Children with acute abdominal pain are among the most difficult patients a general practitioner faces. The symptoms of appendicitis overlap heavily with those of gastroenteritis, constipation and other self-limiting conditions, and more than 90 percent of children who present with acute abdominal pain in primary care turn out to have a benign cause. Yet roughly 5 percent have acute appendicitis, the leading reason for referral to secondary care in this population, and missing it can have serious consequences. A Dutch research consortium has now published the study protocol for a large randomised trial designed to find out whether a structured diagnostic strategy, combining a validated prediction rule with a rapid bedside blood test, can help general practitioners make this decision more safely and efficiently.
The study, known as the ISAAK study, is a pragmatic, hybrid type 1 effectiveness-implementation cluster randomised controlled trial that will take place in Dutch general practice. It is described in a protocol paper published on 13 August 2026 in Diagnostic and Prognostic Research, an open-access journal. The protocol has been registered at ClinicalTrials.gov under identifier NCT06762275, with registration dated 3 December 2024, and was reviewed and approved by the Central Ethical Review Committee of the University Medical Center Groningen. Recruitment of general practitioner practices began in November 2024, the first child was enrolled in March 2025, and inclusion is expected to continue until January 2028, with a planned end date of 1 April 2028 after three months of follow-up.
The trial will enrol children aged 4 to 18 years who present to their general practitioner with acute abdominal pain of no more than seven days’ duration. Children who have previously had an appendectomy, whose abdominal pain is caused by trauma, or who are pregnant are excluded. General practitioner practices, rather than individual children, are the unit of randomisation: practices are allocated one to one, using computer-generated permuted blocks of randomly varying size, stratified by practice size above or below 5000 patients. An independent methodologist performed the randomisation. Blinding of general practitioners and patients is neither feasible nor appropriate in such a pragmatic study, and outcome assessors cannot be blinded either, but the data analysts will remain unaware of group allocation until the analysis is complete.
In the intervention arm, general practitioners will use a diagnostic strategy built around a seven-item prediction rule consisting of routinely available symptoms and signs, each assigned a point score. Based on the total score, children are stratified into three risk groups: low risk at 0 to 7 points, medium risk at 8 to 13 points, and high risk at 14 points or more. The rule was developed and validated specifically in a primary care setting, and its external validation demonstrated good discrimination, with a c-statistic of 0.90 (95 percent confidence interval 0.89 to 0.92), and good calibration. The cut-offs of 7 and 14 points correspond to a sensitivity and specificity for appendicitis of at least 95 percent, according to the protocol authors.
Management advice is tailored to the risk group. For children in the low-risk group, the strategy recommends safety netting, meaning parents are given clear instructions about when to seek further care. Children in the high-risk group are referred immediately. If the CRP value is below 10 milligrams per litre, safety netting is advised; if it is 10 to 50 milligrams per litre, the strategy recommends reassessment or immediate referral depending on the clinician’s judgement and the preferences of the parent and child; and if it is 50 milligrams per litre or higher, immediate referral is recommended. A planned reassessment is considered a continuation of the initial consultation and should take place later the same day or, if not feasible, the following morning.
The CRP cut-offs are drawn from previous primary care research. In that work, a CRP value of at least 10 milligrams per litre showed a sensitivity of 0.87 (95 percent confidence interval 0.77 to 0.94) for appendicitis, rising to 1.00 (95 percent confidence interval 0.89 to 1.00) after 48 hours, while a value of at least 50 milligrams per litre demonstrated a specificity of 0.94 (95 percent confidence interval 0.93 to 0.96). The protocol notes that current Dutch general practitioner guidelines provide limited support for this decision, because they lack explicit referral criteria for appendicitis and do not recommend CRP point-of-care testing due to insufficient evidence. In the control arm of the trial, general practitioners provide usual care at their own discretion, advised but not obliged to follow the national guideline for abdominal pain in children, which does not recommend CRP testing.
The primary outcome of the trial is referral efficiency, defined as the proportion of non-referrals among children without appendicitis during 30 days of follow-up. The diagnosis of appendicitis will be based on imaging, intraoperative findings or histopathology as reported by the secondary care specialist, and after 30 days the development of appendicitis is considered highly unlikely. Safety is a secondary outcome, defined as the proportion of children with appendicitis who are referred during the initial consultation or a planned reassessment. Further secondary outcomes include the proportion of children undergoing CRP point-of-care testing, the proportion with a planned reassessment, child anxiety measured with the Dutch State-Trait Anxiety Inventory for Children, satisfaction with the baseline consultation measured with the Parental Medical Interview Satisfaction Scale, quality of life measured with the EQ-5D instruments, and medical and societal costs assessed with adapted versions of the iMTA productivity and medical consumption questionnaires.
The sample size calculation anticipates an improvement in referral efficiency from 88 percent under current care to 95 percent, the weighted mean of the specificities of the prediction rule cut-off and the CRP threshold. To detect that 7 percentage point difference with 80 percent power at a two-sided 5 percent significance level, assuming a cluster size of 3 and an intraclass correlation of 0.01, the trial requires 510 children without appendicitis, and allowing for 10 percent loss to follow-up the team aims to include 566 such children, 283 per arm. Because children with appendicitis do not contribute to this calculation, all children assessed for appendicitis will be included until the target is reached, and those with appendicitis will contribute to the safety analysis. To reach the target, the researchers aim to recruit at least 170 general practitioner practices, since on average two to six children per practice present with acute abdominal pain during office hours over two years.
Because diagnostic strategies are often not used, or used incorrectly, by general practitioners, the protocol embeds targeted implementation strategies. Barriers and facilitators were identified in advance through qualitative research, literature and stakeholder input using the Consolidated Framework for Implementation Research, and strategies were selected with the CFIR-ERIC Matching Tool. The two main strategies are a 60-minute online training covering acute abdominal pain in children, use of the prediction rule and point-of-care testing, and risk-based communication, and a clinical decision support tool integrated into the primary care electronic health record system. For practices using the Medicom record system, the strategy appears in real time as a pop-up when a gastrointestinal symptom code is entered for a child aged 4 to 18; for other practices, general practitioners access the tool through a REDCap link via a button available in five Dutch record systems. Implementation outcomes, including reach, adoption, implementation and maintenance, will be assessed through project records, decision tool logs and semi-structured interviews with general practitioners, with maintenance assessed two years after inclusion using the Provider REport of Sustainment Scale.
The protocol authors are careful to frame the trial as a test of a hypothesis rather than a report of results. They hypothesise that use of the diagnostic strategy will reduce referrals without increasing delayed diagnoses of appendicitis, which may ultimately lead to better patient outcomes, lower costs and reduced health care resource use. They also acknowledge limitations: the trial cannot include children presenting at out-of-hours services, which may limit generalisability, and the safety outcome is not powered as a co-primary endpoint, so any observed differences in safety must be interpreted with caution. The analysis will be on an intention-to-treat basis, with cost-effectiveness evaluated from both societal and health care perspectives over a 30-day time horizon and secondary analyses over three months. If the hypothesis is confirmed, the findings may support adoption of the strategy in routine practice, contributing to more efficient care for children with acute abdominal pain.
Subject of Research: A cluster randomised trial protocol evaluating a diagnostic strategy for appendicitis in children presenting with acute abdominal pain in primary care
Article Title: Impact of a diagnostic strategy for appendicitis in children with acute abdominal pain in primary care: study protocol for a hybrid type 1 cluster randomised controlled trial
Article References: Hogervorst, E. M., Venekamp, R. P., Knol-de Vries, G. E., Platteel, T. N., Watjer, R. M., Hurkmans, D. P., Numans, M. E., Blok, G. C. G. H., Naber, E. J. A. Q., de Boer, M. R., Vermeulen, K. M., Hogenes, M., Hiddink, E., Tromp, P. C., Gorter, R. R., Holvast, A., Hulscher, J. B. F., Everts-Panman, C. M. C. R., Welker, G. A., … Holtman, G. A. (2026). Impact of a diagnostic strategy for appendicitis in children with acute abdominal pain in primary care: study protocol for a hybrid type 1 cluster randomised controlled trial. Diagnostic and Prognostic Research, 10(1), Article 22. https://doi.org/10.1186/s41512-026-00234-x
Image Credits: AI Generated
DOI: 10.1186/s41512-026-00234-x
Keywords: appendicitis, children, primary care, diagnostic strategy, prediction rule, CRP point-of-care testing, cluster randomised trial, referral efficiency, general practitioners, study protocol, implementation science, ISAAK study
News Source: Ophelia Keating. (October 28, 2026). Dutch trial to test whether a simple bedside rule can cut unnecessary appendicitis referrals in children. Scienmag.
