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.

Three interventions for the treatment of lateral epicondylitis in adult patients were examined. These were corticosteroid injections, physiotherapy and wait-and-see. With corticosteroid injections, patients were given a single local corticosteroid injection of 1 mL triamcinolone (10 mg/mL) and 1 mL lidocaine 2% on up to three separate occasions. The patient was injected until free of pain, with a maximum of 2 mL per injection. Physiotherapy consisted of treatment with pulsed ultrasound (2 W/cm2, applied for 7.5 minutes), deep friction massage and an exercise programme, over a 6-week period. Treatment sessions of approximately 30 minutes were scheduled twice a week, with a maximum of nine treatment sessions. Under a wait-and-see policy, patients visited their general practitioner (GP) once during the 6-week intervention period. During this visit, the GP discussed with the patient any activities that caused pain and disability. Medication (paracetamol or non-steroidal anti-inflammatory drugs) was prescribed, if necessary.

Treatment.

Cost-utility analysis and cost-effectiveness analysis.

The study population comprised patients aged between 18 and 70 years with pain at the lateral side of the elbow for at least 6 weeks, and with pain increasing with pressure on the lateral epicondyl and during resisted dorsiflexion. Patients were excluded if they had received physiotherapy or corticosteroid injections for lateral epicondylitis in the last 6 months, surgery of the elbow, or had bilateral elbow symptoms. They were also excluded if there were contraindications for corticosteroid injections, or if there was evidence of any specific pathology (e.g. malignancy, fracture or inflammation).

The setting was primary care. The economic study was carried out in The Netherlands.

Patients were enrolled between September 1997 and October 1998, during which time the effectiveness and resource use data were presumably gathered. The price year was 1999.

The effectiveness evidence was derived from a single study.

The costing was carried out prospectively on the same sample of patients as that used in the effectiveness study.

Power calculations were not reported. A sample of 185 patients was enrolled. There were 62 patients in the injection group, 64 patients in the physiotherapy group, and 59 patients in the wait-and-see group. There was no information on either the characteristics of the patients or patients who refused to participate in the study.

This was a randomised clinical trial that was carried out at 85 primary care centres in two cities in The Netherlands. The length of follow-up was one year. Only two patients in the injection group were lost to the follow-up assessment. Thus, the rate of complete follow-up was 99%.

The analysis of the clinical study was conducted on an intention to treat basis. The four clinical measures used in the analysis were general improvement, pain during the day, elbow disability, and utility. Patients rated their general improvement on a 6-point ordinal scale, ranging from "much worse" to "completely recovered". Success was defined as "much improved" or "completely recovered". The patients rated their pain during the day, in the week preceding their periodic assessment, on an 11-point numerical scale (0 indicating no pain). Elbow disability was measured using the modified Pain Free Function Questionnaire (PFFQ), which consists of common situations that might cause elbow pain (0 indicating no pain). All outcome values for the pain scale and the modified PFFQ were transformed into a 100-point scale to facilitate interpretation. Quality of life was measured with the EuroQol and expressed as utility values ranging from 0 to 1, where 1 represented perfect health. All outcomes were assessed using self-reported patient questionnaires at baseline and 3, 6, 12, 26 and 52 weeks. The baseline comparability of the study groups was not discussed.

The success rate after 52 weeks was 91% in the physiotherapy group, which was significantly higher compared with that in the injection group (69%) but not compared with that in the wait-and-see group (83%).

The difference in success rates between the wait-and-see policy and injections did not reach statistical significance.

General improvement was 0.90 with physiotherapy, 0.83 with the wait-and-see policy, and 0.71 with corticosteroid injections.

The pain during the day was 39 (+/- 26) in the wait-and-see group, 35 (+/- 26) in the injection group, and 46 (+/- 28) in the physiotherapy group.

The values of elbow disability were 35 (+/- 21) in the wait-and-see group, 27 (+/- 23) in the injection group, and 40 (+/- 22) in the physiotherapy group. These differences were statistically significant in favour of physiotherapy compared with injections, but not for physiotherapy versus the wait-and-see policy or injections versus the wait-and-see strategy.

The utility values were 0.81 (+/- 0.12) in the wait-and-see group, 0.78 (+/- 0.14) in the injection group, and 0.82 (+/- 0.14) in the physiotherapy group. Differences in the utility values were not statistically significant between groups.

The effectiveness analysis showed that physiotherapy was the most effective strategy. Physiotherapy was in general significantly more effective than corticosteroid injections, while the trend in better outcomes compared with wait-and-see did not reach statistical significance.

The summary benefit measures used were the four clinical end points that were estimated from the clinical trial (i.e. general improvement, pain during the day, elbow disability, and utility).

The analysis took the viewpoint of society and included direct medical and non-medical costs. The direct medical costs referred to the interventions (i.e. number of sessions with a physiotherapist or visits to a GP, number of corticosteroid injections), additional visits to a health care provider (GP, physiotherapist, medical specialist or outpatient care, other healthcare professionals), prescribed medications, professional home care, diagnostic interventions and hospitalisations. The direct non-medical costs included out-of-pocket expenses (i.e. over-the-counter medication) and the costs of paid and unpaid help. The unit costs were presented separately from the quantities of resources used for all items. The resource use data were derived from the sample of patients included in the clinical trial using standardised forms and cost diaries. The costs came from multiple sources, including published Dutch guidelines, tariffs of the Dutch Central Organisation for Health Care Charges, fees charged by professional organisations, and the Royal Dutch Society for Pharmacy. Travel costs were estimated using shadow prices. Discounting was not relevant since the costs per patient were incurred during one year. The price year was 1999.

Bootstrapping (2,000 replications) was used for pair-wise comparison of the mean costs. Confidence intervals for differences were calculated.

The indirect costs (i.e. productivity losses associated with absenteeism paid and unpaid labour) were included in the analysis since a societal perspective was adopted. The unit costs were presented separately from the quantities of resources used. Resource consumption came from the sample of patients included in the clinical trial. The costs of paid work were estimated using mean income of the Dutch population, according to age and gender of employees, on the basis of the friction cost method. Unpaid work was estimated using a shadow price. As in the analysis of the direct costs, discounting was not applied and the price year was 1999.

Euros (EUR). The exchange rate from Euros into US dollars ($) was EUR 1 = $0.90.

Sensitivity analyses were carried out using bootstrapping (5,000 replications) for cost-effectiveness and cost-utility ratios, which were represented on cost-effectiveness planes. Cost-effectiveness acceptability curves were then generated to determine the probability that a treatment is cost-effective given a specific ceiling ratio.

See the 'Effectiveness Results' section.

The cost analysis showed that there were no statistically significant differences in the consumption of resources among treatment groups.

The direct health care costs were EUR 56 (+/- 100) in the wait-and-see group, EUR 143 (+/- 187) in the injection group, and EUR 214 (+/- 92) in the physiotherapy group.

The direct non-health care costs were EUR 57 (+/- 182) in the wait-and-see group, EUR 125 (+/- 379) in the injection group, and EUR 96 (+/- 101) in the physiotherapy group.

The indirect costs were EUR 518 (+/- 1,549) in the wait-and-see group, EUR 164 (+/- 507) in the injection group, and EUR 612 (+/- 2,456) in the physiotherapy group.

The total costs were EUR 631 (+/- 1,627) in the wait-and-see group, EUR 430 (+/- 872) in the injection group, and EUR 921 (+/- 2,468) in the physiotherapy group.

The mean total costs in the injection group were 53% (EUR 491) lower than in the physiotherapy group. This difference was statistically significant.

The mean total costs in the wait-and-see group were 47% (EUR 201) higher than in the injection group and 31% (EUR 290) lower than in the physiotherapy group, but none of these differences reached statistical significance.

Incremental cost-effectiveness and cost-utility ratios were calculated to combine the costs and benefits of the alternative treatments.

The additional cost per additional 1-point improvement on the ordinal General Improvement Scale was EUR 2,035 with wait-and-see over injections, EUR 4,675 with physiotherapy over wait-and-see, and EUR 3,089 with physiotherapy over injections.

The additional cost per additional 1-point improvement in pain during the day was EUR 43 with wait-and-see over injections, EUR 64 with physiotherapy over wait-and-see, and EUR 53 with physiotherapy over injections.

The additional cost per additional 1-point improvement on the PFFQ was EUR 29 with wait-and-see over injections, EUR 72 with physiotherapy over wait-and-see, and EUR 46 with physiotherapy over injections.

The additional cost per additional utility gain was EUR 6,807 with wait-and-see over injections, EUR 34,461 with physiotherapy over wait-and-see, and EUR 12,158 with physiotherapy over injections.

The cost-effectiveness acceptability curve showed that with a ceiling ratio of EUR 20,000 per improvement in outcome, the probability that physiotherapy was cost-effective in comparison with injection was approximately 80%. Less clear indications for the comparison between injections versus wait-and-see and physiotherapy versus wait-and-see were observed.

Physiotherapy showed better clinical effects, whereas a wait-and-see strategy was less costly but still produced clinical benefits in comparison with corticosteroid injections. Therefore, no strong conclusion could be reached on the optimal strategy for the treatment of lateral epicondylitis.

The rationale for the selection of the comparators was clear. Wait-and-see was the treatment option recommended by the Dutch College of General Practitioners, while physiotherapy and corticosteroid injections were also commonly used for the treatment of lateral epicondylitis in adult patients. You should decide whether they are valid comparators in your own setting.

The effectiveness evidence came from a published clinical trial, thus limited information on the design and other characteristics of the study was provided. In general, the use of a randomised trial has a high internal validity, owing to the reduced impact of selection bias and confounding factors. An intention to treat analysis of the clinical study and the limited loss to follow-up enhance the robustness of the effectiveness study. However, the reader is referred to the primary trial to assess the validity of the clinical estimates.

The summary benefit measures were specific to the disease considered in the study. They are not comparable with the benefits of other health care interventions. They were derived directly from the effectiveness study. The impact of the interventions on quality of life was investigated.

The adoption of a societal point of view was appropriate as all relevant categories of costs were considered. Both paid and unpaid work lost was considered, which represents a strength of the analysis. The unit costs and the quantities of resources were appropriately reported, which aids replication of the cost analysis in other settings. The source of the data was reported for all items. The costs were specific to the study setting, but extensive sensitivity analyses were carried out to deal with the distribution of cost data and their variability. The price year was reported, which enables reflation exercises to be conducted.

The authors did not compare their findings with those from other studies. They also did not address the issue of the generalisability of the study results to other settings. The extensive use of bootstrapping enhances the robustness of cost-effectiveness estimates. The study referred to adult patients with lateral epicondylitis and this was reflected in the authors' conclusions.

The study results suggested that there is no reason to update or amend the clinical guidelines for GPs in The Netherlands, which recommend a wait-and-see strategy for the management of lateral epicondylitis.

Funded by the Health Insurance Council for Investigative Medicine, and the Netherlands Organisation for Scientific Research.

Smidt N, van der Windt DA, Assendelft WJ, et al. Corticosteroid injections, physiotherapy, or a wait-and-see policy for lateral epicondylitis: results from a randomised controlled trial. Lancet 2002;359:657-62.

Labelle H, Guibert R, Joncas J, et al. Lack of scientific evidence for the treatment of lateral epicondylitis of the elbow: an attempted meta-analysis. J Bone Joint Surg 1992;74B:646-51.

Hudak PL, Cole DC, Haines AT. Understanding prognosis to improve rehabilitation: the example of lateral elbow pain. Arch Phys Med Rehabil 1996;77:568-93.