Protocol for a double-blinded randomised controlled trial and process evaluation of a digital psychotherapeutic app in Singapore to improve symptom burden in patients with dermatological problems

Abstract

Introduction

Chronic inflammatory skin diseases, despite low mortality, significantly impair quality of life (QoL). Up to 80% of patients with dermatological conditions experience severe itch and poor sleep, as well as related mental health challenges such as anxiety and depression. The relationship between skin diseases and mental health highlights the challenges that doctors face in treating these conditions. Existing psychotherapeutics, such as mindfulness training, cognitive behavioural therapy and acceptance and commitment therapy, are widely used and effective in the treatment of mental health illnesses. However, there is limited evidence on the application of such interventions in dermatology, and most mental health apps lack robust clinical evaluation. We report the design of a randomised controlled trial to evaluate the efficacy and implementation of a mobile app containing dermatology-specified psychotherapeutic strategies in reducing QoL burden.

Methods and analysis

English-speaking patients aged 16 years and older with psoriasis, eczema or chronic urticaria will be recruited and randomised into the intervention arm (psychotherapeutic application) or active control group (Healthy365 app, a general wellness application managed by the Singapore Health Promotion Board). This allows a comparative assessment of app-usage-specific outcomes while preserving the blinding of all participants. The primary outcome is the change in the Dermatology Life Quality Index (DLQI) score from baseline to week 8. Secondary outcomes include physician-assessed disease severity at weeks 8 and 16 relative to baseline, differences in other patient-reported measures at weeks 8, 16 and 32, self-reported treatment adherence and initiation/escalation of systemic medications. To understand how patients engage with the app, we will evaluate the implementation process, focusing on key measures such as engagement, satisfaction and willingness to pay. Statistical analysis will be carried out on an intention-to-treat basis, and missing data will be analysed using last observation carried forward.

Ethics and dissemination

All participants will receive both verbal and written study information that aligns with Good Clinical Practice guidelines. Ethical approval has been obtained from the National Healthcare Group’s Domain Specific Review Board (reference number: 2022/00751). Results will be disseminated via publication in a relevant journal. Data will be available from the corresponding author on reasonable request.

Trial registration number

NCT06702293.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This evidence-based intervention is built using the user-centred design technique, with participation from patients, physicians and business.

• It follows a randomised and double-blinded methodology with the use of an active control.

• Limitations include possible loss of blinding or guessing of study hypothesis, limited use of objective measures and potential inter-rater variability of assessments.

• Participants are continued on standard medical care, which may include changes to the medical treatment that may confound the outcome measures.

Introduction

Despite low mortality, skin diseases carry a high quality of life (QoL) impairment and disability.1,3 Globally, skin diseases are the fourth leading cause of non-fatal disease burden.⁴ A multicentre cross-sectional study of over 5000 participants found a EuroQol-5 Dimension (EQ5D) score of 0.66 in psoriasis and atopic dermatitis,⁵ which reflects greater disability compared with patients with stage IV breast cancer (EQ5D=0.73)⁶ and renal failure on dialysis (EQ5D=0.77).⁷ Up to 80% of patients with dermatological conditions experience significant burden from itch and poor sleep,^{8 9} with anxiety and depression afflicting 10–20%.¹⁰ The bidirectional relationship between skin diseases and mental health can be explained by the diathesis-stress model.^{11 12}

The QoL impairment in dermatological conditions is strongly correlated with psychological constructs such as negative social comparison, resilience, stigmatisation, anxiety and depression.^{13 14} This highlights the contribution of psychological and behavioural constructs towards perceptions of itch, skin discomfort and eventually symptom burden. While psychotherapeutics such as mindfulness training, cognitive behavioural therapy and acceptance and commitment therapy are widely used and effective in the treatment of mental health illness,^{15 16} there is limited application of such interventions in specific disease fields like dermatology, although small preliminary studies have shown the efficacy of isolated psychotherapeutic interventions in attenuating QoL impairment and objective disease outcomes.^{17 18}

Unfortunately, there is a large treatment gap and a lack of access to mental healthcare for many patients, whether in primary, secondary or tertiary care.¹⁹ Many non-psychiatric clinicians may lack the skillset and time to deal with the psychosocial aspects of disease in their patients.^{20 21} The time-intensive and often costly services by psychologists and psychiatrists are usually limited to patients who manifest significant downstream psychiatric diseases. The lack of access underscores the value of delivery mechanisms that can go beyond the boundaries of clinicians’ time and expertise.

In addition to improving access, the use of a digital electronic platform to deliver psychotherapeutic interventions can provide anonymity, flexibility and scalability. It can help patients with dermatological disease to better cope by reducing symptom burden without constant contact with a healthcare provider.^{22 23} Industry-wide, digital mental health interventions are increasingly popular, with a compounded annual growth rate of 20%.²⁴ Yet few have been robustly tested and subjected to gold-standard clinical evaluations like randomised controlled trials.²⁵

The primary aim of this study is, therefore, to produce scientific evidence on the effectiveness of a newly developed self-administered psychotherapeutic toolkit for managing skin disease. The secondary aim is to understand the process of implementing the toolkit, including how it is perceived by stakeholders and its associated costs. By delivering it via a mobile app, we aspire to provide a more accessible and sustainable approach to addressing the psychosocial dimensions of dermatological conditions and ultimately improving the overall well-being of patients.

Methods/approach

Aims

The primary aim is to evaluate the efficacy of a digital toolkit of psychotherapeutic strategies for reducing QoL burden in patients with chronic inflammatory skin conditions such as eczema, psoriasis and chronic urticaria using a double-blinded randomised controlled clinical trial.

The hypotheses are as follows:

• Self-administered psychotherapeutic interventions via a mobile application will lower QoL impairment (Dermatology Life Quality Index (DLQI)) at 8 weeks, compared with baseline, in patients with eczema, psoriasis or chronic urticaria.

• An improvement in secondary outcomes, including itch scores, sleep impairment, illness perception and psychiatric comorbidity, will be observed up to 32 weeks post-recruitment.

• Participants’ demographics and psychographic attributes (eg, personality traits) are related to the adoption and effectiveness of the intervention.

The secondary aim is to understand the implementation of the toolkit, including how it is perceived and its associated costs. This will be achieved via a process evaluation methodology, grounded in the established Medical Research Council guidance for complex intervention evaluations.²⁶ The process evaluation will be carried out in the intervention arm of the study; it will assess descriptively the dosage and wider implementation aspects of the toolkit, as well as its associated costs. These data will inform larger future studies and implementation within existing clinical services in Singapore.

Study design and population

A multicentre, randomised controlled, double-blinded, effectiveness-implementation type I hybrid trial will be conducted at three outpatient dermatological centres in Singapore (National University Hospital, National Skin Centre and KK Women’s and Children’s Hospital). To test for generalisability and address potential selection bias, participants will also be recruited from the general public and through word of mouth. The study is scheduled to be conducted between 2025 and 2028.

Inclusion criteria:

• Patients with psoriasis, eczema or chronic urticaria diagnosed by a dermatologist.

• Aged 16 years or older.

• Peak pruritus intensity of 4 or more on an 11-point numerical rating scale (0–10, with 10 representing the worst itch).^{27 28}

• Able to engage with a mobile application in the English language.

Exclusion criteria:

• Patients with active psychiatric symptoms (eg, active suicidal ideation and psychosis).

• Patients with an unstable psychiatric condition, characterised by:

  • Hospital inpatient admission for a psychiatric condition or initiation of a psychotropic medication in the prior 3 months.

• Patients with unstable dermatological conditions, characterised by:

  • A recent flare of the skin condition within the last 2 weeks, including diagnosis of a flare by a doctor, use of systemic steroids, oral antibiotics, wet wrap rescue therapy, oral antivirals or increase in the frequency of phototherapy or the dose of systemic medications such as biologics or conventional immunosuppressants for the dermatological condition within the last 2 weeks OR

    • Any of the following within 3 months:

    • Inpatient admission for a dermatological condition.

    • Initiation of phototherapy.

    • Initiation of systemic therapy (conventional immunosuppressive agents, biologics and Janus Kinase (JAK) inhibitors) within the last 3 months.

Trained study staff will obtain signed informed consent from participants (online supplemental file 1). Participants will be recruited and randomised in a 2:1 ratio into either an intervention arm or a control arm, respectively, using stratified block randomisation by disease (ie, eczema, psoriasis or urticaria). A centralised web-based application will be employed to generate the randomisation list and maintain allocation concealment.

Intervention

This fully automated and self-administered intervention will be delivered via a Health Insurance Portability and Accountability Act-compliant mobile app developed in collaboration with Intellect, a Singapore-based mental health company with broad reach and prior research collaborations in the Asia Pacific region.29,31 A detailed discussion focusing on the engineering of the application, including the user interface, data acquisition and curation technology, security and privacy assurance and user pilot testing, is beyond the scope of this article but will be available from the authors.

The intervention includes modules that aim to address the following interventional domains:

1. Disease understanding and coherence:

   These modules deepen participants’ comprehension and acceptance of the pathophysiology and chronicity of their skin condition. Topics covered include the understanding of disease pathogenesis, allergies and triggers, self-assessment of disease severity and the mechanisms underpinning the skin-brain connection.

2. Disease management strategies:

   Concentrated on empowering individuals to proactively manage their skin conditions, these modules encompass the management of allergies, dietary modifications and strategies for minimising scratching.

3. Emotional representation of disease:

   These modules focus on addressing maladaptive cognitive processes that influence psychological outcomes such as negative social comparisons, anxiety and depression. The aim is to empower patients to recognise that they can lead fulfilling lives by fostering more flexible and positive emotional representations of their skin conditions. These modules cover values training, cognitive restructuring and behavioural therapy, mindfulness strategies and guided visualisations for alleviation of distressing symptoms like itch.

4. Functional impairment due to disease:

   These modules address the social, recreational and work-related challenges commonly faced by patients with skin disease. Content includes tips and recommendations for adapting to life with skin disease, many of which were built around real-life experiences and sharing to improve relatability. Strategies for interaction with others, such as dealing with unwanted comments, are also shared.

5. The patient-healthcare provider interaction:

   This section guides patients through a typical dermatological consult, with guidance on how they may optimise and achieve the most from their limited time with the doctor. Recognising the difficulty patients may face with getting timely support from doctors when required, modules such as itch crises rescue sessions provide some form of 24/7 support when access to a doctor is not possible.

The digital intervention is delivered through learning paths, rescue sessions, guided journals/check-ins and self-assessments as follows:

• Learning paths comprise the bulk of the intervention and are structured modules for knowledge delivery. Walk-through exercises are included for practising the taught skills. A learning path may comprise 1–4 modules, with each module expected to take 3–8 min to complete.

• Rescue sessions are intended for on-demand access by users to alleviate the distressing symptoms of itch or worry. These guided sessions incorporate mindfulness, distraction and other techniques as a digital companion to get the patient through the distressing moment.

• The journaling sessions facilitate tracking of progress. These include guided journals to document triggers and track itch, sleep and disease severity scores.

• The self-assessment component includes validated questionnaires such as the Perceived Stress Scale³²and Self-Compassion Scale,³³ promoting self-reflection and insight. Score interpretations that are contextualised with average population data, along with recommendations for suitable modules.

During the consent-taking and onboarding process, patients will be guided on how to download the app onto their mobile phones and encouraged to use it at least two times a week for the entire 32-week study period. Participants have the autonomy to choose the content and tools, as well as the order and pace at which they are used. This approach recognises the heterogeneity of preferences and needs across individuals. To support personalisation, tailored recommendations are offered based on participants’ interactions with the app. Elements like quizzes are incorporated to promote engagement and quotes from fellow patients to enhance relatability and peer learning.

Control

Participants in the active control group will be similarly instructed to download and engage with the Healthy365 app,³⁴ a different mobile application managed by the Singapore Health Promotion Board. The Healthy365 app offers a variety of dynamic general health-related features that are not specific to skin disease. These include personalised health tracking, participation in fitness challenges, access to nutritional guidance and lifestyle recommendations. The use of an active control allows a comparative assessment of app usage-specific outcomes, while preserving the blinding of all participants.³⁵

Other treatment

In both groups, participants will be encouraged to continue adhering to routine medical advice and care, including scheduled clinic consultations. To ensure that patients continue to receive the best care, there are no limitations or restrictions on the therapy that may be prescribed for their skin conditions.

Blinding

Patients, physicians and clinical assessors will be blinded. To preserve the blinding, participants will be told that the study aims to evaluate the effects of a mobile application on their health behaviours. However, specific details of the psychotherapeutic dermatology app will be withheld. Participants will be instructed by the research assistant not to disclose details of the mobile application to their dermatologist or clinical assessor during their visits. Any instances of unblinding, whether intentional or inadvertent, will be documented and reviewed by the team to assess their influence on study validity and determine any necessary next steps. A summary of the recruitment flow is presented in figure 1.

Figure 1. Flow diagram of study protocol.

Outcomes

The primary clinical outcome will be the difference in the change of DLQI³⁶ from baseline (skin-specific QoL impairment) at week 8 between the intervention and control groups.

Secondary clinical outcomes include the change in:

• Physician assessed disease severity at weeks 8 and 16 relative to baseline.

  • Body surface area affected.

  • Investigator/physician global assessment on a 0–4 numerical rating scale (NRS).³⁷

  • Eczema Area and Severity Index (only for participants with eczema).³⁸

  • Psoriasis Area and Severity Index (only for patients with psoriasis).³⁹

  • Urticaria Activity Score-7 (only for patients with chronic urticaria).⁴⁰

  • Composite measure of scratching (eg, excoriation, lichenification and prurigo).

  • Global, emotional, functional and physical disease severity grading.

  • Perception of patient-reported global disease severity grading.

  • Clinical global change (weeks 8 and 16 only).^{41 42}

• Difference in other patient-reported measures.

  • At baseline and weeks 4, 8, 12, 16, 24 and 32:

    • 24-hour worst itch score, 1-week peak and average itch scores (NRS).^{27 43}

    • Skin pain NRS.⁴³

    • Sleep disturbance NRS.⁴⁴

    • Global disease severity grading.¹⁴

    • Patient Global Impression of Severity. ⁴¹

• Patient Global Impression of Change (weeks 4, 8, 12, 16, 24 and 32).⁴¹

  • At baseline and weeks 8, 16, 24 and 32:

    • Patient-Reported Outcomes Measurement Information System (PROMIS) Depression 8b.^{45 46}

    • PROMIS Anxiety 7a.⁴⁵

  • Modified Work Productivity and Activity Impairment Questionnaire plus Classroom Impairment Questions: Specific Health Problem, V.2 (only at baseline and week 16).⁴⁷

  • EQ-5D-5L (5-level) (only at baseline and week 32).⁴⁸

  • DLQI (at weeks 4, 12, 16, 24 and 32).³⁶

• Initiation or escalation of systemic treatments (inclusive of conventional immunosuppressants, biologics and systemic steroids) (at baseline and weeks 8, 16 and 32).

The following measures will also be collected to account for potential confounders, mediators and moderators. These will be incorporated into statistical models to better understand the relationship between the intervention and the primary and secondary outcomes.

• Age, gender, education and occupation.

• Duration of dermatological disease.

• Comorbidities.

• Type D Scale-14 (DS-14) (Type D personality).⁴⁹

• 5D Pruritus Scale.⁵⁰

• PROMIS Itch Questionnaire—Scratching Behaviour 5a.⁵¹

• Acceptance of Chronic Health Conditions Scale.⁵²

• Revised Illness Perception Questionnaire.⁵³

• Brief Resilience Scale.⁵⁴

• PROMIS Self-Efficacy for Managing Chronic Conditions—Managing Emotions 4a.⁵⁵

• Two questions from PROMIS Self-Efficacy for Managing Chronic Conditions—Managing Social Interactions.⁵⁵

• Two questions from the Cognitive and Affective Mindfulness Scale-Revised.⁵⁶

• Weight Bias Internalisation Scale, modified for skin conditions.⁵⁷

• Dermatology Social Comparison Scale.⁵⁸

• WHO-5 Well-Being Index.⁵⁹

• Physician-reported and self-reported treatment adherence (scale validation in process).

Research questionnaires will be administered electronically at weeks 0, 4, 8, 12, 16, 24 and 32. An in-person physical examination by the study team would be conducted at baseline, week 8 and week 16.

Process evaluation

The process evaluation will address prospectively and descriptively aspects of implementation in the intervention arm of the study—specifically the feasibility, usability and practicality of the intervention, covering areas such as content delivery, user engagement and costs.

Measures of implementation that will be collected:

• Frequency and duration of app usage.

• Percentage of completed learning paths/modules.

• Appropriate responses and engagement to tracking exercises and open-ended questions in the app.

• Patient engagement and adoption using the App Engagement Scale.⁶⁰

• Likely frequency of app usage after the study.

• Likes and dislikes of the app.

• Suggestions and feedback to improve the app.

• We will also track collective server access statistics to determine which pages have the lowest and highest traffic, as well as analyse ‘helpfulness’ ratings of each module. This will guide further modification and improvement of the tool for future research and clinical applications.

• Barriers and facilitators of use and long-term sustainability will be gathered from qualitative interviews (online supplemental file 2) conducted after the exposure to the intervention is completed. These will be carried out by trained interviewers with a subset of patients from the intervention group, purposively sampled to ensure maximum variation in terms of demographic and clinical profiles. The final sample size for this part of the study will be determined via the saturation criterion, that is, participants reporting similar themes in their interviews, which is estimated to be reached within 20–40 participants in total.⁶¹ Interview topic guides will explore factors that facilitated or hindered engagement with the intervention, its impact on the healthcare experience and suggestions for refining the intervention and integrating it into clinical services for broader implementation.

Implementation costs:

• Maximum willingness to pay per month to continue using the tool, evaluated at the final time point (week 32), using a single bounded question: ‘How much would you maximally be willing to pay/month to continue using this mobile app?’ Options range from ‘I would not use it even if free’, to ‘I would pay $41 or more/month.’

• Costs of development and deployment will be assessed using a micro-costing approach. Development costs will be prospectively recorded by tracking personnel hours spent on content creation and direct payments made to engage the app developer. Deployment costs include app infrastructure, software and maintenance costs and personnel hours to promote and facilitate patient onboarding. Costs from the patient perspective will be assessed via retrospective self-reports on time spent using the app, including onboarding, regular usage and troubleshooting, collected at the final time point.

Sample size calculation

We estimate a 3-point difference (SD: 10) in the improvement of DLQI scores at week 8 from baseline.^{22 36} This was based on systematic reviews of psychological interventions in atopic dermatitis showing a pooled effect size of 0.37.¹⁷ As self-administered interventions typically have a lower effect size, a conservative estimate of 0.3 was selected. With an allocation ratio of 2:1 (intervention:control), a total sample size of 525 will be sufficient to detect the difference with a power of 90% and an alpha of 0.05.⁶² A 30% attrition was factored in, leading to a total target of 690 patients (intervention 460, control 230).

Statistical analysis

Statistical analysis of all study outcomes will be carried out on an intention-to-treat basis. The last observation carried forward will be used for missing data. The primary outcome (change in DLQI score from baseline to week 8) would be compared between the two arms using a general linear model, adjusted for baseline DLQI. Secondary outcomes involving repeated measures over time would be analysed using the linear mixed model or generalised estimating equation model, whichever is more appropriate. If no significant interaction effect between treatment and time is detected, a main effect model with group, time and baseline measurement will be used to compare the outcomes between the two arms and among the different times. Sensitivity analysis will be conducted, including a per-protocol analysis, where only patients who used the app at least two times a week for 70% or more of the observed period will be included. A sensitivity analysis using multiple imputation to handle missing data will also be used.

The following subgroup analyses will be performed:

1. By disease type.

2. By baseline disease severity using established strata for mild/moderate/severe disease.^{40 63 64}

3. Presence of psychiatric comorbidities.

4. By usage of the tool (participants who access the application at least once a week with an average weekly usage of at least 15 mins vs those who do not).

For the process evaluation, qualitative data from interviews will be transcribed and coded into themes using thematic analysis.⁶⁵ In brief, descriptive open coding will first be performed to identify initial themes, followed by group discussions to refine and clarify these emerging themes. Subsequently, analytic-focused coding will be performed on the qualitative dataset. Final extracted themes will be tabulated, and illustrative quotes from participants will be extracted to support them. For cost analysis, personnel hours will be multiplied by the respective hourly wage for each role to obtain labour cost. Total expenses, including direct payments to the app developer and monthly maintenance fees, will be aggregated. From the patient’s perspective, their self-reported time spent using the app will be multiplied by the national mean hourly wage, calculated from the national mean gross monthly income from employment divided by 160 hours.

Safety and suicide ideation management plan

On recruitment into the study, participants are provided with contact details of the research team as well as details of emergency services and hotline numbers. Submitted responses in the research data collection sheets would be reviewed at least weekly. Patients with a PROMIS Depression score of 33 and above or an anxiety score of 28 and above will be flagged. These thresholds indicate a severe range, defined as 2 or more SD from the mean and represented by a T score of 70 or higher.^{45 66 67} A research coordinator will attempt to contact the patient up to three times to provide a recommendation to seek a psychiatric consultation. If indicated, the 6-question Columbia-Suicide Severity Rating Scale68,70 will be used to assess suicidal ideation and risk, with patients at high risk recommended to seek urgent psychiatric consultation through emergency services. Patients with a PROMIS Depression score of 23–32 or an anxiety score of 20–27 will also be flagged. These thresholds indicate a moderate range (T score of 60–69.9).^{45 66 67} Support will be provided to all flagged cases to facilitate psychiatric referrals, and participants will be given links and contact numbers for mental health resources. The site Principal Investigator (PI) and the patient’s clinician will also be informed of flagged cases.

Data that are entered into the mobile app are de-identified for data protection and thus not actively monitored for safety signals. Participants will be made aware that their entries in the app will not be reviewed by healthcare providers.

An interim analysis will be conducted at 6 months or after the first 200 recruited participants (whichever is earlier) to assess for any safety signals and allow for recalculation of sample size.

Patient and public involvement

The study incorporates the insights and perspectives of patient and public representatives, from developing and curating the psychotherapeutic modules to providing feedback on the study design. These members will also support the analysis and dissemination of the study results. There are currently three patient representatives involved in this study.

Trial management and oversight

The trial’s coordination and supervision will be led by the overall study PI (EC). The PI, together with co-investigators, will monitor the progress of the trial, protocol adherence, patient safety and review updated information. The intervention content has been collaboratively developed with input from patients and reviewed by dermatologists, a clinical psychologist and a psychiatrist. Thus, we anticipate minimal harm resulting from trial participation.

Nevertheless, precautionary measures include providing emergency service numbers and 24-hour crisis hotlines in both the consent form and within the mobile application. Participants will also receive contact details for the research team, enabling prompt escalation of any concerns or reported adverse events to the study team and principal investigator for swift and secure resolution.

Data analyses will be conducted on pseudonymised data, with identifiable information stored securely under password-protected files or lock and key, accessible solely to the study investigators. Participants will be allowed to discontinue the trial on request to the trial coordinator.

Ethics and dissemination

All participants will receive both verbal and written study information that aligns with Good Clinical Practice guidelines. Ethical approval has been obtained from the National Healthcare Group’s Domain Specific Review Board (reference number: 2022/00751). Results will be disseminated via publication in a relevant journal. Data will be available from the corresponding author on reasonable request.