This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Mobile health apps related to maternal and infant health (MIH) are prevalent and frequently used. Some of these apps are extremely popular and have been downloaded over 5 million times. However, the understanding of user behavior and user adoption of these apps based on consumer preferences for different app features and categories is limited.
This study aimed to examine the relationship between MIH app characteristics and users’ perceived satisfaction and intent to use.
The associations between app characteristics, ratings, and downloads were assessed in a sample of MIH apps designed to provide health education or decision-making support to pregnant women or parents and caregivers of infants. Multivariable linear regression was used to assess the relationship between app characteristics and user ratings, and ordinal logistic regression was used to assess the relationship between app characteristics and user downloads.
The analyses of user ratings and downloads included 421 and 213 apps, respectively. The average user rating was 3.79 out of 5. Compared with the Apple App Store, the Google Play Store was associated with high user ratings (beta=.33;
A majority of MIH apps are developed by non–health care organizations, which could raise concern about the accuracy and trustworthiness of in-app information. These findings could benefit app developers in designing better apps and could help inform marketing and development strategies. Further work is needed to evaluate the clinical accuracy of information provided within the apps.
Increasingly, users are turning to digital technologies, such as Web-based and mobile platforms, where information regarding any topic is obtained at the touch of a button. There have been significant changes in the types of digital technologies that are available for use, and smartphones are increasingly the most popular devices for
The ubiquity of mobile phones offers a unique opportunity to use mobile health (mHealth) for health information seeking [
Compared with other health topics, mobile apps for maternal and infant health (MIH) subjects, such as pregnancy, childbirth, and infant care, are some of the most frequently developed and commonly used [
The rapid proliferation of mHealth apps has not been accompanied by equal attention to understanding the factors that consumers prefer or the real-world usage patterns when selecting from a multitude of available apps [
Considering the popularity of MIH apps, it is important to understand whether app characteristics (eg, price, ratings, or update age) indicated by previous studies remain influential within the context of perceived satisfaction and intent to use these apps. Therefore, the objective of this study was to examine the relationship between MIH app characteristics (app price, update age, app store, developer type, primary category/genre, content rating, in-app purchase, and in-app advertisement) and 2 outcomes, that is, end user’s perceived satisfaction (user ratings) and intent to use (downloads). Using app data from both the Apple App Store and Google Play Store, this study quantifies apps’ features and characteristics that may affect end users’ perceived satisfaction and intent to use. Given the specificity of MIH apps, this study also examined the influence of app developer type (ie, health care vs non–health care) on user behavior, that is, do users frequently download and rate apps developed by health care developers?
We measured the association between app characteristics, ratings, and downloads in a cross-sectional study of MIH apps available in the Apple App Store and Google Play Store. The dataset of MIH apps was built by scraping data from the Apple App Store [
Scraping results returned apps in the same order as if the search was conducted by an end user. Only the first 200 app results for the Apple App Store and the first 250 app results (later reduced to 50 starting January 2017) for the Google Play Store were returned by the scraper program [
We followed a 3-step process to identify a list of popular MIH apps focused on health education or decision-making support to pregnant women or parents/caregivers of infant. The data reflect app store content as of March 2017.
First, we identified a comprehensive list of relevant keywords that users might enter when searching for apps related to MIH. Search terms such as
Next, each of the 34 keywords was entered individually into a separate search to obtain a comprehensive set of apps for potential inclusion in the study. This resulted in a total sample of 6670 apps. The resultant apps were merged and deduplicated first within stores and then across stores for a total of 4753 unique apps in the dataset (
Flowchart detailing the sample selection process.
List of keywords (N=34).
Keywords | Frequency of use within app description |
Pregnant/pregnancy | 170 |
Prenatal | 63 |
Baby/babies | 56 |
Child/children/childhood | 29 |
Parents/parenting/parenthood | 22 |
Birth | 14 |
Mom/mum | 12 |
Labor/delivery | 8 |
Fetus/fetal | 6 |
Maternal/maternity | 5 |
Breastfeeding | 4 |
Mother/motherhood | 4 |
Infant | 3 |
Obstetrics | 3 |
Antenatal | 3 |
Conception | 3 |
Postnatal/postpartum | 2 |
Gestation/gestational | 2 |
Newborn | 1 |
Intrapartum | 1 |
Lactation | 1 |
Overall, 2 reviewers (RB and CH) independently screened the app descriptions of all retrieved apps (n=4753) for inclusion and exclusion. Disagreements were resolved through discussion and consensus. Inclusion criteria were as follows: (1) description written in English language; (2) target users judged to be pregnant women, to-be parents, and other caregivers of infant children (ie, 0-1 year old as defined by Centers for Disease Control and Prevention, 2017) [
For each included app, the data extracted included (1) average user rating (ie,
The first outcome variable, average user ratings, was standardized as z-scores for the purpose of analysis. The Google Play Store offers continuous values to one-tenth of a point, whereas the Apple App Store rounds it to the nearest half point. To maintain consistency across stores, we converted it to a standardized z-score. Critically, the Apple App Store requires a minimum number of reviews before releasing average user ratings (ie, small numbers are suppressed), and the Google Play Store does not report ratings for unreviewed apps. Of the 742 apps in the sample, 43.3% of apps had no or suppressed user ratings. Therefore, these were all coded as missing values (n=321) and omitted from the analysis; hence, the analysis of user ratings reflects 421 apps from both stores (
To categorize app developer type, a manual review of developer website provided by the app stores was conducted by the primary reviewer (RB). On the basis of the description provided, developers were categorized as a health care developer if they were identified as one of the following: government agency, US hospital system, US academic medical institution, medical specialty society, nonprofit health care organization, consumer organization with health focus, US physician, third-party payer, and pharmaceutical and medical technology companies [
First, descriptive statistics were calculated and assessed. Next, the relationship between app characteristics and end users’ perceived satisfaction (user ratings) and intent to use (downloads) were examined in 2 separate regressions models. First, a multivariable linear regression assessed the relationship between app characteristics (app price, update age, app store, developer type, genre, and content rating) and standardized user ratings controlling for all other available app characteristics for both the Apple App Store and Google Play Store apps. Second, the association between app characteristics (standardized user rating, app price, update age, developer type, genre, in-app purchase, and in-app advertisement) and the number of app downloads was modeled using a series of ordinal logistic regressions for the Google Play Store apps only. Given the small sample size, the analysis of downloads could not be examined with all independent variables in a single model. Therefore, 6 models holding user ratings and price as constant with an additional independent variable were run. Statistical significance was assessed at the
From the total of 421 apps that were included, 322 (75.5%) were free. Of the paid apps, the prices ranged from US $0.99 to US $10.92, with an average price of US $3.14 and a median of US $2.99. The number of days since the last update varied from 14 to 2888 (average 582 days). Only 102 (102/421, 24.2%) apps were developed by health care organizations. The average user rating was 3.79 out of 5. Furthermore, the modal category for user downloads was greater than 50,000, with 66 (66/213, 31.0%) apps (
Descriptive statistics for independent and dependent variables.
Variables | Values | ||
|
|||
|
Average user ratings (number of stars out of 5), mean (SD) | 3.79 (0.98) | |
|
App price (paid apps), mean (SD) | 3.14 (2.13) | |
|
Update age (days), mean (SD) | 582 (624.44) | |
|
|
||
|
|
Apple App Store | 221 (52.5) |
|
|
Google Play Store | 200 (47.5) |
|
|
||
|
|
Non–health care | 319 (75.8) |
|
|
Health care | 102 (24.2) |
|
|
||
|
|
Health and fitness | 225 (53.4) |
|
|
Medical | 156 (37.1) |
|
|
Other (books and reference, education, and parenting) | 40 (9.5) |
|
|
||
|
|
Not age restricted | 319 (75.8) |
|
|
Age restricted | 90 (21.3) |
|
|
Unrated | 12 (2.9) |
|
|||
|
|
||
|
|
1-500 | 43 (20.2) |
|
|
501-5000 | 52 (24.4) |
|
|
5001-50,000 | 52 (24.4) |
|
|
50,001-50,000,000 | 66 (31.0) |
|
|
||
|
|
Yes | 39 (18.3) |
|
|
No | 174 (81.7) |
|
|
||
|
|
Yes | 108 (50.7) |
|
|
No | 105 (49.3) |
Compared with the Apple App Store, apps from the Google Play Store had, on average, 0.33 higher star ratings (
Multivariable linear regression for factors associated with standardized user ratings (N=421).
Variables | Estimates | SE | ||
Developer type health carea | −0.20 | 0.11 | .06 | |
Google Play platforma | 0.33 | 0.12 | .005 | |
|
||||
|
Other (books and reference, education, and parenting) | Refb | Ref | Ref |
|
Medical | −0.19 | 0.17 | .23 |
|
Health and fitness | −0.41 | 0.17 | .01 |
Update age | −0.0004 | 0.00008 | <.001 | |
|
||||
|
Age restricted apps | Ref | Ref | Ref |
|
Not age restricted apps | −0.32 | 0.13 | .01 |
|
Unrated apps | −0.51 | 0.32 | .10 |
Price (US $) | 0.03 | 0.03 | .35 |
aThe reference level for platform iOS and for developer type not health care developer.
bRef: reference.
Factors positively associated with user downloads were standardized user ratings (beta=.80;
Ordinal logistic regression for factors associated with user downloads (n=213).
Variables | Estimates | SE | ||
Standardized user rating | 0.80 | 0.20 | <.001 | |
Price (US $) | −0.45 | 0.15 | .003 | |
Developer type health care | −0.14 | 0.30 | .63 | |
|
||||
|
Other (books and reference, education, and parenting) | Refa | Ref | Ref |
|
Medical | −1.63 | 0.46 | <.001 |
|
Health and fitness | −1.29 | 0.42 | .002 |
Update age | −0.0008 | 0.0003 | .009 | |
In-app purchase | 1.12 | 0.36 | .002 | |
In-app advertisement | 0.64 | 0.27 | .02 |
aRef: reference.
This study uses publicly available open-source data to assess the factors related to user ratings (perceived satisfaction) and user downloads (intent to use) for MIH apps. To our knowledge, this is the first study that quantifies app features and characteristics that relate to user ratings and downloads for MIH apps using data from the Apple App Store and Google Play Store.
The Apple App Store and Google Play Store contain hundreds of apps related to MIH, many of which have been downloaded hundreds and thousands of times. Our findings suggest that price, user ratings, in-app purchase options, and presence of in-app advertisements were impactful predictors of user downloads. For instance, less expensive apps and apps with optional in-app purchases were associated with higher user downloads. Consumers tend to prefer apps that are free or of low cost with an ability to purchase additional features or functionalities via in-app purchases, as opposed to paying a higher price upfront [
Furthermore, the number of user downloads also increased with average user ratings, which suggests that perceived satisfaction with these apps is an important indicator related to new user preferences. This corroborates previous findings that most users tend to download apps with high user ratings [
In terms of genre, our findings suggest that apps in the health and fitness category have lower ratings and downloads, whereas apps in the medical category have fewer downloads. However, we cannot ascertain the exact reason behind why users may prefer MIH apps within specific categories over others, thereby calling for further investigation.
In addition, our results reveal that the availability of updates (ie, when was the app last updated) positively influences both user ratings and downloads. This is because updates act as a proxy of the app’s evolution [
These results may provide some correlational information to app developers, including health care organizations, about the types of apps that people tend to download and rate higher.
Our findings could be applied to improve app design mechanisms that are currently in place for the MIH app market. Considering the sensitivity of MIH, we recommend that developers employ ways to increase health expert involvement in app design and content delivery.
A large majority (75.8%) of MIH apps included in this study were developed by non–health care organizations. This is consistent with previous reports on limited or nonexistent health expert involvement in app development within other health domains such as urology [
Although these concerns have garnered attention from public agencies such as the US Food and Drug Administration (FDA), presently, the FDA only regulates apps that act as medical devices [
In this study, we examined MIH apps from only 2 app stores, and information available in these app stores and developers’ websites were collected. However, the app stores and developers’ websites remain the main source of information available to consumers too. Thus, the study uses information similar to what would normally be available to consumers in a
We suggest future studies focus on establishing consistent guidelines for the disclosure of health care professional’s participation and measures to quantify it. We also recommend future studies apply the same approach to other health topics and compare their results with this study. Although we found associations of app characteristics with perceived satisfaction and intent to use, we were not able to identify their impact on learning or behavior change because of app use. Therefore, we recommend future research and inquiry to focus on collecting data from users pertaining to their learning and behavior impact from app use. At present, we lack a standardized format/clinical guideline for the evaluation of accuracy of clinical content or included topics within apps, which necessitates further study and recommendations in this area.
A large majority of MIH apps were developed by non–health care organizations, which raises concern about the clinical accuracy and quality of MIH app content. No differences in ratings or downloads were observed between health care and non–health care organizations. Therefore, if health care organizations, in fact, provide more credible information, fewer consumers may receive this information. Health care providers, app developers, and policy makers may consider strategies to review and promote evidence-based and trustworthy apps to consumers.
mHealth apps are increasingly becoming popular and can be used as a tool for MIH care delivery. However, the design and delivery of effective MIH apps still remain a challenging issue. Considering the lack of standard guidelines for app development, or selection, users typically consider publicly available app characteristics to make decisions pertaining to app use and satisfaction. Therefore, we examined the relationship between app characteristics, perceived satisfaction, and intent to use by using cross-sectional data from 2 app stores. We observed that app price, update age, user ratings, in-app purchases, and in-app advertisements are important predictors for intent to use, whereas update age is an important indicator for perceived satisfaction. Most importantly, our findings revealed that apps developed by health care developers were neither associated with higher perceived satisfaction nor intent to use. Knowledge of factors related to ratings and downloads may benefit app developers and help inform future marketing and development strategies.
Food and Drug Administration
institutional review board
Indiana University-Purdue University Indianapolis
maternal and infant health
mobile health
The authors would like to thank the Department of Health Policy and Management at Richard M. Fairbanks School of Public Health, IUPUI, and Regenstrief Institute, Inc for their ongoing support during the research process. The authors would also like to thank the anonymous reviewers who provided critical feedback and guidance on revising the manuscript before publication.
The primary author RB proposed and completed this study as part of her doctoral dissertation. The coauthors JRV, BED, TC, and CAH (chair), who are all members of the dissertation committee, contributed significantly to the acquisition, analysis, and interpretation of data; to rigor of the methodology and analysis; and to write the manuscript.
None declared.