|An economic evaluation of early versus delayed operative treatment in patients with closed tibial shaft fractures
|Sprague S, Bhandari M
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.
The study compared early versus delayed surgical treatment in patients with closed tibial shaft fractures (CTSFs).
Economic study type
Cost-effectiveness analysis and cost-utility analysis.
The study population comprised adult patients with CTSFs. Further inclusion criteria were considered:
patients aged 17 years and older;
isolated orthopaedic injury;
fracture type (CTSF with or without comminution and unilateral injury);
surgical indications, that is, inability to maintain fracture alignment by conservative methods (>1 cm shortening, >100 varus or valgus and recurvatum or procurvatum), malalignment (>200 external rotation and >100 internal rotation); and
treatment (plate fixation, one-year follow-up by operating surgeon).
The exclusion criteria included significant co-morbidity, cardiovascular disorders (myocardial infarction, angina, hypertension), respiratory disease, and other medical problems that may affect hospital stay or surgery. Further exclusion criteria were multiple trauma or neurovascular compromise of limb, proximal tibial intra-articular and distal tibial intra-articular fractures, and preoperative compartment syndrome (pain swelling, paraesthesias, pallor).
The setting was a university-affiliated trauma centre (secondary care). The economic study was carried out in Canada.
Dates to which data relate
The effectiveness and resource use data were collected between 1989 and 1999. All costs were reported for the fiscal year 1999/2000.
Source of effectiveness data
The effectiveness data were derived from a single study augmented by some authors' assumptions.
The authors reported that their sample size was based upon another report (Bhandari et al., 1999) that identified 20 patients per group as adequate to achieve conventional statistical significance, (p<0.05). Initially, 151 consecutive patients with tibial fractures, surgically managed by two orthopaedic surgeons at one university-affiliated trauma centre, were reviewed. After application of the inclusion and exclusion criteria, 116 patients were excluded. Thus, only 35 patients (23%) were eligible for the study. The 35 eligible patients were divided into two groups, an early treatment group who underwent surgery within 12 hours of injury and a delayed surgical group who underwent surgery more than 12 hours after injury. There were 16 patients (11 males) in the early group and 19 patients (10 males) in the delayed group. The mean age of the patients was 32.0 years in the early group and 31.4 years in the delayed group. The sample of patients in each group was smaller than that identified by power calculations in the other published study.
The analysis was based on a retrospective cohort study that was conducted in a single centre. The duration of follow-up was not explicitly reported, but the authors stated that patients with tibial shaft fractures are typically followed up at regular intervals until fracture union has been achieved, and then yearly after that. No losses to follow-up were reported. The reliability of the assessment of fracture union by two independent reviewers (blinded) was quantified using kappa values (chance-correlated agreement). The method of blinding was not reported.
Analysis of effectiveness
All the eligible patients (n=35) were accounted for in the analysis. The primary outcome used in the analysis was time taken to bone union. The authors defined radiographic union as bridging bone on at least 3 of 4 cortices on anteroposterior and lateral views. Further outcomes used in the analysis were days in hospital, postoperative hospital stay, home care, postoperative complications, number of non-unions of fractures, and number of reoperations. The authors conducted a statistical analysis, using a t-test (alpha = 0.05; two-tailed test, different sized groups) to determine whether potential differences in the baseline characteristics (continuous variables) existed between the two groups. The authors reported that the two groups (non-delayed surgical group, n=16; delayed surgical group, n=19) were similar in all baseline variables, including age, gender, smoking and alcohol consumption, fracture severity and fracture treatment. They also reported that they chose strict eligibility criteria, to ensure that none of the patients included had confounding variables that could alter their postoperative course and bias the results.
The average time to bone union was 28.2 weeks (standard deviation, SD=9.4) for the non-delayed surgical group and 44.2 weeks (SD=7.4) for the delayed surgical group. The difference of 16.0 weeks (95% confidence interval, CI: 5.94 - 0.25) was statistically significant, (p=0.0003).
The days spent in hospital were 4.6 days (SD=1.3) for the non-delayed group and 7.8 days (SD=3.1) for the delayed surgical group. The difference of 3.2 days was statistically significant, (p=0.011).
Postoperative hospital stay was 4.0 days (SD=0.9) for the non-delayed surgical group and 7.1 days (SD=3.2) for the delayed surgical group. The difference was statistically significant, (p<0.05).
The difference in home care need by patients in both groups was not statistically significant.
Two patients in the non-delayed surgical group and 8 in the delayed surgical group developed postoperative complications. The difference was statistically significant, (p=0.0006).
Eight patients in the delayed surgical group, but none in the non-delayed surgical group, developed non-unions. The difference was statistically significant, (p<0.01).
There was 1 reoperation in the non-delayed surgical group and 8 in the delayed surgical group. The difference was statistically significant, (p<0.05).
The authors concluded that the early treatment of CTSFs results in a significantly shorter time of fracture union and fewer postoperative complications.
The costing was carried out retrospectively on the same sample of patients as that used for the effectiveness analysis.
Methods used to derive estimates of effectiveness
The authors used expert opinion to derive utility values for particular health states for both groups.
Estimates of effectiveness and key assumptions
The following utility values were reported:
waiting for an operation, 0.5;
staying in the hospital following surgery, 0.6;
leaving the hospital in crutches with limited activities, 0.7;
experiencing a postoperative complication, 0.5;
delayed union, 0.6;
non-union that requires reoperation, 0.5; and
returning to normal activities, 0.9.
Measure of benefits used in the economic analysis
The measures of benefits used were the number of weeks of bone union and the number of quality-adjusted life-years (QALYs). The authors estimated utility scores for each of the health states for the patients. Since there has been limited research on patient' health utility in orthopaedic trauma, the health utility for each health state was estimated from content expert opinion for both groups. From the utility values, an average utility score was estimated for the non-delayed surgical group and the delayed surgical group. These utility scores were used to calculate the cost per QALY gained for each group.
The costs and the quantities were analysed separately for the two groups, and both the quantities and costs were estimated from actual data. Discounting was not carried out and the authors did not provide any explanation for the time horizon used in the cost analysis. The costs were derived from published sources and from actual hospital records. All costs were reported for the year 1999/2000.
Inpatient costs comprised initial consultation fees, the surgeon's fees (depending on the type of operation), the anaesthetist's fee, surgical overheads, blood tests, radiographs, pain medication, physiotherapy and overhead ward room costs. Initial consultation fees covered the emergency room physician, surgical and anaesthetic consultation. Surgical overheads covered nursing staff, surgical supplies and infrastructure costs. In the case of complications after surgery, additional costs were recorded for diagnostic tests (ultrasound, chest X-rays), the intensive care unit, and additional treatment or procedures. Hospital costs were obtained from a Physician's Schedule of Benefits (published source), while the costs for drugs during hospital stay were obtained from the hospital pharmacy records. The fee for hospital stay was found by taking the weighted average of orthopaedic ward fee per day from a city hospital. The cost of the intensive care unit was obtained by taking the weighted average fee per day for a city hospital.
Outpatient costs comprised follow-up visits with the surgeon (including radiographs and cast removal), home care services, outpatient diagnostic tests required (ultrasound, bone density scan), and additional outpatient therapies. The costs of follow-up visits were derived from the Physician's Schedule of Benefits. The costs of home care services were derived from the billings of the local Community Care Access Centre. The cost of outpatient physiotherapy was not included in the analysis, as the majority of outpatient physiotherapy in Canada is paid for by the patients or their insurance funds, and not by the health care system.
The authors also considered the cost of maintaining and staffing a second operating room for orthopaedic surgery. In the authors' setting, a second operating suite dedicated to trauma was open 3 days a week for 8 hours a day. Thus, they considered the cost of opening and staffing the second operating room 7 days a week for 8 hours a day, to ensure that CTSFs would be operated on within 12 hours, excluding building costs. They also considered the cost of maintaining and staffing a second operating suite for orthopaedic surgery at their centre for an extra 4 days a week. In their analysis they included the annual cost per patient for the second operating suite.
Statistical analysis of costs
The authors reported that after recent recommendations (provided through personal communication with personnel of the Centre for Health Economic and Policy Analysis) their statistical analysis for the economic evaluation was based on an estimation approach, and not on a hypothesis testing approach. They did not report any further statistical tests used in the analysis of costs.
Productivity losses (time lost from work) were determined for each group by assuming that people returned to work at the time their fracture achieved union. The authors used the average earnings for full-time male workers and other workers to estimate the cost of lost productivity, based on the assumption that the majority of patients who fracture their tibiae were males between the ages of 25 and 40 years. The authors also took the unemployment rate for Canadian males between the ages of 25 and 40 years into consideration when determining productivity losses. Discounting was not carried out. It appears that the indirect costs have been reported for the year 1999.
One-way sensitivity analyses were carried out on key variables to test the robustness of the results and to investigate variability in the data. The variables investigated were the upper and lower values of the 95% CI for the mean difference in consequences (time of fraction union), all hospital and outpatient costs (+/- 50%), and the estimated patient utility values (+/- 50%).
Estimated benefits used in the economic analysis
The average time to bone union was 28.2 weeks for the non-delayed surgical group and 44.2 weeks for the delayed surgical group. The number of weeks to bone union gained in the non-delayed surgical group compared with the delayed surgical group was 16.
The mean number of QALYs was 0.79 for the non-delayed surgical group and 0.70 for the delayed surgical group. The number of QALYs gained in the early surgical group versus the delayed surgical group was 0.09.
The total costs were reported per patient. In the non-delayed surgical group, the total cost per patient was Can$20,159.00 when including productivity losses, and Can$6,740.00 when excluding productivity losses. In the delayed surgical group, the total costs per patient were Can$27,489.00 (including productivity losses) and Can$5,667.00 (excluding productivity losses), respectively.
When productivity losses were included, a mean annual cost-savings per patient treated early was Can$7,330.00 (95% CI: 4,148.78 - 16,879.15). This difference included the additional cost of opening a daily dedicated trauma operating suite. When patient productivity losses were not included in the analysis, there was a mean annual increased cost per patient treated early of Can$1,073.00 (95% CI: -3.61 - 2,126.37).
Synthesis of costs and benefits
The non-delayed surgical group dominated when productivity losses were included in the analysis, resulting in a saving of Can$458.00 for each week that a fracture healed sooner in comparison with the delayed group. When productivity losses were not included, the early surgical group resulted in a cost per week of fracture union gained of Can$67.00.
When productivity losses were included in the analysis, the non-delayed surgical group yielded savings of Can$81,444.00 per QALY gained, but when productivity losses were not included, the non-delayed surgical group yielded a cost of Can$11,922.00 per QALY gained.
The sensitivity analysis showed that the cost-effectiveness and cost-utility results were robust. When productivity losses were included in the analysis, all sensitivity analyses performed resulted in savings for the non-delayed surgical group versus the delayed surgical group. From the perspective of the health care system though, the sensitivity analyses resulted in costs of Can$33.16 to Can$180.33 per week for earlier fracture union, and in costs of Can$5,896.67 to Can$23,844.00 per QALY gained.
Compared with delayed treatment, the early treatment of closed tibial shaft fractures (CTSFs) results in a significantly shorter time to fracture union, fewer postoperative complications, significant cost-savings and greater quality-adjusted life-years (QALYs) gained.
CRD COMMENTARY - Selection of comparators
The authors compared early versus delayed operative treatment in patients with CTSFs. The comparators used were justified on the grounds that surgical delays are common in most trauma centres. You should decide if the comparator represents current practice in your own setting.
Validity of estimate of measure of effectiveness
The analysis was based on a retrospective cohort study of consecutive patients with tibial fractures, surgically managed by two orthopaedic surgeons. This seems to have been appropriate for the study question. The study sample was representative of the study population, but the sample size of the two groups seemed to be smaller than that determined by power calculations in another published study. The patient groups were shown to be comparable at analysis. The authors also applied strict eligibility criteria for the patients, so as to avoid potential biases and confounding factors. The two investigators who assessed the fracture union from radiographs were independent and blinded, but the method of blinding was not reported in detail.
Validity of estimate of measure of benefit
The number of weeks to bone union was derived directly from the effectiveness analysis. The authors used expert opinion instead of a validated instrument to estimate the utility values for each of the health states, in order to calculate the cost per QALY. It is therefore possible that these values might have been used selectively, and this limits the generalisability of the results.
Validity of estimate of costs
All the categories of costs relevant to the perspective adopted were included in the analysis. It appears that for each category of costs, all the relevant costs have been included. As reflected in the limitations of the study reported by the authors, productivity losses from other family members were not included, and might have been greater in the delayed surgical group. The analysis did not include the cost of outpatient physiotherapy, as it is paid by the patients or their insurance plans and not by the health care system. It is difficult to estimate, though, whether this cost would have affected the authors' conclusions. It may limit the generalisability of the results to other countries with different systems.
The costs and the quantities were reported separately, which will enhance the reproducibility of the results. A sensitivity analysis of the costs was undertaken, using what appear to have been appropriate ranges. The date to which the prices referred was reported and this will facilitate possible reflation exercises. Discounting was not carried out. However, since the period during which the costs were incurred was unclear, it is therefore not possible to determine whether discounting was relevant.
The authors did not compare their findings with other published results. However, this may have been due to a lack of published literature in this specific area. They mentioned though that studies have reported increased postoperative hospital stays and postoperative complication rates in patients who experienced delayed surgery. The authors did not discuss the generalisability of the results to other settings, although the sensitivity analyses undertaken improve the generalisability of the findings. The authors do not appear to have presented their results selectively. The study involved patients with tibial fractures surgically managed, who were divided into two groups (early and delayed surgical groups), and this was reflected in the authors' results and discussion.
The authors pointed out some further limitations of their study. First, the data were collected retrospectively. Second, there was insufficient information on patients' productivity losses regarding time lost from leisure activities, unpaid production time, and productivity losses from family members or other caregivers, which may have been greater in the delayed surgical group. Further, the authors assumed that patients were not replaced at their jobs and they returned to work when their fracture had healed, which may have resulted in slightly inflated productivity losses in cases where workers were replaced, or continued to work in a less physically demanding occupation. Finally, the authors mentioned that their choice to examine only a single cut-off point (12 hours) for early surgery versus delayed surgery provided little information about other degrees of surgical delay (i.e. 24 or 48 hours). They reported though that another study had suggested that postoperative hospital stay is linearly correlated to the preoperative surgical delay.
Implications of the study
The authors did not make any explicit recommendations for policy change or practice. They suggested that a prospective cohort study with a larger sample size would achieve a more precise estimate of the effects and costs of surgical delay. In addition, it should be mentioned that their discussion highlighted some areas where assumptions were made, implying the need for greater information.
Source of funding
Funded in part by a research grant from the AO Foundation, Switzerland.
Sprague S, Bhandari M. An economic evaluation of early versus delayed operative treatment in patients with closed tibial shaft fractures. Archives of Orthopaedic and Trauma Surgery 2002; 122(6): 315-323
Subject indexing assigned by NLM
Ambulatory Care /economics; Canada; Costs and Cost Analysis; Fractures, Closed /economics /surgery; Fractures, Ununited /economics; Hospital Costs; Humans; Length of Stay /economics; Outcome Assessment (Health Care) /economics; Postoperative Complications /economics; Quality-Adjusted Life Years; Retrospective Studies; Statistics as Topic; Tibial Fractures /economics /surgery; Time Factors
Date bibliographic record published
Date abstract record published