Usage of a workplace health promotion app: an evaluation of app usage data and medical check-up results

Abstract

Smartphone applications (apps) have been used and evaluated in the context of workplace health promotion (WHP) programs. However, there is a lack of studies analyzing actual app usage data and measuring changes in medical markers to evaluate the effectiveness of WHP apps in terms of health improvements in practice. In this study, we evaluated data from 555 employees of an IT company who participated in a WHP program over the course of one year. Participants of the program received a medical check-up as well as a health app to understand their medical results and receive advice for a healthier lifestyle. In addition, 99 of these employees underwent a follow-up medical check-up. It was found that the smartphone app for healthy lifestyle promotion in combination with onsite medical check-ups was effective in improving various health indicators, for example, BMI, body fat, blood pressure and triglycerides. The study further identified influence factors for sustained app usage, and analyzed different usage behaviors among gender and age groups.

Keywords

Adoption effectiveness corporate health mHealth workplace health

Introduction

Mobile health (mHealth) is used in numerous contexts across many countries.¹ The technology was first developed for chronic diseases but has since been employed for health promotion and self-management as well.² Aside from private use, workplaces have also started to use apps as a means of workplace health promotion (WHP) because they are considered a cost-effective solution for supporting employee health and enhancing productivity, as well as employee satisfaction.^3,4

Studies have evaluated the various use cases, identifying measurable effectiveness. For instance, research has demonstrated the effectiveness of mHealth in the use of physical activity trackers and wearables at the workplace.^5,7–11 Besides physical activity and sedentary behavior,^8,12 studies have also targeted mental health (e.g., in terms of stress perception or resilience).^6,13–15 The intervention groups in these effectiveness studies have extended beyond office workers to include, shift workers, truck drivers, and airplane pilots.^10,16,17

Studies have also illustrated perceived advantages of mHealth e.g. flexibility, personalization, possibilities of real-time progress tracking and feedback, the ability to reach a larger target group, nonstigmatization, 24-hour accessibility, usage independently from time and location, the existing integration of smartphones in daily life, prolonged user engagement, and cost-efficiency.^{4,6,11,18–23} Findings have further suggested that employers also benefit because the working conditions and health of employees can be tracked to implement health measures tailored to the workforce, hence improving work safety by preventing accidents.¹⁰

However, up until now, detailed app usage behaviors of employees have not yet been extensively studied. The potential of WHP apps has been mainly examined by qualitative and quantitative studies measuring potential acceptance^18,21 and by developing theoretical frameworks and models for adoption.^24,25 Clearly, more detailed analyses of usage behaviors and actual effectiveness are required. However, according to Bol et al., it is difficult to compare studies focusing on usage behavior patterns because the apps differ widely in terms of their content, focus area, and target groups.²⁶ Still, Bol et al.‘s study emphasizes that those analyses are important, and a focus on the differences among individuals is crucial to develop better apps. Research indicates a need to perform qualitative research regarding the motivation of end-users.²¹ McCallum et al. have argued that researchers should use more log data from the app to investigate the full potential of mHealth and make conclusions about usage behavior.⁷

To fill this knowledge gap, the present study aims to evaluate app usage behavior and effectiveness based on health improvements using an explorative approach in a case study. Subgroup analyses and comparisons were conducted to assess users’ specialties and preferences. The basis for the evaluation is usage data gathered by app log and medical data acquired in medical check-ups before and after app usage. The study answers the following two research questions:

1. What are the differences in usage behavior (duration, frequency) among different demographic groups?

2. Did app usage have a significant impact on the measured medical outcomes?

The present research seeks to identify the factors that contribute to the usage of an app; examining this can help support the development of these apps and incorporate the preferred content across different target groups.

Related work

There have been few contributions in the field of app usage behaviors and the effectiveness of mHealth in terms of health improvements for WHP. The relevance of research arises from the fact that the costs for absent employees are increasing, there is increasing labor market competition, and there is a shortage of qualified workers because of demographic changes.⁴ Thus, employers need measures to increase employee satisfaction, lower employee turnover and support the health of its their employees. One option is corporate health apps, which were first started by various employers and are now being examined by researchers to assess their effectiveness.²⁷ Literature reviews have pointed out increased physical activity, better productivity, and higher engagement; however, often, the effects are not significant.^28,29 Additionally, a study by Lowenstyn et al. has also reported the potential effectiveness of a medical check-up in combination with the use of web-based information³⁰ as a new promising measure, supporting our study aim. Some authors have further argued that more studies on the effectiveness of these apps by using a larger number of participants and longer time frames are required.²⁸ Other studies have contributed in the field of adoption to gain insights into the potential usage and acceptance of end-users.^24,31

To explain a range of usage behaviors and detect usage approaches, studies have combined available theoretical frameworks and models, such as the theory of planned behavior (TPB) and technology acceptance model (TAM).³² Sari et al. modify the unified technology acceptance and use theory (UTAUT) to explain WHP app usage,²⁴ while Melzner et al. further develop a model based on the TPB, the TAM, and the health belief model (HBM).²⁵ For their proposed model, they find limited overall validity in terms of variance explained; however, based on the structural equation model, the importance of normative belief and, thus, colleagues’ support in the context of WHP apps is considered important.³³ Yang et al. have conducted a literature review to gather the factors that are relevant for adherence to mHealth apps; they summarize the predicting factors as being adherence, dimensions of adherence, and potential outcomes.³⁴ The four dimensions to measure adherence to mHealth apps are shown to be the following: “breadth (device functions used and completion of modules), depth (meeting tasks or challenges and behavior change), length (device use time and length of use), and interaction (active or passive interaction with the program elements)”.³⁴ These dimensions match the focus of our study on different usage behaviors in the app, while also combining them with medical outcomes. However, not all dimensions will be covered in our paper.

In addition to these general theories and theoretical frameworks, studies have also underlined the relevance of demographics to WHP effectiveness in practice. Gender is one factor that has been shown to influence behavior. Specifically, women use apps and the internet to search for more in-depth information about health and tend to use nontechnical health promotion programs in the workplace.^35,36 whereas men appear more eager to test new technologies. These findings imply that men are more heavily influenced by perceived usefulness and that women are more attracted by ease of use and subjective norms.^37,38 Additionally, findings have suggested that more men have fitness apps installed on their phones, while women seem to be more interested in nutrition and reproduction apps.²⁶

Age may be another influencing factor of mHealth usage. Older employees have been found to be more engaged in non-technology-based programs, placing greater value on ease of use compared with younger participants, which might be explained by less experience with technology among older individuals. Accordingly, younger individuals have had more mHealth apps installed on their phones.²⁶ Younger employees are also more interested in content related to physical activity and nutrition, stress management, and technology.^39–41

Methods

In accordance with McCallum’s recommendations, the present research considers actual usage data from a WHP app.⁷ In this case, the data were derived from an intervention conducted at a German IT company. The current study analyzes data from medical examinations compared with data from activities executed within an app. This specific app can be considered an information-based intervention that does not involve connected wearable functionalities. All data used in the app were obtained from either the check-up and app provider based on the medical check-up results (medical data), self-administered data from questionnaires (subjective health status), or log data of the user in the app tracked by the app provider (app usage data). No permissions for other tracking apps or gadgets were utilized. After the first medical check-up, the users were equipped with the WHP app, which demonstrated their personal results and provided personalized health recommendations and challenges.

Two groups of participants were distinguished. The first group (456 participants) consisted of users who participated in only one medical check-up and the app afterwards. The second study group (99 participants) completed pre- and post-screening within 1 year (December 2017–December 2018) at two different company locations; this included the initial and follow-up medical check-ups, which were conducted either in December 2017 and December 2018 (one-year follow-up period) or in March 2018 and October 2018 (seven-month follow-up period). However, for the comparisons between the two check-up results, no differences were considered between the two follow-up times.

Application and intervention

The intervention was supported by one supplier who executed the comprehensive medical check-up and provided the app. Neither the check-up nor app was explicitly developed for the study, and the design process was not scientifically reported. The provider was chosen by the company based on the comprehensive medical check-up, as well as the technical component. Generally, the intervention consisted of two parts: the medical check-up and app usage afterward.

Medical check-up: The medical check-ups were performed by educated staff and assessed four main areas: metabolism, body composition, lung function, and cardiovascular system. Aside from measuring lung function, the app also measured pulse pressure, body fat, body water, ratios of body parts, and various blood values. Following the check-up, the results were immediately entered into the app. The participants subsequently underwent teleconsultation with a doctor to obtain feedback on the results and advice about follow-up measures. Body age, which was calculated using a benchmark based on health measures, was compared with real age to visualize the health status of the participants. Body age was calculated according to the various age-related biomarkers that have been described by several authors⁴² using generalized additive models (GAM). For all values, comparisons were made according to gender and age to contextualize the values.

mHealth application: The app was developed to pursue three main objectives: first, to demonstrate personal check-up results; second, to compare the results with others (i.e., normative comparison); and third, to yield advice (e.g., articles and challenges) to improve those outcomes based on the medical check-up results. The challenges, which were self-administered and not tracked, ranged in focus from physical activity and nutrition to mental health (e.g., mindfulness exercises). Screenshots are provided in Appendix 1.

Recruiting and study group

The study group was composed of employees from a German IT company with around 2500 employees. All employees were informed about the program via the usual communication channels of the company (email and intranet). They were also informed about data usage in the check-up and app and for scientific research (in an anonymous form). Data usage was approved by the internal privacy department of the company, as well as employee representatives. Communications were handled by the human resources department responsible for corporate health management.

Participation in the program was voluntary. No information was collected regarding the reasons for why employees decided to participate. The incentives for participation were that the check-ups took place during working hours and that all costs were covered by the employer. The participation rate was limited by the number of contracted check-ups with the provider. Participation was approved on a first-come, first-serve basis, which might have caused selection bias toward more motivated employees.

Statistical analysis

First, the entire study group (555 participants) was analyzed using descriptives and charts to evaluate usage duration and content (Part A). Then, detailed analyses were carried out concerning the differences in medical check-up results between pre- and post-check-ups for only those 99 participants who participated in both screenings (Part B). All factors were tested for significant differences between the baseline and follow-up screenings using Spearman’s correlation. Furthermore, a t-test was employed to evaluate the differences between medical values because a normal distribution was assumed based on the sample sizes in both groups, as well as the characteristics of the medical values. However, the usage rates were not normally distributed and, thus, needed to be evaluated with nonparametric measures, which were performed using SPSS 26.0.

Results

To answer the research questions, the study was divided into Parts A and B. Part A accounted for all users and focuses on usage behavior after the initial medical check-up and initial usage of the app. Part B investigated the effects of the intervention by monitoring only those participants who participated in two health screenings and, thus, provided pre- and posttest values (99 participants).

Study part A

The datasets from a total of 555 participants were analyzed. Thirty datasets with missing data in the demographics were excluded. Of the valid cases, 456 participants participated in only one screening, whereas 99 (16.9%) completed two screenings. The participants were 39.37 years old (SD = 9.55) on average and ranged from 22 to 67 years old. They consisted of 219 female employees and 336 male employees (60.5%). Their body mass index (BMI) ranged from 17.1 to 40.7, with an average of 24.71 (SD = 3.86). Their body fat percentage was 22.72 (SD = 7.1) on average and ranged from 5.1% to 46.7%. Blood pressure (BP) measures ranged from 100 to 198 (systolic BP left) and from 100 to 196 (systolic BP right), with an average of 129.07–130.86 (SD = 13.75). The average cholesterol/HDL ratio was 3.19 (SD = 1.13), and the average triglyceride measurement was 145.90 (SD = 80.23)

Significant differences were found between the participants in terms of their demographics. Some of the medical values revealed positive correlations. Additionally, age was seemingly also positively correlated with medical values, implying that health risks increase with age (see Tables 1 and 2). Because of the differences in medical results between men and women, each group is presented separately.

Table 1.

Correlation table for women (N = 219): Pearson’s correlation.

		BMI	Age	Body fat	Systolic BP (left)	Systolic BP (right)	Trigly- cerides	Cholesterol/HDL
BMI	Correlation	1	0.236**	0.790**	0.301**	0.206**	0.259**	0.469**
BMI	Sig. (2-tailed)		<0.001	<0.001	<0.001	0.002	<0.001	<0.001
Age	Correlation	0.236**	1	0.263**	0.361**	0.178**	0.162*	0.296**
Age	Sig. (2-tailed)	<0.001		<0.001	<0.001	0.008	0.016	<0.001
Body fat	Correlation	0.790**	0.263**	1	0.284**	0.166*	0.171*	0.315**
Body fat	Sig. (2-tailed)	<0.001	<0.001		<0.001	0.014	0.011	<0.001
Systolic BP (left)	Correlation	0.301**	0.361**	0.284**	1	0.630**	0.128	0.276**
Systolic BP (left)	Sig. (2-tailed)	<0.001	<0.001	<0.001		<0.001	0.058	<0.001
Systolic BP (right)	Correlation	0.206**	0.178**	0.166*	0.630**	1	0.071	0.196**
Systolic BP (right)	Sig. (2-tailed)	0.002	0.008	0.014	<0.001		0.298	0.004
Triglycerides	Correlation	0.259**	0.162*	0.171*	0.128	0.071	1	0.388**
Triglycerides	Sig. (2-tailed)	<0.001	0.016	0.011	0.058	0.298		<0.001
Cholesterol/HDL	Correlation	0.469**	0.296**	0.315**	0.276**	0.196**	0.388**	1
Cholesterol/HDL	Sig. (2-tailed)	<0.001	<0.001	<0.001	<0.001	0.004	<0.001

Table 2.

Correlation table for men (N = 336): Pearson’s correlation.

		BMI	Age	Body fat	Systolic BP (left)	Systolic BP (right)	Trigly- cerides	Cholesterol/HDL
BMI	Correlation	1	0.263**	0.858**	0.282**	0.329**	0.342**	0.382**
BMI	Sig. (2-tailed)		<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Age	Correlation	0.263**	1	0.227**	0.137*	0.136*	0.247**	0.209**
Age	Sig. (2-tailed)	<0.001		<0.001	0.012	0.013	<0.001	<0.001
Body fat	Correlation	0.858**	0.227**	1	0.262**	0.312**	0.328**	0.338**
Body fat	Sig. (2-tailed)	<0.001	<0.001		<0.001	<0.001	<0.001	<0.001
Systolic BP (left)	Correlation	0.282**	0.137*	0.262**	1	0.854**	0.128*	0.131*
Systolic BP (left)	Sig. (2-tailed)	<0.001	0.012	<0.001		<0.001	0.019	0.016
Systolic BP (right)	Correlation	0.329**	0.136*	0.312**	0.854**	1	0.157**	0.159**
Systolic BP (right)	Sig. (2-tailed)	<0.001	0.013	<0.001	<0.001		0.004	0.003
Triglycerides	Correlation	0.342**	0.247**	0.328**	0.128*	0.157**	1	0.737**
Triglycerides	Sig. (2-tailed)	<0.001	<0.001	<0.001	0.019	0.004		<0.001
Cholesterol/HDL	Correlation	0.382**	0.209**	0.338**	0.131*	0.159**	0.737**	1
Cholesterol/HDL	Sig. (2-tailed)	<0.001	<0.001	<0.001	0.016	0.003	<0.001

Different usage behaviors in the app

General usage: Significant correlations between the total number of activities executed in the app by the user and the measured medical indicators were found for cholesterol/HDL ratio (r = 0.92; p = .031), BMI (r = 0.085; p = .045), systolic BP left (r = 0.138; p = .001), and systolic BP right (r = 0.125; p = .003) according to Spearman’s rank correlation. These results indicate that those with higher values used the app significantly more often. Interestingly, when split by gender, no significant differences were apparent among men, but BMI (r = 0.176; p = .009) and body fat (r = 0.160; p = .018) were significantly correlated with higher app activity among women.

Usage over time: The usage was normalized based on the first day of usage per user. On average, the app was used for 130 days, with 124 users logging activities only in the first week (or even less). On the first day, a total of 36,719 activities were logged (excluding the variables “time spent on a page” and “time before starting”), but the number of activities dropped to 11,682 and 6137 on the second and third days, respectively. To illustrate development over 1 year, Figure 1 visualizes the normalized activities from 2 days after the first usage onward. There are marked peaks of usage around the first days after the first screening, as well as after 7 and 12 months of use, which can be explained by the check-up dates after 7 and 12 months. The first strong peak lasted around 1 month.

Figure 1.

Total activities within the app over time (normalized per user).

Duration spent on page (intensity): When evaluating the duration of usage, some differences in demographics were found. Tables 3 and 4 display the mean duration of usage. Screenshots of the app are included in the appendix. Here, the mean of each individual’s average duration per activity was used. When correlating these data, each logged interaction with the app was considered to be one case, here adjusting for the frequency of usage (i.e., users who logged more values in the app were considered more heavily in the correlation). Logs that included less than 1 s of usage were excluded. Additionally, the highest 1% of the logged durations were excluded as outliers, which led to the exclusion of all logs with a duration longer than 12.6 min.

Table 3.

Average duration of app usage according to age group.

	<30 years	30–39 years	40–49 years	50–59 years	>60 years	Total
Mean time on the page (in seconds)	22.20	25.38	25.10	26.53	10.87	24.82

*Calculation: mean of average time spent on page per individual.

Table 4.

Average duration of app usage according to gender group.

	Women	Men	Total
Mean time on the page (in seconds)	25.15	24.61	24.82

*Calculation: mean of average time spent on page per individual.

Age groups

The time spent on the page increased with the age of the users (except for ages above 60, although this category included only two users). On average, younger users spent less time on one page compared with older users, as confirmed by the Spearman rank correlation (p = .00; correlation coefficient: 0.127).

Gender Groups

On average, women spent slightly more time on the app pages than men. However, it should be noted that, on average, men logged 40.29 entries that were included in this analysis, whereas this figure was 35.67 for women.

Study Part B: Differences in the medical indicators before and after app usage

Table 5 presents the demographics of the study group that participated in both screenings. No differences were detected when comparing the demographics of those who participated in the two screenings to those who participated in only one. Slight differences were observed between gender groups, with 19.31% of the women and 27.33% of the men participating twice.

Table 5.

Demographics of the study participants who participated in two screenings (N = 99).

Demographics	Mean (SD)	Min–Max	Frequency (%)
Age at initial check-up	38.75 (9.88)	23–58
<30 years			22 (22.2)
30–39 years			34 (34.3)
40–49 years			25 (25.3)
50–59 years			18 (18.2)
>60 years			0
Gender
Male			36 (36.4)
Female			63 (63.6)
BMI	24.66 (3.85)	18.3 – 40.7
Calculated body age at first screening
Body composition	39.94	23–64
Circulatory	38.16	22–66
Metabolic	38.90	23–61
Respiratory	32.13	20–58

After analyzing the demographics, the differences between the two check-ups were evaluated to determine the potential effects of app usage. Table 6 presents the main results. Significant differences were found for BMI, body fat, and systolic BP left. The remaining values further improved over the time of the app usage; however, no significance was found.

Table 6.

Differences in health values between the first and second screenings.

Value	Mean (SD) 1. Screening	Mean (SD) 2. Screening	Paired sample t-test significance	Effect size (r)
BMI	24.66 (3.85)	24.38 (3.44)	0.12*	0.25 (low)
Body fat	21.80 (7.91)	20.95 (7.60)	0.00*	0.40 (medium)
Cholesterol/HDL ratio	3.18 (1.10)	3.25 (0.87)	0.385	0.087 (low)
Systolic BP (left)	129.98 (13.89)	126.12 (13.41)	0.00*	0.35 (medium)
Systolic BP (right)	131.02 (14.22)	129.10 (13.02)	0.050	0.20 (low)
Triglycerides	148.04 (80.78)	133.84 (65.51)	0.083	0.17 (low)

The difference in calculated body age between the first and second screenings was also determined. The calculated body age accounted for a one-year difference (calculation = [Value_Screening1] − [Value_screening2 – 1]) because all participants aged 1 year during the time period. On average, positive developments in the calculation of body age were detected.

As illustrated in Table 7, body age improved in terms of body composition (0.76 years), circulatory system (0.39 years), metabolic system (0.64 years), and respiratory system (0.43 years) when accounting for the age difference of 1 year. Additionally, the comparison of real age versus body age also improved in the second screening.

Table 7.

Differences: Real age at check-up and calculated age.

Differences: Real age at check-up and calculated age (positive values demonstrate positive development)
		Min.	Max.	Mean	SD
Check-up 1	Real age versus body composition age	−8.00	5.00	−1.19	3.00
	Real age versus circulatory age	−10.00	6.00	0.59	2.87
	Real age versus metabolic age	−8.00	8.00	−0.15	2.70
	Real age versus respiratory age	−6.00	11.00	6.62	3.21
Check-up 2	Real age versus body composition age	−4.00	2.00	−0.43	0.98
	Real age versus circulatory age	−6.00	6.00	0.98	2.45
	Real age versus metabolic age	−5.00	6.00	0.49	3.19
	Real age versus respiratory age	−2.00	11.00	7.05	2.65
Average change in calculated age between Check-up 1 and Check-up 2^a
		Min.	Max.	Mean	SD
	Body composition	−6.00	8.00	0.76	2.88
	Circulatory system	−6.00	12.00	0.39	1.91
	Metabolic system	−12.00	11.00	0.64	4.20
	Respiratory system	−8.00	14.00	0.43	2.39

^aThe calculation is based on the following formula: Value Check-up 1 − (Value Check-up 2 − 1 year); 1 year is deducted to account for the year of aging between check-ups

The metabolic system, especially the liver values, improved. For body composition, muscle mass was notably strengthened, whereas body fat and weight also decreased significantly. These results confirm the findings of other studies on weight management.⁴³ On average, the participants lost 0.78 kg. The circulatory values, especially the BP values, also improved, which is in line with other studies demonstrating positive results of screening on BP.⁴⁴ In a study using web-based interventions, a notable reduction of 0.3 years was found for cardiovascular age (including systolic BP and cholesterol).⁴⁵

Discussion

The above results support various conclusions and implications for theory and practice.

When applying the framework of Yang et al. on adherence to mHealth interventions, the app used in the present study provided deep insights by measuring the length and depth of app usage and adherence by duration of usage and content used.³⁴ These factors were complemented by predictor factors (correlations between demographics and adherence, as well as outcome), as well as medical data for the outcomes (using objective measures instead of self-reported outcomes), to provide deep insights into these relations.

During the evaluation of usage frequency over time, the screening dates were identified as incentives. When considering the distribution of activities over 1 year, upcoming screenings caused a higher frequency of interaction with the app, which, in turn, stimulated awareness of one’s health status. A possible explanation for this finding is that a fear of receiving negative results motivated the users to be more active. Indeed, Jimenez and Bregenzer have recommended a combination of instant feedback and direct health expert feedback to enhance motivation for participation.⁴⁶ Balk-Moller et al. have further confirmed that users considered clinical examinations as incentives to change their behavior.⁴⁷ Regardless, app usage clearly dropped after a few months, which has been similarly reported by other studies, including technical components. ³⁰ For instance, in a literature review by Buckingham et al. (2018), the attrition rate regarding a follow-up screening ranged from 0% to 74%, and four of the 30 reviewed studies recorded attrition rates above 40%.⁸ Of course, the follow-up period varied widely.

However, these high attrition rates do not necessarily need to be considered a limitation or as nonadoption of the app. Because apps are meant to support and incentivize a change toward healthy behavior, permanent long-term usage of the app might not be necessary. It is more important that users become educated about their potential health problems, learn how to change them, and sustain healthy behaviors. Still, studies have indicated that more research is needed to understand the relationship between long-term usage and behavioral change.⁴⁸ The incentive of screenings and certain check-up results might reinforce the importance of combining screening with technology, as promoted by studies hypothesizing that workplace screening may be effective in stimulating lifestyle improvements.³⁰

Focusing on our first research question and the effects of demographics, we can state the following: we studied a typical IT group consisting of mostly young male employees.⁴⁹ Men were more likely to participate in a second screening, which, based on other studies, might be explained by the technical component of the intervention.³⁸ However, other research has claimed that women tend to search more extensively for health information.³⁵ Generally, positive correlations between elevated check-up results and app usage have led to the conclusion that alarming results e.g. high BMI triggered usage of the app. This phenomenon supports the purpose of WHP but contradicts assumptions that healthy employees are more likely to participate in WHP.^50,51 Interestingly, certain findings (especially for BMI and body fat) were significant for women but not for men, implying that the potential triggers for usage might differ. For women, elevated weight and fat measures are apparently a more important incentive to use the app compared with other elevated biomarkers such as cholesterol or BP.

Additionally, the women seemed to spend more time on the app pages, even though this difference was small. Research on usage frequency (not duration) confirms these slight differences.^26,52 Moreover, older users also spent more time on individual pages on average. Studies focused only on frequency (e.g., when comparing usage in a meditation app) have confirmed that other content in the app was used more by older users than younger ones.⁵²

The effects of the intervention showed that the participants generally improved their medical results. When examining the check-up results of the participants independently of their app usage, the values for body composition and metabolic system reflected the most significant improvement over the course of the intervention. This observation has been based on the calculated age compared with the real age, as well as the differences between the first and second screenings. A similar study on a website including screenings and technical usage has yielded similar results to support technical usage.³⁰ The greatest improvement of those values can be explained by the fact that a healthy lifestyle (i.e., proper nutrition and physical activity) has the strongest influence on such values. In the individual analysis of certain values, significant differences were found in BMI, body fat, and systolic BP. These results clearly signify the positive outcomes of the intervention. Because of the individual design of the intervention, detailed comparisons with other studies would be difficult.

The present exploratory study has conclusively indicated that users need incentives, such as another screening or low health values, for app usage. The use of screening in combination with the app improved the health of most participants, thus achieving positive results for WHP.

Limitations

When interpreting the results of the analysis, certain limitations must be considered. First, recruitment took place through the company process. Participation was approved on a first-come, first-serve basis, which may have introduced bias by involuntarily excluding participants. Additionally, self-selection bias by more motivated employees might have occurred. Regardless, the representativeness of the samples was tested separately for those who attended only one screening and those who completed two screenings. Besides app usage, the company applied other interventions, such as sports courses—during the same time period, but these were performed independent of the present study—which could have influenced the medical indicators. Regarding the calculated body age, the participants aged during the study; therefore, a standardized value of 1 year was subtracted from the difference between calculated body ages to reflect the aging between check-ups. Generally, the sample consisted of only two company samples derived from the same company. Finally, because the intervention was conducted in a practical setting, its effectiveness cannot be directly proven because other influencing factors might have played a role.

Conclusion

The study is among the first to use usage data of a WHP app and medical check-up results to evaluate the usage and effects of an mHealth intervention. The intervention was found to be effective in improving various health parameters, especially regarding body composition and the metabolic system. In a comparison of the activities performed in the app, the screening itself or the expectancy of an upcoming screening seemed to strengthen the motivation to use the app, which emphasizes the importance of medical check-up elements within technical health solutions. Interestingly, the differences in usage behaviors in the gender groups and age groups were further detected. Thus, the present study contributes to theories on app adoption and the potential motivators for initial app usage. Furthermore, the implications for implementation are also evident because an app intervention as a standalone product appears to be less valuable than an intervention that combines screening with technology and incentives for usage.

Footnotes

Declaration of conflicting interests

The other author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Maren Junker

Appendix

References

Auroy

Sarradon-Eck

. Patients’ perceptions of mHealth apps: meta-ethnographic review of qualitative studies. JMIR mHealth and uHealth 2019; 7: e13817.

Fiordelli

Diviani

Schulz

. Mapping mHealth research: a decade of evolution. Journal of Medical Internet Research 2013; 15: 15.

Steigner

Doarn

Schütte

, et al. Health Applications for corporate health management. Telemedicine and E-Health 2017; 23: 448–452.

Dehkordi

Breitschwerdt

Fellmann

. IT-support in workplace health promotion: mobile apps on the rise. In: International Conference on Exploring Services Science. Springer, 2017, pp. 38–50.

Lennefer

Lopper

Wiedemann

, et al. Improving employees’ work-related well-being and physical health through a technology-based physical activity intervention: A randomized intervention-control group study. Journal of Occupational Health Psychology 2019; 25: 143–158.

Asplund

Andersson

. Stress management for middle managers via an acceptance and commitment-based smartphone application: a randomized controlled trial. Internet Interventions 2014; 1: 95–101.

McCallum

Rooksby

Gray

. Evaluating the impact of physical activity apps and wearables: interdisciplinary review. JMIR Mhealth Uhealth 2018; 6: e58. DOI: 10.2196/mhealth.9054

Buckingham

Williams

Morrissey

, et al. Mobile health interventions to promote physical activity and reduce sedentary behaviour in the workplace: a systematic review. Digital Health 2019; 5: 205520761983988.

Becker

Matt

Widjaja

, et al. Understanding privacy risk perceptions of consumer health wearables–an empirical taxonomy. In: Thirty Eighth International Conference on Information Systems South Korea, 2017, p. 12.

10.

Greenfield

Busink

Wong

, et al. Truck drivers’ perceptions on wearable devices and health promotion: a qualitative study. BMC Public Health 2016; 16: 677.

11.

Giddens

Leidner

Gonzalez

. The role of fitbits in corporate wellness programs: does step count matter? In: Proceedings of the 50th Hawaii International Conference on System Sciences, 2017.

12.

Bond

Thomas

Raynor

, et al. B-MOBILE-A smartphone-based intervention to reduce sedentary time in overweight/obese individuals: a within-subjects experimental trial. PloS One 2014; 9: e100821.

13.

Ahtinen

Mattila

Välkkynen

, et al. Mobile mental wellness training for stress management: feasibility and design implications based on a one-month field study. JMIR mHealth and uHealth 2013; 1: e11.

14.

Carolan

deVisser

. Employees’ perspectives on the facilitators and barriers to engaging with digital mental health in the workplace: a qualitative study. JMIR Mental Health, 2017.

15.

Muuraiskangas

Harjumaa

Kaipainen

, et al. Process and effects evaluation of a digital mental health intervention targeted at improving occupational well-being: lessons from an intervention study with failed adoption. JMIR Mental Health 2016; 3: e13.

16.

Paschou

Papadimitiriou

Nodarakis

, et al. Enhanced healthcare personnel rostering solution using mobile technologies. Journal of Systems and Software 2015; 100: 44–53.

17.

van Drongelen

Boot

Hlobil

, et al. Process evaluation of a tailored mobile health intervention aiming to reduce fatigue in airline pilots. BMC Public Health 2016; 16: 894.

18.

Bregenzer

Wagner-Hartl

Jiménez

. Who uses apps in health promotion? A target group analysis of leaders. Health Informatics Journal 2019; 25: 1038–1052.

19.

Mattila

Orsama

A-L

Ahtinen

, et al. Personal health technologies in employee health promotion: usage activity, usefulness, and health-related outcomes in a 1-year randomized controlled trial. JMIR mHealth and uHealth 2013; 1: 1.

20.

Morris

Kathawala

Leen

, et al. Mobile therapy: case study evaluations of a cell phone application for emotional self-awareness. Journal of Medical Internet Research 2010; 12: 12.

21.

Dunkl

Jiménez

. Using smartphone-based applications (apps) in workplace health promotion: the opinion of German and Austrian leaders. Health Informatics Journal 2017; 23: 44–55.

22.

Guertler

Vandelanotte

Kirwan

, et al. Engagement and nonusage attrition with a free physical activity promotion program: the case of 10, 000 steps Australia. Journal of Medical Internet Research 2015; 17: 17.

23.

Gilson

Pavey

Wright

, et al. The impact of an m-Health financial incentives program on the physical activity and diet of Australian truck drivers. BMC Public Health 2017; 17: 467.

24.

Sari

Othman

Al-Ghaili

. A proposed conceptual framework for mobile health technology adoption among employees at workplaces in Malaysia. In: International Conference of Reliable Information and Communication Technology. Springer, 2018, pp. 736–748.

25.

Melzner

Heinze

Fritsch

. Mobile health applications in workplace health promotion: an integrated conceptual adoption framework. Procedia Technology 2014; 16: 1374–1382.

26.

Bol

Helberger

Weert

. Differences in mobile health app use: a source of new digital inequalities? The Information Society 2018; 34: 183–193.

27.

de Korte

Wiezer

Janssen

, et al. Evaluating an mHealth app for health and well-being at work: mixed-method qualitative study. JMIR mHealth and uHealth 2018; 6: e72. DOI: 10.2196/mhealth.6335

28.

Jung

Cho

. Promoting physical activity and weight loss with mHealth interventions among workers: systematic review and meta-analysis of randomized controlled trials. JMIR mHealth and uHealth 2022; 10: e30682.

29.

Stratton

Jones

Peters

, et al. Digital mHealth interventions for employees: systematic review and meta-analysis of their effects on workplace outcomes. Journal of Occupational and Environmental Medicine 2021; 63: e512–e525.

30.

Lowensteyn

Berberian

Belisle

, et al. The measurable benefits of a workplace wellness program in Canada. Journal of Occupational and Environmental Medicine 2018; 60: 211–216.

31.

Keen

PGW

. Shaping the future: business design through information technology. Cambridge: Harvard University Press, 1991.

32.

Kim

Park

H-A

. Development of a health information technology acceptance model using consumers’ health behavior intention. Journal of Medical Internet Research 2012; 14: 14.

33.

Junker

Böhm

Franz

, et al. Value of normative belief in intention to use workplace health promotion apps. BMC Medical Informatics and Decision Making 2022; 22: 1–11.

34.

Yang

Boulton

Todd

. Measurement of adherence to mHealth physical activity interventions and exploration of the factors that affect the adherence: scoping review and proposed framework. J Med Internet Res 2022; 24: e30817. DOI: 10.2196/30817

35.

Renahy

Parizot

Chauvin

. Determinants of the frequency of online health information seeking: results of a web-based survey conducted in France in 2007. Informatics for Health and Social Care 2010; 35: 25–39.

36.

Jonsdottir

Börjesson

Ahlborg

. Healthcare workers’ participation in a healthy-lifestyle-promotion project in western Sweden. BMC Public Health 2011; 11: 448.

37.

Venkatesh

Morris

. Why don’t men ever stop to ask for directions? gender, social influence, and their role in technology acceptance and usage behavior. MIS Quarterly 2000; 24: 115–139.

38.

Zhang

Guo

Lai

K-h

, et al. Understanding gender differences in m-health adoption: a modified theory of reasoned action model. Telemedicine and E-Health 2014; 20: 39–46. DOI: 10.1089/tmj.2013.0092

39.

Robroek

Lindeboom

Burdorf

. Initial and sustained participation in an internet-delivered long-term worksite health promotion program on physical activity and nutrition. Journal of Medical Internet Research 2012; 14: 14.

40.

Rongen

Robroek

van Ginkel

, et al. How needs and preferences of employees influence participation in health promotion programs: a six-month follow-up study. BMC Public Health 2014; 14: 1277.

41.

Zhao

Zhou

. What factors influence the mobile health service adoption? a meta-analysis and the moderating role of age. International Journal of Information Management 2018; 43: 342–350.

42.

Ueno

Yamashita

Moritani

, et al. Biomarkers of aging in women and the rate of longitudinal changes. Journal of Physiological Anthropology and Applied Human Science 2003; 22: 37–46.

43.

Han

Cho

Kwon

, et al. A mobile-based comprehensive weight reduction program for the workplace (health-on): development and pilot study. JMIR mHealth and uHealth 2019; 7: e11158.

44.

Eng

Moy

Bulgiba

. Impact of a workplace health promotion program on employees’ blood pressure in a public university. PloS One 2016; 11: e0148307.

45.

Lowensteyn

Berberian

Berger

, et al. The Sustainability of a workplace wellness program that incorporates gamification principles: participant engagement and health benefits after 2 years. American Journal of Health Promotion 2019; 33: 0890117118823165–0890117118823858.

46.

Jimenez

Bregenzer

. Integration of eHealth tools in the process of workplace health promotion: proposal for design and implementation. Journal of Medical Internet Research 2018; 20: 20.

47.

Balk-Møller

Poulsen

Larsen

. Effect of a nine-month web-and app-based workplace intervention to promote healthy lifestyle and weight loss for employees in the social welfare and health care sector: a randomized controlled trial. Journal of Medical Internet Research 2017; 19: e108.

48.

Vaghefi

Tulu

. The continued use of mobile health apps: insights from a longitudinal study. JMIR mHealth and uHealth 2019; 7: e12983.

49.

Statistisches Bundesamt . Arbeitsmarkt (statistisches jahrbuch 2019), 2019. https://www.destatis.de/DE/Themen/Querschnitt/Jahrbuch/jb-arbeitsmarkt.pdf?__blob=publicationFile: Statistisches Bundesamt.

50.

Gånedahl

Zsaludek Viklund

Carlén

, et al. Work-site wellness programmes in Sweden: a cross-sectional study of physical activity, self-efficacy, and health. Public Health 2015; 129: 525–530. DOI: 10.1016/j.puhe.2015.01.023

51.

Middlestadt

Sheats

Geshnizjani

, et al. Factors associated with participation in work-site wellness programs: implications for increasing willingness among rural service employees. Health Education & Behavior 2011; 38: 502–509.

52.

Huberty

Vranceanu

A-M

Carney

, et al. Characteristics and usage patterns among 12, 151 paid subscribers of the calm meditation app: cross-sectional survey. JMIR mHealth and uHealth 2019; 7: e15648.