Secondary Logo

Journal Logo

The Spine Blog

Friday, January 17, 2020

Surgical site infection (SSI) is a dreaded complication following spine surgery that results in significant morbidity to patients and high costs to the healthcare system. The topic has been studied extensively, and the past decade has seen a decrease in SSI thanks to the use of intrawound vancomycin powder. In order to provide a broad, epidemiological snapshot of the topic, Dr. Zhou and colleagues from China performed a wide-reaching meta-analysis evaluating the prevalence of SSI and SSI risk factors. They excluded administrative database studies due to their lack of sufficiently detailed clinical information and analyzed 27 studies including over 22,000 patients. They reported an overall SSI rate of 3.1%. The surgical indication with the highest rate was for neuromuscular scoliosis (13%). Somewhat surprisingly, trauma and tumor cases were not associated with an increased rate compared to degenerative cases. The posterior approach had an SSI rate twice that of the anterior approach (5% vs. 2.3%), and instrumentation was also associated with an increased risk of SSI (4.4% vs. 1.4%). Minimally invasive surgery, surgical duration less than 3 hours, and blood loss less than 500 mL also had lower SSI rates. The use of intrawound vancomycin powder had a significant effect on SSI rate reduction (4.8% vs. 1.9%). Staph aureus was the most common pathogen (50% of infections), of which nearly half were methicillin-resistant.

The authors have performed a fairly exhaustive meta-analysis of SSI in spine surgery, and these rates can serve as useful benchmarks. The biggest game changer in SSI prevention—as this study indicates—has been the use of intrawound vancomycin powder, which has markedly reduced the SSI rate, especially in multilevel posterior instrumented fusions. Only 12 of the 27 studies stratified results based on the use of vancomycin powder, and this makes it more difficult to benchmark rates in the vancomycin era. It would have been helpful if the authors had stratified approach (i.e. anterior vs. posterior) by location (i.e. cervical vs. lumbar). Traditionally, the posterior cervical approach has been associated with the highest SSI rate and the anterior cervical approach with the lowest infection rate, but they did not look at this specifically. The overall rate is also difficult to interpret as it is not clear that the sample included is representative of the overall practice of spine surgery. A multivariate approach would have been useful to determine the independent risk factors for infection, as covariance plays a role with many of the variables (i.e. surgical duration and blood loss are tightly related). Overall, this paper provides some useful information for benchmarking SSI rates and again demonstrates the benefit of intrawound vancomycin powder in reducing SSI. Hopefully similar studies in the future including more patients treated with vancomycin powder will allow for new benchmarks for the different approaches and indications.

Please read Dr. Zhou's article on this topic in the February 1 issue. Does this change how you view SSI rates and risk factors? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS
Associate Web Editor

Friday, January 10, 2020

The spine literature makes it clear that medical and psychosocial comorbidities result in worse patient reported outcomes following spine surgery. However, most papers have looked at absolute scores on patient reported outcome measures (PROMs) rather than change scores or recovery ratios. Even fewer have analyzed the treatment effect of surgery—the difference in change scores between patients treated with surgery and non-operative care. Most of this literature has evaluated lumbar surgery patients, and far fewer papers have looked at the relationship between psychosocial factors and outcomes in cervical spine surgery. In order to help fill this void, Dr. Divi and colleagues from Thomas Jefferson University stratified 264 ACDF patients into a depressed and non-depressed group based on SF-36 mental component scale (MCS) scores at baseline. They used two thresholds for depression (MCS < 45.6 and MCS < 35) based on prior literature, and the results were effectively the same for the two depression thresholds. Not surprisingly, they found that the depressed group had worse baseline Neck Disability Index (NDI), VAS Neck, and VAS Arm pain scores and worse post-operative scores at an average follow-up of 19 months. However, the depressed group improved significantly more on NDI and VAS neck and arm pain scores than the non-depressed group. The recovery ratios were similar for the two groups. Multivariate analyses controlling for some baseline differences demonstrated that depression was not an independent predictor of change score for any PROM.

This paper supports prior literature demonstrating that depressed patients have worse baseline and post-operative PROMs compared to non-depressed patients. There is less literature looking at change scores following cervical spine surgery than lumbar surgery, though most literature on lumbar surgery has shown less improvement in the depressed cohort. In the SPORT lumbar disk herniation study, patients with self-reported depression and MCS < 35 had worse baseline and post-operative ODI scores as well as less improvement on the ODI than non-depressed patients and those with MCS > 35.1 However, there was no difference in the treatment effect of surgery as depressed patients also improved less with non-operative treatment compared to non-depressed patients. The current paper did not include a non-operative group, so the treatment effect of surgery could not be determined. Additionally, MCS scores reflect mental health in general and are not specific for depression, so it is possible that the results would have been different if a depression-specific questionnaire had been used to define the groups. The authors also combined radiculopathy and myelopathy patients, who have different post-operative outcomes. However, the depressed and non-depressed cohorts had a similar proportion of myelopathy patients, so this is unlikely to have affected outcomes. The authors did not control for baseline PROM scores, which were worse in the depressed group, so it is possible that the greater degree of improvement seen in the depressed group is due to less of a ceiling effect in the non-depressed group. The recovery ratio helps to take into account the baseline PROM score differences, and there were no differences in those analyses. This paper adds to the growing literature suggesting that while patients with medical and psychosocial comorbidities have worse absolute outcomes, they still benefit significantly from surgery. Many of these characteristics are not easily modifiable, so surgery should still be offered to patients with mental health diagnoses. Both patients and surgeons should be aware that their absolute outcomes may not be as good as patients without such comorbidities.

Please read Dr. Divi's article on this topic in the February 1 issue. Does this change how you consider depression in the surgical decision-making process? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS
Associate Web Editor



1.            Pearson A, Lurie J, Tosteson T, et al. Who should have surgery for an intervertebral disc herniation? Comparative effectiveness evidence from the spine patient outcomes research trial. Spine 2012;37:140-9.



Friday, January 3, 2020

A small proportion of high resource utilization patients tend to drive the majority of healthcare spending, and prior literature has explored these patients' characteristics. Medical comorbidities, psychological problems, low socioeconomic status, low educational achievement, worker's compensation, disability, litigation, smoking, and opioid use are associated with worse outcomes and higher costs across multiple medical conditions. In order to better understand resource utilization and costs following lumbar fusion, Jason Lerner and colleagues analyzed the IBM MarketScan database of commercial insurance claims for one and two level lumbar fusions from 2007-2016, which included over 18,000 patients eligible for the study. They used a machine-learning program to identify three clusters of patients based on pre-operative healthcare utilization that were associated with post-operative resource utilization. The largest cluster included 14,000 patients who were normal utilizers pre- and post-operatively. The second cluster included over 4,000 high utilizer patients, who used the most opioids pre-operatively and had the highest medical comorbidity burden. The third and smallest cluster (500 patients) were also high utilizers, though this group's utilization was primarily for mental health resources. Very limited demographic data were included, though the high utilizer groups both had a significantly higher proportion of females compared to the normal utilization group (67% cluster 3, 62% cluster 2, and 54% cluster 1). After controlling for potential confounders, the normal utilization groups overall two-year post-operative costs were significantly less than the high utilization groups ($34,000 cluster 1, $52,000 cluster 2, and $48,000 cluster 3). These cost differences were present both for spine-related and other medical costs.

This paper clearly demonstrated that patients with high pre-operative resource utilization continue to have high post-operative utilization. While this comes as no surprise, it does provide some data that could help identify these patients pre-operatively and potentially develop approaches to help address their high utilization. Like most administrative database studies, this study has significant limitations. One major limitation is that other patient characteristics known to affect utilization (i.e. psychosocial problems, socioeconomic status, worker's compensation, etc.) were not included in the database and may have been strong predictors of post-operative utilization. Additionally, those characteristics may be much easier to identify than prior resource utilization. While the authors suggested that prior utilization would be easy to identify, many patients see spine surgeons at referral centers who do not have access to their outside records and may be completely unaware of their resource utilization. Their insurance provider does have these data and could make them available to their surgeon, though this would require more complex communication than is currently the norm. Additionally, the database included no information on patient reported outcomes. While prior literature would suggest that the high utilizers likely had worse outcomes overall, they also likely improved more with surgery than they would have with non-operative treatment. The spine surgery community must be careful with data like these as payors and administrators can use them to identify high resource utilizers and create policy that limits their access to care. These high utilizers tend to be our most vulnerable patients who are in the most need of care, and we must be careful not to create any more barriers for them. Instead, we must work to create ways to mitigate the negative effects of their characteristics on outcomes and resource utilization. That is no small task and likely requires addressing major societal issues.

Please read this article in the January 15 issue. Does this change how you consider surgical decision making or post-operative care in the high resource utilization population? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS

Associate Web Editor

Friday, December 27, 2019

Depression is a well-known risk factor for less improvement on patient-reported outcomes (PROs) following cervical and lumbar surgery. Given that depressed patients tend to respond less favorably to health-related questionnaires, it seems likely that they may report less satisfaction with their physicians and hospital. The most commonly used patient satisfaction survey is the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS), which is administered by the Center for Medicare and Medicaid Services (CMS) and is used to adjust hospital reimbursement. Many health systems also use the data to evaluate physicians and adjust compensation. Now that HCAHPS data can affect reimbursement, it is important for physicians and hospitals to better understand factors that affect patient responses to these surveys. In order to evaluate the effect of pre-operative depression on HCAHPS scores following cervical spine surgery, Jay Levin and colleagues from the Cleveland Clinic evaluated 145 cervical spine surgery patients from 2013-2015 who had responded to the HCAHPS survey and also had baseline PROs (i.e. PHQ-9, EQ-5D, neck pain VAS). Forty-one patients were classified as having at least moderate depression (PHQ-9 score over 9), and the depressed patients were about 6 years younger and had worse baseline EQ-5D and neck pain VAS scores. There was a trend towards a higher proportion of female patients in the depressed group (59% in depressed group vs. 41% non-depressed, p = 0.06). In univariate analysis, they found that depressed patients were significantly less likely to report that their doctors always treated them with courtesy and respect (88% depressed vs. 97% non-depressed, p = 0.03) and less likely to report that doctors always listened carefully to them (78% depressed vs. 91% non-depressed, p = 0.03). There was also a trend towards depressed patients being less likely to rate the hospital as a 9 or 10 overall (68% vs. 81%, p = 0.1). Multivariate analysis controlling for demographic characteristics and baseline PROs demonstrated that depression was an independent predictor of reporting that their doctor did not always treat them with courtesy and respect (OR = 0.14, p = 0.04).  

This paper has done a nice job demonstrating that depressed patients are less likely to be satisfied with their doctor's communication style following cervical spine surgery. This result comes as no surprise given the strong negative association between depression and PROs. I would have expected an even more pronounced effect on HCAHPS scores. There were many trends towards less satisfaction with nurses and hospitals that did not reach significance, and this likely reflects that the study was underpowered for these analyses. Only a minority of patients receive and then return the HCAHPS survey, so the authors only had 145 patients in their study, of whom 41 were depressed. This limits the power of all the analyses, especially the multivariate analysis. This paper begs the question of how HCAHPS data should be used, and also how hospitals and providers might be incentivized to game the system. While CMS can control for macro level data (i.e. proportion of patients on Medicaid, socioeconomic data, etc.), they do not collect meaningful data on the individual patient level (i.e. depression, educational attainment, work status, etc.). Controlling for the macro level data in large populations may be reasonable for hospital level results, however, this is probably not sufficient at the individual provider level. If these data are used for reimbursement and compensation, hospital systems and physicians may be incentivized to care for patients without risk factors for dissatisfaction. This could create even greater barriers to healthcare access for our most vulnerable patients, who tend to have medical and psychosocial comorbidities associated with worse outcomes and lower satisfaction. Healthcare administrators should be very cautious about how they use patient satisfaction data in determining reimbursement and compensation.

Please read Mr. Levin's article on this topic in the January 1 issue. Does this change your view of how patient satisfaction data like HCAHPS should be used? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS
Associate Web Editor

Friday, December 20, 2019

Lumbar diskectomy is a highly effective treatment for radiculopathy patients who meet the indications for surgery. Many surgeons include post-operative physical therapy (PT) as part of their protocol, though there is little high quality data to support this. Anecdotally, most lumbar diskectomy patients seem to do quite well without any PT, though a minority have a difficult time returning to work and their regular activities. In order to better assess the effectiveness of post-diskectomy PT, Dr. Paulsen and colleagues from Denmark performed a Level 1 RCT in which lumbar diskectomy patients were randomized to outpatient PT starting 4-6 weeks post-operatively or to no PT. Both groups received instructions about home exercises prior to discharge from the hospital and were instructed to advance their activities as tolerated without restrictions. Seventy-three patients were randomized to each group, and there were no significant baseline differences between the groups. They were followed out to 2 years, and both groups improved significantly on all patient reported outcomes (PROs). There were no significant differences between the two groups on the ODI, EQ-5D, or VAS leg pain or VAS back pain scores out to 2 years. These outcomes were measured at 1, 3, 6, 12, and 24 months after surgery. Ten patients in the PT group and 9 patients in the no PT group underwent reoperation within one year (13% overall), with the vast majority of these being revision diskectomies for recurrent disk herniation. Other complication rates were very low and similar between the two groups. The authors concluded that post-diskectomy PT did not improve PROs.

The authors have performed a simple but high quality study that demonstrated convincingly that post-diskectomy outpatient PT starting 4-6 weeks after surgery did not improve PROs. There was a small amount of crossover between the groups, but the as-treated and per protocol analyses showed the same result. One of the limitations of this study was that PT did not start until 4-6 weeks after surgery, and many patients had likely already returned to completely normal function at this point. One month after surgery was the earliest time at which PROs were recorded, and much of the improvement had occurred prior to this. The authors also did not evaluate when patients returned to work and activities, and these outcomes may have been affected by PT. The revision surgery rate was relatively high in the first year (13% compared to 6% in SPORT), though it was nearly identical for the two groups. The authors' conclusion that outpatient PT starting 4-6 weeks after diskectomy did not affect PROs measured from 1-24 months post-operatively appears completely valid. However, it is possible that starting PT earlier may have hastened return to work and activities. Given that the surgeons did not use activity restrictions, it is unclear why PT did not start earlier. Additionally, PT was provided to all patients randomized to it, and clearly not all patients benefit from it. However, there may be a subgroup of patients at risk for a slower recovery (i.e. those with fear avoidance behavior, worker's compensation claims, heavy work demands, etc.) who would benefit from early PT. This paper suggests that routinely referring all diskectomy patients to PT one month after surgery is probably not helpful. There are likely patients who do benefit from early PT, though who these patients are and how early to start PT remain to be determined.

Please read Dr. Paulsen's article on this topic in the January 1 issue. Does this change how you view the role of post-diskectomy PT? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD, MS
Associate Web Editor