Investigators invited postpartum women from ten OBGYN clinics across Prisma Health – Upstate and Prisma Health – Midlands (in Greenville, and Columbia, South Carolina, respectively) to complete an anonymous survey in the years 2020 and 2021 on their perception of the nutrition counseling provided to them by their obstetric healthcare provider during the course of their pregnancy. The target audience of less than two-week postpartum women was used to ensure recall capture from the entire pregnancy timeline. These clinics served Greenville and Richland, SC counties. Greenville county’s total population was estimated in 2019 to be 523,523 persons, of which 68.0% were white, 18.4% were Black or African American, 9.5% were Hispanic or Latino, 2.8% were Asian or Pacific Islander, and 1.3% identified as another race (13). Richland county’s total population was estimated in 2019 to be 415,759, of which 41.5% were White, 48.7% were Black or African American, 5.3% were Hispanic or Latino, 3.1% were Asian or Pacific Islander, and 1.4% identified as another race (14).
Due to Covid-19 social distancing precautions and restrictions for in-person interviews, investigators sent each participant an email containing a survey invitation. Each participant was invited to participate in a 27-item survey using REDCap software via patient emails obtained from the Prisma Health Office of Patient Experience, Reporting, and Analytics. Surveys were emailed during the timeframe of November of 2020 through May of 2021, for each participant to be completed within two weeks of delivery. The surveys were to be completed within two weeks of each patient’s delivery date in order to ensure the recall period was comprehensive and consisted of the entire gestational period, so that researchers could conclude if counseling had been provided at prenatal encounters during any point of the pregnancy.
This study was approved by the Prisma Health Institutional Review Board. Informed consent was obtained for all participants electronically through the “Invitation to Participate in Research Study,” the first document that appeared upon opening the electronic survey. The first question of the survey was “Are you interested in participating in the study?” and if the participant selected “yes” as their answer, the rest of the study was then available. Potential risks were identified and disclosed, which include psychological distress if the patient realizes that they did not receive certain nutrition recommendations from their healthcare provider, or loss of confidentiality. Participants were informed that participation was voluntary, that they did not have to answer any question they were not comfortable with, and that they could withdraw from participation at any point. Exclusion criteria for participants were women who did not speak English, and women who gave birth to stillborn infants or infants that required Neonatal Intensive Care Unit (NICU) care. Email addresses were securely provided by the Reporting and Analytics manager through the Prisma Health Office of Patient Experience, Analytics, and Reporting based on delivery records from hospitals around Greenville, South Carolina (SC) and Columbia, SC. This and all pertinent study data was maintained on password protected electronic files, accessed through password-protected computers to maintain security and confidentiality. All other individual patient data was collected anonymously via the electronic survey and remained confidential as part of this investigation.
The 27-item, predominately multiple-choice survey was developed and validated according to the process described by Zamanzadeh et al, drawing from nationally accepted guidelines for nutrition during pregnancy (2, 4). The development of the survey involved determining the scope of questioning, generating appropriate items for response, instrument formation, revision and consolidation of instruments, and preliminary team consensus. The process also involved a team of experts from a wide variety of backgrounds in the development and review of the survey. The survey development team included one PhD-trained, professionally board-certified Lifestyle Medicine scientist, one board-certified OBGYN clinician, two second-year medical students, and one senior biostatistician in order to maximize content importance of the survey questions, which ensured the survey questions engaged the participants in topics that were up to date and evidence based (1, 2, 3, 4). Investigators designed the survey to be brief; average time to complete was less than ten minutes, in an effort to respect each participant’s time, keep the participant’s attention, and reduce chance of survey errors due to postpartum fatigue and infant care demands.
Content validity of the survey was established through the following process. Team members proposed topics, and determined an appropriate length and content for a survey. Internal review of this draft survey was performed by the PhD-trained, professionally board-certified Lifestyle Medicine scientist and the board-certified OBGYN clinician, with edits performed. External validity was then completed through pilot testing from external obstetric providers, nursing staff and pregnant women not participating in the study who provided feedback to improve the instrument for optimal patient engagement. This review process ensured the questions were clear to all parties, and that the survey was concise enough to be completed in the estimated time of ten minutes. The survey was designed to be brief in order to respect the time of respondents, reduce recall bias, and maintain the participant’s attention throughout the duration of the survey (15).
Participants completed the survey on a Health Insurance Portability and Accountability Act (HIPAA)-secure, user-friendly, web-based online application that collects and stores data (REDCap, Vanderbilt University). Analyses examined nutrition counseling perception and content, nutrition information obtained by a source other than an obstetric provider, and demographics. Discrete variables are reported as N (%). Discrete variables were stratified by demographics and analyzed using Chi-square or Fisher’s exact test. All analyses are carried out using R statistical software (R Foundation for Statistical Computing, version 4.0.4). A P value of < .05 was considered statistically significant.