Digital Health Applications in the Area of Mental Health

Abstract

Background

The introduction of digital health applications (DiGA) is a fundamental innovation in Germany. In the field of mental health, numerous applications are already available whose efficacy has been tested in clinical trials. We investigated whether, and to what extent, the use of DiGA can be recommended on the basis of the available evidence.

Methods

In this scoping review, we summarize the evidence supporting the use of DiGA in the mental health field through an examination of relevant publications that were retrieved by a systematic literature search. We provide an annotated tabular listing and discuss the current advantages of, and obstacles to, the care of mentally ill patients with the aid of DiGA.

Results

We identified 17 DiGA for use in depression, anxiety disorders, addiction disorders, sleep disorders, stress/burnout, vaginismus, and chronic pain. These DiGA have been evaluated to date in 3 meta-analyses, 39 randomized controlled trials (RCTs), and two single-armed intervention trials. 23 of the 36 published trials were carried out with the direct participation of the manufacturers. 29 of the 39 RCTs were not blinded or contained no information regarding blinding. Active controls were used in 6 of the 39 RCTs. The reported effect sizes, with the exclusion of pre–post analyses, ranged from 0.16 to 1.79.

Conclusion

Most of the published studies display a high risk of bias, both because of the manufacturers’ participation and because of methodological deficiencies. DiGA are an increasingly important therapeutic modality in psychiatry. The available evidence indicates that treatment effects are indeed present, but prospective comparisons with established treatments are still entirely lacking.

cme plus

This article has been certified by the North Rhine Academy for Continuing Medical Education. Participation in the CME certification program is possible only over the internet: cme.aerzteblatt.de.

The deadline for submission is 23 November 2024.

In Germany, the legal framework for digital therapeutics was created with the adoption of the Digital Healthcare Act (DVG, Digitale-Versorgungs-Gesetz) and the Digital Health Applications Ordinance (DiGAV, Digitale-Gesundheitsanwendungen-Verordnung) in 2019. Since 2020, manufactures have been able to submit applications to the German Federal Institute for Drugs and Medical Devices (BfArM, Bundesamt für Arzneimittel und Medizinprodukte) for listing of their digital health applications (DiGA, German: digitale Gesundheitsanwendungen) in the DiGA Directory. DiGA are subject to mandatory CE conformity marking as a medical device. Currently, the BfArM’s three-month fast-track process allows for provisional listing of a DiGA in the DiGA Directory; within one year, the manufacturer then has to provide evidence of positive healthcare effects of the DiGA. So far, 40 DiGA have been included in the DiGA Directory, of these 16 in the area of “Mental Health“ and one in the area of “Stress/Burnout” (e1). A definition of DiGA is provided in eBox 1.

eBox 1. What is a digital health application (DiGA)*?

• The app is a medical device of risk class I or IIa.

• The main function is mostly based on digital technologies.

• The medical purpose is largely achieved through the digital main function.

• It supports the detection, monitoring, treatment, or alleviation of diseases or the detection, treatment, alleviation, or compensation of injuries or disabilities.

• The app is used by patients or jointly used by service providers and patients.

*Content adapted from: Bundesministerium für Arzneimittel und Medizinprodukte: Wissenswertes zu DiGA. (2023) (e31)

Once listed in the DiGA Directory, DiGA can be prescribed by physicians or psychotherapists or, after the diagnosis has been established, directly by the statutory health insurance. According to the German National Association of Statutory Health Insurance Funds (GKV, gesetzlichen Krankenversicherungen), statutorily insured persons used DiGa at least 164 000 times between 2020 and 2022, including at least 51 000 times in the “mental health” field (e1). Since private health insurances are not legally obliged to reimburse DGA and the regulations governing the inclusion of new services in existing insurance policies are not standardized, they pursue different reimbursement strategies (e2).

Manufactures have to provide evidence that the DiGA is able to deliver positive healthcare effects; for provisional listing in the DiGA Directory under the fast-track process, only results from unpublished pilot studies are required (e3).

There must be either a medical benefit (for example, improvements in morbidity, mortality and quality of life) or improvements in structure and processes (for example, improvements in coordination of treatment procedures or health literacy of patients) (e3). Health economic parameters are not assessed, nor does the Federal Joint Committee (G-BA, Gemeinsamer Bundesausschuss) evaluate the benefits and necessity in comparison to established methods of treatment. Thus, the approval requirements are not as strict as those for methods of psychotherapy (e4, e5).

Since DiGA constitute medical devices, the specialist standard applies: When prescribing DiGA, the potential risk should be assessed and patients should be advised on suitable applications and how they work. It is therefore important for treatment providers to understand the purposes and functions of the applications (e6). Surveys among doctors in Germany have revealed that although 24% of them are willing to prescribe DiGA (e7), 56% feel ill-equipped to provide comprehensive app-related advice (e8).

Thus, the aim of this review is to provide an overview of the DiGA currently available for prescription in the mental health field with regard to the areas and types of use and the available supporting evidence. The combination of qualitative and quantitative aspects in our research question and the heterogeneity of the evidence landscape were the reasons why we chose the format of a scoping review.

Method

We extracted the outcomes from the PubMed database, the DiGA Directory, the manufacturer‘s websites and the German Clinical Trials Register, based on the extended PRISMA guidelines (e9). Details of the methodology used are provided in eBox 2; the PRISMA 2020 flow diagram is provided in the eFigure.

eBox 2. Supplementary information about Methods.

This scoping review was conducted following the extended PRISMA guidelines. No review protocol was prepared.

Identification of studies: Based on the DiGA Directory (DiGA, digitale Gesundheitsanwendungen) of the German Federal Institute for Drugs and Medical Devices (BfArM) (as of 30 December 2022), data were systematically extracted by two investigators independently according to the PICO framework. DiGA from the areas “Mental Health” and “Other—Stress and Burnout“ were included. In addition, information on the manufacturers’ websites was taken into account; efficacy studies retrieved by a systematic literature search in the PubMed database (Medline) were also included. The names of the applications were used as search terms. In the case of several applications from the same manufacturer, the name of the indication was added. We used the following search terms: “Deprexis”, “Selfapy”, ”Selfapy” AND Depression, ”Novego”, ”Elona”, “Edupression”, “Invirto”, “Mindable”, “Selfapy“ AND Anxiety, “Velibra“, “Hellobetter”, “Hellobetter” AND panic, „Nichtraucherhelden”, “Vorvida”, “Somio”, “Hellobetter” AND sleep, “Hellobetter” AND stress, “Hellobetter” AND vaginal, “Hellobetter” AND pain. The resulting publications were selected by screening first the title to exclude duplicates, and then the abstract for subject comparison. If the data provided in the DiGA Directory could not be unambiguously matched with studies identified in PubMed, a cross-check was carried out by means of a targeted search in the German Clinical Trials Register (DRKS, Deutsches Register für Klinische Studien). All clinical intervention studies and meta-analyses with a clear reference to the identified DiGA were included in this review.

Data extraction and presentation: Relevant variables on application types and prerequisites, areas of indication and contraindications as well as aspects of therapeutic content were identified. The identified studies were assessed with regard to publication status, manufacturer influence, study type, inclusion criteria, sample size, intervention, control, primary endpoint, sex distribution, retention during the intervention period, blinding, method of analysis, effect size, and additional special features in order to evaluate the quality of the evidence. As far as possible with the available data, effect sizes were extracted in form of Cohen’s d (d) or Hedges‘ g (g) with 95% confidence intervals, otherwise in form of another effect measure as indicated in the publication. Where available, data of the intent-to-treat analysis was shown and highlighted, or else the other analysis principle present. If there were inconsistencies between the results in the DiGA Directory and the respective publication, the authors and the manufacturer were contacted. Effect sizes were classified according to Cohen as small for d between 0.2 and 0.5, medium for d between 0.5 and 0.8 and large for d greater than 0.8. The results were summarized in table format.

eFigure.

PRISMA 2020 – Flow diagram of the selection strategy

*Studies for which the data are only listed in the information for healthcare professionals section of the DiGA Directory

DiGa, digital health application (digitale Gesundheitsanwendung);

RCT, randomized controlled trial

Results

Of the 17 DiGA in the area of “mental health“, eleven apps achieved final listing in the DiGA Directory, while six are currently in the testing phase (provisional listing) (Table 1, eTable 1). The costs of DiGA range from 192 euros to 620 euros for the initial prescription. Six DiGA are also available in languages other than German.

Table 1. Areas of indication and types of application of digital health applications (DiGA) (excerpt: Indication Depression, eTable 1).

Deprexis	Selfapy’s online course for depression	Novego: Coping with depression	Edupression.com	Elona therapy Depression
Type of application
Web application*1	Web application*1 + Smartphone app*2	Web application*1	Web application*1	Web application*1 + Smartphone app*2
Listing status
Final	Final	10.10.2021–09.10.2022*3	26.12.2022–25.08.2022*3	26.12.2022–25.12.2023*3
Indications
● Mild to severe depressive episode (F32.0, F32.1, F32.2) ● Recurrent depressive disorder (F33.0, F33.1, F33.2)	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1)	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1) ● Dysthymia (F34.1)	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1)	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1) ● Dysthymia (F34.1)
Contraindications
● Schizophrenia (F20) ● Acute transient psychotic disorder (F23) ● Schizoaffective disorders (F25) ● Bipolar disorder (F31) ● Other symptoms and signs involving emotional state (R45.8) ● Suicidal risk	● Alzheimer dementia (F00) ● Alcohol dependence and withdrawal syndrome (F10.2, F10.3) ● Schizophrenia (F20) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Acute suicidality	● Organic mental disorder (F0x) ● Dependence disorder (F1x.2) ● Psychotic disorder (F2x) ● Bipolar disorder (F31) ● Mania (F30.0) ● Suicidality in the last 4 weeks	● Schizotypal disorder (F21) ● Persistent delusional disorders(F22) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Suicidality ● Psychotic disorder (F2x)	● Psychotic disorder (F2x) ● Bipolar disorder (F31) ● Mania (F30) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Acute suicidality
Duration of use
At least 90 days (1–2 x/week minimum duration 30 min each)	At least 90 days	At least 90 days	At least 90 days	At least 90 days
Costs/quarter (€)
210.00	217.18	249.00	357.00 (follow-up prescription: 178.50)	535.49
Foreign languages
ELL, ENG, FRA, ITA, POR, SPA, SWE, ZHO	–	–	ENG	–

*¹ Web application (internet-enabled end device)

*² Smartphone app (Android/ iOS smartphone)

*³ In case of provisional listing in the BfArM‘s DiGA Directory, the trial period is stated.

ELL, Greek; ENG, English; FRA, French; ITA, Italian; POR, Portuguese; SPA, Spanish; SWE, Swedish; ZHO, Chinese

eTable 1. Areas of indication and types of application of DiGA.

DiGA	Type of application	Listing status	Indications	Contraindications	Duration of use	Costs (€) per quarter	Foreign language
Depression
Deprexis	Web application*1	Final listing in the DiGA Directory of the German Federal Institute for Drugs and Medical Devices (BfArM)	● Mild to severe depressive episode (F32.0, F32.1, F32.2) ● Recurrent depressive disorder (F33.0, F33.1, F33.2)	● Schizophrenia (F20) ● Acute transient psychotic disorder (F23) ● Schizoaffective disorder (F25) ● Bipolar disorder (F31) ● Other symptoms and signs involving emotional state (R45.8) ● Suicidal risk	At least 90 days(1–2 x per week, minimum duration 30 min each)	210.00	ELL, ENG, FRA, ITA, POR, SPA, SWE, ZHO
Selfapy’s Online course for depression	Web application*1 + smartphone app*2	Final listing in the DiGA Directory	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1)	● Dementia in Alzheimer dis. (F00) ● Alcohol dependence and withdrawal syndrome (F10.2, F10.3) ● Schizophrenia (F20) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Acute suicidality	At least 90 days	217.18	–
Novego: coping with depression	Web application*1	Provisional listing in the DiGA Directory; trial period: 10.10.2021–09.10.2022	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1) ● Dysthymia (F34.1)	● Organic mental disorder (F0x) ● Dependence disorder (F1x.2) ● Psychotic disorder (F2x) ● Bipolar disorder (F31) ● Mania (F30.0) ● Suicidality in the last 4 weeks	At least 90 days	249.00	–
Edupression.com	Web application*1	Provisional listing in the DiGA Directory; trial period: 26.12.2022–25.08.2022	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1)	● Schizotypal disorder (F21) ● Persistent delusional disorder (F22) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Suicidality ● Psychotic disorder	At least 90 days	357.00 Follow-up costs: 178.50	ENG
Elona therapy Depression	Web application*1 + smartphone app*2	Provisional listing in the DiGA Directory; trial period: 26.12.2022–25.12.2023	● Mild and moderate depressive episode (F32.0, F32.1) ● Recurrent depressive disorder (F33.0, F33.1) ● Dysthymia (F34.1)	● Psychotic disorder (F2x) ● Bipolar disorder (F31) ● Mania (F30) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Acute suicidality	At least 90 days	535.49	–
Anxiety disorders
Invirto – The Therapy for Anxiety	Smartphone app*2 *3	Final listing in the DiGA Directory	● Agoraphobia with/without panic disorder (F40.00, F40.01) ● Social phobia (F40.1) ● Panic disorder (F41.0)	● Organic mental disorder (F0x) ● Substance dependence disorder (F1x) ● Psychotic disorder (F2x) ● Mania (F30) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3 and F33.3), H54, H81, H82, H91.3, I08, I09, I20, I22, I25, I45, I49, I50, I69, R42 Relative contraindications: F32.2, F33.2, F06.3, F06.4, F43, F44, G40–47, H90, I10–15, J40–47 ● Pregnancy	90 days, with an average of 8 weeks	620.00	–
Mindable: Panic disorder and agoraphobia	Smartphone app*2	Final listing in the DiGA Directory	● Agoraphobia with/without panic disorder (F40.0) ● Panic disorder (F41.0)	● Acute psychoactive substance intoxication (F1x.0) ● Psychotic disorder (F2x) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Acute suicidal ideation	90 days, at least 8 weeks	576.00	ENG, FRA
Selfapy‘s online course for generalized anxiety disorder	Web application*1 + smartphone app*2	Final listing in the DiGA Directory	● Generalized anxiety disorder (F41.1) ● Approved only for women; for male patients on a self-pay basis	● Schizophrenia (F20) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3, F33.3) ● Dementia in Alzheimer dis. (F00) ● Alcohol dependence and withdrawal syndrome (F10.2, F10.3) ● Acute suicidality	At least 90 days(at least 1 x/w for approx. 30–45 min))	479.52	–
Velibra	Web application*1	Final listing in the DiGA Directory	● Agoraphobia with panic disorder (F40.01) ● Social phobia (F40.1) ● Panic disorder (F41.0) ● Generalized anxiety disorder (F41.1)	● Schizophrenia (F20) ● Acute transient psychotic disorderr (F23) ● Schizoaffective disorder (F25), ● Bipolar disorder (F31) ● Other symptoms and signs involving emotional state (R45.8) ● Suicidal ideation	At least 90 days(1–2 x/week at least 30 min each)	230.00	ENG
HelloBetter Panic	Web application*1	Final listing in the DiGA Directory	● Agoraphobia with panic disorder (F40.01) ● Panic disorder (F41.0)	● Organic mental disorder (F0x) ● Substance dependence disorders (F1x) ● Psychotic disorder (F2x) ● Mania (F30) ● Bipolar disorder (F31) ● Severe depressive episode with psychotic symptoms (F32.3 and F33.3) ● Dissociative disorder (F44) ● Suicidality ● Psychotherapy naivety	12 weeks	599.00	–
Substance use disorders
Nichtraucher-Helden app	Smartphone app*2	Provisional listing in the DiGA Directory; trial period:03.07.2021–02.07.2022	● Nicotine dependence (F17.2)	● Acute transient psychotic disorder (F23) ● Schizoaffective disorder (F25) ● Bipolar disorder (F31) ● Depressive episode or recurrent depressive disorder (F32, F33) ● Delirium (F05) ● Other (non-phobic) anxiety disorders (F41) ● Acute depression or suicidality ● Acute delirium or psychotic disorder	At least 90 days	Initial costs: 329.00 Follow-up costs: 119.00	–
Vorvida	Web application*1	Final listing in the DiGA Directory	● Harmful use of alcohol and alcohol dependence (F10.1, F10.2)	● Schizophrenia (F20) ● Acute transient psychotic disorder (F23) ● Schizoaffective disorder (F25) ● Other symptoms and signs involving emotional state (R45.8)	At least 90 days(1–2 x/week at least 30 min each)	192.01	ENG
Sleep disorders
Somnio	Web application*1 + smartphone app*2 *4	Final listing in the DiGA Directory	● Nonorganic insomnia (F51.0)	● Bipolar disorder (F31) ● Epilepsy ● Suicidality	At least 90 days	224.99	ENG, FRA
HelloBetter Sleep	Web application*1	Provisional listing in the DiGA Directory; trial period: 18.12.2022–17.12.2023	● Nonorganic insomnia (F51.0) ● Disorders of initiating and maintaining sleep (G47.0)	● Acute transient psychotic disorder (F23) ● Bipolar disorder (F31) ● Epilepsy (G40) ● Suicidality	8 units, 45–60 min each	599.00	–
Other
HelloBetter Stress and Burnout	Web application*1	Final listing in the DiGA Directory	● Problems related to life management difficulty (Z73)	● Schizophrenia spectrum disorders ● Dissociative disorder ● Suicidality	At least 90 days	235.00	–
HelloBetter Vaginismus Plus	Web application*1 *5	Final listing in the DiGA Directory	● Nonorganic vaginismus (F52.5) ● Nonorganic dyspareunia (F52.6)	● Acute transient psychotic disorder (F23), ● Suicidality	At least 90 days	599.00	–
HelloBetter ratiopharm Chronic Pain	Web application*1	Provisional listing in the DiGA Directory; trial period:18.12.2021–17.07.2023	● Persistent somatoform pain disorder(F45.40) ● Chronic pain disorder with somatic and psychological factor(F45.41), ● Back pain (M54) ● Fibromyalgia (M79.7) ● Chronic intractable pain(R52.1) ● Other chronic pain (R52.2)	● Suicidality ● Exclusively acute pain ● Cancer pain	At least 90 days	599.00	–

*¹ Web application (internet-enabled end device)

*² Smartphone app (Android-/iOS smartphone)

*³: VR glasses (provided by the manufacturer; reimbursable)

*⁴ Integration with Sleep Tracker app possible, non-reimbursable, approx. 50–150 €)

*⁵ A vaginal dilator set is required for optimum use (reimbursable on request via a separate prescription).

ELL, Greek; ENG, English; FRA, French; ITA, Italian; POR, Portuguese; SPA, Spanish; SWE, Swedish; ZHO, Chinese

Areas of indication

Five DiGA are offered for the area of anxiety disorders (HelloBetter, Invirto, Mindable, Selfapy, Velibra), of these four each for agoraphobia with panic disorder and for isolated panic disorder and two each for agoraphobia without panic disorder, social phobia and generalized anxiety disorder. Five DiGA are available for the treatment of unipolar depression (Deprexis, Edupression.com, Elona, Selfapy, Novego), of these two also for dysthymia. One DiGA each exists for the treatment of nicotine dependence (NichtraucherHelden app) and alcohol dependence, including harmful use (Vorvida). Two DiGA (Somnio, HelloBetter Schlafen) are available for the treatment of non-organic insomnia. HelloBetter offers a DiGA for burnout and stress and also one application each for non-organic vaginismus/dyspareunia and chronic pain syndromes.

Exclusion diagnoses for the above applications are mostly organic mental disorders, psychotic disorders, bipolar disorders, substance use disorders, and suicidality (Table 1, eTable 1).

Therapeutic content

All 17 DiGA are based on methods of cognitive behavioral therapy (CBT). Users can independently access therapeutic content that is individually tailored to their needs and implemented with the aid of videos, audio, texts, exercises, and self-monitoring. For example, Deprexis uses an automated virtual dialogue in which responses and offerings are adapted based on communication behavior. Six of the application offer monitoring or contact with trained staff via messaging function. Information about therapeutic content is provided in eTable 2.

eTable 2. Application content and information for treatment providers.

Digital health application (DiGA)	Panel doctor participation required (designated activities)	Application content/ therapeutic principles	Proof of patient-relevant improvements of structure and processes	Trial access for treatment providers
Depression
Deprexis	No	● Interactive online-based self-help program to complement care as usual ● Based on methods of psychotherapy, especially cognitive behavioral therapy (CBT)	No	For 3 days (https://ehealth.servier.de/diga-manager/suche)
Selfapy’s online course for depression	No	● Personalized online course, based on CBT ● Therapeutic content is worked on independently by the patient ● Individual lessons (e.g., dealing with negative thoughts, creating a positive daily structure, relaxation techniques, sleep problems, strategies for preventing relapse) ● Teaching by means of audio and video clips, texts, and exercises ● Customization of content to the user‘s personal situation ● Continuous monitoring by a personal psychologist (via messaging function)	No	Tel.: +49 30 120834114 E-mail: aerzte@selfapy.de
Novego: Coping with depression	No	● Online support program based on CBT ● Customization on personal circumstances of the participants ● Teaching content in the form of texts, videos, audio, and interactive exercises ● The Novego psychologist team can be contacted by a message once a week ● 24-hour crisis hotline	No	E-mail: info@novego.de
Edupression.com	No	● Digital health application, based on CBT ● User-adapted delivery of therapy content between regular therapy sessions ● Teaching in the form of texts, videos, audio, and interactive exercises ● „”Interlinked” therapy offering alongside ongoing outpatient therapy	● Improvement in level of functioning ● Improvement in treatment adherence	Online dummy: (https://app.edupression.com/welcome?lang=1 Free basic version: https://app.edupression.com/register)
Elona therapy Depression	● Personalization of the treatment program for patients ● Monitoring + evaluation	● Digital health application, based also on CBT ● (Psychoeducational) therapy content in a “playful” way ● Monitoring of therapy using a proprietary emotional state diagram	No	Online form (https://www.elona.health/elona-therapy-fachkreise)
Anxiety disorder
Invirto – The Therapy for Anxiety	● Manualized therapy, free advanced training (approx. 2 h with certificate), provision of accompanying therapy hours in accordance with the current legal requirements ● Overview of patient progress and care data	● Learning with an app and virtual reality glasses, under the guidance of psychotherapists or physicians, to deal with high levels of tension, anxious thoughts and fearful situations, among others ● Based on CBT with exposure training	No	Trial access upon participation in a free 1.5-hour training session (https://invirto.de/fuer-behandler/)
Mindable: Panic Disorder and Agoraphobia	No	● Based on CBT ● Psychoeducation, getting used to physical symptoms (habituation through targeted symptom provocation), in vivo confrontation ● Live recording of the progression of anxiety during confrontation exercises ● Self-monitoring of anxiety-related symptoms, events, and behavior	● Autonomous health action (patient sovereignty) ● Coping with illness-related difficulties in daily life	Online form (https://www.mindable.health/fachkreise/)
Selfapy’s online course for generalized anxiety disorder	No	● Online course, based on CBT ● Independent processing of therapeutic content through lessons (e. g., dealing with automatic thoughts, anxiety reactions and physical symptoms; exposure exercises and mindfulness exercises) ● Teaching by means of audio and video clips, texts, and exercises ● Customization of content ● Continuous monitoring by a psychologist (can be contacted via messaging function)	No	Online form (https://www.selfapy.com/fuer-aerzte)
Velibra	No	● Based on CBT, consists of 6 modules in addition to care as usual	Reducing treatment-related expenses and burdens for patients and relatives	Online form (https://de.velibra.com/kontakt/#bob-frame)
HelloBetter Panic	No	● Web application based on CBT ● Psychoeducation on anxiety, panic disorder and agoraphobia ● Independent practice of exercises (interoceptive and in vivo exposures, cognitive restructuring, relaxation exercises, and relapse prevention) ● Diary, fictitious cases	No	Online form (https://hellobetter.de/fachkreise/)
Substance use disorders
NichtraucherHelden app	No	● Non-smoking coaching with learning modules, relaxation techniques, distraction options, and motivation, based on CBT	No	Online form (https://www.nichtraucherhelden.de/pages/doctors-info.jsf)
Vorvida	No	● Based on CBT ● Virtual dialogue	Autonomous health action (patient sovereignty)	Online form (https://de.vorvida.com/#bob-frame)
Sleep disorders
Somnio	● Step 1: Motivation ● Step 2: Monitoring (optional, by discussion or report of the application 5–10 min); ● Step 3: Analysis (reviewing the report and evaluating the treatment success and initiating further measures, if needed (5–10 min)	● Behavioral therapy self-help program, sleep diary, relaxation techniques, sleep hygiene ● Learning e.g. how to optimize sleep times and the sleep-wake rhythm, management of sleep-preventing thoughts, relaxation techniques	No	Online form (https://somn.io/fachpersonen/)
HelloBetter Sleep	No	● Behavioral therapy self-help program, sleep diary, relaxation techniques, sleep hygiene, fictitious persons as a reference for therapy ● Psychoeducation	No	Online form (https://hellobetter.de/fachkreise/)
Other offerings
HelloBetter Stress and Burnout	No	● Interactive, psychological therapy program in the format of an online course ● Seven units to be completed weekly; one refresher unit after the end of the course ● Psychoeducation by means of text, videos and audio ● Teaching of CBT strategies (problem-solving techniques, behavioral activation, emotion regulation strategies, among others) ● Online diary, companion app, repeated symptom checks, monitoring progress ● Written feedback from psychologists within 24 h after completion of a unit ● Psychologists can be reached via messaging function	No	Online form (https://hellobetter.de/fachkreise/)
HelloBetter Vaginismus Plus	No	● Online program based on CBT (strategies for dealing with distressing thoughts and feelings, relaxation and pelvic floor exercises, sensuality training [“sensate focus”], as well as step-by-step exercises involving vaginal insertion of dilators, among others) ● Psychoeducation by means of text, videos and audio ● Special focus on step-by-step exercises, involving vaginal insertion of dilators with the associated reduction of avoidance behavior and development of positive sexual experiences ● Online diary, symptom checks, companion app ● Fictitious course partners accompany users through the online program and serve as inspiration and assistance in completing the exercises ● Written feedback from psychologists ● Psychologists can be reached via messaging function	No	Online form (https://hellobetter.de/fachkreise/)
HelloBetter ratiopharm Chronic Pain	No	● Interactive psychological online program, based on acceptance and commitment therapy (ACT) and CBT (mindfulness techniques, cognitive restructuring, values work, and instructions for self-reflection, among others) ● Psychoeducation by means of text, videos and audio ● Teaching exercises to be integrated into everyday life ● In addition: online diary, companion app and repeated symptom checks to monitor personal progress	No	Online form (https://hellobetter.de/fachkreise/)

Technical requirements

Of the 17 DiGA, nine can be used as a web application, four as a native smartphone app (via Google Play Store or Apple AppStore) and four as both a web application and a native app (Table 1, eTable 1). The use of Invirto requires additional hardware (virtual reality glasses, headphones) which is eligible for reimbursement. For HelloBetter Vaginismus Plus, a dilatator set can be obtained which may be reimbursed. For Somnio, sleep trackers are available that users have to pay for themselves

The duration of use is 90 days for all applications, with 15 of the 17 DiGA specifying an average or minimum duration of use of two to three months. Velibra states that the minimum period of use should be one week; HelloBetter Sleep consists of eight one-hour course sessions.

Treatment provider involvement

In 14 of the DGA, it is part of the treatment provider’s role to establish the diagnosis, inform the patient and follow up on treatment success and adherence. Panel doctor involvement is only required with Elona therapy Depression, Invirto and Somnio: When Elona therapy Depression is used, treatment providers create an individualized therapy program and are involved in monitoring and evaluation. In the case of Invirto, patients receive manualized therapy within the framework of psychotherapeutic care. Before they start to treat patients with Invirto for the first time, treatment providers receive free advanced training, access to the therapeutic manual and a certificate. In addition, an overview of the progress made by patients is provided. In the case of Somnio, treatment providers support monitoring and evaluation, with an accompanying report.

Safety

Prior to the listing of an app in the DiGA Directory, the BfArM assesses whether the “requirements for the safety and functionality of the medical device“ have been fulfilled. However, no statistical information about adverse events is reported in the pivotal studies in addition to the evaluation of clinical outcomes. When using the Selfapy app, reference is made to a security concept in the DiGA Directory: The users are assigned to psychologists to ensure that the DiGA is used as intended. In addition, patients are monitored for therapeutic progress and the occurrence of adverse events. Signs of symptom aggravation or suicidality are detected by an automatic alarm system which refers users to appropriate points of contact. The use of Invirto is also supported by therapists. The instructions of use of the HelloBetter DiGA state that a psychologist is available via the messaging function. The “Instructions for safe use” of the remaining DiGA highlight the respective contraindications and recommend users to contact their treatment provider if they experience any deterioration or suicidality. No separate security concepts are described.

Evidence

We identified 3 meta-analyses, 39 randomized controlled trials (RCTs) and 2 single-arm interventions (Table 2, eTable 3). The effect sizes reported in controlled interventions ranged between 0.16 and 1.79; besides comparative analyses, several pre-post analyses with effect sizes in the range of 0.22 to 1.88 were also described. 23 of the 37 published trials were carried out with the direct participation of the manufacturers. Information about manufacturer involvement is not provided in the unpublished studies. Of the 39 RCTs, 29 were either not blinded or no information about blinding was provided. Standardized questionnaire-based self-assessment of symptoms was the primary outcome in most of the trials. Six RCTs had an active control group (in-patient or outpatient treatment), in all other studies, waiting list patients served as controls. The proportion of women in the study populations ranged between 24% and 86.7%, but was typically >70%. The retention rates were between 40% and 91%.

Table 2. Criteria-based summary of the clinical evidence supporting DiGA (excerpt: Indication anxiety and panic disorders, eTable 3).

Invirto – The Therapy for Anxiety	Mindable: Panic disorder, agoraphobia*1	Selfapy’s online Course Generalized anxiety disorder	Velibra*1	HelloBetter Panic
Publication
Unpublished (e25)	Unpublished (e26)	Unpublished (e27)	Berger (2017) (33)*2	Ebenfeld (2021) (34)
Type
RCT	RCT	RCT	RCT	RCT
Inclusion criteria
● Agoraphobia +/- Panic disorder ● Panic disorder ● Social phobia	● Panic disorder/ agoraphobia	● Generalized anxiety disorder	● Social anxiety disord. ● Panic disorder +/– agoraphobia ● Generalized anxiety disorder	● Mild to moderate anxiety symptoms (PAS: 9–28)
n (intervention/control)
297 (197/100)	107 (57/50)	156 (78/78)	139 (70/69)	92 (45/47)
Intervention
6 mo	8 w	12 w	9 w U; 6 mo FU	8 w G; 3 mo + 6 mo FU
Control
WL	WL	WL	WL	WL
Primary endpoint
BAI	PAS	GAD-7, WHO-5	DASS-21, BAI, BDI-II, SF-12 MH	PAS
Blinding
–	–	–	No	Yes
% female
62	–	82	70.5	55.4
% retention
82	84.1	76.9	81	91
Effect size of the primary endpoint
● Agoraphobia MD = 6.64 [2.33; 10.95]; p<0.01 ● Panic MD = 6.55 [1.38; 11.71]; p<0.05 ● Social phobia MD = 3.42 [0.01; 6.85] p<0.05	● d = 0.64 (-/-)	● GAD-7 d = −0.85 [−1.19; −0.51] ● WHO-5 d = 0.60 [0.29; 0.91]	● DASS-21 d = 0.47 [0.13; 0.81] ● BAI d = 0.41 [0.07; 0.74] ● BDI-II d = 0.61 [0.27; 0.95] ● SF-12 MH d = 0.49 [0.15; 0.83]	● d = 0.66 [0.24; 1.08] ● 3 mo: d = 0.89 [0.46; 1.31] ● 6 mo: d = 0.81 [0.38; 1.24]
Comment
		No information about statistical analysis		● Exclusion PTSD, dissociative disorder ● Effect size in DiGA D < in the publication

*¹ Studies provide evidence of additional improvements of structure and processes (not shown here). *² Study was carried out with the participation of the manufacturer. % female, proportion of women in the sample (%);% Retention, retention of study participants at the end of the intervention; BAI, Beck Anxiety Inventory; BDI-II, Beck Depression Inventory, second edition; d, Cohen’s d (effect size measure); DASS-21, Depression Anxiety Stress Scale (21 items); DiGA; digital health application (German: digitale Gesundheitsanwendung); DiGA D, DIGA Directory; FU, follow-up; G, guided (personal assistance from trained staff); GAD-7, Generalized Anxiety Disorder Scale (7 items); mo, months; MD, baseline-adjusted mean difference; p, p-value (level of significance); PAS, Parkinson Anxiety Scale; PTSD, posttraumatic stress disorder; RCT, randomized controlled trial; SF-12 MH, Short Form 12 Health Survey – Mental Health (12 items); U, unguided (independent use); w, weeks; WHO-5, World Health Organization Five Well-Being Index; WL, waiting list. If available, 95% confidence intervals are shown in square brackets after the effect size measure. Example: d = 0.23 [0.15; 0.31]. Missing 95% confidence intervals are shown as (-/-).

eTable 3. Criteria-based summary of the clinical evidence supporting digital health applications (DiGA).

DiGA	Publication (first author)	Type	Inclusion criteria	n (intervention/ control)	Intervention	Control	Primary endpoint	Blinding	% female	% retention	Effect size of the primary endpoin	Comment
Deprexis	Twomey (2020) (1)*1	MA	RCTs on Deprexis	2 901; 12 RCTs	8–12 w G or U	WL	Various	NA	–	72.1	g = 0.51 [0.40; 0.62]	Expansion by Twomey (2017) (2) by 4 RCTs
	Twomey (2017) (11)*1	MA	RCTs on Deprexis	2 402; 8 RCTs	8–12 w G or U	WL	Various	NA	–	73.5	g = 0.54 [0.39; 0.69]
	Meyer (2009) (12)*1*2*3	RCT	General population	396 (320/76)	9 w U; 6 mo FU	WL	BDI-I	–	76	55	d = 0.30 [0.05; 0.55]*5
	Berger (2011) (13)*1*2*3	RCT	BDI-II > 13	76 (25/25/26)	10 w G/U, 6 mo FU	WL	BDI-II	–	69.7	90.1	U: d = 0.66 (-/-)G: d = 1.14 (-/-)
	Moritz (2012) (14)*2*3	RCT	Depressive disorder	210 (105/105)	8 w U	WL	BDI-I	–	78.6	78.1	d = 0.36 (-/-)
	Schröder (2014) (15)*2*3	RCT	Epilepsy + depressive symptoms	78 (38/40)	9 w U	WL	BDI-I	Yes	75.6	63	d = 0.29 (-/-)	Diagnosis Epilepsy
	Meyer (2015) (16)*1*2*3	RCT	Moderate or severe depressive symptoms (PHQ-9 > 14)	163 (78/85)	12 w U; 6 mo FU	WL	PHQ-9	–	74.8	78	d = 0.57 [0.22; 0.92) 6 mo: d = 0.3 [−0.03; 0.69]
	Fischer (2015) (17)*2*3	RCT	Multiple sclerosis+ depressive symptoms	90 (45/45)	9 w U; 6 mo FU	WL	BDI-II	Yes	78	78	d = 0.53(-/-)	Diagnosis Multiple sclerosis
	Klein (2016) (18)*2*3	RCT	Mild to moderate depressive symptoms (PHQ-9 = 5–14)	1 013 (509/504)	12 w U (G if PHQ-9 >10); 6 mo FU	WL	PHQ-9	Yes	–	78.4	d = 0.39 [0.13; 0.64] 6 mo: d = 0.32 [0.06; 0.69]	Limited information available on population
	Beevers (2017) (19)*1*2*3	RCT	Moderate depressive symptoms (QIDS-SR ≥ 10)	376 (285/91)	8 w U	WL	QIDS-SR	–	74.4	79.6	d = 0.8 [0.65; 1.04]
	Zwerenz (2017) (20)*3	RCT	Depressive disorder + BDI-II > 13	229 (115/114)	12 w U + SB	SB	BDI-II	No	60.7	88	d = 0.44 (-/-)	Population: inpatients
	Berger (2018) (21)*1*2*3	RCT	Unipolar depression	98 (51/47)	12 w U + PT 6 mo FU	PT	BDI-II	Yes	66.3	70.4	d = 0.51 [0.11; 0.91]
	Bücker (2018) (22)*3	RCT	Problem gambling + subjective sadness/despair	140 (71/69)	8 w U	WL	PHQ-9	No	24	44.3	PP: d = 0.34 (-/-); ηp2 = 0.125; p = 0.03	Population: slot machine gamblers; no ITT analysis
	Fuhr (2018) (23)*1*2	RCT	Subjective depressive symptoms	27(14/13)	10 w U	WL	PHQ-9	No	68	–	I d = 0.22 (-/-)*5 K d = 0.37 (-/-)*5
	Richter (2022) (24)	RCT	IP with depression	69 (36/33)	12 w G + SB	SB	BDI-II	n	46.9	62.5	d = −0.73 [1.4; 0.06]	Population: inpatients
	Zwerenz (2019) (25)	RCT	Diagnosed depressive disorder + BDI-II > 13	229 (115/114)	12 w U + IP; 6 mo FU	IP	BDI-II	–	60.7	87.8	d = 0.58 (-/-)	Population: inpatients
	Klein (2017) (26)*1	RCT	Mild to moderate depressive symptoms (PHQ-9 = 5–14)	1 013 (509/504)	12 w U (G if PHQ-9 > 10); 12 mo FU	WL	PHQ-9	Yes	86.7	77.6	d = 0.13 [−0.18; 0.45]	Follow-up/Klein (2016) (9)
	Schuster (2020) (27)*1	RCT	Mild to moderate depressive symptoms (PHQ-9 = 5–14)	340 (166/174)	12 w U (G if PHQ-9 > 10) + PT; 6 mo FU	PT	PHQ-9	Yes	71.1	88.8	d = –0.32 [−0.11; 0.53] 6 mo: d = –0.31 [−0.09; 0.52]	Follow-up/subgroup analysis from Klein (2016) (9)
	Mullarkey (2020) (28)	RCT	At least moderate depressive symptoms (QIDS-SR ≥ 10)	295 (219/76)	8 w U	WL	QIDS-SR	–	75.2	100	CC: sadness d = 0.81 [0.54; 1.08]; indecision d = 0.72 [0.45; 0.99]; etc	Symptom level analysis Beevers (2017) (10); no ITT analysis
	Klein (2020) (29)*1	SAT	Depression or dysthymia	104	12 w U or G + OP	NA	MADRS, PHQ-9	NA	60.6	79.8	MADRS d = 1.42 [0.08; 2.76]*5 PHQ-9 d = 1.29 [0.60; 1.97]*5
	Gräfe (2020) (6)	RCT	At least mild depressive symptoms (PHQ-9 > 4)	3 805 (1 904/ 1 901)	12 w U 3 mo and 9 mo FU	WL	PHQ-9	–	79	72.6	d = 0.37 [0.29; 0.44] 3 mo: d = 0.23 [0.15; 0.31] 9 mo: d = 0.5 [0.07; 0.23]	Also evidence of reduced healthcare costs
Selfapy’s Online course Depr.	Krämer (2022) (30)*1	RCT	Mild to severe depression	401 (151/150/ 100)	12 w G/U 6 mo FU	WL	BDI-II	Yes	83	75.1	G: d = 1.63 [1.37; 1.93]; U: d = 1.47 [1.22; 1.73]; 6 mo G: d = 1.58 [1.29; 1.95]*5; 6 mo U: d = 1.88 [1.59; 2.24]*5	Effect sizes in DiGA Directory differ from publication*7
Novego: coping with depression	Beiwinkel (2017) (31)	RCT	Mild to moderate depression, dysthymia	180 (100/80)	12 w G 6 mo FU	WL	PHQ-9, BDI-II	No	68.3	48.8	PHQ-9: d = 0.55 [0.25; 0.85] BDI-II: d = 0.41 [0.11; 0.70]
	Moritz (2016) (2)	RCT	Schizophrenia + subjective depressive symptoms	58 (31/27)	12 w U	WL	CES-D	Yes	55.4	84	p = 0.001, ηp2 = 0.179 CC: p = 0.003,. ηp2 = 0.176	Diagnosis Schizophrenia
	Miegel (2019) (32)	RCT	Subjective depressive and pain symptoms	142 (71/71)	12 w U	WL	BDI-II	No	77.5	67.7	p = 0.014, ηp2 = 0.043	Diagnosis pain; study only in German language
elona therapy Depression	Unpublished (e23)	RCT	Unipolar depression	80(-/-)	12 w + PT	PT	PHQ-9, BDI-II, GAD-7	–	–	77.5	PHQ-9: d = 0.27(-/-); BDI-II: d = 0.22(-/-); GAD-7: d = 0.50(-/-)	No information on population
Edupression.com*6	Unpublished (e24)	RCT	Mild to moderate depressive symptoms	62(-/-)	12 w	WL	PHQ-9	–	–	67.2	MD = –1.74 [−3.74; 0.26]; p = 0.09; PP: DOM = –3.67 [−6.35; 1.00]; p = 0.008	No information on population
Invirto - The Therapy for Anxiety	Unpublished (e25)	RCT	Agoraphobia +/- panic disorder, panic disorder, social phobia	297 (197/100)	6 mo	WL	BAI	–	62	82	Agoraphobia: DOM = 6.64 [2.33; 10.95]; p = 0.0029; Panic: DOM = 6.55 [1.38; 11.71]; p = 0.0137; Social phobia: DOM = 3.42 [0.01; 6.85]; p = 0.0499
Mindable: Panic disorder, Agoraphobia *6	Unpublished (e26)	RCT	Panic disorder/ agoraphobia	107 (57/50)	8 w	WL	PAS	–	–	84.1	d = 0.64 (-/-)
Selfapy’s online course for gen. anxiety disorder	Unpublished (e27)	RCT	Generalized anxiety disorder	156 (78/78)	12 w	WL	GAD-7, WHO-5	–	82	76.9	GAD-7: d = –0.85 [−1.19; 0.51] WHO-5 d = 0.60 [0.29; 0.91]	No information about statistical analysis
Velibra *6	Berger (2017) (33)*1	RCT	Social anxiety disorder, panic disorder +/- agoraphobia, generalized anxiety disorder	139 (70/69)	9 w U 6 mo FU	WL	DASS-21, BAI, BDI-II, SF-12 MH	No	70.5	81	DASS-21: d = 0.47 [0.13; 0.81] BAI d = 0.41 [0.07; 0.74] BDI-II d = 0.61 [0.27; 0.95] SF-12 MH d = 0.49 [0.15; 0.83]
HelloBetter Panic	Ebenfeld (2021) (34)	RCT	Mild to moderate symptoms of anxiety (PAS 9–28)	92 (45/47)	8 w G 3 mo + 6 mo FU	WL	PAS	Yes	55.4	91	d = 0.66 [0.24; 1.08] 3 mo: d = 0.89 [0.46; 1.31] 6 mo: d = 0.81 [0.38; 1.24]	Exclusion of PTSD, dissociative disorders; effect size in DiGA D < in publication
Nichtraucher Helden app	Unpublished (e28)	SAT	Nicotine dependence	50	4 mo	NA	Nicotine abstinence	NA	–	–	Smoking abstinence (7-day prevalence) in 30% (n = 15)*5	No information about statistical analysis and population
Vorvida*6	Zill (2019) (35)*1	RCT	Harmful use of alcohol (> 12/24 g/d in m/f or AUDIT-C > 3)	608 (306/302)	6 mo U	WL	QFI (intake in last 30 days), TFB (intake in last 7 days	No	52.5	68	3 mo QFI: d = 0.38 (-/-) 3 mo TFB: d = 0.42 (-/-) 6 mo QFI: d = 0.33 (-/-) 6 mo TFB: d = 0.54 (-/-)
Somnio	Lorenz (2019) (36)*1	RCT	ISI > 8	56 (29/27)	6 w U; 12 mo FU	WL	ISI	Yes	67.2	86.2	d = 1.79(-/-) 12 mo: d = 1.98 [1.31; 2.66]	Exclusion of shift work, OSAS, RLS, parasomnia
HelloBetter Sleep	Unpublished (e29)	IPD-MA	NA	433 (216/217)	8 w	WL	ISI	NA	–	–	d = 1.06 [0.85; 1.26]	Inhomogeneous study population
	Thiart (2015) (37)*1*4	RCT	ISI ≥ 15	128 (64/64)	6 w G; 6 mo FU	WL	ISI	No	74.2	92.2	d = 1.45 [1.06; 1.84] 6 mo: d = 1.43 [1.04; 1.82]	Population: teachers
	Eber (2015) (38)*1*4	RCT	ISI ≥ 15	128 (64/64)	6 w U; 6mo FU	WL	ISI	No	74.2	59.4	d = 1.37 [0.99; 1.77] 6 mo: no control available	Population:teachers
	Behrendt (2020) (39)*1*4	RCT	Interest in iCBT-I	177 (88/89)	6 w U; 6 mo FU	WL	ISI	No	65.5	40	d = 0.97 [0.66; 1.28] 6 mo: d = 0.86 [0.55; 1.17]
HelloBetter Stress/Burnout	Heber (2016) (40)*1	RCT	PSS-10 ≥ 22, healthy employees	264 (132/132)	7 w G; 6 mo FU	WL	PSS-10	No	73.1	69.7	d = 0.83 [0.58; 1.08] 6 mo: d = 1.02 [0.76; 1.27]
HelloBetter Vaginismus Plus	Zarski (2017) (3)*1	RCT	Vaginismus + no vaginal intercourse since >6 mo	77(40/37)	10 w G; 6 mo FU	WL	PEQ item 7	No	100	90	CC: d = 0.16 [−0.32; 0.64] 6 mo: d = 0.37 [−0.17; 0.92]	No ITT analysis
	Zarski (2021) (4)*1	RCT	Nonorganic vaginismus/dyspareunia (PEQ item 7 ≥ 1)	200 (100/100)	12 w G; 6 mo FU	WL	PEQ item 7	No	100	79	Ability of coital vaginal penetration: 12 w: χ2(1) = 9.44; p < 0.01; 6 mo: χ2(1) = 2.19; p = 0.19	Effect size in DiGA D >in publication
HelloBetter ratiopharm Chronic Pain	Lin (2017) (5)*1	RCT	Pain >3 mo + one of the diagnoses: M79.7, M54, F45.40, F45.41, R52.1, R52.2	302 (100/101/ 101)	9 w G/U; 6 mo FU	WL	WHYMPI, (pain interference subscale)	–	84.1	76	G: d = 0.58 [0.30; 0.86] 6 mo guided d = 0.58 [0.30; 0.86] U: no significant effect	Effect sizes of the unguided intervention not shown
	Unpublished (e30)	RCT	Pain >3 mo + one of the diagnoses: M79.7, M54, F45.40, F45.41, R52.1, R52.2	201 (100/101)	9 w G	WL	WHYMPI, (pain interference subscale)	–	–	–	β = −0.756; p<0.001; θ = −0.66	No information about statistical analysis and population

*¹ Trial carried out with participation of the manufacturer. *² Study was included in Twomey (2017). *³ Study was included in Twomey(2020). *⁴ Study was included in an IPD-MA. *⁵ Gray font color: highlights pre-post analysis-based statistics. *⁶ Studies also show improvements in structure and processes (not shown here). *⁷ According to the manufacturer, various ITT processes were used. –, not stated.% female, proportion of women in the sample (%);% Retention, retention of subjects at the end of intervention; OP, outpatient psychiatric/psychotherapeutic/medical treatment; AUDIT-C, Alcohol Use Disorders Identification Test; short version; BAI, Beck Anxiety Inventory; BDI-I, Beck‘s Depression Inventory, version 1; BDI-I, Beck’s Depression Inventory, version 2; CC, complete case analysis; CES-D, Center for Epidemiologic Studies Depression Scale; d, Cohen’s d (measure of effect size); DASS-21, Depression Anxiety Stress Scale (21 items); FU, Follow-up; G, guided (personal assistance from trained staff); g, Hedges’ g (measure of effect size); GAD-7, Generalized Anxiety Disorder Scale (7 items); ICBT-I, Internet-Based Cognitive Behavioral Therapy for Insomnia; IPD-MA, individual patient data meta-analysis; ISI, Insomnia Severity Index; ITT, intention-to-treat analysis; C, control; MA, meta-analysis; mo, months; DOM, baseline-adjusted difference of means; NA, not applicable; OSAS, obstructive sleep apnea syndrome; p, p-value (significance level); PAS, Parkinson Anxiety Scale; PEQ, Primary Endpoint Questionnaire; PHQ-9, 9-question Patient Health Questionnaire; PP, per-protocol analysis; PSS-10; Perceived Stress Scale (10 items); PT, psychotherapy; PTSD, posttraumatic stress disorder; QFI, Quantity Frequency Index; QIDS-SR, Quick Inventory of Depressive Symptomatology – Self-Report; RCT, randomized controlled trial; RLS, Restless Legs Syndrome; SAT, single-arm trial; IP, in-patient psychiatric/psychosomatic treatment; SF-12 MH, Short Form 12 Health Survey – Mental Health (12 items); SvMADRS, Montgomery-Åsberg Depression Rating Scale, short version; TFB, Timeline-followback method; U, unguided (independent use); w, weeks; WHO-5, World Health Organisation Five Well-Being Index; WHYMPI, West Haven Yale Multidimensional Pain Inventory; WL, waiting list; ηp2, partial eta-squared; χ2, chi-square test. If available, 95% confidence intervals were shown in square brackets after the effect size measure. Example: d = 0.23 [0.15; 0.31]. Missing 95% confidence intervals are shown as (-/-).

The broadest evidence is available for Deprexis (2 meta-analyses, 18 RCTs and 1 single-arm interventional study); the most recent meta-analyses reports an effect size for the treatment of g = 0.51 compared to waiting-list controls (1). The program Novego for the treatment of depression was evaluated in three RCTs in various patient populations (depression, pain, schizophrenia [2]): The effect sizes found were moderate/small and moderate and moderate to large, respectively. HelloBetter Sleep cites an unpublished individual patient data (IPD) meta-analysis of three RCTs with an effect size of 1.06 (e10). HelloBetter Vaginismus Plus and HelloBetter ratiopharm Chronic Pain each cite two RCTs with low to medium effect sizes (3–5).

For each of further 11 applications, one RCT with low to medium effect size is available. This includes unpublished results for Invirto, Mindable and Selfapy (generalized anxiety disorder). The evidence in support of the Nichtraucherhelden app is based on one unpublished, uncontrolled observational study with before-and-after comparison.

All 17 DiGA appear to have positive healthcare effects in the form of a medical benefit (for example, decrease in depressive symptoms). In addition, four DiGA demonstrate improvements in quality of life (Invirto, Mindable, Selfapy for generalized anxiety disorder, NichtraucherHelden app). Four DiGA report patient-related improvements of structure and processes (eTable 2).

The only health economic analysis that could be identified was for Deprexis in the form of a study that showed a 32% reduction in health insurance costs in the intervention group compared to 13% in the passive control group in a routine setting (6).

Discussion

Our review provides a summary of the proof-of-efficacy data and application modalities available for the 17 DGA which have been approved for use in the field of mental health.

One point that needs to be critically discussed is the high risk of bias in the available studies favoring inflated effect sizes. The main restrictions on study quality were as follows (Table 2, eTable 3):

• Lack of blinding

• with passive/pragmatic control group in the waiting list setting

• Single-center design

• Use of indirect patient-reported outcome measures (PROMs)—and consequently without assessment of relevant outcomes by trained specialist staff

• With participation of the manufacturers.

Of particular note is that no conclusions on causality can be drawn when effect sizes are calculated from data of pre-post comparisons (gray font color in eTable 3). The development of psychological symptoms over time is influenced by a variety of factors, including high spontaneous fluctuation or being enrolled in a therapy program or even a clinical trial. As a result, the effect size may be considerably overestimated, since a large number of distorting factors are not controlled for (e11).

Another point of criticism is that occasionally unpublished datasets that cannot be assessed by the scientific community and users have served as the basis for the provisional approval of these medical devices. In addition, there are significant dropout rates of up to 60% in some studies, which may lead to misjudgments of the effectiveness and/or acceptance of the product. Another potential bias factor is the significant underrepresentation of male subjects in the study cohorts. In none of the examples included in our review, adverse events were systematically recorded and reported. Finally, the low approval hurdles mean that there are no incentives to already conduct high-quality studies to support applications for provisional approval. Thus, falsely reported high effect sizes can only be corrected years after approval of the DiGA.

In addition to clinical evidence, therapeutic content and technical requirements are key criteria for potential prescription. For example, the availability of a smartphone or computer and an Internet connection is essential. Other points to consider prior to prescribing any DiGA include the digital (health) literacy of potential users as well as problematic digital media use behavior (7). Only six applications are available in German and other languages, none of these in Turkish, Arabic and Farsi/Dari. Since adequate knowledge of German was an inclusion criterion in most of the pivotal studies, it can be assumed that there is insufficient validation of the products for cultural minorities in Germany.

All existing DiGA are based on strategies of cognitive behavioral therapy (CBT). The effectiveness of digital CBT is now well established for several indications: For example, meta-analyses show a mean effect size of 0.65 for the treatment of depression (8) and an effect size of 0.39 for the treatment of insomnia (9). Digital CBT might be as effective as conventional psychotherapy. Although several meta-analyses support this conclusion, particularly in the treatment of depression (10, e12, e13), there is a lack of additional high-quality studies directly comparing digital and conventional CBT (10, e12, e13). Such comparisons are still entirely lacking for the DiGA included in this review; studies comparing the applications with each other or with specific treatments are still almost non-existent.

To date, there are no DiGA available that address severe and chronic mental illnesses. Despite higher barriers to approval and development, first data indicating the efficacy of applications in the treatment of patients with severe mental illness have become available from international studies (e14).

The principle of cost-effectiveness must also be observed when prescribing DiGA (section 12 of the German Social Code [SGB, Sozialgesetzbuch] V); this means that treatment has to be sufficient, expedient and cost-effective. The DiGA price range is between EUR 192 and EUR 620; only for Deprexis, we were able to identify a manufacturer-independent health economic study, showing a reduction in costs of the insured in the intervention group (6).

The DiGA procedure is also intended to promote the establishment of standards: In many cases, insured persons can digitally export treatment-relevant information, for example information about treatment progress, and make it available to their treatment provider. Interfaces to electronic patient records, medical devices and sensors are also being established (e3). Even though all DiGA require patient information, motivation, support, and monitoring of treatment success, direct involvement of the treatment provider is only required with the Elona, Invirto and Somnio apps.

The willingness of treatment providers to prescribe DiGA is significantly affected by both limited familiarity with and skepticism about digitalization solutions (e15). In the expert opinion of the German Advisory Council on the Assessment of Developments in the Health Care System (SVR, Sachverständigenrat zur Begutachtung der Entwicklung im Gesundheitswesen und in der Pflege), it is therefore recommended that digital health literacy be anchored in the training curricula of the healthcare professions, which has been partially implemented to date (e7). The majority of manufacturers offer free trial access or training for treatment providers. The Central Research Institute of Ambulatory Health Care (Zi, Zentralinstitut für die kassenärztliche Versorgung in Deutschland) has in addition launched the online information portal “KV-App-Radar“ (www.kvappradar.de/).

Given various DiGA-related issues with data protection in in recent years and well-known examples of non-transparent handling of personal data, skepticism towards the disclosure of sensitive personal data to a private service provider would appear to be justified: Once consent has been given, collected data can be used for commercial and other purposes. During periods of illness, it can be even more difficult to adequately weigh up the risks associated with sensitive personal data provided in sometimes difficult-to-understand consent forms.

DiGA must ensure that high data protection standards are met: In addition to data encryption and user information, an information security management system (ISMS) has to be provided. Compliance with the requirements has until recently been proven on the basis of declarations in the application for listing in the DiGA directory. With the First Amendment of the Digital Health Applications Ordinance (1^st DiGAVÄndV, erste Verordnung zur Änderung der Digitalen-Gesundheitsanwendungen-Verordnung), stricter requirements have been in force since 2022: Data security must now be confirmed by a certificate from the German Federal Office for Information Security (BSI, Bundesamt für Sicherheit in der Informationstechnik) in accordance with the General Data Protection Regulation (GDPR, Datenschutzgrundverordnung), and proof of penetration tests must also be provided upon submission of the application. Accredited data protection certification in accordance with Article 42 GDPR has been a requirement since 2023 (e19). However, scientific evaluation of data security flaws in DiGA is virtually non-existent. Tools for the comparative assessment of matters such as data security for users are not established, but are being developed by various institutions and could be an important component of informed use decisions (e20).

One of the strengths of this review is the criteria-based presentation and discussion of a comprehensive overview of DiGA. A standardized approach and a meta-analysis of the data would have been of additional benefit. Due to the heterogeneity of the studies and unpublished data, however, this was only possible to a limited extent, and therefore an overview is provided in the form of a scoping review. Considering the increase in involvement and autonomy of patients in their treatment possibly offered by DiGA, participation of patient representatives in the development of DiGA would be an important aspect; however, we could not assess its relevance based on the studies available to us.

DiGA offer the opportunity to include the perspectives of patients and patient organizations in the process of their ongoing development. DiGA strengthen the autonomy of patients in their treatment and are an important additional therapy option, especially in view of the significant gaps in the provision of therapeutic services (e21, e22). The field of digital health applications continues to grow despite the various obstacles discussed in this review—it is increasingly adopted as an additional evidence-based therapeutic modality in psychiatry.

Questions on the article in issue 47/2023:

Digital Health Applications in the Area of Mental Health

The submission deadline is 23 November 2024. Only one answer is possible per question.

Please select the answer that is most appropriate.

Question 1

What are the requirements for digital health applications mentioned in the text?

1. DIN certification and classification as a medical device by the BfArM

2. TÜV certification for medical devices and CE certification

3. Classification as a medical device by the Federal Joint Committee (G-BA) and ISO 9001 certification

4. Mandatory CE conformity marking as a medical device.

5. TÜV certification and assessment by the G-BA

Question 2

How often were digital health applications (DGA) in the area of mental health used in Germany between 2020 and 2022?

1. 22 000

2. 51 000

3. 100 000

4. 164 000

5. 250 000

Question 3

According to a survey, what percentage of physicians in Germany would like to prescribe digital health applications?

1.  5%

2. 11%

3. 24%

4. 33%

5. 55%

Question 4

For which of the following mental illnesses is there currently no digital health application available?

1. Obsessive-compulsive disorder

2. Anxiety and panic disorder

3. Alcohol dependence

4. Nicotine dependence

5. Chronic pain syndromes

Question 5

Which of the following diagnoses is usually considered an exclusion diagnosis for a digital health application?

1. Nonorganic insomnia

2. Burnout and stress

3. Bipolar and psychotic disorders

4. Nonorganic dyspareunia

5. Dysthymia

Question 6

Which of the following statements about the methods used in currently available digital health applications (DiGA) is true?

1. 9 of the 17 DiGA are based on methods of cognitive behavioral therapy.

2. All 17 DiGA are based on methods of psychoanalytic therapy.

3. 6 of the 17 DiGA are based on methods of systemic therapy.

4. 10 of the 17 DiGA are based on psychoanalytic therapy.

5. All 17 DiGA are based on methods of cognitive behavioral therapy.

Question 7

For which of the following DiGA is the broadest evidence (two meta-analyses and 18 randomized controlled trials) available?

1. Mindable

2. Deprexis

3. Invirto

4. Selfapy

5. Velibra

Question 8

Which of the following points is not one of the critical points addressed in the article with regard to the available evidence from studies on the effectiveness of DIGA?

1. Underrepresentation of female participants in the trials

2. In some cases significant dropout rates of up to 60%

3. Many studies without blinding

4. Many studies with participation of manufacturer

5. Outcomes are mostly based on patient-reported data

Question 9

What is the price range for digital health applications?

1. approx. 50 – 80 euros

2. approx. 90 – 180 euros

3. approx. 200 – 600 euros

4. approx. 750 – 1000 euros

5. approx. 1200 – 1500 euros

Question 10

For the indication depression, how long should the DiGA described in the article at least be used?

1. 14 days

2. 30 days

3. 45 days

4. 90 days

5. 170 days