To identify antecedents of acceptance of a mobile app for COVID-19 symptoms recognition and monitoring (hereafter: symptom app), and a mobile app for contact tracing (hereafter: tracing app), an online survey was developed, tested, and distributed among Dutch citizens. This study did not require formal ethical approval (as ruled by CMO Oost Nederland, file number: 2020-6628). At the beginning of the survey, participants were asked for consent to use their data for research purposes.

Distribution of the survey (via QualtricsXM) started on April 15, 2020. Participants were eligible if they were 18 years of age or older. We used a snowball sampling via posts on social media (LinkedIn, Twitter, and Facebook) and personal connections. Next to this, we recruited participants via a Dutch panel of older adults that indicated they were interested in participating in research on the topic of eHealth. The survey was closed on April 30, 2020. Due to the method of recruitment, a response rate could not be calculated.

Data were analyzed using SPSS (version 19; IBM). Descriptive statistics were performed for all outcomes. Cronbach α values were calculated to assess internal consistency for attitude toward technology, perceived health, fear of COVID-19, and intention to use. Next, survey scores were interpreted for these factors as being negative (score 1 or 2), neutral (score of 3), or positive (score 4 or 5). Via a paired (2-tailed) t test, the difference in intention to use score between both mobile apps was tested. To identify antecedents of acceptance of (1) a symptom app and (2) a tracing app, we conducted multiple linear regression analyses (backward model analyses). The intention to use each app was used as the dependent variable. The independent variables were selected based on Pearson correlation coefficients. Demographic characteristics and factors that (borderline) significant correlated (Pearson correlation cut-off level P≤.10) with the dependent variable “intention to use” were included in the multiple linear regression analyses. For the paired t test and regression analyses, the level of significance was set at P<.05. For the final models the r² was calculated, which indicates the percentage of the variance in the dependent variable that the independent variables explain collectively. To support the quantitative results, the responses on the 2 open questions were sorted and counted by the first author and discussed with the second author, taking an inductive approach. Disagreements were discussed until unanimous agreement was reached.

In total, 238 Dutch citizens completed the survey. Fifteen responders only completed the intention to use survey of a tracing app as this app was presented first and these responders stopped with the survey after these questions; 59.2% (141/238) of the responders were female and the average age was 45.6 years (SD 17.4 years). Only 2.1% (5/238) of responders did not own a smartphone and 74.8% (178/238) claimed that they carried their smartphone with them for most of the day. The average age of our sample is higher than the average age of the Dutch population. There was also an overrepresentation of female participants and participants with a high education level [12]. Compared with the statistics of 2018, in the current sample, there is an overrepresentation of participants owning a smartphone and compared with the statistics of 2020 there is an underrepresentation of participants who are unemployed [12]. The internal consistency of the Attitude Toward Technology scale was good (Cronbach α=.85). Most responders (176/238, 73.9%) had a moderate attitude toward technology. Only 3/238 responders (1.3%) claimed to be infected with COVID-19. All demographic characteristics are presented in Table 1.

The internal consistency of the 4 items in this scale was acceptable to good (Cronbach α=.78). The mean score on this topic was 3.3 (SD 0.68). The majority of the responder’s opinion on this topic was neutral (192/238, 80.7%) and 16% (38/238) of the responders were afraid for a COVID-19 infection. Only a few responders (8/238, 3.4%) were not afraid (Table 2).

For the 3 items to assess the perceived health of the responders the internal consistence was acceptable (Cronbach α=.69). The mean score on this scale was 3.8 (SD 0.68). Most respondents were positive about their health (139/238, 58.4%).

The intention to use was assessed for the symptom app and the tracing app. For both scales, internal consistency was excellent (Cronbach α for the symptom app=.96 and Cronbach α for the tracing app=.96). For both apps, the majority’s intention to use was neutral (Table 2). However, an additional paired t test indicated that there was a significant difference in the scores on intention to use for the symptom app (mean 3.38 [SD 1.07]; n=223) and the tracing app (mean 3.27 [SD 1.13]; n=223; t₂₂₂=–2.598 and P=.01), indicating that the responders were more willing to use a mobile app for COVID-19 symptom recognition and monitoring compared with a mobile app for contact tracing.

The intention to use the symptom app was related to income level (r=0.132, P=.05), attitude toward technology (r=0.220, P<.001), and fear of COVID-19 (r=–0.291, P<.001). The intention to use the tracing app was related to age (r=0.135, P=.04), attitude toward technology (r=0.223, P<.001), and fear of COVID-19 (r=–0.303, P<.001). Based on these outcomes, the independent variables within the linear regression analysis were age, income level, attitude toward technology, fear of COVID-19, and perceived health. Table 3 provides an overview of the correlations between all demographics and factors, and the intention to use.

A multiple linear regression analysis was conducted to predict the intention to use a symptom app based on age, income level, attitude toward technology, fear of COVID-19, and perceived health. The final model included the predictors attitude toward technology, fear of COVID-19, and age (F_3,3=12.012; P<.001). The model has an r² of 0.141. It contains 3 factors that affect the intention to use, but only 2 of them are significant predictors:

Another multiple linear regression analysis was conducted to predict the intention to use a tracing app based on age, income level, attitude toward technology, fear of COVID-19, and perceived health. The final model included the predictors attitude toward technology, fear of COVID-19, and age (F_3,3=14.333; P<.001). The model has an r² of 0.155. Intention to use is predicted by:

An overview of all reasons the responders brought forth for using both mobile apps is presented in Tables 4 and 5. The main reason (33/116, 28.4) for responders to use the symptom app was to control the spread of the COVID-19 virus. In addition, respondents were willing to use this mobile app to monitor own complaints (22/116, 19.0%) and to gain more insight into the spread and symptoms of the COVID-19 virus (19/116, 16.4%).

The main reason to use a tracing app was also to control the spread of the COVID-19 virus (45/147, 30.6%). Next to this, respondents were willing to use this mobile app to gain more insight into the spread and symptoms of the COVID-19 virus (34/147, 23.1%) and for one’s own health (19/147, 12.9%).

An overview of the reasons to not use the mobile apps is presented in Tables 6 and 7. For both mobile apps, privacy was mentioned as the main reason (symptom app=56.6% [64/113] and tracing app=64.8% [92/142]) to not use the mobile apps. Other reasons for not using the mobile apps were the expected usefulness of the app (symptom app=23.9% [27/113] and tracing app=13.4% [19/142]) and a fear of becoming over aware of the situation and its potential consequences, leading to unnecessary stress (symptom app=8.0% [9/113] and tracing app=11.3% [16/142]).

Our main finding is that for both mobile apps age, attitude toward technology, and fear of COVID-19 are antecedents of acceptance. A large group of the Dutch citizens (101/223, 45.3%) is willing to use a mobile app for COVID-19 symptom recognition and monitoring. The main reasons to use this mobile app are to (1) control the spread of COVID-19; (2) monitor their own complaints, and (3) gain more insight into the spread and symptoms of the COVID-19 virus. For the case of a mobile app for COVID-19 contact tracing, 41.2% (98/238) of the Dutch adults appear to be willing to use this mobile app. The main reasons for use are (1) to control the spread of the COVID-19 virus, (2) to gain more insight into the spread and symptoms of the COVID-19 virus, and (3) for their own health. Privacy, doubting the usefulness of the mobile app, and a fear of becoming over aware of the situation and its potential consequences, leading to unnecessary stress, are the main reasons to not use the mobile apps. Overall, Dutch citizens were more willing to use a mobile app for COVID-19 symptom recognition and monitoring compared with a mobile app for contact tracing.

It is difficult to relate our findings to the existing literature, as limited technology acceptance studies have focused on mobile apps to be used during a pandemic, and insights into factors that determine the acceptance of COVID-19–related mobile apps are lacking [7]. In general, age and attitude toward technology are widely acknowledged antecedents of acceptance. For age there is evidence that older age is associated with a lower level of acceptance of mobile apps [13]. Previous results also indicated that attitude toward technology is an important antecedent of acceptance of mobile apps [13,14]. The degree to which an individual is willing to try out any new mobile app is related to the intention to use [13]. Since this study, the mobile apps, announced by the Dutch Government in April 2020, have been developed and implemented. In a recent study by Bente et al [15], the contact tracing app (the CoronaMelder) was tested for usability, and was found easy to use. A comparable study was executed in Germany by Blom et al [16]. They analyzed the potential barriers for the large-scale adoption of the official contact tracing app that was introduced in Germany. The foremost barrier toward using the contact tracing app was the lack of willingness to correctly adopt the app. Besides, compared with the younger group (aged 18-59 years), the older age group (aged 60-77 years) was less likely to use a compatible smartphone. Therefore, access was also mentioned as barrier in this study [16]. Another cross-country survey study (participating countries: France, Germany, Italy, the United Kingdom, and the United States) on the acceptance of a contact tracing app is more optimistic [17], as the willingness to install the app was high among all 5 counties and across all subgroups of the population. In addition, this study concluded that epidemiological evidence shows that app-based contact tracing can suppress the spread of COVID-19 if a high enough proportion of the population uses the app [17].

Our results show that fear of COVID-19 is the most important COVID-19–related factor that predicts acceptance of mobile apps to deal with the COVID-19 pandemic. Because it is difficult to translate this fear into a technology design, this finding needs to be seen in a bigger picture. Public health campaigns during the COVID-19 epidemic will need to educate citizens about the dangers of COVID-19 (personally and for society as a whole), and should then offer downloading COVID-19 mobile apps as a personal strategy to deal with this fear. Next, the positive attitude toward technology that precedes a decision to download a COVID-19 app should be taken into consideration when using these innovations. The end-user population might be skewed toward those with interest in technology (traditionally these are younger, highly educated men [18]), which can create a use divide, and thus, a health divide in society. Measures should be installed to support those groups in society that are not, by nature, technically interested, such as having promotional stalls in the community and diverse channels of user support.

The following 4 limitations should be taken into account for this study. First, due to our recruitment method (snowball sampling via social media), our sample could have been affected by a selection bias. Our sample was mainly composed of participants with a high educational level and a moderate attitude toward technology. Therefore, our results are based on the views of a somewhat skewed sample of the Dutch population, which might reduce the generalizability of our findings. Second, for our analysis, the power of our sample was sufficient. However, a larger sample would improve the generalizability of our outcomes, as mainly Dutch citizens from the eastern part of the Netherlands (87.0% [207/238] of our sample) completed our survey. Third, in our survey the 2 mobile apps are introduced by means of a short description of their general aim. It is unclear if this description was sufficient for the responders to understand the purpose of both mobile apps. Our survey was distributed before the development of the CoronaMelder app in the Netherlands. The study by Bente et al [15] indicated that during this period there were many misconceptions concerning contact tracing among the Dutch population. It is likely that these short descriptions of the general aim of the 2 mobile apps were insufficient to take those misconceptions away. Fourth, the explained variance of both our models is relatively low. Normally, in studies such as these, this number is boosted by including the predictors perceived ease of use and perceived usefulness. However, including these 2 factors leads to little practical results, that is, concluding that the apps should be easy to use. By contrast, the identification of COVID-19–related factors remains an important extension of the existing technology acceptance models.

Age, attitude toward technology, and fear of COVID-19 are important predictors of the acceptance of COVID-19 mobile apps for symptom recognition and monitoring and for contact tracing. These predictors should be taken into account during the development and implementation of these mobile apps to secure acceptance.