Increasing Retention in a Trial by Supporting Basic Psychological Needs for Autonomy, Competence and Relatedness: Developing a Theory‐Informed App‐Based Complex Intervention

ABSTRACT

Aims and Objectives

Most interventions to improve retention in randomised trials have low or uncertain evidence of effect and are not informed by behaviour change science even though behaviour change is their aim. Self‐Determination Theory is an evidence‐based theory within behaviour change science which argues that supporting basic psychological needs for autonomy, competence and relatedness leads to high‐quality motivation, active engagement and well‐being. The purpose of this study was to develop a need‐supportive digital intervention supporting these three psychological needs to increase retention in a randomised allergy trial.

Methodological Design

We applied the Medical Research Council's framework for complex interventions and merged the development phase with steps suggested by O'Cathain and colleagues. Development was an iterative process involving patient partners. It included: (1) Evidence review, (2) Identification of theory, (3) Interviews, (4) Workshops, (5) Prototyping, and (6) Programme theory.

Results

We developed a web app intended to support participants' basic psychological needs informed by findings from stages one to six in this development process. It included the following features: (a) the option to only complete all questions on days with symptoms, (b) the possibility to answer up to 48 h later, (c) positive feedback when responding to a questionnaire, (d) pollen counts and forecasts, (e) personal and total response rate, (f) visualisation of data entered, (g) advice about grass pollen allergy, (h) updates on the status of the allergy trial, (i) an ability to change the time of the reminder, (j) download icon and (k) contact information.

Conclusions

This study provided findings on how to support basic psychological needs among participants with grass pollen allergy in a randomised trial through need‐supportive technology. The web app was ready for a randomised feasibility test within the allergy trial during the grass pollen season of 2022.

Trial Registration

Allergy trial: EudraCT 2020–006020‐28

1. Introduction

Retention of participants is a common issue in clinical trials [1]. Retention is when participants continue to participate in a trial after being recruited and randomly assigned [2]. Non‐response and loss to follow‐up will affect retention and can create an imbalance between trial groups, which will reduce the internal validity and generalisability of a trial [2]. According to Schulz and colleagues, bias is introduced at a loss to follow‐up of 5%, and validity is severely threatened if the loss reaches 20% [1]. They therefore recommend that researchers develop and implement interventions to increase retention in clinical trials [1]. Against this background, there is a need to explore effective strategies for increasing retention in randomised trials.

One important line of work has focused on different ways and stages at which trial retention can be supported [3]. Through surveys and a consensus meeting with researchers, patients, and the public, they identified and ranked unanswered research questions on improving trial retention, such as enhancing motivation, reducing burden, and better leveraging routine care, data, patient involvement, and technology, especially mobile devices [3]. This provides a rationale for exploring mobile devices to address retention challenges.

Mobile apps and other technological tools, like text messages and reminders, have been used to improve trial retention [4]. The importance of this approach is evident, as mobile devices are available to a large part of the global population [5]. Their portability makes it easier to access health information anytime and anywhere [6], which could reduce data collection costs [7].

However, despite these advantages, the evidence for the effect of such interventions is limited. A Cochrane Review has found that existing interventions, whether technology‐based or not, show little evidence of effect in improving retention in randomised trials [4]. Many of the identified interventions intended to change the behaviour of participants but were, according to Gillies and colleagues, not informed by behaviour change science [8].

In general, human behaviour has received little attention in clinical trials, even though focusing on how and why people behave as they do has the potential to improve retention and trial success [8]. Self‐Determination Theory (SDT) is an evidence‐based theory of motivation and human behaviour [9]. Fundamental to it is the distinction between controlled and autonomous types of motivation—and thereby the type of motivation that moves people to act [9, 10]. According to the theory, sustained behaviour can be achieved when the three basic psychological needs for autonomy, competence, and relatedness are supported, as this leads to better internalisation and more autonomous actions [9].

No interventions informed by SDT, including digital ones, have been identified to improve trial retention. However, digital approaches are increasingly applied in health research, where they support participants and contribute to intervention quality, suggesting potential for improving retention when guided by SDT [7]. Developing and evaluating digital interventions informed by SDT thus seems relevant for increasing retention in randomised trials.

These general challenges with retention are also evident in the field of allergy and allergen immunotherapy, where only a few studies have followed participants beyond the treatment period, and common to them all were issues with retention [11, 12, 13, 14]. In one of our previous randomised trials with intralymphatic allergen immunotherapy (ILIT), 57% of participants did not attend the final consultation [15]. Attention to retention therefore seemed essential when designing a larger randomised trial testing the efficacy of ILIT with a two‐year follow‐up. Low adherence to treatment is distinct from low retention [4] and is not within the scope of this development study, as an advantage of this treatment is its brevity. Hence, this paper aims to describe the development process of a need‐supportive digital intervention to increase retention in a randomised controlled allergy trial with the official title: “ILIT.NU—Documentation of efficacy for intralymphatic allergen immunotherapy in a phase III randomised, parallel‐group, double‐blind placebo‐controlled multisite trial” (EudraCT 2020–006020‐28).

2. Methods

2.1. Setting

ILIT.NU is a randomised placebo‐controlled multicentre trial to document the effect of intralymphatic immunotherapy. After a baseline year in 2021, over 8 weeks participants received three injections into an inguinal lymph node. From 2022 to 2023, the effect of ILIT for grass pollen‐induced allergic rhinoconjunctivitis was investigated. Participants attended two annual consultations and filled out online questionnaires regarding allergy symptoms and use of medication during the grass pollen seasons. Links were sent using a short message service—daily for the Combined Symptom Medication Score [16] and weekly for the Rhinoconjunctivitis Quality of Life Questionnaire [17]. The data were stored on the secure web‐based platform REDCap [18].

2.2. Study Design

The development drew on the Medical Research Council's (MRC) framework for complex interventions. We considered ours to be a complex intervention as several components could interact and influence the participants' behavior with regard to retention. This interaction and possible influence are characteristic of a complex intervention [19]. Interventions with multiple interacting components are often more difficult to implement successfully in clinical practice because their effects can differ among participants [19]. This variation is due to the complexity of the components, the delivery context and the interactions between these factors. The MRC framework offers a structured approach for developing interventions that considers this complexity, thereby increasing the likelihood of success in health research [19]. Additionally, it is evidence‐based and internationally recognized [19].

We used the framework published in 2006, in which the development phase includes three stages: identifying existing evidence, identifying theory, and modelling process and outcomes [20, 21]. To further strengthen the development phase, we applied the guidance suggested by O'Cathain and colleagues on actions to take during intervention development [22]. This addresses how to iteratively develop an intervention with stakeholder input throughout [22]. We found it suitable as we wanted to involve patients with grass pollen‐induced rhinoconjunctivitis in the research process to develop an intervention relevant to participants in the ILIT.NU trial. Informed by these elements, we developed the intervention in six stages (1) Evidence review (2) Identification of theory (3) Interviews (4) Workshops (5) Prototyping and (6) Programme theory. Figure 1 illustrates the six‐stage process, which was iterative and not completely linear. The illustration is adapted from Sandbæk and colleagues [23]. The process lasted from August 2020 to April 2022. To improve the quality of the reporting, we used the ‘Guidance for reporting intervention development studies in health research (GUIDED)’ [24]. To describe the intervention to enable replication, we used the ‘Template for intervention description and replication (TIDieR)’ [25]. To describe how patients were involved in the process, we used the ‘Short form of the guidance for reporting involvement of patients and the public (GRIPP2‐SF)’ [26]. All checklists are included in the Supporting Information.

FIGURE 1.

The six‐stage iterative development process adapted from Sandbæk and colleagues [23].

2.3. Stage 1: Evidence Review

We identified existing published research evidence through a systematic literature search in PubMed and CINAHL based on criteria for review questions when developing a complex intervention [27]. It included type of population, type of intervention/comparison, and type of outcome of interest:

1. To what extent are participants with allergy retained in clinical trials?

2. What are the essential components of a need‐supportive intervention for clinical trials?

3. Is the use of a need‐supportive intervention more likely to retain participants with allergy in clinical trials?

A librarian helped develop the search strategy, which can be found in the Supporting Information. The overall categories of search terms were allergy/allergen immunotherapy, motivation/theoretical framework, patient participation/involvement, and trial retention/follow‐up/response. The search was conducted between September and November 2020. Only articles reporting clinical trials involving participants with allergy were included for questions one and three, and only articles reporting interventions intended to increase motivation for participation and retention in randomised controlled trials were included for question two. All types of study designs were included, but they had to be published in English. The first author performed the literature search and screened titles and abstracts for relevance, and this was followed by a full‐text screening and data extraction. Reference lists of included articles were hand‐searched. Included articles were assessed for quality using a relevant appraisal tool [28, 29, 30, 31].

2.4. Stage 2: Identification of Theory

During the evidence review, we did not identify any interventions informed by SDT intended to retain participants in trials, but Teixeira and colleagues argue that autonomous forms of motivation are associated with sustained behaviour across several health contexts [10]. We therefore identified SDT as an appropriate theory to draw on during the development of the intervention.

2.5. Stage 3: Interviews

We wanted to understand what it is like to live with grass pollen allergy and to participate in an allergy trial, including daily completion of questionnaires and annual consultations. The research questions focused on how allergy could influence participation, the positive and negative contextual influences on participation, and how the basic psychological needs were met as part of participating in an allergy trial. This should serve the aim of developing an intervention capable of supporting positive contextual factors and participants' basic psychological needs in the ILIT.NU trial. Due to the Covid‐19 pandemic, we planned online interviews using Zoom. The sample size guidance for interview studies indicated that 10 participants would be appropriate [32]. The informants were participants in a former small‐scale ILIT trial (ILIT.SS), which was similar to the ILIT.NU trial in terms of trial activities, but involved fewer participants [33]. The guide for the semi‐structured interviews was based on Kvale and Brinkmann [34], where several research questions were raised based on the output of stages one and two regarding interventions to increase trial retention and the application of SDT in this context. Then each research question was the basis for a series of interview questions. The research team formulated the questions, but they were not pilot tested by participants. The interview guide is provided in the Supporting Information. The interviews were transcribed and then analysed using NVivo software version 1.7.1 for Mac and the systematic text condensation of Malterud [35, 36]. This process included four steps: naïve reading, identifying and sorting meaning units, condensation, and synthesising [36].

2.6. Stage 4: Workshops

To complement the findings from the first three stages, we held two workshops with stakeholders to directly involve them in the development of the digital intervention. These workshops aimed to create and refine content and features that could support motivation and basic psychological needs throughout trial participation. Due to the COVID‐19 pandemic, the two workshops were planned to be held online via Zoom, 1 week apart, with 10 participants. We invited participants from the ILIT.SS trial, as they had already experienced participating and reporting symptoms. We also invited participants from the ILIT titration trial (ILIT.FT) (EudraCT 2020–005823‐35), as they were enrolled in the ILIT.NU trial for the first year only. At the time of the workshops, these participants had not yet undertaken any of the trial activities themselves; however, they were familiar with the planned trial procedures and requirements (including daily questionnaires). They therefore had an informed understanding of what participation in ILIT.NU would involve and were well placed to provide input on the anticipated feasibility and acceptability of the proposed intervention features. The outlines of the workshops are presented in the Supporting Information. Before the workshops, participants received a box with different items for use during the workshops. This approach has been shown to enrich interpretation, creativity and curiosity [37]. The content of workshop one was designed to explore the concept of motivation in relation to trial retention. Exercises encouraged participants to reflect on their own sources of motivation and consider that others may be motivated by different factors. Through brainstorming and structured input, participants contributed suggestions for elements to be included in the intervention. Drawing on the findings from stages one to three, the researchers highlighted the importance of autonomous motivation. They used these insights to guide the discussions and pose follow‐up questions to deepen participants' reflections. There was a check‐in exercise and three additional exercises. The check‐in exercise and the first two additional exercises were about motivation to participate in the ILIT.NU trial while the last exercise was a discussion on how to support the participants' basic psychological needs through the intervention. These needs were operationalised into questions, and for each question, the participants were given examples based on findings from the interview study in stage three. To facilitate some of the exercises in smaller groups, we used the option to create ‘breakout rooms’ in Zoom. Collaborative learning informed both workshops [38], with representatives from the groups presenting their key findings followed by a discussion in plenum to reach a common agreement. The content of workshop two was similarly based on outputs from the first three stages and the common agreements from workshop one. We discussed the design and features of the intervention using mock‐ups in four different exercises. Both workshops were recorded and transcribed. Findings and agreements were discussed further in the research group, after which mock‐ups of the intervention were created.

2.7. Stage 5: Prototyping

A software developer created a prototype intervention based on the mock‐ups. A central focus in the development process was to ensure that the app design itself was need‐supportive. Insights from the interviews consistently highlighted the importance of making the app easy and intuitive to use, with minimal time required for participants to provide their data to the research project. These considerations informed the design decisions and motivated the inclusion of professional designers in the development of the app's features. Specific design principles included the choice of font to enhance readability, the selection of colours that would be accessible to colour‐blind users, and the avoidance of colour schemes that could bias participants' responses. Participants from the workshops, clinicians, and researchers from the ILIT.NU trial tested the prototype. Involving participants in this process was essential to ensure that all elements of the app were developed to be as need‐supportive as possible. All received a list with attention points regarding usability and positive and negative feedback on the features. Errors, inappropriate elements, proposals for changes, and feedback were recorded in writing. Amendments were then made so the intervention was ready for a randomised feasibility test within the ILIT.NU trial during the grass pollen season of 2022.

2.8. Stage 6: Programme Theory

We developed a programme theory that describes the mechanisms that lead to the intervention's outcome, and the contextual factors that can impact on it [22]. This drew on findings from stages one to five of this development process and was iteratively refined as new knowledge was acquired. The new MRC framework was published during this time [19], so the programme theory was also informed by this updated version. The programme theory was developed by the first author but discussed after each iteration in the research group.

2.9. Patient and Public Involvement

Participants, clinicians and researchers—including designers—were all involved in the development phase to empower the intervention. Patient and public involvement in the research process can be defined as “research being carried out ‘with’ or ‘by’ members of the public rather than ‘to’, ‘about’ or ‘for’ them” [39]. Participants were involved in the research process as: (1) it had the potential to impact on the research agenda and design [40], (2) many of the reported reasons for attrition in a clinical trial would potentially be reduced by involving patients in the research process [41], and (3) involvement is important when developing need‐supportive technology [42]. In this development study, participants from the ILIT.SS trial and from the ILIT.FT trial were involved in stages three, four and five. During the interviews in stage three, participants acted as informants and only provided information. In the workshops in stage four, participants entered the discussions on an equal footing, but the research team made the final decisions based on the discussions. During the prototyping in stage five, each participant was in dialogue with and provided feedback to the research team on an individual level. The participants provided important knowledge and insights on living with grass pollen allergy and how to support their basic psychological needs through the intervention. Their involvement was central to the final design and features of the intervention. The participants all had high trial adherence, which is one reason why the development might have benefited from involving a group with a broader composition. Participants did not receive any monetary compensation for their involvement.

3. Results

The development process resulted in a web application (web app). In stage three, one informant suggested that an app would make it easier to answer the daily questionnaire. We approved the suggestion after discussions in the research group because health apps informed by SDT have the potential to support health outcomes [42, 43], and because an app would make it possible to test and evaluate the intervention on a large scale in the trial. We decided on a web app—rather than a native app that must be downloaded from mobile app stores—to make it possible to send reminders through a short message service to follow the protocol, with daily links to the questionnaires. The results from each stage are presented below.

3.1. Stage 1: Evidence Review

As stated in the method section, no retention interventions informed by SDT were identified during the evidence review, but all effective retention interventions include elements of behaviour change techniques according to Duncan and colleagues [44]. This means that all interventions to some extent possibly support the basic psychological needs. Therefore, all types of retention interventions intended for randomised trials were considered for review question two. A Cochrane Review of interventions to increase motivation for participation and retention in randomised clinical trials was identified [2]. Because a high‐quality review existed and was about to be updated [4] and due to technological development, the first author performed a systematic literature search covering papers newer than those included in the reviews. In total, 751 records were screened by title and abstract, of which 705 were excluded. Eligibility was assessed for 46 articles, and 14 studies were included for review question two. Results from review question one are presented in the introduction to this paper, where four papers were included. Only one potential article was identified for review question three. The study suggested that the use of an intervention informed by motivational interviewing helped improve retention and adherence in a clinical trial among participants with asthma [45]. The intervention had been used before informed consent was obtained and was excluded because the participants in the ILIT.NU trial had already been enrolled. Table 1 presents results for review question two, where numerous retention interventions were found to have been tested in many different health care settings.

TABLE 1.

Included studies. Text in quotation marks are direct quotations.

Title, authors, year, study design and population	Intervention and conclusion	Comments
Strategies to improve retention in randomised trials [2] ○Brueton et al. ○2013 ○Cochrane Review ○All health care settings	A small amount of money could in some settings improve returns of completed electronic questionnaires, but evidence on most strategies is still unclear. None of the included trials had high GRADE certainty of evidence.	The review included randomised controlled trials (retention trials) of strategies to improve retention in randomised controlled trials (host trials). Some of the host trials had more than one retention trial embedded. In total, 26 retention trials used questionnaires returned by post. Four trials used test kits, four trials used telephone interviews, four trials used electronic questionnaires and four trials analysed attendance at follow‐up visits, which is why evidence relevant to this development study is sparse. This review informed the interview study in stage three, where the informants were asked about incentives.
Strategies to improve retention in randomised trials [4] ○Gillies et al. ○2021 ○Cochrane Review ○All health care settings	None of the included trials had high GRADE certainty of evidence. The authors were not able to conclude whether it was the retention strategies or other factors that had an impact on retention. They encouraged further research regarding trial retention. The following are the suggestions relevant to electronic questionnaires and follow‐up visits: ○“Monetary incentive compared to no incentive” ○“Electronic reminders compared to usual follow‐up” ○“Newsletter compared to no newsletter” ○“Addition of monetary rewards to both trial arms”	The review included randomised controlled trials (retention trials) of strategies to improve retention in randomised controlled trials (host trials). Some of the host trials had more than one retention trial embedded. In total, 70 retention trials involved the return of questionnaires, mostly by post. Only two were about attendance at follow‐up visits. Which is why evidence relevant to this development study still is sparse. This updated Cochrane Review informed the last part of the development process. The authors use the words ‘strategy’ and ‘intervention’ simultaneously without further definition.
Non‐randomised evaluations of strategies to increase participant retention in randomised controlled trials: a systematic review [46] ○Elfeky et al. ○2020 ○Systematic review ○All health care settings	Six strategies increased the response rate by more than 10%. The following four are the strategies relevant to an electronic intervention: ○“Shortened version of questionnaire compared to longer questionnaire” ○“Electronically transferred monetary incentive compared to cash incentive” ○“Cash compared to no incentive” ○“Reminders to non‐responders (telephone or text messaging)”	A total of 14 studies were included. The four strategies were each evaluated in one non‐randomised trial. It affects the methodological quality, but the strategies may be worth evaluating in future randomised trials. Shortened questionnaires and reminders were incorporated in workshop two in stage four.
Designing and using incentives to support recruitment and retention in clinical trials: a scoping review and a checklist for design [47] ○Parkinson et al. ○2019 ○Scoping review ○All health care settings	The authors identified eight issues to consider, if trialists want to use incentives to increase retention: ○“What are the current incentives and barriers operating in the system?” ○“Who should incentives be directed towards?” ○“What should be incentivised?” ○“What form of incentive should be offered?” ○“How large should the incentive be?” ○“How should the incentive be structured?” ○“When, and how often, should payments be made?” ○“What are the potential unintended consequences?”	They argue that the complexity of incentives is overlooked, which is an important consideration. In our development study, informants from the interview study in stage three were asked about incentives.
Interventions to improve recruitment and retention in clinical trials: a survey and workshop to assess current practice and future priorities [48] ○Bower et al. ○2014 ○Survey and workshop ○All health care settings	Trial staff from the Clinical Trial Units in the United Kingdom were invited to a survey about recruitment and retention strategies. The results from the survey were discussed in a workshop to generate strategies relevant for evaluation in randomised trials. The following strategies were relevant to an electronic intervention: ○“Methods of communication with participants” ○“Incentives”	In our development study, informants from the interview study in stage three were asked about incentives. Methods of communication were incorporated in workshop two in stage four.
Exploring non‐retention in clinical trials: a meta‐ethnographic synthesis of studies reporting participant reasons for drop out [41] ○Skea et al. ○2019 ○Meta‐etnography synthesis ○All health care settings	A total of 11 studies reporting qualitative data of participants' reasons for attrition were included and synthesised. The following reasons were identified: ○“Perceptions of current health state in relation to specific aspects of the trial” ○“Being too well to engage further with trial processes” ○“Lack of compatibility with personal sense of self” ○“Being ‘too ill’ to be able to engage appropriately with trial interventions” ○“The ‘fit’ of aspects of the trial with individual preferences for care and support” ○“The compatibility of aspects of trial processes with individual capabilities” ○“Concerns about or experiences of the trial medication” ○“Considerations around the extent to which trial participation could be appropriately accommodated into their broader lives”	The reasons for attrition were relevant to develop an intervention capable of accommodating them.
Digital tools for the recruitment and retention of participants in randomised controlled trials: a systematic map [49] ○Frampton et al. ○2020 ○Systematic mapping ○All health care settings	A systematic map was conducted following criteria for conducting a systematic review, but the systematic map provided a descriptive conclusion rather than an estimate of effects. The following overall approaches were identified: ○“Prompts or reminders to attend or complete tasks” (17/26 studies) ○“Digital collection of data to improve the completeness” (5/26 studies) ○“Communication approaches to maintain participants engagement” (2/26 studies)	A total of 20 retention studies were included and 26 approaches were evaluated in mainly observational studies. The methodological quality is affected by the absence of randomised trials, but the approaches may be worth evaluating in future randomised trials. The overall approaches were all incorporated in workshop two in stage four.
Using digital tools in the recruitment and retention in randomised controlled trials: survey of UK Clinical Trial Units and a qualitative study [50] ○ Blatch‐Jones et al. ○ 2020 ○ Survey and interviews ○ All health care settings	Trial staff from Clinical Trial Unites in the United Kingdom were surveyed and 16 of them were interviewed about currently used digital tools to increase recruitment and retention. The following were the retention tools: ○“Online patient information/video, etc.” ○“Trial website” ○“Web chat with trial nurse” ○“App (software) providing tailored information for patient”	The strategies were consistent with the web app intervention and were incorporated in workshop two in stage four.
Best practice guidance for the use of strategies to improve retention in randomised trials developed from two consensus workshops [51] ○Brueton et al. ○2017 ○Consensus workshops ○All health care settings	Professional stakeholders associated with randomised trials participated in two consensus development workshops. The following strategy was relevant to a digital intervention: ○ “Alternative ways to complete outcome data (e.g., by post, text, or e‐mail)” Strategies for further research relevant to a digital intervention was: ○ “Electronic follow‐up technologies”	The strategies were consistent with the web app intervention.
Identifying research priorities for effective retention strategies in clinical trials [52] ○Kearney et al. ○2017 ○Delphi Survey ○All health care settings	The priority of retention strategies was identified through a Delphi Survey. The following strategies were identified as being critically important in future retention interventions: ○“Site training on missing data” ○“Use of routinely collected data” ○“Frequency of patient contact during the trial” ○“Frequency and timing of reminders” ○“Length/time needed to complete the questionnaire”	It was not possible to change the ILIT.NU trial concerning the first two strategies. The last three were incorporated in workshop two in stage four.
Assessing effective interventions to improve trial retention: do they contain behaviour change techniques? [44] ○Duncan et al. ○2020 ○Review ○All health care settings	Strategies with some evidence of an effect on retention in randomised trials were investigated for behaviour change techniques using the behaviour change taxonomy. The analysis showed that all strategies included elements of behaviour change techniques, but were not informed by a theoretical rationale. The most common were prompts and instructions on how to perform a behaviour change. They concluded that triallists should be mindful of behaviour change techniques when designing interventions to increase retention in trials.	Their findings highlight the importance of focusing on what aspects of an intervention thwart or support psychological needs.
Patient‐centered recruitment and retention for a randomised controlled study [53] ○Chhatre et al. ○2018 ○Randomised controlled trial ○Prostate cancer	The authors developed a conceptual model with factors influencing recruitment and retention and evaluated them in a randomised trial. A stakeholders' advisory board including patient partners were described as important for the outcome. The following strategies facilitated successful retention and were relevant to an electronic intervention: ○ “Trust (ensuring protection of information)” ○ “Communication (calling/emailing assessment)” ○ “Attitude (emphasising the altruistic value of research, positive attitude of providers and research staff)” ○ “Expectations (full disclosure of study requirements and time commitment, update letters)” “Active engagement, reminders at the offices, and personalised update letters” increased retention during follow‐up.	The authors evaluated a lot of strategies. It is difficult to assess whether they had an individual effect on retention or it was an interaction. The strategies were incorporated in workshop two in stage four.
Development of the IRIS‐AR strategy: an intervention to improve rates of accrual and retention for the VTE‐PRO randomised controlled trial [54] ○Fahim et al. ○2019 ○Survey and interviews ○Venous thromboembolism	Barriers to participation in a trial were identified in a survey and analysed using Theoretical Domains Framework and Behaviour Change Wheel. Interviews refined and confirmed the findings. The following elements of the strategy were relevant to an electronic intervention: ○“Information booklets and counselling” ○“Daily reminders to administer injections”	The strategy has not been evaluated, but the two elements were incorporated in workshop two in stage four.
Qualitative study of barriers to clinical trial retention in adults with recently diagnosed type 1 diabetes [55] ○Henshall et al. ○2018 ○Qualitative study ○Type 1 diabetes	A total of 20 interviews about barriers and facilitators to retention with patients newly diagnosed with type 1 diabetes. Following barriers and facilitators were identified: ○“Lack of time” ○“Work pressures” ○“Level of health professional support” ○“Clarity and consistency of information and feedback” ○“A desire for knowledge about their condition”	The population was comparable to the population in the ILIT.NU trial. The barriers and facilitators were relevant to develop an intervention capable accommodate them.

Overall, Stage 1 demonstrated that existing retention interventions, including digital approaches, have limited and largely atheoretical evidence, underscoring the need for a theory‐informed SDT‐based intervention for the ILIT.NU trial.

3.2. Stage 2: Identification of Theory

During the evidence review, we did not identify any need‐supportive apps that had been developed to increase retention in trials or intended for people with grass pollen allergy, but Peters and colleagues developed—in collaboration with young people with asthma—a self‐management app to support the three basic psychological needs [42]. Around 60% of people with asthma are affected by allergy [56]. Also, participants in the ILIT.NU trial were in general young, with a median age of 29 years. Hence, the development process was inspired by findings from Peters and her colleagues. Features proposed by the young people as part of the study were: “profile, reminders, tracking, social, emergency support, expert access and information, feedback and motivational features” [42]. Peters and her research team also listed supportive design features. Relatedness should be supported by making “salient user belongingness to a larger community while maintaining privacy and safety”, while autonomy and competence should be supported by providing “tools that will allow users to feel more in control and better able to handle asthma impacts and emergencies; support autonomous motivation based on personal values and goals” [42]. Thus, Stage 2 established SDT as the guiding theory and showed that an SDT‐informed app model from a related population could be adapted to inform feature design for the ILIT.NU trial.

3.3. Stage 3: Interviews

Of 16 patients invited, a total of 10 agreed to participate in the interview study. The reasons for rejecting the invitation were lack of time and it being a long time since the patient had participated in the ILIT.SS trial, making them think their contribution would be irrelevant. The relevant characteristics of the participants are presented in Table 2.

TABLE 2.

Participant characteristics.

Characteristics	Interviews, n = 10	Workshops, n = 10	Prototyping, n = 5
	No. (%)	No. (%)	No. (%)
Gender
Female	5 (50)	5 (50)	2 (40)
Age in years
21–30	2 (20)	4 (40)	3 (60)
31–40	3 (30)	5 (50)	1 (20)
41–55	5 (50)	1 (10)	1 (20)
Interview duration in minutes
11–20	1 (10)	—	—
21–30	3 (30)	—	—
31–40	2 (20)	—	—
41–50	4 (40)	—	—
Treatment
Verum	6 (60)	10 (100)	5 (100)

Five themes were derived from the analysis and are presented in Table 3. Overall, the informants participated in the hope of receiving the vaccine to relieve their symptoms, as they felt limited in their daily lives. However, they were also driven by a strong desire to contribute to research for the benefit of other people with grass pollen allergy. The informants thought it was a well‐conducted trial where they felt well informed and experienced a good relationship with the researchers and clinicians. Flexibility was mentioned as being very important in facilitating their active participation.

TABLE 3.

Findings from systematic text condensation.

Theme	Content and quotations
1. Trying to avoid symptoms limiting everyday life	○ Symptoms regarded as debilitating, especially tiredness. ○ Organising life to avoid symptoms. ○ Limitations on daily activities. “When I think of the grass pollen season, I think about tiredness. That kind of exhaustion, which is just really debilitating, and it's especially annoying because it's summer, and you've been looking forward to bright nights to be able to take full advantage of the good weather. But then at half past nine in the evening, I can't do anything anymore, I'm simply completely done. That's probably what bothers me the most (…). Of course, I go to work and take care of my job, but there isn't really room for anything else. So, cycling to the beach and having a swim at eight o'clock in the evening in the lovely summer weather, I just don't do that, even though it's what I want to do the most” (Informant 2). “Many will probably look forward to summer and the freedom that comes with it, but I am more the type who stays indoors with the windows closed as much as possible.” (Informant 9).
2. From personal benefits to altruistic reasons	○Enrolment in the hope of getting the vaccine to relieve symptoms. ○Committing to contribute to research for the benefit of others with grass pollen allergy, out of sympathy for researchers and to be dutiful.
	“Getting the vaccine was the greatest motivation for me.” (Informant 3) “The motivation was perhaps a little less because you no longer have the vaccine to look forward to. But during the consultations, I also got some kind of sympathy for the persons who carried out the project. So of course, I also wished them to finish their project in a good way and that their data would be as good as possible.” (Informant 2)
3. Reconciling trial activities with everyday life as a positive predictor	○Flexibility of trial activities regarded as valuable. ○The less time they had to spend, the better they could fit the activities into their daily schedule, and the more likely they were to participate in the activities. ○A disruption to daily routines was mentioned as one of the main reasons for not reporting. Flexibility and several activities in everyday life were mentioned as well.
	“If it was a normal day where I had no plans, where I just followed my daily routines, then I could easily complete the questionnaire before bedtime. It was immediately worse if I had something planned, then I tended to forget it. (…) I actually think it would have been beneficial if the questionnaire was in an app, but that also requires money.” (Informant 1). “If I was not able to manage it Wednesday, then I could show up Friday instead, it was flexible regarding my personal appearance.” (Informant 6).
4. Experiencing equality as valuable	○Professional and equal treatment important. ○No interest in contact with other participants.
	“I think it was nice that you were taken seriously. I have struggled with hay fever since I was a child. (…) When you are sick for so long, and the medication you receive doesn't help you, and you don't know which options are available, but no matter what you have tried, it doesn't work, then it feels nice to meet someone who takes you seriously. (…) I felt heard—does that make sense?” (Informant 3). “I wasn't interested in e.g., some group conversation with other people with grass pollen allergy (…). I didn't need to talk to them, so we could complain together about how annoying it is.” (Informant 2).
5. Being well informed leading to increased knowledge and ability to complete trial activities despite a need for progress updates	○Capable of completing trial activities. ○Increased knowledge. ○A need for progress updates.
	“It was a fine enrolment letter, and I felt well informed, there was nothing I was in doubt about.” (Informant 6). “I got increased knowledge of what to do, and what is best to do, regarding my allergy.” (Informant 2). “I would have liked more information along the way. Just a status like: ‘We are so and so far.’ Or it could just be an email that said: ‘Hey, there's actually nothing new, but we'll see you in 3 months for the next injection.’” (Informant 1).

In summary, Stage 3 clarified how symptom burden, motivations (personal and altruistic), flexibility of trial activities, perceived equality, and information needs shape participants' ability and willingness to stay in the trial, directly informing which contextual factors and basic psychological needs the intervention should target.

3.4. Stage 4: Workshops

Of the five participants from the ILIT.SS trial invited to the workshops, two were able to participate on the scheduled days. This resulted in eight participants from the ILIT.FT trial being invited, all of whom agreed to participate. All those invited participated in workshop one, but due to illness and forgetfulness, only five participated in workshop two. Two of the absent participants agreed to take part in an individual online conversation where they approved the findings from workshop two. Each workshop lasted 1 hour. Relevant participant characteristics are presented in Table 2. Results from the first three exercises in workshop one about motivation to participate were in agreement with the findings from the analysis of the interviews and are not presented. Additional results are presented in Table 4 in relation to the specific psychological needs. It was important to the participants that the app collected all the information they normally access through other apps and webpages (e.g., pollen counts and advice about grass pollen allergy) to save time. In general, they needed to spend as little time as possible on using the app and its features. They suggested minimising scrolling and clicking, text rather than videos and avoiding games.

TABLE 4.

Findings from workshops one and two.

Psychological need	Features
Autonomy	Flexibility to fit the trial activities into own schedule—this includes the ability to: ○ Answer the questionnaire within a longer time slot ○ Answer the questionnaire a few days later ○ Only answer the questionnaire on the most important days over the period ○ Receive a reminder and decide the time yourself ○ Record other allergy symptoms (e.g., tiredness) ○ Make it optional to use the different features “The most important is flexibility, that you don't have to respond within a short time. (…) It is just important that you can choose it yourself.” (Participant 10)
Competence	○ The app must be designed with usability in mind to make it easy to use ○ Visualise the data entered and compare it with other participants' data ○ Visualise pollen counts and forecasts ○ Advice about grass pollen allergy ○ Updates on the ILIT.NU trial, to establish a better understanding of the trial, the results and why active participation is important for the outcome of the trial “Very good with pollen count, since I check them elsewhere anyway. It would save me time and clicks to have it all together in one place.” (Participant 8)
Relatedness	Create a sense of belonging without knowing each other: ○ Positive feedback when responding to the questionnaire ○ Updates on the ILIT.NU trial, to be reminded that you are a part of something bigger and that your participation is valued ○ Compare your data with other participants' data ○ Form to contact the clinical staff ○ Visualise own and overall response rate “Just being able to see that there are others doing this [responding to the questionnaire], I think that would be nice.” (Participant 9)

Participants were interested in comparing their symptom and medicine data with other participants' data. However, this suggestion was rejected as it could have affected the scientific integrity of the ILIT.NU trial. Knowing that their symptoms were low/high compared to others could have given a hint about whether they were being treated with verum or placebo. It was made possible to receive the daily reminder after 6 pm, to make sure participants had been exposed to the total amount of pollen for the day. Overall, Stage 4 translated theoretical and interview findings into concrete, user‐informed features that support autonomy, competence and relatedness, while also highlighting practical constraints (e.g., trial integrity) that needed to be balanced in the final intervention design.

3.5. Stage 5: Prototyping

The software developer created the prototype intervention. The prototype was tested for 1 week by clinicians, researchers and five of the participants from the workshops. The relevant characteristics of the participants are presented in Table 2. Additionally, it was tested for further errors in five iterations by the first author. Table 5 presents the features of the final intervention and how each feature supported the basic psychological needs. Informed by SDT and Peters and colleagues [42], features supporting autonomy were intended to provide choice or facilitate control. Features supporting competence were intended to increase knowledge and ability to act. The usability of the web app was also intended to support competence. Features supporting relatedness were intended to create a sense of relatedness or the possibility of interacting with the clinicians. Figure 2 is a mock‐up of the menu of the web app, while Figure 3 is a mock‐up of the features within this menu. Only mock‐ups are presented to display the content in English, as the web application was in Danish. Two amendments were made in addition to correcting various errors. Firstly, one of the participants suggested that completing the contact form should generate a message saying that a clinician would respond as soon as possible. Secondly, it was suggested that it would be more meaningful if the graph showed the individual symptom score and the individual medication score as a function of time, instead of showing the individual compound symptom and medication score as a function of time. All participants in the ILIT.NU trial were given the ability to record other symptoms (e.g., tiredness), and these are not listed below. Data entered were stored on the secure web‐based platform REDCap [18], and a password was required to access available data in the web app.

TABLE 5.

Features of the web app and the corresponding needs they were intended to support.

Feature	Content	Psychological need supported
Link received through short message service	When clicking the link: ○The questionnaire was available ○The link was active for 48 h	Autonomy
Questionnaires to be answered: Compound symptom and medication score (CSMS) and Rhinoconjunctivitis Quality of Life Questionnaire	The current day's questionnaire(s) were available from 6 pm Questionnaires not answered were available for 48 h When they clicked on the CSMS questionnaire, participants only had to answer all questions on days with symptoms. The first question was: “Your use of medication in the last 24 h—did you use symptom‐relieving medication?” – Yes: Directed to the CSMS questionnaire – No: Directed to next question: “Your allergy symptoms in the last 24 h—did you have allergy symptoms?” – Yes: Directed to the CSMS questionnaire – No: Directed to the answer button When they had completed the questionnaire, participants were directed to the front page of the web app where each feature could optionally be accessed or the web app could be closed. A message was generated thanking participants and informing them that each response was important for the trial	Autonomy Competence Relatedness
Pollen count and forecast	○Choose between pollen counts for grass, birch and mugwort ○Choose between counts from the east and west of Denmark ○Current pollen count and indication of level – low/medium/high ○Next day's pollen forecast ○Previous day's pollen count	Autonomy Competence
Response rate	Individual response rate and total ILIT.NU response rate	Relatedness
Graph of own data	A graph with both the individual symptom score and the individual medication score as a function of time. Instructions on how to understand the graph were also provided	Autonomy Competence
Advice about grass pollen allergy	During the grass pollen season: ○Use (sun)glasses outside ○Close doors and windows during the daytime ○Doors and windows can be opened once the dew has appeared ○Rinse hair and face before going to bed ○Minimise outdoor activities around noon ○Do not dry laundry outside	Autonomy Competence
Updates on the ILIT.NU trial	Four updates during the grass pollen season were planned. They should contain information about trial progression, trial activities, recognition for effort in the trial and why active participation is important for the trial outcome	Competence Relatedness
Change time of reminder	Choose a time between 6.00 pm and 11.59 pm	Autonomy
Download the icon to the home screen	A guide on how to download an app icon to the home screen for both Android and Apple, allowing the web app to be opened at any time during the day	Autonomy
Contact	A form to contact the trial team with trial‐related questions. It was stated that questions regarding health conditions or any acute need for help should go to a general practitioner or the emergency department. A message was generated to inform a user who completed the form that a clinician would respond as soon as possible	Relatedness
Login/Create user	A code was required to access the features to protect sensitive personal information about the user's health condition	—

FIGURE 2.

Mock‐up of the menu of the web app.

FIGURE 3.

Mock‐up of the features in the web app.

Thus, Stage 5 demonstrated that the SDT‐informed, user‐derived requirements could be operationalised in a usable and secure web app prototype.

3.6. Stage 6: Programme Theory

Aspects of the programme theory are presented in a logic model in Figure 4. The visual design of the model is adapted from Mills and colleagues [57] and concepts from the MRC framework [19] and outlines hypothesis, substudies, mechanisms of change, outcome measures, impact and identified contextual factors that could possibly influence the outcome.

FIGURE 4.

Logic model of the need‐supportive intervention—adapted from Mills and colleagues [57].

Thereby, Stage 6 defined a programme theory that linked the web app's features to mechanisms of change and outcomes, guiding subsequent evaluation in the ILIT.NU trial.

4. Discussion

We developed a web app intended to increase retention in a trial by supporting the three basic psychological needs of autonomy, competence and relatedness as described in SDT. Features intended to support autonomy were a shortcut for reporting days without medication and/or without symptoms, the ability to answer 48 h later, changing the time of the reminder and downloading an icon to the home screen. Features intended to support competence were pollen counts and forecasts, visualisation of data entered, advice about grass pollen allergy and updates on the status of the ILIT.NU trial. Features intended to support relatedness were positive feedback when responding to a questionnaire, personal and total response rate and contact information. Some features were intended to support more than one need (Table 5).

4.1. A Digital Intervention That Meets the Principles of SDT

Peters and colleagues have developed the ‘Motivation, Engagement and Thriving in User Experience’ (METUX) model, which provides a framework for designing need‐supportive technology based on five different spheres of user experience [58]. Applying METUX helps avoid oversimplification when operationalising the research [59] and guards against designs that are need‐supportive at one sphere, yet need‐thwarting in another [58].

In the METUX model, the first sphere, Adoption, focuses on fostering users' autonomous motivation to begin using technology [58]. The web app was aligned with this first sphere because use of the web app and its features was entirely voluntary, minimizing external pressure and supporting autonomous rather than controlled motivation to engage with it.

The second sphere is Interface, where good usability will support autonomy and competence [58]. During development, we focused on usability to facilitate anticipated competence in using the web app. We did this by involving both designers and participants. However, the Interface aspects could have benefited from also involving user experience designers. The description of the Interface sphere also suggests that a lack of social features can actually contribute to better relatedness [58]. Peters and colleagues argue that this is remarkable when considering SDT [58], but it is in agreement with the results from stages three and four of this development study, where no one was interested in a relationship with other participants.

Task is the third sphere. Task features can support or thwart psychological needs [58]. We hypothesise that the features of the web app are generally need‐supportive as their development was informed by evidence and completed in collaboration with participants. However, the feature comparing response rates, which is intended to support relatedness, also contains an element of competition. According to SDT, competition can thwart autonomy when there is high pressure to win and can thwart competence when the primary focus is on avoiding loss [60]. It is possible that some participants, by using this feature, will either feel pressure to do better reporting than others or avoid being worse than others, resulting in a negative effect on their reporting. On the other hand, competition can support competence if it is informational [60]. It can be argued that the feature is informational, as participants receive information on the general reporting within the trial. Similarly, Peters and colleagues added a competition feature to their app to increase relatedness [42].

Behaviour is the fourth sphere [58]. Even if a technology is need‐supportive within the first three spheres, it is not a given that it will impact on behaviour in this fourth sphere [58]. In this development study, the web app has been alpha tested as part of the prototyping. To evaluate whether participants find the web app and its features need‐supportive, we will conduct a randomised feasibility test within the ILIT.NU trial following the MRC framework. This beta testing is important, as many components need to interact to create a satisfying user experience for the end users. Another aspect is whether the collection of research data is acceptable to the research team. It is important to resolve these key uncertainties to know whether to proceed to an evaluation [19]. If the web app is found to be feasible and acceptable, we will evaluate whether it will impact on behaviour and increase retention by testing it in a randomised study within the ILIT.NU trial.

The fifth sphere is Life, where some technologies can impact overall well‐being in life [58]. We do not expect the web app to have this kind of impact. Nevertheless, it is possible that using the need‐supportive features will increase users' knowledge and ability to live with grass pollen allergy, which can lead to increased well‐being according to SDT [9].

To sum up, the web app generally fits within the spheres related to user experience in the METUX model. This represents a strength of the study, as the design and its features appear to be genuinely need‐supportive, fostering autonomy, competence, and relatedness in line with the principles of SDT. However, SDT has been criticised for whether autonomy is truly a basic psychological need. Critics say “autonomy” implies complete independence and freedom from societal norms, but SDT's meaning differs [61]. Autonomy involves acting based on internalised cultural or external demands, aligned with your integrated self, and choosing to continue these actions upon reflection [61]. Hence, understanding this distinction is essential when applying SDT.

5. Limitations

This development study also has a number of limitations. Even though the search strategy was developed in collaboration with a librarian, all relevant publications may not have been identified. Only the first author was involved in the systematic search and the APA PsycInfo database was not searched. It is possible that additional interventions intended to support behaviour change in relation to trial retention could have been indexed there. Furthermore, it would have been relevant to conduct a systematic review analysing retention interventions and whether elements of these interventions respectively supported autonomous or controlled types of motivation. In 2023, Gamble and colleagues published a scoping review with a similar focus on technology to support clinical trial retention [62]. They concluded that most retention interventions rely on controlled types of motivation. If interventions supported any psychological need, this was mainly relatedness [62]. The online interviews and workshops also have some limitations. Computer‐mediated communication has been criticised for lacking social environment cues (e.g., nonverbal communication), affecting the quality of the data [63]. Interaction between the participants is especially difficult to facilitate remotely, resulting in less sharing and comparing of experiences, and difficulties in knowing who is next to speak [63]. This was also our experience during the workshops, but we believe that the final version of the web app reflects the needs of people with grass pollen allergy, as patient partners were involved in the prototyping. Nonetheless, in‐person interviews and workshops might have provided better conditions for interaction and group dynamics, as participants could respond more naturally to one another. Such an approach, however, requires participants to dedicate additional time and effort, but future studies could benefit from adopting in‐person formats where feasible. Finally, this study was conducted with participants aged 20–55 years living with grass pollen allergy. While this group represents individuals who are generally healthy but experience recurring seasonal symptoms, their perspectives may differ from those of people with more complex or chronic health conditions. Likewise, variations in age, life circumstances, and health management strategies within and beyond this age range could influence how the intervention is perceived and used. These factors place certain limits on how broadly the findings can be interpreted, both among individuals with grass pollen allergy and in relation to transferring the results to other populations. However, because the three basic psychological needs are considered universal across developmental periods and cultures, the underlying approach may still be suitable for adaptation to other contexts [64].

6. Implications and Perspectives

This development study provides scientific knowledge on how to increase retention in trials informed by behaviour change science, which is in demand [8]. There is little evidence of the effect of current interventions to increase retention in trials [4], and most are not informed by behaviour change science even though their intention is to change human behaviour [8]. Therefore, future studies where optimisation of retention is of importance will possibly benefit from drawing on our results, especially trials collecting remote research data at multiple time points. Some of the findings specifically relate to supporting the needs of people with grass pollen allergy, but the three basic psychological needs are universal across cultures and developmental periods [64]. To address these, features like the ability to answer 48 h later, changing the time of the reminder, updates on the status of the trial, positive feedback when responding to a questionnaire and contact information will be easy to implement in future studies across all health disciplines and without large financial costs. This study focused solely on need‐supportive technology, but future trials may consider drawing on SDT in communications with the participants (e.g., through need‐supportive patient information and need‐supportive consultations). Focusing on supporting participants' psychological needs according to SDT may have the practical implication that participants will be more likely to continue with their behaviour, which will provide better outcome data and thereby improve the validity and reliability of the results.

The intervention is first being tested in a feasibility study and will subsequently be evaluated in a randomised study within the ILIT.NU trial, with the specific aim of increasing trial retention. This represents an important step towards generating robust evidence, which may in turn allow other studies to benefit from the results. The future perspectives outlined should be considered with caution, as they are based on a selected group of participants and subject to the study's limitations. However, since the three basic psychological needs proposed by SDT are regarded as universal, we would expect the approach to be relevant across other patient groups as well, while remaining mindful of the specific context on which our recommendations are grounded. Therefore, while the findings provide valuable insight into this particular population, further research including more diverse age groups and diagnostic categories is needed to determine the broader relevance of the approach.

7. Conclusion

A web app intended to increase retention in the ILIT.NU trial by supporting the three basic psychological needs of autonomy, competence, and relatedness was developed in six stages in collaboration with participants, clinicians, and researchers. Taken together, the six stages moved systematically from identifying evidence and theoretical foundations, through eliciting and integrating stakeholder perspectives, to iterative prototyping and refinement. Each stage contributed distinct insights, which were translated into concrete features aligned with the principles of SDT. In combination, these stages yielded a need‐supportive web app prototype. The final prototype of the web app was ready for a randomised feasibility test within the ILIT.NU trial during the grass pollen season of 2022. This feasibility study represents an important first step in evaluating the prototype, and future research will be needed to further assess its effect and potential contributions in broader contexts.

Author Contributions

Study design: first author, second author, fourth author, and fifth author. Data collection: first and last author. Data analysis and interpretation of results: all authors. Draft manuscript preparation: first author. Critical review of the article and agreement to this final submitted version: all authors.

Funding

This work was supported by Innovation Fund Denmark, 9090‐00065B. Aarhus University Hospital and Aarhus University.

Ethics Statement

This development study complied with the Helsinki Declaration and the Guideline for Good Clinical Practice (ICH‐GCP) [65, 66]. Among others, it obtained ethical approval from the Danish National Center for Ethics as part of the ILIT.NU trial (1–10–72‐315‐20) and is registered in the Central Denmark Region's internal records of research projects. All participants received oral and written information, gave their written consent, and were informed about their rights. The storage of data complied with the General Data Protection Regulation and the Danish Data Protection Act [67].

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: scs70208‐sup‐0001‐AppendixS1.docx.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.