Article Text
PDF
Abstract
Background and objectives Radiofrequency ablation (RFA) is a common treatment modality for chronic axial spine pain. Controversy exists over its effectiveness, and outcomes in a real-world setting have not been evaluated despite increasing use of RFA. This study examined changes in healthcare utilization and opioid use after RFA in Ontario, Canada.
Methods This retrospective cohort study was conducted in Ontario using administrative data. Ontario residents receiving their initial RFA between 1 January 2009 and 31 March 2015 were included. Physician visits, spinal injections, and opioid dosing/prescriptions in the 12-month periods before and after RFA were compared.
Results The study included 4653 patients. The number of RFA procedures significantly increased from 2009 to 2014 (22.5 cases/1 000 000 person-years to 82.5 cases/1 000 000 person-years). 4465 patients had at least one physician visit pre-RFA; there was a significant 23.89% reduction in physician visits post-RFA (pre-RFA: 29 616 visits; post-RFA: 22 542 visits). All reviewed specialties demonstrated a decrease in physician visits post-RF except neurosurgery. 3445 (85.70%) fewer spinal interventions for axial pain (medial/lateral branch blocks, facet/sacroiliac injections) were performed post-RFA. Significantly fewer epidurals were also performed post-RFA. 198 of 1007 patients (19.66%) on the Ontario Drug Benefit who received opioids pre-RFA did not require a postprocedure opioid prescription. Mean opioid dosing was unchanged post-RFA.
Conclusions Healthcare utilization was significantly reduced in the 12 months following RFA, and some patients eliminated opioid use. Selection criteria for RFA are not standardized in Ontario, and appropriate use guidelines for spine interventions may improve outcomes and reduce unnecessary procedures.
- interventional pain management
- outcomes
- radiofrequency ablation
Statistics from Altmetric.com
Introduction
Chronic pain is a significant burden on healthcare resources, costing the Canadian healthcare system an estimated $7.2 billion annually.1 It is associated with increased healthcare utilization, including an increased frequency of physician visits.1–3 Treatment is often challenging, and there has been a wide reliance on opioids for management.4 5 Recently, there has been significant concern over the potential misuse of opioids and the negative effects of these medications, including physical dependence and addiction.6 7
Management options that can minimize the need for opioid medications and reduce healthcare costs for those with chronic pain are increasingly important. Radiofrequency ablation (RFA) is a common, minimally invasive interventional procedure for the management of chronic pain involving the axial spine. RFA involves the application of thermal energy to denervate putative pain generating structures, including the facet joints and the sacroiliac joint. Some studies have demonstrated excellent benefit following RFA, with 53%–90% of patients achieving >90% reduction in pain severity for at least 12 months.8–12 RFA success has been associated with reduced healthcare follow-up8 9 and decreased analgesic use.8 13
Although some studies demonstrate excellent benefit with RFA, other studies have not replicated this success.14–17 Several reasons have been offered for the lack of treatment success in these studies, including variable criteria for patient selection, different procedural approaches, different technical factors (eg, needle sizes) and the presence of medical comorbidities.8 18 19 As illustrated by variable technical and selection factors across studies, a single management paradigm is not consistently applied in practice.
Despite controversy in terms of efficacy, the use of RFA in the US Medicare population increased by 544% for lumbosacral procedures and 836% for cervicothoracic procedures from 2000 to 2011.20 Given ongoing controversy regarding the benefit of RFA, variations in clinical practice, the current focus on minimizing opioid usage and the need to contain healthcare costs, it is important to evaluate RFA outcomes in a real-world setting. The primary purpose of this study is to characterize the utilization of RFA, and evaluate changes in healthcare utilization and opioid usage before and after RFA using population-level data.
Methods
Design and setting
A retrospective cohort study was conducted using administrative data from the province of Ontario, Canada. The study included residents of Ontario who received their initial RFA for axial spine pain between 1 January 2009 and 31 March 2015. Ontario is Canada’s most populous province and the entire population has access to a single government-funded healthcare system. Universal medication coverage is provided for those 65 years or older. The use of data in this project was authorized under section 45 of Ontario’s Personal Health Information Protection Act, which does not require review by a research ethics board. Individual patient consent was not required.
Data sources
Several linked administrative data sources were used for this study. These included the (1) Registered Persons Database for demographic information; (2) patient contact and eligibility yearly files for prior patient healthcare contact information; (3) Ontario Population estimates and Projects, distributed by the Ontario Ministry of Health and Long-Term Care: IntelliHEALTH Ontario, to calculate rates of RFA procedures over time; (4) Ontario Drug Benefit (ODB) database, containing all outpatient prescription drug use for those on social assistance or aged 65 years or older; (5) Canadian Institute for Health Information Discharge Abstract Database/Same Day Surgery identifying all hospital admissions and procedures; (6) National Ambulatory Care Reporting System identifying all emergency room encounters; (7) Ontario Health Insurance Plan (OHIP) containing all physician billing codes for patient assessment or treatment; and (8) the Institute for Clinical Evaluative Sciences (ICES) Physician Database identifying physician specialties. These data sets were linked using unique encoded personal identifiers and analyzed at ICES Western. The validity of the data elements in the above databases has been previously documented.21–25
Patient population
OHIP billing codes were used to identify patients who had an RFA procedure for axial spine pain between 1 January 2009 and 31 March 2015. The first occurrence of the RFA OHIP billing code was considered the index date. Exclusion criteria for the main cohort were: RFA in the 5 years prior to the index date (to ensure we were identifying index treatments), no healthcare contact between 3 and 5 years prior to the index date (to ensure patients would have a minimum 2-year administrative data history), and non-residents of Ontario.
Characteristics of the population, including age, gender, number of levels ablated (i.e. number of instances of the OHIP codes N556 and E396 within the first 90 days), spinal regions injected prior to RFA, physician specialty performing RFA, and rate of index cases over time were described.
Possible procedural complications were assessed by documenting the number of emergency room visits within 14 days of the index date and the number of CT and/or MRI scans performed in the first 30 days postprocedure.
Within our primary cohort, we further restricted the patient population for our various specific outcomes:
Physician visits subcohort: Additional exclusion criteria included no relevant physician visits in the preperiod.
Spinal interventional procedures subcohort: The subcohort was limited to those who had their initial RFA between 1 January 2013 and 31 March 2015. This is because OHIP billing codes for facet joint injections/medial branch blocks, and sacroiliac joint injections were only introduced in September 2012. Patients with no relevant preperiod spinal interventions were excluded.
Opioid utilization subcohort: Additional inclusion criteria included ODB coverage. Patients with no pre-RFA and post-RFA opioid prescriptions were excluded.
Further exclusion criteria in all subcohorts included no healthcare contact between 1 and 2 years after the index date, or death within 1 year after the index date.
Study outcomes
The primary objectives were to evaluate healthcare and opioid utilization in the immediate 12-month periods before and after RFA. Healthcare utilization was assessed through changes in (1) overall physician visits; (2) physician visits in family medicine, physical medicine and rehabilitation (PMR), orthopedics, neurology, neurosurgery, or anesthesiology, and (3) axial spine interventional procedures other than RFA (cervical/thoracic/lumbar facet injections or medial branch blocks, and sacroiliac joint injections). OHIP billing codes specific to each physician specialty for consultations/assessments/reassessments were used to determine the number of physician visits. OHIP billing codes for the different spinal procedures were used to determine the number of interventional procedures before and after RFA. As the OHIP billing codes for medial branch blocks and facet injections are the same, the number of axial spine procedures in the period from 12 to 24 months before RFA was also compared with the post-RFA period. This was done as any pre-RFA axial spinal procedures in the immediate 12-month period before RFA may have been part of a diagnostic block protocol, rather than therapeutic injections. It was assumed that injections in the pre-RFA period older than 1 year were more likely to be therapeutic injections.
Opioid utilization was assessed by comparing (1) total daily oral morphine equivalent (MEQ) dosing, and (2) total number of opioid prescriptions, in the year before and after RFA (figure 1, time period A). Because of the possibility of a postprocedural opioid prescription for acute pain, we also performed an analysis excluding prescriptions ±30 days around the index date (figure 1, time period B). Analyses assessing 90-day periods pre-RFA (between 1 and 4 months prior to RFA) and two separate periods post-RFA (between 4 and 8.5 months, and 8.5 and 12 months) were done to account for any time required to taper opioid doses, in order to better demonstrate any potential long-term reduction in opioid use (figure 1, time periods C and D). The changes in pre-RFA opioid doses for individual patients following the ablation procedure were also documented.
Opioid analysis time frames. Time period A compares opioid prescriptions and morphine equivalent (MEQ) in the full 12-month periods before and after radiofrequency ablation (RFA); time period B excludes 30-day periods around the index date. Time periods C and D compare a pre-RFA 90-day time frame to two separate post-RFA time frames.
Secondary outcomes included the proportion of people who received a second RFA, time between subsequent RFAs, and changes in opioid dosing/prescriptions for those who had a second RFA. Subsequent RFAs needed to be at least 90 days apart to ensure that the second RFA was not part of the first procedure (ie, left and right sides done on different visits). The number of musculoskeletal ultrasound-guided injections and epidural injections before and after RFA was used as a control group since the anatomical targets and therapeutic rationale for these procedures and RFA differ.
Statistical methods
Differences between pre-post RFA groups were tested using a paired t-test when the normality assumption held. However, in the majority of groups the normality assumption did not hold, and Wilcoxon signed-rank tests were used instead. In all pre-post group comparisons a two-sided p value <0.05 was considered significant. A test for trend on the rate of index RFA procedures between 2009 and 2014 was carried out using linear regression, where a two-sided p value <0.05 was considered significant. The range of differences in total daily oral MEQ dosing pre-post RFA was shown using a waterfall plot and a Wilcoxon signed-rank test was performed to compare these differences. Descriptive statistics were used to describe distributions of the pre/post RFA groups. All analyses were conducted using SAS V.9.4 (SAS Institute).
In accordance with ICES privacy policies, if any categorical value was less than or equal to 5, the exact number was not reportable because of reidentification risk, and the result was reported as ‘≤5’. If only one value in a set was ≤5, the exact number of a second value had to be suppressed, and was reported as ‘≤x’ in any results and totals, where x was any number greater than or equal to the true value. Differences and percentage changes were still calculated using exact numbers.
Results
A total of 4653 patients (2773 female; 1880 male) had an initial RFA between 1 January 2009 and 31 March 2015 and met study inclusion criteria (figure 2). The mean age of patients receiving RFA was 58.25±14.94 years (table 1). Most commonly, six medial/lateral branches were ablated (table 1). When specified, most patients receiving RFA had previously received lumbosacral spinal interventions only (ie, lumbar medial branch blocks, sacral lateral branch blocks or lumbar facet joint injections), with cervical interventions only (medial branch blocks or facet joint injections) the second most common (table 1).
Flowchart of cohort and subcohort size determinations. ODB, Ontario Drug Benefit; RFA, radiofrequency ablation.
Sample characteristics of those receiving their initial RFA in Ontario from January 2009 to March 2015
Anesthesiologists performed 83.1% of all RFA procedures (table 1). Additional physicians performing RFA procedures included physiatrists (4.19%), neurosurgeons (4.51%), radiologists (2.15%) and other specialties (6.04%). There was a significant increase in the number of patients receiving RFA from 2009 to 2014 (table 2), with the rate increasing from 22.5 cases/1 000 000 person-years in 2009 to 82.5 cases/1 000 000 person-years in 2014 (p=0.0001). A total of 151 patients (3.25%) visited an emergency room within the first 14 days post-RFA. Seventy-eight patients (1.68%) had a CT and/or MRI within the first month postprocedure.
Annual population, number of initial RFA cases and rate of index RFA procedures in Ontario, Canada per year from January 2009 until March 2015
Physician visits subcohort
A subcohort of 4465 patients between 1 January 2009 and 31 March 2015 was identified (figure 2). A 23.89% reduction in overall physician visits was seen following RFA (p<0.0001) (table 3). Family medicine visits were reduced by 42.84% (p<0.0001). All specialties that were evaluated, except neurosurgery, demonstrated a statistically significant decrease in patient visits following RFA (specialty, % decrease in visits): neurology, 47.89%; orthopedic surgery, 35.50%; anesthesia, 21.80%; PMR, 41.83% (p<0.0001). Neurosurgery visits demonstrated a non-significant increase of 148.49% (p=0.068) (table 3).
Differences in the number of physician visits pre-RFA and post-RFA
Spinal interventional procedures subcohort
A subcohort of 2193 patients was identified (figure 2). When comparing the immediate 12-month period before RFA to the 12-month period after RFA, there was an 85.70% reduction in spine interventional procedures (medial branch block, lateral branch block, facet joint or sacroiliac joint injection) following RFA (p<0.0001). The largest reductions in interventional procedures were for patients receiving facet injections/medial branch blocks pre-RFA (ranging from an 80.0% reduction in thoracic procedures after RFA to a 90.13% reduction in cervical procedures). For other interventions, the need for procedures post-RF was also reduced, but to a lesser degree (ranging from 38.65% for lumbar epidurals to 69.70% for thoracic epidurals). 62.64% of those receiving a pre-RFA sacroiliac joint injection did not require one post-RFA. Miscellaneous musculoskeletal ultrasound-guided procedures increased by 4.09% after RFA (p<0.0001) (table 4).
Differences in the number of spinal injections pre-RFA and post-RFA
A total of 266 patients had a facet injection, medial/lateral branch block or sacroiliac joint injection in the 24 to 12 months pre-RFA. When this time frame is compared with the immediate 12-month period post-RFA, 81.35% fewer procedures were performed post-RFA. The reductions in specific procedures post-RFA in this subcohort were similar to the immediate 12-month pre-RFA and post-RFA comparison (table 4).
Opioid utilization subcohort
A total of 1219 patients met inclusion criteria for this analysis (figure 2). In this population, average MEQ was comparable in the 12-month periods before and after RFA (mean±SD; 19.19 mg±45.22 vs 20.68 mg±47.94). When examining the patterns of change of MEQ after RFA, 47% had an increase, 15% had no change, and 38% had a decrease in their MEQ dosing post-RFA (figure 3).
Waterfall plot of differences in total daily oral morphine equivalent (MEQ) dosing 1 year pre-RFA and post-RFA. Although there was an overall increase in daily MEQ dosing for the analysis of all RFA patients on Ontario Drug Benefit (ODB) meeting inclusion criteria, this was due to pre-RFA opioid-naïve patients starting opioids. In the analysis restricted to only those patients receiving opioids pre-RFA, there was a significant decrease. In accordance with Institute for Clinical Evaluative Sciences (ICES) privacy policies, patients were grouped in sizes of ≤6. RFA, radiofrequency ablation.
Of 1007 patients who had at least one opioid prescription in the 12-month pre-RFA period, 198 (19.6%) did not require one in the 12-month post-RFA period. No clinically apparent differences in the number of opioid prescriptions provided were seen in the different time periods examined post-RFA, although there was a statistically significant decrease in median MEQ post-RFA (table 5).
Opioid utilization before and after RFA
Repeat RFA procedures
41.35% of the full cohort had a second RFA. The mean time (±SD) to a second RFA was 432.20±375.28 days with a median time of 312 days (IQR 183–545 days). Opioid utilization (pre-RFA MEQ dosing; post-RFA MEQ dosing; percentage not requiring prescriptions post-RFA) in those who had a second RFA procedure (23.75±49.38 mg; 20.67±47.43 mg; 21.08%; n=370) was not substantially different from the full subcohort (23.23±48.80 mg; 22.90±51.83 mg; 19.66%; n=1007).
Discussion
This is the first study, to our knowledge, that documents population-level outcomes following RFA for chronic axial spine pain. As in the US Medicare population,20 a marked increase in the number of RFA procedures in Ontario over time was seen. Despite controversy relating to the effectiveness of RFA in clinical trials, this study identified significantly reduced healthcare utilization, including fewer spinal interventional procedures and physician visits post-RFA. Additionally, 19.66%–24.11% of patients eliminated opioid use after RFA. A previous 10-year prospective clinical audit of a single interventional practice in Ontario, Canada demonstrated a positive benefit from RFA12; this study suggests a positive effect of RFA at a broader population level in Ontario.
Positive findings in this study were observed even though RFA procedures are not currently standardized in Ontario. Criteria for patient selection (including the requirement for diagnostic blocks) and procedural techniques are at the discretion of the treating physician. Reductions in healthcare utilization and opioid use may therefore improve if standardized selection criteria for RFA were introduced.
Healthcare utilization: interventional procedures and physician visits
While the need for spinal interventional procedures decreased post-RFA, this may have been partially due to the performance of diagnostic blocks within the first 12 months pre-RFA. Physician billing codes in Ontario are the same for facet joint injection and medial branch blocks, so it was not possible to distinguish these interventions from each other. To ensure that pre-RFA interventions were less likely to represent diagnostic blocks and more likely to represent therapeutic interventions, an additional 24 to 12-month time frame pre-RFA period was evaluated. It was found that the decrease in the number of spinal injections post-RFA was similar between the two pre-RFA time frames, suggesting that there was a reduced need for therapeutic spinal interventions following RFA.
Unexpectedly, a significant decrease in epidural injections post-RFA was seen, although the therapeutic rationale and anatomical target for epidural injections are different from RFA.26 In contrast, the number of ultrasound-guided musculoskeletal injections was relatively unchanged post-RFA, as expected. While it should be noted that the decrease in epidural injections was not as large as the change in medial branch blocks and facet injections post-RFA, the decrease was still significant. It is possible that epidural injections are still being performed for primarily mechanical spine pain in Ontario, reinforcing the need for standardized selection criteria to ensure appropriate use of interventional modalities. Choosing Wisely Canada, which seeks to reduce the unnecessary use of procedures and treatments, has recommended against the use of epidural injections for those who have primarily axial pain without leg dominant pain originating from the nerve roots.27 Dissemination of this recommendation may improve practice patterns in the future.
A reduction in physician visits was seen in general practice and all specialties except for neurosurgery post-RFA, which demonstrated a non-significant increase in visits. A relatively small number of emergency room visits and CT/MRI scans were seen immediately after RFA, and neurosurgeons performed only 210 of all RFA procedures in the observed time frame. Given this, it is highly unlikely that procedural complications or postprocedural follow-ups accounted for the increase in neurosurgical visits post-RFA.
Opioid use
There has been increasing awareness and acknowledgment of complications associated with opioids.28 Alternatives to opioids for reducing pain intensity in those with chronic pain are therefore important. This study demonstrates that RFA allows some patients to reduce or eliminate opioid use, suggesting that RFA may be an important alternative to opioids for the appropriately selected patient.
Mean overall dosing, however, was comparable before and after RFA due to the average effect of some patients increasing, and others decreasing, their utilization. Forty-seven per cent increased their opioid dosing post-RFA; a likely cause for this was the initiation of opioids in those who had not received opioids previously. Although the underlying reasons for opioid initiation are not clear, opioids for acute procedure-related pain, or the need for additional management options after failure of RFA are possibilities. Fifteen per cent of patients maintained their opioid dosing, which may relate to failure of the procedure or the use of opioids for unrelated pain conditions. In addition, tapering opioids can be challenging29; opioid use may not decrease despite RFA success. This possibility is supported by the finding that those who had a second RFA procedure (and were therefore suspected of having some degree of benefit from initial RFA) had similar mean opioid doses pre-post RFA as the entire opioid cohort.
Changes in opioid dosing could only be evaluated in those on ODB, which includes people over 65 and adults on social assistance. Prescription information for the remainder of the population was not captured in provincial databases during the study. Because opioid use in only a subset of the sample could be analyzed, the results for opioid use are less generalizable.
Limitations
Limitations of this study arise from its retrospective nature. Although a comprehensive administrative provincial database was used, the possibility of misclassification exists. In addition, causation could not be definitively established. Treatment success based on pain intensity reduction could not be directly evaluated; healthcare utilization and opioid usage were used as surrogates of clinical success. Post-RFA changes in these outcomes may be due to reasons other than decreased pain. For example, clinicians and patients may view RFA as a last resort, and patients may not require physician follow-up to the same degree post-RFA if it is felt that other management options would not be useful. Conversely, there are numerous reasons that outcomes may be better than those observed, including:
1)Spinalregion targeted with RFA
Our study was unable to differentiate the site that was targeted for ablation. The physician billing code for RFA in Ontario (N556) is the same for cervical, thoracic, lumbar, and sacroiliac-level procedures and does not require identification of the region targeted. Success following RFA, however, is partially dependent on the site that is ablated; for example, clinical studies of RFA for sacroiliac joint pain have demonstrated less success in reducing pain than those for cervical RFA.30 31 Using the codes for medial/lateral branch block and facet joint injections, we attempted to extrapolate the site that was most likely targeted with RFA. Many patients in our sample had unspecified sites of previous spinal injections (44.36%), or had multiple spinal regions injected previously (10%). Specific billing codes reflecting the spinal region of diagnostic blocks and facet/sacroiliac injections were not introduced until September 2012, accounting for the large proportion without any spinal injection billings prior to RFA. Patients who had lumbosacral-level injections only (35.70%) were the largest group of identified single-region injections pre-RFA. Unfortunately, the same billing code (G912) covers multiple lumbosacral spinal injections (specifically, lumbar facet injection, lumbar medial branch blocks, and sacral lateral branch blocks). This could significantly skew results since RFA of the lateral branches for sacroiliac joint RFA has been less successful in clinical trials than lumbar medial branch RFA.19 31 It was therefore not possible to determine the effectiveness for RFA in different spinal regions. RFA success in the study could be higher than observed, given the potentially wide variety of spinal regions targeted for RFA, with differing expectations for success.
2)Ongoing benefit beyond study timeframe
The average time to a second RFA procedure in our study was similar to the duration of benefit following RFA in clinical trials.9 19 32 33 The proportion of patients who received a second procedure, however, was only 41.35%, which is on the lower end of the range for those achieving success from RFA in clinical trials.30 31 While the percentage of those requiring a second procedure cannot be taken as a direct corollary of treatment success, it represents to some extent the proportion of patients in which effectiveness was sufficient to justify a repeat procedure. Given that the strict conditions within clinical trials are harder to reproduce in a clinical setting, decreased success was expected in a real-world environment. However, it is possible that the proportion achieving significant relief after RFA is higher than those requiring a repeat procedure, as some people may have prolonged relief that did not require a repeat RFA in the observed time frame.
Conclusions
Reduced healthcare utilization, including fewer physician visits and spinal injections, was seen after RFA in Ontario, Canada. Although mean opioid dosing did not change after RFA, 19.66%–24.11% of those receiving an opioid prescription before RFA no longer required one after RFA. The implementation of standardized criteria for interventional pain procedures in Ontario could improve these outcomes. A detailed economic analysis of the potential cost savings associated with reduced healthcare utilization after RFA would be an important next step in this study, and may strengthen the justification for the ongoing clinical use of this modality.
Acknowledgments
We thank IMS Brogan for use of their Drug Information Database. Parts of this material are based on data and/or information compiled and provided by the Canadian Institute for Health Information (CIHI).
References
Footnotes
Presented at Preliminary data from this study were presented at the Spine Intervention Society Meeting in Chicago, IL, on 18 August 2018, and will be published as an abstract in Pain Medicine.
Funding This study was funded by an Earl Russell Trainee Grant awarded to FA and supported by the Institute for Clinical Evaluative Sciences (ICES) Western site. ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Core funding for ICES Western is provided by the Academic Medical Organization of Southwestern Ontario (AMOSO), the Schulich School of Medicine and Dentistry (SSMD), Western University, and the Lawson Health Research Institute (LHRI).
Disclaimer The analyses, conclusions, opinions, results and statements expressed in the material are those of the author(s) and not necessarily those of CIHI, and are independent of the funding sources. No endorsement by ICES, AMOSO, SSMD, LHRI, or the MOHLTC is intended or should be inferred.
Competing interests EL has received in-kind research support from Nimbus Concepts and Diros Technologies for other projects.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.