A randomised controlled trial of the effectiveness of a breastfeeding training DVD on improving breastfeeding knowledge and confidence among healthcare professionals in China

Intervention

The breastfeeding training DVD developed by the researchers in England was the training intervention for healthcare professionals in the study. It covers four educational sessions, including optimal P & A, the process of a satisfying feed shown by a baby of three months, hand expression demonstrated by a mother, and baby-led feeding demonstrated by an experienced baby of ten months. The total length of the DVD is 14 min 32 s [29]. The Mandarin script of the breastfeeding trainers’ voice-over explaining the learning points of the DVD was developed by the lead author based on the English script version, and agreed with two breastfeeding experts and one professional translator. Two staff in Zhejiang University of Media and Communications recorded and produced the Mandarin version of the breastfeeding training DVD according to the Mandarin script.

Design

The study tests the effect of the Intervention Group (IG) watching the Breastfeeding DVD designed to educate clinicians to facilitate the development of knowledge and confidence in the skills of supporting mothers in P & A and HE compared to a Control Group (CG) of clinicians who watch a DVD about vaginal birth of similar duration on clinicians’ total knowledge of PA and HE (primary outcome) and confidence in P & A and HE (secondary outcomes) in a randomised controlled trial. The baseline measures were administered immediately before watching the DVD, and post intervention measures were taken immediately after watching the DVD. A power analysis was carried out to determine the required sample size in the study based on the a result of a pilot study among Chinese professionals conducted by the lead author [31]. The observed effect size was large (d = 1.05) for the t-test on change in total knowledge scores (primary outcome) between pre- and post-DVD. G*Power was used to calculate the required sample size to detect a condition by time interaction in ANOVA with a medium-to-large effect size (F = 0.30) [32]. In order to achieve 90% power, with significance level set at 5%, a total of 119 participants were required. We sought to invite approximately 200 participants to allow for 40% attrition.

Setting

The study was conducted in Zhejiang Province, China. Participants were recruited from those clinicians about to attend a breastfeeding training course. They are all from three hospitals, including Women’s Hospital, School of Medicine, Zhejiang University, Kaihua county maternal and child health care hospital, and Kaihua county general hospital. The Women’s Hospital, School of Medicine, Zhejiang University is the largest and highest level teaching tertiary hospital for maternal health care in Zhejiang Province. The other two hospitals are secondary county hospitals in Kaihua county. All three hospitals are accredited as BFHI hospitals.

Sample

Qualified health professionals (doctors, nurses and midwives) in Zhejiang Province who work with breastfeeding women were included. Staff who were not directly caring for breastfeeding mothers or volunteer for supporting breastfeeding mothers were excluded.

Procedure

The study was conducted before two breastfeeding training courses during March and June 2014. The researcher obtained the lists of names of all the participants who would take part in the provincial and county breastfeeding training courses. There was a staff member in each hospital who co-ordinated the study and emailed potential participants with an invitation to take part in the study. Those professionals who had no email address or did not reply to the invitation email were invited by telephone. The reasons for declining the invitation were recorded by the staff member. Participants who met the eligibility criteria and agreed to participate in each hospital were randomly allocated by a researcher blind to the two conditions to an intervention group (IG) and a control group (CG) using SPSS software by a computer professional in Zhejiang University. Randomisation was for the whole sample not by site. See Fig. 1.

All the participants received a sealed envelope marked Group 1 (IG) or Group 2 (CG), in which there was information for participants about the study, consent form, pre-DVD training questionnaire and post-DVD training questionnaire, both with a coded number sequence. Participants were not aware until they saw the DVD which group they were randomly allocated to in order to minimize disruption to the training courses, and they were informed they could withdraw consent at any time. The two groups were allocated to different rooms at the same time point and separate researchers explained the aims, methods and use of their data in line with the written information provided, and supervised the signing of consent forms. Pre-DVD training assessments (baseline questionnaire) were carried out for all participants before they watched the DVD. The baseline questionnaires were collected by the researchers immediately after being completed. Any one who withdrew consent to participate was asked to leave the room and wait for the training course. Next, the participants in the intervention group watched the Mandarin version of the breastfeeding training DVD. The control group watched a DVD in Mandarin unrelated to breasteeding about vaginal delivery. Both DVDs are of about 15 min. After watching their respective DVDs, participants in both intervention and control groups completed the post DVD training assessments.

Measures

Analysis

Epidata (3.1) was used for double entry of data and the accuracy of entries was checked. Statistical Package for Social Science (SPSS version 22. 0) was used to analyse the data and the level of statistical significance was set at p < 0.05. Cases were excluded if they were missing data on any variable included in the specific analysis. For statistically significant groups, the strength of association was assessed by effect size to provide the magnitude of the differences between groups [33, 34]. Cronbach’s alpha coefficient was examined and reported the internal consistency for the Breastfeeding Support Self-Efficacy Scale. Chi-square tests compared the differences of each personal and job variable between the intervention and control groups. Mixed 2-way analysis of variance (ANOVA) were conducted to explore the interaction between group and time (pre-post DVD) for each dependent variable (PA knowledge, PA confidence, HE knowledge and HE confidence). Independent-Samples T Test was used to examine the differences in the mean score of knowledge and confidence between the two groups, and before and after viewing the DVD. The changes in total score of knowledge and confidence in the two skills in the intervention group by personal and job type variables was analysed by mixed ANOVA.

Characteristics of participants

Between March and June 2014, a total of 216 names were provided by the staff members at the research sites, including118 professionals participating provincial breastfeeding training course and 98 professionals in county training course. Six professionals were not eligible and were excluded as they worked in an administrative department. Nine staff declined to participate before the study with two staff stating they were not interested, and seven staff said they did not have time to participate in the study. The remaining 201 eligible staff were invited to take part in the study. Five professionals who were late for the breastfeeding training course did not participate in the study with 2 staff of IG and 3 staff of CG. Two clinicians did not complete the post-training questionnaires as they left early. Three post-training questionnaires were excluded during analysis due to half or more uncompleted items. The final sample included 191 healthcare professionals who watched the allocated DVD and completed the pre- and post-DVD training questionnaires: 96 in the IG and 95 in CG. The response rate was 91% (See Fig. 2).

Pre DVD knowledge and confidence scores in IG and CG

At pre DVD baseline, there was no statistical difference in P & A, HE and total knowledge score between IG and CG. There was missing data from 48 participants on the confidence measure at either pre or post -test, so confidence analyses were restricted to 143 cases. At pre-test there was missing data for 25/96 participants in the IG, and 19/95 in the CG, in the post-test there was missing data for 4/96 in the IG and 14/95 in the CG. The total confidence score and P & A confidence score at baseline showed no significant difference between IG and CG. However, the HE confidence score at baseline was significantly higher in the CG compared with IG. This data is available from the authors.Therefore change scores on knowledge and confidence are used in further analyses.

Knowledge scores changes between pre- and post-DVD in IG and CG

Confidence scores changes between pre- and post-DVD in IG and CG

Sub group analyses on change in knowledge scores of the intervention group

Knowledge

The study showed a large range of knowledge scores and poor knowledge level at baseline. The mean scores were less than half of the total possible correct scores at baseline among participants in total. Although clinical staff in BFHI accredited hospitals are expected to be trained in breastfeeding support as soon as they are employed in a clinical role with mothers, this study suggests many women will be exposed to clinical staff with insufficient knowledge of these essential skills. Such results are perhaps to be expected, given the acknowledged shortfalls in implememtaion of BFHI in China referred to above [24, 35].

The study indicates the new DVD training method is effective in improving professionals’ knowledge in each area of P & A, HE. Improvements in mean knowledge scores were not moderated by age, time working with breastfeeding mothers, job type, years since qualification, seniority (job title) and prior to breastfeeding training experience.This finding is similar to a study which assessed breastfeeding knowledge among 51 clinicians in a neonatal unit in England using a similar multiple choice test of knowledge, where personal and job variables similar to those in this study were not associated with differences in objectively assessed knowledge [36].

Further sub-group analysis shows a significantly higher mean knowledge score was found in specialist tertiary hospitals than in the less specialist local hospitals before DVD training. This is consistent with the findings from the study by Wang and colleagues who reported the area of maternal service to be one of the crucial factors influencing breastfeeding duration due to the different support services provided by professionals in different settings, based on 47,843 mother-infant pairs under two years of age in ten provinces of China [37]. In our study, significantly improved scores were achieved in both types of hospital, showing that a short training intervention may contribute to reducing this variation in the provision of breastfeeding support.

Other sources of unequal provision arise from the large economic differences between provinces. The eastern coast of China is more developed and has more high quality education in medicine than those in rural areas [38]. If the results of this study can be achieved in these less prosperous areas, it may help to reduce the inequalities in breastfeeding training between low-income regions and richer areas of China. Further studies could test whether the breastfeeding DVD training achieves the same results when offered for home self study, enabling more clinicians to meet their personal learning objectives.

Confidence in practice

For those who completed the confidence measure, significantly increased total confidence scores were found in the intervention group between pre- and post-training, but not in the control group. Also, there were no significant differences in the change of total confidence scores in each sub-group between pre- and post-training. Our results are limited by the 48 participants who did not complete one of the measures of confidence. Neverthelss, these findings are consistent with other studies that show breastfeeding education interventions for clinicians significantly increase their confidence in practice [39–42]. Downie, Rakic and Juliff reported that the increased confidence among nurses and midwives continued after completing the intervention Lactation Advisor Program (LAP) in Australial [39]. Similarly, in England, midwives and health visitors had increased confidence after BFHI training, which they reported enabled them to be more confident in supporting breastfeeding mothers [43]. The authors of a synthesis of breastfeeding training interventions concluded that proficient skills and increased confidence in helping breastfeeding mothers among healthcare professionals can be improved by training interventions, and this may improve rates of duration and exclusivity of breastfeeding [25]. Further study in China is require to examine the effect of the increased confidence on improving professionals’ practice in supporting mothers’ breastfeeding skills, thereby increasing the breastfeeding duration and rate of exclusive breastfeeding.

Limitations

The retest of the breastfeeding knowledge in the study was conducted immediately after both groups were exposed to the DVD, so only a short term effect has been measured. And no measures have been made in this study of changes in clinical practice and in breastfeeding rates. The large amount of missing data in the confidence scales suggests that the measure may not be understood, and needs further methodological development for use in the Chinese population.