Information technology is one of the most significant tools currently available to improve health care quality and productivity. When it comes to IT, however, health care lags behind almost every other industry . Although the health care industry does not lack for technology, major expenditures are limited to profit-making sectors such as surgery and treatment . Other data-driven industries such as insurance or financial services budget more than 10 percent of their finances for IT, whereas health care puts only two to three percent of its budget into IT. Along the same lines, private industry spends an average of about $7,000 per worker on IT hardware, software, and services, with banking approaching $15,000 a worker; health care averages about $3,000 annually per worker for IT .
This contrast in IT investment is important to note as it comes at a time when we are realizing that patients see real benefits from increased investment in technology. In fact, examinations into how users feel about technological applications in health care indicate generally high levels of satisfaction as both patients and providers appreciate the convenience and efficiency that comes with new technologies. Videoconferencing consultations between specialists and patients improves patient satisfaction by reducing burdens associated with travel and scheduling while promoting a feeling of involvement in the physical examination itself . Store-and-forward consultations in record-driven fields such as dermatology lead to positive reviews from all parties involved (patients, referring physicians, and specialists) . Even basic Internet searching for health information makes patients more likely to pose more informed questions to their doctors and to follow prescribed treatment regimens .
The possibility of increased patient satisfaction is an immense consideration for many institutions, especially for larger health care facilities that tend to have lower satisfaction ratings. The importance of measuring, improving, (and achieving) patient satisfaction has not always been a top priority for health care institutions. Whereas in the past, patient satisfaction was not necessarily seen as related to institutional success, today it is recognized that quality of care, customer satisfaction, and financial outcomes are all interrelated. Higher patient satisfaction can help retain a customer base, increase physician loyalty, bring more patients, raise employee satisfaction and retention, cut costs, and reduce length of stay .
On the other hand, patient dissatisfaction is associated with a significant decrease in revenue for the health care organization . In light of this situation, Press Ganey, the largest health care satisfaction measurement firm in the U.S., aims to provide one of the largest pools of comparative data in the nation . This allows hospitals to not only measure and compare patient satisfaction scores, but also to identify areas that are lacking and improve upon them.
Given the ever-growing standards and recognition of satisfaction research, the current environment is ripe for new and more detailed investigations into the antecedents of patient satisfaction. More specifically, a pressing question remains to be explored in the realm of health care satisfaction research: Are hospitals that invest more heavily in health IT more likely to achieve higher patient satisfaction? A growing number of examples seem to point strongly in this direction.
Overall, evidence is continually mounting that there is something special about health care organizations that invest in IT (hospitals that are "wired"). This project seeks to further investigate the relationship between investment in health IT and patient satisfaction in the hospital context through analysis of patient satisfaction data for hospitals with validated Press Ganey satisfaction survey  data that was included in the 2005 Hospitals & Health Networks annual list of the "100 most wired hospitals and health systems." Specifically, we sought to test the following hypotheses:
H1: Patients from the Most Wired Hospitals would report higher satisfaction scores regarding the overall experience in the hospital.
H2: Patients from the Most Wired Hospitals would report higher satisfaction regarding specific aspects of their hospital experience, including: H2a: Their admission process and experience; H2b: Their experiences with hospital-based nurses; H2c: Items related to tests and treatments within the hospital; H2d: Their experience with physicians during their hospitalization; H2e: Their discharge experience; H2f: Personal issues such as sensitivity and pain control.
H3: Status as a most wired hospital would more accurately predict higher patient satisfaction than specific demographic characteristics of the hospitals such as number of patient beds, case mix, number of critical days, payer mix, community size, total number of full-time equivalents, and services provided.
This work is based in part on the revisitation of a theoretical proposition developed almost 30 years ago. Ben-Sira offered a revised model of social interaction regarding relationships in the health context whereby the mode of an interaction may equal or supercede the actual content . Specifically, Ben-Sira suggested that a client's satisfaction may indeed be a consequence of the mode of a professional's response. He argues that an emotional involvement in treatment, a lack of detailed medical knowledge, and an ability to connect treatments to healing will lead patients to judge medical providers on the basis of the physician's behavior toward the patient . Thus, high levels of socio-emotional behavior on the part of the provider can promote patient satisfaction, self-disclosure, and trust . We seek to update this theory to allow mode to encompass the use of information technologies to facilitate communication-related activities in the health setting. With this in mind, we sought to test whether most wired hospital patients report higher levels of satisfaction related to the inpatient experience.
During the time of data collection for this study (inpatient surveys received between Jan. 1, 2004 and Sept. 1, 2005), Press Ganey collected patient satisfaction surveys for 1,382 hospitals in the U.S. Specifically, standardized questionnaires were mailed shortly after discharge to patients hospitalized in an acute care hospital. This procedure yielded over three million survey responses from patients discharged from 1,382 hospitals nationwide.
In 2005, Hospitals & Health Networks magazine published its list of the 100 most wired hospitals in the U.S.1 The annual Hospitals & Health Networks' Most Wired Survey and Benchmarking Study asks hospitals via an eight-page survey to self-report on their use of IT in five key areas: business processes, customer service, safety and quality, work force, and public health and safety. (See Table 1 for evaluation criteria.) Hospitals & Health Networks then reviews the results of the proprietary survey and evaluates the hospitals to determine which hospitals have the highest performance. Top-scoring hospitals are published in the Top 100 list (actual scores and scoring criteria are not included with the published list); 42 of these Top 100 most wired hospitals are Press Ganey clients. Therefore, patient satisfaction scores were compared between the two groups, the 42 hospitals included on the most wired list and the 1,340 not included on the list, referred to as "other" in this article.
The Press Ganey Inpatient Survey was first developed in 1987 and has undergone rigorous validation testing. The conceptual model behind the ratings is real-world-based in that it derives from typical experiences a patient may actually encounter during a hospital stay. Events that occur, (admission, meals, tests or treatments, discharge); personnel encounters (nurses, physicians, and technical staff); the physical surroundings (room and hospital) and the interpersonal aspects of the stay are seen as important contributors to the patient's total experience. They are also believed to be reflections of the quality of the medical care delivered and received.
The survey includes 49 standard questions asked by all organizations that are organized into 10 sections or sub-scales including: Admission, Room, Diet and Meals, Nursing, Tests and Treatment, Visitors and Family, Physician, Discharge, Personal Issues, and Overall Assessment. Within each section, respondents are asked to evaluate a set of attributes that relate to that conceptual area. Responses are coded on a Likert-type scale from 15 (1=Very Poor, 2=Poor, 3=Fair, 4=Good, and 5=Very Good). Scores are linearly transformed to a 0100 scale for ease of interpretation. Chronbach alphas indicating the reliability of the subscales range between 0.78 and 0.95. Reliability for the entire instrument is 0.95. Factor analysis supports the construct areas measured by each of the subscales. More information about the psychometric properties of the Press Ganey Inpatient Survey can be obtained from the authors.
Independent samples t-tests were performed to compare the satisfaction scores between the most wired hospitals and the rest. Patient-level satisfaction survey responses were first aggregated at the facility level to create the two distributions for comparison. The 42 hospital-level mean scores (comprising those hospitals from the Most Wired list) were used to create the mean for `Most Wired Hospitals'. Similarly, the 1,340 hospital-level mean scores for facilities not on the most wired list were use to create the mean for the `Other Hospitals' Group.
The standard parametric t-test requires the assumption that variances are homogeneous between the two groups being compared. When this assumption is met, a pooled variance estimate is used to calculate t. When variances are not equal, the t statistic is calculated using separate variance estimates. Additionally, the Welch correction for degrees of freedom is applied when using the t designed for unequal variances. In each comparison reported, Levene's test for the equality of variance was performed first to determine if the pooled variance or separate variance results should be reported. See Tables 2 and 3 for notation as to which results assumed equal variances per the results of the Levene's test.
Data from the surveys revealed there were significant differences in satisfaction-related issues for clients at the most wired hospitals. These differences were demonstrated for overall satisfaction, as well as for specific aspects of the hospital experience.
First, the data demonstrated consistent support for the first hypothesis that patients receiving care at the most wired hospitals would report higher global satisfaction scores. The measure contains several indicators of global satisfaction including a composite overall performance score, a subscale addressing the patients overall assessment of the care experience, as well as four individual questions geared to global outcome measures. In each case, the most wired hospitals fared better (see Table 2). Notably, wired hospitals scored 0.76 points higher on the composite overall performance score (t=2.116, df=45.72, p=0.040).
A similar pattern of more positive performance was found when looking at the overall assessment subscale of the measurement tool which was 1.24 points higher for the most wired group (t=2.242, df=1380, p=0.025). The overall assessment subscale that appears at the conclusion of the questionnaire asks the respondent to consider their experience from a broad view and evaluate the overall cheerfulness of the hospital, how well staff worked together to care for the patient, likelihood of recommending the hospital to others and the overall rating of care given at the hospital. Three of the four individual items that address global evaluations of the hospital were also significantly higher for the most wired group. The most wired hospital group was higher in patient evaluations of overall rating of care (+1.24 (t=2.340, df=1380, p=0.019)); likelihood of recommending (+1.96 (t=3.049, df=1380, p=0.002); and coordination of care (+0.81 (t=2.183, df=46.06, p=0.034)). There was no significant difference between the two groups in respect to overall cheerfulness of the hospital.
Statistical comparisons regarding specific aspects of patients' hospital experiences were also conducted (see Table 3). The hypothesis that patients from the most wired hospitals would rate their satisfaction as higher with the admission process and experience was supported. At the subscale level, wired hospitals scored 1.52 points higher (t=3.524, df=45, p=0.001), which was a statistically significant difference. At the individual item level, all three individual items were significantly different. Specifically, wired hospital patients reported higher scores for the speed of admission (+1.98 (t=3.969, df=44.89, p=0.001)), courtesy of admission staff (+1.05 (t=3.244, df=45.3, p=0.002)), and the pre-admission process (+1.43 (t=2.509, df=1375, p=0.012)). However, the hypothesis that patients at the wired hospitals would view the experience of hospital discharge in a manner that was statistically more favorable was not supported at the subscale level or for any of the four individual items within this subscale.
In regard to satisfaction with health providers, patients in the most wired hospitals did report statistically reliably higher satisfaction levels with physicians in general. At the subscale level, the most wired hospitals scored 0.80 points higher (t=2.420, df=45.86, p=0.02) than the other group. At the individual item level, three out of five items had significant differences between the two groups. Specifically, wired hospitals' patients were more satisfied with physician's concern with their questions/worries (+0.75 (t=2.158, df=45.89, p=0.036)); friendliness/courtesy (+0.74 (t=2.663, df=46.29, p=0.011)); and skill of physician (+1.27 (t=3.484, df=1379, p=0.001)). Yet, patients from the most wired hospitals did not report statistically higher satisfaction scores regarding their experiences with hospital-based nurses in general at the subscale level. However, at the item level, patients at the most wired hospitals did report higher satisfaction related to nurses for two of the six items. Most wired hospital patients reported significantly higher satisfaction with nurses' attitude to patient requests (+0.65 (t=2.178, df=48.94, p=0.034)) and skill of nurses (+0.57 (t=2.128, df=47.8, p=0.039)).
The hypothesis that patients from the most wired hospitals would view care more favorably related to the personal issues such as sensitivity and pain control did prove to be supported. At the subscale level, wired hospitals scored 0.89 points higher than the other group (t=2.431, df=46.37, p=0.019). At the individual item level, two out of six items were significantly different, with wired hospitals scoring higher in both cases: staff sensitivity to patient's inconvenience (+1.00 (t=2.469, df=45.91, p=0.017)), and pain control (+0.83 (t=2.981, df=46.32, p=0.005)).
Most wired hospital patients did not report higher satisfaction with tests and treatments at the subscale level. At the item level regarding tests and treatments, only one of the seven items was significantly different in favor of the wired hospitals. Specifically, most wired hospitals' patients were more satisfied with the explanations received regarding medical tests and treatments (+0.83 (t=2.596, df=45.28, p=0.013)).
Finally, status as a most wired hospital proved to be an important variable in more accurately predicting higher patient satisfaction than specific demographic characteristics of hospitals. In general, smaller hospitals tend to have higher patient satisfaction scores than the larger hospitals in the Press Ganey Inpatient Database. The Pearson Correlation between hospital bed size and the overall patient satisfaction score is -0.321 (p<0.000). Average bed size of the most wired hospital group in our analysis is 543, and the average bed size for the other group is 260. However, patients treated at the most wired hospital group were significantly more satisfied than those treated at other hospitals (t=2.116, df=45.72, p=0.04).
Hospitals with higher Case Mix Indices (that is, having sicker patients) tend to score lower in patient satisfaction than do hospitals with lower Case Mix Indices. Overall, the Case Mix Index for the most wired hospital group was 359 versus 144 for the other group (with a higher case mix index indicating a sicker population of patients). Yet, patients from the most wired hospitals reported significantly higher levels of satisfaction.
Teaching hospitals (Council of Teaching Hospitals (COTH) members) have lower overall patient satisfaction than the non-members (82.3 vs. 84.3) in the Press Ganey Inpatient Database. In this study, most wired hospitals were more likely to be COTH members (44.8% vs. 13.5%), but nevertheless demonstrated higher patient satisfaction results. A similar relationship holds for non-COTH member teaching hospitals: non-members score significantly higher in patient satisfaction (84.5 vs. 82.9). In this study, most wired hospitals were more likely to be teaching hospitals (52% vs. 27.5%), yet these most wired hospitals demonstrated higher levels of patient satisfaction.
Finally, when examining a number of miscellaneous hospital demographics (the hospitals' community type and size, UHC membership, presence of medical residents in the hospital, types of services provided) we again found that most wired hospital status is a better predictor of higher patient satisfaction than any other hospital demographic variable.
In summary, analysis from this study found that patients from the most wired hospital group report higher levels of overall satisfaction than do patients from the other group of hospitals. Patients from the most wired hospitals also reported higher satisfaction related to the admission process, their experiences with physicians, and personal issues such as sensitivity and pain. However, there was no difference in general satisfaction scores between the two groups for experiences with nurses, the discharge process and tests and treatments (though there were a handful of individual items for these three areas where most wired patients reported higher satisfaction with no items where the other group reported statistically higher satisfaction). Finally, higher satisfaction scores were associated with most wired hospital status more so than for any specific demographic variable tested.
The results from this study are important and thought provoking for a variety of reasons. Health care organizations, often non-profit, are faced with challenging resource allocation. Administrators must make difficult decisions regarding investment in IT in lieu of other critical resources such as personnel or capital equipment. Often, administrators are pressured to allocate resources in ways that demonstrate immediate, short-term benefits. Yet, the results of this study suggest that among the longer-term benefits of IT investment in hospitals may actually be issues related to patient satisfaction. This data suggests IT enhancements don't just affect the way health care professionals work, they also affect the way patients receive and perceive their care. As we move toward a new paradigm of health delivery necessitated by public and private desires to contain health costs, we are moving to a world where many patients will be more knowledgeable about managing health care, better informed about the benefits, risks, costs and alternatives for treatments, more technologically savvy, and more engaged in decision making with providers. As a result, hospitals will need strong IT infrastructure and tools to meet the increasing expectations of these more sophisticated consumers.
Patient satisfaction is a phenomenon determined by expectations and values. These values are important antecedents for patient satisfaction as we ask them to evaluate their care based on what they want and expect from health care providers. In general, patient satisfaction is an evaluative summary of whether a patient likes or dislikes health care services. Raftopoulos explained that patients evaluate care as functions of cognitive (beliefs, expectations and perceptions), affect (feelings) and behavioral intentions (aspirations and expected responses to care provided) . This means that patient satisfaction is a dynamic process determined by the way a patient thinks, observes, and acts. Therefore, patient satisfaction is an attitude based on the way a patient conceives the phenomenon of a health care experience while in the hospital. Patients live in a world where IT and its associated services and benefits abound in almost every sector ranging from banking, entertainment, and communicating with friends and colleagues. It should not be surprising that many of these patients express greater satisfaction in hospitals that also employ IT in significant ways.
This study represents a simple first step to determine if there is merit in further assessing IT as an antecedent for patient satisfaction. The study is not without its limitations. For example, we do not know the level of IT investment in the hospitals included within the other category. There may well be additional explanatory variables that better explain these differences in patient satisfaction. For example, perhaps most wired hospitals inherently possess an innovative and radical culture that permeates all levels of care. Yet, the data in this study repeatedly pointed to instances where patients from the most wired hospitals were more satisfied even in those cases where demographic variables such as hospital size always seem to outweigh other impacts. Also worth noting is that even though there were variables with non-significant results in satisfaction ratings, there was not one single statistical test where the non-wired hospital patients expressed higher satisfaction than those in the most wired hospitals.
We were fortunate to have a large amount of standardized satisfaction data to explore for this study. The existence of a centralized data bank of validated satisfaction results through Press Ganey permits unique and innovative comparisons across hospitals. This study suggests IT may be an important antecedent for patient satisfaction. It is a first step that validates the need for significant future study to better explain this potential. Future work needs to differentiate between the myriad of IT solutions to clarify if some play a more important role in leading to enhanced satisfaction.
Patient satisfaction has emerged as a vital indicator of the quality of medical care, as well as a significant determinant in decisions regarding future health providers. IT investment may well emerge as a strategy to better meet the needs of an evolving hospital patient demographic, ultimately resulting in a hospital's ability to ensure its competitive position.
1. Ben-Sira, Z. The function of the professional's affective behavior in client satisfaction: A revised approach to social interaction theory. Journal of Health and Social Behavior 17 (Mar. 1976), 311.
2. Ben-Sira, Z. Affective and instrumental components in the physician-patient relationship: An additional dimension of interaction theory. Journal of Health and Social Behavior 21 (June 1980), 170180.
4. Harris Interactive. The increasing impact of eHealth on consumer behavior. Health Care News 1, 21 (June 26, 2001), 19; www.harrisinteractive.com/news/newsletters/healthnews/HI_HealthCareNews2001Vol1_iss21.pdf.
9. Reese, B. Statement of the National Alliance for Health Information Technology. National Committee on Vital and Health Statistics Subcommittee on Standards and Security. Sentara Healthcare of Norfolk, VA, 2002.
10. Roberts C. and Aruguete M. Task and socioemotional behaviors of physicians: A test of reciprocity and social interaction theories in analogue physician-patient encounters. Social Science in Medicine 50, 3 (Feb. 2000), 309315.
©2008 ACM 0001-0782/08/0400 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2008 ACM, Inc.