Spinal Cord Stimulation Improves Quality of Life for Patients With Chronic Pain — Data From the UK and Ireland National Neuromodulation Registry

Introduction: Spinal cord stimulation (SCS) is a well-established treatment for chronic pain and is supported by numerous studies. However, some recent articles have questioned its ef ﬁ cacy. This article examines a cohort of > 1800 patients with SCS from the UK and Ireland National Neuromodulation Registry. It is intended to provide a “ real-world ” assessment of ef ﬁ cacy and compare its effects with other procedures performed for painful indications. Materials and Methods: Quality of life (QoL) data (EuroQoL ﬁ ve-level [EQ5D]) and demographic data were extracted from the National Neuromodulation Registry for all patients ( N = 1811) who underwent SCS for chronic pain in 27 centers in the UK between February 2018 and July 2022. These were compared with data from the published literature for other commonly performed elective surgical procedures. Results: The EQ5D utility index increased by a mean of 0.202 in the 1236 patients with paired pre-and postoperative utility scores. The median utility was 0.263 (interquartile range [IQR] = 0.384; n = 1811) preoperatively, whereas at six months after the operation, it was 0.550 (IQR = 0.396; n = 1025), p < 0.0001, Wilcoxon rank sum test. The median utility score at 12 months postoperation was 0.548 (IQR = 0.417; n = 970). There was no difference in utility scores at six months and 12 months after implantation ( p = 0.15, Wilcoxon rank sum test). There was a signi ﬁ cant improvement in QoL in all ﬁ ve domains of the ﬁ ve-level EQ5D tool at six months after baseline ( p < 0.01, for all subcategories), and this was sustained at one year after implantation. The baseline utility was lower than in patients who underwent elective surgery for other painful conditions, and the absolute (and proportionate) increase in utility produced by SCS was greater than that achieved with most other interventions. Conclusions: SCS increases the QoL in patients requiring surgery for pain. Similar results were seen regardless of SCS indication. When comparing analogous data bases, SCS produces a greater percentage improvement in EQ5D utility than do many other elective surgical procedures for painful conditions, including spinal surgery and some joint replacements.


INTRODUCTION
Patients with chronic pain have inferior quality of life to that of the general population. 1,2The Global Burden of Disease study 2021 showed that pain and pain-related disease are the greatest cause of years lived with disability worldwide. 3][7][8][9][10][11][12][13][14] However, some recent publications have asserted that spinal neuromodulation may not provide clinically significant pain relief compared with placebo.Notable among these are two Cochrane systematic reviews. 15,16ne of the Cochrane reviews 15 states that "people treated with the addition of SCS may experience less pain and higher quality of life after one month or six months of stimulation.There is limited evidence to draw conclusions in the long term of one year or more," and that "compared to a sham … SCS may result in small reductions in pain intensity in the short term that may not be clinically important … no evidence [in the] medium or long-term."The other 16 states, "moderate-certainty evidence suggests there is probably no benefit of SCS over placebo on pain, function, or health-related quality of life in the medium term." The findings of these reviews are at odds with the everyday clinical experience of implanting physicians, and their inclusion criteria, methods, presentation of results, analysis, and interpretation have been robustly criticized. 17,18Nevertheless, the authors draw conclusions intended for policymakers and funders, which may directly affect the availability of SCS for some patients.In this context, large "real-world" data sets showing outcome after SCS, in a population reflective of the generality of those receiving the treatment, complementing data from clinical trials, are urgently needed.The National Neuromodulation Registry (NNR) is a UK-wide repository of demographic, diagnostic, device identification, and quality of life (QoL) outcome data for patients who undergo neuromodulation procedures including SCS.It is governed by the Neuromodulation Society of the UK and Ireland, a chapter of the International Neuromodulation Society.To date, 27 implanting centers have joined.The implanting clinicians include both anesthetists and neurosurgeons, and data are collected for both surgical (paddle) and percutaneous implants from all device manufacturers.
When setting up this registry, a conscious decision was taken only to include demographic and device data and a simple healthrelated QoL outcome measure, using the EuroQoL five-level system (EQ5D-5L).This was to ensure the robustness of data collection while this registry was in its early stages, with a plan to expand the amount and type of data subsequently captured.
A vital aspect of the registry is that the patient-reported QoL outcome data are collected by an independent third party (NEC Software Solutions UK, the registry operator) rather than by the implanting clinical teams, to ensure the data are as robust and free from bias as possible.The registry operates on a subscription model in which NEC is funded through an annual charge paid by the implanting centers (most of which are government funded through the National Health Service [NHS], with a few private providers).
The most performed elective surgical procedures include cataract surgery, cholecystectomy, inguinal hernia repair, joint replacements, and spinal surgical interventions. 19Apart from cataract surgery, the primary indication for almost all these procedures is pain-for example, chronic osteoarthritis for joint replacement surgery.Pre-and postintervention QoL data for some of these procedures in the UK are captured in a similar way (using EQ5D-5L) by NHS Digital and are publicly available.The collection of QoL outcome data on a standard scale allows comparison of the results of SCS with those of procedures recorded in other, similar registries.
The purpose of this paper is twofold.First, we analyze real-world data from the NNR to quantify the effect of SCS on QoL, as measured with the EQ5D-5L.Second, we compare the effects of SCS with those of other elective surgery for painful conditions, when analogous registry data exist.

Data Collection
QoL data (EQ5D-5L) and demographic data were extracted from the NNR for patients who underwent SCS for chronic pain in 27 centers in the UK between February 2018 and July 2022.Patients implanted after July 2022 were excluded because they had yet to have the opportunity to provide their one-year follow-up data.Follow-up time points were six and 12 months.The registry extract contained no identifying information.As a retrospective analysis of fully anonymized data, formal ethical approval was not required for this study.

Statistical Analysis
Data were analyzed using a commercially available statistical software platform, R. 20 Analyses were either parametric or nonparametric, depending on data attributes and on assessment of normality by inspection of a QQ plot in addition to the Shapiro-Wilk test.Data are described as proportions (%), and mean ± 95% CI, or median with interquartile range, as appropriate.Parametric data were compared using an appropriate paired or unpaired Student or Welch t-test.Nonparametric data were compared using the Wilcoxon rank sum (unpaired) or signed rank (paired) tests.Categorical data were compared using Fisher's exact test.One-way analysis of variance (ANOVA) was used when multiple groups were compared.Statistical significance was set at p < 0.05.Output of statistical significance on the R platform is limited at <2.2 × 10 -16 ; instances in which the p value is lower than this are described as p < 0.0001.

Visualization
Data were plotted using R. When possible, we have shown the entire data set, with overlaid graphics to provide summary data.Sankey diagrams were used to visualize the changes in the five components of the EQ5D after intervention.

EuroQoL Five-Level System
The EQ5D-5L 21,22 tool comprises five questions, each a selfreported assessment of one aspect of function related to QoL: mobility, self-care, the ability to perform one's usual activities, pain/ discomfort, and anxiety/depression.There are five possible responses to each question: no problem, a slight problem, a moderate problem, a severe problem, or an extreme problem.The responses are assigned an integer score ranging from 1 (no problem) to 5 (extreme problem).These scores feed into an algorithm that generates an overall index, known as the utility.The utility index value is on a continuous scale with a maximum of 1, representing perfect health.A score of 0 is regarded as equivalent to being dead.Scores <0 are possible; the minimum value (using the UK data set) is −0.594, corresponding to "extreme" ratings for all five questions.

Comparison With Other Elective Surgery
The QoL outcomes for SCS were compared with EQ5D-5L data from large data sets for other commonly performed elective painrelated procedures.Comparator data sets from UK sources were used.

Demographics
There were 1811 recorded cases of SCS.Of these cases, 984 (54.3%) were in female patients; 826 patients (45.6%) were male, and one preferred not to disclose.The mean age at implantation was 52.2 years (SD = 13.5 years).
Of the 1811 recorded cases, baseline EQ5D utility was available for all cases.There was a paired EQ5D utility available for 1025 cases at the six-month follow-up time point and for 970 cases at the 12-month follow-up time point.In 1236 of 1811 cases, there were paired pre-and postoperative EQ5D utility scores for at least one of the six-or 12-month follow-up time points.In those cases for which both were available, the 12-month utility was used for the analyses below.
Indications for treatment are listed in Table 1.
The median utility score at 12 months postoperation was 0.548 (IQR = 0.417; n = 970; mean = 0.467).There was a significant difference between the utility scores before and 12 months after implantation (p < 0.0001, Wilcoxon rank sum test).
There was no difference in utility scores at six months and 12 months after implantation (p = 0.15, Wilcoxon rank sum test).

Utility Change
A total of 1236 patients had paired preoperative and postoperative utility indices (Fig. 2).Postoperative utility was calculated from 12-month follow-up data if available; when this was not available, data from the six-month time point were used.This approach is supported by the previously mentioned data showing that the six-and 12-month EQ5D index values are statistically indistinguishable.The mean improvement was 0.202 (95% CI: 0.185-0.219).Of these patients, 930 showed improvement (75.3%); 23 had no change (1.9%), and 283 experienced worsening of their condition (22.8%).
In those cases in which follow-up data were available at both six and 12 months, there was no significant difference between the utility at six months and 12 months after treatment (p = 0.21, t-test).
Of the patients with worsening at six months, 146 also had 12month follow-up data.The mean change at 12 months in this subgroup was −0.031 (i.e., still worse than baseline), and only 58 (40%) of the 146 had a 12-month utility at or above baseline.These findings indicate that a negative outcome at six months is predictive of a poorer outcome at 12 months.

EQ5D Components
For the 1236 patients with paired pre-and postoperative utility scores, the change in each component of the EQ5D-5L was evaluated.The mean score was significantly improved after SCS for each of the five components.
The distributions of each score within each category, and the ways these changed after treatment, are presented in the Sankey diagrams in Figure 3.The proportion of participants with severe or worse problems (ie, EQ5D component score of 1-3 vs 4-5) before and after implantation was investigated.A significant difference (improvement) between these proportions was seen for all the EQ5D components (p < 0.0001 for pain/discomfort, mobility, usual activities; p = 0.0002 for self-care; p = 0.01 for anxiety/depression, Fisher's exact test).This is represented graphically in Figure 3f.
Nearly three-quarters of patients (74.4%, n = 919) reported severe or extreme pain/discomfort before treatment, which decreased to less than one-third of patients (32.7%, n = 404) reporting severe or extreme discomfort after treatment (p < 0.0001, Fisher's exact test).

Outcomes Related to Main Surgical Indications
The most established indications for SCS are persistent spinal pain syndrome (PSPS) and complex regional pain syndrome (CRPS).Together, these accounted for nearly three-quarters of the cases in the registry.Figure 4 indicates the changes in utility for PSPS (split into lumbar and cervical cases) and CRPS.
Within these primary indications, there was a significant increase in median postoperative utility when compared with baseline utility (Fig. 4a; it is noteworthy that only those with paired pre-and postoperative results were included here).There was no significant difference when comparing the change in utility between the different indications (Pr(>F) = 0.297, one-way ANOVA, Fig. 4b).

PSPS (Lumbar)
There were 1152 cases of PSPS related to the lumbar region (Table 2).
The median preoperative utility was 0.258 (IQR = 0.378; n = 1152).The median utility score at six months postoperation was 0.554 (IQR = 0.394; n = 664).There was a significant difference between the utility scores before and six months after implantation (p < 0.0001, Wilcoxon rank sum test).The median utility score at 12 months postoperation was 0.548 (IQR = 0.405; n = 635).There was a significant difference between the utility scores before and 12 months after implantation (p < 0.0001, Wilcoxon rank sum test).There was no difference in utility scores at six months and 12 months after implantation (p = 0.11, Wilcoxon rank sum test).
Of the 1152 cases, 797 patients had paired preoperative and postoperative utility indices.Postoperative utility is defined as the utility at 12 months after implantation if this was available, or if this was not available, the six-month utility was used.The mean improvement was 0.205 (95% CI: 0.188-0.222),as presented in Figure 5.Of these patients, 599 showed improvement (75.2%); 15 had no change (1.9%), and 183 experienced worsening (22.9%) in their condition.
Lumbar PSPS Subcategories.The mean improvement in utility for lumbar region type 2 PSPS (PSPS2) was 0.195 (n = 675) (Fig. 6).When used for PSPS2, SCS is generally considered better for neuropathic leg pain than for back pain.We therefore compared the results of SCS in PSPS2 groups with predominant leg pain, predominant back pain, or equal back and leg pain.
The analysis showed no significant difference (one-way ANOVA, Pr(>F) = 0.51) in results of stimulation by pain location for PSPS2mean improvement for leg pain predominant was 0.192 (n = 296), for back pain predominant was 0.179 (n = 183), and for codominant was 0.213 (n = 196).For the patients not previously operated on (PSPS1), the mean improvement was 0.264 (n = 122).PSPS1 had a significantly better outcome than did PSPS2 after SCS (p = 0.03, Welch's two-sample ttest).

PSPS (Cervical)
There were 45 cases of PSPS associated with cervical spinal surgery.Subcategories are listed in Table 3.
Overall Outcome Cervical PSPS.The median preoperative utility was 0.309 (IQR = 0.458; n = 45).The median utility score six months postoperation was 0.632 (IQR = 0.498; n = 26).There was a significant difference between the utility scores before and six months after implantation (p = 0.003, Wilcoxon rank sum test).The median utility score at 12 months postoperation was 0.559 (IQR = 0.488; n = 24).There was a significant difference between the utility scores before and 12 months after implantation (p = 0.007, Wilcoxon rank sum test).There was no difference in utility scores at six months and 12 months after implantation (p = 0.74, Wilcoxon rank sum test).
Of the 45 cases, 31 patients had paired preoperative and postoperative utility indices, shown in Figure 7. Postoperative utility is defined as the utility at 12 months after implantation when available, and if this was not available, a six-month utility was used.The mean improvement was 0.191 (n = 31).Of these patients, 24 showed improvement (77.4%); one had no change (3.2%), and six showed worsening (19.4%).
The data set was not large enough to allow meaningful comparison of subgroups.Overall, the mean improvement for PSPS1 was 0.322 (n = 6), and for PSPS2 was 0.160 (n = 25)-these were not significantly different (p = 0.08, Wilcoxon rank sum test).

Complex Regional Pain Syndrome
In this series, 151 patients with CRPS underwent SCS.
Overall Outcomes in CRPS The median preoperative utility was 0.262 (IQR = 0.341; n = 151).The median utility score at six months postoperation was 0.531 (IQR = 0.453; n = 89).There was a significant difference between the utility scores before and six months after implantation (p < 0.0001, Wilcoxon rank sum test).The median utility score at 12 months postoperation was 0.528 (IQR = 0.432; n = 81).There a significant difference between the utility scores before and 12 months after implantation (p < 0.0001, Wilcoxon rank sum test).There was no difference in utility scores at six months and 12 months after implantation (p = 0.85, Wilcoxon rank sum test).
Of the 151 cases of CRPS, 107 patients had paired preoperative and postoperative utility indices.Postoperative utility is defined as the utility at 12 months after implantation if available, and when this was not available, a six-month utility was used.The mean improvement was 0.175 (Fig. 8).Of these patients, 78 showed improvement (72.9%); three had no change (2.8%), and 26 showed worsening (24.3%).

Impact of Loss to Follow-Up
In the cohort reported here, follow-up data were obtained in 68.2% of cases in which baseline data had been entered into the  registry.To investigate whether compliance with follow-up registry data collection was likely to be related to outcome, we compared patients who provided both six-and 12-month follow-up data (defined as completely compliant) with those who provided either six-or 12-month follow-up data but not both (defined as partially compliant).

Change in
The mean changes in EQ5D-5L utility in the complete compliance (n = 759) and partial compliance (n = 477) groups were virtually identical (0.202 complete compliance group vs 0.201 partial compliance group, p = 0.92 Welch's two-sample t-test, Fig. 9).
This comparison provides reassurance that compliance is unrelated to outcome in this data set for those individuals who have been at least partially compliant.
To further assess the effect of missing follow-up data points (ie, individuals who were noncompliant), a "worst-case scenario" was assumed in which all those without follow-up data were assumed to obtain zero benefit from SCS.A distribution of n = 575, with a mean change in utility of 0.0 and SD equal to that of the cohort with results (ie, SD = 0.300), was superimposed on the cohort of paired results (n = 1236), as presented in Figure 10.This new worst-case distribution of n = 1811 showed an overall mean improvement in EQ5D utility of 0.146 (95% CI: 0.132-0.160).This finding shows that even if all missing data points are assumed to be of net no improvement, there is still a benefit of SCS for the cohort overall.

Comparison of Change in Utility Before and After Treatment With SCS and Other Elective Operations
We identified the most common elective operations in the UK to see how the change in utility after SCS compares with that seen after other surgical procedures.We selected those performed for conditions with a painful component.We then compared the baseline utility scores in the largest and most comparable studies in the literature, and the percentage improvement in mean EQ5D utility.These data are presented in Table 4.To maximize comparability, the data sets are all from the UK, and all use the five-level (rather than three-level) EQ5D tool.Data for spinal surgery, joint replacements, and hernia repairs are from national registries (British Spine Registry for spinal surgery or NHS Digital for joint replacements and hernia repairs).We were unable to locate suitable  comparator data for cholecystectomy.The comparator data for percutaneous coronary intervention for stable angina pectoris are sourced from a UK randomized controlled trial owing to the absence of a suitable registry data set; they are therefore of a qualitatively different nature and should be regarded as illustrative only.

Lumbar PSPS Subcategories
The preoperative QoL for patients in the NNR who underwent SCS was poorer than that reported for any of the other listed commonly performed procedures (Fig. 11, Table 4).Mean population utilities in England at different ages 27 are included for illustration in Figure 11b.
SCS compares favorably in terms of change in EQ5D utility, both in proportional and absolute improvement, with almost all other procedures.It is appreciably surpassed only by joint replacement surgery regarding absolute health gain, and in terms of proportional increase in utility, its efficacy is second only to total hip replacement.

DISCUSSION
Data from the UK NNR show that most patients treated with SCS for chronic pain substantially improved their QoL.This finding was consistent across indications.In all cases, there was a highly significant change in the index due to SCS, and the results at six and 12 months postoperatively were statistically indistinguishable.
SCS improved the scores in each of the five domains of the EQ5D-5L.Not surprisingly, given that SCS is an intervention to treat chronic pain, the "pain/discomfort" domain change is the most obvious.However, chronic pain can influence all facets of life, and preoperatively, most patients also describe problems in their ability to undertake their usual daily activities, difficulties with their mobility and self-care, and some degree of anxiety or depression.SCS approximately halved the proportion of patients in the worst ("severe" or "extreme") categories in every domain except anxiety/ depression, where it reduced the proportion by a quarter.
SCS produces a large percentage improvement in EQ5D utility owing to a large effect size and a low baseline utility index (85% mean difference and an improvement of 109% in median utility).This compares favorably with other well-established surgical interventions for painful conditions with analogous registry data, surpassed only by hip replacement (eg, spinal surgery 59.5% improvement at 12 months, 26 total knee replacement 76.4% improvement, 25 total hip replacement 139.0%improvement 25 ).The absolute change in the EQ5D index with SCS is closely comparable with that after spinal surgery, and only joint replacements show an appreciably higher absolute change.The proportion reporting an improvement in their QoL after SCS (75.2%) compares well with that reported for hip replacement (69.7%) and knee replacement (59.5%). 25lthough a score of 1.0, representing a state of perfect health, is the maximum that the algorithm that calculates the index can generate, at the other end of the scale, the most extreme responses to the survey questions can generate scores less than zero (zero being notionally a state equivalent to death).Index weights, and therefore the exact value of the minimum, vary from country to country because the index is normalized on each population.Using UK values, the index calculator has a lower limit of −0.59 for the worst health state possible.This is why a small number of individuals, who went from an extremely poor QoL before treatment to a very good QoL afterward, show a change in utility index that is >1.0.

Change in Utility
A conscious decision was made in the initial setup of the NNR to keep the data set as simple as possible while the initial uptake of registry use was encouraged.EQ5D-5L was selected as a widely accepted measure of health-related QoL validated in many disease states.The utility index is the key measure of efficacy used in the UK and many other countries in cost-effectiveness analyses for national decisions regarding commissioning health care interventions.We have now introduced other measures into the core data set.These include the 29 item patient-reported outcomes measurement information system (PROMIS-29), 28 and data from these measures will be included in future reports.
Although this data set is reasonably large, it does not encompass the full five-year UK SCS caseload.This is partly because implanting centers are still coming on board-the data presented here come from 27 centers, but it is believed the total number of implanting centers in the UK is >40.It also is partly because the 27 centers that are included have begun to use the registry at various points in the reported period, some of them quite recently, so their contributions will be, in some cases, only a small proportion of their output in the reported period.Although they limit the data set size, these factors would not be expected to bias the data (in statistical terms, data "lost" owing to centers not using the registry can likely be regarded as missing at random).N is the number of cases in the series denoted by the reference.The EQ-5D utility before and after surgery is listed.PCI, percutaneous coronary intervention.The most obvious criticism of this data set is the incompleteness of follow-up data, with postoperative EQ5D scores available in just under 70% of cases.This is partly due to the data collection method, which is performed through email by providing a link to patients to complete the EQ5D questions online.A single reminder is sent if this is not done.A higher response rate could be achieved if implanting centers collected the data themselves and pursued missing data more aggressively.However, the independence of the collection process of follow-up data from the implanting centers is one of the strengths of the registry because it guards against implicit biases or patient suggestibility.

SCS
The respondent rate with paired pre-and postoperative data points is, however, higher than in the most recent reported year for the most closely analogous registries (63.1% for the total hip replacement section and 59.4% for the total knee replacement section of the UK patient-reported outcome measures reports 25 ).
We have focused analysis on cases with paired pre-and postintervention data.We also have attempted to assess the impact of incompleteness by comparing results in those deemed fully compliant (returned data at both follow-up time points) with those who were partially compliant (returned data at one time point but not both).We found no difference in outcome between these groups.This suggests that single missing follow-up data points in the partially compliant group may be regarded as missing at random.Although this is encouraging, caution must be exercised in extrapolation to the noncompliant group (ie, those who did not provide postoperative data at all), and it remains an open question whether patients who were noncompliant fared better, the same, or worse than those who did comply.A strong tendency one way or the other in the noncompliant group could affect the analysis.To provide a worst-case assessment of the effect of this possibility, we imputed data based on the assumption that there was zero net benefit for the noncompliant group.A distribution was generated with a mean utility improvement of 0.0, with the same variance as those that were compliant.This "worst case" data set still showed a substantial and significant net benefit of SCS on QoL, with an EQ5D index improvement of 0.146 (reduced by 27.7% compared with the original data set without imputed results).
The assumption that patients lost to follow-up (LTFU) were nonresponders is extremely conservative.A recent study has addressed the reason patients do not engage with follow-up after SCS.Ho et al 29 made determined efforts to locate a cohort of 49 patients with SCS who had been LTFU and managed to find 24 of these.Of the 24, 14 (58%) cited successful pain relief as the primary reason they had not engaged with follow-up, and only three had stopped using their system owing to lack of effect.
The NNR aims to capture data on all neuromodulation implants in the UK, including outcome data.The ideal real-world data set would be the entire population having an intervention, not a sample of the population.We are optimistic that procedural and device data for the entire UK practice will soon be routinely entered into the registry because the NHS has recently mandated the usage of registries for all medical implants, and centers that do not enter their implant data will face decommissioning.Baseline QoL data and demographics will continue to be entered at the same time.The focus is now on the completeness of follow-up data, and we are taking active steps to improve this.A new NNR portal app has been produced that is believed will make it easier for patients to enter their data, and all patients are being issued with a card with access instructions to the portal, including a quick response (QR) code shortcut.It is envisaged that with increased ease and availability of digital and remote access for patients and clinicians, compliance will improve, without sacrificing the independent nature of data collection.The effect of these changes will be audited, and adjustments made as and when necessary depending on patient and clinician feedback.
The power of registry data increases with the size of the data set.Efforts to form international registries are critically important in the field and are under way.Merging data from different localities requires care: For example, EQ5D data are defined and validated in large studies on a country-by-country basis, given different social and cultural norms can cause greater or lesser emphasis and importance to perceived QoL being assigned to different facets of the EQ5D-5L tool.

CONCLUSIONS
This study comprises real-world data from a large patient population.It shows that SCS has a significant benefit in terms of overall health-related QoL for up to three-quarters of patients who are selected to undergo this therapy with currently used patient selection protocols.
Neuromodulation for chronic pain increases patients' QoL in line with or more than many other typically performed surgical procedures for painful conditions and from a lower baseline QoL.

Figure 1 .Figure 2 .Figure 3 .Figure 4 .
Figure 1.QoL is improved six months after SCS and is sustained 12 months after implantation.Bee swarm plot showing the frequency and median ± IQR of the EQ5D utility for patients before treatment, six months after implantation, and 12 months after implantation.Each black point represents one patient; the red horizontal lines are the median, and the vertical lines are the IQRs.QoL was significantly higher six months after implantation (EQ5D utility 0.263 vs 0.550, p < 0.0001, Wilcoxon rank sum test), and this change was sustained at 12 months after implantation (EQ5D utility 0.550 vs 0.548, p = 0.15, Wilcoxon rank sum test).[Color figure can be viewed at www.neuromodulationjournal.org]

Figure 5 .
Figure 5. Change in utility after SCS with lumbar PSPS.Bee swarm plot shows each individual's change in utility-red dot is the mean value; the red line shows the 95% CI.The mean change was an improvement in the utility of 0.205; 75.2% of patients showed a positive response to SCS. [Color figure can be viewed at www.neuromodulationjournal.org]

Figure 6 .
Figure 6.Change in utility after SCS with lumbar PSPS subcategories.Bee swarm plots of change in utility for each subcategory.The red dot shows the mean value.[Color figure can be viewed at www.neuromodulationjournal.org]

Figure 7 .Figure 8 .
Figure 7. Change in utility after SCS with cervical PSPS.Bee swarm plot of change in utility overall for cervical PSPS.The red dot shows the mean value; whiskers show the 95% CI. [Color figure can be viewed at www.neuromodulationjournal.org]

Figure 9 .Figure 10 .
Figure 9. Change in utility SCS-patient compliance.Bee plot of change in utility; the red shows the mean values.This shows the baseline and postoperative utility for participants with complete and partial compliance with follow-up.[Color figure can be viewed at www.neuromodulationjournal.org]

Figure 11 .
Figure 11.Panel a shows preoperative QoL for common elective procedures.The average preoperative EQ5D utility is lower for patients with SCS than in other commonly performed operations.Panel b shows change in QoL for common elective procedures.Bottom of arrow shows baseline; tip of arrow shows postoperative utility for commonly performed operations.Dashed blue lines show the UK (England) mean EQ-5D-5L utility for ages 20, 50, and 80 years as annotated.Panel c shows absolute improvement in EQ5D utility before and after treatment.SCS ranks among the most effective surgical treatments for improving QoL regarding absolute change in EQ5D utility.Panel d shows percentage improvement in EQ5D utility before and after treatment.SCS ranks as the most effective surgical treatment for improving QoL regarding percentage change in EQ5D utility.[Color figure can be viewed at www.neuromodulationjournal.org]

Table 1 .
Indications for SCS and Their Relative Frequency in the Registry.

Table 2 .
The Proportion of Patients With Each Reported Subtype of Lumbar PSPS.

Table 3 .
Proportion of Patients With Each Reported Subtype of Cervical PSPS.

Table 4 .
Most Performed Elective Procedures in the UK.