Parents of a new baby can access a wealth of information and support from the internet through multiple electronic devices [1-3]. There is evidence to suggest, however, that web-based advice is not always evidence-based, even in the critical area of infant nutrition [4]. Smartphone ownership has expanded worldwide, with 81% of Australian adults and 97% of Australians aged 18 to 34 years owning a smartphone in 2018 [5]. Recent national data suggest that 46% of Australian adults accessed the internet for health information [6]. In 2017, 84% of Australian adults used mobile phones to access the internet, exceeding access through laptop computers (69%) and desktop computers (54%) [7]. Furthermore, smartphone ownership has now surpassed the ownership of desktop and laptop computers [8].

The proliferation of web-based health information sources is reflected by the growing literature for health care professionals discussing and advising the use of new technology [9-12]. Studies have shown that parents and pregnant women trusted hospital, government, and university websites as accurate, regulated, useful, and current sources of pregnancy and parenting information [13,14]. Parents in a video education study on introducing solid foods preferred the internet as a source of information for infant nutrition and felt that public authorities were important information providers [15].

Governments and nonprofit organizations have developed smartphone apps to promote and enhance breastfeeding [16-20]. A recent content analysis of social support in 31 breastfeeding apps found that the most common topics were managing breastfeeding problems (informational support) and locating where to express or breastfeed in public (instrumental support) [21]. Increasingly, clinical trials have assessed mobile health (mHealth) and smartphone apps as interventions to promote and support breastfeeding among mothers and their partners [22-24] through text messaging, goal setting, access to information and videos, provision of online support groups, and troubleshooting breastfeeding difficulties. A review by Tang et al [25] on digital interventions that support breastfeeding found that client communication systems to communicate breastfeeding information, facilitate communication, and provide on-demand information services through text messages, phone calls, email, smartphone apps, and websites may improve breastfeeding adherence.

In Australia, government health services use mHealth apps to support routine child and family health nursing practice in fields such as child literacy and development [26,27], immunization [28], and safe infant sleeping [29]. An early childhood obesity prevention trial will test an app developed for parents and caregivers [30] to be integrated into an Australian statewide pregnancy coaching service [31]. Clearly, apps used as part of routine care must meet practice standards for providing understandable, reliable, current, and evidence-based information independent of commercial associations. This highlights the need for evidence-based, well-developed, and updated apps.

Despite the proliferation of apps and their increasing popularity with parents, their quality may not have kept pace with their quantity. Earlier research on infant feeding apps and websites available in Australia found that 78% of the apps were of poor quality, with deficits in the breadth and completeness of information, author credibility, and readability [4]. A similar review of infant feeding apps in China found that most apps advertised infant formula and parenting products and rated poorly on the availability of information, author credibility, and transparency in disclosing advertising policy, app ownership, and app sponsorship [32]. A recent review of mHealth apps for parents of infants in neonatal intensive care units found that smartphone apps were functional but had low quality and credibility, with only 2 apps rated good by nurse and information scientist reviewers and only 5 apps deemed trustworthy [33].

Since the 2015 study [4], the Australian ownership of smartphones has increased from 64% to 81%, and the proportion of users accessing the internet through mobile phones has increased from 42% to 84% [7,34]. In 2017, there were 325,000 health, fitness, and medical apps available [35]. The increasing interest in infant feeding and physical activity apps indicated by popular search queries in the Google search engine (Multimedia Appendix 1) demonstrates the need to systematically assess and update the current smartphone app landscape. Furthermore, there is a continual turnover of smartphone apps, with several apps from the 2015 study [4] subsequently removed from distribution.

Given the rapid expansion of digital technology into the realm of child health, this study aimed to evaluate the quality of information on infant nutrition and physical activity currently available to Australian parents via smartphone apps. It updates and expands the 2015 systematic assessment, examining the comprehensibility, suitability, and readability of information in free and paid apps available in Australia.

This study used systematic methods to identify, select, and evaluate infant feeding and activity apps that were available in Australia between August 2018 and January 2019. It assessed aspects of their quality and utility using validated and purpose-specific instruments to replicate and update an earlier study [4]. Full details of the methods are given in Multimedia Appendix 2, and evaluation tools are described in Multimedia Appendix 3.

We undertook interrater reliability (IRR) testing with 2 or more reviewers assessing at least 20% of the selected apps using all rating tools. We tested apps that were available on both iOS and Android platforms. Discrepancies were discussed until reviewers reached a consensus on their final ratings.

IRR was calculated for the readability scores, MARS scores, SAM, and the evaluation of information content using Krippendorff α (α), which is appropriate when there are missing or incomplete data. Using Krippendorff standards for data reliability, .667≤α<.80 was accepted as tentatively reliable, and α≥.80 was accepted as reliable.

Statistical analyses were performed using IBM SPSS Statistics for Windows, version 25.0 (IBM Corporation). Descriptive results on the ratings from the various instruments are reported as median (IQR). We calculated the correlations among different measures of readability using the Spearman rank-order correlation. We also compared the 5-point rating scores in the MARS tool with the user ratings of the apps presented in the Apple App Store and Google Play Store using the Pearson correlation coefficient. Krippendorff α was calculated using the KALPHA macro for IBM SPSS [51]. Significant values were indicated at P<.05.

This study did not require ethics approval.

This study evaluated 47 apps on infant feeding and physical activity and found that the information content within these apps was largely poor for coverage and depth of information presented. This study updated a systematic assessment of 46 apps available from 2013 to 2014, which similarly found poor quality information for parents [4]. Although the quality of apps, rated through the MARS, improved since the previous study, the credibility of advice did not improve; of the 59 evaluations, almost two-thirds reported poor coverage of information, and over half the app evaluations reported incorrect or incomplete information on the topics addressed.

Reviewers identified information contrary to the Australian guidelines on infant feeding [40] and physical activity [41] in 21 apps. Many of these apps were developed outside of Australia, that is, in America, the United Kingdom, and the European Union, where the official guidelines may be different from those in Australia. Although 1 app with incorrect information was developed in Australia and involved health care professionals, it was developed by a medical device company whose promotion of its breast pumps likely interfered with the provision of correct breastfeeding information.

The results using the MARS tool were largely positive. In this systematic assessment, the apps were rated as moderately engaging, functional, and visually appealing, and these apps clearly reported whether the information came from credible sources. One strength of the MARS tool is that it provides a multidimensional overview of app characteristics, all of which contribute to the user’s experience of an app and the capacity to learn from it. In this case, most apps rated well on engagement, functionality, and aesthetics. However, in MARS, the quality and quantity of the information content constitute only 2 out of the 7 domains that contribute to the final score. Therefore, the information items were inadequate to reflect the importance of the content in this context. In other words, the MARS tool gave relatively little weight to the quality and accuracy of the information provided to consumers about infant nutrition and activity.

Therefore, although many apps were qualitatively evaluated as good to excellent on the objective and modified MARS score, the information contained within these apps was poor. This is reflected in the subjective scores for the apps, with a median of 2.50 and most apps falling between inadequate and acceptable. Reviewers reported that nearly half of all apps were poor or inadequate for recommending to others.

Brown et al [55,56] systematically reviewed the content of pregnancy apps against the Australian national recommendations on healthy eating for pregnancy. Similar to this study, they found that the apps were of moderate quality (mean objective MARS score was 3.05 [SD 0.66] and 3.52 [SD 0.58] for iOS and Android apps, respectively), with the functionality scale (mean 3.32 [SD 0.66]and 4.06 [SD 0.67], for iOS and Android apps, respectively) rated highest. These reviews also found incorrect and potentially harmful information in 3 apps about avoiding alcohol during pregnancy, fasting, restricting dairy products, and using dairy alternatives to meet calcium requirements.

Similarly, a review by Richardson et al [33] on apps for parents of infants in neonatal intensive care units highlighted the inconsistencies in quality ratings according to the method used. They identified a discrepancy between high app quality ratings using the MARS tool and those using the Trust It or Trash It tool [57] for evaluating clearly reported sources of health information.

Although most apps were rated suitable for their potential users on many dimensions, they scored less on readability and cultural appropriateness. Using the readability tools, only a handful of apps met the Australian government’s requirement of year 4 reading level (3-5 apps, depending on what readability tool was used), although a majority met the less stringent South Australian government’s requirement of year 8 level. However, the reviewers’ SAM ratings suggested that most apps used a suitable vocabulary and appropriate style and ordering.

The app market is highly dynamic, and this study captured a cross-sectional snapshot of infant feeding and play apps available at one time point. Several apps in the previous study [4] were not available for download in 2018 and 2019. Even in this study, some apps had been removed from the marketplace by the time of writing, demonstrating the volatility of this information source. This changing availability makes the process of systematically reviewing apps challenging and potentially limited if key resources are not available during the selection phase.

Changes in the app search engines impacted the search process. The removal of the iOS App Store feature on iTunes in September 2017 [37] meant that the reviewers could only search using an iPhone instead of a desktop computer and prevented reliable double-checking of apps found for screening. As with other mHealth reviews conducted after September 2017 [56], double-checking of apps available in the App Store was conducted on Fnd, a web-based search tool [58], which reduces the flexibility of consumers to search for apps on desktop and synchronize the app download to their smartphone.

Search optimization, which refers to optimizing smartphone apps for visibility in search engines and browsing [59], also affected the types of apps found. To reflect the external validity of apps found by users searching through a smartphone app search engine, the authors did not include handsearching or searching for apps outside of the app search engine. Oversaturation of search engines with apps that are malware, spam, counterfeit, or contain copyrighted and farmed content [60-62] results in excessive number of apps; over 5000 apps were found during screening (Figure 1). This affects the apps found through Google Play, as search results are limited to 250 apps per search; therefore, if irrelevant apps are found in keyword searches and listed in the first 250 apps found, this is a nonoptimal search strategy, and relevant apps will be displaced from the search and not included in the analysis. This excessive number of apps caused relevant apps to be displaced by irrelevant apps; for example, a nongovernment organization–developed app on tummy time (Red Nose Safe Sleeping) was available on both search engines, but poor search optimization for keywords resulted in the app not being found through Google Play searches.

This study used keywords in English and evaluated apps available with English as the main language option. We acknowledge the multicultural background of the Australian population and the need for apps suitable for culturally and linguistically diverse parents [63], and a limitation of our study may exclude app users without English language proficiency (approximately 3.5% of Australians aged 15-49 years in 2016, data from Australian Bureau of Statistics [64]). Of the 102 relevant apps screened (Figure 1), only 3 were excluded for not being available in English. It is likely, however, that non-English–speaking parents may search for non-English sites. Zhao et al [32] used the 360 App Store and Chinese language keywords to conduct similar research, focusing on apps used by non-English–speaking parents. This shows the potential for this approach to include different app marketplaces and language search terms to explore apps available for different groups of parents.

The IRR results (Multimedia Appendix 4) showed considerable discrepancies on some instruments. There was low agreement on readability scores, which may reflect the different passages of text selected for evaluation by the pairs of reviewers. This is likely as reading grade level was calculated using objective tools, rather than individual reviewer assessment.

Although there was acceptable agreement on the modified MARS scores between the pairs of reviewers, there was far less agreement on the subjective MARS items. This difference in scoring may reflect the reviewer’s perspectives (a dietitian researcher’s evaluation compared with a child and family health nurse clinician’s evaluation), affecting their likelihood of recommending an app and also their evaluation of its suitability for consumers using the SAM tool.

There was also mixed assessment on the coverage and depth of information on infant feeding, introduction to solids, and infant activity subtopics. Although reviewers evaluated content against the same government guidelines, they may have varied with regard to stringency in certain subtopics, such as deciding whether infants with appropriate signs of readiness can start solid foods between 4 and 6 months or if exclusive breastfeeding for 6 months was imperative [65]. Similarly, reviewers varied on how much information constituted complete or partially complete advice, affecting IRR.

In replicating an earlier study of infant feeding and activity apps, we were able to compare the quality and content of apps available during 2018 to 2019 and those available 5 years earlier. The increased access to smartphones and the utilization of app-based health information among the Australian population over that time suggest that this study is highly relevant and timely. However, notwithstanding the increased suitability and functionality of these apps (measured with the SAM and MARS tools, respectively), the quality and accuracy of much of the information in apps have not greatly improved over time.

Since the 2015 study, this study identified growth in apps that are from reputable sources such as the government, universities, and health professionals: only 2 apps with university endorsements were found in the previous study [4] compared with 14 apps with university or government development or affiliation in this study. In addition, 3 apps were also used in research [52-54]. This finding indicates a positive transition of trustworthy sources that leverage the increased usage of technology and offer credible information to a wider population. This study also indicated that 9 apps were available in languages other than English, compared with none in the previous study. This demonstrates improved multilingual resources for a growing culturally diverse population.

This study was more comprehensive in that it looked at smartphone apps in Android and iOS, both paid and free to access. However, the previous study also looked at websites available on this topic, which might still be a widely used tool given the widespread use of Dr Google to search for pregnancy, birthing, and parenting information, which may not be accurate, credible, reliable, or safe [3,13,66-68].

Despite an increased proportion of apps published by reputable sources, this study also found relatively few good quality apps available. We reiterate the earlier recommendation [5] to establish a certified endorsement for apps similar to the Health on the Net Foundation Code of Conduct used on the website. This code of conduct is used to standardize the reliability of medical and health information available on the World Wide Web [69] and encourages website developers to maintain the quality standards of the organization. The 2015 study found that websites that subscribed to this code of conduct certificate had higher quality scores [4].

Only 2 apps (WebMD Baby and Pregnancy and Baby Tracker [formerly What to Expect]) were included across both studies, highlighting that the app marketplace changes continuously. A potential reason for the short shelf life of some apps could be the high maintenance cost required to keep the apps updated with the evolving smartphone operating systems. Intervention studies of health apps have reported technical issues with the implementation of their app [70], such as updates in the operating system or app impeding participant access [71-74]. This indicates that app functioning requires ongoing maintenance. This might be a challenge for apps that are developed with limited funding from universities or governments compared with commercial companies that often have higher budgets. Further research is required to explore factors that impact the shelf life of health-related apps and to develop suggestions on sustainable ways to leverage technology to share health-related information.

The use of the MARS tool, specifically developed for app evaluation, is a strength of this study. However, the MARS tool had not been published at the time of the 2015 study [4]. The original study adapted the Quality Component Scoring System to assess apps; however, this tool was originally developed to evaluate the quality of medical websites [75], and not all items were appropriate for apps.

Although the national infant feeding guidelines in Australia have remained consistent since the previous study [40], the uptake of this information has not improved. Research indicates that gaps exist in the current infant feeding practices. Begley et al’s [76] research with mothers, using focus groups, in Western Australia found that less than half of the participants had heard of the Australian Infant Feeding Guidelines or were aware that the recommended age for the introduction of solid foods was around 6 months; many participants believed that the guidelines were based on opinion rather than scientific research. A survey of mother-infant dyads in Western Australia and South Australia during 2010 to 2011 found that the feeding behaviors of participants fell short of Australian feeding guidelines, where although 93% of mothers initiated breastfeeding, only 42% of infants were breastfed to 6 months, and 97% of infants received solid food by 6 months [77].

It is well established that early childhood experiences have a significant impact on optimal child development [40]. Child and family health nurses in the community play an integral role in monitoring children’s growth and development and providing guidance and support to parents. Child and family health nurses and lactation consultants work in an environment that has limited staffing and financial resources.

Many parents live in isolation from extended families and turn to social networking sites for advice from peers for health-related information. Conflicting advice and lack of continuity of care from health professionals often adds to their confusion about how they should care for their children [78,79]. Increasingly, parents require direction and guidance to seek evidence-based educational resources. Parents seek information that is easily accessible and affordable. Information provided during consultation in the child and family health nurse clinics often dispels the parents’ concerns, but parents may be unable to absorb all the information at one time and often require educational resources that they can refer back to once they have gone home. Appropriate internet websites and smartphone apps are key to meeting this need.

Parents, particularly first-time parents, are bombarded with information from a variety of sources, especially from websites and smartphone apps. Although information evaluation tools supported by librarians, academics, and the government [57,80,81] for critically assessing health information quality can support the health and digital literacy of parents, these tools may not be widely known to the layperson. Similarly, clinicians who regularly support new parents often receive requests for advice about apps and cannot confidently provide a recommendation if they are unaware of app evaluation tools [82] or have insufficient time to evaluate apps. One way to overcome this challenge is to establish a trusted app or similar logo (logo similar to that of the Health on the Net Foundation Code of Conduct’s logo) or a repository of approved apps, such as the United Kingdom’s National Health Service Apps Library [83], which can be applied to evidence-based apps that do not promote particular products.

Improved functionality, suitability, and user engagement of smartphone apps in recent years are welcome developments for parents seeking guidance on infant feeding and activity. However, the high-quality content of apps is critical to good health outcomes. Assessment of available apps revealed that some provided very useful information, but there was wide disparity in reliability and consistency with evidence-based knowledge. Many apps provided additional information outside their focal topic or area of expertise; others offered information largely designed to promote their products. In some instances, this information was limited and did not provide comprehensive advice consistent with evidence-based guidelines.