We conducted a written, in-person survey among a convenience sample of pregnant women presenting for obstetrical ultrasound at a university-based, tertiary care center in Central North Carolina between March 1, 2011 and May 31, 2011. The ultrasound unit provides care to a racially and socioeconomically diverse group of pregnant women from hospital-based clinics at the university and 14 health departments in the surrounding area. Approximately 300 women per month are referred for ultrasound evaluation. Women were eligible to participate if they were 18 years of age or older and able to provide written informed consent in English. We excluded women presenting for genetic counseling and women with a nonviable pregnancy or a pregnancy complicated by a fetal anomaly. We aimed for a sample size of 100 women based on the assumption of 80% power to detect a 10% difference in our main outcome, which was willingness to participate in a behavioral intervention program via Internet on the computer versus on a mobile phone. The Institutional Review Board of the University of North Carolina approved the study.

The Institute of Medicine access to care model was used as a framework for the study. A questionnaire was developed to examine women’s access to computers and mobile phones according to predisposing (ie, age, race/ethnicity), enabling (ie, income, number of children in the home), and clinical (ie, gestational age) factors [17]. A brief survey was developed following a review of published articles on Internet use among pregnant women. The survey was initially reviewed by 10 pregnant women attending our university-based prenatal clinics. Based on the feedback provided, we made minor edits to the wording of several questions for clarity. The final version consisted of eight items on access to, and use of, computers, mobile phones, and Internet and social media tools, as well as preferences for the delivery of a postpartum weight-loss intervention and demographic information. Each participant was asked the following questions (see Multimedia Appendix 1 for the full survey):

1. “Do you have access to a home phone with Internet? (Yes/No)”

2. “Do you have access to a mobile phone with Internet? (Yes/No)”

3. “Do you have access to a computer for at least 8 hours per week (at home, at work, or in a public setting such as your local library)? (Yes/No)”

4. “How often do you use the following Internet programs (Internet/Web, email, Facebook, blogs, Twitter, chat rooms, Skype) on your computer? (Not at all, Not very often, Often, Very often)”

5. “How often do you use these items (Internet/Web, email, Facebook, blogs, Twitter, chat rooms) on your cell phone? (Not at all, Not very often, Often, Very often)”

6. “How willing are you to participate in a postpartum weight-loss program? (Not at all willing, Not very willing, Fairly willing, Very willing)”

7. “How willing are you to participate in an Internet-based postpartum weight-loss program delivered via the computer? (Not at all willing, Not very willing, Fairly willing, Very willing)”

8. “How willing are you to participate in an Internet-based postpartum weight-loss program on your cell phone? (Not at all willing, Not very willing, Fairly willing, Very willing)”

Women were approached consecutively by trained research staff to ascertain their interest in participating in the study as they presented for their ultrasound appointment. If women expressed an interest, they were escorted to a private conference room within the ultrasound unit where they were screened for eligibility. If they were deemed eligible, informed consent was obtained and they were officially enrolled in the study. The average time to complete the survey was 10 minutes. The research staff reviewed each survey for completeness in the presence of the participant. Participants who completed the survey received a US $10 gift card. Data were entered into a master spreadsheet by a study staff member. As part of our quality control protocol, a second staff member conducted an audit of the data each month to ensure correct entries. There were no missing data.

Maternal demographic and clinical characteristics were summarized using means and standard deviations for continuous variables and numbers with proportions for categorical variables. The proportion of participants with access to computers and mobile phones across predisposing (ie, age, race/ethnicity), enabling (ie, income, number of children in the home), and clinical (ie, gestational age) factors was assessed using chi-square statistics. Given the performance of four tests and following the Bonferroni correction for multiple testing [18], results were considered statistically significant if P<.013 (.05 divided by three). For analyses stratified by race/ethnicity, we recategorized African American, Asian, and Hispanic women into one category, termed nonwhite, due to small numbers.

Willingness to participate in the various types of interventions was defined as those who answered fairly willing or very willing and was compared to the number who answered not at all willing or not very willing. Willingness to participate in a postpartum weight-loss intervention across sociodemographic groups was assessed using Fisher’s exact test. The P values were corrected for multiple testing with a Benjamini-Hochberg correction [19]. To determine which demographic characteristics were associated with women’s willingness to participate in computer-based or mobile phone-based weight-loss interventions, we first performed univariate logistic regression. The models looked at willingness to participate in interventions via computer and mobile phone according to age (continuous), race/ethnicity (nonwhite versus white), income level (≥US $25,000 versus <US $25,000), and number of children in the home (one or more versus none). We then included all the above variables in a multiple logistic regression model to determine if any were predictive of increased willingness to participate in an Internet intervention by computer or by mobile phone. Again, based on a Bonferroni correction [18], results were considered statistically significant if the P value was <.006. Analyses were conducted using Stata version 10 statistical software (StataCorp) and R (R Core Team 2013).

Of the 120 women approached, 110 (91.7%) were eligible for the study and 100 (83.3%) agreed to participate. Of those approached, 10 out of 120 (8.3%) women were not eligible to participate in the study due to a nonviable pregnancy (9/120, 7.5%) or pregnancy complicated by a fetal anomaly (1/120, 0.8%). Of those approached, 10 out of 120 (8.3%) women declined to participate due to a lack of time to complete the survey. The study sample represents approximately 10% of the patients presenting to the ultrasound unit during the 3-month study period.

The mean age of participants was 31.0 (SD 5.1) years (see Table 1) with a range of 22 to 44 years. Of the participants, 61.0% (61/100) were white, 26.0% (26/100) were African American, 7.0% (7/100) were Asian, and 6.0% (6/100) were Hispanic. Of the participants, 45.0% (45/100) reported a yearly household income of more than US $50,000. Median gestational age at the time of the survey was 26.0 weeks ± 9.2 with a range of 7 to 41 weeks. Most participants (64/100, 64.0%) had one or more children in the home. See Table 1 for further demographic information.

Most participants reported access to computers (89/100, 89.0%) and mobile phones (88/100, 88.0%). White women were more likely to report access to both computers and mobile phones compared to nonwhite women (P=.007, see Table 2). There were no statistically significant differences in access to computers or mobile phones by age.

Women with one or more children in the home reported slightly less access to both mobile phones and computers than those with no children in the home, but this difference was not statistically significant (P=.07)—94.0% (94/100) versus 81.0% (81/100), respectively (see Table 3). There were no statistically significant differences in access to computers or mobile phones by income or by number of children in the home.

Of the women surveyed, 82.0% (82/100) were very willing or fairly willing to participate in an online weight-loss intervention program after delivery. Bivariate analysis did not show any statistically significant differences in the willingness versus nonwillingness to participate in an online intervention by age, race, or income categories. Women with no children in the home were more likely to be very willing or fairly willing to participate in a postpartum program compared to women with one or more children in the home (P=.07)—93.0% (93/100) versus 77.0% (77/100), respectively. When asked whether they were willing to participate in a postpartum Internet-based intervention delivered via mobile phone or computer, women were significantly more willing to participate via computer (83/100, 83.0%) compared to by mobile phone (49/100, 49.0%) (P<.001).

To explore the individual contributions of demographic factors on willingness to participate in a computer-based or mobile phone-based intervention, we developed separate logistic regression models for each modality (Table 4). In both bivariate and adjusted analyses, there were no statistically significant findings. Women of different ages, races, income levels, and number of children at home were similarly willing to participate in computer-based or mobile phone-based interventions.

The majority of participants reported that they often, or very often, used the Internet (94/100, 94.0%), email (90/100, 90.0%), and Facebook (59/100, 59.0%) through a computer. While the use of specific technologies among sociodemographic groups was not significantly different, there were several important trends noted. In an analysis stratified by age, women under 30 years of age were more likely to report the use of Twitter and chat rooms compared to women 30 years and older (Table 5). Nonwhite women were more likely to report using a computer to access Facebook (26/39, 67%) or chat rooms (5/39, 13%) compared to their white counterparts—54% (33/61) and 5% (3/61), respectively. While the Internet/Web, email, and Facebook were commonly accessed through computers, a smaller proportion of women reported they often or very often used these technologies via mobile phone—49.0% (49/100), 43.0% (43/100), and 36.0% (36/100), respectively. Women aged 30 and older were less likely to report the use of Facebook often or very often via mobile phone compared to women under age 30.

Women with one or more children in the home were less likely to report the use of the Internet via computer compared to women with no children in the home. Women with one or more children in the home reported less use of a mobile phone to access email, Facebook, and other technologies compared to women with no children in the home. The lowest-income women were less likely to access most Internet technologies via computer and mobile phone than women in higher-income levels, as seen in Table 6.

Investigations among nonpregnant, adult populations suggest that behavioral interventions delivered through computers and mobile phones are effective [4-10]. Some groups of pregnant women frequently access health information via the Internet and may do so in lieu of conversations with health care providers [16,20]. Text4Baby is an example of a mobile phone educational campaign accessed by thousands of pregnant women across the US [21,22]. At least one small trial of an Internet-based intervention with pregnant women improved compliance with medical therapy and physical activity [9]. Health care providers and researchers are interested in scalable computer-based or mHealth interventions for larger populations of pregnant women, but more information about access to, and use of, Internet technologies among pregnant women is needed.

Pregnant women in our study reported broad access to the Internet through computers and mobile phones and use of a variety of Internet technologies, including Internet/Web browsers, email, and Facebook. Although nonwhite women and women with one or more children in the home may have less access to the Internet than white women with no children at home, access to, and use of, the Internet was quite high across all groups. Less access by nonwhite women and women with children in the home may be due, in part, to financial restraints or competing demands on time [23].

The majority of women in the study sample were willing to participate in an Internet-based behavioral intervention via computer (83/100, 83.0%), but not as willing to participate via mobile phone (49/100, 49.0%). A recent US study found that 50% of the pregnant women surveyed used the Internet to access health information. However, the survey was administered online and may not be generalizable to larger groups of women [16]. Although other studies have found that pregnant women use the Internet to access pregnancy and childbirth-related information, guidance on physical activity, and newborn health, the data are largely limited to Caucasian women and international populations [14,15,24,25]. It may be that women use mobile phones to access specific screening or treatment information, but may not be as willing to use mobile phones for ongoing weight-management interventions. We did not ascertain the extent to which women used computers or mobile phones to gather health information in the current study. Future studies that examine women’s access and use of the Internet will need to explore what types of information women are searching for when using computers or mobile phones.

In our study, willingness to participate in a mobile phone intervention was not dependent on age, race, income level, or having a child at home. Some trends were noted which may be important to investigate in a larger sample. Younger women and women with no children at home tended to be more willing to participate in a mobile phone-based intervention. Our findings support the development of Internet-based clinical and educational interventions for expectant mothers. Developing mHealth programs or interventions that can be delivered through computer-based technologies may promote better uptake based on individual patient preferences. For example, only 8% of Dutch pregnant women used an email-based intervention [26]. Additionally, further work is needed to better understand the influence of other children in the home on less access to the Internet and, possibly, lower willingness to engage in computer- or mobile phone-based interventions.

The broad access and use of the Internet by women in our sample is similar to rates reported by a 2011 Pew Research Center report [27,28]. In the Pew report, minority women reported slightly lower access to the Internet compared to their white women counterparts (73% versus 94%, respectively). Our findings differ from the Pew study in that we did not find substantially less access to the Internet among lower-income women compared to higher-income women. This difference may be due to the younger age range of our sample of study participants compared with the national sample used in the Pew study. Our findings suggest less of a “digital divide” based on income levels in younger, reproductive-age women. These findings further emphasize the need for targeted studies to better describe access to, and use of, the Internet and other computer technologies in different patient groups (eg, pregnant women) and geographical areas (eg, Central North Carolina).

Our study has several limitations that deserve attention. First, our descriptive study includes a relatively small sample size. Therefore, there is limited power to detect differences between sociodemographic groups. We found some small differences in access between white and nonwhite women and also between women with and without children at home. The survey was investigator developed and, therefore, had not been extensively tested prior to its use. Our data also relies on self-reporting. However, given the nature and setting of the survey, it is unlikely that women were biased in their responses. Also, we did not provide specific information on the weight-loss intervention, so participants may have responded differently if they had been given additional information. Another limitation is that our survey does not explore women’s health information-seeking behavior on the Internet. Access to Internet technologies and use of such technologies is not the same as using such technologies for health applications. For example, in a recent study of Internet users in the US, adults of lower socioeconomic status were less likely to use the Internet for health information seeking specifically [29]. However, use of the Internet specifically for health information seeking is not a prerequisite to participating in an Internet-based behavioral intervention. It is also not a known indicator of uptake of such interventions.

Because our study was conducted in one tertiary care center, it can be argued that the findings have limited generalizability. However, we recruited women referred for ultrasound examination from a wide range of community settings and economic backgrounds. Nevertheless, these findings should be confirmed in a larger sample of women living in a diverse community and receiving care in multiple tertiary care settings.

This study represents a first step in characterizing access to the Internet among a diverse group of pregnant women in the United States. Larger scale studies across geographical regions are needed to better inform and tailor the development of Web-based interventions to take advantage of this critical period in a woman’s lifespan. Integrating Web-based educational interventions into pregnancy and the postpartum periods may lead to improved maternal and newborn outcomes through improved access and education. Such interventions could change the current paradigm of perinatal care, giving clinicians an opportunity to provide ongoing care to pregnant women between prenatal visits [9]. Internet-based interventions to improve clinical care of pregnant women should also be considered. For example, Internet-based interventions might include feedback and communication between patients and providers about a variety of prenatal disease conditions, screening tests, and shared decision making. Our findings suggest broad access to the Internet through computers and mobile phones for pregnant women across multiple demographic strata. Also, our findings suggest that interventions designed to be delivered using contemporary technologies should be accessible via computers or mobile phones and able to be scaled to meet the needs and preferences of women based on age, race, and the presence of other children in the home that may influence access and sustainable use of the Internet.