Association Between Nonmodifiable Demographic Factors and Patient Satisfaction Scores in Spine Surgery Clinics

• Updated on: 2019-06-24
• Read original article here

Patient satisfaction surveys are gaining clinical importance as they are used as a metric to compare quality of care delivery between institutions and guide reimbursement. Press Ganey Associates is the largest administrator of satisfaction surveys in the United States, with more than 10,000 centers and approximately 50% of hospitals using their services. In the ambulatory care setting, the Press Ganey Clinician and Group Consumer Assessment of Healthcare Providers and Systems survey is widely used to assess patient satisfaction. Medicare's value-based purchasing program provides financial rewards to hospitals with higher patient satisfaction survey scores, and there is evidence that modifiable factors such as improved communication with nurses and physicians, shorter length of hospital stay, and pain management are associated with higher satisfaction.

The value-based purchasing program model is predicated on satisfaction as a modifiable attribute used to compare institutions; however, there is an increasing body of evidence indicating that nonmodifiable factors are significantly associated with patient satisfaction. Studies in both nonorthopedic specialties and non-spine orthopedic subspecialties have suggested that age, education, and sex influence patient satisfaction. The relationship between nonmodifiable patient characteristics and Press Ganey survey scores for ambulatory spine surgery patients remains unclear. Understanding the factors that influence Press Ganey surveys of outpatient spine surgery patients may permit better use of survey results and may direct interventions to increase spine patient satisfaction.

The purpose of this study was to determine whether nonmodifiable demographic factors affected the satisfaction scores on the Press Ganey Clinician and Group Consumer Assessment of Healthcare Providers and Systems survey of ambulatory spine patients in a large tertiary care network.

The Press Ganey database (Press Ganey Associates, Wakefield, Massachusetts) was queried for all Clinician and Group Consumer Assessment of Health-care Providers and Systems surveys administered to ambulatory spine surgery clinic patients at the authors' institution between May 2016 and September 2017. All visits were within an academic tertiary care network of ambulatory centers staffed by 11 fellowship-trained orthopedic surgeons. Surveys were randomly administered to 20% of patients, either on paper or online at each patient's discretion, after their visit. Patient and physician demographic factors were gathered for each visit linked with a completed survey. The six age ranges of the Press Ganey database were used: 0 to 17 years, 18 to 34 years, 35 to 49 years, 50 to 64 years, 65 to 79 years, and 80 to 100 years. Responses to “overall provider rating” (OPR) and “recommend this provider's office” (RP) survey questions were used as surrogates for overall patient satisfaction (Table 1). The OPR question was answered on a scale of 0 to 10 by the respondent. The responses to the RP question were converted to ordinal variables to allow scaling of the satisfaction score, such that 3 was the “Yes, definitely” response and 1 was the “No” response. The results for each question were then normalized to a 0 to 100 scale to allow direct comparison between the two questions. Press Ganey Clinician and Group Consumer Assessment of Healthcare Providers and Systems Questions Analyzed Mean comparison testing was performed to calculate associations between survey responses, patient demographics, and physician demographics. A multivariate regression analysis was performed to determine independent factors. Statistical analysis was performed using SPSS Statistics for Windows version 23.0 software (IBM, Armonk, New York). Significance level was set at α<0.05.

A total of 1400 survey responses were included (Table 2). There were 605 female patients (43.2%) and 781 male patients (55.8%). Sex was not reported for 14 patients (1%). Female patients tended to have higher satisfaction scores than male patients; however, this difference was not significant. The mean age of the patients was 54.9±17.3 years. The most common age range of the patients was 50 to 64 years, which represented 499 responses (36.0%). Patients 18 to 34 years old had significantly lower scores for the OPR (mean, 85.8) and RP (mean, 81.0) questions than all other patients (OPR mean, 95.3, P<.001; RP mean, 92.6, P<.001). Among adult patients, a weak correlation was noted between increasing patient age and improved satisfaction scores (OPR: R=0.113, P<.001; RP: R=0.108, P<.001) (Figure 1). In contrast, pediatric patients had higher mean satisfaction scores than nonpediatric patients (OPR, 97.3 vs 94.4, respectively, P=.01; RP, 93.0 vs 91.7, respectively, P=.684). Education level showed a significant inverse correlation with satisfaction scores; however, the correlation was weak (OPR: R=−0.114, P<.001; RP: R=−0.122, P<.001). Satisfaction scores tended to trend downward after attending college (Figure 2). Patients who had attained graduate level education had the lowest satisfaction scores (OPR, P=.001; RP, P<.001). Insurance type was significantly associated with satisfaction score. Patients with commercial insurance had significantly lower mean scores for the RP (P=.042) and OPR (P=.022) questions than patients with other insurance types. Following commercial insurance, which had the lowest mean scores (OPR, 93.6; RP, 90.0), satisfaction increased in the following order: workers' compensation (OPR, 94.1; RP, 93.5), Medicaid (OPR, 95.1; RP, 94.0), Medicare (OPR, 96.4; RP, 94.0), and no insurance (OPR, 98.4; RP, 96.5). Most patients chose to complete the survey online, with only 193 completing it on paper (13.9%). Patients who completed the survey on paper had significantly lower ratings than those who completed the survey online (OPR, P<.001; RP, P=.006). Ethnicity was associated with significant differences in satisfaction scores when one ethnic group was compared with the rest of the cohort. Whites represented 75% of the patient population and had significantly higher satisfaction scores than the rest of the patient population (OPR mean, 95.3 vs 92.2, P=.006; RP mean, 93.0 vs 88.5, P=.013, respectively). Asians had lower mean satisfaction scores (OPR, 86.6; RP, 83.5) than the rest of the cohort (OPR, 95.1; RP, 92.0), although they represented only 5% of the total study sample. When the sex and the ethnicity of the patient were in concordance with the sex and the ethnicity of the provider (28.1% of visits), mean satisfaction was significantly higher than when there was no concordance (OPR, 96.1 vs 94.0, P=.021; RP, 94.0 vs 91.0, P=.023, respectively). These results should be interpreted with caution because the sample was small for many of the relationships. Multivariate analysis showed that age, education, concordance of sex and ethnicity, and survey medium (for OPR) maintained significant relationships with satisfaction scores in the presence of one another (Table 3).

As health care shifts to a value-based model, patient satisfaction is an increasingly important metric for quality of care. The authors' goal was to determine if nonmodifiable factors inherent to specific patient populations were associated with patient satisfaction survey results. This study showed that independent, nonmodifiable factors such as age, education level, and survey mode were associated with Press Ganey satisfaction survey results at a large tertiary care institution. Patients 18 to 34 years old had lower satisfaction scores than older patients, and increasing age showed a significant positive correlation with satisfaction among adult patients. In a cross-sectional sample of 20,901 multispecialty outpatients, elderly patients (>65 years) had significantly higher satisfaction scores than non-elderly patients. Similarly, Martin et al found that in a large cohort of surgical patients, younger patients were less likely to be satisfied with their care. Younger age has also been associated with lower satisfaction scores in spine surgery clinics. In contrast, Mistry et al found that in an inpatient hospital cohort of patients undergoing total hip arthroplasty, age was not a significant factor influencing satisfaction; however, this was an inpatient cohort that examined associations with communication as well as pain. Finally, the pediatric patients in the study also had high satisfaction. It is possible that pediatric patients score higher in satisfaction; however, this is a heterogeneous population from a developmental standpoint, and older guardians may have assisted in responding to the survey questions. Patients with a college education or higher had significantly lower scores than less educated patients, and increasing age showed a significant inverse correlation with satisfaction. Studies in nonorthopedic specialties have shown that lower education level is associated with higher satisfaction. In addition, Abtahi el al found this relationship to be true among orthopedic patients responding to outpatient Press Ganey surveys. Paper surveys were associated with higher satisfaction scores than online surveys. There was the potential for confounding in this result because the survey medium was not distributed randomly, although the relationship remained significant on multivariate analysis. Little has been published about the effect of survey medium on patient satisfaction surveys, with one study showing no difference between paper and online satisfaction surveys in a homogeneous population of Canadian school children. Race was associated with differing satisfaction scores within the current data set, with Asians having the lowest satisfaction scores. Male sex was associated with a trend toward lower satisfaction, although this did not reach significance. When the patient's race and sex were concordant with the race and sex of the treating physician, satisfaction was significantly higher. Young et al reported that racial background was associated with satisfaction scores. Furthermore, several multispecialty studies have shown that patients are more likely to choose race-concordant physicians and that race concordance leads to higher satisfaction. These findings reinforce the importance of cultural competence among physicians. The value-based purchasing model used by the Centers for Medicare & Medicaid Services to compare institutions assumes that patient populations given the same level of care at separate institutions will respond similarly to satisfaction surveys. This study showed that within a single tertiary care network, patients with different demographic backgrounds can have significantly different levels of satisfaction. These results suggest that it may be challenging to use patient satisfaction surveys to compare institutions whose patient populations have different demographics. Patient satisfaction is multifactorial and may reflect the ability of a provider and an organization to address the needs of the patient population served. For example, highly educated patients may be frustrated by practices in centers adapted to care for less educated patients. The best satisfaction scores may result from concordance of patient demographics with the type of patient a center is best suited to serve. Future studies will need to compare the effect of nonmodifiable demographics at multiple institutions to elucidate the effect of institutional differences on satisfaction survey results. This study had several limitations, including the flaws inherent to a nonrandomized, retrospective analysis conducted at a single center. The survey results may not be generalizable to other institutions adapted to serve a different subset of patients. The anonymized nature of the Press Ganey data does not allow visit-specific details, such as visit length and wait time, to be collected and analyzed. Similarly, it is not possible to collect patient-specific characteristics, such as diagnosis and comorbid conditions. The surveys may also have nonresponse bias, as survey completion is encouraged but optional. For example, Tyser et al found that younger and male patients were less likely to respond to the Press Ganey survey. In the current study, survey completion percentage varied by surgeon, suggesting that there may be confounding factors affecting patient response rates. Finally, there is a possibility that allowing patients to complete the survey using the medium of their choice (paper vs electronic) biases results; however, the authors attempted to control for this in their multivariate analysis.