = –4.38, P < 0.001), but not by T3 or T4. Several SF–36 scales also improved. Reported energy levels increased from T1 to T5 (F = 20.97, d.f. = 1,138, P < 0.001) and the increase was apparent by T4 (F = 20.81, d.f. = 1,113, P < 0.001). Mental health improved from T1 to T5 (F = 21.42, d.f. = 1,137, P < 0.001) and by T4 (F = 22.85, d.f. = 1,113, P < 0.001). Mental role limitation improved from T1 to T4 (McNemar, 2 = 11.81, P = 0.001) but the improvement was not sustained at T5. Social function improved from T1 to T5 (F = 18.63, d.f. = 1,146, P < 0.001) and by T4 (F = 7.38, d.f. = 1,127, P < 0.01). Four dimensions of the SF–36 (health perception, physical function, physical role limitation and pain) did not change.
Is outcome related to type of training?
There were no significant main effects of programme type on any variable. Furthermore, no interaction of programme type with time reached significance, indicating that patients in the two programmes did not differ at any point and changed similarly over time.
Is outcome related to attendance?
Since, in the multiple linear and logistic regression analyses, programme type did not interact with attendance on any variable, the analyses were repeated with this interaction term excluded. In general, and as would be expected, the value of each variable at T5 was strongly related to the value at T1. Controlling for this, attendance at treatment sessions predicted reduced GP attendance during the 6-month follow-up (b/ß = –0.17/–0.29, t = 4.94, d.f. = 151, P < 0.001) as well as a reduction of number of prescription items (b/ß = –0.20/–0.20, t = –3.15, d.f. = 151, P < 0.01) and the number of types of drug prescribed (b/ß = –0.10/–0.24, t = 3.66, d.f. = 151, P < 0.001). The total number of symptoms, somatization (HSCL, MSPQ), HAD and SF–36 scales and whether patients were referred or whether they had any secondary care contacts were not related to attendance.
The three health care use variables that were available for the period of training (number of GP consultations, number of prescription items, number of secondary care contacts) were analysed in a similar way. Attendance at training was unrelated to each.
Relationship between different measures of outcome
Self-reported outcome variables were inter-correlated at both T2 and T5 (66/78 and 69/78 significant correlations, respectively). Health care use variables were also inter-correlated at both times (19/21 significant correlations at T2 and 20/21 at T5). However, of the correlations linking patient and GP reported variables, only 8/84 were significant at T2 and none was significant at T5.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
A number of factors, including their mental state, perceived disability and disaffection with medical care, increase the difficulty of recruiting patients with a broad range of unexplained symptoms to a controlled study of physical exercise. Nevertheless, the present sample is the largest group of such patients with persistent symptoms so far reported from primary care in the UK. In terms of characteristics and patterns of health care use they strongly resemble patients with PUPS previously reported.4,11 Nevertheless, the conclusions which can be drawn from the present study are necessarily limited to those primary care patients with PUPS willing to be engaged in exercise programmes.
After excluding a few individuals with physical contraindications, 74% of patients identified by GPs were available for assessment and randomization into aerobic exercise or stretching training. Seventy-eight per cent of randomized patients attended their first training session, and completed a median of 11 one-hour sessions (out of a maximum of 20). Differences in populations sampled in secondary care studies, prevents direct comparison with previously reported levels of recruitment into exercise programmes. Some secondary care studies fail to provide relevant details of the selection processes that determine entry into the study,31 whilst others indicate stringent inclusion criteria that restrict the trial to patients who are most likely to engage. For example, in a trial of exercise training for patients with CFS, more than half of those screened were excluded because of insomnia or a comorbid psychiatric disorder, and several were excluded because they were too incapacitated to attend training.20 Equally, comparison of the level of attendance at exercise sessions in the present study with levels reported from secondary care,20,21,31 is likely to be misleading since greater compliance would be expected in the more highly selected samples used in those studies.
A key issue in controlled studies of aerobic exercise is the intensity of exercise undertaken by intervention and control groups. Data which enable judgements to be made about the intensity of exercise have not been reported in previous studies; one study simply reporting that ‘most patients’ did not achieve their targets.32 In contrast, we recorded heart rates during training: the majority of our patients exercised within the maximum or above the minimum heart rate parameters of the stretching and aerobic programmes respectively. Aerobic exercise was intended to keep heart rate within a very narrow range, based on previous evidence of the most beneficial level of training.23 In the event, a significant minority of the aerobic patients exceeded their prescribed maximum.
In assessing progress and outcome, we chose two groups of measures that reflect principal features of patients with PUPS: symptomatology and use of health care documented by GPs as part of ongoing care; and emotional state and disability self-reported by patients directly to us. Both documented health care use and the self-report measures improved significantly over time. However, there was no advantage of aerobic exercise training beyond that achieved by the control group trained in stretching techniques. Thus the specific benefits of aerobic exercise in CFS reported by Fulcher and White20 and in fibromyalgia by Wigers et al.31 and McCain et al.21 were not replicated among the wider group of symptoms found in primary care patients with PUPS. Possible reasons for this discrepancy include differences in recognition, referral and selection between studies of ‘specific’ syndromes emanating from specialist centres,33 differences in the level of exercise training achieved both within and between studies, and variation in the outcome measures employed. Whatever the explanation, our study provides no support for the view that aerobic exercise training is an effective specific treatment for PUPS in primary care.
Although in the absence of a ‘treatment as usual’ control arm it is impossible to say with complete certainty, the significant improvement over time in self-reported measures is unlikely to represent a continuing spontaneous recovery or regression to the mean because, for most variables, it occurred during the training period and did not continue during the six month period after training (Fig. 2). Neither was it due to increased fitness since VO2 max did not improve for either group (furthermore, previous evidence indicates that emotional benefits from exercise are not related to improvements in fitness19).
There is clearer evidence that the reduction in health care use was a product of training, because the extent of reduction depended on the level of attendance at training sessions. It cannot be explained, simply, as the training distracting patients from consulting their GP, since reduction in health care use during the six months following training, but not during the period of training, was related to attendance. Moreover, this finding is of practical significance for primary care: if the observed benefits were sustained for a full year, on average, patients with PUPS who attended 10 sessions of exercise training (less than the average achieved in the present study) would each require three fewer GP consultations.
Therefore, whereas reduction in health care use appears to be a product of attendance at training sessions, subjective improvement does not. Both intervention and control programmes offered patients distinctive explanatory models linking training with amelioration of their problems, involved group activities with fellow sufferers and provided support and counselling by physiotherapists.22 Thus, the potential existed for patients to reconceptualize or reattribute their problems and to find alternative solutions under the influence of individuals in the group and the therapists. It is plausible that the number of training sessions attended, and consequent level of exposure to these influences, proportionately reduced patients’ dependency on their GPs, while having no sustained effect on their disability and mental state.
It should not be surprising that health care use and subjective state improve independently: in the present study, these two groups of measures were minimally correlated at recruitment and not at all at follow-up. This finding is hard to reconcile with the view that health care use (including symptomatology) reflects the subjective difficulties of patients with PUPS and, specifically, with the widespread assumption among clinicians that somatization of mental illness is the underlying problem.34,35 Recent accounts of patients with unexplained symptoms suggest a complex biopsychosocial explanatory model in which individual vulnerabilities, social influences and medical care all interact.6,25,36,37 From this perspective, whereas exercise training (aerobic or otherwise) is likely to exert only an indirect effect on the problems of primary care patients with PUPS, reattribution and related cognitive approaches16,17 and empowering explanations18 that directly address medicalization appear to offer both PUPS patients and their doctors alternative and theoretically sounder approaches.
|
Acknowledgments |
|---|
We are very grateful for the co-operation of patients, GPs and practice staff. The Department of Physiotherapy, University of Liverpool provided space and facilities for the training programmes, for which we are indebted to Ms E Thornton. The study was funded by the Medical Research Council.
|
References |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionReferences |
|---|
1 Blackmore RA. Treatise of the Spleen and Vapours. Kent: Doddington Press, first published 1725, facsimile edition 1985.
2 Mettler L. Aboulia in relation to hysteria. J Am Med Ass 1898; 31: 1200–1203.
3 Mott F. Neurasthenia: the disorders and disabilities of fear. Lancet 1918; ii: 608–613.
4 Shorter E. From paralysis to fatigue: a history of psychosomatic illness in the modern era. New York: Free Press, 1992.
5 Verhaak PFM, Tijhuis MAR. The somatizing patient in general practice. Int J Psychiatry Med 1994; 24: 157–177.[ISI][Medline]
6 Peters S, Stanley I, Rose M, Salmon P. Medically unexplained symptoms: sources of patients’ authority and implications for demands on medical care. Soc Sci Med 1998; 45: 559–563.
7 Colgan S, Creed F, Klass H. Symptom complaints, psychiatric disorder and abnormal illness behaviour in patients with upper abdominal pain. Psychol Med 1988; 18: 887–892.[ISI][Medline]
8 Creed F, Firth D, Timol M, Metcalfe R, Pollock S. Somatization and illness behaviour in a neurology ward. J Psychosom Res 1990; 34: 427–437.[ISI][Medline]
9 Simon G, Gater R, Kisely S, Piccinelli M. Somatic symptoms of distress: an international primary care study. Psychosom Med 1996; 58: 481–488.
10 O’Dowd T. Five years of heartsink patients in general practice. Br Med J 1988; 297: 528–530.[ISI][Medline]
11 Smith GR. The course of somatization and its effects on the utilization of health care resources. Psychosom 1994; 355: 263–267.
12 Muris J, Starmans R, Fijten G, Knottnerus J. One-year prognosis of abdominal complaints in general practice: a prospective study of patients in whom no organic cause is found. Br J Gen Pract 1996; 46: 715–719.[ISI][Medline]
13 Gureje O, Simon GE. The natural history of somatization in primary care. Psychol Med 1999; 29: 669–676.[ISI][Medline]
14 O’Brien M. Medically unexplained neurological symptoms. Br Med J 1998; 316: 564–565.
15 Crimlisk H, Bhatia K, Cope H, David A, Marsden C, Ron M. Slater revisited: 6 year follow-up study of patients with medically unexplained motor symptoms. Br Med J 1998; 316: 582–586.
16 Goldberg D, Gask L, O’Dowd T. The treatment of somatization: teaching the skills of reattribution. J Psychosom Res 1989; 33: 689–695.[ISI][Medline]
17 Speckens A, van Hermert A, Spinhoven P, Hawton K, Bolk J, Rooijmans H. Cognitive behavioural therapy for medically unexplained symptoms: a randomized controlled trial. Br Med J 1995; 311: 1328–1332.
18 Salmon P, Peters S, Stanley I. Patients’ perceptions of medical explanations for somatisation disorders: qualitative analysis. Br Med J 1999; 318: 372–376.
19 Salmon P. Effects of physical exercise on anxiety, depression and sensitivity to stress: a unifying theory. Clin Psychol Rev 2001; 21: 33–62.[ISI][Medline]
20 Fulcher K, White P. Randomized controlled trial of graded exercise in patients with chronic fatigue syndrome. Br Med J 1997; 314: 1647–1652.
21 McCain G, Bell D, Mai F, Halliday P. A controlled study of the effects of a supervised cardiovascular fitness training program on the manifestations of primary fibromyalgia. Arthritis Rheum 1988; 31: 1135–1141.[ISI][Medline]
22 Crook P, Stott R, Rose M, Peters S, Salmon P, Stanley I. Adherence to group exercise: physiotherapist-led experimental programmes. Physiotherapy 1998; 84: 366–372.
23 Steptoe A, Edwards S, Moses J, Mathews A. The effects of exercise training and perceived coping ability in anxious adults from the general population. J Psychosom Res 1989; 33: 337–348.
24 Department of the Environment. The 1991 deprivation index: a review of approaches and a matrix of results. London: HMSO, 1995.
25 Stanley I, Peters S, Salmon P. A primary care perspective on prevailing assumptions about persistent medically unexplained physical symptoms. Int J Psychiat in Med 2002; 32: 125–140.
26 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatrica Scand 1983; 67: 361–370.[ISI][Medline]
27 Main CJ. The modified somatic perception questionnaire. J Psychosom Res 1983; 27: 503–514.[ISI][Medline]
28 Derogatis LR, Lipman RS, Rickels K, Uhlenhuth EH, Covi L. The Hopkins symptom checklist: a self-report symptom inventory Behav Sci 1971; 19: 1–15.
29 Ware JE, Snow KK, Kosinski M, Gandek B. SF–36 health survey: manual and interpretation guide. Boston, MA: The Health Institute; New England Medical Center, 1993.
30 Lange Anderson K, Shephard RJ, Denolin H, Varnauskas E, Masironi R. Fundamentals of Exercise Testing. Geneva: World Health Organization, 1971.
31 Wigers SH, Stiles TC, Vogel PA. Effects of aerobic versus stress management treatment in fibromyalgia. Scand J Rheumatol 1996; 25: 77–86.[ISI][Medline]
32 Nøregrad J, Lykkegaard JJ, Mehlsen J, Danneskiold-Samsøe. Exercise training in treatment of fibromyalgia. J Musculoskeletal Pain 1997; 5: 71–79.
33 Fitzgibbon E, Murphy D, O’Shea K, Kelleher C. Chronic debilitating fatigue in Irish general practice: a survey of general practitioners’ experience. Br J Gen Pract 1997; 47: 618–622.[ISI][Medline]
34 Wright A. A study of the presentation of somatic symptoms in general practice by patients with psychiatric disturbance. Br J Gen Pract 1990; 40: 459–463.[ISI][Medline]
35 Craig T, Boardman A. Somatisation in primary care settings. In Bass CM (ed.). Somatisation: Physical symptoms and psychological illness. Oxford: Blackwell, 1990.
36 Wessely S, Nimnuan C, Sharpe M. Functional somatic syndromes: one or many? Lancet 1999; 354: 936–939.[ISI][Medline]
37 Barsky AJ, Borus JF. Functional somatic syndromes. Ann Intern Med 1999; 130: 910–921.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||