weight. For comparing the differences, this study also calculated the CS’ based on a 5-points original Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) but standardization plus the same AHP weight.
()
100) ) (
(
level scalelower level
scaleupper
W level scalelower Y
CS x
x x
j j j
X (16)
Where x implies must-be, attractive or one-dimension
)
j( x is the number of x quality attribute
)
j( x
w is the weight of the jth quality element of x quality attribute
)
j( x
Y is the average score of the jth quality element of x quality attribute
X
X W
CS
CS
(17)Where WX is the weight of x quality attribute.
The main stages of survey on this study are as shown on Figure 7.
Section 3 Empirical Example of Customer
three major local hospitals as the objects of this empirical study.
Reviewed previous studies to develops the quality attributes of nursing care
Phase 1:
←Pilot study
To confirm pair wise Kano questionnaire
← Samples responded by mail To category nursing service attributes
Phase 2:
← Evaluation table by QEs To measure service performance then transfers to
patients satisfaction of each quality attributes
← AHP by QEs
To find the weight of each quality attributes
To compare CS (based on this approach) and CS’ (according to Likert Scale)
Figure 7 The processes of survey
Phase 1
The source of subjects is inpatient for the first questionnaire. Based on the Chou, Chen, Woodard, and Yen (2005) and Jacobi, Triemstra, Rupp, Dinant, and Van Den Bos (2004), the study also considers the current medical environments in Taiwan to select nursing care service quality elements where proposed on questionnaires. The objective is to understand that, in Taiwan area, the inpatient’s perception of nursing care service. After reviewing several measurement tools developed by previous researchers and considering patients' ability to answer the questions, the study
selected 27 items to describe the service quality elements of nursing care. 80 discharged patients joint a small-scaled pilot test to answer the standardized pre-questionnaire. All patients have discharged from hospital approximately three months. The outcome data of pilot test would be applied to measuring the reliability of the service characteristics by Cronbach’s α and the validity by principle component analysis. Statistical analysis was done with SPSS for Windows, version11.5. The overall alpha for each independent sample and the overall sample was between 0.92 and 0.96. The principle component analysis showed that four items had no aspects in common with any other categories. The analysis was, therefore, carried out with the remaining 23 items. The correlation matrix between the 23 items showed positive correlation coefficients ranging from 0.179 to 0.724 (P < 0.001 in all cases).
After confirmed the questionnaire structure, the next step of this phase is to classify the 23 quality elements into which Kano category. Postal questionnaire were undertaken because of consideration of costs. There were 387 returns correctly completed of the 1000 sent out surveys, giving a response rate of 38.7%. Of the total sample, who responded to a questionnaire administered eight weeks post discharge were included in the analysis. It uses the Kano’s evaluation table to categorize all service attributes. Satisfaction increment index (SII) and dissatisfaction decrement index (DDI) can clearly figure out which quality element can result in maximum impact. The SII and the DDI are defended as the following formula according to Berger, et al. (1993). A minus sign of DDI means to stress its negative influence on customer satisfaction if this quality element is not functional (see Table 16).
I M O A
O SII A
(18)
A O M I
M
DDI O (19)
Table 16
The results of the categorization of quality elements
The main descript of items SII DDI
M1 Have enough knowledge and skills 0.2 -0.933
M2 Inform on course of symptoms 0.533 -0.733
M3 Provide comfortable and quiet surroundings 0.4 -0.733 M4 Explanation of hospital routines 0.467 -0.667
M5 Being treated with respect 0.4 -0.8
M6 Responding to patients’ request 0.333 -0.867
M7 Use modern looking equipment 0.5 -0.938
M8 Physical facilities are visually appealing 0.5 -0.875 M9 Provide care at the time they promised 0.53 -0.933
A1 Inform about home adjustments 0.938 -0.25
A2 Assure good care coordination 1 -0.333
A3 Encourage patient to take care of himself 0.8 -0.4 A4 Post-discharge service arrangements 0.938 -0.313 A5 Have the patients-best interests at heart 0.867 -0.267 A6 Inform about looking after condition 1 -0.267 O1 Keep patient’s file confidentially 0.917 -0.583
O2 Have a friendly attitude 0.933 -0.6
O3 Coming in not only to give mediation or display treatment 0.8 -0.667 O4 Give positive and corrective feedback 0.8 -0.733
O5 Individualization of care 0.733 -0.8
O6 Privacy 0.667 -0.867
I1 Make sure that the patient sees the same provider at
each visit 0.133 -0.333
I2 Inform in plain words 0.375 -0.125
Phase 2
The survey team recruited fifteen 18-year-older volunteers as QEs who were inpatient non-ICU nursing care during 2005 at all three target hospitals located in Hisn-Chu City of Taiwan. Taiwan, as known as a place where the hospital care market is highly competitive and patients can choose any hospital to receive care under a universal health insurance scheme. All volunteers were cancer patient and inpatient at all these three hospitals more than three days. In order to guarantee the fulfillment of the hypothesis of stochastic independence of the judgments for each QE (Erto & Vanacore, 2000), the fifteen quality experts stayed over three days at each hospital to fulfill the Evaluation Table and AHP table. The Evaluation Table for the nursing care service contains 23 quality elements selected form Phase 1 as representatives. The survey would collect data form QEs’ responses after 20 days of their discharge. Consequently, 45 returns would be collected from 15 QEs. Five-level
response categories were designed respectively for each of the 23 questions, such as
“agree” to “disagree”, “strongly agree” to “strongly disagree”, or “neither agree nor disagree”. We let the QEs indicate whether the extent to which they agree with the statements about the nursing care service they received should exceed their expectation. Scores ranging were assigned to the five categories of answers as we early suggested (see Table 15). Missing values in each item would be replaced by the mean scores for each of the five diagnoses, thus, it would maximize the amount of useful information. Not only did the QEs answer the Evaluation Tables but also AHP table. The study collected outcome data of both tables.
2 Result and Discussion
The results of pair wise Kano questionnaire to classify the quality categories are shown in Table 16. According to the results as shown in Table 16, there are 9 must-be quality elements, 6 attractive quality elements, 6 one-dimensional quality elements and 2 indifferent quality elements of all 23 service quality elements. Both SII and DDI, when the value is close to 0 means that the influence is low, meanwhile if it is close to 1 means that the element can increase customer satisfaction; if the value is close to -1 means that the element has a positive influence on decreasing customer dissatisfaction. This research has calculated the SII and DDI of service quality of the nursing care service as shown in Table 16. The values of weight of each elements obtained by AHP. The higher weights demonstrate that patients concern about the psychological or empathetic aspects more than the physical feature.
According to the results in Table 17 and 18, the suggestions to three hospitals respectively are as follows: Comparing to the patient satisfaction degree of must-be attributes, although Hospital B performed better whereas Hospital C and A got similar performance below Hospital B did, however, all of them had to do more efforts to decrease dissatisfaction to reach ultimate performance. Comparing to the patient satisfaction degree of one-dimensional attributes, Hospital B performed best, Hospital A in second and Hospital C in worst. Hospital C had to benchmark both A and B to increase satisfaction and decrease dissatisfaction. Comparing to the patient satisfaction degree of attractive attributes, Hospital A attained to very high degree whereas Hospital B and Hospital C missed reach the three-quarter point. For the overall performance of patients satisfaction degree, Hospital A was the top one, Hospital B the second one and Hospital C the last one. In Table 17 and 18, under the
same weights and different calculation method, the satisfaction sequence of “must-be attribute” in three hospital is B > A > C. The sequence of “attractive attributes” is A >
B > C. The sequence of “one-dimensional attribute” is B > A > C. However, after calculating and standardizing by 5-points Likert scale, the highest satisfaction of overall is to hospital B, and the second is to hospital B, while hospital C is on the third. Moreover, the satisfaction divergence of A and B in overall is not big. However, the satisfaction sequences of patients will reverse as A > B > C by applying the suggested evaluating weights in the study. The main reason to cause the different is because when calculating with Likert scale, there equidistance in each numeric and linear relation between satisfaction and quality performance.
Table 17
Results of the patient satisfaction (CS)
Quality elements
Weights (w)
Average Satisfaction (S)
Weighted Average Satisfaction (w×S)
A B C A B C
M1 0.050 -0.171 -0.073 -0.319 -0.855 -0.366 -1.595 M2 0.052 -0.192 -0.054 -0.403 -0.998 -0.281 -2.096 M3 0.060 -0.208 -0.045 -0.417 -1.248 -0.270 -2.502 M4 0.202 -0.309 -0.036 -0.272 -6.242 -0.727 -5.494 M5 0.142 -0.267 -0.018 -0.376 -3.791 -0.253 -5.339 M6 0.181 -0.309 -0.089 -0.319 -5.593 -1.611 -5.774 M7 0.080 -0.268 -0.128 -0.062 -2.144 -1.024 -0.496 M8 0.133 -0.281 -0.018 -0.417 -3.737 -0.237 -5.546 M9 0.100 -0.309 -0.036 -0.389 -3.090 -0.360 -3.890 Must-be (weight = 0.321): -27.70 -5.13 -32.73 A1 0.196 0.942 0.333 0.025 18.463 6.527 0.490 A2 0.188 0.942 0.417 0.121 17.710 7.834 2.275 A3 0.171 0.883 0.225 0.154 15.099 3.848 2.633 A4 0.066 0.913 0.375 0.042 6.026 2.475 0.277 A5 0.159 0.833 0.354 0.054 13.245 5.629 0.859 A6 0.220 0.825 0.288 0.029 18.150 6.325 0.638 Attractive (weight = 0.377): 88.69 32.64 7.17 O1 0.189 0.067 0.266 -0.530 1.266 5.027 -10.017 O2 0.063 0.567 0.533 -0.430 3.572 3.358 -2.709 O3 0.083 0.533 0.567 -0.300 4.424 4.706 -2.490 O4 0.218 0.033 0.133 -0.047 0.719 2.899 -1.018 O5 0.172 0.266 0.400 -0.370 4.575 6.880 -6.364 O6 0.275 0.133 0.600 -0.233 3.658 16.500 -6.408 One-dimensional (weight = 0.302): 18.21 39.37 -29.01
Overall satisfaction: 30.05 22.55 -16.56
Table 18
Results of the patient satisfaction based on likert scale (CS’)
Quality elements
Weights (w)
Average Satisfaction (S’)
Weighted Average Satisfaction (w×S’)
A B C A B C
M1 0.050 3.800 4.466 2.933 3.500 4.333 2.416 M2 0.052 3.667 4.600 2.460 3.467 4.680 1.898 M3 0.060 3.533 4.667 2.400 3.800 5.501 2.100 M4 0.202 3.000 4.733 3.200 10.100 18.852 11.110 M5 0.142 3.200 4.866 2.600 7.810 13.724 5.680 M6 0.181 3.000 4.333 2.933 9.050 15.082 8.747 M7 0.080 3.200 4.533 2.667 4.400 7.066 3.334 M8 0.133 3.133 4.866 2.400 7.092 12.854 4.655 M9 0.100 3.000 4.733 2.533 5.000 9.333 3.833 Must-be (weight = 0.321): 54.22 91.42 43.77 A1 0.196 4.867 3.267 1.400 18.948 11.108 1.960 A2 0.188 4.867 3.533 2.133 18.175 11.905 5.325
A3 0.171 4.733 2.866 2.400 15.959 7.977 5.985
A4 0.066 4.800 3.400 1.530 6.270 3.960 0.875
A5 0.159 4.600 3.330 1.600 14.310 9.262 2.385
A6 0.22 4.600 3.066 1.460 19.800 11.363 2.530
Attractive (weight = 0.377): 93.46 55.58 19.06 O1 0.189 3.133 3.533 1.930 10.078 11.968 4.394
O2 0.063 4.130 4.060 2.133 4.930 4.820 1.784
O3 0.083 4.060 4.130 2.400 6.350 6.495 2.905
O4 0.218 3.066 3.267 2.067 11.260 12.355 5.815 O5 0.172 3.533 3.800 2.267 10.892 12.040 5.448 O6 0.275 3.267 4.200 2.530 15.586 22.000 10.519
One-dimensional (weight = 0.302): 59.095 69.678 30.866
Overall satisfaction: 70.49 71.34 30.56
3 Conclusion
Some focus on patient expectation consists of what factors; some try to figure out patient satisfaction belongs to what categories; some attempt to explain patient perception and characteristics by what definitions. However, most of them applied the methods of close question, directly rating and sum of ratings. The majority studies were used to detecting the relationship between customer satisfaction and quality performance by Likert scale. But the conceptual complexes of customer satisfaction or quality performance make it much harder by direct measurement or summing up. For example, if the category contains two elements: Element 1 and Element 2. There are five options of Element 1 and Element 2 selected by Respondent A and Respondent B. Respondent A evaluates Element 1 as “strongly
agree” (5 points) and Element 2 as “strongly disagree” (1 point). Respondent B evaluates both of Element 1 and Element 2 as “neither agree nor disagree” (3 points).
According to original Likert scale, both of A and B get same score as 6 points (A = 5+1, B = 3+3), that means, both of A and B have the same evaluation about this category. Thus, the facts cannot really reflect on the outcome.
When seven-point Likert-type scale is adopted on a questionnaire, customer satisfactions corresponding to the different types of the categories proposed by Kano can also be derived from equations (11)-(15) and listed as shown as Table 19.
Table 19
Modified seven-points likert scale list
Judgment Customer Satisfaction
Sm So Sa Sr Si
Strongly agree C1 0 1 1 -1 0
Agree C2 1
6 5
6 4
36 25
6
4 0
Slightly agree C3 1
6 4
6 2
36 16
6
2 0
Neutral C4 1
6
3 0
36
9 0 0
Slightly disagree C 1
6 2
6
2
36 4
6
2 0
Disagree C6 1
6 1
6
4
36 1
6
4 0
Strongly disagree C7 -1 -1 0 1 0