The study was a pre-post design using online surveys to evaluate changes in knowledge of eating disorders in the perinatal period after completing an online education module. The study also collected feedback from participants for the further development and improvement of the modules.

Ethical approval was obtained from the Human Research Ethics Committees from both hospital sites (HREC/2022/QGC/90364). Participants were sent a link to the information and consent form, and provided electronic consent by checking a box prior to commencing the first online survey.

Purposive sampling was used to recruit clinicians working in maternity services and perinatal mental health services at two of Queensland’s largest maternity hospitals and peer workers employed at a perinatal wellness centre. Participants included peer workers, midwives, medical specialists, nursing staff, and allied health staff who work in maternity and perinatal mental health services, and screen and/or manage perinatal individuals who may experience an eating disorder.

The eligibility criteria included any clinician or peer worker who screens and/or manages perinatal individuals at one of the participating sites. Participants responded to intranet advertisements, emails and flyers distributed at team meetings and posted in communal areas, using QR codes, and were sent links to the information and consent form, online surveys, and education modules. Pre- and post- surveys were embedded into the online module using an open-source learning management platform (Moodle), and participants were asked to complete all activities in approximately one week. Participants were also given a follow-up survey at three months to determine the sustainability of the education. That aspect of the study is still ongoing. On completion of the study, clinicians were presented with a $30 supermarket voucher.

The online education package was created using the online training tool platform, Articulate [13]. Using this platform, the research team created a learning package that was broken up into five modules. Each module contained case studies, quotes from individuals with lived experience, and a short questionnaire to reinforce learning. A number of interactive elements were embedded into the modules: videos and audio of individuals with lived experience and clinicians; matching words or questions to definitions and answers; clicking on images that reveal additional information; and links that opened new tabs to related information and resources (see Fig. 1). Once completed, a certificate of achievement was generated. Section topics are described in Table 1.

Screenshots from the ‘eating disorders in the peripartum’ online education course

The evaluation of the educational modules was designed using the Kirkpatrick Evaluation Model of Education, which rates four stages of learning: (1) reaction, the subjective experience of the learner; (2) learning, the acquisition of skills, knowledge or confidence; (3) behaviour, the application of learning, and; (4) results, the impact of training [14]. Surveys were created to assess stages one to three of the Kirkpatrick model. Stage Four could not be measured without an evaluation of ED screening, management, and referrals to measure practical changes due to education, which was outside the scope of the present study. A bespoke confidence and knowledge survey, based on the key learnings and objectives of the online learning package, was designed in collaboration with researchers and clinician experts (see Supplementary Files 1 and 2 for surveys). Section A collected demographic information about employment experience and previous ED training. Section B of the survey asked participants to rate their level of agreement with statements regarding their confidence in their ability to screen, manage, identify, and refer a person with an ED on a scale of 1 (strongly disagree) to 7 (strongly agree). Section C of the survey aimed to determine participants’ level of knowledge by asking 20 short answer or multiple-choice questions about ED signs and symptoms, risk factors, screening, and management considerations.

The post-education survey included sections B and C, along with an additional 16 questions asking participants how they felt about the module. This included questions relating to the organisation and usability of the module, whether the module increased their knowledge, whether or not the module contained adequate and relevant information, and suggestions for improvement.

Data analysis was conducted using IBM Statistical Package for Social Sciences (SPSS) version 29.0. Normality was investigated using a visual inspection of P-P plots. Scores for confidence in knowledge and skill were averaged to give a score of 0 to 7 out of 7. Knowledge scores were summed to give a total score out of 20. Reliability of the scales was evaluated by Cronbach’s alpha. The Cronbach’s alpha for the confidence scale was 0.884, and for the knowledge scale was 0.740, indicating good internal consistency. Descriptive statistics were displayed using means and standard deviations. Paired t-tests, with statistical significance set at p < 0.05, were used to compare mean and total survey scores before and after completion of the online education module. Subgroup differences and interaction effects were unable to be investigated due to the limited demographic data collected.

Qualitative data were analysed using thematic content analysis [15]. Respondent feedback was read through and key words were identified and used to create codes, categories, and themes relating to what participants did and did not like about the module. Data was managed in an excel spreadsheet. Codes were refined using a reiterative process of reading and refining codes. Data that fit under each code was represented in a summative manner in the content analysis. This form of analysis was chosen as it limited researcher interpretation of included data. Quotes provided were selected based on their representativeness of findings and of included participant healthcare roles.

A power calculation was conducted prior to the commencement of the study using G*Power software for matched pairs t-tests. As no previous research using these measures exists to provide approximations of effect size, the authors chose to include into sample the size calculation, the conservative estimate of a moderate effect size (d = 0.5), with 90% power, and an alpha error probability of α = 0.05. Given these parameters, a total sample size of at least 44 was required.