The Treatment of Vertigo With a Digital Health App

Findings of the Prospective Randomized Controlled GEVE-I Trial

Background: Vestibular vertigo is common and frequently leads to inability to work. We studied the clinical efficacy and safety of a digital health app (VH90D) for the treatment of vestibular vertigo.

Methods: We conducted a confirmatory, prospective, single-blinded, randomized, group-controlled, single-center, two-arm superiority trial (n = 212). The 106 patients in the intervention group were treated with the app, and the other 106 received treatment as usual (TAU) with standard physiotherapy. The primary endpoint was the vertigo symptom score (VSS-sf-VER) at 12 weeks. The trial was registered in the German Clinical Trials Register (DRKS00024188).

Results: The mean intergroup difference in the change of VSS-sf-VER score from baseline to week 12 was −7.9 points (95% confidence interval [−9.5; −6.2]; Cohen’s d 1.55). The vertigo intensity score fell by an average of 12.7 points (clinically important difference [CID] −5 points; [−14.1; −11.2]) in the intervention group and 2.7 points (CID −5 points; [−3.7; −1.7]) in the TAU group (physiotherapy).

Conclusion: The digital health app VH90D markedly reduced vertigo, with greater efficacy than the median of six 20-minute physiotherapy sessions.

Cite this as: Wirth M, Pieper J, Heller U, Bulitta M, Schmitz D, Wollenberg B, Ruck A, Löwenheim H, Lehmacher W, Wolpert S: The treatment of vertigo with a digital health app: Findings of the prospective randomized controlled GEVE-I trial. Dtsch Arztebl Int 2026; 123: 157–62. DOI: 10.3238/arztebl.m2025.0232

Dizziness and vertigo are common reasons for medical consultations (1, 2). Among adults, the lifetime prevalence of dizziness of any type is estimated at 23.2% (3) and the lifetime prevalence of vestibular vertigo severe enough to interfere with daily activities is 7.8% (1). The clinical standard of care for chronic vestibular vertigo includes balance-improving exercise therapy (4).

The digital health app used in this study (VH90D) involves a sensorimotor, multimodal vertigo treatment program for use at home. The major therapeutic principle of the app comprises diagnosis-adjusted sensorimotor vestibular exercise programs: the Adaptive Balance Eye and Vision (ABEV) exercises (5, 6). The performance of these exercises stimulates the sensory organs and therefore the brain. This sensorimotor brain stimulation, the principal mechanism of action of the app’s ABEV program, is intended to induce central vestibular compensation (CVC), i.e., a neuronal learning process (4, 7, 8). Thus, as long as central vestibular compensation is achieved, the app may be effective irrespective of the underlying peripheral disease. The app includes:

• Specific cognitive behavioral therapy interventions for coping with vertigo-associated stress, vertigo-induced sleep disorders, and vertigo-related anxiety
• Health education providing adequate information about the underlying disease models, with the aim of improving therapy adherence
• Relaxation techniques such as progressive muscle relaxation and autogenic training to reduce unwanted somatosensory responses
• The Otago Exercise Program (9) as a training option in muscle atrophy of the lower extremities

The training program of the digital app is adapted to the underlying disease, e.g., in patients with benign paroxysmal positional vertigo.

To investigate the clinical efficacy and safety of the digital app, the present German Vertigo trial (GEVE I) was designed as a confirmative, prospective, single-blinded, randomized, group-controlled, single-center, two-arm trial with active safety surveillance.

Materials and methods

The Vertidisan app (VH90D-1.0.2) was provided by the manufacturer (Digitineers, Tübingen, Germany). The clinical investigation plan was approved by the ethics committee of the Medical Association of the Federal State of Baden-Württemberg, under the appraisal number F-2021–157. The study was registered at the German Registry for Clinical Studies (DRKS00024188, https://www.drks.de/search/de/trial/DRKS00024188) where the clinical investigation plan, including the statistical analysis plan, was deposited.

Briefly, this study was a randomized controlled trial (RCT) with treatment as usual (TAU; physiotherapy) in the control group and 12 weeks’ treatment with the app (VH90D) in the experimental group. Patients in the TAU group received physical therapy/physiotherapy (at least six individual sessions, 20 minutes each, corresponding to the usual number of exercise units per prescription for patients with statutory health insurance in Germany) and were offered the app after 12 weeks. No influence was exerted on timing, possible waiting time or additional physiotherapy sessions. Patients could go to the physiotherapist of their choice, so the physiotherapists’ previous experience of treating vestibular disorders was not controlled for. Figure 1 shows the overall study design and procedures. The principal steps included enrolment, allocation, and follow-up. The patients were recruited via advertisements in local newspaper and by informing local physicians’ offices. The single-center study was performed at the clinical study center Tübingen Kelternturm (Tübingen, Germany) and was conducted from 10 October 2022 to 22 March 2023. A CONSORT 2025 (10) checklist on the trial is provided in eTable 1.

Figure 1

Flowchart of patient allocation

Randomization and blinding are described in the eMethods.

A total of 326 individuals were screened based on inclusion and exclusion criteria. The main inclusion criteria were benign paroxysmal positional vertigo (BPPV), labyrinthine trauma, labyrinthitis, persistent postural-perceptual dizziness (PPPD), tumors of the skull base and cerebellopontine angle, presbyvestibulopathy, Menière’s disease, vestibular migraine, superior semicircular canal dehiscence syndrome, acute vestibular syndrome, vestibular neuritis, vestibular paroxysmia, unilateral or bilateral vestibulopathy, sudden hearing loss with vestibular involvement, toxic vestibulopathy, and idiopathic vertigo.

The clinical variables were vertigo, stress, autonomic responses/anxiety, and quality of life (QoL). The primary efficacy variable was the score on the German version of the Vertigo Symptom subscale, VSS-sf-VER (11, 12) after 12 weeks (T3). The primary objective was the intergroup comparison of the changes from T0 to T3. The secondary endpoints were evaluated in a predetermined confirmatory test sequence (eMethods). This article focuses on the primary outcome, i.e., analysis of the VSS-sf-VER results.

The primary evaluation was conducted as an intention-to-treat (ITT) analysis. Dropouts were firstly replaced by reference-based multiple imputations (jump to reference [JTR]). Other substitution methods were applied for sensitivity analyses (e.g., last observation carried forward, baseline carried forward, and total mean replacement).

Analysis of covariance (ANCOVA) was performed, with baseline value as covariate and treatment group as factor. Wilcoxon–Mann–Whitney tests were performed for sensitivity analysis. All primary endpoints were analyzed using two-sided tests, with α = 5%. All other statistical analyses were performed exploratorily with descriptive interpretation of p-values and 95% confidence intervals; for subgroups the separate effects are described by means of 95% confidence intervals.

Demographic results

Our analysis included all 212 ITT patients (100 men and 112 women). The TAU group was older (64.7 years) than the APP group (61.6 years) (eTable 2). The largest diagnosis groups were patients with unilateral or bilateral vestibulopathy (N = 77, 36%), BPPV (N = 34, 16%), PPPD (N = 29, 14%), vestibular migraine (N = 27, 13%), vestibular neuropathy (N = 23, 11%), presbyvestibulopathy (N = 22, 10%) and Menière’s disease (N = 12, 6%). The baseline demographics did not differ meaningfully between the per-protocol (PP) population and the ITT population.

eTable 2

Demographic distribution

eTable 3

Treatment adherence in the two groups during the study period

eTable 4

Summary of mean adherence and compliance of the two PP groups

eTable 5

Analysis of vertigo (VSS-sf-VER score, ITT) at each time point and in comparison with baseline

eTable 6

Diagnoses of the study participants

Efficacy results

Vertigo intensity was measured using the German version of the validated Vertigo Symptom Subscale, VSS-sf-VER. A change of five points (15% of the total range of 32, cf. IQWIG 2022) (13) was defined as a clinically important difference (CID) in the VSS score. Thus, a score difference of five or more points indicates a clinically relevant change. Using JTR imputations, the statistical efficacy comparison of the two ITT groups resulted in a least-square mean difference (adjusted by ANCOVA) between the groups of –7.9 ([−9.5; –6.2]; Cohen’s d: 1.55) points.

Both the APP group and the TAU group (physiotherapy) exhibited a statistically significant improvement in the VSS-sf-VER score from baseline to week 12 in the ITT population. From baseline to week 12 (Figure 2a, eTable 1) the VSS-sf-VER score decreased by a mean of –12.7 points (-66.2%, [−14.1; –11.2]; p < 0.001) in the APP group and by a mean of –2.7 points in the TAU (physiotherapy) group (−16.9%; [−3.7 to –1.7]; p = 0.001). The APP group, but not the physiotherapy group (TAU), reached CID status (Table).

Figure 2

Change in vertigo intensity during the study. VSS-sf-VER score (score range: 0–32); the bars indicate the confidence interval (95% CI)

Table

Analysis of vertigo on the basis of the score on the Vertigo Symptom Scale–Short Form vertigo (VSS-sf-VER)*1

In the responder analysis of ITT (JTR) patients (Figure 3), 93 (87.7%) of the app users were responders after 12 weeks. Responders were defined as patients who achieved the above-mentioned CID, i.e., improvement of the VSS score by 5 points. In the TAU group (physiotherapy), 27 participants (25.5%) were responders. The group comparison demonstrated differences between the APP group and the TAU group (physiotherapy) (response rate difference 62.3% [51.9; 72.6]; p < 0.001) (Figure 3).

Figure 3

Percentages of responders and non-responders. Responders were predefined as exhibiting improvement of at least 15% on the VSS-sf-VER scale (or at least 5 points).

Furthermore, a difference was found between the absolute VSS-sf-VER scores of the two groups after 12 weeks. Within the APP intervention group (ITT, JTR), the scores exhibited a mean end value of 6.47 (SD 6.18) and a median of 4.48 points. The mean and median values in the TAU group were 13.30 (SD 5.67) and 13.50, respectively, after the 12-week period.

Moreover, a higher number of completed digital app sessions seemed to contribute to an increased efficacy of the app, as depicted in Figure 2b. In a post-hoc analysis after 12 weeks, patients who had completed 1–30 sessions with the app showed an average improvement of 9.93 points on the VSS-sf-VER scale. For patients who used the app 31–60 times and 61–90 times, respectively, this improvement increased to 11.40 points and 13.63 points.

In a subgroup analysis of underlying diseases, the app showed clinical efficacy in different peripheral vertigo diseases. Figure 4 shows the forest plot of the group comparison of T0–T3 changes for different subgroup analyses. The diagnoses investigated included BPPV, bilateral and unilateral vestibulopathy, vestibular neuropathy, and vestibular migraine, as well as PPPD as comorbidity. Furthermore, sex, a pretreatment tendency towards falling, duration of illness, paroxysmal or persistent forms of dizziness type as well as permitted occasional adjunctive treatment by means of physiotherapy (in the experimental group) or permitted occasional use of medications had no major influence on the efficacy of the app (eFigure).

Figure 4

Forest plot of APP treatment effect for VSS-sf-VER subgroups (ITT): Adjusted LS mean and 95% confidence intervals of APP treatment effect as determined by group comparison of changes from T0 to T3. The black square indicates the mean value of the observed treatment effect. The horizontal line represents the 95% confidence interval. The vertical line (no-effect line, right) marks the point at which there is no treatment effect on the measured end point for the respective subgroup. With the exception of the subgroups female/male, all presented subgroup analyses were planned a priori. A forest plot depicting additional subgroups can be found in the eFigure.

eFigure

Exhaustive forest plot (cf. Figure 4) of APP treatment effect for VSS-sf-VER subgroups -(ITT): Adjusted LS mean and 95% confidence intervals of app treatment effect determined by group comparison of T0-T3 changes. Dose-dependent response: days. Physiotherapy/pharmacotherapy: i.e., additional physiotherapy/pharmacotherapy. The black square indicates the mean value of the observed treatment effect. The horizontal line depicts the 95% CI. The vertical line (no-effect line) marks the point at which there is no treatment effect on the measured end point for the respective subgroup. With the exception of the subgroups female/male, all presented subgroup analyses were planned a priori. Selected subgroups of the forest plot are shown in Figure 4.

To investigate safety, we performed active surveillance of adverse events (AE). Among 29 patients of the APP group, we observed the following reversible adverse events: dizziness, nausea, balance problems, neck pain, headache, blurred vision, tunnel vision, tinnitus, musculoskeletal pain, back tension, muscle tension, fatigue, knee pain, and loss of appetite. In the physiotherapy group, 20 participants exhibited similar reversible AE as in the APP group. Participants in the physiotherapy group (TAU) experienced two additional permanent AE, namely sudden hearing loss and tinnitus, which were categorized as physiotherapy-induced serious AE.

Discussion

The clinical standard of care for vertigo may include medication, physiotherapy, psychotherapy, and, rarely, surgery. In chronic vertigo, however, evidence-based drug treatment is frequently not available. In the acute phase of peripheral vertigo, physicians regularly prescribe cortisone or antivertigo drugs that interact with histamine, muscarinic, dopamine, serotonin, and/or GABA receptors (14). Many of these drugs should only be given for a limited period of time (4). In the chronic treatment scenario, drug therapy is often based purely on a physician’s personal experience, as there is often no published scientific evidence (4, 15, 16, 17, 18). In addition, medication may be limited by AE (14).

Here we present the results of a single-center, randomized, clinical trial using a digital health app to treat patients suffering from peripheral vertigo. We found that 87.7% of the app users were responders, compared with 25.5% in the physiotherapy group.

The amelioration of vertigo symptoms was similar irrespective of the underlying peripheral vertigo diagnosis. This may be explained by the expected central vertigo compensation mechanism of action in the brain (4, 19), which may be similar in peripheral vestibular diseases. In terms of age, there is a tendency towards lower effectiveness in older patients, which can be explained by the expected delayed central compensation at advanced age. Furthermore, 34 patients (16%) with BPPV were included in the study. Instruction in the positional maneuvers was therefore possible via the app we used, but not verification of the correct performance of the exercises. In the event of doubt about proper execution of the exercises, additional physiotherapy or verification of compliance by the treating physician may be beneficial.

One potential reason for the poorer efficacy of physiotherapy in comparison with the app is the brevity of the treatment, from six to 20 individual sessions of 20 minutes each, compared with 20 minutes’ treatment daily for 90 days with the app. In this study, performance of exercises at home by patients in the physiotherapy group was not documented and the suitability of the physiotherapy exercises was not monitored. Patients could visit a physiotherapist of their own choice irrespective of the latter’s expertise in treating vestibular disorders, reflecting routine clinical care in Germany. Moreover, only 51.4% of patients had started with physiotherapy at T1 (after 2 weeks), again mirroring the routine care situation in Germany. Specialized (and daily) physiotherapy or daily non-digital treatment from day 1 as control may have been more effective than the physiotherapy in this study. Indeed, a Cochrane review reported a statistically significant effect of vestibular exercises on rehabilitation (odds ratio (OR) 2.67, [1.85; 3.86; four studies, 565 participants) (20). The study therefore does not per se prove that the app is more efficacious than a non-digital standardized daily multimodal vestibular rehabilitation therapy for 90 days, which was not performed in the TAU (physiotherapy) group—and on practical as well as economic grounds is not routinely carried out in Germany. Another potential reason for the difference in results between the APP group and the TAU group (physiotherapy) is the differing distribution of vestibular disorders in the two groups. There were higher proportions of patients with bilateral vestibulopathy (26% vs 17%) and idiopathic vertigo (12% vs 3%) in the TAU (physiotherapy) than in the APP group, and both are more difficult to treat with physiotherapy alone than, for example, BPPV. In addition, the patients that used the app were slightly younger (mean 62 vs. 65 years) with a potentially faster rate of recovery.

In addition, this study compared the multimodal approach of the app with physiotherapy alone, rather than with multimodal treatment. Another bias of the present study is the single blinding, a double-blind design being methodologically infeasible. Potential further sources of bias are the higher baseline values of the primary variable in the APP group and the above-mentioned differing distribution of the vertigo diagnoses in the two groups. However, these two factors cannot fully explain the difference seen between the groups. Moreover, selection bias could confound the results in the APP group, since 32% of patients completed 60 sessions or fewer. Furthermore, the study was limited to 3 months’ follow-up, with no data on longer-term effects.

Data sharing
Individual participant data underlying the results reported will be shared after anonymization upon reasonable request (research proposal) to the first author within 5 years of publication.

Note
The work for this study was part of the PhD thesis of JP. A summary of the results has already been published in German Registry for Clinical Studies: https://www.drks.de/search/de/trial/DRKS00024188.

Conflict of interest statement
This study was sponsored by Digitineers GmbH & Co KG, Tübingen, Germany. The sponsor was involved in the planning and monitoring of the study. The management and shareholders of the sponsoring company overlap. Michael Bulitta, Daniel Schmitz, and Jannik Pieper were paid for performing statistical analyses and Walter Lehmacher for statistical consulting in the planning and analysis of this study. The study was conducted in facilities provided by the sponsor, including diagnostic services. Manuscript writing was assisted by the sponsor.

The authors declare that no further conflict of interest exists.

Manuscript received on 4 June 2024, revised version accepted on 26 November 2025

Corresponding author
Prof. Dr. med. Markus Wirth

References

1. Neuhauser HK, von Brevern M, Radtke A, et al.: Epidemiology of vestibular vertigo: A neurotologic survey of the general population. Neurology 2005; 65: 898–904

2. Murdin L, Schilder AG: Epidemiology of balance symptoms and disorders in the community: A systematic review. Otol Neurotol 2015; 36: 387–92

3. Kroenke K, Price RK: Symptoms in the community. Prevalence, classification, and psychiatric comorbidity. Arch Intern Med 1993; 153: 2474–80

4. AWMF guideline 053–018: Akuter Schwindel in der Hausarztpraxis S3. 2015/2019. https://register.awmf.org/de/leitlinien/detail/053-018 (last accessed on 4 December 2025).

5. AWMF guideline 017 – 078: Vestibuläre Funktionsstörungen S2k2021. https://register.awmf.org/assets/guidelines/017-078l_S2k_Vestibulaere-Funktionsstoerungen_2021-05.pdf (last accessed on 4 December 2025).

6. Szturm T, Ireland DJ, Lessing-Turner M: Comparison of different exercise programs in the rehabilitation of patients with chronic peripheral vestibular dysfunction. J Vestib Res 1994; 4: 461–79

7. Han BI, Song HS, Kim JS: Vestibular rehabilitation therapy: Review of indications, mechanisms, and key exercises. J Clin Neurol 2011; 7: 184–96

8. Shepard NT, Telian SA, Smith-Wheelock M: Habituation and balance retraining therapy. A retrospective review. Neurol Clin 1990; 8: 459–75

9. Campbell AJ, Robertson MC, Gardner MM, Norton RN, Tilyard MW, Buchner DM: Randomised controlled trial of a general practice programme of home based exercise to prevent falls in elderly women. BMJ 1997; 315: 1065–9

10. Hopewell S, Chan AW, Collins GS, et al.: CONSORT 2025 statement: Updated guideline for reporting randomised trials. BMJ 2025; 389: e081123.

11. Gloor-Juzi T, Kurre A, Straumann D, de Bruin ED: Translation and validation of the vertigo symptom scale into German: A cultural adaption to a wider German-speaking population. BMC Ear Nose Throat Disord 2012; 12: 7

12. Mendel B, Bergenius J, Langius A: Dizziness symptom severity and impact on daily living as perceived by patients suffering from peripheral vestibular disorder. Clin Otolaryngol Allied Sci 1999; 24: 286–93

13. Institute for Quality and Efficiency in Health Care (IQWiG): General methods. 2022; Draft version 7.0. www.iqwig.de/methoden/general-methods_version-7-0.pdf (last accessed on 4 December 2025).

14. Strupp M, Dlugaiczyk J, Ertl-Wagner BB, Rujescu D, Westhofen M, Dieterich M: Vestibular disorders. Dtsch Arztebl Int 2020; 117: 300–10

15. Harcourt J, Barraclough K, Bronstein AM: Meniere‘s disease. BMJ 2014; 349: g6544

16. Adrion C, Fischer CS, Wagner J, Gürkov R, Mansmann U, Strupp M: Efficacy and safety of betahistine treatment in patients with Meniere’s disease: Primary results of a long term, multicentre, double blind, randomised, placebo controlled, dose defining trial (BEMED trial). BMJ 2016; 352: h6816

17. Maldonado Fernández M, Birdi JS, Irving GJ, Murdin L, Kivekäs I, Strupp M: Pharmacological agents for the prevention of vestibular migraine. Cochrane Database Syst Rev 2015; 2015: Cd010600

18. Brandt T, Dieterich M, Danek A: Vestibular cortex lesions affect the perception of verticality. Ann Neurol 1994; 35: 403–12

19. Zenner BP, Schmitz D, Zenner HP, Wirth M: [Structured ABEV exercises for the Treatment of vestibular dysfunction]. Laryngorhinootologie 2024; 103: 207–212

20. McDonnell MN, Hillier SL: Vestibular rehabilitation for unilateral peripheral vestibular dysfunction. Cochrane Database Syst Rev 2015; 1: Cd005397

e1. Doig GS, Simpson F: Randomisierung und Verschleierung der Allokation: Ein praktischer Leitfaden für Forscher. J Crit Care 2005; 20: 187–91; Diskussion 91–3

e2. Kurre A, van Gool CJ, Bastiaenen CH, Gloor-Juzi T, Straumann D, de Bruin ED: Übersetzung, interkulturelle Adaption und Zuverlässigkeit der deutschen Version des Dizziness Handicap Inventars. Otol Neurotol 2009; 30: 359–67

e3. Busner J, Targum SD: Die klinische Skala der globalen Eindrücke: Anwendung eines Forschungsinstruments in der klinischen Praxis. Psychiatrie (Edgmont) 2007; 4: 28–37.

e4. Mehnert-Theuerkauf A, Müller D, Lehmann C, Koch U: Die deutsche Version des NCCN Distress-Thermometers: Empirische Prüfung eines Screening-Instruments zur Erfassung psychosozialer Belastung bei Krebspatienten. Z Klin Psychol Psychiatr Psychother 2006; 54: 213–23

e5. Gloor-Juzi T, Kurre A, Straumann D, de Bruin ED: Übersetzung und Validierung der Schwindelsymptomskala ins Deutsche: Eine kulturelle Anpassung an eine breitere deutschsprachige Bevölkerung. BMC Hals-Nasen-Ohren-Störung 2012; 12: 7

e6. Mendel B, Bergenius J, Langius A: Schwere der Schwindelsymptome und Auswirkungen auf das tägliche Leben, wie sie von Patienten mit peripherer vestibulärer Störung wahrgenommen werden. Clin Otolaryngol Allied Sci 1999; 24: 286–93

e7. Maurischaft C, Bestmann C, Hasenbein U, Bullinger M: Strukturelle Validität des Short Form 36 Health Survey (SF-36) bei neurologischen Patienten. Neurol Rehabil 2004; 10: 13–20 .

e8. Hopewell S, Chan AW, Collins GS, et al.: CONSORT 2025 statement: Aktualisierte Richtlinie für die Berichterstattung über randomisierte Studien. BMJ 2025; 389: E081123