Introduction

Microdiscectomy is the most frequently performed surgical procedure for patients with sciatica caused by lumbar disc herniation. This open surgical approach aims to remove the herniated disc fragment causing nerve root compression.

Another indirect way of decompressing the compromised nerve root is to reduce pressure in the intervertebral disc by removing herniating disc material. Several percutaneous techniques are based on this principle, including percutaneous laser disc decompression (PLDD). An attractive feature of PLDD is its minimally invasive, non-surgical nature.

In a randomized clinical trial, we compared PLDD with conventional microdiscectomy in a subgroup of patients. The trial showed that a strategy of PLDD—followed by surgery when needed—in patients with sciatica due to lumbosacral disc herniation was not inferior to open surgery at 1 year, at the cost of higher re-treatment rates and a longer time to recovery.

PLDD is associated with lower short-term healthcare costs than conventional surgery, as it does not require hospitalization or general anaesthesia. To date, no economic evaluations comparing PLDD with surgery have been performed; therefore, the cost-effectiveness of PLDD remains to be established.

We carried out a cost-utility analysis of our randomized controlled trial, comparing observed 1-year quality-adjusted life years (QALYs) with observed 1-year societal costs to determine whether favourable cost-effectiveness for a strategy of PLDD can be achieved compared with a direct surgical strategy.

Methods

Patients with sciatica due to nerve root compression caused by lumbar disc herniation participated in a multicentre randomized controlled trial comparing conventional microdiscectomy with PLDD. Institutional Review Boards of the participating hospitals approved the research protocol, and participants provided written informed consent.

A total sample size of 110 was chosen based on equivalence in effect on the Roland Disability Questionnaire for Sciatica (RDQ) between microdiscectomy and PLDD. Between January 2005 and September 2007, 115 patients were enrolled and 112 were analysed (55 allocated to PLDD and 57 to surgery). Three patients were excluded after randomization: two withdrew before treatment and one was excluded before treatment because of protocol violation. Both groups were comparable in baseline characteristics.

Patients and treatment

Eligible patients were 18–70 years old with a radiologically confirmed disc herniation and corresponding lumbosacral radicular syndrome lasting more than 6–8 weeks. The herniated fragment had to be smaller than one-third of the spinal canal. We excluded patients with cauda equina syndrome, previous spinal surgery at the same disc level, spondylolisthesis, bony spinal stenosis, pregnancy, severe somatic or psychiatric disease, inadequate knowledge of the Dutch language, or planned emigration within 1 year of inclusion.

Treatment details are reported elsewhere. Briefly, microdiscectomy was performed via an ipsilateral approach with midline paravertebral muscle retraction and removal of the herniated disc through a transflaval approach, with no or minimal bone removal. Surgery was finished once nerve root decompression was visible and loose disc fragments had been removed.

In PLDD, laser energy is delivered via an optical fibre passed through an 18G needle placed into the nucleus pulposus via a posterolateral approach. In contrast to other minimally invasive surgical procedures, this treatment is performed under local anaesthesia and does not require an anaesthesiologist. As a result of laser disc decompression—whereby the water content of the nucleus pulposus decreases due to vaporisation—the herniated disc shrinks, resulting in reduced nerve root compression.

Utilities and QALYs

Utilities represent valuation of patients’ quality of life, anchored at 0 (as bad as death) and 1 (perfect health). Patients described their quality of life using the EuroQol classification system (EQ-5D), from which we calculated utilities for the United States (US) and The Netherlands (NL). Similarly, patients reported their quality of life using the 36-item Short Form Health Survey (SF-36), from which we calculated SF-6D utilities. Both EQ-5D and SF-6D provide societal valuation, which is preferred for economic evaluations from a societal perspective.

In addition, we obtained valuations from the patients themselves using a visual analogue scale (VAS) ranging from 0 (worst imaginable health) to 100 (perfect health). We transformed the values to a utility scale using the power transformation: 1 − [1 − (VAS/100)].

We obtained EQ-5D and VAS measurements at randomization and at 3, 4, 6, 8, 12, 26, 38, and 52 weeks after randomization. SF-36 measurements were obtained less often: at randomization and at 4, 8, 26, and 52 weeks. For EQ-5D, SF-36, and VAS, respectively, 5%, 9%, and 6% of items were missing. From the area under the utility curves, we calculated the average utility during each quarter of the year after randomization and during the entire year (QALYs).

Costs

We estimated costs from the societal perspective over 1 year of follow-up. Because of the 1-year time horizon, costs were not discounted. Costs were converted to 2010 price levels using the general Dutch consumer price index.

Using cost diaries, patients reported hospital admissions, visits (specialists, general practitioner, physiotherapy, and alternative healthcare), home care, paid domestic help, informal care, drugs and aids, out-of-pocket expenses due to sciatica, and hours of absenteeism from work.

At follow-up at 4, 8, 26, and 52 weeks after randomization, the research nurse reviewed the diary with the patient. At these follow-up intervals, 2%, 2%, 3%, and 7% of diaries were missing, respectively.

Integral costs of surgery were based on cost prices of participating hospitals. Costs of surgery were estimated at €2,428 excluding hospital stay. Integral costs of PLDD were estimated at €868 based on a micro-cost calculation of direct and indirect personnel, equipment, materials, and overhead costs.

For scheduled treatments cancelled at the last moment (for example because the patient recovered), costs of personnel, equipment, and overheads were included (as no other intervention could be scheduled), but the cost of consumables was not included.

For other healthcare, we used Dutch standard prices designed to represent societal costs and to standardize economic evaluations. Healthcare costs are reported including patients’ time and travel costs.

We valued hours of absenteeism from work and hours worked less in case of reduced working hours during the 1-year follow-up period according to the friction cost method using a friction period of 22 weeks, at standard productivity costs of €34 per hour for women and €43 per hour for men.

Analysis

All analyses followed the intention-to-treat principle. Statistical analyses were conducted with Stata version 9.2 (StataCorp, College Station, TX, USA).

To reduce possible bias due to missing data, we used multiple imputation by chained equations with five iterations for the switching regression model. For each missing utility or cost measure, an imputation regression model was used that included patient age, sex, body mass index, randomization group, functional disability (RDQ), VAS leg and back pain, duration of complaints, and all other utility and cost measures at all other time points.

Group differences in QALYs and costs were analysed using unequal-variance Student’s t-tests. The base-case cost-utility analysis compared 1-year societal costs with 1-year QALYs based on the US EQ-5D. Sensitivity analyses were carried out using different utility measures (NL EQ-5D, SF-6D, or VAS) and perspectives (societal or healthcare).

Depending on willingness to pay for obtained effectiveness, a strategy is cost-effective compared with an alternative if it has a better average net benefit (willingness to pay × QALYs − costs). Acceptability curves were constructed in which the y-axis represents the probability that PLDD is cost-effective compared with conventional microdiscectomy. A value of 0.5 represents equality in cost-effectiveness between two treatments.

Results

Utilities and QALYs

According to the EQ-5D, quality-of-life valuation for PLDD 2 weeks after randomization was better than for surgery. However, after that time the utility measures were consistently worse for PLDD than for surgery.

The VAS and SF-6D provided similar patterns, although quality of life of PLDD patients according to SF-6D was better until 8 weeks after randomization, while according to the VAS quality of life of PLDD patients was only better at 8 weeks.

Figure 1. Cost-effectiveness acceptability curves for PLDD compared with conventional microdiscectomy
EQ-5D: EuroQol five dimensions questionnaire; NL: Netherlands; PLDD: percutaneous laser disc decompression; QALY: quality-adjusted life year; US: United States


Figure 2. Utilities (including confidence intervals) according to US, NL EQ-5D, SF-6D, and VAS
CD: conventional discectomy; EQ-5D: EuroQol five dimensions questionnaire; NL: Netherlands; PLDD: percutaneous laser disc decompression; SF-36: 36-item short form health survey; US: United States; VAS: visual analogue scale.

Utilities during all four quarters and according to all measures were consistently less favourable after PLDD, except for the first quarter according to SF-6D (Table 1). The largest difference in utilities was observed in the second quarter: 0.103 according to the NL EQ-5D (95% CI 0.016–0.189). QALYs during the entire 1-year follow-up period as measured by the VAS were significantly lower for PLDD patients (p = 0.02).

Healthcare costs

Mean costs of initial PLDD treatment were €2,751 lower (95% CI €2,414 to €3,088) than mean costs of surgery, including an initial hospital admission of an average of 3.0 days (Table 2). Part of this cost difference was offset by a higher reoperation rate (44% for PLDD versus 16% for conventional discectomy). Consequently, costs per patient of reoperations were significantly higher for PLDD patients (mean difference €1,643, 95% CI €780 to €2,506).

Over the first year, total healthcare costs for PLDD were significantly lower than for conventional microdiscectomy, with a difference of €1,771 (€303 to €3,238) per patient.

Societal costs

Total non-healthcare costs after PLDD were lower than after conventional microdiscectomy with a non-significant difference of €609 (−€3,840 to €5,058). The difference in total societal costs amounted to €2,379 (−€2,860 to €7,618) per patient, non-significantly in favour of PLDD.

Cost-utility analysis

The combination of lower societal costs and less favourable QALY outcomes means that the probability that PLDD is cost-effective compared with surgery depends on willingness to pay. For low values, PLDD is more likely to be cost-effective; for higher values, between €30,000 and €70,000, conventional microdiscectomy becomes more favourable.

In the base-case analysis (US EQ-5D, societal perspective), PLDD was more likely to be cost-effective than surgery for a willingness to pay of €70,000 per QALY and lower. From the healthcare perspective, PLDD was more likely to be cost-effective for a willingness to pay of €50,000 and lower. Using the NL EQ-5D, PLDD was more likely to be preferred below €40,000 and €30,000 per QALY for the societal and healthcare perspectives, respectively.

Discussion

Although for higher willingness-to-pay values surgery becomes more likely to be cost-effective, acceptability curves in this randomized study show that for low values of willingness to pay, PLDD is more likely to be cost-effective. At internationally used threshold values for willingness to pay per QALY, there is no clear preference between methods from a societal perspective; therefore, treatment can be based on patient preference.

This makes PLDD a viable non-surgical treatment option for patients with sciatica caused by small, contained herniated discs who wish to postpone the decision for surgery.

Reference

INR Interventional Neuroradiology (2017), Vol. 23(5), 538–545