The mobile app platform was developed by technical experts, is currently used, and has previously been used in studies of postoperative outcome with a high rate of use [1]. A multidisciplinary team of health care professionals and technical experts developed a proof-of-concept version of the app that included specific adjustments for an older population of patients. A digital questionnaire consisting of indicators of quality of care from the Dutch Hip Fracture Audit was developed to remotely monitor postoperative functional outcome at 3 months. This questionnaire was implemented in the mobile app, and the technology was pretested with 2 patients with fractures of the hip who had been chosen at random.

Patients with a hip fracture who had undergone surgical treatment between July 2017 and December 2017 at the Centre for Geriatric Traumatology of the Department of Trauma Surgery at Ziekenhuisgroep Twente hospital were recruited to participate in the study and asked to download the app in addition to their regular 3-month outpatient visit (the recruitment process is summarized in Figure 1). The population of interest consisted of older adults, among whom information and communication technology literacy or low motivation to use technology may be factors that hinder implementation of a mobile app and which could suggest the need to focus on patient spouses in addition to the patients themselves [19]. For the purpose of this study, both patients and spouses who decided to participate were considered participants. During admission to the surgical ward of the Centre for Geriatric Traumatology, a nurse informed potential participants about the study, use of the app, and how to download instructions for using the app. After verbally providing informed consent, participants received an information leaflet and provided their email address for further information.

One week later, participants received a code by mail to activate the questionnaire in the downloaded app. Completion of the questionnaire was restricted to a period between 12 weeks and 18 weeks after their operation. A push notification with a request to complete the questionnaire was sent to the participant 12 weeks after they had been discharged from the hospital. A push notification was also sent to the health care provider at 17 weeks for unfilled questionnaires.

Completed questionnaires were saved in OpenLine (a specialized health care hosting center) in accordance with Dutch legislation with respect to security standards. The local researcher applied for the data from the hosting center. Participants were anonymized and coded using a study number without any reference to patient number or date of birth. Only the local researcher had access to the participant study numbers. All data were treated confidentially and saved to the secured hospital network with a password.

To investigate usability and participant acceptance of the mobile app, an interview questionnaire was developed (Multimedia Appendix 1) based upon the Unified Theory of Acceptance and Use of Technology [20]; the model investigates user intentions and usage behavior in technology systems [20].

Two questions regarding participant recollection of the intended purpose of the study and feedback on the use of the app were added to the interview. These questions were added because we were interested in obtaining feedback on the app and on the duration of the interval from when the information was given (from July 2017 to December 2017) to when the telephone interview took place (in May 2018). A single researcher conducted all interviews. Participants were given the option to stop the telephone interview at any time.

Data were collected from the clinical charts of the patients who participated themselves or whose caregivers participated. Age, gender, type of fracture, American Society of Anesthesiologists physical classification status, Charlson Comorbidity Index [21], dementia, prefracture Katz Index of Independence in Activities of Daily Living score [22], prefracture mobility score, and prefracture living situation were recorded as baseline characteristics. In April 2018, the app usage data from the hosting center were collected. Participants were divided into two groups—those who downloaded the app (use group) and those who did not download the app (nonuse group). Mobile app usability and acceptance telephone interviews were conducted with participants who could be reached by telephone within 3 attempts. The number of participants in both groups was adjusted to the lowest number of participants accessible by telephone of either group (use group, n=24); therefore, in the nonuse group, 24 participants were selected randomly. Participant answers were fully transcribed in individual and anonymized Office Word (version 2007; Microsoft Inc) documents and saved on a secure hospital server.

Statistical analyses were performed using SPSS software (version 22.0; IBM Corp). We used thematic analysis with a deductive theoretical approach to analyze the written answers to the recalled purpose of the study and feedback questions [23]. Identification of patterns and themes within the data was performed by one researcher, and a second researcher was consulted to reach agreement; the data were then coded by themes. Categorical data were analyzed using the chi-square test or Fisher exact test when appropriate (ie, Fisher exact test was used when frequency was less than 5). Functional outcomes were analyzed using two-tailed paired t tests. Continuous data were analyzed using two-tailed independent t tests. If significant differences were found in categorical variables with two or more subgroups, Pearson chi-square test was performed post hoc. P<.05 was considered statistically significant.

This study was been designated as an observational study not subject to the Dutch Medical Research Involving Human Subjects Act by the appropriate medical research ethics committees.

Categorical variables are described as numbers with corresponding percentages. Continuous variables are described as the mean with standard deviation, or for nonparametric data, as the median with interquartile range.

Patient characteristics are shown in Table 1. Patients with fractures of the hip (N=110) were a mean age of 80.5 (SD 10.4) years and were 71.8% (79/110) female and 28.2% (31/110) male. No significant differences were found between those who downloaded the app and those who did not download the app for age (P=.21), gender (P>.999), type of fracture (P>.999), American Society of Anesthesiologists physical classification status (P>.999), Charlson Comorbidity Index (P>.999), dementia (P=.05), prefracture Katz Index of Independence in Activities of Daily Living score (P=.10), prefracture mobility score (P=.10), and prefracture living situation (P=.73).

Of the participants (29/110, 26.4%) who downloaded the mobile app, only 1 (1/29, 3.4%) completed the app questionnaire.

Completion of 3-month mandatory functional monitoring is poor among patients with fractures of the hip, which may result in a suboptimal monitoring of quality of care. This single-center pilot study to investigate the use and to analyze the usability and acceptance of a mobile app for monitoring postoperative functional recovery after hip fracture revealed poor results for actual use of the mobile app despite high self-reported intention to use the mobile app, high self-reported expertise in using mobile apps, and conditions that facilitated the use of mobile apps. This suggests that participants had the goal of using the mobile app, but that better support was needed to properly implement the technology in health care.

For many years, apps have been regarded as an alternative to paper questionnaires, but the use of apps may have difficulties as well, especially when implemented in a population of community-dwelling older patients [16]. This study demonstrated implementation difficulties; only 26.4% (29/110) participants downloaded the mobile app. This demonstrated that implementation of the app may have required that sufficient attention be given to education of the community-dwelling older patient users.

The low percentage of app downloads could partially be explained by an inability of the patients or caregivers to correctly remember the information that was provided to them in the hospital possibly as a result of stress [24]. Receiving information in a state of stress has been associated with suboptimal information processing and reduced cognitive efficiency [25,26]; therefore, correct timing of information provision is essential. This study provided both oral and written information, but more emphasis should be given to written information or video instructions, as this has been shown to lead to better information retention [27]. Among participants who are elderly, an inverse correlation has been reported between age and recall of medical information which could also have influenced the findings of this study [24,28]. The 3-month time period between when the information was provided and when the questionnaire was to be completed which also could have negatively affected information recall and recollection of the study’s purpose.

One participant completed the app questionnaire after downloading the app. This participant showed an active approach by contacting the app developers and completed the questionnaire with assistance from the developers.

A high percentage of the participants (34/48, 70.8%) who were interviewed were caregivers who were first-degree relatives of the patient. Study information was provided independently of whether a caregiver was present at the time of information provision; therefore, it is possible that some first-degree relatives were not provided with the study information if they were absent during recruitment.

The telephone interview findings demonstrated that many in the use group had the intention of completing the questionnaire. This indicates that those participants were motivated to complete the app questionnaire. In the nonuse group (11/24, 45.8%), participants remembered the study, and 10 out of the 11 intended to download the app. Given this result, there seems to be a good level of intention in both groups. Facilitating conditions, such as facilitated help, were high in both groups and were not a restrictive factor for app usage [29]. Some participants in the nonuse group (13/24, 54.2%) were unable to remember the study, and they could not complete the interview. Difficulties in patients or caregiver recollection of study information may have been influenced by the previously noted patient-related factors such as cognitive impairment, anxiety, or stress [24]. Approaching multiple caregivers when providing information and conducting the telephone interview may also be a reason for some participants reporting that they did not remember the study. Respondents (18/48, 41.6%) also reported technical problems. The app developers suggested the start-up phase of the app as a possible explanation for the technical problems. The developers also suggested that a lack of received notifications could have been as a result of participants not enabling the appropriate permissions for notifications when downloading the app. Providing help in the hospital with downloading of the app could assist with this issue. Another way to decrease the frequency of technical problems while also optimizing usability and acceptance would be to frequently evaluate the mobile app during the implementation process [16].

Findings revealed intention to use the mobile app, but very low actual usage. The use of a mobile app as it was implemented in this study was not feasible, but the study findings suggested a potential for use if implemented properly. First, technical issues should be solved, and a helpdesk should be made available. Second, it is recommended to involve participants in the development and implementation phases—doing so can optimize ease of use and acquiring feedback during implementation is a feasible goal. Third, information provision needs to be optimized in terms of timing and method of dissemination. It is important to supply additional information after discharge in order to prevent low download rates as a result of patient or caregiver stress during admission [27]. Written information, video instructions, or fact sheets are preferred to oral information [2,3]. Fourth, in studies involving caregivers, a single contact person is recommended.

Selection bias in the downloading group represents a threat to validity, as patients or caregivers already intended to participate in the study by downloading the app.

The use of a mobile app to monitor 3-month postoperative functional outcome of hip fracture was low. Despite intention, expertise, and sufficient facilitating conditions for using smartphones, the implementation of the mobile app in this study was demonstrated to be infeasible. Reasons for this included a technical problem, the implementation process, and population of interest having cognitive disorders or a dependency on caregivers for mobile technology.