Patient-related Factors and Perioperative Outcomes Are Associated with Self-Reported Hospital Rating after Spine Surgery

Abstract

Background

Since 2013, the Centers for Medicare & Medicaid Services has tied a portion of hospitals’ annual reimbursement to patients’ responses to the Hospital Consumer Assessment and Healthcare Providers and Systems (HCAHPS) survey, which is given to a random sample of inpatients after discharge. The most general question in the HCAHPS survey asks patients to rate their overall hospital experience on a scale of 0 to 10, with a score of 9 or 10 considered high, or “top-box.” Previous work has suggested that HCAHPS responses, which are meant to be an objective measure of the quality of care delivered, may vary based on numerous patient factors. However, few studies to date have identified factors associated with HCAHPS scores among patients undergoing spine surgery, and those that have are largely restricted to surgery of the lumbar spine. Consequently, patient and perioperative factors associated with HCAHPS scores among patients receiving surgery across the spine have not been well elucidated.

Questions/purposes

Among patients undergoing spine surgery, we asked if a “top-box” rating on the overall hospital experience question on the HCAHPS survey was associated with (1) patient-related factors present before admission; (2) surgical variables related to the procedure; and/or (3) 30-day perioperative outcomes.

Methods

Among 5517 patients undergoing spine surgery at a single academic institution from 2013 to 2017 and who were sent an HCAHPS survey, 27% (1480) returned the survey and answered the question related to overall hospital experience. A retrospective, comparative analysis was performed comparing patients who rated their overall hospital experience as “top-box” with those who did not. Patient demographics, comorbidities, surgical variables, and perioperative outcomes were compared between the groups. A multivariate logistic regression analysis was performed to determine patient demographics, comorbidities, and surgical variables associated with a top-box hospital rating. Additional multivariate logistic regression analyses controlling for these variables were performed to determine the association of any adverse event, major adverse events (such as myocardial infarction, pulmonary embolism), and minor adverse events (such as urinary tract infection, pneumonia); reoperation; readmission; and prolonged hospitalization with a top-box hospital rating.

Results

After controlling for potential confounding variables (including patient demographics), comorbidities that differed in incidence between patients who rated the hospital top-box and those who did not, and variables related to surgery, the patient factors associated with a top-box hospital rating were older age (compared with age ≤ 40 years; odds ratio 2.2, [95% confidence interval 1.4 to 3.4]; p = 0.001 for 41 to 60 years; OR 2.5 [95% CI 1.6 to 3.9]; p < 0.001 for 61 to 80 years; OR 2.1 [95% CI 1.1 to 4.1]; p = 0.036 for > 80 years), and being a man (OR 1.3 [95% CI 1.0 to 1.7]; p = 0.028). Further, a non-top-box hospital rating was associated with American Society of Anesthesiologists Class II (OR 0.5 [95% CI 0.3 to 0.9]; p = 0.024), Class III (OR 0.5 [95% CI 0.3 to 0.9]; p = 0.020), or Class IV (OR 0.2 [95% CI 0.1 to 0.5]; p = 0.003). The only surgical factor positively associated with a top-box hospital rating was cervical surgery (compared with lumbar surgery; OR 1.4 [95% CI 1.1 to 1.9]; p = 0.016), while nonelective surgery (OR 0.5 [95% CI 0.3 to 0.8]; p = 0.004) was associated with a non-top-box hospital rating. Controlling for the same set of variables, a non-top-box rating was associated with the occurrence of any adverse event (OR 0.5 [95% CI 0.3 to 0.7]; p < 0.001), readmission (OR 0.5 [95% CI 0.3 to 0.9]; p = 0.023), and prolonged hospital stay (OR, 0.6 [95% CI 0.4 to 0.8]; p = 0.004).

Conclusions

Identifying patient factors present before surgery that are independently associated with HCAHPS scores underscores the survey’s limited utility in accurately measuring the quality of care delivered to patients undergoing spine surgery. HCAHPS responses in the spine surgery population should be interpreted with caution and should consider the factors identified here. Given differing findings in the literature regarding the effect of adverse events on HCAHPS scores, future work should aim to further characterize this relationship.

Level of Evidence

Level III, therapeutic study.

Introduction

The Centers for Medicare & Medicaid Services (CMS) reimburse hospitals in part based on how patients rate their hospital experiences, as assessed by a survey titled the Hospital Consumer Assessment of Healthcare Provider Systems (HCAHPS). As part of the Value-Based Purchasing program, which was introduced in 2013, a hospital’s scores dictate 2% of its annual reimbursement from CMS [48]. Hospitals are required to give the survey to a random sample of adult patients who have spent at least one inpatient night for non-psychiatric indications, regardless of payer status [4].

The HCAHPS survey contains 32 questions pertinent to a patient’s hospital stay, 22 of which ask for Likert-scale responses. Three questions guide patients through the survey and the remaining seven are related to demographics. The Likert-scale questions focus on different aspects of an inpatient’s experience, from communication with nurses and physicians to pain management to hospital cleanliness. The most general question on the survey asks patients to rate their hospital on a scale from 0 to 10, “where 0 is the worst hospital possible and 10 is the best hospital possible.” A rating of 9 or 10 is considered a high rating, or “top-box” [4].

HCAHPS is intended to be an unbiased, objective measure of hospital quality. However, it has been suggested that patients’ HCAHPS scores may vary according to patient factors existing before hospitalization [31, 46]. For example, a study of patients undergoing elective colorectal surgery found that those with lower overall health ratings were more likely to rate their overall hospital experience as worse [46]. Similarly, a study of varying types of inpatients found that patients with greater illness severity had lower HCAHPS scores across all five of the survey’s domains [31]. Several other studies have considered the relationship between perioperative outcomes and HCAHPS scores. Studies regarding this relationship, however, are divided, with some reporting a relationship between perioperative outcomes and HCAHPS scores [16, 19, 29, 35, 43, 44] and others finding no association [1, 5, 7, 17, 18, 21, 25, 37-39, 47].

Some studies of 200 to 500 patients undergoing lumbar spine surgery have evaluated the correlation of HCAHPS scores with patient factors and postsurgical outcomes [24-28]. Two other studies evaluated HCAHPS responses for patients undergoing mixed types of spine surgery, but each only considered a few potential indicators of HCAHPS scores [15, 22].

The existing evidence studying HCAHPS scores in relation to spine surgery has identified several patient and surgical factors related to HCAHPS responses. However, most of this research is focused on the lumbar spine, and the evidence on HCAHPS scores and spine surgery remains sparse. Consequently, patient and perioperative factors associated with HCAHPS scores among patients undergoing surgery across the spine have not been well elucidated. Seeking to address these gaps, the present study aimed to identify independent factors associated with a top-box overall hospital rating. Compared with previous studies of patients undergoing spine surgery, the current study considered a wider range of demographics, comorbidities, and perioperative outcomes in a large sample.

Among patients undergoing spine surgery, we asked if a “top-box” rating on the overall hospital experience question on the HCAHPS survey was associated with (1) patient-related factors present before admission; (2) surgical variables related to the procedure; and/or (3) 30-day perioperative outcomes.

Patients and Methods

Study Design and Study Population

We performed a retrospective, comparative study after obtaining approval from our institution’s institutional review board. Between 2013 and 2017, 5517 patients underwent spine surgery at our institution, all of whom were sent HCAHPS surveys. Surveys were sent to patients undergoing both emergent and non-emergent procedures, and included those undergoing surgery on the cervical, thoracic, or lumbar spine, or on multiple regions. Of those surveyed, 27% (1480 patients) returned their HCAHPS surveys and answered the question of interest, and were therefore included in the study. This response rate mirrors that seen nationally [4], and falls within the wide range of response rates reported in other studies, which range from 5% to 60% [2, 6, 15]. While in most survey studies, a non-response rate of 73% would be somewhat alarming, the results are worth considering in spite of patients’ low response rate as the survey nonetheless continues to be used nationally to determine hospital reimbursement.

The observation that 75% of patients rated the hospital top-box is consistent with reports from CMS that on average, 72% of patients rate major teaching hospitals as top-box, with an equal number doing so for hospitals in the Northeast [4]. Mean (± SD) time to survey response was 34 ± 27 days after discharge.

Patient data were collected as per the guidelines of the American College of Surgeons National Surgical Quality Improvement Program, including a range of demographics, comorbidities, and perioperative outcomes. All patients, regardless of their discharge status, were followed for 30 days postoperatively by specifically trained nurses.

Demographic variables assessed included age, gender, height and weight (used to calculate BMI), American Society of Anesthesiologists (ASA) classification, and functional status.

Comorbidities assessed were those provided in the dataset. They included diabetes mellitus (including insulin-dependent diabetes mellitus), history of smoking within 1 year preoperatively, pulmonary diseases (dyspnea at rest or with moderate exertion or history of severe chronic obstructive pulmonary disease), hypertension, current dialysis use, bleeding disorders, disseminated cancer, open wound, ascites, steroid use within 30 preoperative days, and preoperative blood transfusion.

Surgical Variables and Perioperative Outcomes

Surgery-related variables included the surgeon’s specialty (orthopaedics or neurosurgery) and operative site (cervical, thoracic, lumbar, or multi-region) (Table 1). The operative site was identified based on surgical Current Procedural Terminology codes, with a manual review of medical records as needed. We also determined if the procedure was nonelective. Data were analyzed with nonelective cases excluded, and the results were nearly identical (not shown). Length of hospital stay was assessed as days from admission to discharge. Prolonged hospital length of stay was defined as longer than one SD greater than the mean of the population.

Table 1.

Demographics, comorbidities, and surgical variables by overall hospital rating with multivariate analysis

We extracted 30-day postoperative adverse events from the dataset and aggregated into any, major, and minor adverse events. Major adverse events included deep or organ-space surgical site infection, unplanned reintubation, pulmonary embolism, deep vein thrombosis, acute renal failure, sepsis or septic shock, stroke, cardiac arrest, myocardial infarction, and death. Minor adverse events included superficial surgical site infection, wound dehiscence, pneumonia, urinary tract infection, Clostridium difficile infection, and renal insufficiency.

Data on reoperation and readmission within 30 days, as well as prolonged hospital stay, were also extracted from the dataset, but did not contribute to the number of major or minor adverse events. Any adverse event was counted when a patient had a major or minor adverse event, reoperation, readmission, or prolonged length of stay.

HCAHPS Responses

Patients were dichotomized based on their responses to HCAHPS Question 21 (overall hospital rating), which reads “Using any number from 0 to 10, where 0 is the worst hospital possible and 10 is the best hospital possible, what number would you use to rate this hospital during your stay?” Patients who rated their hospital as “top-box” (a score of 9 or 10) were grouped together, while patients who rated the hospital an 8 or lower were grouped together.

For validation of the above technique of using Question 21 as an overview response, each of the responses to the other 22 Likert-scale HCAHPS questions was also assessed and summed for comparison with the overall hospital rating question with the numerical scores for each question added together. For all questions but Question 21, possible scores ranged from 1 to 4 points. To validate the choice of Question 21 as an indicator of patient satisfaction, we assessed concordance between the total score and score of Question 21 using a two-tailed t-test and linear regression. Patients who rated the hospital top-box had higher overall HCAHPS scores by t-test (53 ± 5 versus 42 ± 9; p < 0.001). Similarly, a linear regression showed concordance between Question 21 and overall HCAHPS score (r² = 0.6, p < 0.001).

Statistical Analysis

We compared patient demographics, comorbidities, surgical variables, and perioperative outcomes between patients who rated the hospital as top-box and those who did not. Categorical demographic and surgical variables were compared between the groups using chi-squared or Fisher’s exact tests. Any, major, and minor postoperative adverse events; reoperation; readmission; and prolonged length of stay were also compared between the groups using a chi-squared test.

We performed univariate analyses to identify variables to include in multivariate analyses. Variations in continuous variables were compared between the groups using Levene’s test. Thereafter, continuous demographic variables (age and BMI) were compared between groups using a t-test, assuming either equal or unequal variation depending on the result of Levene’s test.

In total, 1480 HCAHPS surveys from patients undergoing spine surgery were returned. Of these, 1106 (75%) patients gave a top-box response (9 or 10 of a possible score of 10) to the overview HCAHPS question about hospital rating, while 374 (25%) did not.

In terms of demographics, top-box respondents were more likely to be older, men, and healthier based on ASA classification than low-box respondents (Table 1). With regard to comorbidities, top-box respondents were more likely to be nonsmokers and to not have disseminated cancer than low-box respondents.

To evaluate the effect of patient demographics, comorbidities, and surgical variables on a patient’s likelihood of rating his or her hospital stay as “top-box,” a multivariate logistic regression was performed. To limit collinearity, in the model, we only controlled for comorbidities that differed between groups on univariate analysis. Additionally, a separate series of five multivariate logistic regression analyses was performed, similarly controlling for patient and surgical factors, to evaluate the effect of any, major, and minor adverse events; reoperation; readmission; and prolonged length of stay on patients’ overall hospital ratings.

A power analysis assumes random sampling. Given that our sample was not randomly chosen (there are factors that predispose patients to return HCAHPS surveys), we determined that a power analysis would not be valid for this analysis. Statistical significance was established at p < 0.05 for all comparisons.

Results

Association Between Top-Box Rating and Patient Factors

After controlling for demographics, comorbidities, and surgical variables that differed by hospital rating, we found that several patient factors were independently associated with a top-box rating (Table 1). Patients were more likely to give a top-box rating if they were older (compared with age ≤ 40 years, for age 41 to 60 years: odds ratio, 2.2 [95% CI 1.4 to 3.4]; p = 0.001; for age 61 to 80 years: OR 2.5 [95% CI 1.6 to 3.9]; p < 0.001; for age > 80 years: OR 2.1; [95% CI 1.1 to 4.1]; p = 0.036), or men (OR 1.3 [95% CI 1.0 to 1.7]; p = 0.028). Conversely, patients were less likely to give a top-box rating if they had an ASA Class of II (OR 0.5 [95% CI 0.3 to 0.9]; p = 0.024), Class III (OR 0.5 [95% CI 0.3 to 0.9]; p = 0.020), or Class IV (OR 0.2 [95% CI 0.1 to 0.5]; p = 0.003), compared with an ASA Class of I.

Association Between Top-Box Rating and Surgical Variables

Controlling for demographics, comorbidities, and surgical variables as above, we found that undergoing a cervical procedure compared with a lumbar procedure was associated with a top-box hospital rating (OR 1.4 [95% CI 1.1 to 1.9]; p = 0.016), while patients undergoing nonelective spine surgery were less likely to rate their hospital as top-box (OR 0.5 [95% CI 0.3 to 0.8]; p = 0.004). Patients’ hospital ratings did not vary by surgeon’s specialty.

Despite finding an association between non-elective surgery and hospital rating, when excluding emergent cases, the other associations observed were unchanged, except that male sex was no longer associated with top-box ratings.

Association Between Top-Box Rating and Perioperative Outcomes

Controlling similarly for patient and surgical factors, we found that experiencing any adverse event within 30 days (OR 0.5 [95% CI 0.3 to 0.7]; p < 0.001), requiring readmission within 30 days (OR 0.5 [95% CI 0.3 to 0.9]; p = 0.023), and having a prolonged hospital stay (OR 0.5 [95% CI 0.4 to 0.8]; p = 0.004) was associated with a non-top-box hospital rating. Major adverse events, minor adverse events, and reoperation were not associated with patients’ hospital ratings (Table 2).

Table 2.

Odds ratios for overall hospital rating considering postoperative adverse events

Discussion

To incentivize high-quality health care delivery, hospital reimbursement has become tied in part to measures of the quality of care delivered. Since 2013, patients’ responses to the HCAHPS survey dictate nearly 2% of hospitals’ reimbursements from the Centers for Medicare and Medicaid Services [12, 36, 48]. As with any instrument, a survey measuring hospital quality should provide an unbiased, robust measure of the care delivered. Previous studies, however, have suggested that HCAHPS responses may vary based on patient demographics and comorbidities present before admission [25, 28, 31, 46]. While previous studies have examined predictors of HCAHPS scores among patients undergoing orthopaedic surgery [1-3, 7-9, 14, 30, 34, 40, 45, 49], and among those receiving surgery of the lumbar spine [24-28], few studies to date have evaluated patient factors and perioperative outcomes associated with patients’ responses on HCAHPS for those undergoing surgery at all spine levels. Consequently, it remains unclear how patient and perioperative factors impact patients’ HCAHPS scores when considering the entire spine surgery population. In so doing, we found that a top-box hospital rating on HCAHPS is associated with older age, male sex, and cervical surgery, while a non-top-box hospital rating is associated with worse overall health status by ASA classification, non-elective surgery, a prolonged hospital stay, and experiencing any adverse event or readmission within 30 days. Finding that patient factors present preoperatively influence hospital ratings after spine surgery highlights a crucial limitation of the HCAHPS survey in measuring the quality of spine care delivered. HCAHPS surveys for patients undergoing spine surgery should be interpreted in light of the identified factors associated with patient responses.

There are a number of limitations to the present study. Foremost, the overall response rate (27%) is low for a survey study. While our response rate is comparable to that seen across the United States for HCAHPS surveys, [4] previous studies have suggested that higher response rates are associated with more positive ratings of patient satisfaction [41]. Specific to HCAHPS, CMS reports that each increase in response percentile accounts for an approximately 0.2% increase in the incidence of top-box hospital ratings [4]. This suggests that lower response rates could artificially suppress the evaluation of patients’ satisfaction and should be considered in interpreting this study.

The literature around nonresponse bias—the systematic variation in survey responses by the demographics of (non-)respondents—for HCAHPS surveys is limited [10]. One study finds that among those receiving spine surgery, patients who were male, required urgent/emergent surgery, and who had longer hospitalizations were less likely to return HCAHPS surveys [15]. Analogously, a study of nonresponse bias in another patient satisfaction survey finds that patients who were younger, male, and had longer hospitalizations were less likely to respond [23]. Taken together, these findings suggest there is likely bias in which patients returned surveys in this study. Applying the findings just described to our study, hospital rating could be affected both positively and negatively by nonresponse bias, as patients with longer hospitalizations and those who were younger rated the hospital less favorably, while men offered more positive appraisals.

The study has several additional limitations. The HCAHPS survey has been recently critiqued in the literature for the influence of patient factors on its scores [8, 9, 49], which was again demonstrated here. Additionally, while the survey polls patients on a range of domains, it remains limited in the specificity with which patients can respond, restricting patients to four-point Likert scale responses for most questions, save for the one (Question 21) analyzed in this study, which offers a scale from 0 to 10. Although these limitations make HCAHPS a less-than-ideal instrument for measuring patient satisfaction, we feel that the survey merits study in the spine surgery population given its role in determining hospital reimbursement from CMS. In light of these shortcomings, rather than providing a definitive analysis of patient satisfaction per se, we feel this study more compellingly demonstrates the limitations of the HCAHPS survey for evaluating patient satisfaction after spine surgery, with results that can guide interpretation of the results of this widely used and influential survey.

Further, it could be argued that including patients undergoing surgery at any spine level could muddy the results, as patients undergoing surgery at different spinal levels could have different expectations about the results of treatment. Despite this concern, we included patients undergoing operations at all spinal levels, as doing so highlighted the impact of surgical site on patient-reported satisfaction.

Regarding the data collected, there are several variables that this study does not consider. While we would have liked to include information on indication for surgery, duration of symptoms, and additional demographic characteristics (such as employment, payer status), these data were not available in our dataset. Further, several comorbidities examined in the study could be considered contraindications to spine surgery. The data provided did not include a discussion on how these comorbidities were managed before surgery.

The timing of adverse events relative to the time surveys were returned was not analyzed, though it is possible that response time could influence our interpretation of patient satisfaction [15]. However, patients who have postoperative complications may have had factors that predisposed them to perioperative outcomes and thus were relevant before and after returning the survey. Additionally, given that the present study used data from a single high-volume urban tertiary-care teaching institution in the Northeast, the findings might not represent a national patient population. However, studying patients at a single institution controls for institutional variability because the care received is likely more uniform.

The current study was observational, and so we cannot infer cause-and-effect relationships here, only associations. However, we believe that the findings of the present study remain relevant in describing how patient and perioperative factors relate to HCAHPS scores in the spine surgery population.

Finally, a potential negative consequence of this study. By identifying patients who are less likely to rate the hospital favorably, and thereby decrease a hospital’s reimbursement through providing lower HCAHPS ratings, a scenario can be envisioned in which institutions could be incentivized to delay or withhold care from these individuals.

We found that older age and being a man were positively associated with top-box overall hospital ratings. These findings are consistent with those of some previous studies, which have suggested that older patients [20] and men [1, 11, 30, 32, 33] report higher satisfaction after inpatient hospitalization compared with younger patients and women. However, at least one study noted that male inpatients rated their hospital less favorably [20], with another finding no effect of age on HCAHPS scores [13]. These studies’ conclusions likely differ from our own due to differences in the patient populations and the interventions patients received. More definitively, considering all patients returning HCAHPS surveys, CMS finds that older patients and male surgical patients are more likely to rate their hospital favorably [4]. In agreement with CMS’ report, this study extends the existing literature by demonstrating that older age and male sex are associated with more favorable hospital ratings after spine surgery at all levels [4].

We also found that patients who are sicker by ASA classification were less likely to rate their hospital favorably. Specifically, ASA Classes II to IV were associated with a lower likelihood of a top-box hospital rating than was ASA Class I. There is also debate in the literature regarding the impact of patients’ health status on their satisfaction. Studies have reported a negative association [29-31, 46] (including in a lumbar spine cohort [28]) or no relationship between the two [11]. Again, more definitively, CMS reports that a one-point increase in overall health status graded on a scale from 1 to 5 (from “poor” to “excellent”) is associated with a 6% increase in patient-reported overall hospital rating [4]. Our study mirrors these results in the spine surgery population, showing an independent association between worse overall health status and lower hospital ratings. It is possible that patients who have worse overall health status have both more numerous and more complex medical needs while inpatient, and might feel that those needs are not as readily met by healthcare staff as they are for healthier patients with fewer concerns.

Patients receiving cervical (rather than lumbar) surgery and those having elective (rather than emergency) surgery were also more likely to rate their hospital experience as top-box. This differs from a previous report, which found no difference in satisfaction between patients receiving lumbar spine surgery and those undergoing cervical spine surgery [42]. Results might differ between the prior and present studies because of differences in case mix and outcome measures used across study populations, as well as differences in sample size.

Regarding the effect of treatment acuity, it has been suggested that patients who are admitted urgently [20] or emergently [46] are less likely to rate the hospital favorably. A similar phenomenon was observed in this study. While not quantified in this analysis, we theorize that patients might experience psychological distress when requiring urgent spine surgery, and could have worse health status overall predisposing to the need for urgent treatment, both of which could be related to worse hospital ratings.

We also found that any adverse event, readmission, and prolonged hospital stay were associated with non-top-box hospital ratings. We believe we observed an association between adverse events and patient satisfaction because patients could view adverse outcomes as resulting from worse care, as opposed to a risk inherent to spine surgery. These findings are consistent with multiple existing studies, which highlighted both complications [16, 19, 29, 35, 43, 44] and readmissions [29] as negatively affecting HCAHPS scores. Other studies, however, have found no association between complications and HCAHPS scores [1, 5, 7, 17, 18, 21, 25, 37-39, 47], including in spine surgery. Integrating earlier reports, it is difficult to evaluate the effect of complications on HCAHPS scores overall given the heterogeneity in patients, interventions, and outcomes evaluated across studies. However, regarding spine surgery specifically, it has been suggested that adverse post-surgical outcomes do not lead to worse HCAHPS scores [15, 25]. Differences between our study and these could result from differences in outcomes tracked. A study of patients receiving lumbar spine surgery found no association between patient-reported health status and hospital rating, though unlike our report this study did not examine specific adverse events [25]. Similarly, another study of patients receiving spine surgery found no association between any complication or a range of adverse events with patient-reported satisfaction. This study from Hopkins et al. [15] appears reliable as it uses a large sample and rigorous analysis. However, it is likely that our results differ from these authors’ because although they considered a range of adverse outcomes, these authors did not evaluate the impact of readmission on HCAHPS scores, which we identify as independently associated with hospital rating in our study.

Given that there were relatively few adverse events among our study population, the study might be underpowered in terms of detecting whether major and minor adverse events are associated with hospital rating. Given the discrepancies between this study and the limited existing literature around spine surgery and HCAHPS, as has been suggested in previous analyses [15, 25], future studies should evaluate the impact of adverse outcomes and their severity on patient’s hospital ratings after spine surgery.

In conclusion, we found three factors that independently predicted a top-box hospital rating (older age, sex [men], and cervical surgery), and several that predict lower hospital ratings (ASA Classes II to IV, nonelective surgery, any adverse event, readmission, and prolonged hospital stay). Identification of patient factors present preoperatively that are independently associated with HCAHPS scores suggests that the survey has limited utility in measuring the quality of care delivered to patients undergoing spine surgery. HCAHPS responses for this patient population should be interpreted cautiously and should consider the factors associated with hospital rating identified here. Additionally, given conflicting reports regarding the effect of adverse events on HCAHPS scores after spine surgery, future work should aim to more fully characterize this relationship.