Health anxiety (HA) by proxy is defined as parents’ excessive fear that their child is experiencing a serious illness although this is not confirmed by a medical professional [1]. It is characterized by obsessive rumination about the health of one’s child and is often accompanied by control or avoidance behavior, such as facilitating repeated medical evaluations of the child or avoiding all information about child illnesses [2]. More severe cases may be observed in parents diagnosed with HA [3], but milder cases on the continuum are also likely to be recognized by health professionals working in pediatric settings.

A recent qualitative study investigating the lived experiences of parents with HA by proxy found that these parents experience significant distress as a result of their anxiety, and they struggle with finding the right balance in worrying about their child’s health [4]. Furthermore, the parent’s relationships with their child, other parent, and health professionals are affected [4].

HA by proxy may also have consequences for the child. Evidence suggests that HA symptoms in parents with health worries and increased focus on somatic concerns play a significant role in HA symptoms in children [5]. Furthermore, the parent’s response to the child’s health concerns [6] and how the parent copes with their own symptoms have been found to influence the health attitudes and behaviors of the child [7-10]. This risk may be especially high in HA by proxy because the child is exposed to a particular parental preoccupation with symptoms and fear of illness.

The members of the research team behind this study have previously investigated the assessment of HA by proxy by developing the Health Anxiety by Proxy Scale (HAPYS) with promising results [2]. Now that parents with excessive worries over their child’s health can be identified, the next important step is to develop effective treatment. Existing treatments do not specifically address HA by proxy. A treatment program has already been developed to target parents with anxiety to prevent anxiety in their offspring [11], and effective psychological treatments for HA also exist in both face-to-face and internet-delivered formats [12-14]. Specifically, a recently developed treatment program for HA based on the principles of acceptance and commitment therapy (ACT) and delivered over the internet (iACT) [15] has shown promising results [14,16] and was usable as a template for the development of specialized treatment for parents with HA by proxy.

PROXY is inspired by the existing iACT program for HA and uses the same web platform [17]. The treatment content for PROXY was developed de novo in several steps by 4 of the authors: 3 psychologists (KI, DHF, and LF) and a child and adolescent psychiatrist (CUR) with clinical and research experience with HA. Furthermore, 2 of these authors (LF and CUR) are also trained specialists in cognitive behavioral therapy (CBT). The development process included a number of steps. First, ideas for content, themes, exercises, videos, and audio files were brainstormed. These were presented as sketches to 3 parents diagnosed with HA and with worries about their children’s health who subsequently provided group feedback.

Next, the treatment manual went through several rounds of feedback from the project group. Further written feedback on the text manual was provided by 2 (67%) of the aforementioned 3 parents with individual comments in Microsoft Word document format. These 2 parents also participated in patient videos about living with HA by proxy, how to involve relatives, and how to speak to their children about their anxiety. Finally, the content was programmed on a web platform with close collaboration between the project group and web developers.

The thinking-aloud method was used to test the usability of specific selected parts of PROXY. The method involved participants verbalizing their thoughts while using the web-based program [18,19]. A male layperson aged 25 years without prior knowledge about the treatment and a female patient aged 47 years with HA by proxy and prior ACT group treatment of HA explored how they experienced (1) navigating the treatment and (2) potential technical issues, as well as (3) how they understood the program. Generally, they navigated the platform easily. Minor adjustments to the instructions, navigation, the choice of wording, and models were made.

PROXY is an 8-week therapist-supported internet-delivered program containing written psychoeducation, audio files, behavioral exposure exercises, homework, and videos distributed in 8 modules (Textbox 1). The therapist answers questions and provides feedback every week using an embedded secure message system.

The content framework is based on a combination of ACT [15] and CBT [20]. ACT aims to create greater psychological flexibility that allows (1) full awareness of here-and-now experiences, with an attitude of openness and curiosity; and (2) conscious and deliberate decisions inspired by core values, that is, deepest desires for who one wants to be and what one wants to stand for in life [15,21]. Thus, PROXY targets parents’ maladaptive anxiety-driven behaviors toward their child with focus on their gradual exposure to the things or situations that trigger their anxiety. This means that the CBT concept of exposure is urged to be based on the patient’s personal value. Furthermore, practicing acceptance helps to reduce the influence and impact of the painful thoughts and feelings that arise during the exposure.

Each of the first 7 modules in PROXY offers a helper, which is a treatment component (eg, awareness of values, acceptance, and self-compassion) that will help the patient move in a value-based direction in life instead of being trapped trying to control and avoid anxiety (Textbox 1).

In patient groups in which the prevalence of a health condition is low or unknown, as is the case with HA by proxy, randomized controlled trials are either not possible or unlikely to succeed [22]. Single-case experimental design (SCED) studies are used in clinical psychology to determine whether an effect of an intervention at an individual level has occurred [23-25]. Thus, a SCED study allows testing the effect of treatment for parents with HA by proxy in a small sample size while still controlling for external factors and upholding internal validity [26].

Consequently, this paper aims to describe the design for the first testing of PROXY using a SCED. The study investigates the potential effect of PROXY on parents’ worries about their children’s health. We hypothesize an initial increase in worries and anxiety, followed by a delayed and gradual reduction.

A SCED study implies control of the independent variable (here, PROXY) while measuring the dependent variable (here, parental worries about their child’s health) repetitively and often [24]. SCEDs should be distinguished from both case studies, which are not experimental but descriptive, and designs based on group comparisons where the experimental unit is the participant and where participants are assigned to different groups [27]. In a SCED, the experimental units are the repeated measures of a specific variable under investigation. Various types of SCEDs exist, but only the replicated randomized single-case AB phase design type will be elaborated here [23,27].

The project was registered on ClinicalTrials.gov on April 2, 2021 (NCT04830605).

This study will apply a replicated randomized single-case AB phase design for testing the treatment effects. AB designs are particularly useful when testing a psychological treatment that cannot be subjected to withdrawal or to repeated reversal between treatment and baseline. By adding a randomization feature and a replication feature to this basic design, both the internal and external validity are strengthened [27-29]; for example, potential confounding variables related to time, such as natural changes in the participants’ anxiety level or events that may affect anxiety, are statistically controlled for when the time point for the introduction of the intervention is randomized. In addition, replicating the design adds to the generalizability.

Specifically for this study, the intervention is randomly introduced after 7 to 26 days in the baseline phase (phase A). After the intervention (phase B), a follow-up phase (phase C) lasting between 14 and 33 days, depending on when the intervention was introduced, ensures that every participant has the same number of daily measurement points throughout the study (96 in total; Figure 1).

Five parents assessed with HA by proxy using the HAPYS [2] will be recruited for the study. No more than 3 participants may be allocated to a baseline duration of between 7 and 10 days to ensure that the participants will have different start times. The specific inclusion and exclusion criteria are presented in Textbox 2.

Participants are recruited in two ways: (1) participants self-refer to the project through the secure email system at the Research Clinic for Functional Disorders and Psychosomatics (hereinafter the Research Clinic), Aarhus University Hospital, and complete the project questionnaire (Table 1) in Research Electronic Data Capture (REDCap; Vanderbilt University), which includes written consent to assessment [30,31], and, after a diagnostic interview, they are included in the project, if eligible; and (2) participants are recruited at their ordinary assessment for HA at the Research Clinic if they are also clinically assessed as experiencing excessive worries about their child’s health.

Participants are assessed using a short standardized diagnostic interview based on schedules for clinical assessment in neuropsychiatry (SCAN) [32,33] and supplemented with an assessment of HA by proxy using the HAPYS.

All participants receive written and verbal information about the project before signing the consent to participate. The project has been approved by the Danish research ethics committee (1-10-72-296-20) and registered with the Danish data protection agency (1-16-02-921-17).

HA by proxy is a novel research area where much still remains to be investigated. Recent studies suggest that parents experiencing this type of anxiety struggle with how to cope with their health-related worries about their child [4]. When a parent directs special attention toward bodily symptoms and demonstrates health worries regarding their child, research suggests negative consequences for the child, such as negative illness perception and maladaptive symptom coping [5,6,67,68]. Together, this underlines the need for the development and testing of more specific treatment options. Therefore, we developed the first systematic treatment program for HA by proxy, named PROXY.

In this study protocol, we described the treatment components of PROXY and how these components were established in the web-based treatment program. It is a collaboratively developed treatment program that involved clinical experts, web developers, and users in the process. PROXY received positive feedback from patients during development, and it seems to be a usable treatment program. Conducting this first testing and having patients receive treatment in the program may shed light on how to further improve the treatment content of PROXY.

The SCED approach was chosen for the first testing of PROXY because it is particularly useful when testing novel treatments for smaller patient groups [22]. It enables us to measure potential effect with a limited number of participants and to investigate treatment courses for each participant.

The randomization and replication across participants are important methodological strengths of this SCED study. By using randomization and randomization tests, confounding factors that are time related are statistically controlled for [27-29]. The external validity is strengthened when this procedure is repeated for several participants [69]. This accommodates the criticism against the standard single-case AB phase design of having low internal validity and a lack of multiple phases without data collected concurrently [70,71].

The selected test statistics for the randomization test are mean comparison and predicted response functions. Using test statistics based on predicted response functions for SCED studies is a newly described framework with the purpose of offering statistical analysis for data with nonlinearity and delayed or gradual changes to the primary outcome measure, as is often seen in psychological treatment [60]. The advantage of response functions is the ability to test very specific developmental trajectories—this is central when testing therapeutic interventions where the introduction of therapeutic techniques can be hypothesized to cause sudden and temporary increase or decrease in symptoms (eg, exposing oneself to feared stimuli may initially increase one’s anxiety symptoms). However, the great specificity of the trajectory described by a response function also means that there is a high risk of failing to reject the null hypothesis. Therefore, primary analyses include visual analysis. Furthermore, using the multiverse approach enables the testing of whether slight adaptations of the characteristics of the trajectory are more fitting.

As the treatment for patients with HA by proxy has not been previously examined, this study is particularly important in terms of describing the treatment development and content as well as the research protocol for using a SCED for the first testing of PROXY. Applying a SCED provides us with the opportunity to investigate the effect of the treatment in a small sample and explore individual treatment courses in detail. Furthermore, the detailed visual analysis provides implicit tests of treatment mechanisms. This has relevance in clinical practice and can potentially be used to inform larger studies on the mechanisms of change and when and how the treatment may be offered to patients in the future [22]. In addition, this study may help to further refine the treatment content of PROXY.