Decision rules and clinical prediction of pneumonia: Evaluation of low-yield criteria

doi:10.1016/S0196-0644(89)80304-X

Annals of Emergency Medicine

Volume 18, Issue 1, January 1989, Pages 13-20

https://doi.org/10.1016/S0196-0644(89)80304-X Get rights and content

Prediction rules have been recommended for guiding the ordering of diagnostic tests. Such rules can be used to define low-yield criteria (LYC) for the purpose of identifying patients with an extremely low probability of disease and hence discouraging test ordering by the physician on patients meeting LYC. In this study, community hospital emergency department populations of adults (n = 255) and children (n = 78) were evaluated prospectively for the presence of predictive clinical parameters and the physician's estimate of pneumonia prior to obtaining a chest film. We developed LYC and analyzed published LYC for obtaining chest films on patients considered at risk for pneumonia by means of logistic regression, receiver operating characteristics curve, and negative predictive value analyses. We were unable to derive or validate clinically useful LYC to improve on the seasoned clinician's probability estimate of pneumonia. We discuss the inherent limitations in the development and application of LYC that must be understood by those who seek to limit the ordering of chest films by the application of guidelines developed from decision rules.

References (13)

GrossmanLK et al.
Clinical, laboratory, and radiological information in the diagnosis of pneumonia in children
Ann Emerg Med
(1988)
DiehrP et al.
Prediction of pneumonia in outpatients with acute cough: A statistical approach
J Chronic Dis
(1984)
PaukerSG et al.
The threshold approach to clinical decision making
N Engl J Med
(1980)
FeinsteinAR
The “chagrin factor” and qualitative decision analysis
Arch Intern Med
(1985)
HanleyJA et al.
The meaning and use of the area under a receiver operating characteristic (ROC) curve
Radiology
(1982)

There are more references available in the full text version of this article.

Cited by (97)

Clinical prediction rule to predict pneumonia in adult presented with acute febrile respiratory illness
2019, American Journal of Emergency Medicine
Citation Excerpt :
While experienced clinicians may better use their clinical skills to determine the necessity of ordering a CXR, a risk stratifying prediction rule may assist all emergency physicians in whether a CXR should be ordered with better consistency and reproducibility. The possibility of predicting pneumonia from clinical parameters was studied by a number of researchers in the literature [2,9,12-16]. They derived an association between the diagnosis of pneumonia and its presenting symptoms and physical signs, using an abnormal CXR as confirmation of the diagnosis.
To derive a clinical prediction rule to predict pneumonia in patients with acute febrile respiratory illness to emergency departments.
This was a prospective multicentre study. 537 adults were recruited. Those requiring resuscitation or were hypoxaemic on presentation were excluded. Pneumonia was defined as new onset infiltrates on chest X-ray (CXR), or re-attendance within 7 days and diagnosed clinically as having pneumonia. A predictive model, the Acute Febrile Respiratory Illness (AFRI) rule was derived by logistic regression analysis based on clinical parameters. The AFRI rule was internally validated with bootstrap resampling and was compared with the Diehr and Heckerling rule.
In the 363 patients who underwent CXR, 100 had CXR confirmed pneumonia. There were 7 weighted factors within the ARFI rule, which on summation, gave the AFRI score: age ≥ 65 (1 point), peak temperature within 24 h ≥ 40 °C (2 points), fever duration ≥3 days (2 points), sore throat (−2 points), abnormal breath sounds (1 point), history of pneumonia (1 point) and SpO2 ≤ 96% (1 point). With the bootstrap resampling, the AFRI rule was found to be more accurate than the Diehr and Heckerling rule (area under ROC curve 0.816, 0.721 and 0.566 respectively, p < 0.001). At a cut-off of AFRI≥0, the rule was found to have 95% sensitivity, with a negative predictive value of 97.2%. Using the AFRI score, we found CXR could be avoided for patients having a score of <0.
AFRI score could assist emergency physicians in identifying pneumonia patients among all adult patients presented to ED for acute febrile respiratory illness.
Adult Outpatients With Acute Cough Due to Suspected Pneumonia or Influenza: CHEST Guideline and Expert Panel Report
2019, Chest
Patients commonly present to primary care services with upper and lower respiratory tract infections, and guidelines to help physicians investigate and treat acute cough due to suspected pneumonia and influenza are needed.
A systematic search was carried out with eight patient, intervention, comparison, outcome questions related to acute cough due to suspected pneumonia or influenza.
There was a lack of randomized controlled trials in the setting of outpatients presenting with acute cough due to suspected pneumonia or influenza who were not hospitalized. Both clinical suggestions and research recommendations were made on the evidence available and CHEST Expert Cough Panel advice.
For outpatient adults with acute cough due to suspected pneumonia, we suggest the following clinical symptoms and signs are suggestive of pneumonia: cough; dyspnea; pleural pain; sweating, fevers, or shivers; aches and pains; temperature ≥ 38°C; tachypnea; and new and localizing chest examination signs. Those suspected of having pneumonia should undergo chest radiography to improve diagnostic accuracy. Although the measurement of C-reactive protein levels strengthens both the diagnosis and exclusion of pneumonia, there was no added benefit of measuring procalcitonin levels in this setting. We suggest that there is no need for routine microbiological testing. We suggest the use of empiric antibiotics according to local and national guidelines when pneumonia is suspected in settings in which imaging cannot be performed. Where there is no clinical or radiographic evidence of pneumonia, we do not suggest the routine use of antibiotics. There is insufficient evidence to make recommendations for or against specific nonantibiotic, symptomatic therapies. Finally, for outpatient adults with acute cough and suspected influenza, we suggest that initiating antiviral treatment (according to Centers for Disease Control and Prevention advice) within 48 hours of symptoms could be associated with decreased antibiotic use and hospitalization and improved outcomes.
Fever in the pediatric patient
2013, Emergency Medicine Clinics of North America
Citation Excerpt :
Seven percent of children younger than 2 years with temperatures higher than 38°C have pneumonia.94 Most children older than 3 months with pneumonia have clinical signs of infection (tachypnea, nasal flaring, grunting, oxygen saturation <95%, cyanosis, abnormal lung sounds), but approximately 3% of young patients without these symptoms have radiographic evidence of occult pneumonia.93,95–97 Studies have also reported high rates of occult pneumonia (20%–30%) among highly febrile children older than 3 months with temperatures greater than 39°C and WBC counts greater than 20,000/mm3, without clinical signs of pneumonia.67
Diagnosis and Management of Pneumonia in the Emergency Department
2008, Infectious Disease Clinics of North America
The added value of C-reactive protein measurement in diagnosing pneumonia in primary care: A meta-analysis of individual patient data
2017, CMAJ
Citation Excerpt :
To quantify the added value of CRP measurement to signs and symptoms for diagnosing pneumonia in primary care, we first developed a basic diagnostic prediction model to mimic daily clinical practice using a prespecified set of commonly used and clinically relevant signs and symptoms. These clinical predictors were selected from guidelines on lower respiratory tract infections, 17,18 literature9,19–23 and consultation of experts. The predictors were age, dyspnea, tachypnea, not previously existing chest signs (physician-reported wheezing, rhonchi, crackles, diminished vesicular breathing, pleural rub or dullness), cough, (increased) sputum production, chest pain, ear-nose-throat symptoms (sore throat or rhinorrhea), (current) smoking, fever (reported by patient or physician) and comorbidity (defined as heart failure, diabetes mellitus, chronic obstructive pulmonary disorder, asthma, immunodeficiency, malignant disease or renal failure).
C-reactive protein (CRP) is increasingly being included in the diagnostic work-up for community-acquired pneumonia in primary care. Its added diagnostic value beyond signs and symptoms, however, remains unclear. We conducted a meta-analysis of individual patient data to quantify the added value of CRP measurement.
We included studies of the diagnostic accuracy of CRP in adult outpatients with suspected lower respiratory tract infection. We contacted authors of eligible studies for inclusion of data and for additional data as needed. The value of adding CRP measurement to a basic signs-and-symptoms prediction model was assessed. Outcome measures were improvement in discrimination between patients with and without pneumonia in primary care and improvement in risk classification, both within the individual studies and across studies.
Authors of 8 eligible studies (n = 5308) provided their data sets. In all of the data sets, discrimination between patients with and without pneumonia improved after CRP measurement was added to the prediction model (extended model), with a mean improvement in the area under the curve of 0.075 (range 0.02–0.18). In a hypothetical cohort of 1000 patients, the proportion of patients without pneumonia correctly classified at low risk increased from 28% to 36% in the extended model, and the proportion with pneumonia correctly classified at high risk increased from 63% to 70%. The number of patients with pneumonia classified at low risk did not change (n = 4). Overall, the proportion of patients assigned to the intermediate-risk category decreased from 56% to 51%.
Adding CRP measurement to the diagnostic work-up for suspected pneumonia in primary care improved the discrimination and risk classification of patients. However, it still left a substantial group of patients classified at intermediate risk, in which clinical decision-making remains challenging.
External validation of prediction models for pneumonia in primary care patients with lower respiratory tract infection: An individual patient data meta-analysis
2016, PLoS ONE
Pneumonia remains difficult to diagnose in primary care. Prediction models based on signs and symptoms (S&S) serve to minimize the diagnostic uncertainty. External validation of these models is essential before implementation into routine practice. In this study all published S&S models for prediction of pneumonia in primary care were externally validated in the individual patient data (IPD) of previously performed diagnostic studies.
S&S models for diagnosing pneumonia in adults presenting to primary care with lower respiratory tract infection and IPD for validation were identified through a systematical search. Six prediction models and IPD of eight diagnostic studies (N total = 5308, prevalence pneumonia 12%) were included. Models were assessed on discrimination and calibration. Discrimination was measured using the pooled Area Under the Curve (AUC) and delta AUC, representing the performance of an individual model relative to the average dataset performance. Prediction models by van Vugt et al. and Heckerling et al. demonstrated the highest pooled AUC of 0.79 (95% CI 0.74–0.85) and 0.72 (0.68–0.76), respectively. Other models by Diehr et al., Singal et al., Melbye et al., and Hopstaken et al. demonstrated pooled AUCs of 0.65 (0.61–0.68), 0.64 (0.61–0.67), 0.56 (0.49–0.63) and 0.53 (0.5–0.56), respectively. A similar ranking was present based on the delta AUCs of the models. Calibration demonstrated close agreement of observed and predicted probabilities in the models by van Vugt et al. and Singal et al., other models lacked such correspondence. The absence of predictors in the IPD on dataset level hampered a systematical comparison of model performance and could be a limitation to the study.
The model by van Vugt et al. demonstrated the highest discriminative accuracy coupled with reasonable to good calibration across the IPD of different study populations. This model is therefore the main candidate for primary care use.

View all citing articles on Scopus

: Presented at the University Association for Emergency Medicine Annual Meeting in Cincinnati, May 1988.

View full text

Original contributionDecision rules and clinical prediction of pneumonia: Evaluation of low-yield criteria

Ann Emerg Med

J Chronic Dis

The threshold approach to clinical decision making

N Engl J Med

The “chagrin factor” and qualitative decision analysis

Arch Intern Med

The meaning and use of the area under a receiver operating characteristic (ROC) curve

Radiology

Original contribution
Decision rules and clinical prediction of pneumonia: Evaluation of low-yield criteria