Abstract
Study Design:
Retrospective cohort study
Objective:
To assess non-responder biases for the HCAHPS survey following spine surgery.
Summary of Background Data:
The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey is a publicly reported patient satisfaction survey. In addition to having the potential of impacting a hospital’s reputation, it is directly linked to government reimbursement. However, it is known that a minority of patients return this survey, and it is expected that there are non-responder biases.
Methods:
All adult inpatient spine surgery patients at a single institution between January 2013–August 2017 at a single institution were selected for retrospective analysis. Patient demographics and perioperative outcomes were assessed as potential predictors of not returning HCAHPS surveys. Univariate and multivariate analyses were performed.
Results:
Of 5,517 spine surgeries analyzed, 1,505 (27.3%) patients returned the HCAHPS survey. Response rate was variable based on patient characteristics (with statistically significant differences based on age, functional status, race, and ASA score) but not variable based on anatomic region of the spine surgery. Multivariate analysis revealed that patients who did not return the HCAHPS survey were more likely to be black/African American (OR=2.8, p<0.001), have a higher ASA score (OR 1.76, p<0.001), and have had a major adverse event (OR=1.66; p=0.001), minor adverse event (OR=2.50; p<0.001), discharged to a destination other than home (OR=2.16, p<0.001), hospital readmission (OR=2.58; p<0.001), and a long hospital length of stay (OR=1.28, p=0.001).
Conclusion:
For spine surgery patients, patient characteristics and perioperative outcomes were found to be significantly associated with the non-responder bias for HCAHPS surveys. Although the potential resultant bias in HCAHPS scores cannot be directly determined, this must be considered in interpreting the results of such satisfaction surveys given that less than one third of patients actually completed this survey in the study population.
Keywords: spine surgery, The American College of Surgeons National Surgical Quality Improvement Program, NSQIP, HCAHPS, surgical outcomes, spine surgery, patient satisfaction, CAHPS, survey return, non-responder bias, Hospital Consumer Assessment of Healthcare Providers and Systems survey
INTRODUCTION
Over the past thirty years, increasing attention has been placed on the patient’s perception of their health care.1 Based on this, in 2008, the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) was implemented as the first national, standardized survey of patients’ perception of hospital care and public reporting of HCAHPS scores became mandatory.2 In addition to the importance of patient satisfaction and the effects of such surveys on the reputation of hospitals, the Medicare’s value-based program has linked a portion of a hospital’s Medicare compensation to HCAHPS survey performance.3
Despite being well recognized as a metric for “healthcare quality”, studies have shown that HCAHPS survey scores are affected by a variety of confounding factors that appear to lie outside of health-related outcomes. For example, a randomized sample of 7,555 responders in a study of a nationally representative sample of patients demonstrated that those who were given telephone and interactive voice response surveys provided more positive evaluations that those who took mail or mixed surveys (mail surveys with telephone follow-ups).4 Additionally, a study involving 580 knee arthroplasty patients at a single institution showed that patients who were male, African American, of lower socioeconomic status, and had a shorter length of stay were more likely to report higher levels of inpatient satisfaction.5 A further study involving 2,758 surgical patients across 8 different subspecialties showed that certain preadmission medications such as chronic NSAIDs and less benzodiazepines, were associated with higher patient satisfaction scores.6 Although the validity of the HCAHPS survey and correlation with clinical outcomes may be questioned,7–9 it is part of current clinical practice across the United States.
Beyond the above variables affecting HCAHPS responses, nonresponse bias may affect study results. Nonresponse bias occurs when a significant fraction of the surveyed population does not respond to the survey, and non-responders differ significantly in results from responders. This becomes particularly notable as the national HCAHPS survey response rate for all inpatient hospital encounters between April 2016 and March 2017 was reported to be only 28%,10 and lower response rates may introduce a higher risk of non-responder bias.11–13
In order to accurately interpret HCAHPS patient satisfaction data, it is important to increase our understanding of differences between those patients who do and do not return the HCAHPS survey. While there are studies in literature looking at the impact of patient factors on HCAHPS scores, to our knowledge, there is has been no prior published research that examines how perioperative and postoperative outcomes, in addition to patient characteristics, cumulatively affect HCAHPS response rates. The current study evaluates the non-response rate for HCAHPS surveys in a spine surgery patient population at a single academic institution and uses multivariate analysis to assess potential patient characteristics and perioperative outcomes associated with this nonresponse rate.
METHODS
Patient Inclusion/Exclusion Criterion
The current protocol was approved by our Institutional Review Board (IRB). Consecutive patients who underwent an inpatient spinal procedure at a single academic institution between January 2013–August 2017 were selected for retrospective analysis (all had been sent the HCAHPS questionnaire).
The cohort of patients was identified using the Current Procedural Terminology (CPT) codes pertaining to spinal procedures. Based on these CPT codes, cases were determined to be cervical, thoracic, lumbar, or multi-region. The individual surgeons performing the procedures was also anatomized and recorded as surgeon number. Surgeon specific HCAHPS return rate was studied for surgeons who had performed a minimum of 100 procedures in our dataset.
The exclusion criterion used were: patient’s age<18 years, patients that died during their hospital stay, and patients that were left the hospital against medical advice (AMA). The last two exclusion criterion were used because those patients would not receive the HCAHPS survey. Finally, any readmissions were not separately assessed for HCAHPS responses.
HCAHPS survey responses
The HCAHPS survey is mailed to all the patients who have an inpatient spinal procedure (defined as a hospital length of stay of at least 1 day) performed at our institution. The survey includes 27 questions, used nationwide, that are broadly divided into categories: communication with doctors, responsiveness of hospital staff, cleanliness of hospital environment, quietness of hospital environment, pain management, communication about medicines, discharge information, overall rating of hospital and recommendation of hospital.14 HCAHPS surveys are sent to patients within 2 days of discharge from our institution. The responses to these HCAHPS surveys are mailed back to our institution.
Patient characteristics and perioperative outcomes
Patient characteristics were tabulated. These included ages, sex, height and weight (used to calculate boney mass index [BMI]), preoperative functional status, American Society of Anesthesiologists (ASA) classification, and race. The ASA score was used as a marker of comorbidity, consistent with prior literature in orthopedics.15,16
For classification of the anatomic level of spinal procedure, CPT codes were used to determine if the procedure was “cervical”, “thoracic”, and “lumbar”. If this still did not lead to clarity, chart review was completed to determine the procedure location as one single spinal level or designated as multilevel.
Thirty-day perioperative outcomes (hospital length of stay, adverse events, and readmissions) were also recorded by our National Surgical Quality Improvement Program (NSQIP) team. These data were available for the study population and used to correlated variables with HCAHPS survey response rates.
The occurrence of a major adverse event was defined as the occurrence of any of the following: death, unplanned return to operating room, deep vein thrombosis, ventilator use >48 hours, stroke/cerebrovascular accident, cardiac arrest, unplanned intubation, acute renal failure, sepsis/septic shock, pulmonary embolism, deep infection, return to the operating room, and readmission within 30 days. The occurrence of a minor adverse event was defined as the occurrence of any of the following: wound disruption, urinary tract infection, superficial infection, pneumonia, and progressive renal insufficiency. The occurrence of any adverse event was defined as the occurrence of a major or minor adverse event.
Statistical Analysis
Patients were divided into two groups: survey returned, and survey not returned, based on the HCAHPS survey return status. Statistical differences between the two groups on preoperative and perioperative metrics were studied.
For the univariate analysis, Chi-squared tests or Fisher’s exact tests were used for categorical variables. Statistical significance was set at α = 0.05 initially and Bonferroni corrected wherever applicable (corrected significance level set at α = 0.007 for the patient demographic analysis, and α = 0.008 for the surgical event outcomes analysis).
Logistic regressions were performed for both demographic and surgical variables affecting HCAHPS survey return. The patient cohorts studied were based on HCAHPS survey return status. The first regression was studied patient demographic factors and nature of procedure and included: ASA score, functional, BMI, sex, and age, race, and spinal level of procedure. The second regression studied post-operative factors: major adverse event occurrence, minor adverse event occurrence, readmission, hospital discharge destination, and long hospital length of stay (>3 days). In addition, the second regression controlled for all the variables studied in the first regression. For these multivariate analyses, statistical significance was initially set at α = 0.05 (and subsequently Bonferroni correctly; with statistical significance set at α = 0.008 for Table 3, and α = 0.01 for Table 4), and 95% confidence intervals [CI] are reported.
Table 3:
Multivariate Analysis of Demographic factors associated with not returning the HCAHPS survey
| Type | Likelihood of Not Returning Survey | ||
|---|---|---|---|
| Total Cases (N= 5,517) | |||
| OR | 95% CI | P-value | |
| Age | |||
| 18–34 | 1 | – | – |
| 35–54 | 0.71 | [0.52–0.97] | 0.029 |
| 55–74 | 0.42 | [0.31–0.58] | <0.001 |
| 75+ | 0.4 | [0.29–0.57] | <0.001 |
| Sex | |||
| Male | 1 | – | – |
| Female | 0.92 | [0.81–1.05] | 0.214 |
| BMI | |||
| <25 | 1 | – | – |
| 25–30 | 0.88 | [0.74–1.04] | 0.139 |
| 30–35 | 0.90 | [0.75–1.08] | 0.262 |
| >35 | 0.99 | [0.80–1.23] | 0.946 |
| ASA | |||
| 1 | 1 | – | – |
| 2 | 1.11 | [0.85–1.46] | 0.438 |
| 3+ | 1.75 | [1.31–2.35] | <0.001 |
| Functional Status (Prior to Surgery) | |||
| Independent | 1 | – | – |
| Partially/Totally Dependent | 2.7 | [1.67–4.38] | <0.001 |
| Race | |||
| White | 1 | – | – |
| Black/African American | 2.8 | [2.10–3.72] | <0.001 |
| Asian | 1.18 | [0.60–2.30] | 0.636 |
| Procedural Location | |||
| Cervical | 1 | – | – |
| Thoracic | 1.21 | [0.89–1.64] | 0.219 |
| Lumbar | 1.06 | [0.93–1.22] | 0.383 |
| Multilevel | 1.02 | [0.56–1.87] | 0.945 |
Factors in Model: age, sex, BMI, ASA class, functional status, race, procedural location
Bolding indicates statistical significance at p<0.008 (Bonferroni corrected)
Table 4:
Multivariate Odds ratio of HCAHPS return status following spine procedures as a functional of post-operative outcomes
| Type | Likelihood of Not Returning Survey | ||
|---|---|---|---|
| Total Cases (N= 5,517) | |||
| OR | 95% CI | P-value | |
| Any Adverse Event | 1.94 | [1.49–2.52] | <0.001 |
| Major Adverse Event | 1.66 | [1.24–2.21] | 0.001 |
| Minor Adverse Event | 2.50 | [1.65–3.78] | <0.001 |
| Readmission | 2.58 | [1.90–3.50] | <0.001 |
| Discharged Place Other Than Home | 2.16 | [1.82–2.55] | <0.001 |
| Long Hospital Length of Stay (> 3 Days) | 1.28 | [1.11–1.49] | 0.001 |
Separate regressions above also controlled for: age, sex, BMI, ASA class, functional status, race, procedural location
Bolding indicates statistical significance at p<0.01 (Bonferroni corrected)
Statistical analysis was performed in Stata version 13.1 (StataCorp, LP, College Station, TX).
RESULTS
Patient Procedure / Procedure Univariate Analyses:
A total of 5,517 patients met criteria for inclusion in this study. Table 1 displays differences between responders and non-responders based on patient and procedural characteristics. In total, 1,505 patients completed and returned an HCAHPS satisfaction survey (“Survey Returned”), and 4,012 patients did not respond “Survey Not Returned”). This led to a survey response rate of 27.3% (Figure 1).
Table 1:
Demographics of patients undergoing spine procedures organized by HCAHPS survey return status
| Type | Survey Returned | Survey Not Returned | Univariate P-value |
|---|---|---|---|
| Cases (N= 5,517) (%) | 1,505 (27.28%) | 4,012 (72.72%) | |
| Age: Mean [SD] | 60.70 [0.34] | 57.48 [0.23] | <0.001 |
| 18 – 34 | 63 (4.19%) | 301 (7.50%) | |
| 35 – 54 | 411 (27.31%) | 1,400 (34.90%) | |
| 55 – 74 | 828 (55.02%) | 1,826 (45.51%) | |
| 75+ | 203 (13.49%) | 485 (12.09%) | |
| Sex | 0.279 | ||
| Male | 756 (50.23%) | 2,081 (51.87%) | |
| Female | 749 (49.77%) | 1,931 (48.13%) | |
| BMI: Mean [SD] | 29.21 [0.14] | 29.67 [0.10] | 0.022 |
| < 25 | 340 (22.59%) | 912 (22.73%) | |
| 25 – 30 | 561 (37.28%) | 1,379 (34.37%) | |
| 30 – 35 | 391 (25.98%) | 1,026 (25.57%) | |
| > 35 | 213 (14.15%) | 695 (17.32%) | |
| Functional Status | <0.001 | ||
| Independent | 1,481 (98.41%) | 3,852 (96.01%) | |
| Partially/Totally dependent | 24 (1.59%) | 160 (3.99%) | |
| ASA: Mean [SD] | 2.32 [0.02] | 2.41 [0.01] | <0.001 |
| 1 | 92 (6.11%) | 246 (6.13%) | |
| 2 | 862 (57.28%) | 1,963 (48.93%) | |
| 3+ | 551 (36.61%) | 1,803 (44.94%) | |
| Race | <0.001 | ||
| Asian | 12 (.80%) | 33 (.82%) | |
| Black/African American | 58 (3.85%) | 421 (10.49%) | |
| Unknown/Unreported | 75 (4.98%) | 314 (7.83%) | |
| White | 1,360 (90.37%) | 3,244 (80.86%) | |
| Procedure Location | 0.203 | ||
| Cervical | 454 (30.17%) | 1,246 (31.06%) | |
| Thoracic | 71 (4.72%) | 246 (6.13%) | |
| Lumbar | 964 (64.05%) | 2,468 (61.52%) | |
| Multi-region | 16 (1.06%) | 52 (1.30%) |
Underlining indicates most common subgroup
Bolding indicates statistical significance at p < 0.007 (Bonferroni corrected)
ASA = American Society of Anesthesiologists classification
Figure 1: HCAHPS Survey Return Rate for Spine Surgeries.
Pie chart depicting HCAHPS return rate in our dataset. The grey region depicts HCAHPS surveys not-returned while the black region depicts the surveys returned
In the univariate analysis (Table 1), survey non-responders were statistically more likely to be younger (p<0.001), more functionally dependent (p<0.001), have higher ASA scores (p<0.001), be minority/non-white (p<0.001). Responders did not differ from non-responders in terms of sex (p=0.279), BMI (p=0.022), or anatomical location of procedure (p=0.203).
The 5,517 procedures were performed by 27 surgeons. Out of this group, fifteen surgeons met the criterion for performing more than 100 procedures in the dataset. Figure 2 displays the number of surgeries performed by each of the numbered surgeons (shown on the bottom of the figure) and the respective HCAHPS response rates, ranging from 15.1%–34.3% (mean=26.2%, SD=5.49) and not clearly related to the number of surgeries performed.
Figure 2: HCAHPS Response Rate by Surgeon.
Bar chart depicting survey response rate on the y-axis and anonymized surgeon code (in descending order of number of procedures performed) on the x-axis. Surgeons who performed more than 100 procedures in the dataset were included in this figure. Number of procedures performed per surgeon is denoted below the x-axis.
Perioperative Outcome Univariate Analyses:
For the overall study population, 9.4% had any adverse event, 7.0% had a major adverse event, 4.3% had a minor adverse event, 7.8% were readmitted, 27.6% had a non-home discharge, and 32.5% had a hospital length of stay greater than 3 days. The breakdown of these by HCAHPS response status are shown in Table 2.
Table 2:
Surgical event outcomes following spine procedures organized by status of HCAHPS survey return
| Type | Survey Returned | Survey Not Returned | Univariate P-value |
|---|---|---|---|
| Cases (N= 5,517) (%) | 1,505 (27.28%) | 4,012 (72.72%) | |
| All Adverse Events | 83 (5.51%) | 433 (10.79%) | <0.001 |
| Major Adverse Events | 67 (4.45%) | 317 (7.90%) | <0.001 |
| Deep Infection | 10 (0.66%) | 69 (1.72%) | |
| Sepsis/Septic shock | 10 (0.66%) | 80 (1.99%) | |
| Ventilator >48 hrs | 4 (0.27%) | 40 (1.00%) | |
| Unplanned Intubation | 2 (0.13%) | 41 (1.02%) | |
| Acute Renal Failure | 1 (0.07%) | 4 (0.10%) | |
| Pulmonary Embolism | 15 (1.00%) | 41 (1.02%) | |
| Deep Vein Thrombosis | 11 (0.73%) | 54 (1.35%) | |
| Cardiac Arrest | 1 (0.07%) | 12 (0.30%) | |
| MI | 2 (0.13%) | 15 (0.37%) | |
| Stroke | 1 (0.07%) | 8 (0.20%) | |
| Unplanned Return to OR | 29 (1.93%) | 131 (3.27%) | |
| Death | 0 (0.00%) | 14 (0.35%) | |
| Minor Adverse Events | 30 (1.99%) | 209 (5.21%) | <0.001 |
| Superficial Infection | 3 (0.20%) | 47 (1.17%) | |
| Wound Disruption | 2 (0.13%) | 8 (0.20%) | |
| Pneumonia | 14 (0.93%) | 89 (2.22%) | |
| UTI | 12 (0.80%) | 70 (1.74%) | |
| Progressive Renal Insufficiency | 0 (0.00%) | 8 (0.20%) | |
| Readmissions | 54 (3.59%) | 374 (9.32%) | <0.001 |
| Discharge Disposition | <0.001 | ||
| Home | 1,229 (81.66%) | 2,765 (68.92%) | |
| Other | 276 (18.34%) | 1,247 (31.08%) | |
| Long hospital length of stay (> 3 days) | 411 (27.31%) | 1,384 (34.50%) | <0.001 |
Underlining indicates most common subgroup
Bolding indicates statistical significance at p < 0.008 (Bonferroni corrected)
ASA = American Society of Anesthesiologists classification
In the univariate analysis of aggregated adverse outcomes (Table 2), survey non-responders were statistically more likely to have suffered any, major, and minor adverse events (p<0.001). In addition, non-responders were statistically more likely to have a hospital readmission (p<0.001), be discharged to a site other than home (p<0.001), and have a long hospital length of stay (>3 days) (p<0.001). A significantly higher occurrence of both major (7.90% vs. 4.45%) and minor (5.21% vs. 1.99%) adverse events were seen in patients who did not return the survey (p<0.001). Non-responders were also more likely to be discharged to a destination other than home (p<0.001), be readmitted (9.32% vs. 3.59%, p<0.001), and have a longer (>3 days) hospital length of stay (p<0.001).
Multivariate Analysis:
The multivariate analysis was set up in order to showcase factors that are associated with non-response. The odds ratios higher than 1 indicate factors directly associated with non-response, while odds ratios lower than 1 indicate factors associate with a higher response rate. In the first multivariate regression (Table 3) that studied patient preoperative characteristics and procedure location, non-responders were more likely to have a higher (3+) ASA score (OR=1.75; 95% CI=1.31–2.35; p<0.001), be black/African American (OR=2.80; 95% CI= 2.10–3.72; p<0.001), and have a partially/totally dependent functional status (OR=2.70; 95% CI=1.67–4.38; p<0.001). In addition, age greater than 75 (OR=0.40, 95% CI= 0.29–0.57, p<0.001), and patients aged 55–74 (OR=0.42, 95% CI=0.31–0.58, p<0.001) was associated with a higher rate of survey response,; thereby showing how lower age was associated with non-response which agrees with the results of the univariate analysis.
On the second multivariate regression (Table 4) that studied surgical outcomes while controlling for preoperative factors, the incidence of any adverse event (OR=1.94; 95% CI=1.49–2.52; p<0.001) major adverse event (OR=1.66; 95% CI=1.24–2.21; p=0.001), minor adverse event (OR=2.50; 95% CI=1.65–3.78; p<0.001), readmission (OR=2.58; 95% CI=1.90–3.50; p<0.001), discharge to a place other than home (OR=2.16; 95% CI=1.82–2.55; p<0.001), and long (>3 days) hospital length of stay (OR=1.28; 95% CI= 1.11–1.49; p=0.001) were found to be associated with non-responders of the HCAHPS patient satisfaction survey. The multivariate analyses have been summarized in a forest plot (Figure 3).
Figure 3: Forest plot depicting the significant variables from the multivariate regression on both demographic and post-operative factors highlighting HCAHPS non-responder bias following spine procedures.
Forest plot with the odds ratios for the significant variables in the multivariate regression studying patient characteristics and post-operative factors against HCAHPS response status
DISCUSSION
The HCAHPS is used to assess patient satisfaction after hospitalization. The goal of the HCAHPS program, created by the Centers for Medicare and Medicaid Service (CMS), was for better transparency, improved consumer decision making, and to incentivize hospitals to deliver high quality care and to monitor patient satisfaction trends overtime.2,17 Nonetheless, it remains unclear whether there is an association between HCAHPS scores and health care quality.18,19
The current study found that the response rate for HCAHPS surveys sent to over five thousand spine surgery patients was 27.3%. This is very much in line with the overall national average return rate of 28%.10 With less than a third of the population returning the survey, there is clearly the potential for a bias in those who return the survey, which could bias the responses received. There was also identified variability of patients of different surgeons to return the HCAHPS survey (range 15.1–34.3%), but this did not seem to correlate with the volume of surgeries that those surgeons performed. Further, the rate of survey return did not correlate with the spinal anatomic region where the surgery was performed.
Patients who did not respond to the survey were more likely to be younger, be partially or totally dependent prior to surgery, have a higher ASA classification, be discharged to destination other than home, and be non-white in race. No association between patient’s sex and response rate was seen. The fact that different portions of the surgical population are more or less likely to return the HCAHPS survey is of interest because the actual HCAHPS responses have been shown to vary based on patient characteristics.5,6,9,20,21 To that end, in a study of primary care patients returning patient satisfaction surveys (predating the HCAHPS era), a positive correlation was found between mean patient satisfaction rating and response rate of the data.22
A strong non-response bias skews the results of the HCAHPS survey towards certain demographics and subgroups. For example, the opinions of older adults are more strongly represented in the average result simply because they are more likely to respond the survey. Similarly, patients discharged to home are more likely to return the survey, and hence the end results are more representative of this sub-population compared to the entire patient population. Particularly, since the surveys are sent to patients within 2 days of discharge, it is possible that patients not discharged to home did not see/receive the survey causing a lower response rate in this patient population. In addition, surveys are sent to patients either as a physical copy or as an e-survey at our institution, this can further create confounders regarding who responds to surveys based on reliable access to physical mail/technological proficiency.
Patients who did not respond to the survey were also more likely to have experienced both minor and major adverse event during their hospitalization, and were more likely to have a longer length of hospitalization. They were also more likely to have had a readmission, either related or unrelated to their index spine surgery. There is limited literature on the impact of perioperative factors on HCAHPS response rate, however, studies in literature have shown shorter length of hospitalization, higher hospital surgical volume, and lower pain scores to be associated with higher HCAHPS scores.6,23 Thus, these results show that patients with poorer postoperative outcomes are less likely to return the survey and thus, the overall survey results are not as representative of this patient subpopulation.
Response rates are an important metric for hospitals. For example, hospitals receiving 100 or more completed HCAHPS surveys within a reporting period to be eligible for Hospital value-based purchasing, and HCAHPS star ratings.24 To help hospitals obtain more patient population representative HCAHPS results, suggestions have been made by the Center for Medicare and Medicaid services to help improve response rates. Limiting supplemental items added to the end of the survey, offering surveys in multiple languages, verifying patient contact information, offering mixed mode survey administration, and scheduling patient callbacks (for telephone-based surveys) are associated with higher response rates.24
The current study has some limitations. As with any retrospective study, there is a possibility of unmeasured confounding variables. The reason for returning or not returning the HCAHPS surveys could not be assessed. Further, the reason for the perioperative outcomes affecting the return rates could not be assessed. Finally, the trends reported here are form a single institution that might not be fully generalizable.
Nonetheless, the current study has clear strengths. There was a large, representative spine surgery patient population surveyed. Specific attention was given to the rate of return of the HCAHPS surveys. While the rate of return of such surveys can be lost in the description of actual responses, the current study isolated this variable to highlight the variability in simply returning the survey. Furthermore, both patient demographics and post-operative characteristics were studied using multivariate analysis to establish variables independently associated with increased/decreased rate of survey completion.
Overall, for spine surgery patients, the current study confirmed that the response rate for HCAHPS surveys is low and that patient characteristics and perioperative outcomes were significantly associated with the non-responder bias for HCAHPS surveys. Additionally, via univariate and multivariate analysis, HCAHPS responders are shown to be a non-representative subpopulation of the total patient sample studied. This must be considered in interpreting the results of such satisfaction surveys given that less than one third of patients actually completed this survey in the study population.
Acknowledgments
The manuscript submitted does not contain information about medical device(s)/drug(s).
The National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health funds were received in support of this work.
Relevant financial activities outside the submitted work: consultancy.
REFERENCES
- 1.Sitzia J, Wood N. Patient satisfaction: a review of issues and concepts. Soc Sci Med 45(12): 1829, 1997 [DOI] [PubMed] [Google Scholar]
- 2.Mann RK, Siddiqui Z, Kurbanova N, Qayyum R. Effect of HCAHPS reporting on patient satisfaction with physician communication. J Hosp Med 11(2): 105, 2016 [DOI] [PubMed] [Google Scholar]
- 3.Hospital Value-Based Purchasing. In. Medicare Learning Network: Department of Health and Human Services Centers for Medicare and Medicaid Services. 2015
- 4.Elliott MN, Zaslavsky AM, Goldstein E, Lehrman W, Hambarsoomians K, Beckett MK, Giordano L. Effects of survey mode, patient mix, and nonresponse on CAHPS hospital survey scores. Health Serv Res 44(2 Pt 1): 501, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Peres-da-Silva A, Kleeman LT, Wellman SS, Green CL, Attarian DE, Bolognesi MP, Seyler TM. What Factors Drive Inpatient Satisfaction After Knee Arthroplasty? J Arthroplasty 32(6): 1769, 2017 [DOI] [PubMed] [Google Scholar]
- 6.Maher DP, Wong W, Woo P, Padilla C, Zhang X, Shamloo B, Rosner H, Wender R, Yumul R, Louy C. Perioperative factors associated with HCAHPS responses of 2,758 surgical patients. Pain Med 16(4): 791, 2015 [DOI] [PubMed] [Google Scholar]
- 7.Day MS, Hutzler LH, Karia R, Vangsness K, Setia N, Bosco JA 3rd. Hospital-acquired conditions after orthopedic surgery do not affect patient satisfaction scores. Journal for healthcare quality : official publication of the National Association for Healthcare Quality 36(6): 33, 2014 [DOI] [PubMed] [Google Scholar]
- 8.Anil U, Elbuluk AM, Ziegler J, Schwarzkopf R, Long WJ. Hospital Consumer Assessment of Healthcare Providers and Systems Scores Do Not Predict Outcomes After Total Hip Arthroplasty. The Journal of arthroplasty 33(2): 337, [DOI] [PubMed] [Google Scholar]
- 9.Levin JM, Winkelman RD, Smith GA, Tanenbaum J, Benzel EC, Mroz TE, Steinmetz MP. The association between the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey and real-world clinical outcomes in lumbar spine surgery. The spine journal : official journal of the North American Spine Society 17(11): 1586, 2017 [DOI] [PubMed] [Google Scholar]
- 10.Summary of HCAHPS Survey Results, April 2016 to March 2017 Discharges. In.: Centers for Medicare & Medicaid Services, Baltimore, MD: 2017 [Google Scholar]
- 11.Tyser AR, Abtahi AM, McFadden M, Presson AP. Evidence of non-response bias in the Press-Ganey patient satisfaction survey. BMC Health Serv Res 16(a): 350, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Groves RM, Peytcheva E. The Impact of Nonresponse Rates on Nonresponse BiasA Meta-Analysis. Public Opinion Quarterly 72(2): 167, 2008 [Google Scholar]
- 13.Lewis EF, Hardy M, Snaith B. Estimating the effect of nonresponse bias in a survey of hospital organizations. Eval Health Prof 36(3): 330, 2013 [DOI] [PubMed] [Google Scholar]
- 14.CAHPS® Hospital Survey (HCAHPS) Quality Assurance Guidelines Version 12.0. In.: Department of Health and Human Services Centers for Medicare and Medicaid Services. 2017
- 15.Bovonratwet P, Nelson SJ, Ondeck NT, Geddes BJ, Grauer JN. Comparison of Thirty-Day Complications Between Navigated And Conventional Single-Level Instrumented Posterior Lumbar Fusion: A Propensity Score Matched Analysis. Spine, 2017 [DOI] [PubMed] [Google Scholar]
- 16.Buerba RA, Fu MC, Grauer JN. Anterior and posterior cervical fusion in patients with high body mass index are not associated with greater complications. The spine journal : official journal of the North American Spine Society 14(8): 1643, 2014 [DOI] [PubMed] [Google Scholar]
- 17.Giordano LA, Elliott MN, Goldstein E, Lehrman WG, Spencer PA. Development, implementation, and public reporting of the HCAHPS survey. Med Care Res Rev 67(1): 27, 2010 [DOI] [PubMed] [Google Scholar]
- 18.Isaac T, Zaslavsky AM, Cleary PD, Landon BE. The relationship between patients’ perception of care and measures of hospital quality and safety. Health Serv Res 45(4): 1024, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Malpani R, Hilibrand AS, Grauer JN. Evolution and Use of Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) Surveys and Their Application for Spinal Surgery Patients. Contemporary Spine Surgery 19(5): 1, 2018 [Google Scholar]
- 20.Dad T, Tighiouart H, Fenton JJ, Lacson E Jr., Meyer KB, Miskulin DC, Weiner DE, Richardson MM. Evaluation of non-response to the In-Center Hemodialysis Consumer Assessment of Healthcare Providers and Systems (ICH CAHPS) survey. BMC Health Serv Res 18(1): 790, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Levin JM, Winkelman RD, Smith GA, Tanenbaum JE, Benzel EC, Mroz TE, Steinmetz MP. Impact of Preoperative Depression on Hospital Consumer Assessment of Healthcare Providers and Systems Survey Results in a Lumbar Fusion Population. Spine 42(9): 675, 2017 [DOI] [PubMed] [Google Scholar]
- 22.Mazor KM, Clauser BE, Field T, Yood RA, Gurwitz JH. A demonstration of the impact of response bias on the results of patient satisfaction surveys. Health Serv Res 37(5): 1403, 2002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tevis SE, Kennedy GD. Patient satisfaction: does surgical volume matter? The Journal of surgical research 196(1): 124, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Improving Response Rates of HCAHPS Hospitals. Hospital CAHPS Podcast Series: Health Services Advisory Group, 2018. Accessed on 8/7/19 at https://innovation.cms.gov/initiatives/bundled-payments/index.html [Google Scholar]