Emergency readmissions are substantially determined by acute illness severity and chronic debilitating illness: a single centre cohort study

Reduced 30-day in-hospital but increased long-term mortality for weekend vs weekday acute medical admission

BACKGROUND

Acute medical admission at the weekend has been reported to be associated with increased mortality. We aimed to assess 30-day in-hospital mortality and subsequent follow-up of all community deaths following discharge for acute medical admission to our institution over 21 years.

METHODS

We employed a database of all acute medical admissions to our institution over 21 years (2002-2023). We compared 30-day in-hospital mortality by weekend (Saturday/Sunday) or weekday (Tuesday/Wednesday) admission. Outcome post-discharge was determined from the National Death Register to December 2021. Predictors of 30-day in-hospital and long-term mortality were analysed by logistic regression or Cox proportional hazards models.

RESULTS

The study population consisted of 109,232 admissions in 57,059 patients. A weekend admission was associated with a reduced 30-day in-hospital mortality, odds ratio (OR) 0.70 (95%CI 0.65, 0.76). Major predictors of 30-day in-hospital mortality were acute illness severity score (AISS) OR 6.9 (95%CI 5.5, 8.6) and comorbidity score OR 2.4 (95%CI 1.2, 4.6). At a median follow-up of 5.9 years post-discharge, 19.0% had died. The strongest long-term predictor of mortality was admission AISS OR 6.7 (95%CI 4.6, 9.9). The overall survival half-life after hospital discharge was 16.6 years. Survival was significantly worse for weekend admissions at 20.8 years compared to weekday admissions at 13.3 years.

CONCLUSION

Weekend admission of acute medical patients is associated with reduced 30-day in-hospital mortality but reduced long-term survival.

Short- and long-term mortality following acute medical admission

BACKGROUND

Short-term in-hospital mortality following acute medical admission has been widely investigated. Longer term mortality, particularly out-of-hospital mortality, has been less well studied.

AIM

The aim of this study is to evaluate short- and long-term mortality, and predictors of such, following acute medical admission.

DESIGN

Retrospective database study.

METHODS

We evaluated all acute medical admissions to our institution over 10 years (2002-11) with a minimum of a further 10 years follow-up to 2021 using the Irish National Death Register. Predictors of 30-day in-hospital and long-term mortality were analysed with logistic and Cox regression, with loss of life years estimated.

RESULTS

The 2002-11 cohort consisted of 62 184 admissions in 35 140 patients. 30-Day in hospital mortality (n = 3646) per patient was 10.4% and per admission was 5.9%. There were an additional 11 440 longer-term deaths by 2021-total mortality was 15 086 (42.9%). Deaths post hospital discharge had median age at admission of 75.4 years [interquartile range (IQR) 63.7, 82.8] and died at median age of 80 years (IQR 69, 87). The half-life of survival following admission was 195 months-representing a short fall of 8 life years (32.9%) compared with the projected population reference of 24.3 years. Age [odds ratio (OR) 1.73 (95% confidence interval (CI) 1.64, 1.81)], acute illness severity score [OR 1.39 (95% CI 1.36, 1.43)] and comorbidity score [OR 1.09 (95% CI 1.08, 1.10)] predicted long-term mortality.

CONCLUSION

Similar factors influence both short- and long-term mortality following acute medical admission, the magnitude of effect is attenuated over time.

Red Cell Distribution Width as a Prognostic Indicator in Acute Medical Admissions

The red cell distribution width (RDW) is the coefficient of variation of the mean corpuscular volume (MCV). We sought to evaluate RDW as a predictor of outcomes following acute medical admission. We studied 10 years of acute medical admissions (2002-2011) with subsequent follow-up to 2021. RDW was converted to a categorical variable, Q1 < 12.9 fl, Q2-Q4 ≥ 12.9 and <15.7 fL and Q5 ≥ 15.7 fL. The predictive value of RDW for 30-day in-hospital and long-term mortality was evaluated with logistic and Cox regression modelling. Adjusted odds ratios (aORs) were calculated and loss of life years estimated. There were 62,184 admissions in 35,140 patients. The 30-day in-hospital mortality (n = 3646) occurred in 5.9% of admissions. An additional 15,086 (42.9%) deaths occurred by December 2021. Admission RDW independently predicted 30-day in-hospital mortality aOR 1.93 (95%CI 1.79, 2.07). Admission RDW independently predicted long-term mortality aOR 1.04 (95%CI 1.02, 1.05). Median survival post-admission was 189 months. For those with admission RDW in Q5, observed survival half-life was 133 months-this represents a shortfall of 5.7 life years (33.9%). In conclusion, admission RDW independently predicts 30-day in-hospital and long-term mortality.

Prognostic value and clinical utility of NT-proBNP in acute emergency medical admissions

BACKGROUND

NT-proB-type natriuretic peptide (NT-proBNP) is a frequently utilized test in congestive cardiac failure. There is little data on its utility in unselected emergency medical admissions.

AIM

This study aims to investigate the clinical utility and prognostic value of NT-proBNP in emergency medical admissions and to determine whether such testing influenced downstream investigations and length of stay (LOS).

METHODS

We report on NT-proBNP tests performed in emergency medical admissions in a 2005/2006 and subsequent 7-year (2014-2020) retrospective cohort. We assessed 30-day in-hospital mortality with a multivariable logistic regression model. The utilization of procedures/services was related to LOS with zero-truncated Poisson regression.

RESULTS

There were 64,212 admissions in 36,252 patients. Patients with a NT-proBNP test were significantly older at 75.3 years vs. 63.0 years and had longer LOS -9.4 days vs. 4.9 days. They had higher acute illness severity and comorbidity scores. Thirty-day in-hospital mortality was higher in those with a NT-proBNP test (8.8%) vs. no request (3.2%). NT-proBNP test level was prognostic in univariate - OR 2.87 (2.61, 3.15), and multivariate analyses - OR 1.40 (1.26, 1.56). Higher NT-proBNP levels predicted higher 30-day in-hospital mortality. Multivariable thirty-day in-hospital mortality was 3.8% (3.6%, 3.9%) for those without a test, increasing to 4.9% (4.7%, 5.2%) for ≥ 250 ng/L and 5.8% (5.8%, 6.3%) for ≥ 3000 ng/L. LOS was linearly related to the total number of procedures/services performed.

CONCLUSION

NT-proBNP is prognostic in emergency medical admissions. Downstream resource utilization differed following an NT-proBNP test; this may reflect different case complexity or the 'uncertainty' surrounding such admissions.

Non-Specific Clinical Presentations are Not Prognostic and do not Anticipate Hospital Length of Stay or Resource Utilization

AIM

To investigate whether a specific (SP) or non-specific (NSP) clinical presentation, predicts prognosis and in-hospital resource utilization in emergency medical admissions.

METHODS

We studied admissions over 5 years (2015-2019) and classified the symptom presentation as SP or NSP. The predictive capacity of the NSP category was related to 30-day in-hospital mortality with a multivariable logistic regression model. Utilization of procedures/services was related to hospital length of stay (LOS) with zero truncated Poisson regression.

RESULTS

There were 39,776 admissions in 23,995 patients. A NSP occurred in 18.2% of our top 20 clinical presentations; the top five being shortness of breath (12.8%), 'unwell' (7.1%), collapse (4.1%), abdominal pain (3.6%) and headache (2.7%). Baseline demographic characteristics were similar and unrelated to type of presentation; the model adjusted mortality by SP 4.0% (95% CI: 3.8%, 4.2%) or NSP 3.9% (95% CI: 3.5%, 4.4%) was identical. LOS was a dependant quantitative function of procedures/services undertaken; for the top two presentations of shortness of breath (SP) or unwell (NSP) there was no relationship between a SP or NSP presentation and hospital utilization of procedures/services or LOS.

CONCLUSION

Our data suggest no utility for a categorisation of presentations as specific or non-specific in terms of provision of prognostic information nor as an indicator of the pattern of hospital investigation or LOS.

Is there excessive troponin testing in clinical practice? Evidence from emergency medical admissions

AIM

To investigate whether excessive high-sensitivity cardiac troponin T (hscTnT) testing, in non-cardiac presentations, increases hospital length of stay (LOS) by driving down-stream investigations.

METHODS

We report on all hscTnT tests in emergency medical admissions, performed over a 9-year period between 2011-2019. Troponin testing frequency in different risk cohorts was determined and related to 30-day in-hospital mortality with a multivariable logistic regression model adjusted for other outcome predictors. Downstream utilization of procedures/services was related to LOS with zero truncated Poisson regression.

RESULTS

There were 66,475 admissions in 36,518 patients. hscTnT was tested in 24.4% of admissions, more frequently in the elderly (>70 years 33.4%, >80 years 35.9%), cardiovascular presentations (33.6%) and in those with high comorbidity (42.2%), and reduced in those with neurologic presentations (20%). A hscTnT request predicted increased 30-day in-hospital mortality OR 3.33 (95% CI: 3.06, 3.64). The univariate odds ratio (OR) of hscTnT test result was 1.45 (95% CI: 1.42, 1.49) and was semi-quantative with worsening outcomes as hscTnT increased. It remained predictive in the fully adjusted model OR 1.17 (95% CI: 1.09, 1.26). LOS was linearly related to the number of procedures/services performed. hscTnT testing did not increase LOS or number of procedures/services CONCLUSION: : A clinical request for hscTnT testing is prognostic and risk categorises. Subsequent resource utilization, if increased, appears an epiphenomenon related to risk categorisation, rather than being driven by inappropriate hscTnT testing.

Air pollution and comorbidity burden influencing acute hospital mortality outcomes in a large academic teaching hospital in Dublin, Ireland: a semi-ecologic analysis

OBJECTIVE

To investigate the extent to which air pollution interacts with comorbidity in determining mortality outcomes of emergency medical admissions.

STUDY DESIGN

Routinely collected data were used to study all emergency medical admissions to an academic teaching hospital in Dublin, Ireland, from 2002 to 2018. Air pollution was measured by particulate matter with aerodynamic diameter ≤10 μm (PM₁₀) and sulphur dioxide (SO₂) levels on the day of admission. Comorbidity Score was measured using a previously derived score.

METHODS

A multivariable logistic regression model was used to relate air pollutant levels, Comorbidity Scores, and their interaction to 30-day in-hospital mortality.

RESULTS

There were 102,483 admissions in 58,127 patients over 17 years. Both air pollutant levels and Comorbidity Score were associated with 30-day in-hospital mortality. On admission days with PM₁₀ levels above the median, mortality was higher (Odds ratio [OR] 1.09; 95% confidence interval [CI] 1.06, 1.18) at 11.2% (95% CI 10.5, 12) compared with 10.4% (95% CI 10, 10.7) on days when PM₁₀ levels were below the median. On admission days with SO₂ levels above the median, mortality was higher (OR 1.13; 95% CI 1.10, 1.16) at 12.2% (95% CI 11.4, 13) compared with 10.7% (95% CI 10.3, 11.1) on days when SO₂ levels were below the median. Comorbidity Score was strongly associated with mortality (mortality rate of 8.9% for those with a 6-point score vs mortality rate of 30.3% for those with a 16-point score). There was limited interaction between air pollutant levels and Comorbidity Score.

CONCLUSION

Both air pollution levels on the day of admission and Comorbidity Score were associated with 30-day in-hospital mortality. However, there was limited interaction between these two factors.

Comparative influence of Acute Illness Severity and comorbidity on mortality

BACKGROUND

The extent to which illness severity and comorbidity determine the outcome of an emergency medical admission is uncertain. We aim to quantitate the relative effect of these factors on mortality.

METHODS

We evaluated all emergency medical admission to our institution between 2002 and 2018. We derived an Acute Illness Severity Score (AISS) and Comorbidity Score from admission data and International Classification of Diseases codings. We employed a multivariable logistic regression model to relate both to 30-day in-hospital mortality.

RESULTS

There were 113,807 admissions in 58,126 patients. Both AISS, Odds Ratio (OR) 4.4 (95%CI 3.5, 5.5), and Comorbidity Score, OR 1.91 (95%CI 1.67, 2.18), independently predicted 30-day in-hospital mortality. The two highest AISS risk groups encompassed 46.5% of admissions with predicted mortality of 5.9% (95%CI 5.7%, 6.1%) and 14.4% (95%CI 13.9%, 14.8%) respectively. Comorbidity Score >=10 occurred in 17.9% of admissions with a predicted mortality of 13.3%. AISS and Comorbidity Score interacted to adversely influence mortality; the threshold effect for Comorbidity Score was reduced at high levels of AISS.

CONCLUSION

High AISS and Comorbidity Scores were predictive of 30-day in-hospital mortality and relatively common in emergency medical admissions. There is a strong interaction between the two scores.

Construct validity of acute morbidity as a novel outcome for emergency patients

OBJECTIVES

Validation of acute morbidity as a novel outcome in emergency medicine.

METHODS

Construct validity of acute morbidity was established by comparison to other outcomes using linear and logistic regression models.

RESULTS

Data of 4608 patients were analysed. 1869 patients (40.6%) fulfilled the criteria for acute morbidity. Using multivariate analyses, acute morbidity was associated with outcomes such as hospitalisation (OR: 11, 95%-CI 9-13), mortality (OR 15, 95%-CI 6-49), and ICU admission (OR: 46, 95%-CI 25-96). Reliability of the construct "acute morbidity" was estimated using Cohens Kappa, which was 0.96 for intra-rater and 0.94 for inter-rater reliability.

CONCLUSION

Reliability of the framework for acute morbidity was high. Construct validity was shown by associations with hospitalisation, mortality, and ICU admission.

Meaningful Use and Hospital Performance on Post-Acute Utilization Indicators

OBJECTIVES

To examine trends in hospital post-acute utilization indicators and to determine whether improvement in these indicators is associated with attesting to meaningful use (MU).

DATA SOURCES

Medicare claims-based, repeated measures on 30-day hospital-wide all-cause readmission and emergency department (ED) utilization rates for 160 short-stay hospitals (2009-2012); Medicare EHR Incentive Program Payments files (2011-2012); and other hospital and market data.

STUDY DESIGN

Interrupted time series with concurrent comparison group.

PRINCIPAL FINDINGS

Propensity score-weighted multilevel models for change demonstrate that 30-day readmission rates (unadjusted) fell from 13.4 percent in 2009 to 12.1 percent in 2012. Similarly, 30-day ED utilization declined from 18.9 percent to 17.3 percent during the same period. However, MU and non-MU hospitals were indistinguishable vis-à-vis performance. Controlling for hospital and market characteristics, MU was unrelated to 30-day readmission. In contrast, 30-day ED utilization deteriorated.

CONCLUSIONS

Hospitals with MU Stage 1 designation did not show significantly higher improvement on post-acute utilization compared to their counterparts without. To achieve gains in quality and safety, potentially associated with EHRs, and to advance care coordination and patient engagement, the regulators should strengthen accountability by linking comprehensive, outcomes-based performance measures to specific MU objectives.

Emergency department visits in older people: pattern of use, contributing factors, geographical differences and outcomes

AIMS

To assess the pattern of use of Emergency Departments (EDs), factors contributing to the visits, geographical distribution and outcomes in people aged 65 years or more living in the Italian Lombardy Region in 2012.

METHODS

Based on an administrative database the study population was divided into groups according to the number of ED visits. A multinomial logistic regression model was performed to compare the characteristics of each group. The Getis-Ord's G statistic was used to evaluate the clusters of high and low visit prevalence odd ratios (OR) at district level. To estimate the severity of the disease leading to ED attendance, visits were stratified based on the level of emergency and outcome.

RESULTS

About 2 million older people were included in the analyses: 78 % had no ED visit, 15 % only 1, 7 % 2 or more. Male sex, age 85 years or more, high number of drugs, ED visits and hospital admissions in the previous year and the location of an ED within 10 km from the patient's place were all factors associated with a higher risk to have more ED visits. Clusters of high and low prevalence of visits were found for occasional users. Overall, 83 % of ED visit with a low emergency triage code at admission had as visit outcome discharge at home.

CONCLUSIONS

In older people several variables were associated with an increased risk to have a high number of ED visits. Most of the visits were done for non-urgent problems and significant geographic differences were observed for occasional users.

The Association of Geriatric Syndromes with Hospital Outcomes

BACKGROUND

Frailty, history of dementia (HoD), and acute confusional states (ACS) are common in older patients admitted to hospital.

OBJECTIVE

To study the association of frailty (≥6 points in the Clinical Frailty Scale [CFS]), HoD, and ACS with hospital outcomes, controlling for age, gender, acute illness severity (measured by a Modified Early Warning Score in the emergency department), comorbidity (Charlson Comorbidity Index), and discharging specialty (general medicine, geriatric medicine, surgery).

DESIGN

Retrospective observational study.

SETTING

Large university hospital in England.

PATIENTS

We analyzed 8202 first nonelective inpatient episodes of people aged 75 years and older between October 2014 and October 2015.

MEASUREMENTS

The outcomes studied were prolonged length of stay (LOS ≥10 days), inpatient mortality, delayed discharge, institutionalization, and 30-day readmission. Statistical analyses were based on multivariate regression models.

RESULTS

Independently of controlling variables, prolonged LOS was predicted by CFS ≥6: odds ratio (OR) =1.55; 95% confidence interval [CI], 1.36-1.77; P ⟨ 0.001; HoD: OR = 2.16; 95% CI, 1.79-2.61; P ⟨ 0.001; and ACS: OR = 3.31; 95% CI, 2.64-4.15; P ⟨ 0.001. Inpatient mortality was predicted by CFS ≥6: OR = 2.29; 95% CI, 1.79-2.94; P ⟨ 0.001. Delayed discharge was predicted by CFS ≥6: OR = 1.46; 95% CI, 1.27-1.67; P ⟨ 0.001; HoD: OR = 2.17; 95% CI, 1.80-2.62; P ⟨ 0.001; and ACS: OR = 2.29; 95% CI: 1.83-2.85; P ⟨ 0.001. Institutionalization was predicted by CFS ≥6: OR = 2.56; 95% CI, 2.09-3.14; P ⟨ 0.001; HoD: OR = 2.51; 95% CI, 2.00-3.14; P ⟨ 0.001; and ACS: OR 1.93; 95% CI, 1.46-2.56; P ⟨ 0.001. Readmission was predicted by ACS: OR = 1.36; 95% CI, 1.09-1.71; P = 0.006.

CONCLUSIONS

Routine screening for frailty, HoD, and ACS in hospitals may aid the development of acute care pathways for older adults. Journal of Hospital Medicine 2017;12:83-89.

Deprivation index and dependency ratio are key determinants of emergency medical admission rates

BACKGROUND

Patients from deprived backgrounds have a higher in-patient mortality following an emergency medical admission; there has been debate as to the extent to which deprivation and population structure influences hospital admission rate.

METHODS

All emergency medical admissions to an Irish hospital over a 12-year period (2002-2013) categorized by quintile of Deprivation Index and Dependency Ratio (proportion of population <15 or ≥ 65 years) from small area population statistics (SAPS), were evaluated against hospital admission rates. Univariate and multivariable risk estimates (Odds Ratios (OR) or Incidence Rate Ratios (IRR)) were calculated, using logistic or zero truncated Poisson regression as appropriate.

RESULTS

66,861 admissions in 36,214 patients occured during the study period. The Deprivation Index quintile independently predicted the admission rate/1000 population, Q1 9.4 (95%CI 9.2 to 9.7), Q2 16.8 (95%CI 16.6 to 17.0), Q3 33.8 (95%CI 33.5 to 34.1), Q4 29.6 (95%CI 29.3 to 29.8) and Q5 45.4 (95%CI 44.5 to 46.2). Similarly the population Dependency Ratio was an independent predictor of the admission rate with adjusted predicted rates of Q1 20.8 (95%CI 20.5 to 21.1), Q2 19.2 (95%CI 19.0 to 19.4), Q3 27.6 (95%CI 27.3 to 27.9), Q4 43.9 (95%CI 43.5 to 44.4) and Q5 34.4 (95%CI 34.1 to 34.7). A high concurrent Deprivation Index and Dependency Ratio were associated with very high admission rates.

CONCLUSION

Deprivation Index and population Dependency Ratio are key determinants of the rate of emergency medical admissions.