Abstract
Background Individual items within the Patient Health Questionnaire-9 (PHQ-9) have not been assessed as predictors of postoperative outcomes. Our objective is to study the relationship between responses to individual PHQ-9 items and achievement of a minimum clinically important difference (MCID) following anterior cervical discectomy and fusion (ACDF).
Methods A prospective surgical database was reviewed for primary, single-level ACDF procedures performed for degenerative spinal pathology. Patient demographics, preoperative spinal pathology, and perioperative characteristics were recorded. Patient-reported outcome measures (PROMs) including PHQ-9, visual analog scale (VAS) neck and arm, Neck Disability Index, 12-item Short Form physical component score (SF-12 PCS), and Patient-Reported Outcomes Measurement Information System Physical Function were administered at preoperative and 6-week, 12-week, 6-month, 1-year, and 2-year postoperative timepoints. MCID achievement was determined by comparing postoperative PROM improvement from baseline to previously established values. Logistic regression assessed responses to each individual question of the preoperative PHQ-9 as predictors of MCID achievement in each other PROMs.
Results Sixty-six ACDF patients were included with a mean age of 47.2 years. Herniated nucleus pulposus was the most common preoperative spinal diagnosis (95.6%). The mean operative duration was 50.3 minutes, the mean estimated blood loss was 27.5 mL, and most patients were discharged on postoperative day 0 (81.8%). A majority of patients achieved MCID for all measures except SF-12 PCS. PHQ-9 question 3 significantly predicted MCID achievement for VAS neck (P = 0.045), VAS arm (P = 0.049), and SF-12 PCS (P = 0.037). No other PHQ-9 items or overall PHQ-9 scores significantly predicted MCID achievement.
Conclusion Question 3 of the PHQ-9 regarding “trouble falling asleep, staying asleep, or sleeping too much” significantly predicted clinically meaningful improvement in neck pain, arm pain, and physical function following ACDF, although overall PHQ-9 scores did not. Providers should inform patients experiencing significant sleep-related difficulties that they may be especially likely to benefit from ACDF surgery.
Clinical Relevance Evaluation of sleep from the PHQ-9 predicts clinically relevant improvement in neck pain, arm pain, and physical function in patients undergoing ACDF.
Level of Evidence 3.
- anterior cervical discectomy and fusion (ACDF)
- patient health questionnaire-9 (PHQ-9)
- patient-reported outcome measures (PROMs)
INTRODUCTION
Anterior cervical discectomy and fusion (ACDF) is one of the most common surgical treatments for the degeneration of the cervical spine when conservative management is insufficient.1 Many who suffer from cervical pathology experience adversities in daily functioning, decreased productivity, and limitation of roles in their personal lives.2 Because patients choose to undergo ACDF largely in the hope of improving overall health, clinicians track pre- and postoperative physical and mental symptom progression with great interest. The role that preoperative mental health, in particular, plays on postoperative outcomes has been previously explored in spinal fusion patients, and the results have been contrasting in nature. Several investigations report that those with greater preoperative depression, as measured by the Patient Health Questionnaire-9 (PHQ-9), experienced poorer functional outcomes and smaller improvements in postoperative quality of life.3,4 However, Jenkins et al have demonstrated that preoperative PHQ-9 scores had minimal impact on postoperative outcomes among ACDF patients.5
The PHQ-9 is a patient-reported outcome measure (PROM) designed to assess a patient’s depressive symptom severity using 9 questions based on the DSM-IV criteria for depressive disorders. These inquiries evaluate specific symptom criteria such as energy level, appetite, concentration, hopelessness, and sleep.6 Each question is scored based on the frequency of symptom occurrence from 0 (never) to 3 (nearly every day), and outcome analyses are typically based on the total combined score ranging from 0 to 27. However, independent analysis of the individual questions of the PHQ-9 may also provide valuable information pertaining to various aspects of daily life. With a sense of the specific individual categories that are most predictive of postoperative outcomes, physicians may promote more targeted screenings of prospective surgical candidates and better account for those preoperative indicators.
For our present study, in order to determine whether aspects of the PHQ-9 truly have an impact on outcomes, we assessed postoperative conditions in terms of the minimum clinically important difference (MCID). MCID has gained traction in spine surgery because of its ability to quantify outcomes in terms of the level of improvement at which patients truly notice meaningful changes in their symptoms.7 To our knowledge, studies have yet to consider the separate questions of the PHQ-9 in relation to postoperative MCID achievement in cervical spine populations. With greater insight on this topic, clinicians may identify ways to optimize postsurgical outcomes as well as administer more efficient and applicable survey questionnaires. Therefore, we aim to study the relationship between responses to individual PHQ-9 items and the achievement of MCID following ACDF.
METHODS
Patient Population
Prospectively collected data were retrospectively reviewed for patients who underwent primary, single-level ACDF procedures from April 2016 to November 2018. Exclusion criteria were patients lacking preoperative PHQ-9 data or whose procedures were performed for trauma, malignancy, or infection. All procedures were performed at a single academic institution by the same attending spine surgeon. All aspects of the current study were approved by the Institutional Review Board, and all included patients provided written informed consent.
Data Collection
Collected data included patient age, gender, body mass index, smoking status, diabetes mellitus status, American Society of Anesthesiologists physical classification, Charlson Comorbidity Index, ethnicity, insurance/payment received, preoperative spinal pathology, operative duration (from skin incision to skin closure, in minutes), estimated blood loss (in minutes), postoperative length of stay (in hours), and postoperative day (POD) of discharge. PROMs were administered at preoperative and 6-week, 12-week, 6-month, 1-year, and 2-year postoperative timepoints. PROMs consisted of PHQ-9, visual analog scale (VAS) neck, VAS arm, Neck Disability Index (NDI), 12-item Short-Form physical component score (SF-12 PCS), and Patient-Reported Outcomes Measurement Information System Physical Function (PROMIS-PF). While the overall survey metrics were collected for the aforementioned PROMs, responses to each separate question of preoperative PHQ-9 were also collected (the details of each individual question are listed in Table 1). We then used MCID achievement rates as a way of assessing the overall success of the procedure, as it resembles the proportion of patients achieving clinically meaningful improvement across PROMs based on established literature values for delta PROMs (change from pre- to postoperative scores). Next, each individual question of the 10 PHQ-9 items was assessed as a predictor of MCID attainment for other PROMs (VAS neck, VAS arm, NDI, SF-12 PCS, and PROMIS-PF) using logistic regression analysis with OR and 95% CI. Other published manuscripts have highlighted the importance of studying the predictive ability of PROMs on MCID achievement: Leyton-Mange et al sought to demonstrate preoperative PROM threshold scores that were predictive of MCID achievement for adult spinal deformity patients,8 while Lynch et al utilized linear regression to determine the influence of PHQ-9 on change in PROMs (ie, delta PROMs), which directly influenced MCID achievement rates.9 In the present manuscript, we are specifically honing in on individual questions of the PHQ-9 to evaluate whether these may hold a significant predictive value of MCID achievement.
The Patient Health Questionnaire-9.
Statistical Analysis
Achievement of MCID was determined by comparing postoperative change from preoperative baseline in each PROM to the following previously established MCID values: VAS neck (2.6),10 VAS arm (4.1),10 NDI (8.5),10 SF-12 PCS (8.1),10 and PROMIS-PF (4.5).11 Descriptive statistics were performed to report means and SD for continuous variables and percentages for categorical variables. Individual responses to each question of the PHQ-9 were assessed as predictors of MCID achievement for each of the other PROMs using logistic regression with ORs and 95% CIs.
RESULTS
A total of 66 primary, single-level ACDF patients with complete preoperative PHQ-9 data were identified. The cohort’s mean age was 47.2 years, 42.4% were female, and 47.0% were obese ( body mass index ≥30 kg/m2) (Table 2)(). A majority of the patients were of white ethnicity (71.2%) and made payments through private insurance (71.2%). Herniated nucleus pulposus was present in nearly all patients (95.5%). The mean operative duration was 50.3 minutes, the mean estimated blood loss was 27.5 mL, and the mean postoperative length of stay was 10.1 hours (Table 3)(). Most patients were discharged on the day of surgery (81.8%).
Patient demographics.
Perioperative characteristics.
Mean preoperative PROM scores were as follows: 6.1 ± 2.4 (VAS neck), 6.0 ± 2.8 (VAS arm), 37.2 ± 16.0 (NDI), 36.1 ± 8.1 (SF-12 PCS), and 39.6 ± 7.2 (PROMIS-PF). Mean postoperative scores ranged from 2.5 to 5.0 (VAS neck), 2.4 to 4.0 (VAS arm), 20.2 to 30.1 (NDI), 35.3 to 43.5 (SF-12 PCS), 40.7 to 47.9 (PROMIS-PF) (Table 4) (). A majority of patients achieved MCID in all PROM scores except SF-12 PCS (47.4%).
Postoperative outcomes.
Question 3 of the PHQ-9 (trouble falling asleep, staying asleep, or sleeping too much) was a significant predictor of MCID achievement for VAS neck (OR: 2.0, 95% CI [1.0, 3.9], P = 0.045), VAS arm (OR: 1.7, 95% CI [1.0, 2.9], P = 0.049), and SF-12 PCS (OR: 1.7, 95% CI [1.0, 2.9], P = 0.037) (Tables 5 and 6). No other PHQ-9 question or overall PHQ-9 score significantly predicted MCID achievement.
PHQ-9 items as predictors of pain and disability MCID achievement.
PHQ-9 items as predictors of physical function MCID achievement.
DISCUSSION
The relationship of preoperative mental health with the postoperative outcomes of spine surgery has been a topic of great interest in recent years. The PHQ-9 has been widely used in a variety of clinical and research settings and has been validated for quantification of depressive symptoms in patients undergoing cervical spine surgery, demonstrating strong correlations with SF-12 mental component score (MCS) and Veterans RAND-12 (VR-12) MCS among ACDF and cervical disc replacement patients.12 As depression rates rise and preoperative management of modifiable factors shows promise, comprehensive prediction tools specifically evaluating depressive symptoms, such as PHQ-9, have become a critical component of preoperative screening.12 Furthermore, the value-based care movement is gaining traction within the medical community as it is centralized on the needs of patients.13 This model allows for shared-decision making grounded on each individual goal of care, and in turn, optimized resource utilization and asset allocation with potentially decreased costs, as more suitable interventions may then be chosen that more directly align with the patient’s goals of treatment.13,14 Mental health impairments are highly prevalent in the spine surgical population undergoing cervical and lumbar procedures, with an estimated depression prevalence of >50% among surgical candidates.3 Still, few authors have addressed the utility of measuring the predictive value of preoperative mental health scores (ie, PHQ-9) on ACDF outcomes, while many have evaluated this relationship with lumbar fusion outcomes (demonstrating a pattern of poorer outcomes among those with lower baseline mental functioning).3 This highlights the importance of continued projects such as the present study aimed at illustrating the influence of mental health on ACDF outcomes. While greater than 30% of spine surgeons do not use PROMs routinely, this is primarily due to lack of time, funding, and/or available staffing for the collection of data.15 Such limitations are noteworthy, and continued studies may highlight the importance to stakeholders and hospitals to allocate more resources and funding to the implementation of PROMs, including PHQ-9. Ultimately, with an emphasis on preoperative mental health management and discussion of evidence-based patient expectations, routine use of PHQ-9 serves as an important avenue for optimization of postoperative PROMs and patient satisfaction.
A number of studies have assessed PHQ-9 scores as a predictor of outcomes in patients undergoing ACDF specifically, although published findings are conflicting. Alvin et al demonstrated that preoperative PHQ-9 scores, along with antidepressant use, PDQ scores, and EQ-5D scores, were significantly associated with MCID achievement in the EQ-5D quality of life score by 1 year following ACDF.3 However, it should be noted that when significant factors were assessed together in a multiple regression model, only baseline EQ-5D scores remained a significant predictor of EQ-5D MCID achievement. These results essentially suggest that MCID achievement may be most closely dependent upon the available “room for improvement,” such that patients with worse preoperative health are more likely to see greater postoperative improvements.
Jenkins et al also studied the relationship between preoperative PHQ-9 scores and postoperative MCID achievement in a variety of physical health PROMs.5 Their results indicated that while preoperative pain, disability, and physical function were significantly associated with preoperative PHQ-9 scores, when postoperative physical health outcomes were assessed based on preoperative PHQ-9, scores were largely similar for pain and only differed at short-term or intermediate timepoints for disability and physical function. Jenkins et al’s methodology differed in their choice to categorize patients based on high (≥5) or low (<5) preoperative PHQ-9 scores, rather than retaining the measure’s native continuous status. Their analysis demonstrated that a significantly greater proportion of patients with high preoperative PHQ-9 scores were able to achieve an MCID in NDI compared to those with lower preoperative depression scores. However, for VAS neck, VAS arm, and SF-12 PCS, Jenkins et al observed no significant differences in MCID achievement on the basis of preoperative PHQ-9 score. Interestingly, the 3 PROMs for which we observed a significant relationship with question 3 of the PHQ-9 are also the same 3 for which Jenkins et al reported no difference based on the overall PHQ-9 score.
Largely in line with the results of Jenkins et al, our analysis revealed no significant associations between preoperative PHQ-9 score and achievement of MCID in pain or physical function, although our results did not replicate their finding regarding the achievement of MCID in NDI. Based on these results and our review of the literature, we can suggest that although PHQ-9 scores may be strongly correlated with mean scores in other health-related quality of life measures, the overall score of the PHQ-9 may not be the best predictor of the postoperative change in these outcomes that are quantified by the MCID.
Although overall PHQ-9 scores did not significantly predict MCID achievement in our cohort, we determined that one specific response item within the PHQ-9 did consistently predict MCID achievement for VAS neck, VAS arm, and SF-12 PCS. Specifically, the response item asking patients to quantify how often they had been bothered by “trouble falling asleep, staying asleep, or sleeping too much” within the last 2 weeks was a statistically significant predictor of meaningful postoperative improvement in these measures of pain and physical function.
Problems with sleep may be relatively common among patients suffering from spinal pathology.16–18 While many cervical surgery candidates report a chief complaint of neck pain (due to various etiologies such as disc herniation, spinal stenosis, or fibromyalgia), recent studies have suggested that such pain is often coupled with sleep disorders, at a rate of 50% to 80%.19 As the pain has important psychological and functional implications and is intricately associated with difficulties in achieving and maintaining sleep,19 our results indicate that ACDF candidates with poor baseline sleep may clinically benefit in postoperative pain and physical functioning following surgical correction. Spinal surgeons should inform patients that pain and sleep problems can coexist and encourage individuals with poorer sleep quality who undergoing ACDF will likely reap clinically significant improvements in their arm and neck pain and physical health. Furthermore, disordered sleep may be strongly associated with other health-related symptoms in patients undergoing spine surgery. In an analysis of a national sample of cervical spine surgery patients, Pennings et al demonstrated strong correlations between NDI and PROMIS sleep disturbance, both preoperatively and at postoperative follow-up.20 Other studies have demonstrated similar associations between sleep disturbance and disability and other health-related outcomes.17,21
Interestingly, a number of animal studies have demonstrated that sleep deprivation is not only associated with but may actually induce increased pain sensitivity in rats.22,23 Wang et al studied the impact of short-term sleep disturbance on the postsurgical recovery of rats in a controlled laboratory setting and demonstrated that repeated sleep interruption resulted in increased postoperative pain response.24 Furthermore, the authors were able to demonstrate that persistent disruption of the rats’ REM sleep phase resulted in significantly decreased expression of mu and kappa opioid receptors in the ipsilateral dorsal root ganglion following surgical intervention. One important point to note is that animal studies have indicated that the effects of sleep deprivation on pain sensitivity are likely reversible. Onen et al demonstrated that decreased pain thresholds among rats were completely reversed following recovery of normal sleep.23
Human studies have also demonstrated direct links between sleep deprivation and increased pain sensitivity.25–27 Azevedo et al demonstrated that following 2 nights of total sleep deprivation, VAS pain responses to a laser stimulus were significantly increased and continued to increase with successive nights of sleep deprivation.26 Of note, this study also demonstrated that VAS ratings decreased after just 1 night of restored sleep. In fact, a recent systematic review by Stroemel-Scheder et al indicated that restoration of normal sleep may effectively reduce the pain sensitivity associated with sleep deprivation.28
Literature regarding problems in patients undergoing cervical spine surgery is relatively limited. However, the results of available studies indicate that sleep-related issues are likely to improve significantly following spinal surgery. Ogden et al reported significant improvements in all subscales of the Pittsburgh Sleep Quality Index, including sleep time, sleep disturbances, daytime functionality loss, and sleep quality following ACDF.19 Löfgren et al demonstrated temporary improvements in sleep, measured via the Sickness Impact Profile; however, these authors reported that while many other postoperative benefits were lasting, these improvements in sleep/rest were only temporary in their cohort.29 Among a cohort of lumbar stenosis patients, significant improvements in sleep disturbance were likewise reported following spinal fusion and/or decompression.30
Given the documented associations of sleep disturbance with increased perceptions of pain, it may initially seem counterintuitive that patients endorsing problems with sleep via the PHQ-9 would actually be more likely to achieve MCID following ACDF. However, these results make sense in the context that sleep disturbance may be a modifiable factor. In fact, results of previous studies in spine surgery indicate that patients with greater severity of preoperative symptoms may essentially have more “room for improvement” and therefore, may actually be more likely to achieve a meaningful degree of change following surgery. For example, Parrish et al demonstrated that patients with poorer preoperative PROMIS-PF scores demonstrated higher rates of MCID achievement in pain, disability, and physical function following ACDF, compared to those with more favorable preoperative scores.31 Therefore, patients who report significant issues with sleep prior to surgery may be well positioned to benefit from operative intervention, given that they have a substantial unmet need that may be amenable to operative intervention.
Limitations
While this is likely the first study of its kind to assess the predictive qualities of individual PHQ-9 response items in the context of cervical spine surgery, our methodology is not without limitations. First, the self-reported nature of our study measures may introduce an element of bias. However, we feel that this is a necessary consequence of quantifying outcomes with a patient-centered focus. Additionally, the single-surgeon, single-institution aspect of our study design may limit generalizability to broader populations. Furthermore, as seen in Tables 5 and 6, question 9 was unable to be analyzed as a predictor of MCID achievement across PROMs due to a limited sample size, reducing power for this portion of our study. Finally, while the PHQ-9 itself is well-validated for use in cervical spine surgery,12 the use of question 3 specifically to quantify sleep-related issues has not been directly assessed. Future studies should include a direct comparison of responses to question 3 with one or more measures specifically designed to assess disordered sleep.
CONCLUSION
Overall, PHQ-9 scores did not significantly predict the achievement of MCID in pain, disability, or physical function. However, a statistically significant relationship was identified between responses to question 3 of the PHQ-9, which assesses the frequency of sleep-related problems, and clinically meaningful improvements in arm pain, neck pain, and physical function. Patients with cervical disease and concomitant preoperative sleep difficulties should be encouraged that ACDF may result in clinically significant pain relief and physical health improvements postoperatively. Additional high-quality studies examining the relationship of preoperative sleep with postoperative outcomes of cervical spine surgery are warranted.
Footnotes
Funding The authors received no financial support for the research, authorship, and/or publication of this article.
Disclosures Kern Singh discloses that he has received grants or contracts from the Cervical Spine Research Society; royalties or licenses from RTI Surgical, Zimmer Biomet, Stryker, Lippincott Williams & Wilkins, Theime, Jaypee Publishing, and Slack Publishing; consulting fees from K2M and Zimmer Biomet; patents planned, issued, or pending with TDi LLC; and leadership or fiduciary role on Vitals 5 LLC, TDi LLC, Minimally Invasive Spine Study Group, Contemporary Spine Surgery, Orthopedics Today, and Vertebral Columns. The remaining authors have no disclosures.
IRB Approval ORA #14051301
- This manuscript is generously published free of charge by ISASS, the International Society for the Advancement of Spine Surgery. Copyright © 2022 ISASS. To see more or order reprints or permissions, see http://ijssurgery.com.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.