Abstract
Objectives:
Multiple factors have been associated with lower satisfaction scores. We hypothesize that patients with a more deprived socioeconomic status will have different patient satisfaction scores than patients of higher socioeconomic status.
Methods:
We reviewed Press Ganey satisfaction scores for new, outpatient visits between January 1, 2014, and December 31, 2018. Due to the high ceiling effects of the survey, “satisfaction” was defined as achieving a perfect score of 100. We determined social deprivation using the 2015 Area Deprivation Index (ADI) which serves as a proxy for socioeconomic status derived from zip codes. Patient satisfaction was defined as a binary variable. Univariate and multivariate binary logistic regression analyses were used to identify factors correlated with patient satisfaction.
Results:
There were 3239 unique new patients with completed surveys. Univariate analysis demonstrated decreased odds of achieving satisfaction for both the Total Score and Provider Sub-Score for each decile increase in ADI (OR 0.94; CI = 0.908-0.981; P = .003 and OR 0.94; CI = 0.91-0.98; P = .002 respectively). Multivariate analysis revealed the odds for reporting satisfaction for each decile increase in ADI were 0.96 for Total Score (CI = 0.921-0.998; P = .038) and 0.96 for Provider Sub-Score (CI = 0.92-0.993; P = .019). Patients in the most deprived quartile, compared to the least deprived, were significantly less likely to be satisfied with their care for both Total Score (OR 0.70; CI = 0.564-0.865; P = .001) and Provider Sub-Score (OR 0.69; CI = 0.558-0.852; P = .001).
Conclusions:
Increased social deprivation was an independent predictor of lower patient satisfaction in otolaryngology outpatients using the Press Ganey survey. This non-modifiable outcome should be taken into consideration when evaluating patient satisfaction scores and offers further support to the need of addressing and improving healthcare discrepancies in the field of otolaryngology.
Level of Evidence:
Level 3
Introduction
There is an increasing emphasis in medicine on improving the patient experience and patient satisfaction. These elements are often measured by patient-reported surveys. In an effort to provide more transparent access to physician ratings, many institutions make these evaluations publicly available on their websites. This has the potential to influence a patient’s likelihood to seek care (or not) from a particular provider. The data from these surveys has also been used by hospitals and healthcare institutions in determining physician compensation and advancement as well as hospital reimbursement. 1 Although these ratings have been made easier for patients to access, Ryan et al 2 demonstrated that online-based reviews of otolaryngologists at major academic institutions do not have statistically significant correlations with the Press Ganey Survey (PGS) and may not provide patients with accurate enough information to make the most informed decisions. Furthermore, various nonmodifiable factors including patient age, sex, race, and clinic visit site have been shown to impact a patient’s perception of their healthcare. 3 These factors, as well as provider gender, race and time in practice have also been shown to influence patient satisfaction in academic otolaryngology settings.4,5 Similarly, decreased socioeconomic status (SES) and longer patient wait times have also been shown to negatively affect patient satisfaction in otolaryngology and other surgical subspecialties.6,7
Social deprivation can be understood as the level of socioeconomic disadvantage that any individual experiences in a particular geographical area. Social deprivation is a factor that has largely been unexplored in otolaryngology and encompasses much more than just SES. Social determinants of health vary widely within communities and the composition of individual neighborhoods influences health outcomes, disease rates, higher utilization of healthcare services 8 and overall satisfaction with care. Social deprivation is routinely measured using the Area Deprivation Index (ADI) which uses multiple factors accounting for various aspects of SES.
Our primary study aim was to determine whether social deprivation can predict patient satisfaction from outpatient visits in an otolaryngology practice measured by the Total Score of the PGS. We also aimed to establish how social deprivation impacts patient satisfaction with their providers by specifically analyzing provider specific questions. Our alternative hypothesis was that social deprivation has no effect on patient satisfaction.
Methods
This study was approved by our institutional review board (IRB 00045048). Our institution utilizes the services of the Press Ganey Corporation in order to measure and track patient satisfaction scores in conjunction with outpatient encounters. After each clinic appointment, patients are sent an e-mail requesting the PGS to be filled out electronically and is accessible for a 30 day period. The Press Ganey Corporation compiles the responses, calculates the relevant scores and reports these to our institution. The PGS is composed of 6 subdomains: access, care provider, moving through your visit, nurse or assistant, personal issues, and overall assessment totaling 25 questions. 9 Survey questions are measured using a Likert scale from 1 (representing very poor) to 5 (representing very good). The Press Ganey Corporation converts these responses to a 0 to 100 point scale. The Total Score is computed using the mean scores from each of the subdomains and the Provider Sub-Score is calculated similarly, using the provider-specific questions.
We included patient responses from all new, adult (>18 years old) otolaryngology outpatient visits that occurred between January 2014 and December 2018 at a tertiary academic institution. Inclusion also required completion of the PGS with a Total Score and residence within Utah and immediate surrounding states (Idaho, Nevada, Arizona, New Mexico, Colorado, and Wyoming) due to the large catchment area of our institution. Only the first patient visit was included during the study period and any postoperative and return patient visits were excluded. Patients who submitted an incomplete PGS preventing the generation of a Total Score, those who speak a primary language other than English (given the survey is distributed in English), anyone who lacked a listed address, or only had a PO box listed as their address were also excluded. The Electronic Medical Record was used to identify eligible patient visits with associated PGS scores as well as to collect corresponding visit characteristics and demographic data.
Patients are also asked to self-report wait time in a free entry text box for both the waiting room and examination room. Total wait time was calculated as a sum based on patient responses for each category. Because of the high ceiling effect and rightward-skew that the PGS has demonstrated in previous studies, “satisfaction” was determined only if a patient gave a perfect total score. 10 This perfect score variable mimics the “Top Box” score reported by the Press Ganey Corporation and represents the percentage of responses with the highest possible score for a particular question or section such as “Very Good” or “Always” responses. This type of Top Box model, similar to what is used in this study, has been utilized in previous otolaryngology patient satisfaction studies.11,12
Social deprivation was determined using the 2015 Neighborhood Atlas® ADI, which provides a neighborhood-disadvantage measure of all neighborhoods in the country. 13 The ADI evaluates 17 components that impact SES, including variables such as education level, income, employment, and quality of housing for any particular Zip + 4 code. 14 ADI has been used in several surgical studies in recent years.15,16 The Zip + 4 codes cover an average of 10 to 20 homes. 8 As a result, the higher the ADI, the greater the social deprivation an individual experiences in that given area.
Categorical variables were summarized as count and percentages (%) using Minitab Statistical Software (Minitab LLC, Pennsylvania, USA). Continuous variables were summarized as mean ± standard deviation. Additionally, the median and interquartile ranges were calculated for ADI. Univariate binary logistic regression analyses were used to distinguish associations between predictor variables and patient satisfaction for both Total Score and Provider Sub-Score. We ran separate univariate binary logistic regression analyses with ADI as a continuous variable to predict odds of satisfaction and compared patients from the top (scores of 76-100) versus bottom (scores of 0-24) quartiles for ADI on a national level. In addition to ADI, other predictor variables analyzed included total wait time, patient sex and age. For predictor variables that were statistically significant in the univariate analyses, multivariate binary logistic regression models were then used to determine factors associated with patient satisfaction for Total Score and Provider Sub-Score, as defined by receiving a perfect score versus not receiving a perfect score.
Results
Following the application of inclusion and exclusion criteria, we identified a total of 3239 unique patient visits with associated PGS Total Scores during the study period. Of the included patients, the mean age was 54 ± 16.6 years and 58.8% were female. The PGS Total Score had an average of 91.9 ± 11.4 (Figure 1) and Provider Sub-Score average of 93.7 ± 13.6 (Figure 2). The observed mean ADI was 32.3 ± 19.7 (median 29; interquartile range 18-45; range 1-100; Figure 3). Additional patient demographics and visit data can be found in Table 1.
Histogram of Press Ganey total score.
Histogram of Press Ganey provider sub-score.
Histogram of ADI values. Abbreviations: ADI, area deprivation index.
Baseline Patient Characteristics.
Note. Continuous data presented as mean ± SD; categorical data presented as number of patients and (percentage).
Histogram of ADI Values. ADI = Area Deprivation Index. For the PGS Total Score, a 6% decrease in achieving satisfaction was shown with each decile increase in ADI, which was significant (odds ratio [OR] 0.94; 95% confidence interval [CI] 0.91-0.98; P = .003; Table 2). Univariate analysis also demonstrated patients in the most deprived quartile were significantly less likely to be satisfied compared to the least deprived quartile (OR 0.70; 95% CI 0.56-0.87; P = .001). Total Score was also significantly decreased for each 10 minutes increase in wait time (OR 0.59; 95% CI 0.54-0.64; P < .001). With each decile increase in patient age, a 6% increase in achieving satisfaction was notable and significant (OR 1.06; 95% CI 1.01-1.1; P = .021).
Analysis for the Press Ganey Total Score.
Abbreviations: ADI, area deprivation index; CI, confidence interval; OR, odds ratio.
Note. Bold values indicate significance.
Per 10 years of additional age.
Per additional 10%ile points.
Per additional 10 minutes.
Univariate analysis of PGS Provider Sub-Score, ADI revealed a 6% decrease in achieving satisfaction for each decile increase in ADI (OR 0.94; 95% CI 0.91-0.98; P = .002; Table 3). Patients in the most deprived quartile, similarly, were significantly less likely to report being satisfied compared to the least deprived quartile (OR 0.69; 95% CI 0.56-0.85; P = .001). Provider Sub-Score also exhibited a significant decrease in satisfaction for each 10 minutes increase in total wait time (OR 0.8; 95% CI 0.77-0.84; P < .001). Similarly to Total Score, an 8% increase in reported satisfaction was seen with each decile increase in patient age associated with Provider Sub-Score (OR 1.08; 95% CI 1.03-1.13; P = .001).
Analysis for the Press Ganey Provider Sub-score.
Abbreviations: ADI, area deprivation index; CI, confidence interval; OR, odds ratio.
Note. Bold values indicate significance.
Per 10 years of additional age.
Per additional 10%ile points.
Per additional 10 minutes.
A 4% decrease in achieving satisfaction was observed between the PGS Total Score and each decile increase in ADI using multivariate analysis, independent of patient age, sex, wait time, and insurance category, which was significant (OR 0.96; 95% CI 0.92-0.998; P = .038; Table 2). Additionally, multivariate analysis demonstrated a notable negative association between the PGS Provider Sub-Score and ADI, independent of patient age, sex, wait time, and insurance category. Specifically, a 4% decrease in achieving satisfaction was seen for each decile increase in ADI (OR 0.96; 95% CI 0.92-0.99; P = .019; Table 3).
Discussion
Our primary finding showed that greater levels of social deprivation was an independent predictor of decreased outpatient satisfaction in an otolaryngology practice. This result was independent of factors that have previously been shown to influence patient satisfaction including advancing patient age and patient wait time7,9 for the PGS Total Score and Provider Sub-Score.
The impact of social deprivation on patient satisfaction has not been extensively explored in otolaryngology. Our findings are consistent with what has been established in previous studies in other surgical subspecialties. In the orthopedic setting, one of our co-authors, Stephens et al 10 similarly demonstrated that increasing ADI was an independent predictor of lower patient satisfaction. Similarly in the pediatric population, lower Press Ganey scores were shown to be associated with poor SES. 6 On the opposite side of the age spectrum, previous work has shown that in elderly patients, average income level based on geographic location was correlated with decreased patient satisfaction. 17 SES has often been evaluated by proxy measures such as education level. Bible et al 18 described how less formal education was related to lower patient satisfaction as measured by a 25-question phone survey.
As the American healthcare system continues to evolve from traditional disease-centered care toward patient-centered care, it’s imperative to recognize how patients interact with our healthcare system. Physician awareness of how patients access, utilize, and engage in their healthcare and how specifically SES impacts this interaction, is important and can lead to improved patient experience. Cognizance of patient social deprivation, when possible, may allow providers to properly establish expectations and mitigate concerns for this patient population. Although research on this topic in otolaryngology is ongoing, several studies have shown that SES and racial/ethnic differences in prevalence, treatment, and access to healthcare exist for various otolaryngologic conditions. Race/ethnicity and SES have been shown to interact significantly and have a combined influence on health outcomes. 19 Previous work has shown that SES was the most common risk factor for otitis media in a pediatric population. 20 Similarly, social/economic and racial/ethnic disparities were associated with a higher prevalence of sleep-disordered breathing in children. 21 Additionally, Kirkham et al 22 found that following cochlear implantation, patients from lower SES populations were more likely to experience worse speech and language outcomes. Furthermore, studies have demonstrated that minority populations, as well as patients who experience increased alienation, are less likely to respond to surveys, including the PGS.12,19,23 This last point is concerning because sufficient data is essential to properly address disparities among these patient populations. Therefore, without proper data collection, disparities may likely be perpetuated.
There has been limited research regarding the factors that impact patient satisfaction in otolaryngology with the majority of available data coming from the pediatric setting. Our study findings are therefore significant in contributing to available literature in the field. A strength of our study is the large sample size and power. Although our study was conducted at an academic center, patients were seen in both general and specialty clinics at satellite clinics and the academic hospital making the findings more generalizable. Furthermore, the large, unique catchment area of our institution requires that many patients travel several hours, often from rural areas, to receive care by specialists. This considerable sacrifice of time and resources may prime certain expectations and therefore impact satisfaction. Also, exclusion of patient responses with only a PO box listed creates further potential bias and prevents proper accounting for patients in these areas, which were often found to have lower ADI scores.
To complete this study, we used a well-validated survey instrument in the PGS, allowing for comparison with other studies and institutions that utilize this metric. However, the use of survey-based practices presents its own limitations. For example, the high ceiling effect is seen in this study because most patients report high satisfaction in most categories while completing the survey. This ultimately results in a highly skewed distribution of results, which may impact sensitivity and potentially oversimplify a more complex outcome. An additional limitation was the lack of other clinical factors and demographic information such as chief complaint, disease severity, patient income, race, among others that were not available for this study given we only analyzed PGS scores. Furthermore, survey-based studies have an inherent selection bias and low response rates. Prior studies based on PGS results conducted at our institution have reported response rates varying from 8.9% to 16.5%.24,25 Also, because the survey is distributed via email, patient responses from certain socioeconomic groups may be further precluded from participating due to unreliable computer or internet access. Additionally, patients are allowed several weeks to complete the survey which may likely introduce recall bias. Finally, we only included survey results that patients completed in English which further limits the diversity of the population under study. Despite the shortcomings of patient surveys, this method continues to be widely used by hospital administrators and policy-makers to drive decision-making, influence hospital and physician reimbursement and advancement, and serve as a proxy for healthcare “quality.” However, because of the continued, widespread use of this method, otolaryngologists are directly impacted by the results of these surveys.
Although there are several previously mentioned non-modifiable factors that impact patient satisfaction, there are reassuring data suggesting there are various interventions providers can implement to increase patient satisfaction. One study showed that the patient-physician relationship and several physician-dependent characteristics are significant indicators of positive patient satisfaction and a higher “likelihood of recommending a practice.” 26 Additional studies have shown that eye contact, communication and attention had more influence than wait time or time spent with the provider. 27 Increased satisfaction has also been shown to be the result of even simple introductions. 28 Although increased wait times have consistently been shown to negatively impact patient satisfaction scores,24,29,30 data has shown that more time discussing therapeutic options and health education versus history taking dulled the negative influence of increased wait time. 31 Providers may be able to improve patients’ negative experiences with increased wait time by clearly addressing the concern before the visit or before the patient arrives at the clinic via email, phone call or letter reminding them providers will do their best to stay on schedule but that each patient has the right to have their provider’s undivided attention and to have all of questions and concerns addressed.
Further investigation is necessary to gain a clearer understanding of the factors that impact satisfaction in patients who experience greater social deprivation. Future work may include examining the impact of specific diagnoses, available treatment options, and evaluation of specific subdomains of the PGS and how they impact patient satisfaction.
Conclusion
Our study found that increased social deprivation experienced by patients is associated with lower Press Ganey patient satisfaction scores in otolaryngology patients. Awareness of this nonmodifiable factor can help improve provider-patient interactions, guide necessary institutional adjustments and policy decisions. With increased efforts to minimize healthcare discrepancies, especially in more vulnerable populations, future directions for this work specifically include investigating which elements of ADI (income, poverty and education level, etc.) have the greatest influence on patient satisfaction. The contribution of these elements to patient perception of their healthcare will likely provide a clearer understanding of the complex interaction of social determinants of health and their impact on patient satisfaction.
Footnotes
Authors’ Note
2021 Combined Sections Virtual Meeting, January 29 and 30, 2021
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: NIH/NIA - Research Funding (R21 AG067403-01A1) - Contracted Researcher Med-El - Other Financial Benefit - Surgical Advisory Board Cochlear Corp and Advanced Bionics - Institutional Research Funding - Contracted Researcher Harmonson Law Firm, Sigfried and Jenson Law Firm, Burbidge and White Law Firm, US Attorney General - Consulting Fee - Medical Legal Consulting
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
