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 Journal of Participatory Medicine, is properly cited. The complete bibliographic information, a link to the original publication on http://jopm.jmir.org, as well as this copyright and license information must be included.
Advances in information communication technology provide researchers with the opportunity to access and collect continuous and granular data from enrolled participants. However, recruiting study participants who are willing to disclose their health data has been challenging for researchers. These challenges can be related to socioeconomic status, the source of data, and privacy concerns about sharing health information, which affect data-sharing behaviors.
This study aimed to assess healthy non-Hispanic white mothers’ attitudes in five areas: motivation to share data, concern with data use, desire to keep health information anonymous, use of patient portal and willingness to share anonymous data with researchers.
This cross-sectional study was conducted on 622 healthy non-Hispanic white mothers raising healthy children. From a Web-based survey with 51 questions, we selected 15 questions for further analysis. These questions focused on attitudes and beliefs toward data sharing, internet use, interest in future research, and sociodemographic and health questions about mothers and their children. Data analysis was performed using multivariate logistic regressions to investigate the factors that influence mothers’ willingness to share their personal health data, their utilization of a patient portal, and their interests in keeping their health information anonymous.
The results of the study showed that the majority of mothers surveyed wanted to keep their data anonymous (440/622, 70.7%) and use patient portals (394/622, 63.3%) and were willing to share their data from Web-based surveys (509/622, 81.8%) and from mobile phones (423/622, 68.0%). However, 36.0% (224/622) and 40.5% (252/622) of mothers were less willing to share their medical record data and their locations with researchers, respectively. We found that the utilization of patient portals, their attitude toward keeping data anonymous, and their willingness to share different data sources were dependent on the mothers’ health care provider status, their motivation, and their privacy concerns. Mothers’ concerns about the misuse of personal health information had a negative impact on their willingness to share sensitive data (ie, electronic medical record: adjusted odds ratio [aOR] 0.43, 95% CI 0.25-0.73; GPS: aOR 0.4, 95% CI 0.27-0.60). In contrast, mothers’ motivation to share their data had a positive impact on disclosing their data via Web-based surveys (aOR 5.94, 95% CI 3.15-11.2), apps and devices designed for health (aOR 5.3, 95% CI 2.32-12.1), and a patient portal (aOR 4.3, 95% CI 2.06-8.99).
The findings of this study suggest that mothers’ privacy concerns affect their decisions to share sensitive data. However, mothers’ access to the internet and the utilization of patient portals did not have a significant effect on their willingness to disclose their medical record data. Finally, researchers can use our findings to better address their study subjects concerns and gain their subjects trust to disclose data.
Advances in information communication, electronic health (eHealth), and mobile health technologies are increasingly used to access and collect personal health data. These have contributed to the expansion of research in health care and public health. The eHealth apps can be Web-based or mobile apps that include a range of features, such as tracking changes in health behaviors, chronic care management by health professionals or the patients themselves, or location tracking. In this regard, many eHealth studies have focused on adult women, and specifically mothers, who have often been used as proxy respondents for studying the diverse health care issues of their families [
Privacy and data disclosure are considered serious challenges for health care researchers, and patients are hesitant about sharing their health data as they may expect a loss of privacy while sharing their personal information. This issue encourages patients to retain control over their personal information and disclosure of their data. In fact, patients undergo a cost-benefit analysis to assess different factors that influence their preferences and decisions to share their data, which is known as the privacy calculus [
Although a substantial body of literature has examined individuals’ willingness to share their health data, most of these studies have focused on health information exchanges in the general population [
An attempt to find a solution to greater privacy control has increased the number of studies addressing privacy on the basis of differentiation between sensitive and less sensitive personal health data. In this regard, consumers have more choices to share their personal data with whom they feel more comfortable [
A total of 622 women were randomly selected from a commercial opt-in panel with several million US members. The participants were non-Hispanic white women with children from the Centers for Disease Control and Prevention–funded study on the care burden of mothers of children with autism spectrum disorder (ASD), implemented collaboratively by the Department of Health Management and Informatics of the University of Missouri School of Medicine and the Kennedy Krieger Institute’s Interactive Autism Network (IAN). The study compared care burdens and associated factors between a US representative sample of mothers and the sample of mothers of children with ASD from the IAN registry. As the proportion of nonwhite mothers in the IAN registry was very small, the design required the selection of a comparative US sample of white women raising children without disabilities. According to the sampling method of the original study, the samples in this study were representative of white mothers aged 25 years or older living in an urban area. The majority of the investigated individuals had an educational level of college (4 years) and were employed or self-employed.
The research tool was a survey questionnaire with different domains related to attitudes and beliefs, trust in data sharing, data sharing through mobile phone apps and devices, internet use, and interest in future research, as well as questions related to caregivers and their children. The survey questionnaire consisted of 51 questions, of which 15 were selected for further analysis. The questions selected were related to attitudes and beliefs about data sharing, internet use, and interest in future research, sociodemographic information, and health questions about caregivers and their children.
On the basis a review of empirical studies [
Similarly, mothers’ familiarity with patient portals was investigated in this study, as familiarity with the source of health care data may influence patients’ willingness to share their data. For instance, a study conducted by the United States Department of Veterans Affairs indicated that veterans had a higher level of willingness to share their health information when they gained a higher level of familiarity with the utilization of a Web-based portal [
The respondents were also required to answer the question, “Which type of health-related data are you willing to share anonymously with researchers?” They could choose from the following choices:
To address respondents’ privacy calculus [
Similarly, to assess mothers’ motivation to share data with researchers, the participants were asked, “What is your motivation for sharing your health information?” The mothers could select any of the following choices:
We dichotomized the variables of age (18-49 and >50 years), education (less than 4 years of college and 4 or more years of college), occupational status (self-employed and other occupational status), income level (household income ≤US $74,999 and ≥US $75,000), marital status (married and other status), age of the child/young adult (≤14 years and ≥15 years), number of children (1 child and more than 1 child), health status of the youngest child/adult, and mothers’ health status in general (excellent to good and fair to poor) based on their frequency distribution.
Data analysis was conducted using SAS (version 9.3). Moreover, the frequency analysis was carried out to describe the demographic characteristics of the surveyed mothers, their data-sharing preferences, and their privacy concerns [
Mothers’ motivation to share their data was split nearly equally between less motivated (302/622, 48.6%) and motivated (320/622, 51.4%). The mothers were motivated to share their data to contribute to science (326/622, 52.4%) and to benefit patient health (387/622, 62.2%)—these results are available upon request.
The majority of the respondents were concerned about the misuse of personal health information (507/622, 81.5%). In fact, the respondents were very concerned that their data would be stolen by unknown individuals or companies (360/622, 57.9%) or if their health data would be used without their consent and knowledge (340/622, 54.7%)—these results are available upon request.
Respondents were more willing to share their data with the researchers provided through Web-based surveys (509/622, 81.8%) and collected through their mobile phones (423/622, 68.0%) compared with their medical record data (224/622, 36.0%) and GPS locations (252/622, 40.5%).
Frequency of mothers’ demographic characteristics (N=622).
Demographics | Values, n (%) | |
|
||
|
18-49 | 444 (71.4) |
|
>50 | 178 (28.6) |
|
||
|
Less than 4 years of a college degree | 275 (44.2) |
|
4 or more years of college | 347 (55.8) |
|
||
|
Employed or self-employed | 452 (72.7) |
|
Other occupational status | 170 (27.3) |
|
||
|
Married | 485 (78.0) |
|
Other marital status | 137 (22.0) |
|
||
|
≤74,999 | 234 (37.6) |
|
≥75,000 | 388 (62.4) |
|
||
|
Fair to poor | 62 (10.0) |
|
Excellent to good | 560 (90.0) |
|
||
|
Fair to poor | 28 (4.5) |
|
Excellent to good | 594 (95.5) |
|
||
|
I don’t have | 52 (8.4) |
|
Have more than 1 HCP | 190 (30.5) |
|
Yes, just 1 HCP | 380 (61.1) |
|
||
|
≤14 | 381 (61.1) |
|
≥15 | 241 (38.7) |
|
||
|
One child | 326 (52.4) |
|
More than one child | 296 (47.6) |
|
||
|
Yes | 553 (96.0) |
|
No | 21 (3.4) |
|
||
|
Yes | 361 (58.0) |
|
No | 251 (40.4) |
|
||
|
Less motivated | 302 (48.6) |
|
Somewhat motivated | 248 (39.9) |
|
Very motivated | 72 (11.6) |
|
||
|
Less concerned | 115 (18.5) |
|
Somewhat concerned | 227 (36.5) |
|
Very concerned | 280 (45.0) |
|
||
|
Never used it | 228 (36.7) |
|
Used it more than once | 394 (63.3) |
|
||
|
Not at all | 13 (2.1) |
|
Somewhat | 169 (27.2) |
|
Extremely | 440 (70.7) |
|
||
|
Yes | 224 (36.0) |
|
No | 398 (64.0) |
|
||
|
Yes | 509 (81.8) |
|
No | 113 (18.1) |
|
||
|
Yes | 423 (68.0) |
|
No | 199 (32.0) |
|
||
|
Yes | 252 (40.5) |
|
No | 370 (59.5) |
aHCP: health care provider.
b48 participants did not respond to this question.
The chi-square test results showed that the HCP status (
Results of multivariate logistic regression regarding the association between mothers’ desire to keep their health information anonymous and a set of predictors, including their demographic characteristics.
Effect | aORa (95% CI) | |||
|
|
|||
|
18-49 | 0.95 (0.55-1.64) |
|
|
|
>50 | 1.00 (Reference)c |
|
|
|
|
|||
|
4 or more years of college | 0.75 (0.5-1.13) |
|
|
|
Less than 4 years of a college degree | 1.00 (Reference)c |
|
|
|
|
|||
|
Employed | 1.34 (0.86-2.1) |
|
|
|
Unemployed | 1.00 (Reference)c |
|
|
|
|
|||
|
Married | 1.27 (0.77-2.1) |
|
|
|
Unmarried | 1.00 (Reference)c |
|
|
|
|
|||
|
≥75,000 | 1.19 (0.77-1.84) |
|
|
|
≤74,999 | 1.00 (Reference)c |
|
|
|
|
|||
|
Excellent to good | 1.03 (0.52-2.03) |
|
|
|
Fair to poor | 1.00 (Reference)c |
|
|
|
|
|||
|
Excellent to good | 0.61 (0.21-1.75) |
|
|
|
Fair to poor | 1.00 (Reference)c |
|
|
|
|
|||
|
More than 1 HCP | 0.82 (0.37-1.84) |
|
|
|
Just 1 HCP | 0.69 (0.32-1.47) |
|
|
|
No HCP | 1.00 (Reference)c |
|
|
|
|
|||
|
≤14 | 1.00 (0.59-1.7) |
|
|
|
≥15 | 1.00 (Reference)c |
|
|
|
|
|||
|
More than one child | 1.01 (0.66-1.54) |
|
|
|
One child | 1.00 (Reference)c |
|
|
|
|
|||
|
Yes | 2.06 (0.8-5.33) |
|
|
|
No | 1.00 (Reference)c |
|
|
|
|
|||
|
Yes | 0.35 (0.04-2.91) |
|
|
|
No | 1.00 (Reference)c |
|
|
|
|
|||
|
Somewhat motivated | 0.66 (0.43-1) |
|
|
|
Very motivated | 0.66 (0.36-1.22) |
|
|
|
Less motivated | 1.00 (Reference)c |
|
|
|
|
|||
|
Somewhat concerned | 2.50 (1.52-4.1) |
|
|
|
Very concerned | 4.77 (2.85-7.96) |
|
|
|
Less concerned | 1.00 (Reference)c |
|
aAdjusted odds ratios (aORs) of reporting desire to keep health information anonymous from a multivariable logistic regression model, conditional on mothers’ motivation and concerns and all other characteristics
b
cReference group does not have CI.
dHCP: health care provider.
Bivariate analyses indicated that child health status (
Results of multivariate logistic regression regarding the relationship between mothers’ use of the patient portal and a set of predictors, including mothers’ demographic characteristics.
Effect | aORa (95% CI) | |||
|
|
|||
|
18-49 | 0.86 (0.52-1.41) |
|
|
|
>50 | 1.00 (Reference)c |
|
|
|
|
|||
|
4 or more years of college | 0.98 (0.67-1.44) |
|
|
|
Less than 4 years of a college degree | 1.00 (Reference)c |
|
|
|
|
|||
|
Employed | 0.87 (0.57-1.35) |
|
|
|
Unemployed | 1.00 (Reference)c |
|
|
|
|
|||
|
Married | 0.97 (0.6-1.58) |
|
|
|
Unmarried | 1.00 (Reference)c |
|
|
|
|
|||
|
≥75,000 | 1.40 (0.93-2.12) |
|
|
|
≤74,999 | 1.00 (Reference)c |
|
|
|
|
|||
|
Excellent to good | 0.88 (0.46-1.68) |
|
|
|
Fair to poor | 1.00 (Reference)c |
|
|
|
|
|||
|
Excellent to good | 0.43 (0.14-1.3) |
|
|
|
Fair to poor | 1.00 (Reference)c |
|
|
|
|
|||
|
More than one HCP | 4.3 (2.06-8.99) |
|
|
|
Just one HCP | 3.47 (1.73-6.94) |
|
|
|
No HCP | 1.00 (Reference)c |
|
|
|
|
|||
|
≤14 | 1.29 (0.79-2.11) |
|
|
|
≥15 | 1.00 (Reference)c |
|
|
|
|
|||
|
More than one child | 1.14 (0.76-1.72) |
|
|
|
One child | 1.00 (Reference)c |
|
|
|
|
|||
|
Yes | 2.01 (0.77-5.2) |
|
|
|
No | 1.00 (Reference)c |
|
|
|
|
|||
|
Yes | 2.27 (0.55-9.32) |
|
|
|
No | 1.00 (Reference)c |
|
|
|
|
|||
|
Somewhat motivated | 1.75 (1.18-2.59) |
|
|
|
Very motivated | 2.09 (1.12-3.91) |
|
|
|
Less motivated | 1.00 (Reference)c |
|
|
|
|
|||
|
Somewhat concerned | 1.43 (0.85-2.39) |
|
|
|
Very concerned | 1.46 (0.88-2.42) |
|
|
|
Less concerned | 1.00 (Reference)c |
|
aAdjusted odds ratios (aORs) of reporting use of patient portals from a multivariable logistic regression model, conditional on mothers’ motivation and concerns and all other characteristics.
b
cReference group does not have CI.
dHCP: health care practitioner.
The results of the bivariate analysis indicated that mothers’ willingness to share their electronic medical record (EMR) data was significantly related to mothers’ marital status (
When running the multivariate logistic regression analysis (
The bivariate analysis revealed that the status of HCP (
The findings of this study indicated that the use of mobile phones (
The independent variables of children’s age (
Results of multivariate logistic regression regarding mothers’ willingness to share different types of data with researchers and a set of predictors, including mothers’ demographic characteristics.
Effect | Willingness to share anonymous data from | |||||||||||||||||
|
Electronic medical data | Web-based surveys | Health app and device | GPS | ||||||||||||||
|
aORa (95% CI) | aOR (95% CI) | aORa (95% CI) | aORa (95% CI) | ||||||||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
18-49 | 0.84 (0.49-1.43) |
|
1.09 (0.57-2.07) |
|
1.93 (1.12-3.32) |
|
1.05 (0.63-1.76) |
|
|||||||||
|
>50 | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
4 or more years of college | 1.14 (0.76-1.7) |
|
0.68 (0.41-1.14) |
|
1.13 (0.74-1.73) |
|
1.04 (0.70-1.54) |
|
|||||||||
|
Less than 4 years of a college degree | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
Employed | 1.15 (0.74-1.81) |
|
0.66 (0.36-1.21) |
|
0.95 (0.59-1.53) |
|
1.3 (0.84-2.02) |
|
|||||||||
|
Unemployed | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
Married | 0.69 (0.42-1.12) |
|
0.86 (0.44-1.68) |
|
0.93 (0.53-1.63) |
|
0.95 (0.59-1.53) |
|
|||||||||
|
Unmarried | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
≥75,000 | 0.73 (0.48-1.13) |
|
1.07 (0.61-1.86) |
|
0.79 (0.49-1.27) |
|
0.67 (0.44-1.01) |
|
|||||||||
|
≤74,999 | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
Excellent to good | 0.67 (0.36-1.28) |
|
1.57 (0.71-3.49) |
|
1.21 (0.61-2.42) |
|
0.87 (0.46-1.62) |
|
|||||||||
|
Fair to poor | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
Excellent to good | 0.38 (0.15-0.93) |
|
1.51 (0.45-5.12) |
|
1.31 (0.5-3.48) | .58 | 0.61 (0.25-1.5) |
|
|||||||||
|
Fair to poor | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c | .59 | 1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
More than one HCP | 2.23 (0.92-5.43) |
|
4.47 (1.94-10.3) |
|
3.03 (1.38-6.67) |
|
2.11 (0.93-4.8) |
|
|||||||||
|
Just one HCP | 2.69 (1.15-6.27) |
|
4.22 (1.97-9.05) |
|
3.64 (1.73-7.65) |
|
2.04 (0.93-4.5) |
|
|||||||||
|
No HCP | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
≤14 | 1.61 (0.95-2.73) |
|
1.35 (0.7-2.6) |
|
1.02 (0.58-1.79) |
|
1.6 (0.96-2.66) |
|
|||||||||
|
≥15 | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
More than one child | 0.84 (0.55-1.28) |
|
0.92 (0.53-1.61) |
|
0.95 (0.6-1.52) |
|
0.68 (0.44-1.02) |
|
|||||||||
|
One child | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|
||||||||||
|
Yes | 1.07 (0.36-3.19) |
|
2.85 (0.97-8.41) |
|
5.47 (1.76-17) |
|
2.01 (0.63-6.94) |
|
|||||||||
|
No | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|||||||||||
|
Yes | 1.58 (0.27-9.27) |
|
1.37 (0.28-6.7) |
|
0.73 (0.16-3.4) |
|
1.93 (0.35-10.69) |
|
|||||||||
|
No | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|||||||||||
|
Somewhat motivated | 2.42 (1.6-3.67) |
|
5.94 (3.15-11.2) |
|
3.37 (2.16-5.26) |
|
3.1 (2.08-4.60) |
|
|||||||||
|
Very motivated | 3.64 (2.00-6.63) |
|
2.87 (1.18-6.94) |
|
5.30 (2.32-12.1) |
|
5.15 (2.83-9.38) |
|
|||||||||
|
Less motivated | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
|||||||||
|
|
|
|
|
|
|
|
|||||||||||
|
Somewhat concerned | 0.59 (0.35-1.00) |
|
0.75 (0.34-1.68) |
|
0.88 (0.47-1.68) |
|
0.50 (0.27 0.77) |
|
|||||||||
|
Very concerned | 0.43 (0.25-0.73) |
|
0.57 (0.26-1.25) |
|
0.58 (0.31-1.08) |
|
0.40 (0.22-0.60) |
|
|||||||||
|
Less concerned | 1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
1.00 (Reference)c |
|
aAdjusted odds ratios (aORs) of reporting willingness to share different type of data from a multivariable logistic regression model, conditional on mothers’ motivation and concerns and all other characteristics.
b
cReference group does not have CI.
dHCP: health care provider.
When adjusting for mothers’ characteristics through multivariate logistic regression, only mothers’ motivation to share data (
This study explored mothers’ motivation to share health data, concerns with potential misuse of personal health information, and willingness to share different types of data with researchers, their utilization of patient portals, and their desire to keep their health information anonymous.
Our study results revealed that about half of the mothers were less motivated to share their data with researchers. Our results contradict the findings of a previous study that found that more than 78% of the surveyed respondents were more willing to share their data with researchers [
Our findings on the respondents’ desire to keep their health information anonymous have also been reported elsewhere [
We found that a relatively high proportion of patient portal use (63%) among women in the general population may seem unusual. Although health portal use by patients is becoming more prevalent, a recent study estimated that only 32% of outpatients of a Dutch academic health center used a patient portal [
Our study found that the majority of the mothers were not willing to share anonymous data from medical records and their GPS location using their mobile app (or device). However, these mothers were willing to share anonymous data through a Web-based survey. Our review of the literature cannot corroborate these findings as most of the studies on individuals’ willingness to share their health data focused on health information exchanges [
To our knowledge, this is the first study that employed the opt-in panel to assess non-Hispanic white mothers’ attitudes and perceptions toward data sharing [
In an era dominated by mobile apps and wearable devices, researchers should focus on the value of the privacy calculus in the context of data sharing for research [
Second, to facilitate and improve participation in citizen research, which requires recruiting a large number of individuals to participate in a health research study, a priori market segmentation studies should be implemented to assess consumers’ data-sharing behavior. Such methods are more rigorous than extrapolating the findings from the general population. Consumers’ data-sharing behavior is warranted in part because of the digital divide that is due to the difference in socioeconomic status exhibited within the general population [
Supplemental results (bivariate analysis).
adjusted odds ratio
autism spectrum disorder
electronic health
electronic medical record
health care provider
Interactive Autism Network
odds ratio
The work of EJS was made possible with support from Washington University in St. Louis CDTR (Grant Number P30DK092950 from the NIDDK). The content is solely the responsibility of the authors and does not necessarily represent the official views of the CDTR or NIDDK.
None declared.