|
Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls. 2: Controlled trial in multiple centres |
Robertson M C, Gardner M M, Devlin N, McGee R, Campbell A J |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology The use of a home-based exercise programme, as delivered by a trained nurse from general practice, for preventing falls and injuries among elderly people. The nurses were supervised by a physiotherapist and they attended team meetings to discuss their progress. The intervention consisted of muscle strengthening and balance retraining exercises of progressive difficulty, and a walking plan. All programmes were prescribed individually.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised the general population of elderly people aged 80 years and older. Patients who were unable to walk around their own residence, receive physiotherapy or to understand the trial requirements, were excluded from the study.
Setting The setting was the community. The economic study was carried out in seven towns in New Zealand.
Dates to which data relate Patient enrolment took place in 1998 and the follow-up for the outcome measures lasted one year (for data on both effectiveness and resource use). The price year was 1998.
Source of effectiveness data The effectiveness evidence came from a single study.
Link between effectiveness and cost data The costing was performed prospectively on the same patient sample as that used in the effectiveness study.
Study sample Power calculations were based on the proportion of elderly people who fell once or more in an earlier study, an expected reduction from 0.50 to 0.30 in the proportion of patients falling, an allowance for a multi-centre design, and a 20% drop-out rate. The patients were identified from computerised registers at general practices in the study area, and they received a letter from their doctor inviting them to participate in the trial. Of an initial sample of 989 patients invited to participate, the study sample used in the effectiveness study comprised 450 patients (120 in the control group and 330 in the intervention group). The mean age in the control group was 84.2 (+/- 3.1) years, 29% were men, and 46% had fallen in the previous year. The patients in the intervention group were split into three groups depending on the enrolment centre.
At centre 1 there were 115 patients. The mean age was 82.7 (+/- 2.5) years, 28% were men, and 41% had fallen in the previous year.
At centre 2 there were 120 patients. The mean age was 83.6 (+/- 2.8) years, 31% were men, and 45% had fallen in the previous year.
At centre 3 there were 95 patients. The mean age was 84.1 (+/- 3.0) years, 34% were men, and 56% had fallen in the previous year.
Study design This was a non-randomised controlled trial, which was carried out in seven towns in New Zealand. The exercise programmes were conducted in three towns by one general practice in each, by one nurse. The patients were followed for one year. Fourteen patients in the control group and 65 patients in the intervention group were lost to follow-up, mainly due to death or health problems. The assessors who collected and classified the data on falls were blinded to the patients' treatment group.
Analysis of effectiveness The basis of the analysis of the clinical study was intention to treat. However, it was not stated what outcomes were attributed to the drop-outs. The primary health outcome estimated in the analysis was the number of falls and injuries. The incidence rates, injurious falls and falls requiring medical care were also recorded. Incidence rate ratios were calculated to compare the incidence rates between the intervention and the control groups. This was performed through negative binomial regression models, taking into account patient clustering. The SF-12 questionnaire was used to estimate self-perceived health status at entry to the trial. Falls were defined as "unintentionally coming to rest on the ground, floor, or other lower level" and were monitored for one year by asking participants to return pre-addressed and pre-paid postcard calendars every month. An independent assessor then telephoned the participants to record the circumstances of the falls. Fall events were classified as serious, moderate, and no injury. Serious falls were those resulting in fracture, admission to hospital with injuries, or a requirement for stitches. Moderate falls were those where patients required medical help, or where bruising, sprains, cuts, abrasions, or a reduction in physical function of at least three days occurred. Compliance was evaluated using postcard calendars similar to those used to monitor falls. The study groups were well matched at baseline, but more participants from the control centres (88%) than the exercise centres (80%) completed the trial, (p=0.066).
Effectiveness results After a mean follow-up period of 11.2 months in the control group, the number of falls was 105. For the intervention groups, after 10.4 months in group 1, 11.1 months in group 2 and 10.3 months in group 3, the number of falls was 57 (group 1), 87 (group 2) and 54 (group 3), respectively. The resulting incidence rate ratio was 0.70 (95% confidence interval, CI: 0.59 - 0.84; p=0.001). This meant that, during the trial, there was a 30% reduction in falls in the intervention groups as a whole compared with the control group.
The falls per 100 person years were 93.9 in the control group, 57.4 in intervention group 1, 78.4 in intervention group 2, and 66.4 in intervention group 3.
The number of injurious falls was 46 (2 serious and 44 moderate) in the control group, 27 (3 serious and 24 moderate) in intervention group 1, 37 (9 serious and 28 moderate) in intervention group 2, and 22 (4 serious and 18 moderate) in intervention group 3. The incidence rate ratio was 0.72 (95% CI: 0.62 - 0.82; p=0.001).
The injurious falls per 100 person years were 41.1 in the control group, 27.2 in intervention group 1, 33.3 in intervention group 2, and 27.1 in intervention group 3.
The number of falls requiring medical care was 24 (23%) in the control group, 14 (25%) in intervention group 1, 24 (28%) in intervention group 2, and 9 (17%) in intervention group 3.
Clinical conclusions The effectiveness study showed that the nurse-delivered home programme was highly effective in reducing the number of falls and serious injuries in elderly people.
Measure of benefits used in the economic analysis The benefit measure used in the economic analysis was the number of falls occurring in the study groups.
Direct costs No discounting was performed since the costs were incurred over one year. The unit costs were reported separately from the quantities of resources used. The health service costs included in the economic evaluation were the training course costs, recruitment, programme prescription, follow-up costs, supervision programme costs (time for the physiotherapist and exercise nurse) and overhead costs. The training course costs were for nurse and physiotherapist time, accommodation and materials. The follow-up costs were for the general practitioner and nurse's time, telephone calls and postage. The cost of hospitalisation due to falls was not included, as no statistically significant differences between the two study groups were found in terms of these hospital admissions. Research costs were also not included in the economic evaluation. The cost/resource boundary adopted in the study was unclear. The resource use was estimated from data collected alongside the clinical trial, while the unit costs were estimated from actual data such as hospital and trial records. The costs were estimated for 1998 to 1999 and 1998 prices (exclusive of government goods and service tax) were used.
Statistical analysis of costs No statistical analyses of the costs were performed.
Indirect Costs The indirect costs were considered to be zero since the exercises were assumed to take place during the patients' leisure time.
Currency New Zealand dollars (NZ$). The average exchange rate in 1998 was NZ$1 = 32 p.
Sensitivity analysis One-way sensitivity analyses were conducted to assess the robustness of the estimated cost-effectiveness ratios to variations in the baseline costs. These were conducted using the 125th and the 75th percentiles of the total programme costs, assuming that training and supervision took place in the same centre (then excluding both travel and accommodation costs). Also, using the 125th percentile of home visit costs, quadrupling the ankle cuff weight costs, assuming no extra overhead costs.
Estimated benefits used in the economic analysis See the 'Effectiveness Results' section.
Cost results The costs of implementing the programme were NZ$137,878 or NZ$418 per person. Since the programme was additional to existing practice, the cost of the alternative option of 'no programme' was zero.
Synthesis of costs and benefits An incremental cost-effectiveness ratio was calculated to combine the costs and benefits of the exercise programme, compared with standard care. The incremental cost per fall prevented was NZ$1,519 (adjusted ratio, using fall events per 100 person years: NZ$1,734). The incremental cost per injurious fall prevented was NZ$3,404 (adjusted ratio: NZ$3,846). The sensitivity analyses showed that the cost-effectiveness ratio per injury prevented ranged from NZ$2,553 to NZ$4,255.
Authors' conclusions Falls can be successfully reduced in men and women aged at least 80 years, using an exercise programme delivered by trained nurses based in general practices. However, the value for money of such an intervention was unclear, as the hospital costs were not reduced in comparison with standard care.
CRD COMMENTARY - Selection of comparators The rationale for the choice of the comparators was clear. Standard care with no additional programme was selected as it represented the routine care for elderly patients. You should decide whether it represents a valid comparator in your own setting.
Validity of estimate of measure of effectiveness The analysis of the effectiveness used a non-randomised controlled trial in which the intervention and control were tested in different towns. The authors said this avoided contamination from the sharing of information. However, it may also have resulted in bias due to different conditions in different towns. The study sample was representative of the study population. Power calculations were based on an earlier prospective study. The length of follow-up was reported, as well as the number of patients who were lost to follow-up assessment. The study groups were comparable at baseline. The basis of the analysis was intention to treat, although it was not stated what outcomes were attributed to the drop-outs. Statistical analyses were also conducted in the effectiveness study. These issues enhanced the internal validity of the analysis. However, blinding was limited to event classification and the authors acknowledged that some assessment bias might have occurred, although its impact was likely to have been minimal. A short-term effectiveness analysis would have been more helpful.
Validity of estimate of measure of benefit The benefit measure used in the economic analysis was the number of falls. This was estimated in the effectiveness study. No measures of utility and patient satisfaction were used, although these would have been helpful.
Validity of estimate of costs The perspective adopted in the study was clearly stated and a wide range of costs was included in the analysis. You should decide whether these cost categories are appropriate for your setting. The unit costs were reported separately from the quantities of resources used and a detailed breakdown of the costs was provided. The price year was reported, thus facilitating reflation exercises in other settings. The costs were treated deterministically in the base-case analysis, but several sensitivity analyses were conducted. The authors stated that the distribution of costs was skewed and this could have led to high costs in the intervention group.
Other issues The authors did not compare their findings with those from other studies, but the present study was based on the results of an earlier study conducted by the same authors. The study referred to a sample of elderly people and this was reflected in the conclusions of the analysis. Although sensitivity analyses were performed, the authors did not explicitly address the issue of the generalisability of the study results to other settings. The authors commented on some potential limitations of their analysis. First, they acknowledged that the duration of follow-up might have been insufficient since, in a previous trial, they found that the reduction in falls continued for two years and involved very little extra use of resources.
Implications of the study The authors recommend the implementation of a patient-tailored, home-based exercise programme delivered to elderly people by trained nurses from within general practice.
Source of funding Funded by the Accident Rehabilitation and Compensation Insurance Corporation of New Zealand; the Health Research Council; the Lottery Grants Board, New Zealand; and the Trustbank Otago Community Trust.
Bibliographic details Robertson M C, Gardner M M, Devlin N, McGee R, Campbell A J. Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls. 2: Controlled trial in multiple centres. BMJ 2001; 322: 701-704 Other publications of related interest Robertson M C, Devlin N, Gardner M M, Campbell A J. Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls. 1: Randomised controlled trial. BMJ 2001;322:697-701.
Indexing Status Subject indexing assigned by NLM MeSH Accidental Falls /economics /prevention & Aged; Aged, 80 and over; Cost-Benefit Analysis; Costs and Cost Analysis; Exercise Therapy /economics /methods; Female; Follow-Up Studies; Home Care Services /economics; Humans; Male; Nursing Care; control AccessionNumber 22001008054 Date bibliographic record published 31/07/2003 Date abstract record published 31/07/2003 |
|
|
|