A pilot study on nurses' attitudes toward perinatal bereavement support: a cluster analysis.

AIM: Nurses' attitudes towards perinatal bereavement care are explored by identifying profiles of nurses working in a Hong Kong Obstetrics and Gynaecology (OAG) unit. Relationships between nurses' attitudes towards bereavement support, need for bereavement training and hospital policy are explored. RESEARCH
METHOD: 110 nurses recruited from the OAG unit of a large Hong Kong public hospital completed a structured questionnaire. OUTCOME MEASURES: Attitudes towards perinatal bereavement support; required support and training needs for nurses on bereavement care.
RESULTS: Two-step cluster analysis yielded two clusters. Cluster A consisted of 55.5% (n = 49) and cluster B consisted of 44.5% (n = 61) of nurses. Cluster A nurses were younger, had less OAG experience, more junior ranking and less education than cluster B nurses. Cluster B nurses had additional midwifery and bereavement care training, personal grieving experiences and experience handling grieving clients. The majority held positive bereavement care attitudes. Significant differences towards perinatal bereavement support were found. Only 25.5% (n = 28) had bereavement related training. Attitudes towards bereavement care were positively correlated with training needs (rs = 0.59) and hospital policy support (rs = 0.60).
CONCLUSION: Hong Kong nurses emphasized need for increased bereavement care knowledge and experience, improved communication skills, and greater hospital and team members' support. Findings may be used to improve support of nurses, to ensure sensitive bereavement care in perinatal settings, and to enhance nursing curricula.

Introduction

During the last 20 years, there have been numerous publications concerned with bereavement counselling services in Western countries (MacCarthy and MacKeith, 1965; Beem et al., 1998; Gardner, 1999; Gensch and Midland, 2000; Engler and Lasker, 2000). Gensch and Midland (2000) have been addressing the needs of parents whose babies die during pregnancy or shortly after birth since the early 1980s. The `Resolve through Sharing (RTS) Program' was the first well-developed model. They use it to support bereaved couples and families experiencing pregnancy or perinatal loss. In 1982, Gardner (1982) developed a questionnaire to assess the needs and feelings of 44 nurses who cared for bereaved parents. One year later, Gardner (1983) surveyed 100 couples that had experienced stillbirth. Her study revealed that the most helpful care practices were the ones least used by the nurses who cared for them. Another important resource for bereavement care services was written in 1986 by Limbo and Wheeler (1998). It is considered the `Bible' for newcomers interested in this area of nursing. Pregnancy and childbirth are times of great joy and rejoicing, and women generally approach motherhood with happiness. Because it is usually sudden and contrary to the expected pattern of events, a perinatal loss can be considered particularly traumatic for women who experience it (Symes, 1991). Induced or spontaneous pregnancy loss may cause sustained psychological morbidity (Lewis, 1976; Lloyd and Laurence, 1985). It is likely that the attitudes of medical and nursing staff can influence recovery from a pregnancy loss and that nurses with a positive attitude to bereavement care can help bereaved families to cope during their grieving period (Lewis, 1976).

In 1999, more than 50,000 babies were born in Hong Kong. Sadly, in that year, the perinatal mortality rate was 5.7 per 1000 total births (Hospital Authority, 2003). Consequently, many hospitals in Hong Kong set-up grief counselling and support teams composed of physicians, nurses and clinical psychologists. The role of these units is to offer bedside emotional support (care) to bereaved couples. The support service facilitates the recovery of bereaved couples going through the period of grieving. This is congruent with the organizational mission of the units, which is to enhance quality patient-centred services through teamwork. However, caring for and supporting parents whose baby has died is extremely demanding, difficult and stressful (Nussbaumer and Russell, 2003; Billson and Tyrrell, 2003). Therefore, nurses involved in the care and support of bereaved parents need to be trained with the relevant knowledge, skills and understanding to acquire sufficient confidence in their ability to provide adequate and appropriate care (de Groot-Bollüjt and Mourik, 1993). It is recommended that pre-registration, post-registration and in-service training be provided for nurses who will care for those who have lost a baby. Thus, special skills are needed to help bereaved parents. After appropriate training, it is logical to assume that nurses are better equipped to cope with perinatal bereavement (Stillbirth and Neonatal Death Society, 1991). In dealing with bereavement care, the most important attitude is caring. The attitudes of nurses can affect the quality of care provided to support bereaved parents and families experiencing perinatal loss. Nursing care may not be based on individual care, but rather on the attitudes of nurses (Shaw and Heyman, 1982). When nurses provide bereavement care, negative attitudes may distract from the provision of good care; whilst positive attitudes can help bereaved couples to cope with the grieving process and create memories for the future. This study was undertaken to address the paucity of information available on nurses' attitudes towards perinatal bereavement support in Hong Kong. This quantitative study was conducted with the aim of helping to fill this gap and the findings may provide valuable baseline information for future studies.

The aim of this research was to study nurses' attitudes towards perinatal bereavement support. Three specific research objectives were formulated:

To identify profiles of nurses based on their demographic variables and their attitudes toward perinatal bereavement support;

To explore any differences among nurses' attitudes towards bereavement support in terms of demographics, practical experience and training factors in the different profile groups;

To explore the relationship between nurses' attitudes towards bereavement support, their need for bereavement training and hospital policy.

Methods

Sample

Convenience sampling was chosen for this study, which involved nurses working in the Department of Obstetrics and Gynaecology (OAG) in one of the largest hospital in Hong Kong. The nurses included student nurses/midwives, registered nurses/midwives, nursing officers, ward managers and nurse specialists and nurse educators. In this study, a structured self-reported questionnaire was used to collect data. Permission was obtained from the Hong Kong Hospital Authority. Prior to the study, an information sheet was given to all nurses explaining the purpose and procedures of the survey. In order to answer any possible questions that potential participants might have about the study, a contact person and telephone number were provided. All participants were asked to sign a consent form attached to the questionnaire. Participants were told that all information would be treated confidentially. Names and staff IDs were not indicated on the returned questionnaires. The questionnaire was distributed to nurses at their work place through their ward managers. Instructions were given to complete and return the questionnaire to the researcher by mail within one week. Totally 125 questionnaires were sent out, 110 were completed and returned, for a response rate of 88%.

Instrument

An intensive study of the existing literature describing nurses' attitudes towards perinatal bereavement revealed 21 items considered to be important to train nurses for caring and supporting bereaved parents and families. Conceptually, three dimensions could be distinguished. A panel of experts (a nurse specialist, a senior medical officer and a clinical psychologist) established content validity. A pilot study was performed with five nurses working in the OAG of a hospital other than the target hospital. After the pilot and repeated discussions with the expert panel, three aspects of the questionnaire were confirmed: (1) nurses' attitudes towards perinatal bereavement support, (2) the importance of hospital policy, and (3) training on perinatal bereavement support. The content validity indices (CVI) ranged from 0.78 to 0.89 for sub-aspects, with an overall score of 0.86. Shelley (1984) stated that a CVI below 0.50 for a measure indicated an unacceptable level of content validity. Since these values were greater than this level, the measure was deemed acceptable. Respondents were asked to rank each item, in each part, on 5 point Likert-type scales. Part one comprised demographic data (i.e., age, education level, recent ranking and religious background), experience aspects (i.e., personal grieving, handling grieving clients and years of work in the OAG unit) and training aspects (i.e., midwifery and bereavement care). This information provides knowledge about participants, and is used to determine differences between or within groups related to the data obtained (i.e., attitude scores). Part two comprised 9 attitude statements formulated to measure nurses' general attitudes towards bereavement care (1=strongly disagree, 5=strongly agree). Part three is composed of two sections; the first section comprised four statements to evaluate nurses' attitudes on the importance of hospital policy on perinatal bereavement (1=very unimportant, 5=very important). The second section comprised eight statements to evaluate nurses' attitudes on the importance of formal training to deal with perinatal bereavement care (1=very unimportant, 5=very important). For all statements in part two, a high score indicates a positive and favourable attitude towards perinatal bereavement support. In part three, a high score indicates attitudes that are highly influenced by policy and a high demand for nurses with perinatal bereavement care training.

Statistical methods

Descriptive statistical analysis of the quantitative data was conducted using SPSS 11.5 (SPSS Inc., 2002). Several statistical techniques were employed for analysis of the data. A cluster analysis was used to group the study sample into several groups. This analysis, also called segmentation or taxonomy analysis, is a way to create groups of objects, or clusters in such a way that the profiles or characteristics of nurses in the same cluster are very similar and that the profiles of nurses in different clusters are quite distinct (Chan et al., 2003). In this study, the analysis was based on four main variables: (1) demographic variables; (2) attitudes towards bereavement care variables; (3) attitudes on the importance of hospital policy on perinatal bereavement variables; and (4) attitudes on the importance of formal training to deal with perinatal bereavement care variables. The analysis used standardized values (z) based on these variables. Two-step cluster analysis was used for this study. In this technique, nurses are sequentially merged into a decreasing number of clusters until the conjoint set contains all nurses (SPSS Inc., 2002). The choice of a similarity measure and the determination of the number of clusters were based on the Log-likelihood distance and Schwarz's Bayesian information criterion (BIC), respectively (SPSS Inc., 2002). Differences between the cluster groups were delineated descriptively. Due to ordinal data and skewed patterns in most of the outcomes, Chi-square test and Mann–Whitney U-test were used to test differences in various nurses' characteristics between the clusters (Hall-Lord et al., 1999). In addition, factor analysis was used to identify the number of factors, and Cronbach's alpha coefficients were used to examine the internal reliability of each part of the instrument. Spearman's correlation coefficients (r s) were used to quantify the relationship between the three attitude scores. Results were considered significantly different at p<0.05.

Results

Cluster analysis

The Two-step Cluster analysis yielded two clusters based on Schwarz's BIC=2843.975 and the highest Log-likelihood distance measures (ratio=1.875). Of the study nurses, cluster A consisted of 55.5% (n=61) and cluster B consisted of 44.5% (n=49). The two subgroups of nurses (clusters) were formed on the basis of the similarity of their responses to the instrument. The two nurse profiles were compared with regard to four main variables. The variables included: (1) demographic data (i.e., age, education level, recent ranking, etc.); (2) attitudes toward bereavement care (9 items); (3) attitudes toward hospital policy to bereavement care (4 items); and (4) attitudes on training for bereavement care (8 items). Profiles are depicted in Table 1, Table 2 .

Table 1

Demographic data of sample by clusters.

Variables	Total (n=110)		Cluster
	n	(%)	A (n=61)		B (n=49)			Chi-squared test
			n	(%)	n	(%)		χ2	P value
Age (years)
20–25	37	(33.6)	35	(57.4)	2	(4.1)	]	62.89	<0.001
26–30	33	(30.0)	23	(37.7)	10	(20.4)
>30	40	(36.4)	3	(4.9)	37	(75.5)
Experience in OAG
<5 years	63	(57.3)	58	(95.1)	5	(10.2)	]	80.62	<0.001
5–10 years	26	(23.6)	3	(4.9)	23	(46.9)
>10 years	21	(19.1)	0	(0.0)	21	(19.1)
Midwifery training
Yes	93	(84.5)	45	(73.8)	48	(98.0)	]	12.17	<0.001a
No	17	(15.5)	16	(26.2)	1	(2.0)
Recent rank
S	24	(21.8)	24	(39.3)	0	(0.0)	]	45.48	<0.001
RN	65	(59.1)	37	(60.7)	28	(57.1)
SN	21	(19.1)	0	(0.0)	21	(42.9)
Education level
Diploma or below	71	(64.5)	53	(86.9)	18	(36.7)	]	29.86	<0.001
BSc or higher	39	(35.5)	8	(13.1)	31	(63.3)
Religious belief
Yes	45	(40.9)	25	(41.0)	20	(40.8)	]	0.01	0.986
No	65	(59.1)	36	(59.0)	29	(59.2)
Personal grieving experiences
Yes	33	(30.0)	10	(16.4)	23	(46.9)	]	12.07	0.001
No	77	(70.0)	51	(83.6)	26	(53.1)
Taken training/courses related to bereavement care
Yes	28	(25.5)	7	(11.5)	21	(42.9)	]	14.10	<0.001
No	81	(74.5)	54	(88.5)	28	(57.1)
Experience in handle grieving clients
Yes	55	(50.0)	10	(16.4)	45	(91.8)	]	61.86	<0.001a
No	55	(50.0)	51	(83.6)	4	(8.2)

S, student nurse/midwife.

RN, registered nurse/midwife.

SN, nursing officer/ward manager/specialist/educators.

OAG, obstetric and gynaecology unit.

Fisher's exact test.

Table 2

The responses of nurses on 21 statements and the results from factor and reliability analysis for each statement.

Statements	Cluster A (n=61)						Cluster B (n=49)						Chi-squared test		Factor analysisa										Reliability analysis
	(1)–(3)		(4)		(5)		(1)–(3)		(4)		(5)		χ2	P value	Total			Cluster A			Cluster B				Total α	Cluster A (α)	Cluster B (α)
	n	(%)	n	(%)	n	(%)	n	(%)	n	(%)	n	(%)			19.2% (factor 1)	15.6% (factor 2)	23.0% (factor 3)	17.2% (factor 1)	12.6% (factor 2)	22.6% (factor 3)	20.0% (factor 1)	25.0% (factor 2)	15.8% (factor 3)
Attitudes toward perinatal bereavement careb
1 I believe that the Grief Counseling Programme can provide psychological support to the bereaved couple.	8	(13.1)	46	(75.4)	7	(11.5)	5	(10.2)	26	(53.1)	18	(36.7)	9.9	0.007	0.66			0.56			0.53		0.37	]	0.86	0.80	0.84
2 I agree that the `Good-bye Baby' Parent Support Group provides support to parents with similar experience.	9	(14.8)	51	(83.6)	1	(1.6)	6	(12.2)	34	(69.4)	9	(18.4)	9.2	0.010	0.61			0.45			0.57
3 I agree that parents should be supported in making their own decisions about what happens to them.	15	(24.6)	38	(62.3)	8	(13.1)	7	(14.3)	24	(49.0)	18	(23.6)	8.7	0.013	0.53			0.65			0.38
4 I respect the bereaved parent's feelings and needs.	12	(19.7)	36	(59.0)	13	(21.3)	1	(2.0)	28	(57.1)	20	(40.8)	10.6	0.005	0.71			0.74			0.79
5 I will communicate with parents in a clear, sensitive and honest manner.	15	(24.6)	36	(59.0)	10	(16.4)	2	(4.1)	25	(51.0)	22	(44.9)	15.3	<0.001	0.69			0.68			0.66
6 I agree that parents should be given time to grieve.	8	(13.1)	45	(73.8)	8	(13.1)	0	(0.0)	19	(38.8)	30	(61.2)	30.4	<0.001	0.65		0.37	0.59			0.57	0.38	0.47
7 I agree that parents should be treated with respect and dignity.	13	(21.3)	44	(72.1)	4	(6.6)	1	(2.0)	21	(42.9)	27	(55.1)	34.6	<0.001	0.70	0.38		0.69			0.71	0.35
8 All those involved in the care of bereaved parents should be well informed.	11	(18.0)	43	(70.5)	7	(11.5)	0	(0.0)	28	(57.1)	21	(42.9)	20.1	<0.001	0.60	0.44		0.60			0.68	0.50
9 All those who care for and support the bereaved parents should have access to support for themselves.	12	(19.7)	39	(63.9)	10	(16.4)	3	(6.1)	28	(57.1)	18	(36.7)	8.3	0.016	0.49			0.37			0.72
Sub-total, mean (SD) [score range from 9 to 45]	35.2 (3.5)						39.2 (3.5)						631.5d	<0.001
Attitudes toward hospital policy to bereavement carec
10 Nurses should feel assured that they are working within an operational policy which is adequate and appropriate.	29	(47.5)	30	(49.2)	2	(3.3)	5	(10.2)	32	(65.3)	12	(24.5)	23.1	<0.001		0.49			0.48			0.36		]	0.83	0.76	0.84
11 The unit should have a clear policy for the management of bereavement.	11	(18.0)	40	(65.6)	10	(16.4)	2	(4.1)	21	(42.9)	26	(53.1)	18.2	<0.001		0.71	0.41		0.64	0.46		0.82
12 All staff involved should be well informed about the policy.	7	(11.5)	42	(68.9)	12	(19.7)	1	(2.0)	26	(53.1)	22	(44.9)	10.0	0.007		0.89			0.91			0.88
13 All staff involved should understand the policy.	6	(9.8)	40	(65.6)	15	(24.6)	0	(0.0)	28	(57.1)	21	(42.9)	7.9	0.019		0.79			0.77			0.82
Sub-total, mean (SD) [score range from 4 to 20]	15.7 (1.8)						17.47 (1.9)						752.5d	<0.001
Attitudes on the training for bereavement carec
14 Nurses involved in the care and support of bereaved parents need to be equipped with relevant knowledge, skills and understanding.	7	(11.5)	35	(57.4)	19	(31.1)	0	(0.0)	25	(51.0)	24	(49.0)	8.0	0.018			0.60			0.60			0.43	]	0.90	0.89	0.89
15 Nurses need to feel confident that they are providing adequate and appropriate care.	7	(11.5)	36	(59.0)	18	(29.5)	0	(0.0)	30	(61.2)	19	(38.8)	6.2	0.042			0.74			0.71		0.36	0.74
16 Nurses need to know that they have limitation when provide perinatal bereavement care.	14	(23.0)	35	(57.4)	12	(19.7)	3	(6.1)	30	(61.2)	16	(32.7)	6.8	0.033			0.63			0.58			0.76
17 Nurses need opportunities to express their own feelings and needs.	10	(16.4)	34	(55.7)	17	(27.9)	0	(0.0)	25	(51.0)	24	(49.0)	11.4	0.003			0.68			0.72			0.40
18 Joining training programme on bereavement care.	9	(14.8)	32	(52.5)	20	(32.8)	1	(2.0)	29	(59.2)	19	(38.8)	5.3	0.070			0.80			0.85			0.46
19 Participating in bereavement care.	12	(19.7)	33	(54.1)	16	(26.2)	1	(2.0)	32	(65.3)	16	(32.7)	8.1	0.017			0.74			0.75			0.35
20 Sharing experience with colleagues and working as a team.	10	(16.4)	28	(45.9)	23	(37.7)	0	(0.0)	30	(61.2)	19	(38.8)	9.3	0.010			0.74			0.71	0.36	0.47	0.59
21 Seeking support when feeling `burnout'.	10	(16.4)	30	(49.2)	21	(34.4)	0	(0.0)	27	(55.1)	22	(44.9)	9.0	0.011			0.81			0.83			0.69
Sub-total, mean (SD) [score range from 8 to 40]	33.0 (4.2)						35.14			(3.1)			1041.0d	0.006
Total																									0.92	0.89	0.92

% are rotated percentage of variance.

α, Cronbach's alpha.

Rotated factor loading of each item use of factor analysis, principal component method, and value less than or equal to 0.35 were not shown.

(5) Strongly agree, (4) agree, (3) uncertain, (2) disagree, (1) strongly disagree.

(5) Very important, (4) important, (3) uncertain, (2) unimportant, (1) very unimportant.

Mann–Whitney U-test.

Comparison of demographic data

The sample consisted of 110 nurses, their personal characteristics are shown in Table 1 and significant differences were noted between the nurses in the two clusters for all demographic variables except their religious beliefs. The nurses in cluster A were younger in age (age 20–25: n=35, 57.4%; χ 2=62.9, p<0.001), were less experienced in the OAG unit (<5 years: n=58, 95.1%; χ 2=80.6, p<0.001), were at a more junior level (student nurse: n=24, 39.3%; χ 2=45.5, p<0.001) and had a lower education level (diploma: n=53, 86.9%; χ 2=29.9, p<0.001) than the nurses in cluster B. By contrast, the nurses in cluster B had additional training in both midwifery (yes: n=48, 98.0%; χ 2=12.2, p<0.001) and bereavement care (yes: n=21, 42.9%; χ 2=14.1, p<0.001), had experiences in personal grieving (yes: n=23, 46.9%; χ 2=12.1, p=0.001) and had handled grieving clients (yes: n=45, 91.8%; χ 2=61.9, p<0.001).

Comparison of attitudes toward bereavement care data

Table 2 displays the results in which nurses responded to items on nurses' attitudes towards bereavement care (statements 1–9). There were significant differences in the nurses' preferences between the two clusters for all items. Most nurses in cluster A agreed that item (1): the Grief Counselling Programme (75.4%, n=46) and item (2): the `Good-bye Baby' Parent Support Group (83.6%, n=51) can provide support to bereaved couples. However, 24.6% (n=15) of nurses in cluster A were uncertain/disagreed/strongly disagreed about items (3) and (5). Significant differences were found compared with the nurses in cluster B for item (3) [14.3%, n=7; χ 2=8.7, p<0.013] and item (5) [4.1%, n=2; χ 2=15.3, p<0.001]. In cluster A, 13.1% (n=8) and 21.3% (n=13) of nurses were uncertain about whether to give parents time to grieve or not (item 6) and whether they should be treated with respect and dignity (item 7), respectively. By contrast, 61.2% (n=30) and 55.1% (n=27) of nurses in cluster B were in strong agreement with item 6 (χ 2=30.4, p<0.001) and item 7 (χ 2=34.6, p<0.001), respectively. For the total sum scores, nurses in cluster B showed significantly higher scores (mean=39.2, SD=3.5) than nurses in cluster A (mean=35.2, SD=3.5; U=631.5, p<0.001).

Comparison of attitudes on the importance of support from hospital policy data

Four items (statements 10–13) were related to hospital policy support for bereavement practices (see Table 2). For the four items and their total scores, comparisons were made between the nurses in the two clusters. There were significant differences between the two clusters in the nurses' preferences for all items. For example, 24.5% (n=12) of nurses in cluster B claimed that it is very important that there should be an operational policy on bereavement care in the obstetric unit, but of nurses in cluster A, only 3.3% (n=2; χ 2=23.1, p<0.001) were in agreement. Over 40.0% of nurses in cluster B claimed that it was very important that the policy should be well informed (44.9%, n=22) and understood (42.9%, n=21) by all staff, which was different from nurses in cluster B where 19.7% made that claim (n=12; χ 2=10.0, p=0.007) and 24.6% (n=15; χ 2=7.9, p=0.019), respectively. For the total sum scores, nurses in cluster B showed significantly higher scores (mean=17.5, SD=1.9) than the nurses in cluster A (mean=15.7, SD=1.8; U=752.5, p<0.001).

Comparison of attitudes on the importance of training on bereavement care data

Eight items (statements 14–21) reflected the responses of nurses on the need for formal education and training in grief counseling (see Table 2). In general, most of the nurses in the two clusters perceived training to be very important/important. However, more than 14.0% of nurses in cluster A claimed that their need to participate in bereavement care (19.7%, n=12), join relevant training programmes (14.8%, n=9), and to be equipped with relevant knowledge, skills and understanding required to support and care for bereaved parents was uncertain/unimportant/very unimportant. This was significantly different from nurses in cluster B. For the sub-total scores, nurses in cluster B showed significantly higher scores (mean=35.1, SD=3.1) than the nurses in cluster A (mean=33.0, SD=4.2; U=1041.0, p=0.006).

Instrument reliability

For all nurses, three factors were generated that accounted for 57.8% of the variance. Factor 1 accounted for 19.2% of the variance (factor loading ranged from 0.49 to 0.71) and reflected concerns related to nurses' attitudes towards perinatal bereavement care. Factor 2 contributed 15.6% of the variance (factor loading ranged from 0.49 to 0.89) and focused on the need for bereavement care training. Factor 3 was comprised of items related to hospital policy related to bereavement care and accounted for 23.0% (factor loading ranged from 0.60 to 0.81) (see Table 2). Overall, the Cronbach's alpha of the instrument for 21 items was 0.92. The internal reliability of the subscales was 0.86, 0.83 and 0.90, respectively for the nurses' attitudes towards perinatal bereavement support scales, the importance of hospital policy scales, and perinatal bereavement support training scales.

For nurses in cluster A, three factors were generated that accounted for 52.4% of the variance. Factor 1 accounted for 17.2% of the variance (factor loading ranged from 0.37 to 0.74). Factor 2 contributed 12.6% of the variance (factor loading ranged from 0.48 to 0.91). Factor 3 accounted for 22.6% (factor loading ranged from 0.58 to 0.85). Overall, the Cronbach's alpha of cluster A for 21 items was 0.89. The internal reliability of the subscales was 0.80, 0.76 and 0.89, respectively for the nurses' attitudes towards perinatal bereavement support scales, the importance of hospital policy scales, and the perinatal bereavement support training scales.

For nurses in cluster B, three factors were generated that accounted for 60.8% of the variance. Factor 1 accounted for 20.0% of the variance (factor loading ranged from 0.38 to 0.79). Factor 2 contributed 25.0% of the variance (factor loading ranged from 0.36 to 0.88). Factor 3 accounted for 15.8% (factor loading ranged from 0.35 to 0.76). Overall, the Cronbach's alpha of cluster B was 0.92, and the internal reliability of the subscales was 0.84 (attitude towards perinatal bereavement support scales), 0.84 (the importance of hospital policy scale) and 0.89 (perinatal bereavement support training scales). All scales for all nurses and for nurses in each cluster demonstrated satisfactory internal reliability based on the typical alpha threshold for a new instrument (Nunnally, 1978).

Relationships between nurses' attitudes towards bereavement care, hospital policy and need for training

The relationships between the three attitude sum scores were analyzed by Spearman's correlation coefficient (r s). The results showed that nurses' attitudes towards bereavement care were significant positively correlated with both the hospital policy on bereavement management of (r s=0.60, p<0.001) and the need for education and training in bereavement care (r s=0.59, p<0.001).

Discussion

This paper aimed to explore some major attitudes of nurses working in the OAG ward. This study identified two clusters on the basis of four aspects of attitudes toward bereavement care (i.e., demographic variables and three attitude scores). There were statistically significant differences between the two clusters for most of the demographic variables and the three attitudes scores, except for religious beliefs (see Table 2). The findings suggest that for these two groups of nurses there were differences in the level of attitudes towards bereavement care. Nurses in cluster B were more prepared with the knowledge needed to handle/carry out bereavement care, while nurses in cluster A were `green' in this area and needed more training and support. They also believed that nurses working in the unit could offer comfort and institutional support, and provide information that would help bereaved parents make acceptable plans and decisions for themselves.

Another study finding is that nurses have very positive attitudes towards grief counselling programmes. However, almost all of the respondents viewed their relevant level of grief counselling knowledge and understanding as insufficient. In total, only 25.5% (n=28) had taken courses related to bereavement care. Because they lacked training in grief counselling, they felt inadequately prepared to provide bereavement care independently and confidently. This finding is supported by studies (e.g. McGuinness, 1988; Szgalsky, 1989; Wright, 1991; Rybarik, 1996; Yam et al., 2001) that have found that support and education were necessary to help nurses in their work with the bereaved.

These results highlight a number of areas that need to be addressed by nurse educators, and the outcomes challenge them to emphasize the importance of grief counselling on caring behaviour in the nursing curriculum. Attitudes are influential in determining behaviour (Todman and Jauncey, 1986). It is believed that a greater awareness and knowledge of grief counselling would lead to more sympathy, understanding and practical help from nurses. In this study, nurses in cluster A urgently needed training and institutional support. Recently, Hong Kong's Severe Acute Respiratory Syndrome (SARS) crisis showed the significant impact that a lack of institutional support, including additional training, as well as resource limitations and heavy workload can have as they result in major sources of job stress. In such an environment, bereavement care is necessary, but produces uncertainty, anxiety and often frustration for nurses.

The goal of holistic quality care can only be achieved when nurses' education and training needs are addressed. Otherwise, caring for a family that has experienced a perinatal death creates a crisis situation for nurses. In this study, the attitudes of nurses in Hong Kong towards perinatal bereavement care emphasized their need for increased knowledge and experience, improved communication skills, and greater support from team members. Hong Kong health care professionals may use these findings to improve the support of nurses, to ensure the delivery of sensitive bereavement care in the perinatal setting, and to enhance curricula in schools of nursing. Nursing educators must focus on the affective domain, as well as the cognitive learning needs of nurses when perinatal bereavement is addressed.

Limitation and further studies

In line with most research, this study had a few limitations that affect its outcomes. One potential limitation is that data for this study used a self-report questionnaire, which may cause possible response bias from each responder (Polit and Hungler, 1995). Second, sample sizes chosen in this study were relatively small and were restricted to one hospital only. Thus, the results cannot represent the attitudes of all classes of nurses in Hong Kong. Truly, this study can be seen as a pilot study only. Further study should proceed at other hospitals. Third, respondent honesty (or lack thereof) is a potential source of error in self-report surveys (Siegel et al., 1998; James et al., 1991). Although replies did not include the names or staff IDs of individual nurses, they might have perceived pressure from senior management or the hospital to respond. This might account for the high response rate seen in this study and for the positive attitudes given in their responses. To reduce such errors, face-to-face interviews could be a better method to improve the accuracy of the results but it is more costly in terms of time, manpower, and money (Oei and Zwart, 1986). Further research is needed to explore common feelings and coping behaviours of bereaved parents in caring programmes offered by nurses working in OAG units.