Agreement and predictive value of the Rockwood Clinical Frailty Scale at emergency department triage.

AIM: To determine the agreement and predictive value of emergency department (ED) triage nurse scoring of frailty using the Rockwood Clinical Frailty Scale (CFS) when compared with inpatient medical assessment using the same scale. METHODS: Prospective, dual-centre UK-based study over a 1-year period (1 April 2017 to 31 March 2018) of CFS recorded digitally at nursing triage on ED arrival and on hospital admission by a medical doctor. Inclusion criteria were emergency medical admission in those aged ≥65 staying at least one night in hospital with a CFS completed in both ED and at hospital admission. Agreement between ED triage nurse and inpatient hospital physician was assessed using a weighted Kappa statistic and Spearman's correlation coefficient. The ability of the ED to diagnose frailty (defined by a CFS ≥5) was assessed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and receiver operating characteristic (ROC) curves. At both time points the ability of the CFS to predict inpatient mortality was also assessed. RESULTS: From 29 211 admissions aged ≥65 who stayed at least one night in hospital, 12 385 (42.3%) were referred from the ED. Of the ED referrals, 8568 cases (69.2%) were included with paired CFS performed. Median age was 84 (IQR 77 to 89) with an inpatient mortality of 6%. Median CFS in ED was 4 (3 to 5) and on hospital admission 5 (4 to 6). Agreement between the ED CFS and admission CFS was weak (Kappa 0.21, 95% CI 0.19 to 0.22, rs 0.366). The area under the ROC curve (AUC) was 0.67 (95% CI 0.66 to 0.68) for the ED CFS ability to predict an admission CFS ≥5. To predict inpatient mortality the ED CFS AUC was 0.56 (0.53 to 0.59) and admission CFS AUC 0.70 (0.68 to 0.73). CONCLUSION: Agreement between ED CFS and inpatient CFS was found to be weak. In addition the ability of ED CFS to predict clinically important outcomes was limited. NPV and PPV for ED CFS cut-off value of ≥5 were found to be low. Further work is required on the feasibility, clinical impact and appropriate tools for screening of frailty in EDs.

Validation of the National Emergency Department Overcrowding Score (NEDOCS) in a UK non-specialist emergency department.

INTRODUCTION: Emergency department (ED) crowding has significant adverse consequences, however, there is no widely accepted tool to measure it. This study validated the National Emergency Department Overcrowding score (NEDOCS) (range 0-200 points), which uses routinely collected ED data. METHODS: This prospective single-centre study sampled data during four periods of 2018. The outcome against which NEDOCS performance was assessed was a composite of clinician opinion of crowding (physician and nurse in charge). Area under the receiver operating characteristic curves (AUROCs) and calibration plots were produced. Six-hour stratified sampling was added to adjust for temporal correlation of clinician opinion. Staff inter-rater agreement and NEDOCS association with opinion of risk, safety and staffing levels were collected. RESULTS: From 905 sampled hours, 448 paired observations were obtained, with the ED deemed crowded 18.5% of the time. Inter-rater agreement between staff was moderate (weighted kappa 0.57 (95% CI 0.56 to 0.60)). AUROC for NEDOCS was 0.81 (95% CI 0.77 to 0.86). Adjusted for temporal correlation, AUROC was 0.80 (95% CI 0.73 to 0.88). At a cut-off of 100 points sensitivity was 75.9% (95% CI 65.3% to 84.6%), specificity 72.1% (95% CI 67.1% to 76.6%), positive predictive value 38.2% (95% CI 30.7% to 46.1%) and negative predictive value 92.9% (95% CI 89.3% to 95.6%). NEDOCS underpredicted clinical opinion on Calibration assessment, only partially correcting with intercept updating. For perceived risk of harm, safety and insufficient staffing, NEDOCS AUROCs were 0.71 (95% CI 0.61 to 0.82), 0.71 (95% CI 0.63 to 0.80) and 0.70 (95% CI 0.64 to 0.76), respectively. CONCLUSIONS: NEDOCS demonstrated good discriminatory power for clinical perception of crowding. Prior to implementation, determining individual unit ED cut-off point(s) would be important as published thresholds may not be generalisable. Future studies could explore refinement of existing variables or addition of new variables, including acute physiological data, which may improve performance.

An external validation study of a clinical prediction rule for medical patients with suspected bacteraemia.

OBJECTIVE: The objective of this study was to externally validate a clinical prediction rule (CPR)-the 'Shapiro criteria'-to predict bacteraemia in an acute medical unit (AMU). METHODS: Prospectively collected data, retrospectively evaluated over 11 months in an AMU in the UK. From 4810 admissions, 635 patients (13%) had blood cultures (BCs) performed. The 100 cases of true bacteraemia were compared with a randomly selected sample of 100 control cases where BCs were sterile. RESULTS: To predict bacteraemia (at a cut-off score of two points), the Shapiro criteria had a sensitivity of 97% (95% CIs 91% to 99%), specificity 37% (28% to 47%), positive likelihood ratio 1.54 (1.3 to 1.8) and a negative likelihood ratio of 0.08 (0.03 to 0.25). The area under the receiver operating curve was 0.80 (0.74 to 0.86), and the Hosmer-Lemeshow p value was 0.45. CONCLUSIONS: A cut-off score of two points on the Shapiro criteria had high sensitivity to predict bacteraemia in a study of acute general medical admissions. Application of the rule in patients being considered for a BC could identify those at low risk of bacteraemia. Though the model demonstrated good discrimination, the lengthy number of variables (13) and difficulty automating the CPR may limit its use.

Clinical decision-making in older adults following emergency admission to hospital. Derivation and validation of a risk stratification score: OPERA.

OBJECTIVES OF THE STUDY: Demographic changes alongside medical advances have resulted in older adults accounting for an increasing proportion of emergency hospital admissions. Current measures of illness severity, limited to physiological parameters, have shortcomings in this cohort, partly due to patient complexity. This study aimed to derive and validate a risk score for acutely unwell older adults which may enhance risk stratification and support clinical decision-making. METHODS: Data was collected from emergency admissions in patients ≥65 years from two UK general hospitals (April 2017- April 2018). Variables underwent regression analysis for in-hospital mortality and independent predictors were used to create a risk score. Performance was assessed on external validation. Secondary outcomes included seven-day mortality and extended hospital stay. RESULTS: Derivation (n = 8,974) and validation (n = 8,391) cohorts were analysed. The model included the National Early Warning Score 2 (NEWS2), clinical frailty scale (CFS), acute kidney injury, age, sex, and Malnutrition Universal Screening Tool. For mortality, area under the curve for the model was 0.79 (95% CI 0.78-0.80), superior to NEWS2 0.65 (0.62-0.67) and CFS 0.76 (0.74-0.77) (P<0.0001). Risk groups predicted prolonged hospital stay: the highest risk group had an odds ratio of 9.7 (5.8-16.1) to stay >30 days. CONCLUSIONS: Our simple validated model (Older Persons' Emergency Risk Assessment [OPERA] score) predicts in-hospital mortality and prolonged length of stay and could be easily integrated into electronic hospital systems, enabling automatic digital generation of risk stratification within hours of admission. Future studies may validate the OPERA score in external populations and consider an impact analysis.

Persistently elevated early warning scores and lactate identifies patients at high risk of mortality in suspected sepsis.

OBJECTIVE: In the UK, the National Early Warning Score (NEWS) is recommended as part of screening for suspicion of sepsis. Is a change in NEWS a better predictor of mortality than an isolated score when screening for suspicion of sepsis?. METHODS: A prospectively gathered cohort of 1233 adults brought in by ambulance to two UK nonspecialist hospitals, with suspicion of sepsis at emergency department (ED) triage (2015-2017) was analysed. Associations with 30-day mortality and ICU admission rate were compared between groups with an isolated NEWS ≥5 points prehospital and those with persistently elevated NEWS prehospital, in ED and at ward admission. The effect of adding the ED (venous or arterial) lactate was also assessed. RESULTS: Mortality increased if the NEWS persisted ≥5 at ED arrival 22.1% vs. 10.2% [odds ratio (OR) 2.5 (1.6-4.0); P < 0.001]. Adding an ED lactate ≥2 mmol/L was associated with an increase in mortality greater than for NEWS alone [32.2% vs. 13.3%, OR 3.1 (2.2-4.1); P < 0.001], and increased ICU admission [13.9% vs. 3.7%, OR 3.1 (2.2-4.3); P < 0.001]. If NEWS remained ≥5 at ward admission (predominantly within 4 h of ED arrival), mortality was 32.1% vs. 14.3%, [OR 2.8 (2.1-3.9); P < 0.001] and still higher if accompanied by an elevated ED lactate [42.1% vs. 16.4%, OR 3.7 (2.6-5.3); P < 0.001]. CONCLUSION: Persistently elevated NEWS, from prehospital through the ED to the time of ward admission, combined with an elevated ED lactate identifies patients with suspicion of sepsis at highest risk of in-hospital mortality.

Systematic review of prognostic prediction models for acute kidney injury (AKI) in general hospital populations.

OBJECTIVE: Critically appraise prediction models for hospital-acquired acute kidney injury (HA-AKI) in general populations. DESIGN: Systematic review. DATA SOURCES: Medline, Embase and Web of Science until November 2016. ELIGIBILITY: Studies describing development of a multivariable model for predicting HA-AKI in non-specialised adult hospital populations. Published guidance followed for data extraction reporting and appraisal. RESULTS: 14 046 references were screened. Of 53 HA-AKI prediction models, 11 met inclusion criteria (general medicine and/or surgery populations, 474 478 patient episodes) and five externally validated. The most common predictors were age (n=9 models), diabetes (5), admission serum creatinine (SCr) (5), chronic kidney disease (CKD) (4), drugs (diuretics (4) and/or ACE inhibitors/angiotensin-receptor blockers (3)), bicarbonate and heart failure (4 models each). Heterogeneity was identified for outcome definition. Deficiencies in reporting included handling of predictors, missing data and sample size. Admission SCr was frequently taken to represent baseline renal function. Most models were considered at high risk of bias. Area under the receiver operating characteristic curves to predict HA-AKI ranged 0.71-0.80 in derivation (reported in 8/11 studies), 0.66-0.80 for internal validation studies (n=7) and 0.65-0.71 in five external validations. For calibration, the Hosmer-Lemeshow test or a calibration plot was provided in 4/11 derivations, 3/11 internal and 3/5 external validations. A minority of the models allow easy bedside calculation and potential electronic automation. No impact analysis studies were found. CONCLUSIONS: AKI prediction models may help address shortcomings in risk assessment; however, in general hospital populations, few have external validation. Similar predictors reflect an elderly demographic with chronic comorbidities. Reporting deficiencies mirrors prediction research more broadly, with handling of SCr (baseline function and use as a predictor) a concern. Future research should focus on validation, exploration of electronic linkage and impact analysis. The latter could combine a prediction model with AKI alerting to address prevention and early recognition of evolving AKI.