Minimal Clinically Important Difference in Patient-Reported Outcome Measures with the Transforaminal Endoscopic Decompression for Lateral Recess and Foraminal Stenosis

INTRODUCTION

Making a case for transforaminal endoscopic spinal decompression surgery for a lumbar herniated disc or foraminal stenosis causing sciatica-type low back and leg pain hinges on providing comparative clinical evidence to alternative translaminar decompression surgeries. The commonly used patient-reported outcome measures (PROMs) include the visual analog score (VAS)^1–12 and the Oswestry Disability Index (ODI).^13–18 Understanding the ability of these PROM scores to detect changes in health status in response to an intervention meaningful to the patient is critical to support conclusions in favor of one treatment over another. While PROMs were designed to more directly involve patients in their care and improve their participation in the judgment of its outcomes, the minimal clinically important difference score (MCID) was developed by Jaeschke et al¹⁹ in 1989 in an attempt to deal with the problem that analysis of a PROM instrument measuring change after intervention in some cases may show statistical significance without any clinical relevance. Therefore, MCID is determined as the threshold value of a spine care outcome instrument that spine surgeons and their patients perceive as clinically meaningful.

With endoscopic spinal surgery moving more into the mainstream,^{4,5,9–11,13,20–35} the debate on its long-term clinical outcomes and recommended indications will likely intensify. It may be fueled by technology advancements but more likely by innovative spine surgeons who advocate to replace traditional lumbar open translaminar with transforaminal endoscopic decompression surgeries on the merits of improved outcomes, and lower cost due to fewer peri- and postoperative complications and reoperations. In the hands of a well-trained and experienced endoscopic spine surgeon, this outpatient procedure is associated with an overall lower burden to patients, and excellent long-term durability of favorable outcomes, especially since the procedure can be performed under local anesthesia and eliminate the use of general anesthesia.²⁹ The authors of this study stipulated that the MCIDs for VAS and ODI as a result of the outpatient transforaminal endoscopic decompression may be different from those published^36–52 for open translaminar decompression as the perception of its clinical benefit well known to patients is driven by the stark differences between the two procedures: a clean small stab versus an open incision with risk of infection, in most cases an outpatient procedure at an ambulatory surgery center) under local anesthesia versus an inpatient surgery at a hospital, minimal bleeding and incisional discomfort versus wound pain with potential for blood loss, and, last but not least, faster recovery and social reintegration versus longer narcotic dependence, delayed return to work, and higher direct and indirect cost. With the advent of the internet and social media this easy-to-understand context of improved standards of care with the endoscopic over traditional translaminar surgery is becoming common knowledge among patients and one wonders how patient perception and, hence, the respective MCID threshold values are impacted by these trends. It is well known that MCIDs are not static numbers and are heavily influenced by a myriad of patient demographic factors, the individual baseline severity of the disease, and the dynamic of a recall bias of the intrinsic nature of their prior condition while comparing the current functional status against expectations which are heavily impacted by the public discourse about contemporary minimally invasive spine surgery techniques. Moreover, MCIDs are also influenced by the type of procedure performed.

One crucial premise for endoscopic spine procedures primarily offered for pain and symptom reduction is the concept of a staged surgical procedure that typically provides more effective and longer-lasting symptom relief when compared to nonsurgical and pain management ablation options, which by their nature are intended as an intermediate step before considering surgery by providing a temporary reduction of symptom.²⁹ Traditionally trained spine surgeons are often focused on employing image-based threshold criteria to define the surgical indication for correcting symptoms stemming from spinal stenosis, deformity, and instability. Traditional surgical spine care tends to be more expensive and associated with higher perioperative risk since definitive treatment tends to be directed at the patients' symptoms when the degenerative spine disease has reached its end stage by performing an aggressive salvage surgery that often involves instrumented fusion. The concept of early and staged surgical pain management with an endoscopically visualized treatment of established and validated pain generators is getting traction amongst those spine surgeons who recognize the need for a paradigm shift toward more personalized and cost-effective spine care to meet the demands by patients and payors alike. This ongoing paradigm shift, however, creates an entirely new clinical context of spine care for patients and surgeons who provide it. Therefore, the authors decided to take a fresh look at MCIDs with VAS and ODI in transforaminal endoscopic spinal surgery patients in an attempt to validate these commonly used PROMs with the intent to improve the clinicians' ability to identify those endoscopic treatments of common painful conditions of the lumbar spine associated with better clinical outcomes when directly compared with traditional open translaminar surgeries.

MATERIALS AND METHODS

Patients

All 406 consecutive patients included in this retrospective study were treated between 2012 and 2017 for claudication sciatica-type low back and leg pain due to contained herniated disc with or without bony and soft tissue lateral recess and foraminal stenosis due to age-related degeneration of the spinal motion segment demonstrated on preoperative magnetic resonance imaging (MRI). The associated lumbar disc herniations, facet joint hypertrophy, and the overall ligamentous and bony overgrowth was treated with a directly visualized outpatient endoscopic transforaminal decompression procedure in an ambulatory surgery center which often employs an initial foraminoplasty to gain access to the triangular safe zone between the exiting and traversing nerve root. The procedural details employed by the authors have been described elsewhere.^{1,2,17,18,53,54} All patients sought out the endoscopic spine surgeons who authored this study and provided informed consent. Patients were matched to age, gender, and diagnosis to avoid the introduction of additional confounding factors or unforeseen biases. The mean follow-up was 33.59 months, ranging from 24 to 85 months, with a standard deviation (SD) of 12.79. The mean age was 41.08 years ranging from 30 to 84 years with an SD of 12.74 with a trimodal distribution (Figures 1 and 2) with patients between the ages of 30 and 35 years making up 53.8% of the entire study population. The second-largest group of patients was between the ages of 39 to 46 years of age (13.9% of the study population), followed by the third group of patients between the ages of 51 to 65 of years (17.6% of the study population). There were 229 females (56.4%) and 177 males (43.6%) in the study population. The inclusion and exclusion criteria for this study have been published elsewhere²⁹ in detail and are briefly described in the following section.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Age distribution of the 406 study patients with the superimposed expected normal distribution if it existed (black line). Patient's age ranged from 30 to 84 years of age and averaged 41.08 years with an SD of 12.74. There was with a trimodal distribution with patients between the ages of 30 and 35 years making up 53.8% of the entire study population (frequency table not shown). The second largest group of patients was between the ages of 39 to 46 years of age (13.9% of study population) followed by third group of patients between the ages of 51 to 65 of years (17.6% of study population).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

The trimodal age distribution of the 406 study patients is illustrated in the detrended normal P-P plot of endoscopy patients' age with an estimated location parameter of 3.67.

RESULTS

The 406 patients who underwent an endoscopic transforaminal decompression procedure reported significant reductions of the mean ± SD VAS from preoperative (8.0813 ± 1.46255) to postoperative (2.2463 ± 1.55823) values (P < .0001; Table 1) on paired t testing. Similar statistically significant reductions were observed from the preoperative ODI of 47.46 ± 8.624 to a postoperative ODI 13.98 ± 6.197 (P < .0001). Patients rated how their expectations were met by the achieved clinical outcomes from the transforaminal endoscopic decompression procedure by selecting one of the following PSI options with the following frequency: excellent, 224 patients (55.2%); good, 112 patients (27.6%); fair, 41 patients (10.1%); and poor, 29 patients (7.1%). The dichotomized PSI results showed 377 patients (92.9%) as improved, and the remaining 29 patients (7.1%) as failed (Table 2). Preoperatively, 31.3% (137) of the 406 study patients had a significant limitation of walking endurance of fewer than 300 m, whereas 50.2% (214) of patients could not walk farther than 500 m (Table 3). Only 132 patients (32.5%) were essentially able to walk without limitations by indicating that their walking endurance was higher than 5000 m. Postoperatively, the overwhelming majority of patients (315/406; 77.6%) indicated unlimited walking endurance (≥5000 m), a statistically significant improvement on paired t testing (P < .0001; Table 3). Cross-tabulation of dichotomized PSI versus walking endurance showed that of the 377 improved patients, 69 patients were able to walk up to 1000 m, and the remaining 299 improved patients were able to walk more than 5000 m. More than half (16/29) of the failed patients were also able to walk more than 5000 m after their successful transforaminal endoscopic decompression procedure. These differences in walking endurance between improved- and failed-patient dichotomized PSI ratings were statistically significant (P = .002).

The ROC curve for postoperative VAS scores given by patients who underwent outpatient transforaminal endoscopic decompression surgery is shown in Figure 3. The AUC individual test results for postoperative VAS are listed in Tables 5A and 5B. The AUC was 0.926, with an asymptotic 95% CI lower limit of 0.882 and an upper limit of 0.97. The ROC curve for postoperative ODI scores given by the same patients is shown in Figure 4. The AUC individual test results for postoperative ODI are listed in Tables 6A and 6B. The AUC was 0.751, with an asymptotic 95% CI lower limit of 0.663 and an upper limit of 0.840. The VAS MCID and ODI MCID values using Youden optimization for equally important sensitivity and specificity were 3.0 and 15, respectively. Left-upper-corner methodology indicated a range of VAS MCIDs from 2.5 to 3.5 and of ODI MCIDs of 14 to 17 to be practical with the highest sensitivity and lowest false positive rate (Tables 5A, 5B and 6A, 6B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Receiver operating characteristic (ROC) curve for postoperative visual analog scale (VAS) scores given by patients who underwent outpatient transforaminal endoscopic decompression surgery. The area under the curve (AUC) individual test results for postoperative VAS are listed in Tables 5A and 5B. The AUC was 0.926 with an asymptotic 95% confidence interval lower limit of 0.882 and upper limit of 0.97.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Receiver operating characteristic (ROC) curve for postoperative Oswestry Disablity Index (ODI) scores given by patients who underwent outpatient transforaminal endoscopic decompression surgery. The area under the curve (AUC) individual test results for postoperative ODI are listed in Tables 6A and 6B. The AUC was 0.751 with an asymptotic 95% confidence interval lower limit of 0.663 and upper limit of 0.840.

Pearson χ² testing and cross-tabulation of the postoperative VAS score with postoperative walking endurance for the 3 age groups identified in the examination of the age distribution of endoscopy patients of this study (Table 7) showed a statistically significant asymmetric distribution of low VAS scores ranging from 0 to 2 in patients 30 to 35 years of age (P = .013), 36 to 47 years of age (P = .009), and 47 to 84 years of age (P = .023). These low VAS score patients were constituted mostly by patients with a walking endurance of greater than 5000 m. An analogous cross-tabulation for the ODI reported by endoscopy patients showed a similar asymmetric but not statistically significant distribution with the majority of patients reporting ODI values between 5 to 19 a final follow-up (Table 8).

DISCUSSION

This study aimed to analyze the MCID score for VAS and ODI for the directly visualized transforaminal endoscopic lumbar decompression surgery in patients who suffered from lumbar foraminal and lateral recess soft tissue and bony stenosis. The purpose of the study was simple: to better delineate the utility of these 2 commonly used PROMs with the lumbar endoscopic spinal surgery for practical comparisons between spine outcome studies while taking patients' judgment of the treatment effect in the context of their expectations into account.

While there is little doubt that the directly visualized transforaminal endoscopic decompression techniques have become mainstream in most countries and are considered a viable alternative to both open and other types of minimally invasive translaminar decompression surgeries,^{4,5,9–11,13,20–35} high-grade scientific clinical evidence is needed to consolidate its role in the treatment of spinal stenosis and herniated discs by replacing empiric knowledge with well-executed meaningful clinical trials. Making comparisons between such clinical trials relies on understanding the sensitivity, specificity, accuracy, and—as illustrated by this study— optimal threshold values of the outcome instruments used to make a consequential distinction between treatments and study groups. Without such understanding, the most sophisticated study designs, such as those employing prospective randomization protocols, may be flawed from the outset and of little consequence to the day-to-day decision making by the individual practitioner despite statistical significance testing suggesting otherwise. Since patient satisfaction, or lack thereof, drives utilization in spine surgery and pain management, and treatment failures are instantly and always known to the patient who remains in pain and experiences disability with limited walking endurance due to neurogenic claudication, understanding the MCID for commonly used PROMs in the context of patients' expectations with each procedure is crucially important when making decisions for cost-effective, safe, and efficacious spine care.

In this situation, the concept of MCID is useful. However, it is complex and deserves further discussion as it should not be limited to a single static number. In their original description, Jaeschke et al¹⁹ defined MCID as “…. the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient's management.”¹⁹ There are many ways to calculate MCIDs, and no single standard exists, making comparison methodologically challenging. For example, change in clinician reporting, disease state, clinical parameters, effect size, baseline and postintervention data from patients may be anchored to external criteria, such as the PSI used in this study, or to the measurement of internal values of another instrument (distribution approach). The VAS and ODI MCID analyses of this study were developed from patient data and relied on the patients to accurately reflect on improvements from the transforaminal endoscopic decompression procedure from the preoperative baseline. As predicated by others, any seasoned spine surgeon knows that patients, instead of reporting improvements of their current health status in relation to their preoperative function, often make a comparison to their present-day expectations,⁷¹ or the functional status of their “normal” peers⁷² and fall victim to recall bias by failing to honestly remember the prior extent of their intrinsic spinal disability.⁷³ Therefore, such a retrospective comparison renders the use of instruments, such as VAS and ODI, that query the amount of change from the preoperative baseline value, inherently imprecise.

Additionally, MCID values may also be influenced by the severity or extent of the baseline symptoms,⁷⁴ as well as differences between age, education, or socioeconomic status of study populations.^75–77 The impact of these variations may carry forward and play out in expectation-driven patient response measures and therefore render different MCID values and ranges for different spine procedures.⁷⁴ One of the most relevant and common problems, though, is the inability of patients to understand the context of improvement,⁷⁸ a problem that may worsen with more complex or lengthy outcome instruments that consist of multiple questions with many multiple-choice answers. The ODI, for example, is a 10-item instrument versus the VAS being a single-integer instrument with responses ranging from 0 to 10. The more complex ODI can induce more recall bias than the simple VAS score. One could contemplate how patients' understanding of the proper context of improvement may be more impaired with lengthier instruments, such as the SF-36. The ROC curve plots (Figures 3 and 4) and data (Tables 5A, 5B and 6A, 6B) for both PROMs corroborated this problem. The AUC is equivalent to the percentage of randomly drawn pairs, which were correctly identified as having a determined outcome or not. Employing previously published criteria,⁷⁹ the AUC analysis for both PROMs of this study showed that the single-integer VAS score had excellent accuracy (area 0.926; Table 5A) compared to the ODI, which produced only fair accuracy (area 0.751; Table 6A) in assessing MCID with the transforaminal endoscopic decompression procedure.

The MCID data of this study were anchored in the PSI, which was based on the modified Macnab criteria, a well-validated outcome instrument employed in countless endoscopic spine surgery outcome studies.⁶⁸ The analysis of dichotomized PSI showed postoperative improvement in 92.9% (Table 2) of patients who underwent the transforaminal endoscopic decompression procedure. The PROMs employed in this study showed statistically significant reductions (P < .0001) from a mean preoperative VAS of 8.0813 to a mean postoperative VAS score of 2.2463 (Table 1). The ODI reductions were equally significant (P < .0001) from a mean preoperative ODI of 47.46 to a mean postoperative ODI score of 13.98 (Table 1). On the surface, the preoperative VAS scores obtained in the authors' study may appear higher than previously reported numbers. In the authors' opinion, this observation is a representation of the fact that endoscopic transforaminal surgery is employed for different indications than translaminar surgery in younger patients who do not receive surgical treatment by traditional surgeons. Hence, their preoperative PROMS, including VAS, are not reported in comparative clinical studies. One of these diverse indications for endoscopic spine surgery in the lumbar spine includes foraminal HNP, which no matter how small may produce pain often out of proportion to its appearance on an advanced cross-sectional imaging study. Toxic annular tears underneath a highly inflamed dorsal root ganglion that can be debrided endoscopically are another such example of a very painful condition affecting younger patients, many of them are in their 30s. In total, the authors have identified 17 validated symptomatic conditions in the spinal segment, from the central canal to the extraforaminal lateral recess, that are not entirely appreciated by traditional spine surgeons who attempt to decompress the spinal segment but are lacking due to the dependence in indirect decompression and the lack of direct endoscopic visualization of the pain generator in the lateral recess and the foraminal zone. Some of the most prevalent pain generators include (1) an inflamed disc, (2) an inflamed nerve, (3) a hypervascular scar, (4) a hypertrophied superior articular process and ligamentum flavum, (5) a tender capsule, (6) an impacting facet margin, (7) a superior foraminal facet osteophyte, (8) a superior foraminal ligament impingement, (9) a hidden shoulder osteophyte, and many others.^29,33,34

As an additional objective to enable patients to more easily recall functional outcome measure, pre- and postoperative walking endurance was recorded, since it may not be obvious to the reader how the ROC plot shows clinically tangible information in terms of improvement of patient functioning with respect to VAS and ODI. Patients reported a statistically significant improvement of walking endurance following the endoscopic decompression procedure, with 77.6% of patients indicating greater than 5000 m of walking distance without pain—essentially an unlimited walking endurance (P < .0001; Table 3). Dichotomization also demonstrated higher walking endurance at a statistically significant level (P = .002) in improved patients when compared to fair patients (Table 4). Walking endurance was also cross-tabulated against PROMs for the 3 observed age groups because of the trimodal instead of normal age distribution observed in our study group (Tables 7 and 8). Only the VAS instrument was able to demonstrate statistically significant associations between low scores (0 to 2) and postoperative improvements in walking endurance following the endoscopic decompression procedure (up to 1000 m or >5000 m) in all 3 age groups (Table 7). This is consistent with the higher accuracy seen with the AUC testing for the VAS score (area 0.926; Table 5A) than with the ODI (area 0.751; Table 6A).

As the determination of MCID is governed by the desired specificity and sensitivity, reporting MCID as a rigid number is of little use to the discussion of endoscopic spinal surgery outcomes.⁷⁸ Ideally, PROM instruments have a sensitivity and a specificity of 1, which equals a false positive rate (1-specificity) of 0.⁷⁸ The upper left corner of the ROC curve is typically closest to this situation. The Youden index is best employed when sensitivity and specificity are diagnostically equally important as desirable with the endoscopic spinal surgery. The optimal cutoff or MCID threshold value is where the Youden index is closest to 1: the ROC curve itself represents it. Given the difficulties in controlling the limitations of the MCID determination described above, the authors of this study recommend reporting these calculated MCID threshold/cutoff values for the VAS and ODI PROMs with the transforaminal endoscopic surgery as a range. The PSI satisfaction data analysis of this study obtained on patients who successfully underwent the directly visualized outpatient transforaminal endoscopic decompression procedure 92.9% of the time suggested that meaningful MCIDs for the 2 commonly used PROMs were VAS score reductions ranging from 2.5 to 3.5 and an ODI score reduction between 14 and 17 points. These VAS and ODI MCID numbers are known to be context-specific⁷³ and, therefore, needed to be researched for the transforaminal endoscopic procedure in a primary outcome study. The authors could only find 1 study to date—a meta-analysis on full endoscopic decompression for lumbar central and lateral recess stenosis using the interlaminar approach—that investigated the context-specific MCIDs for VAS and ODI.⁴⁰ In the authors' opinion, however, that meta-analysis may have overestimated treatment benefits measured in VAS and ODI reductions (ODI 41.71, 95% CI 39.80-43.62; VAS leg 5.95, 95% CI 5.70-6.21; and VAS back 4.22, 95% CI 3.88-4.56)⁴⁰ since they were statistically significantly higher than the MCIDs reported by its specific studies. In comparison, several original studies employing other types of lumbar decompression procedures for sciatica-type low back and leg pain reported PROM reductions and MCIDs similar to ours. It is conceivable though that these variable numbers are in fact a consequence of a real variability of patients' expectations with this still relatively new operation that is a less-known procedure to the public— a position that has not been researched and independently validated by the authors but should be the focus of future work. The premise of providing sufficient pain relief with early and staged endoscopic surgical pain care with the prospect of that care lasting for years rather than months—as seen with ablation-based interventional spine care—by allowing the natural healing process to occur is a real paradigm shift that is currently being embraced by an increasing number of spine surgeons. This reset of the understanding of the indications and appropriate timing of surgical spine care when employing the transforaminal endoscopic decompression surgery is shifting expectations with patients making the investigation and discussion of MCIDs with commonly used PROMS timely and more relevant than ever.

CONCLUSIONS

Management of patient expectations by explaining the probability of a successful outcome following the directly visualized transforaminal endoscopic decompression surgery to them is crucially essential to direct spine care safely and efficiently. The MCID ranges for VAS and ODI PROMs found in this study may be used in clinical outcome research involving endoscopic spinal surgery since they were context-specific to spinal endoscopy and were obtained on patients who specifically sought out endoscopic spine surgeons to treat their sciatica-type low back and leg pain symptoms. The single-integer VAS score had excellent accuracy and was more reliable in capturing patients' judgment of their outcome than the 10-item ODI questionnaire. It is conceivable that the simple VAS score is less susceptible to recall bias or better captures health status improvements relevant to patients who underwent transforaminal endoscopic decompression than the more complex ODI. The VAS could preferably be employed as an easy-to-understand and more accurate outcome measure by spinal endoscopy patients and their surgeons.