Background
Depression is a common psychiatric disorder with a prevalence reaching more than 300 million people around the world [
1]. Its impact on patients’ psychosocial functioning and quality of life makes it a leading cause of disability worldwide [
2,
3]. Despite the existence of effective treatments, such as antidepressants and cognitive-behavioural therapy, almost half of patients with depression stay untreated [
4]. The appearance of computers and internet development in the ‘90s brought about the idea that technological devices could be used as therapeutic tools [
5‐
8] so as to improve treatment rates [
9].
In recent decades, the development of new technologies and their worldwide spread have led the way to new therapeutic and screening tools in mental health [
10]. This field, called “mobile health” or “m-health” [
11,
12], has seen exponential growth with more than 10,000 downloadable mental health smartphone applications (apps) associated with the extensive use of wearables such as smartbands or smartwatches [
13]. Although m-health is a promising field for increasing access to mental health programmes, their current use in clinical practice is limited and most of the available apps have not yet consistently demonstrated their effectiveness in the management of depression [
14]. Several causes may be considered regarding these two issues. First, most of the apps for depression fail to incorporate evidence-based practices or clinical expertise into their design [
15‐
18]. Conversely, most of the scientifically validated apps are not available in apps stores [
16,
19,
20]. Finally, m-health suffers from a low retention rate and engagement by users [
21,
22] and is rarely integrated into clinical practice, relegating apps to a “self-medication tools” status [
23,
24].
To overcome these issues, it could be of interest to implement a user-centred design approach to develop m-health tools to remain as close as possible to patient and physician expectations; this would facilitate both improved retention rates and app implementation in professional healthcare practice [
21,
23].
Qualitative analysis is commonly used to assess patients’ expectations in various domains. It has mostly been used to evaluate patients’ use of pre-existing apps for depression [
25,
26] or to explore the expectations of an app for depression in young people [
27,
28]. To the best of our knowledge, no qualitative study has investigated both patient and physician expectations of a smartphone app dedicated to depression including patients with previous major depressive episodes (MDE), general practitioners and psychiatrists.
Methods
Study design
The perceptions and expectations of patients, general practitioners and psychiatrists concerning a hypothetical smartphone app dedicated to depression were investigated by using a qualitative design with a focus group methodology. The focus group method has been chosen because it is a reliable way to assess the participants’ expectations by facilitating the sharing of ideas and experiences among them.
The focus groups were conducted between November 2018 and May 2019 in France. Patients and physicians were allocated to separate groups. Psychiatrists and general practitioners were distributed randomly in the physicians groups.
Sample and recruitment
Patients included were adults with a diagnosis of MDE in the last 12 months according to the Diagnostic and Statistical Manual of Mental Disorders - 4th edition, Text Revision (DSM-IV-TR) criteria [
29]. They were required to understand and be fluent in French.
Physicians were psychiatrists or general practitioners working in the private and/or public sectors and dealing with patients with MDE in their clinical practice.
Eligible participants (patients and physicians) were screened by the investigator centres hosted by academic Departments of Psychiatry (Clermont-Ferrand, Lyon, Grenoble).
Physicians were solicited by email or by phone to participate.
Patients were recruited among in- and outpatient services of investigator centres. Those who gave consent to be contacted were followed up to arrange participation. The focus groups were held in the centre where the participants were recruited.
The study was carried out in accordance with ethical principles for medical research involving humans (WMA, Declaration of Helsinki). The assessment protocol was approved by the relevant ethical review board (CPP EST I, 2018-A01469–46). All subjects provided written informed consent to participate.
Data collection
After a literature search on m-health and smartphone apps for depression, mirrored semi-structured interview guides have been established for patients and physicians (Additional file
1).
Before starting the session, patients completed a survey with their sociodemographic information, including age, gender and living place (urban or rural). They also indicated their smartphone and app use habits. Finally, the severity of their depressive symptoms was measured using the Inventory of Depressive Symptomatology (IDS-SR) questionnaire. The IDS-SR is a 30-item self-rated questionnaire assessing all the criterion symptom domains designated by the DSM-IV-TR [
30].
Physicians’ age, gender, type of practice (public or private), number of visits per week, number of patients with depression seen each week and smartphone and app use habits were assessed through a questionnaire.
Each focus group included 6 to 8 participants and lasted from 60 to 90 min. There was an interviewer and an observer present for each group (LS, MCP or OB), all familiar with and well trained in the focus group methodology. All focus groups were audio recorded and transcribed verbatim.
Analysis
Data collection and analysis were conducted at the same time in accordance with established qualitative methodologies [
31]. After a focus group was transcribed verbatim, it was fully read then independently and manually coded by two researchers (LS and MCP). To gain familiarity with the content, the transcripts were read several times. Each unit of text was then coded, a code being defined as a meaningful unit describing a section of text (for example, the code “Helping patient’s self-evaluation” described the following text section: “With a self-administered survey on the app, the patient could do self-assessments”). Codes were organised into categories (for example: “data provided by the app”, including codes such as “exercises” or “therapeutic tools”) and themes (for example, “content of the app”, including categories such as “features” or “data collected by the app”). The codes, categories and themes were compared and agreed upon among the research team. In the case of discrepancies between researchers, agreements were reached by individually clarifying the meaning of a code and discussing its interpretation until mutual consent was achieved. If necessary, the codes, categories and themes were updated
. Team meetings were held to discuss and monitor coding consistency and to address the analytic validity of the identified themes. Moreover, the research team met to ensure that the findings were internally consistent and supported by the data from the participants’ interviews. After four patient focus groups and four physician focus groups, no new codes or categories were emerging, indicating the reaching of data saturation.
Patients and physicians respective focus groups have been analysed separately to identify the discrepancies between them. The two codebooks were then merged into a single codebook.
Sociodemographic data of the sample are presented as the mean (Standard Deviation, SD) for continuous variables and frequency distribution for categorical variables.
Discussion
This qualitative study is the first to assess both physician and patient expectations of an app dedicated to depression. The use of the focus group method provided a range of experiences and opinions among the participants and a relationship of trust among the group. The discussions it allows increased the role of the participants who collectively built the results of the research.
All themes and categories were the same for patients and physicians, highlighting a shared interest and mutual needs regarding this tool. However, some code differences pointed out the potential conflict between patients’ needs and physicians’ constraints. Direct access to a professional through the app, which was a strong wish expressed in the patient group, was not mentioned in the physician group and echoed the worry about physicians’ availability to use an app. Nevertheless, these conflicts remained scarce and this study revealed strong similarities between patient and physician expectations, revealing that an app suitable for both patients and physicians could easily be developed. Patients and physicians expressed expectations regarding its content, its operating characteristics and discussed potential barriers to its implementation in real-world clinical practice. Content considered by participants included data provided to and delivered by the app, as well as features thought to be useful for that kind of tool. Data collection methods must gather information about the patient using the app to evaluate their mental state and to inform physicians of the evolution of depressive symptoms. Data delivered by the app should provide psychoeducation elements, therapeutic tools and various functionalities to aid the management of daily life. Features considered for the app were meant to facilitate its use by physicians and ensure patient care in the case of an emergency. The “operating characteristics” theme showed strong heterogeneity between participants’ expectations regarding target users, frequency of use and aims of the app. This heterogeneity emphasised the interest of a tailorable tool to meet all the needs and desires of patients and physicians. Finally, this study highlighted doubts and limitations that both patients and physicians may have regarding an app dedicated to depression. These barriers included concerns about users of the app, its accessibility, safety, side effects, utility and functioning.
Most of our results are consistent with previous qualitative studies on health apps, whether they focus on apps for depression or not. Regarding content, information on the disorder and day-to-day life-supporting activities like music, breathing exercises and videos are often cited as expected items [
27,
32,
33]. Self-tracking is a highly rated activity of health apps [
27,
32] and is described by patients as a reason to return to the app [
34]. Similarly to our findings, among possible app features, personalisation is the most requested and liked, both by patients and professionals [
26,
27,
32‐
35]. The expected aims of such apps found in the literature included psychoeducation [
35], monitoring [
27,
32,
35], providing a presence between face-to-face visits [
27] and social connectivity [
26,
32]. Consistently with our result, other studies highlighted two main barriers to the use of apps cited by physicians: lack of time and medicolegal responsibility [
27,
28]. Financial aspects and deficiency in technological competencies are common barriers to the use of apps for patients [
26,
33,
35].
An important issue considered by the participants in our study was the type of population that could benefit from an app dedicated to depression. First, in line with other studies, patients and physicians agreed to focus on an app for people suffering from mild-to-moderate depression [
24]. This statement tended to be confirmed by a recent meta-analysis focusing on the efficacy of app interventions for depressive symptoms, the post hoc subgroup analysis showing that significant benefits from smartphone apps were only found for patients with self-reported mild-to-moderate depression [
36]. Those results should be considered carefully, with the variations in subgroup sample sizes leaving the analyses for major depression underpowered to detect significant effects. Moreover, a more recent study identified that more severe depression led to enhanced information seeking, counteracting the theory that severe depression keeps patients from using apps [
37].
An app to screen potential depression in the general population was also discussed. Participants emphasised the significant role that this app could provide in facilitating access to mental health for depressed people. They assumed that the large-scale use of such an app could decrease the mean duration of untreated depression and hence the recurrences and the duration of MDEs [
38,
39]. In line with our results, several studies pointed out the possible interest of apps for depression screening in the general population, showing that a large number of people from different countries were searching for, and willing to use, that kind of tool [
40]. Additionally, several apps using text analysis have shown their ability to improve the immediate detection of depressive symptoms [
41,
42]. Finally, several studies highlighted the fact that depression screening apps could motivate some users to discuss the obtained results of the tests with healthcare professionals for further diagnosis and management [
43,
44].
Our findings also emphasised the interest in an app that would not be time-consuming for physicians and could help treatment decision-making in patients with depression based on the most updated and high-quality evidence. To the best of our knowledge, there are very few apps currently available providing evidence-based guidance for treatment decision-making to physicians.
One study highlighted that an app could be an effective tool for both increasing confidence in depression treatment and educating physicians [
45], pointing out the interest to develop more connected tools for healthcare professionals.
Finally, participants mentioned the usefulness of an app for informal caregivers, to inform them, help them in supporting their ill relative and to destigmatise mental health. This demand is supported by the results of a recent systematic review focusing on apps dedicated to caregivers, showing most of the included studies proved their effectiveness in the overall well-being of the caregivers [
46]. Apps dedicated to caregivers, who are often suffering themselves from depression or anxiety, could significantly improve mental healthcare regarding their essential role for patients suffering from depression [
47].
These findings highlight that a single app is not enough: multiple versions of the app are needed to encompass the support and care objectives of patients, mental health professionals and informal caregivers.
One crucial feature expected by participants was self-monitoring. The main interest of it is to improve mental health and wellbeing by increasing emotional self-awareness [
48,
49]. The use of apps for self-monitoring allows precise, easy and quick ecological momentary assessment with the possibility of providing real time feedback for patients and alerts for clinicians in case of emergencies. It would also allow clinicians to monitor the efficacy of treatment over time, predict short-term mood changes and detect the worsening of symptoms early on [
50]. There are many apps for depression integrating a self-monitoring feature and several studies have examined their usability, acceptability, adherence and effectiveness. Self-monitoring of depressive symptoms on patients’ phones has been shown to be easy and reliable [
51,
52] and several studies highlighted its effectiveness to improve depressive symptoms [
36,
53,
54]. A study on untreated patients with symptoms of depression and anxiety also showed that access to daily self-monitoring helped them to translate their intention to seek treatment into actual treatment-seeking behaviour [
55]. Even if this effect was small, it defines self-monitoring as a promising tool to decrease the number of untreated patients suffering from depression. However, the main bottleneck of self-monitoring is the low retention rate of apps offering this feature. For this particular point, studies showed inconsistent results, where some highlighted a quick drop out rate of self-monitoring apps [
56,
57], while others noted, conversely, a fairly high retention rate for these tools [
52,
54,
58].
In addition to self-monitoring and patients’ active input of data, passive data collection has been suggested by physicians in our study. This is consistent with the new research allowed by advanced technologies such as digital phenotyping [
59], aiming to determine clinical phenotypes by measuring patient behaviours from smartphone sensors [
60]. Requiring no active participation from the patient, the collection of passive data has been increasingly studied in recent years and seems to be a promising field for the future of m-health. This method could indeed allow ecological momentary assessment of several parameters and could be a strong tool to improve depression care. An interesting application domain for this method is the use of an algorithm to facilitate the detection of new MDEs that could be of significant help for clinicians in the follow up of patients with mood disorders [
61‐
64]. A study focusing on therapists’ ability to detect negative changes in their patients showed that clinical judgement allowed the detection of only 21% of symptom worsenings [
65]. Early detection of an episode with passive data collection could then facilitate the quick reaction of physicians and improve patients’ outcomes [
66]. Despite being promising, passive data collection is, as shown in our study, not yet accepted by everyone, whether patients or physicians, and remains a strong barrier to the use of apps.
The analysis of the barriers identified the potential replacement of the physician by the app as a major concern in psychiatry, where state-of-the-art methods require human interaction. This barrier is raised in several studies who mentioned the lack of therapist contact as a negative point of apps [
24,
34,
67]. This issue could be overcome with the use of mixed methods or adjunctive apps, integrating the app in face-to-face therapy. This kind of approach has shown better efficacy than self-guidance therapy only through web intervention or smartphone apps [
68‐
71], thus demonstrating the essential role of the therapist in patient care. All these elements emphasised an essential point: complementarity between therapists and m-health tools. In Western countries, the main aim of these devices is to support existing care by providing, for example, better information regarding the patient’s daily state through momentary ecological assessment. The use of new technologies should not be seen as a replacement tool for physicians but as an opportunity to increase the quality of care provided. The inclusion of apps in therapy could then be compared to the development of imagery devices in radiology, improving diagnosis without removing the need for clinical examination and physicians’ knowledge and expertise.
Perspectives
The key findings of this study allow us to build a list of suggestions for app developers in the field of depression to fill both patients’ and physicians’ expectations:
-
The use of the app should be easy and intuitive.
-
The app should be personalised. The content and functioning of the app should be tailored to each patient and should adapt to the patient’s condition over time.
-
A self-monitoring function should be included to both increase the patient’s self-awareness and sharpen the evaluation of the physician. This function should focus on key symptoms and offer the patient and the physician the possibility of choosing other symptoms to monitor.
-
The app should be able to deal with emergency situations. At the least, it should include a tailorable crisis procedure and, at best, include a partnership with the local emergency service.
-
The app should provide the patient with information about depression and/or psychoeducational messages.
Furthermore, our results highlighted several key points that should remove the potential barriers to the use of the app:
-
Relevant users should be precisely targeted to ensure their capacities to use the app, with particular attention given to the intensity of depression.
-
To reduce the potential difficulties bound to the use of the app, physicians should plan a dedicated educational time with the patient to explain the functioning of the app. A user guide should also be included and delivered to the patient.
-
Mixed approaches should be preferred. The app should be fully integrated with the usual therapy and be an adjunctive tool rather than an independent one. The use of the app should not lead to any reduction of the frequency of visits to the therapist.
-
The use of the app should be free of charge.
-
Access should be protected by a password and data stored in a secured server.
Strengths and limitations
This study has a number of limitations. Selection bias may have occurred: patients’ recruitment was limited to France for practical reasons. Furthermore, most of the physicians practiced in urban areas. Therefore, the findings may not be transferable to practitioners in rural areas. This study was also limited to some degree in the use of focus groups as a methodology: group dynamics might have, in some way, shaped the expectations expressed by participants and the interviewers’ personal skills and attributes could also have influenced the nature and quality of the gathered data. However, the use of focus group is also one of the strengths of this study as it allows for interaction among participants and facilitates discussion and sharing of ideas.
Competing interests
Dr. Pacchiarotti has received CME-related honoraria, or consulting fees from ADAMED, Janssen-Cilag and Lundbeck.
Dr. Murru has served as a consultant, adviser or speaker for Adamed, AstraZeneca, Bristol-Myers Squibb, Janssen-Cilag, Lundbeck, Otsuka and Sanofi-Aventis and has received a grant (PI19/00672) from the Instituto de Salud Carlos IIISubdirección General de Evaluación y Fomento de la investigación, Plan Nacional 2019–2022.
Prof. Vieta has received grants and served as a consultant, advisor or CME speaker for the following entities: AB-Biotics, Abbott, Allergan, Angelini, AstraZeneca, Bristol-Myers Squibb, Dainippon Sumitomo Pharma, Farmindustria, Ferrer, Forest Research Institute, Gedeon Richter, Glaxo-Smith-Kline, Janssen, Lundbeck, Otsuka, Pfizer, Roche, SAGE, Sanofi-Aventis, Servier, Shire, Sunovion, Takeda, the Brain and Behaviour Foundation, the Spanish Ministry of Science and Innovation (CIBERSAM), the EU Horizon 2020 and the Stanley Medical Research Institute.
Prof. Llorca has received grants, honoraria or consulting fees from ESAI, Gedeon Richeter, Janssen, Lundbeck, Otsuka and Sanofi.
Dr. Samalin has received grants, honoraria or consulting fees from AstraZeneca, Bristol-Myers Squibb, Janssen-Cilag, Lundbeck, Otsuka, Sanofi-Aventis, and Takeda.
The other authors declare that they have no conflicts of interest.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.