DETERMINANTS OF PATIENT SATISFACTION

Document Sample
DETERMINANTS OF PATIENT SATISFACTION Powered By Docstoc
					Proceedings of the 3rd INFORMS Workshop on Data Mining and Health Informatics (DM-HI 2008)
J. Li, D. Aleman, R. Sikora, eds.


                 DETERMINANTS OF PATIENT SATISFACTION
                                     Sahil Ashok Hebbar
                      Department of Health Management and Informatics
                                    University of Missouri
                                       Columbia, MO

                                    Kalyan S. Pasupathy
                      Department of Health Management and Inforamtics
                                    University of Missouri
                                       Columbia, MO
                                 PasupathyK@missouri.edu

                                       Mary M. Williamson
                                      Patient Care Services
                                 University of Missouri Healthcare
                                          Columbia, MO

                                              Abstract
The focus on patient satisfaction has increased considerably in the past decade. Patient
satisfaction surveys are valuable tools in monitoring and improving the services currently
provided by the hospital. With competition in the health care industry increasing and data
reporting becoming more transparent, patient satisfaction data will become an important tool for
attracting patients. The purpose of this study was to identify predictors of quality improvement
and patient satisfaction. Patients admitted to the university hospital over two years were surveyed
using a close-ended questionnaire. The data were analyzed using SPSS and statistical data
mining was done using Chi-square Automatic Interaction Detector (CHAID). Predictor variables
significantly impacting the patients’ perception regarding overall rating of care of the hospital
and likelihood of recommending the hospital were ascertained. The results indicate that
coordination between staff members and overall cheerfulness of the hospital were the two most
important variables. Using the predictor variables identified in this article, managers could
design specific interventions to enhance patient satisfaction.

Keywords: Patient satisfaction, overall rating of care, likelihood of recommending, CHAID

Introduction
The focus on total quality management and value of the care provided to patients has increased
considerably in the past decade. Today, quality of care and patient satisfaction with care have
become critical issues in the health care industry, drawing the attention of all the stakeholders –
regulatory bodies, policy makers, purchasers, providers and patients. Patient satisfaction is
considered an important outcome in itself [1]. The validity of all other measures of quality
depends on whether the patient was satisfied with the care provided [1]. Patient satisfaction is
associated with increased profitability, increased employee satisfaction and decreased turnover,
and decreased likelihood of malpractice suits [2]. Patient satisfaction surveys are being used to
                                                   1
involve and empower patients with the assumption that it leads to better adherence to treatment
guidelines, and hence to improved patient outcomes [3]. They are valuable tools in monitoring
and improving services currently being provided. These factors have led to a spurt in research
targeting patient opinion and satisfaction vis-à-vis the care given. Ervin (2006) detected 1,778
articles related to patient satisfaction between 1975 and 2004 [4].

There have been studies that have linked patient satisfaction with technical quality of care [2].
However, more research needs to be undertaken before patient satisfaction is accepted as an
indicator for technical quality, since patients have been found to be poor at evaluating the
technical quality [5, 6]. This paper focuses on the association between patient satisfaction and the
perceptional quality of care, which relates to how the patients perceive the services being
provided. The results of this particular study were intended to be used to identify areas that
needed improvement, learn what is important to patients admitted to the hospital, determine if
they were satisfied with the service provided, and share compliments with the employees. The
paper discusses predictors of patient care services that significantly affect likelihood of
recommending the hospital and the overall rating of care. It talks specifically on certain aspects
of care that health care providers and hospitals should concentrate on. Future interventions to
improve patient satisfaction and quality of care should target these areas.

Methodology
More than 4000 patients admitted to the hospital over a period of two years were surveyed using
a close-ended questionnaire with 71 questions. A scale of 5 was interpreted as “very good” and a
scale of 1 was “very poor.” The study results were analyzed using SPSS. Statistical data mining
was done using Chi-squared Automatic Interaction Detector (CHAID). CHAID is a type of
decision tree method originally proposed by Kass (1980). It is an exploratory algorithm that is
used to study the relationship between a dependent variable and several predictor variables.
Various independent variables are evaluated to see if splitting the sample based on these
predictors leads to a statistically significant discrimination in the dependent measure. CHAID
goes over all possible variable combinations in the data and splits them into branches. Then, for
each of the branches formed, we see if the branches could be further significantly split by
another predictor variable. The CHAID algorithm is particularly well suited for the analysis of
large datasets, because the algorithm will often effectively yield many multi-way frequency
tables [8]. In this study, CHAID was used to determine which of the predictor variables
significantly affect the overall rating of care as well as the likelihood of recommending the
hospital. CHAID has been used in the past to evaluate patient satisfaction with home care
services [2]. However, this is the first time that data mining techniques like CHAID have been
used to ascertain predictors of patient satisfaction with quality of care provided in hospitals.

Findings
A total of 4380 patients were surveyed over a period of two years. Almost 52% patients surveyed
were female. The sample included patients over a wide range of ages from 1 to 100, with an
average age of 56 years. 52% of the patients had an emergency admission and 58% patients said
that the admission was unexpected. A high percentage of people are likely to recommend the
hospital. 53% patients said that the likelihood of them recommending the hospital to others is
very good, and 30% said it is good. In contrast, only 5% patients said the likelihood of
recommending is very poor, and 3% said it is poor. Similarly, a majority of the patients said that
                                                2
the overall rating of care provided is very good. 52% patients said that the overall rating of care
given at hospital is very good and 33% said that it is good. On the other hand, only 3% felt the
rating is very poor and 3% said it is poor. Both, the likelihood of recommending the hospital as
well as the overall rating of care, are affected significantly by similar predictor variables. These
predictor variables are: 1) Staff working well together, 2) Overall cheerfulness of the hospital, 3)
Response to concerns/complains, 4) Staff efforts to include patient in decisions regarding
treatment, 5) Staff sensitivity to inconvenience that health problems and hospitalization can
cause, and 6) Nurses keeping the patients informed. In addition, overall rating of the care is also
affected by the attention paid by nurses to special/personal needs of the patient.

Staff working well together: Staff working well together is the most important predictor of
patients likely to recommend the hospital. This is evident from the fact that, when staff working
well together is rated very good, it is very likely (90%) that the patients would recommend the
hospital. In addition to staff working well together, if the overall cheerfulness of the hospital is
also rated very good, then the likelihood increases to 95%. Other important predictors are rating
of resident, intern and house staff, noise level in and around the room, and the extent patients felt
ready for discharge. These predictors increase the likelihood further to 99%. In case the overall
cheerfulness is not very good, the room temperature is the most important predictor. Staff
working well together comprises of various aspects. Nearly 50% of the patients surveyed felt that
the coordination between staff members is very good, whereas only 2% felt that it is very poor.
Nurse keeping the patients informed is the most significant variable affecting the patients’
perception regarding coordination between the staff. The other variables affecting staff
coordination are physician communicating with each other about patient care and treatment,
instruction regarding home care, physician keeping the patient informed, and physician
communicating with the referring or family physician. If the nurses are very good in keeping the
patients informed, then the coordination of care is very good (81% patients). In addition to
nurses keeping the patient informed, if the communication between the physicians is also very
good, then the rating of coordination of care increases to 92%. Other predictors include
instruction given to patients by various professionals about home care, and the physician
communication with referring or family physician. These additional predictor variables increase
the rating of coordination of care to 96%. On the negative side, if the nurses are very poor in
keeping the patient informed, then 62% patients are likely to say that the coordination between
staff members is either very poor or poor. This is consistent with the findings in other studies in
this field [9, 10].

Overall cheerfulness of the hospital: The overall cheerfulness of the hospital plays a vital role in
influencing the likelihood of patients recommending the hospital to others. If the overall
cheerfulness is very good, then a large percentage of patients (97%) said that they would
recommend the hospital. Other significant predictor variables include rating of resident or intern,
staff working well together, noise level in and around the room, and extent that patients felt
ready for discharge. If these predictors are rated very good, the likelihood of recommending
increases to 99%. In contrast, if the overall cheerfulness is very poor, then not a single patient is
likely to recommend the hospital. Forty five percent of the respondents said that the overall
cheerfulness of the hospital is very good, whereas only 1% patients said that the cheerfulness is
very poor. Overall cheerfulness of hospital also has several predictors. Friendliness/courtesy of
the nurses is the most important variable that influences patient perception regarding overall
                                                 3
cheerfulness of the hospital. If the friendliness/courtesy of the nurses is very good, then 67%
patients are likely to say that the overall cheerfulness of hospital is very good. Additionally, if
the staff attitude towards visitors, pleasantness of room décor as well as the courtesy of the
person starting the IV is very good, then this likelihood increases to 97%. On the other hand, if
the friendliness/courtesy of the nurses is very poor, then approximately 31% patients said that the
cheerfulness of the hospital is very poor. In addition, if the friendliness/courtesy of the
physicians is very poor, poor, or fair, then this percentage increases to 47%.

Response to concerns/complains: Response to concerns and complains of the patient is the third
most significant predictor of likelihood of recommending the hospital. If the response to
concerns/complains is very good, then 88% of patients said that they are likely to recommend the
hospital. Other important predictors are the overall cheerfulness of the hospital, rating of the
resident, intern and house staff, noise level in and around the room, and the accommodation and
comfort provided to visitors. If these predictors are rated very good, then 100% patients are
likely to recommend the hospital. On the negative side, if the response to concerns/complains is
very poor, then majority of the patients (52%) said that the likelihood of them recommending the
hospital is very poor. Moreover, if the coordination between staff members is also very poor,
then this likelihood increases to 88%.

Staff efforts to include patient in decisions regarding treatment: Staff efforts to include patient in
decisions regarding treatment is the fourth predictor variable significantly influencing patient
likelihood of recommending the hospital. If the staff did a good job of including patients in
decisions regarding treatment, then a high percentage (88%) said the likelihood of
recommending is very good. Furthermore, if they felt that the overall cheerfulness is very good,
this likelihood increases to 95%. The other predictors affecting the likelihood of recommending
are communication of the physicians with each other, rating of the resident, intern or house staff,
and the noise level in and around the room. If these predictors are rated very good, then the
likelihood increases further to 99%. In contrast, if the staff is very poor in including patients in
decisions regarding treatment, then 56% patients said that the likelihood of recommending the
hospital is very poor. Additionally, if the coordination between staff members is also very poor,
then this percentage increases drastically from 56% to 90%.

Staff sensitivity to inconvenience that health problems and hospitalization can cause: Staff
sensitivity to inconvenience that health problems can cause is the fifth most important predictor
of likelihood of recommending the hospital. If the staff sensitivity to inconvenience is very good,
then a significant majority (88%) of the patients are likely to recommend the hospital. Other
predictor variables of significance are overall cheerfulness of hospital, communication of the
physicians with each other, staff attitude towards visitors, and the room temperature. If these
variables are rated very good, then 100% patients said that they are likely to recommend the
hospital. On the other hand, if the staff sensitivity to inconvenience is very poor then 74%
patients said that the likelihood of recommending is poor. Furthermore, if the coordination
between staff members is either poor or very poor, the percentage increases to 98%.

Nurses keeping the patients informed: Finally, nurse keeping the patients informed is the sixth
predictor that significantly affects the likelihood of recommending. If the nurses keep the
patients very well informed, then 81% patients are likely to recommend the hospital. In addition
                                                4
to that, if the rating of the resident, intern or house staff is very good, then the likelihood of
recommending increases to 91%. Furthermore, if the rating of other predictor variables like room
cleanliness, instructions for home care, and the skill of the ICU/CCU nurses is very good, then
this likelihood increased to 100%. On the negative side, if the nurses are very poor in keeping the
patients informed, then 57% patients said that the likelihood of them recommending the hospital
is very poor. Moreover, if the rating of the resident, intern or house staff is also very poor, this
likelihood increases to 92%. The overall rating of patient care is also affected significantly by the
six predictors mentioned above, but with different percentages. Additionally, it is also
significantly influenced by the rating of the nurse attention to special needs of the patient.

Nurse attention to special/personal needs of the patient: This is a predictor variable that does not
influence the likelihood of recommending the hospital, but significantly affects the overall rating
of care. If the attention paid to special/personal needs is very good, then a high percentage (79%)
of patients said that the overall rating is very good. The other important predictors are rating of
resident/intern, instructions for home care, and cleanliness of the room. If these variables are
rated very good, then the percentage increases from 79% to 98%. On the other hand, if the
attention paid to special/personal needs is very poor, then 46% of the patients said that the
overall rating of care is very poor. Moreover, if the staff are either poor or very poor in
addressing the emotional needs of the patients, and physicians are very poor in keeping the
patients informed, nearly 90% of the patients said that the rating of care is very poor.

Implications
Since nurses are the staffs who spend the most time with inpatients, they are the professionals
who can most influence patient satisfaction. For example, a nurse keeping the patient informed is
more likely to satisfy patients to the fullest extent and increase likelihood of recommendation.
Providing timely information is important for the patient and family, and communicates to the
patient that staff members “know what they are doing”. Nurses can foster positive "transitions"
when patients are being transferred from one nursing unit to another (for example from ICU to
general care), as well as when patients are sent from the nursing units to ancillary departments
for testing or treatment. If the nursing staff cannot give patients information regarding when they
will receive visits from certain specialists or be transported for diagnostics, then the patients
perceive this as a sign of the hospital not being well coordinated. For the nurses to provide the
best care, keep patients informed, and facilitate the coordination of care between all disciplines,
they must develop positive working relationships with other ancillary departments, clinical
disciplines, and with each other. Allied health professionals and other specialists do plan their
rounds, but each group often uses a different scheduling system located in different physical
areas of the hospital. Often these different systems do not communicate with one another. Hence,
there is not a way for nurses in a specific patient care unit to obtain the scheduling
information. A process that would enable nurse managers to obtain patient rounding schedules
from various professions and specialists would enable them to better coordinate care for patients,
and assist patients in planning their activities. Developing a quality improvement team that focus
specifically on improving communication and coordination of care between professionals,
nursing units, and patients and families would result in greatly improved patient satisfaction.
Another target intervention could be to train physicians to communicate effectively with their
patients. If the physicians keep the patients well informed, the overall rating of care and the
likelihood of patient recommending the hospital is very good. Past studies have indicated that
                                                   5
effective physician communication is associated with increased patient satisfaction [11, 12]. If
the patients feel that the physician is friendly and courteous and communicates well with other
physicians, then it could have many positive long-term effects on physician-patient relationship
and patient satisfaction.

One of the limitations of our study is that the evaluation of patient satisfaction took place through
a survey two weeks after the patient was discharged. However, any evaluation on the perceived
quality of care needs to be undertaken immediately on discharge from the hospital. Past studies
examining impact of physician behavior on patient satisfaction concluded that if the survey is
conducted after two weeks, then the results could be influenced by the health outcomes [5].
Hence, the results of our study could have been impacted because of the distribution of
questionnaire after discharge. Also, our study focuses more on the perceived quality of care, and
not on the technical quality of care. More studies need to be undertaken to see if the patient
satisfaction with quality of care is an indicator of the technical quality of care.

Reference and Citations
   1) Donabedian, A., 1966, “Evaluating the quality of medical care”, Milbank Mem. Fund. Q.,
       44, 166-206.
   2) Mylod, D. E., and Kaldenberg, D. O., 2000, “Data Mining Techniques for Patient
       Satisfaction Data in Home Care Setting”, Home Health Care Management & Practice,
       12(18), 17-29.
   3) Grol, R., 2001, “Improving the Quality of Medical Care: Building Bridges Among
       Professional Pride, Payer Profit, and Patient Satisfaction”, JAMA, 284, 2578-2585.
   4) Ervin, N. E., 2006, “Does Patient Satisfaction Contribute to Nursing Care Quality”, The
       Journal of Nursing Administration, 36(3), 126-130.
   5) Jackson, J. L., Chamberlin, J., and Kroenke, K., 2001, “Predictors of patient satisfaction”,
       Social Science and Medicine, 52, 609-620.
   6) Haas-Wilson, D., 1994, “The relationships between the dimensions of health care quality
       and price: the case of eye care”, Med. Care, 32, 175-182.
   7) Kass, G. V., 1980, “An exploratory technique for investigating large quantities of
       categorical data”, Applied Statistics, 29(2), 119-127.
   8) Pasupathy, K. S., and Triantis, K. P., 2007, “A Framework to Evaluate Service
       Operations: Dynamic Service-Profit Chain”, The Quality Management Journal, 14(3), 36-
       49.
   9) McColl, E., Thomas, L., and Bond, S., 1996, “A study to determine patient satisfaction
       with nursing care”, Nurs Stand., 10(52), 34-38.
   10) Lewin, S. A., Skea, Z. C., Entwistle, V., Zwarenstein, M., and Dick, J., 2001,
       “Interventions for providers to promote a patient-centered approach in clinical
       consultations”, Cochrane Database Syst Rev, 4, CD003267.
   11) Carter, W. B., Inui, T. S., Kukull, W. A., and Haigh, V. H., 1982, “Outcome-based
       doctor-patient interaction analysis: II. Identifying effective provider and patient
       behavior”, Med Care, 20, 550-566.
   12) Rowland-Morin, P. A., and Carroll, G. J., 1990, “Verbal communication skills and
       patient satisfaction: a study of doctor-patient interviews”, Evaluation and the Health
       Professions, 13, 168-185.

                                                 6

				
DOCUMENT INFO