A randomised controlled trial of empowerment training for Chinese abused pregnant women in Hong Kong


Dr A. Tiwari, Department of Nursing Studies, Faculty of Medicine, 4/F, Academic and Administration Block, Faculty of Medicine Building, 21, Sassoon Road, Pokfulam, Hong Kong, China.


Objective  To evaluate the effectiveness of an empowerment intervention in reducing intimate partner violence (IPV) and improving health status.

Design  Randomised controlled trial.

Setting  Antenatal clinic in a public hospital in Hong Kong.

Sample  One hundred and ten Chinese pregnant women with a history of abuse by their intimate partners.

Methods  Women were randomised to the experimental or control group. Experimental group women received empowerment training specially designed for Chinese abused pregnant women while the control group women received standard care for abused women. Data were collected at study entry and six weeks postnatal.

Main outcomes measures  IPV [on the Conflict Tactics Scale (CTS)], health-related quality of life (SF-36) and postnatal depression [Edinburgh Postnatal Depression Scale (EPDS)].

Results  Following the training, the experimental group had significantly higher physical functioning and had significantly improved role limitation due to physical problems and emotional problems. They also reported less psychological (but not sexual) abuse, minor (but not severe) physical violence and had significantly lower postnatal depression scores. However, they reported more bodily pain.

Conclusion  An empowerment intervention specially designed for Chinese abused pregnant women was effective in reducing IPV and improving the health status of the women.


The prevalence of intimate partner violence (IPV) during pregnancy has been assessed in a number of studies.1–3 A large review of the prevalence of abuse during pregnancy in the United States suggested that between 0.9% and 20.1% of pregnant women were abused by their partners.2 A similar prevalence was also reported in international studies.4–7

IPV is a significant public health concern with negative health consequences.8 Abuse during pregnancy threatens the health and safety of the mother and the fetus.9–11 However, despite the call for increased assessment and interventions for IPV in health care settings,8 the screening of women by health professionals has been questioned, partly because of the lack of evidence about the effectiveness of health-service-based interventions.12

A review of the literature has revealed only six studies on the effectiveness of health-service-based interventions for women experiencing domestic violence.13–18 While the studies have provided the much-needed information, they have been criticised because none of them were randomised controlled trials.12 Furthermore, only two of the studies included the amount of violence as a measure of the effectiveness of the intervention and none considered health outcomes.

Providing help to abused pregnant women in Hong Kong, where remnants of Chinese culture prevail despite Western socialisation, is challenging. Not only is there a lack of information about the effectiveness of interventions, but the predominant form of IPV appears to be different to that documented in international studies. Previous local studies revealed a predominance of psychological abuse among the forms of violence reported by Chinese pregnant women with a history of IPV.7,19 As the existing models of health-service-based intervention are primarily designed to target women who have experienced physical and sexual violence, they may not be suitable for Chinese pregnant women. Furthermore, in Chinese society traditional patriarchal values still prevail. It is not uncommon, therefore, for women to submit to their husbands and endure humiliation for the sake of keeping their family together, and IPV may not be taken seriously. As a result, Chinese pregnant women may not voluntarily report partner violence, even to health care professionals, even though pregnancy provides a unique opportunity for the detection of IPV.

Therefore, when making a decision as to what kind of intervention should be used for Chinese abused pregnant women, consideration should be given to the possibility that the women might be unwilling to report their abusive experience, and that their abuse may be predominantly psychological. In this study, an empowerment intervention, which was primarily designed for abused women in the United States,13 was modified with an additional component known as empathic understanding.20 The addition was designed to provide a safe environment for the women to tell their abusive experience and for the negative impact of psychological abuse to be addressed. It was hypothesised that women who received the empowerment intervention (experimental group) would have significantly less IPV and a higher health status than those who received the standard care for abused women (control group).


The Abuse Assessment Screen (AAS) was used to screen potential subjects for abuse. It was developed by the Nursing Research Consortium on Violence and Abuse to determine abuse status and perpetrator within a defined period.21 Respondents who answered ‘yes’ to being physically or emotionally hurt by someone or forced to have sexual activities within the last year were considered abused. In the present study, the AAS was translated into Chinese. We also added ‘emotionally hurt’ to the questions where ‘physically hurt’ appeared, as an acknowledgement of the prevalence of psychological abuse among Chinese pregnant women in Hong Kong.

For the purpose of this study, psychological abuse was defined as the use of threats by the intimate partner to hurt the woman or the use of verbal or non-verbal acts that symbolically hurt her. Physical violence was defined as the use of physical force against the woman by her partner as a means of resolving conflicts. Sexual abuse referred to the partner forcing the woman to have sex when she did not want to.

The form and amount of IPV were assessed using the original Conflict Tactics Scale (CTS), Form R.22 The CTS is a 19-item self-report scale listing 19 behaviours that the respondent's partner might use while in family conflicts. The woman was asked to report on the frequency of certain behaviours by her partner over the past year on a seven-point scale, where 0 = never, 1 = once, 2 = twice, 3 = 3–5 times, 4 = 6–10 times, 5 = 11–20 times and 6 = 20 or more times. The CTS consists of the scales of reasoning, psychological abuse and physical violence. The Chinese translation of the CTS showed satisfactory reliability (α ranged from 0.70 to 0.86).23 For the present study, an extra item was added for the assessment of sexual abuse and the physical violence scale was subdivided into minor and severe. The Cronbach's alpha of the modified CTS ranged from 0.65 to 0.71.

The Short Form Health Survey24 (SF-36) was used to assess abused women's health-related quality of life. The SF-36 consists of 36 items with one measuring health transition and the remaining 35 grouped under eight scales: physical functioning (PF), role limitation due to physical health problems (role-physical; RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role limitation due to emotional health problems (role-emotional; RE) and mental health (MH). The scores of the items in each scale are summated and transformed into a scale score that has a standardised range from 0 to 100. Higher scale scores indicate better health-related quality of life. The Chinese (Hong Kong) version of the SF-36 was validated and showed satisfactory reliability (α ranged from 0.65 to 0.83).25 In the present study, α ranged from 0.69 to 0.74.

The Edinburgh Postnatal Depression Scale26 (EPDS) was used to detect postnatal depression. The Chinese version of the EPDS was validated and had satisfactory psychometric properties, and a cutoff score of 9/10 was recommended for screening depressive illness in a Chinese postnatal population.27

A demographic questionnaire was included to record the woman's age, parity, marital status, number of years married, number of years living in Hong Kong, occupation of the woman and her partner, family income, smoking, alcohol and drug status and whether the pregnancy was planned.

The study was undertaken between May 2002 and July 2003, at a public hospital in Hong Kong. In a private room and without the presence of a partner, all pregnant women over 18 and less than 30 weeks of gestation attending their first antenatal appointment were informed about the study. For those who agreed to participate, written consent was obtained. Screening for abuse was undertaken using the modified AAS. Those who screened negative were thanked for their participation and no further contact was made. Women whose abusers were not their male partners were excluded from the study.

Women identified as abused and whose perpetrator was an intimate partner were enrolled in the study. At the time of enrolment, the modified CTS, SF-36 and demographic questionnaire were administered. Telephone numbers for contacting the woman for the follow up interview were obtained. A statement was also stamped on the front cover of the case notes indicating that the woman was a participant in a clinical trial and the research nurse should be contacted when she was admitted for delivery. The name and telephone number of the research nurse were displayed on the front cover.

At this point, the woman was randomly assigned to either the experimental or control group. The randomisation allocation schedule was generated by computer and concealed in consecutively numbered, sealed envelopes by a researcher not involved in the study. The sequence was concealed until interventions were assigned. Women assigned to the experimental group immediately received the intervention while those assigned to the control group were given the standard care for abused women.

The intervention was based on an empowerment protocol developed by Parker et al.13 designed to enhance abused women's independence and control. It consisted of advice in the areas of safety, choice making and problem solving. In the current study, some of the items in the protocol were modified to ensure cultural congruence. For example, as domestic violence is seen as a shame to the family and one would not normally disclose it to someone outside the family, we have suggested that the woman might wish to establish a code as a safety precaution with ‘trusted’ friends and neighbours (instead of ‘friends and neighbours’ as in the original protocol). In addition, as gambling is popular in Hong Kong and heavy losses at the races or similar activities could be an excuse for the abuser to take out his anger on his partner and/or children, we have included ‘heavy loss in gambling’ as a risk factor.

A component of empathic understanding, derived from Roger's client-centred therapy,20 was also added to the empowerment protocol. Empathic understanding emphasised the need to take in and accept the woman's perceptions and feelings. It was designed to help women positively value themselves and their own feelings. This was important because abusive experiences are likely to be ignored by others, particularly if the abuse is psychological and the deleterious effects of partner violence on self-esteem are not recognised.

The intervention was administered to each woman in the experimental group at entry into the study by a senior research assistant who was a midwife with a master's degree in counselling. The one-to-one interview lasted about 30 minutes. At the conclusion of the intervention, the woman was given a brochure reinforcing the information provided.

Standard care for abused women was provided to the control group women. This consisted of a wallet-sized card with information on community resources for abused women, which included shelter hotlines, law enforcement, social services and non-government organisations.

At six weeks post-delivery, an experienced research nurse who was not involved in the pre-intervention assessment or the intervention contacted the woman by telephone. If the abuser was present at the time of the call, the research nurse, who was also a midwife in the unit where the woman delivered her baby, would disguise the call as one to follow up the progress of the baby, and signal to the woman that she would call back at a time when the abuser would not be present.

For consistency, the same research nurse conducted all the postnatal telephone interviews. During the telephone interview, the modified CTS, SF-36 and the EPDS were administered. For this interview, the CTS focussed on the partner's behaviours in dealing with conflicts with the woman since the last interview at enrolment. To ensure allocation concealment, the research nurse was instructed to solicit the woman's satisfaction with the intervention/standard care after the completion of all the questionnaires. The concealment appeared to be successful as none of the women revealed what they received until they were asked at the end of the telephone interview.

As far as the researchers are aware, the Chinese translated versions of the CTS, SF-36 and EPDS have not been validated for telephone administration. Therefore, when administering the questionnaires over the phone, the research nurse took extra care to ensure that the woman understood each of the items and allowed the woman to answer at her own pace. We were mindful that by conducting the initial interview face-to-face and then assessing the final outcomes by telephone, there would be inconsistency in the method of data collection. We had hoped to conduct face-to-face interviews for the follow up but this was not possible as only some of the women would return for their postnatal follow up at our hospital. We tried to minimise the inconsistency by having the items of the questionnaires read out to the respondents at both recruitment and follow up interviews.

The investigators made a number of deliberate efforts to minimise the possibility of contamination of the groups. Firstly, the two researchers responsible for the collection and analysis of data were not involved in the design of the study and did not know about the hypotheses. Second, we tried to minimise clinic waiting time for the participants so that they had little time to meet each other. Finally, the women's allocation was not recorded in the hospital patient file. Therefore, other health workers were not able to detect from the file whether the woman belonged to the experimental or control group.

The primary outcome measure was IPV. The secondary outcome measures were health-related quality of life and postnatal depression. The sample size was calculated based on the primary comparisons of the five subscales of the CTS between the experimental and control groups. The standard deviation of the change was estimated as 6 from previous data. Taking four units difference between the two groups as clinically important, 55 women were required in each group in order to have 80% power by an independent t test. Therefore, a total of 110 women were recruited into the study. The four units difference corresponds to 10% difference out of the possible difference of ±20 between the pre- and post-test.

Data analysis was conducted using the Statistical Package for Social Sciences (SPSS version 11.0, Chicago, Illinois, USA). To compare the continuous and categorical characteristics of the experimental and control groups, the relative risk and mean difference [both with 95% confidence intervals (CI)] were used. Those where the 95% CI for relative risk fell outside unity or for mean difference fell outside zero were considered statistically significant. For participants lost to follow up, missing values were replaced by the pre-test observations.

The institutional review board of the hospital where the research was conducted approved the study, and all participants provided written informed consent.


Figure 1 shows the progress through the phases of the study. Two thousand three hundred and sixty-one women were interviewed during their first antenatal visit and 117 (5%) were identified as abused. One hundred and ten women consented to randomisation and were enrolled in the study. Four women in the experimental group were lost to follow up, so a total of 106 women completed the study.

Figure 1.

Flow diagram of the progress through the phases of a randomised controlled trial.

The demographic data on the study participants are shown in Table 1 and the types of abuse, CTS and SF-36 scores for the experimental and control groups at recruitment are shown in Table 2. All the women in the study continued to stay with their partners throughout the research study including the period between the pre- and post-test.

Table 1.  Demographic characteristics of participants. Values are presented as n (%) or mean [SD].
 Experimental group (n= 55)Control group (n= 55)
  • a

    Average national monthly wages for non-production workers HK$11,717, for production workers HK$10,199 (Census & Statistics Department, HKSAR Government, 2002).

Age (years)30 [5.1]31 [5.2]
Nulliparous24 (44)25 (46)
Marital status
Married48 (87)52 (95)
Single5 (9.1)3 (5.5)
Divorced2 (3.6)0
Number of years married4.3 [4.0]5.4 [4.3]
Number of years living in Hong Kong19 [13]18 [14]
Woman in paid job25 (46)17 (31)
Partner in paid job43 (78)36 (66)
Family income (HK$)a
<500002 (3.6)
5000–10,0007 (13)8 (15)
10,001–15,0008 (15)14 (26)
15,001–20,0007 (13)7 (13)
20,001–25,00013 (24)5 (9.1)
25,001–30,0003 (5.5)3 (5.5)
>30,00014 (26)12 (22)
Planned pregnancy26 (47)28 (51)
Smoker6 (11)6 (11)
Alcohol user4 (7.3)2 (3.6)
Drug abuser1 (1.8)2 (3.6)
Table 2.  Baseline types of abuse, CTS and SF-36 mean scores for trial participants at recruitment. Values are presented as n (%) or mean [SD].
ScaleExperimental group (n= 55)Control group (n= 55)
  1. For the CTS, the possible range is from 0 to 6 for each of the scales. Higher scores indicate higher levels of IPV. For the SF-36, the range of possible scores is from 0 to 100 with higher scores indicating better health-related quality of life.

Types of abuse
Psychological abuse32 (58)35 (64)
Physical violence23 (42)20 (36)
Sexual abuse4 (7.4)8 (15)
Psychological abuse3.1 [2.8]2.8 [2.5]
Minor physical violence1.3 [3.0]0.7 [1.6]
Severe physical violence0.82 [3.0]0.35 [1.2]
Sexual abuse0.16 [0.63]0.18 [0.80]
Physical functioning80 [17]80 [14]
Role-physical43 [43]35 [35]
Bodily pain22 [21]22 [20]
General health60 [12]60 [15]
Vitality65 [12]65 [16]
Social functioning48 [19]45 [18]
Role-emotional53 [33]53 [40]
Mental health60 [13]64 [10]

Table 3 contains the CTS and SF-36 mean scores for the experimental and control groups after the intervention. Following the intervention, the experimental group had significantly higher physical functioning and significantly improved scores on the role limitation measures for both physical and emotional problems. They also reported significantly less psychological (but not sexual) abuse and minor (but not severe) physical violence. There was however more bodily pain reported in this group.

Table 3.  The CTS and SF-36 mean scores for both groups after the intervention. For details of the scales please see Table 2.
ScalesExperimental group, mean [SD]Control group, mean [SD]Mean difference (95% CI)
  • *

    Significant at the 5% level.

Psychological abuse0.79 [1.0]1.6 [2.2]−1.1 (−2.2, −0.04)*
Minor physical violence0.05 [0.4]0.51 [1.3]−1.0 (−1.8, −0.17)*
Severe physical violence0.25 [1.2]0.17 [0.54]0.08 (−0.26, 0.42)
Sexual abuse0.03 [0.11]0.12 [0.55]−0.07 (−0.30, 0.16)
Physical functioning90 [15]80 [20]10 (2.5, 18)*
Role-physical73 [38]45 [40]19 (1.5, 37)*
Bodily pain14 [19]27 [25]−13 (−23, −2.2)*
General health50 [7.0]50 [7.5]−1.3 (−6.4, 3.9)
Vitality55 [11]55 [13]0.45 (−5.4, 6.3)
Social functioning49 [12]43 [15]3.1 (−4.3, 11)
Role-emotional77 [37]47 [43]28 (9.0, 47)*
Mental health60 [2.4]64 [10]0.28 (−4.4, 5.0)

Significantly fewer women in the intervention group had postnatal depression at follow up as assessed by the EPDS. Twenty-five women from the control group had EPDS scores of ≥10 compared with 9 from the experimental group (relative risk 0.36, 0.15–0.88).

The analysis was conducted on an ‘intention-to-treat’ basis. In addition, a per-protocol analysis was conducted on the 106 participants whom we were able to follow up. We found no differences when comparing the results of the two analyses. Thus, attrition did not appear to have caused any effect on the results.

We monitored the occurrence of adverse events carefully, as enhancing abused women's awareness of their abusive relationship could potentially increase their danger (e.g. increased assertiveness from a woman may lead to a violent backlash from the abuser). In the follow up telephone interview, we asked the women to recount if the frequency of violence had increased since the last interview, and if so, whether it was the result of their taking part in this study. No adverse events were reported by the women in the experimental or control group as a result of receiving the intervention or standard care.


This report describes the use of an empowerment intervention in a group of Chinese abused pregnant women in Hong Kong. We found that after receiving the intervention, the women reported significantly less psychological abuse, less minor physical violence and an improved health status compared with those who received the standard care for abused women.

This is one of the few studies conducted to date to evaluate the effectiveness of health-service-based interventions for abused women and the first to examine the appropriateness of an empowerment intervention for Chinese pregnant women with a history of partner violence. The results suggest that the intervention was appropriate for Chinese abused pregnant women as evidenced by the reduction of partner violence and enhancement of health status.

The study from where the current intervention was adapted reported a reduction of violence following the intervention in a group of abused women in the United States.13 The present study also found a reduction of violence after the implementation of the modified intervention in a group of Chinese women. Our experience suggests that the component of empathic understanding was important as it offered, to most of the women, the first real chance to tell their problems to someone, and to do so without the fear of being ignored or ridiculed. Many of them were relieved to hear that they were not alone and that the problem was not their fault. The ability to reflect on their own feelings may also have helped the women to identify and explore their distorted perceptions about IPV. Even for those whose abuse was psychological and did not involve any physical violence, the knowledge of predicting risk, making a safety plan, choosing among the options and locating appropriate support was found to be reassuring, positive and essential. This was probably because many of these women were fearful about the escalation of psychological abuse to a more serious level of physical violence. Thus, the original intervention with its components of safety, choice making and problem solving appeared to suit the needs of Chinese abused pregnant women as well. Future studies may further explore the effectiveness of the different components of the empowerment intervention for the various forms of IPV.

Questions may be raised about the value of a 30-minute intervention for abused women. From our experience in implementing the intervention, we have learned that if we offer the woman something that she needs (information and support), which was previously denied, even a short intervention was valuable. Many of the women in this study reported being ignored and even ridiculed when they tried to discuss their problem of partner violence with others previously. The opportunity to talk about their problems, to be listened to and to have their feelings acknowledged had a cathartic effect as bottled-up tension was released. This was particularly the case when psychological abuse was the only form of partner violence reported. Apparently, when the woman was not physically hurt, her suffering was even more likely to be ignored by others.

This study revealed the predominance of psychological abuse among the forms of partner violence reported by Chinese pregnant women, as did previous studies.7,19 As studies have shown that the effects of psychological abuse are often as deleterious as those of physical violence,28 the need for intervention is obvious. Providing interventions to women who experience psychological abuse, however, requires careful attention as existing screening tools and interventions are primarily directed at women who experience physical and sexual abuse. The appropriateness of existing interventions in addressing psychological abuse should be addressed in further studies.

A previous study hypothesised that as the nature of violence was almost entirely psychological among Chinese pregnant women, the impact would be on their psychological health.19 It further confirmed the hypothesis by the identification of significantly higher EPDS scores in the abused group of women. This study found that nearly one-third of the abused women had EPDS scores of ≥10, and most of them came from the control group. This finding not only justifies the need for intervention for our abused pregnant women but also the need for vigilance for signs of partner violence when a high EPDS score is detected.

The women's willingness to report their relationship problems to a stranger is noteworthy. It is often said that partner violence is considered a family shame in the Chinese culture and not disclosed to outsiders.29 However, in the present study, we have found that the women were willing to talk to our research nurse and accept professional help from her. This has implications for health care professionals. If screening and intervention for IPV is conducted by properly prepared personnel, using appropriately designed tools, eliciting the information from Chinese abused women is possible.

This study has several limitations. Firstly, it should be acknowledged that although the AAS is a validated screening tool in English, it may not be in the adapted and translated version as used in the current study. We recommend that validation should be undertaken in future studies. Another limitation is the CTS. Although the CTS is a widely used instrument for research on intrafamily violence, it has its critics. One of the most frequent criticisms of the CTS is that it counts acts of violence and ignores the context within which the violence occurs.30 Another criticism is that it relies on self-reports, which are subject to memory errors and also to a variety of conscious and unconscious distortions of which is reported.30 Furthermore, in the present study, we used the original CTS instead of the revised CTS (CTS2). The latter has the advantage of enhanced content validity and reliability, increased clarity and specificity and additional items for the psychological aggression scale.31 As psychological abuse has been found to be the most prevalent form of violence among Chinese abused women, the CTS2 would probably be a more suitable tool for this study. We chose the original CTS as it contained fewer items and required less time for administration, the latter being an important consideration for a busy antenatal clinic located in a very confined space. Future studies should use the CTS2 and allow more time for administration with a longer period for the recruitment of subjects. The concurrent use of another measure of psychological aggression, for example, the Index of Psychological Abuse, could also be considered.32

The small sample size, single site and the focus of this study on a group of Chinese women limit the generalisability of the findings. Future studies should consider using a larger and more diverse population of women. Finally, although the women's partners will have played a key role in the abusive relationship, we did not involve them in the study. Thus, the reported outcomes only reflect the women's effort in reducing partner violence without taking into account the actions of the abuser. Future studies could also consider including men undergoing voluntary programmes for abusers.


This study found that an empowerment intervention specially designed for Chinese abused pregnant women can significantly reduce IPV and improve health status.


This study was supported by a Low Budget High Impact Research Grant awarded by the University of Hong Kong. The researchers would like to thank Professor Farideh Salili and Dr Daniel Fong for their expert guidance.

Accepted 3 May 2005