Recruitment

The population of the catchment area comprised women who were in a local FM association. Eligible women were those who had FM diagnosed by a rheumatologist in accordance with the diagnostic criteria of the American College of Rheumatology (ACR) [1]. A total of 40 potentially eligible participants responded and sought further information (Figure 1). Once the study protocol was explained, 38 people gave their written informed consent. The following exclusion criteria were applied: history of severe trauma, frequent migraines, peripheral nerve entrapment, inflammatory rheumatic diseases, severe psychiatric illness, other diseases that prevent physical loading, pregnancy, participation in another psychological or physical therapy programme, or engaging in regular physical exercise more than once a week for 30 min or longer during a 2-week period in the last 5 years. The participants in our study of a 12-week aquatic programme [11] were excluded from the current trial to avoid the influence of re-training. Participants' clinical conditions were checked and a rheumatologist confirmed the diagnosis of FM. After excluding 5 candidates due to their participation in other therapies, 33 female patients, aged 37 to 71 years of age, were selected to participate. They were randomly assigned to either the exercise group (EG; n = 17) or a control group (CG; n = 16). Two patients in the EG failed to attend at least 95% of the treatment sessions due to personal reasons. Nevertheless, these patients were included in the current study to apply an intent-to-treat analysis. The trial was exclusively developed and performed at the facilities of the University of Extremadura, Spain, with the approval of the Committee on Biomedical Ethics of the University and following the updates of the Declaration of Helsinki.

Study design

Interventions

Usual care and the addition of a water-based exercise programme were compared in the CG and EG, respectively. The usual care included standard medical attention in the public system (hospital and outpatient clinic including primary care) and the social support of the local FM association. This care could be considered the average standard of care or better for patients with FM.

The intervention added an exercise programme in a in a waist-high pool of warm water (33°C). A qualified exercise leader instructed and trained the intervention group three times a week for 1 h per session over a period of 8 months. Each session included 10 min of warming up with slow walking and easy movements of progressive intensity, 10 min of aerobic exercises at 60–65% of maximal heart rate, 20 min of overall mobility and lower limb strength exercises using water resistance, another set of 10 min of aerobics at 60–65% of maximal heart rate, and 10 min of cooling down with low intensity exercises. Heart rate was monitored using a pulse meter (Polar Accurex Plus, Kempele, Finland). During this 8-month period participants in the control group continued their daily activities, which did not include any form of physical exercise similar to that in the programme. This programme was designed without reference to any explicit behavioral model or theory, and was intended as a pragmatic intervention that could be easily organized for a large population.

Data collection

Participants completed questionnaires, including the EuroQol EQ-5D health status instrument [17] at the beginning of the programme and after 3 and 8 months. During the same period, private and public health care was recorded, including hospital stays, drug usage, secondary and primary care appointments.

Unit costs

The expense and time needed for travel from the patient's residence to the rehabilitation pool varied, because this facility is a scarce health resource serving a large area. To allow for a range in such additional costs, we performed two economic analyses, one from a health service perspective and another from a societal perspective. The first perspective is recommended by the National Institute for Clinical Excellence (NICE) in the UK to inform decisions on health care policy for an expensive condition. This perspective could help to decide whether to finance the addition of the programme to the health system. The second perspective is recommended to consider the combination of the burden to the patient and the health care system. The unit costs are expressed in Euros (€) based on prices in 2005.

Costs were not adjusted or discounted for changes in currency value over time, as we focused solely on effects over less than 1 year. The programme's cost was calculated based on the following: salaries at the level for a university graduate, cost of staff to run the programme, salaries at minimum wage for the patient's time (based on the 2005 official bulletin of the regional government), cost of renting a pool at a university at public prices without a grant, public bus prices, and private external management costs of the programme (insurance, monthly retrievals from patients and withdrawals to employees). Health care prices (consultations, etc.) were based on the 2005 official bulletin of the regional government. Drug prices were obtained from the Spanish version of Vademecum International [18].

Health outcomes

The EQ-5D [17] was used to assess five dimensions of health related quality of life: (1) mobility, (2) self-care, (3) daily activities, (4) pain and discomfort, and (5) anxiety or depression. The scale for each dimension is from 1 to 3 (with 1 no problems, 2 some problems, and 3 extreme problems). Using a combination of these dimensions, a total of 243 possible health states exist. Each health state has been previously defined using the time trade-off method of utility analysis based on the response of a sample of the Spanish population [19]. This total score of utility was scaled from 1 = fully functional quality of life to 0 = death. The quality-adjusted life years (QALYs) that participants experienced over the 8-month period were estimated by calculating areas under health utility curves [20]. To avoid bias, data were adjusted by regression analysis for differences in baseline EQ-5D scores [21].

Cost utility analysis

First, we estimated the incremental mean costs of the water-based programme and the mean QALYs added by the programme from a health care and societal perspective. Secondly, the incremental cost effectiveness ratio for the water-based programme was calculated by dividing the incremental costs by incremental QALYs.

To report the uncertainty due to sampling variation, we calculated the 95% confidence interval using the non-parametric bootstrapping technique (1,000 replicates re-sampled with replacement from treatment and control populations) and plotted a cost effectiveness acceptability curve [22, 23]. This curve shows the probability that the intervention is cost effective compared with the alternative, across the range of values that decision makers are willing to pay to achieve an additional QALY. The "investment ceiling" is the level of spending that should not be exceeded, even assuming unlimited funding availability. For the health care system in Spain, the 2005 adjusted investment ceiling was set at € 34,729/QALY [24]. Decision makers should compare this upper limit of acceptable payment with estimated incremental cost effectiveness ratios to determine whether a given treatment is cost effective relative to the alternatives.

For the health system and societal perspectives, seven sensitivity analyses were performed to explore the robustness of the estimates and how dependent the results were on estimates of participants' unit costs and efficacy. From the health system perspective, the first analysis examined the influence of participation rate in the programme as this could influence the productivity by affecting the number of participants per unit of time provided by the technician. A second analysis explored the variations due to the salary changes of the technician, since this is a major source of variability in economic studies [25]. From a societal perspective, in addition to two previous analyses the third analysis estimated the cost of increasing the mean distance (in terms of time spent and the number of bus tickets purchased) from the patient's residence to the rehabilitation pool. Finally, from both perspectives, the robustness of cost effectiveness was examined by exploring scenarios combining the influence of the variations in staff salary, rate of participation, distance to the facility and effectiveness, from the lowest to the highest limit of the 95% confidence interval.

Costs

Health outcome

Table 3 shows that the water-based programme was associated with a greater increase in the EQ-5D utilities than the usual care during the first 3 months and this difference was preserved during the subsequent 5 months.

Cost utility analysis

Table 3 shows the cost utility analyses from both perspectives. From the health service perspective, the Spanish Health System Efficiency Threshold was set at € 23,153/QALY for 8 months by multiplying the published threshold of 34,729 for 12 months by 8/12 [24]. From the health service perspective, each additional QALY gained by the exercise group cost in average € 3,947. However, the cost effectiveness acceptability curves (Figure 2) showed a 95% probability that the addition of the water-based programme is a cost effective strategy if the ceiling of inversion is € 14,200/QALY and a 97.5% probability if the ceiling is set at € 21,233/QALY.

From a societal perspective, the mean cost per QALY was € 7,878/QALY and there is a 95% probability that the addition of the aquatic training is cost effective if the ceiling of inversion is € 28,300/QALY. A 97.5% probability requires an inversion higher than € 42,000/QALY.

Sensitivity analysis

Principal findings

Previous studies reported the efficacy of aquatic training on patients with fibromyalgia [11–15, 26, 27] and the cost-utility of a 2.5 week spa treatment [28], but to our knowledge the present study is the first to report cost-utility. The major finding of this study was that the water-based programme was a cost-effective addition to usual care from both health system and societal perspectives. More precisely, an investment in this aquatic training for a similar population (sedentary women with FM) has a greater than 95% probability of being efficient according to the investment ceiling in Spain.

Strengths and weakness

The acceptable efficiency threshold, investment ceiling or maximum willingness to pay for each gained QALY varies among countries or societies because of differences in salaries, priorities, etc. The current study applied the commonly lower threshold of € 34,729 (€ 23,153 for 8 months) used in the Spanish literature [24], but similar conclusions about the efficiency of the addition of aquatic training to usual care could be achieved using the threshold updated to year 2005 (annual inflation of 5%) often reported in American literature ($ 50,000 to $ 60,000) or Dutch literature (€ 28,940) [29].

The retention rate of patients in the our programme (88%) was similar to rates previously reported in community group-based exercise programmes in fibromyalgia (70–90%); however, aquatic training programmes usually report lower retention rates (55–75%) [9]. The social support provided by physicians, research teams, and peers with FM from the local association may have contributed to this high retention rate and the improvement in the psychosocial dimensions of health related quality of life and QALY in the exercise group. Particularly, the patient's affiliation with the local FM association brought them additional care (social support, information, etc.) in comparison to what is offered by the Public Health Care System. In this sense, the care received by the control group could be considered better than usual. By contrast, care that combines the study programme with other therapies may be even better than the programme alone. This issue could not be addressed in the current study because patients were excluded if they used other therapies (standardized behavioral or physical therapies such as massages, etc.).

The small sample size led us to use non-parametric bootstrapping techniques to treat the confidence intervals and probability curve. Health economists recommend bootstrapping techniques, rather than standard deviation-based methods, for treating the uncertainty of cost-effectiveness ratios [22, 23, 30–33]. The small sample, the fact that subjects were self-selected according to bioethics requirements and the catchment throughout local patient associations may limit the generalization of our findings to treatment of less motivated patients.

Use of health care

The current study did not find any evidence for decreased use of health care services during the study period. However, the lack of change in the ratio of frequency (consultations/month) can be explained partially by the limits of supply and the management of free appointments in the general practices of the National Health System in Spain. A study in a non-limited supply setting could address the question of whether an aquatic programme could reduce the use of other health care services.

The increase in the medication cost in both groups may be partly explained because the perception of pain is slightly increased in the summer in persons with fibromyalgia [34]; with a change in the average temperature in Extremadura from 14°C at baseline to 22°C at the end of programme.

By contrast, the aquatic training in facilities with warm water was a cost-effective addition to usual care but it was not compared to other physical therapies that could reduce geographic inequalities (e.g., land-based therapies such as low-impact aerobics, walk-based exercise, tai chi, etc.) because their facilities are cheaper and easier available in more municipalities.