The participants were nurses who had at least 1 years experience in a hospital located in the same administrative district covered by national health insurance in Taiwan. Each of 9 district teaching hospitals recruited 50 eligible nurses during August 2012. A total of 450 selfadministered questionnaires were collected, and, of these, 436 were valid.

In order to investigate the key magnetic factors influencing nursing practice and identify the gaps between nurses’ perception and hospital performance, this study was carried out with a questionnaire using the 14 forces of magnetism drafted by the ANCC [22] and others from a nurse opinion questionnaire [23] as guides. Because it needed to account for a Taiwanese background and health system, the draft was reviewed by academic specialists and practitioners, and pre-tested and revised before the final version was distributed. The contents of the questionnaire included three dimensions: nurses’ perceptions of the importance of magnetism, nurses’ perceptions as to how the hospital performed, and nurse characteristics. The resulting 18 items were: 1) nursing leadership, 2) power-responsibility assignment, 3) management style, 4) salary structure, 5) welfare, 6) manpower deployment, 7) flexible scheduling, 8) professional care model, 9) facility and equipment, 10) quality improvement, 11) consultation and resources, 12) autonomy, 13) Nurse-patient relationships, 14) clinical teachers’ attitudes, 15) interdisciplinary relationships, 16) professional knowledge, 17) opportunity for education, 18) handling nurse-patient disputes. Each item was scored on the nurse’s perception of importance (1 = very unimportant to 5 very important) and the nurse’s perception of hospital performance (1 = very dissatisfied to 5 very satisfied).

Nurse characteristics included gender, age, education, salary, marital status, experience, work setting, clinical ladder, and work shift. Hospital attributes were public or private.

Descriptive statistics, such as mean, standard deviation, frequency and rank order, were used to investigate the key magnetic factors which influenced nursing practice.

Paired sample t-tests were used to analyze the gaps between nurses’ perceptions and hospital performance. Factors with statistical significance (P < 0.05) were defined as performance gaps.

Independent samples t-tests and ANOVA were used to examine the relationships among nurses’ characteristics, hospital attributes and performance gaps.

The IPGA matrix was used to develop strategic management for magnet hospitals. The IPGA tool is a simple graphic that enables a comparison of perceived importance against performance, and is expected to help make service decisions through a simple strategic matrix. It also includes a gap theory to identify hospital failures based on nurse’s expectations vs. their perception of hospital services. Understanding nurse’s demands should help hospital managers to provide effective leadership.

The IPGA model includes the following six steps:

Step 1: Collect nurses’ perceptions of importance and hospital performance for all 18 forces of magnetism.

Step 2: Calculate the average value of importance of each item (), the average value of performance of each item (), the average value of importance of all items (), and the average value of performance of all items ().

Step 3: Use paired sample t-tests to determine whether the gap between perception and performance for each item of the 18 forces of magnetism was a positive gap (performance > importance), negative gap (performance < importance) or no gap (performance =  importance).

Step 4: Compute the relative importance (RI) and relative performance (RP), RI =. Determine if

and the t-test was significant, RP(j) =., if and the t-test was significant, RP(j) =, or if or and the t-test was non-significant, RP(j) = 0.

Step 5: Draw the IPGA strategic matrix (Figure 1), where relative importance (nurses’ perceptions of importance) is the vertical axis and relative performance (hospital service) is the horizontal axis, and the intersec-

tion is fixed at (0,1). The IPGA grid represents the different strategies for resource allocation and management, as illustrated below:

1) Quadrant I is composed of high relative performance and high relative importance items and corresponds to “Keep up the good work”’.

2) Quadrant II is composed of low relative performance and high relative importance items, and corresponds to “Concentrate here”. A point farther away from coordinate (0,1) indicates a greater need for improvement.

3) Quadrant III is composed of low relative performance and low relative importance items, and corresponds to “Low priority”.

4) Quadrant IV is composed of high relative performance and low relative importance items, and corresponds to “Possible overkill”. A point farther away from co-ordinate (0,1) indicates a greater need to re-allocate resources; and Step 6: Determine the priorities for resource allocation for the items in Quadrant II.

The distance D(j) indicates the degree of priority for improvement:

Among the 436 participants, 45.2% worked in a public hospital and 54.8% in a private hospital; 98.8% were women. Those aged 25 - 34 accounted for 60.7%. Being single accounted for 51.8% and being married for 48.2%; 50.2% had Bachelor’s degrees. Salaries of 40,001 - 50,000 NTD were the majority (45%), followed by 30,001 - 40,000 NTD (40.7%). Twelve or more years of experience accounted for 30.9%, and 66.5% worked on internal and surgical wards. A rotating shift was the majority (57.3%) type of work. Clinical ladders N1 and N2 accounted for 28.5% and 38.7% respectively (Table 1).

For the18 forces of magnetism perceived as important and hospital performance as seen by nurses, Cronbach’s α were all >0.9. The rankings are shown in Table 2. The mean values of the nurses’ perceptions of importance for all 18 forces of magnetism were above 4.26. These forces were critical factors influencing nursing practice, especially salary structure, welfare, and manpower deployment. They were also associated with lower hospital performance as perceived by the nurses.

The 18 forces of magnetism had a statistically significant (P < 0.05) gap (Table 3). According to the IPGA model, 8 forces of magnetism in Quadrant II needed urgent improvement (Figure 1). The first priority was salary structure, the second was welfare, and the third was manpower deployment. Other priorities in sequence were flexible scheduling, handling nurse-patient disputes, clinical teachers’ attitudes, the Nurse-patient relationships and management style.

Table 1. Nurses characteristics and hospital attributes.

n

%

n

%

n

%

Hospital

Education

Clinical ladder

Public

197

45.2

216

49.8

N

51

11.8

Private

239

54.8

≥Bachelor

218

50.2

N1

123

28.5

Gender

Salary(NT)

N2

167

38.7

Female

431

98.9

30,001 - 40,000

174

40.7

≥N3

91

21.1

Male

5

1.1

40,001 - 50,000

192

45

Work setting

Age

50,001 - 60,000

47

11

Emergency

54

12.4

≤24

48

11.4

≥60,001

14

3.3

ICU

92

21.1

25 - 29

121

28.7

Experience

Internal Medicine

163

37.5

30 - 34

135

32

≤3

82

18.9

Surgery

126

29

35 - 39

74

17.5

4 - 6

79

18.2

Work shift

≥40

44

10.4

7 - 9

70

16.1

rotating shift

239

57.3

Marital

10 - 12

68

15.7

Evening/ night shift

78

18.7

Single

216

51.8

≥13

134

30.9

day shift

100

24

Married

203

48.2

Table 2. Key forces of magnetism influence nursing practice.