Excluding serious illness in feverish children in primary care: restricted rule-out method for diagnosis
BMJ 2009; 338 doi: https://doi.org/10.1136/bmj.b1187 (Published 20 April 2009) Cite this as: BMJ 2009;338:b1187
- Matthew J Thompson, senior clinical scientist and general practitioner1,
- Anthony Harnden, university lecturer1,
- Chris Del Mar, dean2
- 1Department of Primary Health Care, University of Oxford, England
- 2Faculty of Health Sciences and Medicine, Bond University, Australia
- Correspondence to: M J Thompson matthew.thompson{at}dphpc.ox.ac.uk
Case scenarios
Consider two common scenarios in primary care.
Child 1: A mother calls the out of hours service in the early evening about her 3 year old son. He was seen earlier in the week with cough and runny nose, and the general practitioner (GP) diagnosed an upper respiratory tract infection. He has been getting more miserable and irritable, is lying on the sofa reluctant to move, looks “washed out,” and has a fever. His mother is worried: “Could it be meningitis, doctor?”
Child 2: A 2 year old girl is brought to your morning surgery with a history of a fever that started during the night. You have already seen several children this morning with a non-specific but seemingly mild viral illness.
The diagnostic dilemma
The list of possible diagnoses for febrile children seen in primary care is long. Identifying children who may have a serious illness can be difficult and is at the heart of decisions to prescribe, investigate, and refer to hospital. Serious infections (including pneumonia, meningitis, septicaemia, appendicitis) account for less than 1% of children presenting to primary care,1 2 yet they are leading causes of morbidity and mortality in children. One of the key challenges for primary care practitioners is trying to balance the risk of missing a serious disease against unnecessary investigation or referral. The diagnostic process hinges on the need to rule out (exclude) serious illnesses.
The diagnostic approach: the restricted rule-out
Many doctors come to use some variant of the restricted rule-out method, which recognises that we cannot rule out all of the alternative diagnoses for each presenting complaint, but that there is a short list of serious ones that we absolutely must rule out. The method involves constructing a limited list of serious diagnoses to be ruled out, in addition to constructing a conventional list of the most likely differential diagnoses.3 This process utilises features from the history, examination, and investigations.
In the above scenarios of children with febrile illness, consider the processes that GPs use to exclude a rare yet serious illness such as meningococcal disease. Although imperative to rule out, this condition is difficult to diagnose at an early stage as it starts with non-specific clinical features before progressing rapidly to invasive disease.4 5 Accordingly, about half of children with meningococcal disease do not receive an accurate diagnosis at first consultation in primary care,4 6 7 and complications and mortality rates are lower if meningococcal disease can be recognised and treated expeditiously.7 8
History
The history provides the initial critical information to build up a mental list of possible diagnoses, and establishes whether serious infections are considered. The evolution of the symptoms may help identify disease—for example, Child 1’s recent history of sore throat and runny nose might help to rule out meningococcal disease, whereas unusual or possible “red flag” symptoms (such as a rash or laboured breathing) could suggest serious infection. The speed of onset of illness can be helpful: meningococcal disease, for example, progresses rapidly—nearly 90% of children with it are admitted to hospital 24-48 hours after onset of symptoms.4 The long history of illness of Child 1 makes meningococcal disease less likely (although pneumonia often presents over several days, and other forms of bacterial meningitis can develop over a couple of days). The short illness in Child 2 is of concern, or it could simply indicate an “overly concerned” parent. Previous consulting behaviour and eliciting particular concerns from the parent about this illness being “different from others” can be important.9
Overall assessment of severity of illness
Observing the overall or “global” appearance of the child from the moment he or she enters the consultation room can alert us to the possibility of severe illness. A survey of GPs reported that the features that they use most often to assess overall severity of illness are related to the child’s activity or behaviour.10 These included level of alertness, response to social interaction (including eye contact or engaging with the doctor), and interest in their surroundings, such as playing or moving around the consultation room. Other helpful features in this initial assessment include skin colour (pallid, mottled, blotchy, cyanosed), temperature, respiratory rate, increased work of breathing, hydration, and peripheral circulation (such as capillary refill time, cold extremities, heart rate).4 5 10 These features should prompt a more comprehensive assessment to search for specific features of serious infections. Absence of these features is not sensitive enough to exclude serious infection. When Child 1 is reviewed at the evening out of hours surgery, he walks into the examination room and immediately heads for the toy box, smiling occasionally—indications that serious infection is less likely. Child 2 is carried in by her parent, does not look at you, and seems less alert than usual, observations that raise the possibility of serious illness and the need to try to exclude this.
Localising features
Depending on findings during the first two stages, primary care practitioners will check for localising features of both serious and minor infections. In the early stages most patients with meningococcal disease have non-specific symptoms: fever, nausea, vomiting, decreased appetite, being miserable or lethargic.4 5 11 12 It is difficult to exclude or confirm the presence of serious infection on the basis of these early clinical features. In general, rule-in features are missing for most diseases and typically we focus on rule-out features—for example the absence of tachypnoea in children presenting with respiratory symptoms largely excludes a diagnosis of pneumonia.13 When no features in the history and overall assessment give concern, the search for an inflamed eardrum or chickenpox rash may be all that is needed. Child 2 presents with features in the history and initial assessment that could cause concern, and needs a more detailed assessment before serious illness can be ruled out.
Early markers of possible meningococcal disease
What can the primary care practitioner do at this stage to rule out serious infection? One aspect to consider is the speed with which the illness is evolving. As we mentioned previously, meningococcal disease progresses very rapidly, whereas typical upper respiratory tract infections can linger for 1-2 weeks. In addition it is worth looking for the early “red flag” features of meningococcal disease such as leg pain, confusion, cold extremities (box 1).4 14 15 These are clinical features that primary care practitioners might consider as uncommon in self limiting infections, so may have some value at discriminating them from meningococcal disease.
Box 1: Early “red flags” for meningococcal disease
Lethargy—most patients with meningococcal disease will be drowsy or unusually lethargic
Confusion—nearly half of older children and adults will be acutely confused
Headache—less than half of children overall will have headache, but this symptom is more common in older children and adults4 15 16 17
Leg pain occurs in approximately one third of preschool age children and up to two thirds of older children and young adults with meningococcal disease.4 18 These pains are often so severe that the patient is reluctant to move or walk
Cold hands or feet—peripheral perfusion may be impaired; about half of children will have cold hands and feet (even though they have a fever), and about a fifth will have abnormal skin colour, such as pallor or cyanosis4
Classic features
The classic “red flag” features of meningococcal disease such as neck stiffness, photophobia, and haemorrhagic rash are more common later in the clinical course.4 5 11 12 14 16 17 They are less common in younger children and infants, at least in the prehospital course of illness.4 15When they are present, they are helpful to rule in possible meningococcal disease—but even then, they are not highly specific in isolation: for example, only a small proportion of children presenting with petechial rash will have septicaemia.
When these features are absent it is difficult to be sure whether meningococcal disease is absent, or the symptom simply has not developed yet in the course of the evolving infection. Box 2 outlines some of the pitfalls of trying to use these features to rule out meningococcal disease.
Box 2: Pitfalls in the classic features of meningococcal disease
Non-blanching rash, most commonly haemorrhagic (including petechial or larger purpuric or ecchymotic lesions)
Most patients with meningococcal disease develop a rash at some point in their illness, but in the early stages of the disease, the rash may be absent or may be a blanching macular or papular rash19
Haemorrhagic lesions may be missed if they are sparse or in dark skinned patients, so it may be worthwhile to look in areas of pressure or the conjunctivae
In primary care, petechiae are only rarely caused by meningococcal disease—most will have minor or self limiting causes20
Photophobia
This is not common in children, occurring in only a quarter
Fever, neck stiffness, and altered mental state in adults
The effect of age
Infants (under 1 year old) with meningococcal disease are more likely to present with non-specific features, such as cold hands and feet, an abnormal skin colour, or breathing difficulty rather than signs of meningeal irritation.4 Preschool children (1-4 years old) also have non-specific features of infection, but nearly one third will have leg pain or neck stiffness, half will have cold hands and feet, and just under half will be confused. The presentation in older children and teenagers tends to be more similar to that of adults, who usually have a combination of headache, neck stiffness, or an altered mental state.4 15 16 17
Safety netting
The restricted rule-out process depends on being aware of the relative frequency and time course of clinical features in the serious diseases which need to be ruled out. Unfortunately, this process is not 100% sensitive; inevitably, some serious illnesses will be incorrectly ruled out (or not even considered) at an initial consultation. Specific advice for parents about which clinical features to look out for, an appropriate time-line for reconsultation, and information about how to access care are important components of safety netting.
Case review
At consultation, Child 1 appears generally well and playful, with no increased work of breathing or tachypnoea. The most likely diagnosis is a persisting upper respiratory tract infection, and the out of hours GP considers that pneumonia or meningitis should be ruled out. Focused examination confirms typical signs of an upper respiratory tract infection and no features that give concern, ruling out serious infections. Exploring the parents’ concerns and safety netting, with explanation of the clinical course of upper respiratory tract infections and what to watch for at home, are all that is needed.
Child 2 has a high temperature and seems withdrawn and much less alert than usual. She does not have a rash, difficulty breathing, or any obvious focus of infection on examination. Dip stick urine testing is negative and probably excludes urinary infection, but the GP is unable to rule out serious infection, which in this child would include meningococcal disease. The child is referred urgently to the paediatric team for assessment, and she is admitted to hospital for further investigation.
Learning points
The restricted rule-out approach to diagnosis recognises that it is difficult for clinicians to rule out all alternative diagnoses for each presenting complaint
For any clinical presentation, a short list of rare serious diagnoses usually must be ruled out, as must a list of more common but less critical diagnoses, so thinking explicitly about serious diagnoses (and, for any plausible ones, checking for rule-out clinical features) and safety netting (explaining features to watch for, an appropriate timeframe for re-consultation, and how to access care), may be the best protection.
The case study of children presenting with febrile illness shows how primary care practitioners may rule out serious infections such as meningococcal disease by using clinical assessment and investigations and being aware of discriminatory factors such as the relative frequency and time course of clinical features in serious disease
Notes
Cite this as: BMJ 2009;338:b1187
Footnotes
This series aims to set out a diagnostic strategy and illustrate its application with a case. The series advisers are Kevin Barraclough, general practitioner, Painswick, and research fellow in community based medicine, University of Bristol; Paul Glasziou, senior clinical research fellow and lecturer in medical statistics, Department of Primary Health Care, University of Oxford; and Peter Rose, university lecturer, Department of Primary Health Care, University of Oxford
Contributors: MT drafted the article and made critical revisions to the article; AH and CDM drafted, revised, and commented on various drafts of the article and read and approved the final draft. MT is guarantor.
The Department of Primary Health Care is part of the NIHR School of Primary Care Research. The study sponsors had no role in the study design; in the collection, analysis, or interpretation of data; in the writing of the report; nor in the decision to submit the article for publication.
Competing interests: None declared.
Provenance and peer review: Not commissioned; not externally peer reviewed.
Patient consent not required (patient anonymised, dead, or hypothetical).