Evaluating the effects of two parental self-management eHealth applications on antibiotic use, number of primary care visits and gut microbiota of febrile Dutch children with acute, uncomplicated upper respiratory tract infections: a study protocol for a randomised controlled trial Authors: Eefje Belt-van Zoen, Noortje van Steenbergen, Liesbeth Veldman, Max Nieuwdorp, Erik W Baars Published: 2025 ABSTRACT AbstractIntroductionPhase-specific fever management (ie, tailoring care to the rising, plateau and defevering stages of fever) and the cautious use of evidence-based natural medicinal products (NMPs, eg, herbal preparations with demonstrated evidence of safety and effectiveness) are hypothesised to reduce antibiotic (AB) use for acute, uncomplicated upper respiratory tract infections (URTIs). The objective of the trial is to investigate the effects of a parental phase-specific FeverApp and an NMP WebApp on AB use, the number of primary care visits and the gut microbiota of children with acute, uncomplicated URTIs.Methods and analysesThe study design is a randomised controlled trial with three arms: a FeverApp group, an NMP WebApp group and a control group. Participants are parents of children aged 3 months to 12 years with fever (≥38°C) and symptoms of an acute, uncomplicated URTI. Parents (n=320) will complete a baseline questionnaire and collect a baseline stool sample from their child. The intervention groups then use one of the apps, and all parents report URTI symptoms of their child on days 2, 5 and 7. On day 10, they complete a follow-up questionnaire, and they collect a second stool sample from their child. The FeverApp contains evidence-based information about fever and how to manage it, tailored fever advice and an option to monitor a child’s health status during a fever episode. The NMP WebApp is an evidence-based online decision-making tool that helps parents navigate the use of NMPs with reported safety and effectiveness for URTI symptoms. The primary endpoint is AB use, and secondary endpoints include recovery time, complications/adverse events, the number of primary care visits/telephone consultations and changes in gut microbiota composition. Participants were recruited nationwide (urban and rural) via online advertising and community flyers.Ethics and disseminationA waiver was obtained from a recognised Institutional Review Board, participants signed for informed consent, and all data are pseudonymised and stored securely. Results will be published in scientific, peer-reviewed journals.Trial registration numberNCT06140446 (www.clinicaltrials.gov). Strengths and limitations of this study The study uses a three-arm randomised controlled design, allowing robust comparison between two digital interventions (FeverApp and the natural medicinal product WebApp) and a control group. The study uses a three-arm randomised controlled design, allowing robust comparison between two digital interventions (FeverApp and the natural medicinal product WebApp) and a control group. The inclusion of both subjective (parent-reported symptoms and antibiotic use) and objective (gut microbiota composition) outcome measures strengthens internal validity and enables triangulation of findings. The inclusion of both subjective (parent-reported symptoms and antibiotic use) and objective (gut microbiota composition) outcome measures strengthens internal validity and enables triangulation of findings. Standardised data collection at defined time points (baseline, days 2, 5, 7 and 10) minimises recall bias and supports temporal analyses of symptom progression. Standardised data collection at defined time points (baseline, days 2, 5, 7 and 10) minimises recall bias and supports temporal analyses of symptom progression. Nationwide online and community-based recruitment increases external validity but may introduce selection bias toward digitally literate or health-motivated parents. Nationwide online and community-based recruitment increases external validity but may introduce selection bias toward digitally literate or health-motivated parents. As parents report symptoms and collect stool samples at home, variations in adherence and data quality could affect the precision of outcome measurement. As parents report symptoms and collect stool samples at home, variations in adherence and data quality could affect the precision of outcome measurement. Introduction Resistance to antibiotics (AB) increases mortality and healthcare costs.1 2 In Europe, primary care accounts for about 80%–90% of all AB prescriptions, placing general practice at the heart of antimicrobial resistance (AMR) stewardship.3 Upper respiratory tract infections (URTIs), with symptoms like cough, sore throat, earache, nasal congestion, runny nose or headache, are common in children.46 Although international guidelines state that ABs are unnecessary for uncomplicated URTIs,7 they remain overprescribed.810 In 2021, about 17% of patients (children and adults) registered in Dutch general practice, received at least one prescription for an oral AB,11 and inappropriate prescriptions reached 59% of URTI prescriptions in daycare and 76% in out-of-hours general practitioner (GP) practices in 2016.12 Parental concerns often drive unnecessary AB prescriptions. To combat ‘low-value healthcare’, a list with do-not-do recommendations for GPs from 2022 advised on avoiding ABs for uncomplicated URTIs and emphasised the need for alternative treatments.7 Beyond AMR, ABs can cause allergic reactions13 and reduce gut microbiota diversity, leading to, for instance, type 1 and type 2 diabetes in later life.1416 AB exposure is furthermore linked to gastrointestinal, immunological and neurocognitive disorders,14 underlining the urgency to limit inappropriate AB use in children. Integrative medicine (IM) provides alternative treatment options for URTI symptoms when ABs are unnecessary and not useful. IM combines conventional medicine with complementary therapies such as natural medicinal products (NMPs).17 Although not yet included in AMR policies, UK general practitioners (GPs) trained in IM prescribed fewer ABs in 2016.18 IM strategies, alone or combined with delayed-prescription methods, could reduce patient demand for ABs, GP prescription rates and on the longer AMR.18 19 Therefore, encouraging primary care providers and parents to integrate IM strategies, particularly NMPs for URTIs, may be useful. NMPs include herbal substances, tinctures, extracts, oils and processed exudates, processed into pastilles, sprays, drops, syrups or supplements in pharmacies and online. Most are sold without a prescription and fall under ‘self-care’. NMPs can alleviate URTI symptoms and may also boost the immune response.19 In this study, NMPs refer to herbal preparations with European Medicine Agency (EMA) ‘traditional’ or ‘well-established use’ status for URTI symptom relief (eg, Pelargonium sidoides, Hedera helix, Thymus/Primula combinations and Andrographis paniculata). However, concerns about their safety and effectiveness limit their use. Yet, when used responsibly within guidelines, NMP-related adverse events (AEs) are rare.19 Several NMPs reported promising results for relieving URTI symptoms and have EMA classifications as ‘traditional’ or ‘well-established’.20 Their mechanisms include immunomodulatory, antiviral, antibacterial, antipyretic and expectorant effects.21 The UK National Institute for Health and Care Excellence guideline on acute cough recognises Pelargonium (for people aged ≥12 years) as an option that some may wish to try based on limited evidence, while not recommending it for children.22 The German S3 cough guideline (2021) similarly discusses Pelargonium and reports moderate evidence but does not make a firm recommendation.23 Other herbal products such as A. paniculata, ivy leaf, primrose and thyme were reviewed, but no specific recommendations were made.22 23 By contrast, the 2024 Dutch acute cough guideline states that evidence for herbal extracts, including P. sidoides, remains uncertain/insufficient.24 This underscores the need for accessible, reliable, evidence-based information on NMPs.25 Aside from NMPs, phase-specific and supportive fever management may also reduce unnecessary AB use for URTIs. Fever progresses through three phases: (1) rising, (2) plateau and (3) 'defevering’, with management strategies including suppressive (cooling), permissive (acceptance) or active (supporting fever) approaches.26 The immune system fights infections, making ABs unnecessary in most cases, except for complications or vulnerable children.27 Fever may also strengthen the immune system, possibly reducing the risk of future infections.27 28 From an integrative perspective, fever is a natural resource and should only be limited when necessary.28 A systematic review assessed international guidelines on fever management in children.29 Although the evidence was limited, experts across countries agreed that parents should be educated on fever benefits and alarm symptoms.29 Guideline recommendations focus on distress reduction rather than temperature reduction, and antipyretics (eg, acetaminophen) should not be given to prevent febrile seizures but for reducing discomfort.29 External cooling may increase discomfort and metabolic strain and should be avoided when possible.29 However, Dutch parents tend to lower their child’s temperature with antipyretics, fewer blankets, drinking cold fluids or undressing the child, influenced by patient information and fever advice.30 Research suggests that suppressing fever at least negatively impacts survival rates in sepsis patients.28 Supporting fever, rather than suppressing it, could be more effective, with different strategies for each fever phase. For instance, allowing fever to reach 40°C instead of 38.5°C in the rising phase has been linked to reduced mortality in critically ill patients.31 While Dutch GPs generally follow a permissive fever management strategy, they do not typically distinguish fever phases.30 32 33 In contrast, Germany’s newest fever guidelines (due in 2025) separate the distinct phases of fever and advocate for an active approach for uncomplicated infections.29 A 2019 Dutch study found that parental fever education via a booklet reduced primary care visits and AB prescriptions for uncomplicated URTIs.32 Since 2018, the WHO has recognised eHealth’s potential to improve public health by enhancing health behaviours, disease outcomes and self-management.34 Digital tools allow parents to monitor their child’s health and can reduce healthcare costs,34 though eHealth requires a certain level of health literacy.35 Given the AMR challenge and the potential of NMPs and fever education to mitigate it, eHealth tools could support informed decision-making in primary care. A combination of a fever self-management tool and an NMP decision-making tool would benefit patients, parents and healthcare professionals seeking reliable and integrative guidance for URTI treatment. Although the FeverApp is not yet formally endorsed by a Dutch paediatric or infectious diseases society, its Dutch adaptation involved consultation with Dutch physicians, and acceptability studies showed majority support among primary care professionals. Both apps underwent acceptability evaluation among Dutch primary care professionals, where the majority indicated they would use or recommend them (NMP App (n=414): 75%–82%; FeverApp (n=240): 76%–94%). The present randomised controlled trial (RCT) will provide empirical evidence required for potential future endorsement. Therefore, the University of Applied Sciences Leiden (UASL) developed an evidence-based NMP WebApp, drawing on systematic reviews,20 21 European prescription rates,19 expert opinions19 and international URTI guidelines (eg, from Germany and the UK). Additionally, Witten-Herdecke University (WHU) developed a phase-specific FeverApp, implemented in national German paediatric care to reduce healthcare visits and AB use for febrile children through better self-care.36 37 Based on current fever management insights,26 28 29 the app guides parents in safe, active fever management. This app was translated into Dutch by the University of Applied Sciences Leiden and was adapted to national guidelines, healthcare structures and user preferences. Both the FeverApp and NMP WebApp can be used for (1) professionalisation of healthcare providers, (2) self-management by parents and (3) shared decision-making in primary care. Based on international experiences, both apps could complement Dutch primary care practice, self-care by parents, the national patient information website (www.thuisarts.nl)30 and GP fever guidelines.33 In a German trial, the FeverApp reduced AB use by 50% (unpublished data: German Clinical Trials Register (DRKS) with registration number: DRKS00016591), suggesting that a decrease may also be expected in the Netherlands, although healthcare systems differ. Likewise, UK GPs trained in IM prescribed fewer ABs,18 supporting the potential of an NMP WebApp to lower AB prescriptions and AMR in the Netherlands. Both apps may furthermore contribute to healthier gut microbiota through reduced AB use and less unnecessary fever suppression, lowering future metabolic disease risks.1416 Additionally, the FeverApp may enhance parental confidence and strengthen parent–child bonding, promoting the overall well-being of the child.38 According to eHealth development guidelines, app acceptability and usability should be evaluated before testing effectiveness and implementation.39 40 Acceptability of both apps was therefore assessed in 2022 via online surveys among Dutch parents, GPs, nurse practitioners, physician assistants and pharmacists. Preliminary results showed that 79% would use or recommend the FeverApp, and 45%–55% expected reduced AB use and primary care visits by this (see Chapter 5 of this thesis). For the NMP WebApp, 76% indicated willingness to use it, with 49% and 42% expecting lower AB use and primary care visits (see Chapter 2 of this thesis). Usability of the FeverApp was tested positively in 2023 by means of an interactive survey among parents (n=20), while usability of the NMP WebApp will be assessed during the pilot RCT resulting from this study protocol (see Chapter 8 of this thesis). The research question for the RCT is: ‘What is the effect of the Dutch FeverApp and NMP WebApp’s on AB use, primary care visits/telephone consultations and gut microbiota in febrile children aged 0–12 years with uncomplicated URTIs?’. Methods analyses Study design and setting Dutch parents (citizens) of children (aged 3 months–12 years) with fever (≥38°C) and symptoms of acute, uncomplicated URTIs are randomly assigned to one of three study arms; two intervention groups, where parents use either the FeverApp or the NMP WebApp alongside usual (self-)care and a control group, where parents just follow usual (self-)care without an app. On the day of inclusion (T0), parents receive an email with patient information and a link to the online baseline questionnaire (5–10 min). They are also instructed to collect a baseline stool sample from their child as soon as the sample materials arrive by post, typically the next day (T1). For intervention groups, the email additionally includes study group assignment, guidance on using the assigned app and a request to use the assigned app from inclusion day onwards for URTI management. On days 2, 5 and 7 (T1, T2 and T3), participants receive a short online follow-up questionnaire (2 min) to report URTI symptoms. On day 10 (T4), they complete the final questionnaire (5–10 min) and collect a second stool sample from the child. Stool samples are retrieved within 6 weeks by a courier. A visual trial timeline is shown in figure 1, and a more detailed flowchart of the RCT is provided in online supplemental appendix 1. Timeline of the study. URTIs, upper respiratory tract infections. Participants Recruitment of participants Recruitment of parents runs from September 2023 to May 2027 (or until n=320 parents are included). Parents are recruited via Facebook and Instagram ads, LinkedIn, parental WhatsApp groups, university newsletters and paper flyers with QR codes distributed in daycare centres, primary schools, family cafes and GP/youth healthcare centre waiting rooms. A QR code or link directs potential participants to an online landing page with study details, including a 20-euro gift voucher incentive and poststudy access to both apps. Interested parents register via an online form, answering eligibility questions and providing contact details. If eligible, a researcher/research assistant preferably contacts them the same day to confirm eligibility, provide study details and collect the home address for sending stool sample materials by post. The first parent was included in December 2023, and in March 2025, n=52 parents were included in the (pilot) trial. Recruitment is conducted nationwide across the Netherlands through a digital campaign (social media and institutional channels) and paper flyers with QR codes distributed in primary care and community settings, ensuring inclusion of both urban and rural regions. Inclusion and exclusion criteria Inclusion criteria for parents are defined as follows. Eligible parents must have a child aged between 3 months and 12 years who has a fever (≥38°C) and at least one URTI symptom (acute cough, sore throat, earache or runny ear, nasal congestion or runny nose or headache). In addition, parents must master the Dutch language and possess a smartphone or tablet. Parents are excluded if they or their child have already been included in this trial before, or if they are currently participating in another trial. Children are excluded if their URTI symptoms have existed for longer than 2 weeks, if they are younger than 3 months, or if they have a history of complications or hospitalisation due to a previous URTI. Children are also excluded if they suffer from asthma, diabetes or other autoimmune diseases, or have a history of pneumonia, have Down syndrome or have a less functioning immune system for other reasons. For instance, children with immunocompromising conditions (eg, malignancy under treatment/recent, transplant recipients, systemic immunosuppression including high-dose steroids, uncontrolled HIV, asplenia/primary immunodeficiency and cystic fibrosis) are excluded. Parents are also excluded if they are assessed as being very worried about their child’s URTI during a telephone screening by the researcher/research assistants. Marked parental concern (‘very worried’) is an exclusion criterion, as this is an indication for GP assessment in febrile children under Dutch guidance. Finally, children are excluded if they have already started AB for the current URTI. Although no separate exclusion criterion for potential interactions is applied, interaction risks are addressed within the NMP WebApp itself through product-specific safety guidance (eg, contraindications and interaction warnings), thereby increasing safe self-care use. Selection of participants Selection and inclusion of parents are conducted by phone by the researchers/research assistants from UASL (EBvZ, NVS and LV), under the supervision of the principal investigator (EBvZ). If a parent meets the inclusion criteria and remains interested in participating after being fully informed about the study, they receive a patient information leaflet (PIL) by email, a respondent number and a link to the online baseline questionnaire, including a digital informed consent question for usage and publication of the data (T0). Interventions FeverApp Parents in the FeverApp group receive instructions to download and use the app, along with a link to Dutch national patient information for febrile children30 and advice to contact their GP or dial 112 for urgent concerns. The app explains fever, its purpose and, for instance, how to measure temperature accurately and includes short, animated videos with voiceovers. Parents receive information on febrile seizures, recognising serious symptoms, monitoring the child’s health status and advice on when to call a doctor through the app. The app furthermore provides personalised home care recommendations while fostering a strong parent–child bond.29 36 37 The FeverApp also provides targeted guidance on fever before and after vaccination, helping parents to interpret postvaccination fever as part of a normal immune response and to recognise when medical assessment is required in these cases. The detailed decision logic is proprietary to the original German development team and therefore not openly published. Its structure is documented in the cited German validation studies. The FeverApp used in this study is a Dutch-adapted tool aligned with national guidance and epidemiology; it has not been validated for tropical settings or dengue-endemic regions, and external use would require localisation and separate validation. NMP WebApp Parents in the NMP group receive access to an online, digital decision aid for NMPs for URTIs (NMP WebApp), in addition to national patient information and emergency contact advice. The NMP WebApp is accessible through a link and provides evidence-based advice on NMPs for URTIs, including safety, effectiveness and regulatory (EMA) status. It helps parents select NMPs for their child and informs them on where to purchase these compounds. To support safe self-care, the NMP WebApp provides product-specific safety profiles, age limits and dosage information as stated in official product labelling (EMA/over-the-counter guidance), without recommending a particular product or brand. As the NMP WebApp is a prototype, its algorithm is still being refined and is not publicly available. Controls Parents in the control group receive only the national patient information link and emergency contact advice, making them comparable to the general Dutch population of parents of children with acute, uncomplicated URTIs. Study endpoints Primary endpoint The primary endpoint of the study is AB use for the URTI during the study period. This endpoint, named AB use, is measured by a self-reported questionnaire completed by parents 10 days after inclusion (T4). AB use and primary care visits were parent reported. Recall bias is expected to be minimal given the short 10-day window. It reflects actual consumption during the episode rather than only the prescription. Medical record linkage (GP/pharmacy) is not included due to regulatory and logistical constraints. Secondary endpoints Secondary endpoints include app-related complications and adverse events (AEs). This endpoint, named complications/AEs, is measured by a self-reported questionnaire by parents 10 days after inclusion (T4). Another secondary endpoint is normal recovery time from URTI symptoms within 7 days. This endpoint, named recovery time, is assessed using self-reported questionnaires completed by parents and children at baseline and at 2, 5, 7 and 10 days after inclusion (T0–T4). In addition, the number of primary care visits and telephone consultations for the URTI during the study period is recorded. This endpoint, named primary care visits/telephone consultations, is measured by a self-reported questionnaire by parents 10 days after inclusion (T4). Finally, the diversity and abundance of the gut microbiota are analysed. This endpoint, named gut microbiome diversity, is measured by comparing 16S sequencing results of self-collected fresh morning stool samples at baseline (T0) and 10 days after inclusion (T4). Exploratory outcome measures Exploratory outcome measures include the use of NMPs for URTIs. This endpoint, named NMP use, is measured by self-reported questionnaires by parents at baseline (T0) and at 10 days after inclusion (T4). Another exploratory measure is the way of dealing with fever. This endpoint, named fever management, is measured by self-reported questionnaires by parents at baseline (T0) and at 10 days after inclusion (T4). Finally, parental self-efficacy (self-confidence) regarding the management of an URTI of their child is assessed. This endpoint, named self-efficacy, is measured by self-reported questionnaires by parents at baseline (T0) and at 10 days after inclusion (T4). Sample size The required number of participants per study group was calculated using the following formula for sample size calculation for longitudinal experimental studies: N1=N2 =(10.51*0.29*(1–0.29)*2)/(0.19*0.19)≈137 per group.41 The estimated incidence of AB prescription for uncomplicated URTIs in the control group is 35%.12 With a 50% reduction in AB prescriptions using the FeverApp in Germany (unpublished data: German Clinical Trials Register (DRKS) with registration number: DRKS00016591), we aimed for a (conservative) estimated reduction of 40% in AB use with the Dutch FeverApp. This would result in an intended incidence of 21% AB use in both intervention groups, compared with 35% in the control group. Since there are no reasons to expect differences in effectiveness between the two apps, we will compare a combined intervention group of FeverApp users and NMP WebApp users (n=134) to a control group that does not use either app (n=137). To ensure sufficient power, a total of n=271 participants are required in the analysis. In a 1:1:2 allocation, this corresponds to the smallest possible whole-number distribution of n=272. Namely: n=68 in the FeverApp group, n=68 in the NMP WebApp group, and n=136 in the control group. Considering an estimated 15% loss to follow-up, the goal is therefore to recruit n=80 parents in the FeverApp group, n=80 in the NMP WebApp group, and n=160 in the control group, resulting in a total of n=320 inclusions. The German FeverApp data reported approximately a 50% reduction in AB use (unpublished data). To avoid overestimation, we used a more conservative expected reduction of 40% in the Dutch context. Since the trial is WMO-exempt and minimal-risk, no interim efficacy monitoring is foreseen. To account for potential overestimation of the effect size, sensitivity analyses using smaller effect sizes (eg, 30%) will be performed to evaluate the robustness of the findings and to refine powering assumptions for a potential larger-scale effectiveness trial. The primary comparison pools both intervention groups (FeverApp+NMP WebApp) versus the control to maximise statistical power and avoid multiplicity, as differentiating between the two applications is not a primary aim of the trial. The sample size was powered for the primary endpoint (AB use), not gut microbiota outcomes. The assumed 15% loss to follow-up is based on the experience of the study statistician and supervising epidemiologist in comparable eHealth/self-report trials in the Netherlands. In addition, the pilot phase of this study (n=52) showed an observed loss to follow-up below 15%. Completion rates are further supported by mandatory completion of closed questions before progression in the questionnaires, daily monitoring of responses and personalised text-message reminders. We therefore consider the 15% attrition estimate to be realistic. Randomisation and blinding Participants were assigned according to a pre-specified, repeating allocation sequence prepared in Microsoft Excel. The sequence followed a fixed four-entry cycle (FeverApp—Control—NMP—Control), yielding a 1:2:1 ratio (25% FeverApp, 50% Control and 25% NMP). No random number generator was used. The allocation list was stored in a password-protected file. Research staff enrolled participants and assigned the next sequential respondent number to determine group allocation. Because the sequence was fixed and predictable, allocation concealment was limited. Blinding of study participants is not possible, but all parents receive the same study information and only learn their group assignment after agreeing to participate. Also, the Microbiota Centre Amsterdam (MiCA) lab staff who sequence the samples are blinded to the participants’ study groups. Data collection and management Most data are collected through online questionnaires. An email with the first questionnaire link is sent on inclusion day (T0), with automated follow-up emails scheduled for T1, T2, T3 and T4 via Microsoft Outlook. The questionnaires are compatible with laptops and smartphones to facilitate completion. Responses are monitored daily by the principal investigator (EBvZ), and reminders are sent via text messages if needed to optimise response rates. Respondents report their assigned respondent number in each questionnaire, allowing data merging, including stool sample data and associated paper questionnaires. Only three researchers (EBvZ, LV and NVS) have access to the data. A separate key file linking respondent numbers to personal identifiers (name, phone, address and email) is securely stored in a protected Microsoft Excel sheet, accessible only to these researchers. Personal data are used solely for sending links, coordinating sample collection and sending reminders. After data collection, all personal identifiers are deleted, retaining only respondent numbers and email addresses for potential future research. Baseline questionnaire On inclusion day (T0), parents complete an online baseline questionnaire covering sociodemographic variables, child’s gender, age, delivery mode, breastfeeding history, maternal AB use during pregnancy, prior AB and NMP use by the child, and parents’ current fever management approach. Self-efficacy in managing fever is assessed on a five-point Likert scale. Parents report baseline URTI symptom severity using a 10-point numeric rating scale (NRS), while children, if possible, use a visual six-point emoji scale (figure 2). Child-reported upper respiratory tract infections severity scale (0–10) with facial anchors (custom illustration by the authors; not adapted from copyrighted materials). Baseline stool sample Parents take a baseline stool sample from their child by collecting 2–3 spoons of fresh morning faeces, collected in a vial using a faecal collector (except for diapered children). This preferably occurs on day 2 (T1) or after receiving instructions and materials by post, and they store it in their own freezer at −20°C, and no longer than 2 days after the planned sampling day. In case this is not possible (eg, due to constipation of the child), the child will be excluded from the stool sample analysis. A paper questionnaire included with the sample materials records stool characteristics using the Bristol Stool Form Scale (figure 3), providing metadata for analysis. Dietary intake on the days of stool collections was also assessed by this questionnaire. Bristol stool chart (schematic) (categories according to Heaton and Lewis (1977), illustrated by the authors). Follow-up measurements On days 2 (T1), 5 (T2), 7 (T3) and 10 (T4), parents and children again report URTI symptoms using the NRS and emoji scale. On day 10 (T4), parents also complete the final follow-up questionnaire on AB use, app usage (for intervention groups), fever management methods, NMP use and self-efficacy. And they collect a second stool sample from their child (and complete the belonging paper ’stool questionnaire’), to store this again in their freezer until pick-up of both samples. App adherence will be assessed by parent self-report at T4 (yes/no, and features used). We do not collect in-app analytics because participants use test accounts, and no app data is stored (due to privacy reasons). Adherence will be described and included in exploratory sensitivity analyses (as-treated and dose-response by reported frequency), while the primary analysis follows an intention-to-treat approach. Finally, parents in the NMP WebApp group receive an additional usability questionnaire for providing user feedback on the prototype NMP App used. Stool sample pick-up and transportation Parents are contacted via text, email and/or phone to schedule sample pick-up. To maintain DNA quality, samples are collected within 6 weeks using a portable −20°C freezer by the researchers/research assistants or a courier. Samples are first brought to UASL for short-term storage at −80°C by Leiden Centre for Applied Bioscience (LCAB), before final transport to MiCA at Amsterdam University Medical Centres (AUMC) for long-term storage at −70°C until analysis. Stool sample analysis Gut microbiota diversity is analysed using Illumina 16S sequencing at MiCA. Biodata is securely transferred to LCAB for further R-based analysis, generating operational taxonomic units (OTUs) at family, genus or species levels. Microbiota composition changes will be compared within subjects (T0 vs T4) between AB users and non-users. Remaining stool is destroyed post-sequencing, but extracted DNA is stored at −70°C and is transported back to LCAB for long-term storage for future research. Statistical methods The R Statistical Package and SPSS V.29.0 will be used for all statistical analyses. Analysis of primary endpoints The primary endpoint, AB use, will be recorded as a dichotomous variable (0=no AB use; 1=AB use). The effect of FeverApp and NMP WebApp will be analysed using logistic regression. A dummy variable (0=control and 1=intervention) will assess variance in AB use, adjusting for possible confounders.41 Covariates (eg, gender, age, education, prior AB use, delivery mode, breastfeeding and diet) will be included. A 95% CI and 5% margin of error will be applied, targeting 90% power. Exploratory, non-powered comparisons between the FeverApp and the NMP WebApp will be conducted, but these analyses are secondary to the pooled intervention comparison. Analysis of secondary endpoints Dichotomous outcomes (eg, normal recovery within 7 days, AEs) will be analysed using logistic regression. Continuous outcomes (eg, gut microbiota diversity and primary care visits/telephone consultations) will be assessed using linear regression. The intervention dummy variable will evaluate the effects of app usage. Covariates (eg, demographics, prior AB use, delivery mode, breastfeeding and diet) will be included to correct biases. A 95% CI and 5% margin of error will be applied, aiming for 90% power. Exploratory analyses Exploratory analyses will examine relationships between app usage, NMP use, fever management and parental self-efficacy. Correlations will be explored for breastfeeding, delivery mode, maternal AB use during pregnancy, prior AB use of the child, food supplement use, diet, fever, NMP use and gut microbiota diversity. Relevant covariates will be included to address biases, with a 5% margin of error. Patient and public involvement Patients and the public were involved from the early stages of the research process. Patients, GPs, NPs and pharmacists contributed to the development of the interventions, including testing the acceptability and usability of the apps, after which adjustments were made accordingly. They were also involved in the development and testing (piloting) of the study questionnaires. The research questions and outcome measures were informed through discussions with general practitioners (GPs), ensuring alignment with clinical priorities, patient experiences and practical relevance in routine care. Patients and healthcare professionals, including GPs, nurse practitioners and paediatricians, were involved in the study set-up and design, providing input on feasibility and relevance. Recruitment to the study took place online and in public-sector settings, including primary schools, daycare centres, youth health care centres and GP practices, facilitating broad public engagement. The burden of the intervention and the time required for participation were explicitly assessed during a pilot phase of the RCT, based on participant feedback. For dissemination, participants were given the opportunity in the questionnaires to indicate whether they wished to be informed about the study results. Those who expressed interest will be informed about the outcomes of the study in an appropriate format. Ethics and dissemination The study is conducted according to the principles of the Declaration of Helsinki, in compliance with the Standard Protocol Items: Recommendations for Interventional Trials statement and all other relevant (inter)national guidelines, regulations and acts applicable to this research. Approval and trial registration The study does not fall under the Dutch Medical Research Involving Human Subjects Act (WMO), as the apps pose no risk or burden to participants, the study subjects are parents of generally healthy children, and the questionnaires do not cover sensitive topics. Therefore, a waiver was obtained from the Institutional Review Board (IRB) Medische Ethische Toetsings Commissie Brabant (decision reference: NW2023049). As the study received an IRB waiver, the appointment of a Data Monitoring Committee was deemed unnecessary. However, the data management plan was approved by the privacy officer of UASL, in accordance with the Dutch General Data Protection Regulation. A Data Protection Impact Assessment (DPIA), conducted by a data steward from the Faculty of Healthcare of UASL, indicated a low risk of data breach. The study protocol was registered at the European level on 14 November 2023, on www.clinicaltrials.gov under trial number: NCT06140446. Consent Participating parents receive an email with the PIL attached. After reviewing this document and before completing the first questionnaire, parents sign for informed consent. An example of the participant/parental informed consent form is provided (online supplemental file S1, Dutch; online supplemental file S2, certified English translation). The study obtained informed consent prior to enrolment. After reading this form, provided by email to respondents along with study information and a link to the first online questionnaire, parents sign digitally for consent in the baseline questionnaire, stating that they have read the study information and that they agree with the content of the informed consent file. Participants may withdraw from the study at any time, for any reason, without consequences. Upon withdrawal, their data will be deleted if requested. Researchers may also withdraw participants for medical reasons, with such decisions discussed with the study sponsor: AUMC. Confidentiality of data All answers to the questionnaires are pseudonymised using respondent numbers in a secure Microsoft Forms environment. No data will be collected through the study apps, as participants will use test accounts. To ensure data confidentiality, a data contract was signed between UASL, WHU and AUMC following recommendations from the privacy officer of UASL and the data steward responsible for the DPIA. Risks and burden of the study The risks associated with participation include potential complications or AEs from the child’s URTI, though these are expected to be rare. The likelihood of such events occurring more frequently in intervention groups compared with the control group is minimal. The study burden includes completing two online questionnaires (5–10 min each), reporting symptom severity at three additional time points (approximately 2 min each), and collecting and storing two stool samples from the child for up to 6 weeks in own freezer at home. Participants must also be available for sample pick-up. Intervention group participants are furthermore asked to download and use the FeverApp or NMP WebApp, but actual usage of the apps will be voluntary. Management of adverse events This study does not provide a 24/7 research contact line. Participants are instructed to contact their GP (office hours), out-of-hours GP service (evenings/weekends) or emergency department/112 for urgent concerns or alarm symptoms. In rare instances, AEs leading to hospitalisation due to URTI complications may occur; however, they are unrelated to app usage. Parents should report AEs immediately to the principal investigator (EBvZ), who will notify the sponsor (AUMC). Depending on the severity, parents are advised to contact their GP or visit an emergency department. In such cases, the child’s participation will be discontinued for safety reasons. Benefits of participation Participants in the intervention groups will receive additional support through the apps for managing their child’s URTI. While the control group will not receive this support during the study, all participants will gain access to both apps upon completing the questionnaires. Additionally, participants who collect two stool samples on top of responding to all the questionnaires will receive a €20 gift voucher as an incentive. Those unable to provide two samples through no fault of their own will still receive this voucher. Data availability statement Deidentified 16S rRNA/metagenomic sequence data will be deposited in the European Genome-phenome Archive (EGA) at publication; analysis-ready tables and deidentified clinical metadata will be available on reasonable request via the HELIUS/Amsterdam UMC data access procedure under a data use agreement. Analysis workflows will be publicly available on GitHub. Supplementary material