Clinical Prediction Rule Research Articles

Clinical Decision Rules: A Starting Place in Medical Education, Not a Destination

Clinical Decision Rules: A Starting Place in Medical Education, Not a Destination

338 Inter-Rater Reliability and Acceptability of a Clinical Prediction Rule for Opioid-Associated Out-of-Hospital Cardiac Arrest

338 Inter-Rater Reliability and Acceptability of a Clinical Prediction Rule for Opioid-Associated Out-of-Hospital Cardiac Arrest

A cross-sectional analysis of four common clinical decision rules for pulmonary embolism, Mashhad, Iran.

Pulmonary embolism (PE) is a potentially fatal condition. Several non-invasive clinical decision rules (CDRs) were developed for the safe exclusion of PE. All CDRs used to safely rule out PE have been created and tested within hospital or acute care environments. However, CDRs that are designed in one specific setting may not perform as effectively when used in a different setting. In this study, we aimed to compare the performance of four common CDRs; Wells Score, Simplified Wells Score, revised Geneva Score, and simplified revised Geneva Score. This was a cross-sectional study in which patients suspected of PE presenting to Imam Reza Hospital or Ghaem Hospital were recruited from September 23, 2013, to March 19, 2016 in Mashhad, Iran. The specificity, sensitivity, and accuracy were utilized as metrics to compare the CDRs in our region. Two hundred and forty patients were included in the study. The mean age of patients was 57.91±19.97 years, and 54.16% of them (n=130) were female. 120 patients were confirmed to have PE with CT angiography. Wells score showed the highest sensitivity (90.4%) and revised Geneva score represented the highest specificity (84.9%). The highest accuracy belongs to the simplified Wells score (62.3%). In this study, we demonstrated that the Wells criteria with its high sensitivity, can be used as a score for screening, and the revised Geneva score with its high specificity, can be used in the second stage for healthy people who have been diagnosed as unhealthy by the Wells score.

Current practices in the diagnosis of acute pulmonary embolism: a comprehensive analysis of adherence to contemporary practice guidelines.

Acute pulmonary embolism (PE) poses significant diagnostic challenges with potentially fatal outcomes if not promptly identified and treated. Despite the availability of clinical guidelines, adherence to diagnostic strategies varies globally, impacting patient outcomes. To investigate the current diagnostic practices for PE among Polish physicians and their adherence to the contemporary European Society of Cardiology (ESC) 2019 Guidelines. In this cross-sectional study utilizing a web-based survey developed through a Delphi methodology, we included hospital-based physicians across various specialties in Poland. The survey was focused on diagnostic practices, risk stratification, and adherence to the ESC 2019 Guidelines for PE. The study found limited use of guideline-recommended clinical prediction rules, with a large reliance on D-dimer assessments and CT pulmonary angiography. Risk stratification practices varied, often based on individual experience rather than evidence-based strategies. The vast majority of physicians (80.5%) had never contacted a Pulmonary Embolism Response Team (PERT), but 88% stated that they would if it were available. Cardiologists were more likely to employ guideline-recommended methods compared to other specialties. There is a need for increased adherence to guideline-recommended diagnostic and risk stratification strategies of acute PE among physicians in Poland. Enhancing the availability of Pulmonary Embolism Response Teams (PERTs) and promoting guideline-directed practices could improve diagnostic accuracy and patient outcomes.

Derivation of a clinical decision rule for termination of resuscitation in non-traumatic pediatric out-of-hospital cardiac arrest

AimPrehospital termination of resuscitation (ToR) rules are used to predict medical futility in adult out-of-hospital cardiac arrest (OHCA), however, the available evidence for pediatric patients is limited. The primary aim of this study is to derive a Pediatric Termination of Resuscitation (PToR) prediction rule for use in pediatric non-traumatic OHCA patients. MethodsWe analyzed a retrospective cohort of pediatric OHCA patients within the CARES database over a 10-year period (2013–2022). We split the dataset into training and test datasets and fit logistic regressions with Least Absolute Shrinkage and Selection Operator (LASSO) to select predictor variables and estimate predictive test characteristics for the primary outcome of death and a secondary composite outcome of death or survival to hospital discharge with unfavorable neurologic status. ResultsWe analyzed a sample of 21,240 children where 2,326 (11.0%) survived to hospital discharge, and 1,894 (8.9%) survived to hospital discharge with favorable neurologic status. We derived a PToR rule for death demonstrating a specificity of 99.1% and a positive predictive value (PPV) of 99.8% and a PToR rule for death or survival with poor neurologic status with a specificity of 99.7% and PPV of 99.9% within the test dataset. ConclusionWe derived a clinical prediction rule with high specificity and positive predictive value in prehospital settings utilizing Advanced Life Support (ALS) providers which may inform termination of resuscitation considerations in pediatric patients. Further prospective and validation studies will be necessary to define the appropriateness and applicability of these PToR criteria for routine use.

Predicting mortality in geriatric patients with fever in the emergency departments: a prospective validation study

ObjectiveEmergency physicians are always faced with the challenge of choosing the appropriate disposition for elderly patients in order to ensure an acceptable care plan and effective use of resources. A clinical decision rule, Geriatric Fever Score (GFS) has been proposed but not validated to help ED physicians with decision-making. This rule employs leukocytosis, severe coma, and thrombocytopenia as predictors of 30-day mortality. Through our study determines the performance of this clinical prediction rule in a prospective study in a setting different from where it was developed.Method and materialsIn this prospective cohort study in a 1200-bed tertiary care, patients older than 65 years old who visited the ED with fever were enrolled. All elements of the rule were collected and the total score was calculated for each patient. Patients were also categorized as low risk (score 0–1) or high risk (score ≥ 2). Thirty-day follow-up was performed to determine the patient outcome (survival or mortality).ResultsA total of 296 patients were included in our final analysis. The mortality rate was 33.1% for patients with a Score of 0, 42.1% for a score of 1, 57.1% for a score of 2, and 100% for a score of 3. When divided into two risk groups, patients’ mortality rates were as follows: low risk group 37.9% and high-risk group 40.5%.ConclusionOur study showed that elderly patients who present to ED with fever and have a score of 2 or higher on the Geriatric Fever Score are at higher risk of mortality at 30 days.

Physical therapy in burn care: Development of clinical prediction rules to determine the efficacy of low-level laser therapy

This study aimed to demonstrate the benefits of low-level laser therapy (LLLT) on burn healing and to investigate whether patients' age, burn wound size, wound stage, and total burned surface area influence the burn wound healing response to LLLT. This was a quasi-experimental study with a single-group design that included eighty male and female patients with partial-thickness burn wounds recruited from burn units. The participants were placed in a single intervention group receiving LLLT. The duration of the intervention was six weeks, divided into 18 sessions (three sessions per week). The statistical analysis was conducted using version 25 of the SPSS statistical package for Windows. The results showed a statistically significant negative relationship between wound improvement from LLLT and age (p < 0.05) and between total body surface area (TBSA) and wound improvement (p < 0.05). Additionally, a statistically significant positive relationship was found between initial wound size and wound improvement (p < 0.05). There was no statistically significant relationship between wound improvement and wound stage (p > 0.05). The current study revealed that age, TBSA, and initial wound size can predict the efficacy of low-level laser treatment for burn wounds.

Interpretable Machine Learning Models for Predicting Critical Outcomes in Patients with Suspected Urinary Tract Infection with Positive Urine Culture.

(1) Background: Urinary tract infection (UTI) is a leading cause of emergency department visits and hospital admissions. Despite many studies identifying UTI-related risk factors for bacteremia or sepsis, a significant gap remains in developing predictive models for in-hospital mortality or the necessity for emergent intensive care unit admission in the emergency department. This study aimed to construct interpretable machine learning models capable of identifying patients at high risk for critical outcomes. (2) Methods: This was a retrospective study of adult patients with urinary tract infection (UTI), extracted from the Medical Information Mart for Intensive Care IV Emergency Department (MIMIC-IV-ED) database. The critical outcome is defined as either in-hospital mortality or transfer to an intensive care unit within 12 h. ED visits were randomly partitioned into a 70%/30% split for training and validation. The extreme gradient boosting (XGBoost), random forest (RF), and support vector machine (SVM) algorithms were constructed using variables selected from the stepwise logistic regression model. The XGBoost model was then compared to the traditional model and clinical decision rules (CDRs) on the validation data using the area under the curve (AUC). (3) Results: There were 3622 visits among 3235 unique patients diagnosed with UTI. Of the 2535 patients in the training group, 836 (33%) experienced critical outcomes, and of the 1087 patients in the validation group, 358 (32.9%) did. The AUCs for different machine learning models were as follows: XGBoost, 0.833; RF, 0.814; and SVM, 0.799. The XGBoost model performed better than others. (4) Conclusions: Machine learning models outperformed existing traditional CDRs for predicting critical outcomes of ED patients with UTI. Future research should prospectively evaluate the effectiveness of this approach and integrate it into clinical practice.

Chest Pain in Primary Care: A Systematic Review of Risk Stratification Tools to Rule Out Acute Coronary Syndrome.

Chest pain frequently poses a diagnostic challenge for general practitioners (GPs). Utilizing risk stratification tools might help GPs to rule out acute coronary syndrome (ACS) and make appropriate referral decisions. We conducted a systematic review of studies evaluating risk stratification tools for chest pain in primary care settings, both with and without troponin assays. Our aims were to assess the performance of tools for ruling out ACS and to provide a comprehensive review of the current evidence. We searched PubMed and Embase for articles up to October 9, 2023 concerning adult patients with acute chest pain in primary care settings, for whom risk stratification tools (clinical decision rules [CDRs] and/or single biomarker tests) were used. To identify eligible studies, a combination of active learning and backward snowballing was applied. Screening, data extraction, and quality assessment (following the Quality Assessment of Diagnostic Accuracy Studies-2 tool) were performed independently by 2 researchers. Of the 1,204 studies screened, 14 were included in the final review. Nine studies validated 7 different CDRs without troponin. Sensitivities ranged from 75.0% to 97.0%, and negative predictive values (NPV) ranged from 82.4% to 99.7%. None of the CDRs outperformed the unaided judgment of GP's. Five studies reported on strategies using troponin measurements. Studies using high-sensitivity troponin showed highest diagnostic accuracy with sensitivity 83.3% to 100% and NPV 98.8% to 100%. Clinical decision rules without troponin and the use of conventional troponin showed insufficient sensitivity to rule out ACS in primary care and are not recommended as standalone tools. High-sensitivity troponin strategies are promising, but studies are limited. Further prospective validation in primary care is needed before implementation.

Clinical prediction rules for surgical site infection after minor surgery in general practice.

Surgical site infection (SSI) after dermatological surgery is associated with poor outcomes. Developing clinical prediction rules based on the predicted probability of infection might encourage guided prophylaxis and judicious prescribing. The purpose of this study was to develop a clinical prediction rule based on identified risk factors for SSI in a large general practice patient cohort. We examined a large, pooled dataset from four randomised controlled trials performed in a regional centre of North Queensland, Australia. Multivariable logistic regression identified a prediction model. Bootstrapping was used for internal validation. A scoring system was based on predicted probabilities of infection. The final prediction rule included age >55years and the anatomical site, histology and complexity of the excision. The area under the curve was 0.704. Our prediction rule encourages judicious use of prophylaxis in clinical practice.

Pericardial tamponade: a new perspective on echocardiographic features and application of a prediction score.

Few clinical decision rules have been used to guide clinical management and predict outcomes in patients with pericardial tamponade. The objectivesof this study are to identify the echocardiographic features associated with adverse outcomes in patients with pericardial effusions requiring pericardiocentesis and to apply a previously described four-point clinical and echocardiographic score to predict clinical outcomes over 24-hr, 30-day, and 1-year intervals. We performed a retrospective cohort review of patients who had transthoracic echocardiogram (TTE) performed and underwent pericardiocentesis within 48h of emergency department presentation at two large tertiary care institutions. We constructed different stepwise logistic regression models and examined the associations of TTE characteristics and clinical features with ICU admission, hospital length of stay (h-LOS), and survival. The data set was then employed against a previously proposed scoring system to predict factors associated with clinical outcomes over 24hr, 30days, and 1year. Two hundred thirty-nine patients were included in the final analysis. Echocardiographic characteristics of patients with pericardial tamponade who underwent pericardiocentesis are as follows: 69.1% right ventricular (RV) diastolic collapse, 62.3% exaggerated mitral valve (MV) inflow velocities, 56.4% inferior vena cava (IVC) plethora, and 53.4% right atrial (RA) systolic collapse. Increase in systolic blood pressure and increased variation in MV inflow velocity were associated with reduced ICU admission [OR: 0.94 (CI 0.90, 0.99), 0.28 (CI 0.09, 0.89), respectively]. In addition, a history of malignancy increased the length of hospital stay by about 3.89days (CI 1.43-6.35, p < 0.01) and prior pericardiocentesis history was associated with 4.82-day increase in hospital stay (CI 1.19-8.45, p = 0.01). In utilizing the previously published prediction score, we found no statistically significant correlation in predicting survival. RV diastolic collapse and exaggerated MV inflow velocity were the most common echocardiographic findings in patients requiring pericardiocentesis. Contrary to prior studies, exaggerated MV inflow velocity was associated with reduced ICU admission. In addition, a previously described prediction score did not correlate with decreased survival in this cohort.

A Multicentre Prospective Cohort Study to Identify High-Risk Transient Ischemic Attack/Minor Stroke Patients Benefitting from Echocardiography

BackgroundWe aimed to derive a clinical decision rule to identify patients with transient ischemic attack (TIA) or minor stroke most likely to benefit from echocardiography. MethodsThis multicentre prospective cohort study enrolled adults diagnosed with TIA/minor stroke in the emergency department who underwent echocardiograms within 90 days, from 13 Canadian academic emergency departments from October 2006 to May 2017. Our outcome was clinically significant echocardiogram findings. ResultsIn 7149 eligible patients, a clinically significant finding was found in 556 (7.8%). There were a further 2421 (33.9%) with potentially significant findings. History of heart failure (adjusted odds ratio [OR], 3.9) or coronary artery disease (OR, 2.7) were the factors most strongly associated with clinically significant echocardiogram findings, whereas young age, male sex, valvular heart disease, and infarct (any age) on neuroimaging were modestly associated (OR, 1.3-1.9). The model combining these predictors into a score (range: 0-15), had a C-statistic of 0.67 (95% confidence interval [CI], 0.65-0.70). A cut point of 6 points or more classified 6.6% of cases as high likelihood, defined as > 15% for clinically significant echocardiogram findings. ConclusionsEchocardiography is a very useful test in the investigations of patients with TIA/minor stroke. We identified high-risk clinical features—combined to create a clinical decision rule—to identify which patients with TIA/minor stroke are likely to have clinically significant echocardiogram findings requiring an immediate change in management. These patients should have echocardiography prioritized, whereas others may continue to have echocardiography conducted in a less urgent fashion.

Development and application of a serious adverse events risk model for concurrent chemoradiotherapy in patients with nasopharyngeal carcinoma.

The objective of this study was to construct a concise prediction model for serious adverse events (SAEs) in order to assess the likelihood of SAE occurrence among hospitalized patients undergoing concurrent chemoradiotherapy. An electronic database of a Cancer Centre was utilized to conduct a cross-sectional review survey. Our research involved the recruitment of 239 patients who were undergoing concurrent chemoradiotherapy in the Department of Nasopharynx and Radiotherapy. The clinical prediction rule was derived using logistic regression analysis, with SAE serving as the primary outcome. Internal verification was conducted. The occurrence rate of SAE in the derivation cohort was 59.4%. The ultimate model used had 3 variables, namely cystatin C, C-reactive protein, and serum amyloid A. The model exhibited an area under the curve of 0.626 (95% CI: 0.555-0.696; P < .001). The model accurately predicts the occurrence of SAE, and the variable data can be easily obtained, and the assessment technique is straightforward.

Acute pulmonary embolism pretest probability estimation by d-dimer test, our modified, new ECG score and clinical prediction rules

ObjectivesWe investigated whether a sufficiently sensitive D-dimer test could exclude acute pulmonary embolism (acPE) as a stand-alone diagnostic test and compared our previously published, modified ECG score with the Wells and Geneva scores in the estimation of acPE pretest probability. MethodsWe retrospectively evaluated 345 patients who underwent chest CT angiography (CTA) for the suspicion of acPE. The pretest probability of acPE was assessed in 120 D-dimer negative [DD (−)] and 225 D-dimer positive [DD (+)] patients. ResultsChest CTA verified acPE in 57/345 (16.5 %) patients and in 1/120 (0.8 %) DD (−) patient. In DD (−) patients the test accuracy (TA) and specificity (SP) of the ECG score (98 %, 99 %) were better than those of the Wells score (92.5 %, 92.4 %) (p = 0.063 and p < 0.05 respectively) and the Geneva score (76.7 %, 76.5 %) (p < 0.001 for both), the Wells score TA and SP were greater than those of the Geneva score (p < 0.001 for both). In DD (+) patients the SPs, TAs and positive predictive values (PPV) of the ECG score (94 %, 78.6 %, 69 %) and the Wells score (91.8 %, 75.1 %, 48 %) were greater than those of the Geneva score (71.3 %, 64.9 %, 38.2 %) (p < 0.001 for both SP and TA respectively, and p < 0.001 for PPV of the ECG score vs. the Geneva score and p < 0.05 for PPV of the Wells score vs. Geneva score), their sensitivities (SE) (36.4 %, 23.6 %) were less than that of the Geneva score (47.5 %) (p < 0.05 and p < 0.001 respectively). The ECG score's TA in a trend, its SE and PPV were significantly (p < 0.01 and p < 0.001) better than those of the Wells score. ConclusionIn contrast to the current guidelines, a stand-alone high sensitivity DD (−) test, without prediction rules, could reliably exclude acPE. Our ECG score slightly outperformed the Wells score, the ECG score and Wells score far outperformed the Geneva score in the estimation of acPE pretest probability. An acPE diagnosis with the ECG score, in addition to the supportive diagnosis with the clinical prediction rules, may further increase the chance of true DD positivity.

Development and validation of a clinical prediction rule for walking independence in hospitalized older adults with a vertebral compression fracture.

No reports on factors or Clinical prediction rules (CPRs) associated with walking independence among patients with vertebral compression fractures (VCFs) are available. Evidence regarding epidemiological walking independence rates is also sparse. Here, we sought to (i) obtain epidemiological data on the probability of inpatients with VCFs achieving walking independence, and (ii) develop and validate a CPR to determine walking independence in hospitalized patients with VCFs. We conducted a retrospective cross-sectional observational study of patients aged ≥60years who were hospitalized for VCF at four hospitals in Japan in 2019-2022. The outcome was walking independence at discharge. We performed a binomial logistic regression analysis to assess predictors of walking independence. Five independent variables were entered: age, American Society of Anesthesiologists physical status, cognitive function, Berg Balance Scale (BBS), and 10-m walking test. Among the independent variables that were significant, we converted the continuous variables to binary data by calculating cut-off values and then created the CPR. The area under the curve (AUC) was calculated as the measure of the CPR's diagnostic accuracy, and internal validation was conducted by bootstrapping. Of the 240 patients, 188 (78.3%) achieved walking independence. Cognitive function and the BBS score (with a cut-off of 45 points) were identified as significant predictors. We created a CPR using these two items (0-2 points). The CPR's AUC was 0.92 (0.874-0.967), and internal validation by bootstrapping yielded a mean AUC of 0.919 with a slope of 0.965. The walking independence rate of patients with a VCF during hospitalization was 78.3%, with cognitive function and BBS being predictors. The developed CPR performed well enough to retrospectively predict walking independence in VCF patients. The BBS cut-off value and the CPR may serve as useful indicators for clinicians to predict VCF patients' walking independence.

Application of the Kaiser score on contrast-enhanced mammography in the differential diagnosis of breast lesions: comparison with breast magnetic resonance imaging.

The Kaiser score (KS) as a clinical decision rule has been proven capable of enhancing the diagnostic efficiency for suspicious breast lesions and obviating unnecessary benign biopsies. However, the consistency of KS in contrast-enhanced mammography (CEM-KS) and KS on magnetic resonance imaging (MRI-KS) is still unclear. This study aimed to evaluate and compare the diagnostic efficacy and agreement of CEM-KS and MRI-KS for suspicious breast lesions. This retrospective study included 207 patients from April 2019 to June 2022. The radiologists assigned a diagnostic category to all lesions using the Breast Imaging Reporting and Data System (BI-RADS). Subsequently, they were asked to assign a final diagnostic category for each lesion according to the KS. The diagnostic performance was evaluated by the area under the receiver operating characteristic curve (AUC). The agreement in terms of the kinetic curve and the KS categories for CEM and MRI were evaluated via the Cohen kappa coefficient. The AUC was higher for the CEM-KS category assignment than for the CEM-BI-RADS category assignment (0.856 vs. 0.776; P=0.047). The AUC was higher for MRI-KS than for MRI-BI-RADS (0.841 vs. 0.752; P =0.015). The AUC of CEM-KS was not significantly different from that of MRI-KS (0.856 vs. 0.841; P=0.538). The difference between the AUCs for CEM-BI-RADS and MRI-BI-RADS was not statistically significant (0.776 vs. 0.752; P=0.400). The kappa agreement for the characterization of suspicious breast lesions using CEM-KS and MRI-KS was 0.885. The KS substantially improved the diagnostic performance of suspicious breast lesions, not only in MRI but also in CEM. CEM-KS and MRI-KS showed similar diagnostic performance and almost perfect agreement for the characterization of suspicious breast lesions. Therefore, CEM holds promise as an alternative when breast MRI is not available or contraindicated.

Clinical prediction model for pediatric lymphadenopathy: enhancing diagnostic precision and treatment decision making.

Lymph node enlargement is common in children, with 90% of physiologically palpable lymph nodes. This study aimed to develop a predictive model based on clinical characteristics to enhance the diagnosis of pediatric lymphadenopathy and provide insights into biopsy outcomes. A clinical prediction rule was developed using a retrospective, cross-sectional design for patients under 15years who underwent lymph node biopsy from 2012 to 2022. Multivariable risk regression was used to analyze benign and malignant lesions, presenting results through risk difference and AUROC for each group. Predicted probabilities were applied in a logistic regression equation to classify patients' lymphadenopathy as reactive hyperplasia, benign, or malignant. Of 188 children, 70 (37.2%) had benign lymphadenopathy beyond reactive hyperplasia, and 27 (14.4%) had malignant lymphadenopathy. The predictive model included 12 characteristics such as size, location, duration, associated symptoms, and lymph node examination. Predictive accuracy was 92.2% for benign cases (AUROC = 0.92; 95% CI 0.87-0.96) and 98.6% for malignancy (AUROC = 0.98; 95% CI 0.94-0.99). Overall accuracy for predicting both benign and malignant tumors was 68.3%. The model demonstrated reasonably accurate predictions for the clinical characteristics of pediatric lymphadenopathy. It tended to overestimate malignancy but did not miss diagnoses, aiding in reducing unnecessary lymph node biopsies in benign cases.

Relationship between the HeartMate Risk Score category on admission and outcome in patients with acute heart failure referred to a cardiac intensive care unit.

The HeartMate Risk Score (HMRS), a simple clinical prediction rule based on the patients' age, albumin, creatinine, and the international normalized ratio of the prothrombin time (PT-INR), is correlated with mortality in the cohort of left ventricular assist device (LVAD) recipients. However, in an aging society, an LAVD is indicated for only a small proportion of patients with acute heart failure (AHF), and whether the HMRS has prognostic implications for unselected patients with AHF is unknown. This study aimed to assess the prognostic value of HMRS categories on admission in patients with AHF. We analyzed 339 hospitalized patients with AHF who had albumin, creatinine, and the PT-INR recorded on admission. The patients were categorized as follows: the High group (HMRS > 2.48, n = 131), Mid group (HMRS of 1.58-2.48, n = 97) group, and Low group (HMRS < 1.58, n = 111). The endpoints of this study were all-cause death and readmission for heart failure (HF). During a median follow-up of 247days, 24 (18.3%) patients died in the High group, 7 (7.2%) died in the Mid group, and 8 (7.2%) died in the Low group. In a multivariable analysis adjusted for highly imbalanced baseline variables, a high HMRS was independently associated with survival, with a hazard ratio of 2.90 (95% confidence interval 1.42-5.96, P = 0.004). With regard to the composite endpoint of all-cause death and readmission for HF, the Mid group had a worse prognosis than the Low group, and the High group had the worst prognosis. A high HMRS on admission is associated with all-cause mortality and readmission for HF, and a mid-HMRS is associated with readmission for HF after AHF hospitalization. The HMRS may be a valid clinical tool to stratify the risk of adverse outcomes after hospitalization in unselected patients with AHF.

The Florida Geriatric Head Trauma CT Clinical Decision Rule.

Several clinical decision rules have been devised to guide head computed tomography (CT) use in patients with minor head injuries, but none have been validated in patients 65 years or older. We aimed to derive and validate a head injury clinical decision rule for older adults. We conducted a secondary analysis of an existing dataset of consecutive emergency department (ED) patients >65 years old with blunt head trauma. The main predictive outcomes were significant intracranial injury and Need for Neurosurgical Intervention on CT. The secondary outcomes also considered in the model development and validation were All Injuries and All Intracranial Injuries. Predictor variables were identified using multiple variable logistic regression, and clinical decision rule models were developed in a split-sample derivation cohort and then tested in an independent validation cohort. Of 5776 patients, 233 (4.0%) had significant intracranial injury and an additional 104 (1.8%) met CT criteria for Need for Neurosurgical Intervention. The best performing model, the Florida Geriatric Head Trauma CT Clinical Decision Rule, assigns points based on several clinical variables. If the points totaled 25 or more, a CT scan is indicated. The included predictors were arrival via Emergency Medical Services (+30 points), Glasgow Coma Scale (GCS) <15 (+20 points), GCS <14 (+50 points), antiplatelet medications (+17 points), loss of consciousness (+16 points), signs of basilar skull fracture (+50 points), and headache (+20 points). Utilizing this clinical decision rule in the validation cohort, a point total ≥25 had a sensitivity and specificity of 100.0% (95% CI: 96.0-100) and 12.3% (95% CI: 10.9-13.8), respectively, for significant intracranial injury and Need for Neurosurgical Intervention. The Florida Geriatric Head Trauma CT Clinical Decision Rule has the potential to reduce unnecessary CT scans in older adults, without compromising safe emergency medicine practice.

PECARN prediction rule for cervical spine imaging of children presenting to the emergency department with blunt trauma: a multicentre prospective observational study.

Cervical spine injuries in children are uncommon but potentially devastating; however, indiscriminate neck imaging after trauma unnecessarily exposes children to ionising radiation. The aim of this study was to derive and validate a paediatric clinical prediction rule that can be incorporated into an algorithm to guide radiographic screening for cervical spine injury among children in the emergency department. In this prospective observational cohort study, we screened children aged 0-17 years presenting with known or suspected blunt trauma at 18 specialised children's emergency departments in hospitals in the USA affiliated with the Pediatric Emergency Care Applied Research Network (PECARN). Injured children were eligible for enrolment into derivation or validation cohorts by fulfilling one of the following criteria: transported from the scene of injury to the emergency department by emergency medical services; evaluated by a trauma team; and undergone neck imaging for concern for cervical spine injury either at or before arriving at the PECARN-affiliated emergency department. Children presenting with solely penetrating trauma were excluded. Before viewing an enrolled child's neck imaging results, the attending emergency department clinician completed a clinical examination and prospectively documented cervical spine injury risk factors in an electronic questionnaire. Cervical spine injuries were determined by imaging reports and telephone follow-up with guardians within 21-28 days of the emergency room encounter, and cervical spine injury was confirmed by a paediatric neurosurgeon. Factors associated with a high risk of cervical spine injury (>10%) were identified by bivariable Poisson regression with robust error estimates, and factors associated with non-negligible risk were identified by classification and regression tree (CART) analysis. Variables were combined in the cervical spine injury prediction rule. The primary outcome of interest was cervical spine injury within 28 days of initial trauma warranting inpatient observation or surgical intervention. Rule performance measures were calculated for both derivation and validation cohorts. A clinical care algorithm for determining which risk factors warrant radiographic screening for cervical spine injury after blunt trauma was applied to the study population to estimate the potential effect on reducing CT and x-ray use in the paediatric emergency department. This study is registered with ClinicalTrials.gov, NCT05049330. Nine emergency departments participated in the derivation cohort, and nine participated in the validation cohort. In total, 22 430 children presenting with known or suspected blunt trauma were enrolled (11 857 children in the derivation cohort; 10 573 in the validation cohort). 433 (1·9%) of the total population had confirmed cervical spine injuries. The following factors were associated with a high risk of cervical spine injury: altered mental status (Glasgow Coma Scale [GCS] score of 3-8 or unresponsive on the Alert, Verbal, Pain, Unresponsive scale [AVPU] of consciousness); abnormal airway, breathing, or circulation findings; and focal neurological deficits including paresthesia, numbness, or weakness. Of 928 in the derivation cohort presenting with at least one of these risk factors, 118 (12·7%) had cervical spine injury (risk ratio 8·9 [95% CI 7·1-11·2]). The following factors were associated with non-negligible risk of cervical spine injury by CART analysis: neck pain; altered mental status (GCS score of 9-14; verbal or pain on the AVPU; or other signs of altered mental status); substantial head injury; substantial torso injury; and midline neck tenderness. The high-risk and CART-derived factors combined and applied to the validation cohort performed with 94·3% (95% CI 90·7-97·9) sensitivity, 60·4% (59·4-61·3) specificity, and 99·9% (99·8-100·0) negative predictive value. Had the algorithm been applied to all participants to guide the use of imaging, we estimated the number of children having CT might have decreased from 3856 (17·2%) to 1549 (6·9%) of 22 430 children without increasing the number of children getting plain x-rays. Incorporated into a clinical algorithm, the cervical spine injury prediction rule showed strong potential for aiding clinicians in determining which children arriving in the emergency department after blunt trauma should undergo radiographic neck imaging for potential cervical spine injury. Implementation of the clinical algorithm could decrease use of unnecessary radiographic testing in the emergency department and eliminate high-risk radiation exposure. Future work should validate the prediction rule and care algorithm in more general settings such as community emergency departments. The Eunice Kennedy Shriver National Institute of Child Health and Human Development and the Health Resources and Services Administration of the US Department of Health and Human Services in the Maternal and Child Health Bureau under the Emergency Medical Services for Children programme.