Probability Of Hospital Death Research Articles

A comparison of two frameworks for multi-state modelling, applied to outcomes after hospital admissions with COVID-19.

We compare two multi-state modelling frameworks that can be used to represent dates of events following hospital admission for people infected during an epidemic. The methods are applied to data from people admitted to hospital with COVID-19, to estimate the probability of admission to intensive care unit, the probability of death in hospital for patients before and after intensive care unit admission, the lengths of stay in hospital, and how all these vary with age and gender. One modelling framework is based on defining transition-specific hazard functions for competing risks. A less commonly used framework defines partially-latent subpopulations who will experience each subsequent event, and uses a mixture model to estimate the probability that an individual will experience each event, and the distribution of the time to the event given that it occurs. We compare the advantages and disadvantages of these two frameworks, in the context of the COVID-19 example. The issues include the interpretation of the model parameters, the computational efficiency of estimating the quantities of interest, implementation in software and assessing goodness of fit. In the example, we find that some groups appear to be at very low risk of some events, in particular intensive care unit admission, and these are best represented by using ‘cure-rate’ models to define transition-specific hazards. We provide general-purpose software to implement all the models we describe in the flexsurv R package, which allows arbitrarily flexible distributions to be used to represent the cause-specific hazards or times to events.

Perioperative Management and Outcomes of Hip Fracture Patients with Advanced Chronic Kidney Disease.

To explore clinical characteristics, perioperative management and outcomes of Hip Fracture patients with advanced Chronic Kidney Disease (HF-aCKD) compared to the general Hip Fracture population without aCKD (HF-G) within a large volume tertiary hospital in Western Australia. Retrospective chart review of patients admitted with hip fracture (HF) to a single large volume tertiary hospital registered on Australian and New Zealand Hip Fracture Registry (ANZHFR). We compared baseline demographic and clinical frailty scale (CFS) of HF-aCKD (n = 74), defined as eGFR < 30mls/min/1.73m2, with HF-G (n = 452) and determined their outcomes at 120days. We identified 74 (6.97%) HF patients with aCKD. General demographics were similar in HF-aCKD and HF-G populations. 120-days mortality for HF-aCKD was double that of HF-G population (34% vs 17%, P = .001). For dialysis patients, 120-days mortality was triple that of HF-G population (57%). Except for the fit category of HF-aCKD group, higher CFS was associated with higher 120-days mortality in both groups. Of all HF-aCKD patients, 96% had operative intervention and 48% received blood transfusion. There were no new starts to dialysis peri-operatively. Each point reduction in eGFR below 12mL/min/1.73m2 was associated with 3% increased probability of death in hospital. 120-days mortality was double in HF-aCKD and triple in HF-dialysis that of the HF-G within our institution. Clinical frailty scale can be useful in predicting mortality after HF in frail aCKD patients. High rate of blood transfusions was observed in HF-aCKD group. Further studies with larger HF-aCKD numbers are required to explore these associations in detail.

Hospital mortality in hematogenous vertebral osteomyelitis

Objective. To analyze lethal outcomes in patients with hematogenous vertebral osteomyelitis.Material and Methods. Study design: retrospective analysis of medical records. A total of 209 medical records of inpatients who underwent treatment for hematogenous vertebral osteomyelitis in 2006–2017 were analyzed. Out of them 68 patients (32.5 %) were treated conservatively, and 141 (67.5 %) – surgically. The risk factors for lethal outcomes were studied for various methods of treatment, and a statistical analysis was performed.Results. Hospital mortality (n = 9) was 4.3 %. In patients who died in hospital, average time for diagnosis making was 4 times less (p = 0.092). The main factors affecting mortality were diabetes mellitus (p = 0.033), type C lesion according to the Pola classification (p = 0.014) and age over 70 years (p = 0.006). To assess the relationship between hospital mortality and the revealed differences between the groups, a regression analysis was performed, which showed that factors associated with mortality were Pola type C.4 lesions (OR 9.73; 95 % CI 1.75–54.20), diabetes mellitus (OR 5.86; 95 % CI 1.14–30.15) and age over 70 years (OR 12.58; 95 % CI 2.50–63.34). The combination of these factors increased the likelihood of hospital mortality (p = 0.001). Sensitivity (77.8 %) and specificity (84.2 %) were calculated using the ROC curve. In the group with mortality, the comorbidity index (CCI) was significantly higher (≥4) than in the group without mortality (p = 0.002). With a CCI of 4 or more, the probability of hospital death increases significantly (OR 10.23; 95 % CI 2.06–50.82), p = 0.005. Long-term mortality was 4.3 % (n = 9), in 77.8 % of cases the cause was acute cardiovascular pathology, and no recurrence of vertebral osteomyelitis was detected.Conclusion. Hospital mortality was 4.3 %, and there was no mortality among patients treated conservatively. The main risk factors were diabetes mellitus, type C lesion according to Pola and age over 70 years. There was a significant mutual burdening of these factors (p = 0.001). With CCI ≥4, the probability of death is higher (p = 0.005).

What mediates end-of-life care choices?

In The Lancet Public Health, Joanna Davies and colleagues examine the relationship between socioeconomic position and receipt of hospital-based care towards the end of life for older people,1 showing that lower wealth is associated with increased hospital admissions in the last 2 years of life and a higher probability of death in hospital. This is an important contribution to the literature as it adds weight to the growing understanding of the importance of non-clinical factors such as socioeconomic status to patterns and quality of health care and its usage.

Socioeconomic inequalities in the place of death in urban small areas of three Mediterranean cities

BackgroundDying at home is the most frequent preference of patients with advanced chronic conditions, their caregivers, and the general population. However, most deaths continue to occur in hospitals. The objective of this study was to analyse the socioeconomic inequalities in the place of death in urban areas of Mediterranean cities during the period 2010–2015, and to assess if such inequalities are related to palliative or non-palliative conditions.MethodsThis is a cross-sectional study of the population aged 15 years or over. The response variable was the place of death (home, hospital, residential care). The explanatory variables were: sex, age, marital status, country of birth, basic cause of death coded according to the International Classification of Diseases, 10th revision, and the deprivation level for each census tract based on a deprivation index calculated using 5 socioeconomic indicators. Multinomial logistic regression models were adjusted in order to analyse the association between the place of death and the explanatory variables.ResultsWe analysed a total of 60,748 deaths, 58.5% occurred in hospitals, 32.4% at home, and 9.1% in residential care. Death in hospital was 80% more frequent than at home while death in a nursing home was more than 70% lower than at home. All the variables considered were significantly associated with the place of death, except country of birth, which was not significantly associated with death in residential care. In hospital, the deprivation level of the census tract presented a significant association (p < 0.05) so that the probability of death in hospital vs. home increased as the deprivation level increased. The deprivation level was also significantly associated with death in residential care, but there was no clear trend, showing a more complex association pattern. No significant interaction for deprivation level with cause of death (palliative, not palliative) was detected.ConclusionsThe probability of dying in hospital, as compared to dying at home, increases as the socioeconomic deprivation of the urban area of residence rises, both for palliative and non-palliative causes. Further qualitative research is required to explore the needs and preferences of low-income families who have a terminally-ill family member and, in particular, their attitudes towards home-based and hospital-based death.

Determinants of place of death for end-stage cancer patients: evidence from China.

To determine factors influence place of death (POD) for end-stage cancer patients and investigate how the healthcare utilization mediates on the effect of socioeconomic status (SES) on POD. A population-based, retrospective study from July 2015 to June 2017. Yichang, China. 894 end-stage cancer patients. POD. Patients of hospital death experience more inpatient hospitalization services (IHS) and emergency department visits. Patients enrolled in the New Rural Cooperative Medical Scheme (OR = 7.60, P < 0.001) and Urban Employee Basic Medical Insurance (OR = 28.0, P < 0.001) have higher rates of hospital death than those in the Urban Resident-based Basic Medical Insurance. Living with spouse (OR = 1.72, P = 0.019) and receiving higher education (OR = 1.92, P = 0.004), increase the likelihood of hospital death by 72% and 92%, respectively. The probability of hospital death will increase by 14% and decrease by 4% per IHS and outpatient services occur, respectively. Outpatient services (Z = -2.28, P < 0.001), and IHS (Z = 2.17, P < 0.001) mediate 1.81% and 1.89%, respectively, of the effect of health insurance on POD. The overall effect of the mediators is non-statistically significant (Z = 0.09, P = 0.825). POD is mainly driven by SES. The relationship between health insurance and POD is partly mediated by outpatient services and IHS, respectively. The results corroborated that hospital and home services should be coherently bridged. Furthermore, benefit packages for end-stage cancer patients could be redesigned.

Reduced Food Intake Is Associated with Mortality in Hospitalized Patients (P12-026-19)

ObjectivesInadequate food intake is associated with mortality in hospitalized patients. Even with the emphasis on malnutrition-related dangers, investigating all hospitalized patients in Korea for food intake and factors affecting food intake is difficult but we need to food intake in hospitalized patients for efficient nutrition management. Therefore, A large-scale nutrition investigation was conducted to examine food intake in all hospitalized patients.We aimed to examine food intakes in hospitalized patients in Korea, to evaluate the relationship between inadequate food intake and mortality, and to ascertain mortality-related risk factors. MethodsA one-day cross-sectional study on food intake and mortality was conducted on 1074 hospitalized adult patients on June 23, 2011.Food intake, assessed using questionnaires, was divided into four categories: 0%, 25%, 50%, and 100%. We collected disease-related factors and meal-related factors from electronic medical records. We also evaluated "length of hospital stay" and "probability of death in hospital" using a post-survey after 60 days. We identified risk factors for mortality using multivariate analysis.Gray’s test was adopted for survival analysis of “probability of death in hospital” according to food intake. ResultsWe included 1074 patients from 56 wards, including the intensive care unit. The following characteristics were observed: average age, 57 years; average body mass index, 22.8 kg/m2; and sex ratio, 55.2% male. On survey day, 35.4% patients ate a whole meal, 56.4% ate half of the meal, 20.2% ate quarter of the meal, and 8.2% ate nothing. Patients who consumed nothing had a 7.2-times higher mortality risk (HR) than patients who consumed a whole meal (P = 0.001). The mortality risk increased by 2.9 (P = 0.056) and 2.6 times (P = 0.102) in patients who consumed 50% and 25% of the meal, respectively. Risk factors contributing to mortality risk included heart/circulation (HR 18.026, P = 0.005), cancer (HR 17.990, P < 0.001), decreased consumption than last week (HR 10.120, P = 0.027), weight loss within 3 months (HR 4.820, P = 0.010), walking inability (HR 4.245, P = 0.020), walking only with assistance (HR 3.358, P = 0.010), and myocardial infarction (HR 2.571, P = 0.022). Contrastingly, “nutrition support” resulted in a significantly reduced mortality risk (HR 0.332, P = 0.015). ConclusionsReduced food intake can affect mortality rates in hospitalized patients. Based on factors significantly impacting mortality, patient groups can be selected for intensive nutrition care. Funding Sourcesnone. Supporting Tables, Images and/or Graphs▪

Survival benefits among patients with end-stage renal disease receiving dialysis versus no dialysis.

55 Background: Retrospective study of the outcomes of Stage IV Cancer Patients with End Stage Renal disease Receiving Dialysis and comparing the survival benefits among the patients who received Hemodialysis versus who didn’t at a Comprehensive Cancer Center. Methods: Retrospectively review all patients admitted to MDACC from 2005-2014 and diagnosed with stage IV solid tumor admitted with Acute kidney injury and a nephrology consult were included (N = 903) 19.82% received dialysis (n = 179) versus 80.17% did not receive dialysis (n = 724). The main endpoints of the study to be evaluated would be survival &amp; hospital death. Results: Kaplan-Meier overall survival curve of the dialysis group vs the non-dialysis p-value &lt; 0.001 indicating that patients who underwent dialysis had a poorer overall survival. This was further explored in a multivariate model where at any given time; a patient who underwent a dialysis is 1.318 times as likely to experience a death as a patient who did not undergo a dialysis with the same other clinical factors. The association between dialysis and death in hospital was explored using multivariate model indicated dialysis (odds ratio = 3.339, 95% CI (2.212, 5.04), p &lt; 0.001, ICU (odds ratio = 2.624, 95% CI (1.905, 3.613), p &lt; 0.001), and a lower serum albumin level (p &lt; 0.001) are factors associated with a higher probability of death in hospital. Matched sample where evaluated using propensity score matching &amp; Weighting and the dialysis group still shows a significant negative effect on Overall Survival with (p = 0.003). ICU patients Subset Analysis &amp; SOFA Score: 49.9% had ICU visits in our database with SOFA scores (37.9% were dialyzed). Using propensity score weighting method, the estimated hazard rate is 0.892 with 95% CI (0.775, 1.027) and p = 0.113, showing NO evidence that dialysis is a prognostic factor for ICU patients after controlling SOFA score. In the ICU cohort, the median survival time for dialysis 18 days (95% CI: 14-23) and no dialysis 19 days (95% CI: 15-25). Conclusions: We can conclude that Dialysis when offered to advanced cancer patients does not improve mortality. Creating a hemodialysis prognostic tool relevant to cancer population would be of great benefit to Nephrologists, Oncologist and patients.

Dying among older adults in Switzerland: who dies in hospital, who dies in a nursing home?

BackgroundInstitutional deaths (hospitals and nursing homes) are an important issue because they are often at odds with patient preference and associated with high healthcare costs. The aim of this study was to examine deaths in institutions and the role of individual, regional, and healthcare supply characteristics in explaining variation across Swiss Hospital Service Areas (HSAs).MethodsRetrospective study of individuals ≥66 years old who died in a Swiss institution (hospital or nursing homes) in 2010. Using a two-level logistic regression analysis we examined the amount of variation across HSAs adjusting for individual, regional and healthcare supply measures. The outcome was place of death, defined as death in hospital or nursing homes.ResultsIn 2010, 41,275 individuals ≥66 years old died in a Swiss institution; 54 % in nursing homes and 46 % in hospitals. The probability of dying in hospital decreased with increasing age. The OR was 0.07 (95 % CI: 0.05–0.07) for age 91+ years compared to those 66–70 years. Living in peri-urban areas (OR = 1.06 95 % CI: 1.00–1.11) and French speaking region (OR = 1.43 95 % CI: 1.22–1.65) was associated with higher probability of hospital death. Females had lower probability of death in hospital (OR = 0.54 95 % CI: 0.51–0.56). The density of ambulatory care physicians (OR = 0.81 95 % CI: 0.67–0.97) and nursing homes beds (OR = 0.67 95 % CI: 0.56–0.79) was negatively associated with hospital death. The proportion of dying in hospital varied from 38 % in HSAs with lowest proportion of hospital deaths to 60 % in HSAs with highest proportion of hospital deaths (1.6-fold variation).ConclusionsWe found evidence for variation across regions in Switzerland in dying in hospital versus nursing homes, indicating possible overuse and underuse of end of life (EOL) services.Electronic supplementary materialThe online version of this article (doi:10.1186/s12904-016-0156-x) contains supplementary material, which is available to authorized users.

Effekt des Zeitpunkts der Aufnahme im Krankenhaus auf das Behandlungsergebnis nach schwerem Schädelhirntrauma in Österreich

The goal of this study was to compare outcomes of patients with severe traumatic brain injury (TBI) who had been admitted either during workdays from 7a.m. to 7p.m. ("regular service") or during any other time ("on-call service"). Between March 2002 and April 2012, 17 Austrian centers enrolled TBI patients into two observational studies that focused on effects of guideline compliance (n = 400) and on prehospital and early hospital management (n = 777), respectively. Data on trauma severity, clinical status, treatment, and outcomes were collected prospectively. All patients with severe TBI (Glasgow Coma Scale score < 9) were selected for this analysis. Secondary transfers and patients with unsurvivable injuries were excluded. The International Mission for Prognosis and Analysis of Clinical Trials in TBI core model was used to estimate probabilities of hospital death and unfavorable long-term outcome (Glasgow Outcome Scale score < 4). Based on time of arrival, patients were assigned to groups "regular service" or "on-call service." Data from 852 patients were analyzed (413 "regular," 439 "on-call service"). "On-call" patients were younger (45 vs. 51 years, P < 0.001) and had a higher rate of alcohol intoxication (41 vs. 11 %, P < 0.001). Trauma severity was comparable; the probabilities of death and unfavorable outcome were identical. There were no differences regarding computed tomography findings or treatment. Hospital mortality (24 vs. 28 %, P = 0.191) and rate of patients with unfavorable outcome at 6 months (43 vs. 48 %, P = 0.143) were comparable. In Austria, the time of hospital admission has no influence on outcomes after severe TBI.

Sensitivity of hospital-based surveillance for severe disease: a geographic information system analysis of access to care in Kilifi district, Kenya

To explore the relationship between homestead distance to hospital and access to care and to estimate the sensitivity of hospital-based surveillance in Kilifi district, Kenya. In 2002-2006, clinical information was obtained from all children admitted to Kilifi District Hospital and linked to demographic surveillance data. Travel times to the hospital were calculated using geographic information systems and regression models were constructed to examine the relationships between travel time, cause-specific hospitalization rates and probability of death in hospital. Access to care ratios relating hospitalization rates to community mortality rates were computed and used to estimate surveillance sensitivity. The analysis included 7200 admissions (64 per 1000 child-years). Median pedestrian and vehicular travel times to hospital were 237 and 61 minutes, respectively. Hospitalization rates decreased by 21% per hour of travel by foot and 28% per half hour of travel by vehicle. Distance decay was steeper for meningitis than for pneumonia, for females than for males, and for areas where mothers had less education on average. Distance was positively associated with the probability of dying in hospital. Overall access to care ratios, which represent the probability that a child in need of hospitalization will have access to care at the hospital, were 51-58% for pneumonia and 66-70% for meningitis. In this setting, hospital utilization rates decreased and the severity of cases admitted to hospital increased as distance between homestead and hospital increased. Access to hospital care for children living in remote areas was low, particularly for those with less severe conditions. Distance decay was attenuated by increased levels of maternal education. Hospital-based surveillance underestimated pneumonia and meningitis incidence by more than 45% and 30%, respectively.

Is the SAPS II score valid in surgical intensive care unit patients?

We investigated the performance of the simplified acute physiology score II (SAPS II) in a large cohort of surgical intensive care unit (ICU) patients and tested the hypothesis that customization of the score would improve the uniformity of fit in subgroups of surgical ICU patients. Retrospective analysis of prospectively collected data from all 12,938 patients admitted to a postoperative ICU between January 2004 and January 2009. Probabilities of hospital death were calculated for original and customized (C1-SAPS II and C2-SAPS II) scores. A priori subgroups were defined according to age, probability of death according to the SAPS II score, ICU length of stay (LOS), surgical procedures and type of admission. The median ICU LOS was 1 (1-3) day. ICU and hospital mortality rates were 5.8% and 10.3%, respectively. Discrimination of the SAPS II was moderate [area under receiver operating characteristic curve (aROC) = 0.76 (0.75-0.78)], but calibration was poor. This model markedly overestimated hospital mortality rates [standardized mortality rate: 0.35 (0.33-0.37)]. First-level customization (C1-SAPS II) did not improve discrimination in the whole cohort or the subgroups, but calibration improved in some subgroups. Second-level customization (C2-SAPS II) improved discrimination in the whole cohort [aROC = 0.82 (0.79-0.85)] and most of the subgroups (aROC range 0.65-86). Calibration in this model (C2-SAPS II) improved in the whole cohort and in subgroups except in patients with ICU LOS 4-14 days and those undergoing neuro- or gastrointestinal surgery. In this large cohort of surgical ICU patients, performance of the original SAPS II model was generally poor. Although second-level customization improved discrimination and calibration in the whole cohort and most of the subgroups, it failed to simultaneously improve calibration in the subgroups stratified according to the type of surgery, age or ICU LOS.

Case-mix-adjusted length of stay and mortality in 23 Finnish ICUs

To create a tool for benchmarking intensive care units (ICUs) with respect to case-mix adjusted length of stay (LOS) and to study the association between clinical and economic measures of ICU performance. Observational cohort study. Twenty-three ICUs in Finland. A total of 80,854 consecutive ICU admissions during 2000-2005, of which 63,304 met the inclusion criteria. None. Linear regression was used to create a model that predicted ICU LOS. Simplified Acute Physiology Score (SAPS) II, age, disease categories according to Acute Physiology and Chronic Health Evaluation III, single highest Therapeutic Intervention Scoring System score collected during the ICU stay and presence of other ICUs in the hospital were included in the model. Probabilities of hospital death were calculated using SAPS II, age, and disease categories as covariates. In the validation sample, the created model accounted for 28% of variation in ICU LOS across individual admissions and 64% across ICUs. The expected ICU LOS was 2.53 +/- 2.24 days and the observed ICU LOS was 3.29 +/- 5.37 days, P < 0.001. There was no association between the mean observed - mean expected ICU LOS and standardized mortality ratios of the ICUs (Spearman correlation 0.091, P = 0.680). We developed a tool for the assessment of resource use in a large nationwide ICU database. It seems that there is no association between clinical and economic quality indicators.

Validation of the 2000 Bernstein-Parsonnet Score Versus the EuroSCORE as a Prognostic Tool in Cardiac Surgery

Intradepartmental and interdepartmental benchmarking requires scoring systems with reliability (calibration) and stability over the complete spectrum of periprocedural risk. The aim of this single-center study was to assess the performance of the 2000 Bernstein-Parsonnet risk stratification model in cardiac surgery, by itself and against the EuroSCORE. A prospective observational design was used. The study group consisted of 1,639 consecutive patients of mean age 64.6 +/- 12.04 years who underwent elective or emergency cardiac surgery from January 2003 to June 2004. The probabilities of hospital death were estimated with the 2000 Bernstein-Parsonnet and EuroSCORE algorithms. The correlation of predicted and observed mortality was compared between the two models, and score validity was assessed by calculating the area under the receiver operating characteristic (ROC) curve. The patients were stratified into five risk groups according to their scores in the two models. For the 2000 Bernstein-Parsonnet model, findings were as follows: score 0-10: predicted mortality 0%-2.2%, observed mortality 0.6%; score 10.5-20: predicted 2.3%-4.7%, observed 2.3%; score 20.5-30: predicted 4.8%-10%, observed 6.7%; score 30.5-40: predicted 10.1%-23%, observed 11.5%; and score greater than 40: predicted 23.1%-80%, observed 29.9%. For the EuroSCORE, findings were as follows: score 0%-2%: predicted mortality 1.1%, observed mortality 0.6%; score 3%-5%: predicted 2.1%, observed 3.0%; score 6%-8%: predicted 4.1%, observed 3.5%; score 9-11: predicted 7.6%, observed 6.6.%; and score greater than 12: predicted 13.8%, observed 14.0%. There was good agreement between the observed and expected number of deaths, with both models. The area under the ROC curve was higher for the Bernstein-Parsonnet model (0.83, odds ratio [OR] 2.01, 95% confidence interval [CI] 1.75-2.31, p < 0.0001) than for the EuroSCORE (0.73, OR 1.05, 95% CI 1.04-1.07, p < 0.001). The 2000 Bernstein-Parsonnet model is a simple, objective system for the estimation of hospital mortality in patients undergoing cardiac surgery, with slightly higher calibration and discrimination than the EuroSCORE additive model.

Hospital costs in patients receiving prolonged mechanical ventilation: does age have an impact?

The aging of the population is one of the causes of the increase in healthcare costs in the past few decades. It is controversial whether chronological age alone should be used in making healthcare decisions. To determine the association between age and hospital costs in patients receiving mechanical ventilation (MV). Prospective, observational study. Intensive care units at a teaching hospital. A total of 813 adults who received prolonged (> or =48 hrs) mechanical ventilation. None. Severity of illness, comorbidities, length of stay, hospital costs, and mortality. We evaluated the independent association of age with hospital costs using linear regression. Mean (+/-sd) age of patients was 60.4 +/- 18.8 yrs. Median Acute Physiology Chronic Health Evaluation III score and probability of hospital death at intensive care unit admission were 64 and 0.31, respectively. Hospital mortality was 36%. Median total hospital costs and daily costs were $ 56,056 and $2,655 US dollars, respectively. Older age was associated with lower total hospital costs after controlling for sex, intensive care unit type, severity of illness, length of stay, insurance type, resuscitation status, and survival. Hospital costs were significantly less in older patients in all cost departments examined, except for respiratory care and intensive care unit room costs. Daily and total hospital costs were lower in older patients. Decreased hospital resource use in older patients may be related to a preference for less aggressive care by older patients and their families or by healthcare providers.

Place of death: analysis of cancer deaths in part of North West England.

Relatively little work of a detailed geographical nature has been undertaken on the distribution of place of death. In particular, given evidence that most cancer patients would prefer to die at home there is a need to examine the extent to which this preference is met differentially from place to place. Using data on cancer deaths for a single Health Authority in North West England we conducted both small area and individual analyses of place of death, using binomial and binary logistic regression models, respectively. Results from the small area analysis show that in more deprived areas cancer patients are more likely to die in hospital or hospice, and less likely to die at home, but that the effect disappears for home and hospice deaths once other factors are controlled for. At the individual level, the probability of death at home decreases among those living in deprived areas, whereas the probability of death in hospital increases as area deprivation increases. Age, gender, type of cancer, and proximity to hospital or hospice all have some effect on the probability of dying in a particular setting. There is significant place-to-place variation in place of death among cancer patients in part of North West England. However, studies of place of death among cancer patients need to consider the full range of settings and, if examining the impact of deprivation or social class, need to adjust for other factors, including proximity to different settings.

Intensive care unit admissions with cirrhosis: risk-stratifying patient groups and predicting individual survival.

Prognosis for acutely ill patients with cirrhosis is influenced by the severity of hepatic abnormalities and by dysfunction of other organ systems. The purpose of this study was to examine the usefulness of the Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) prognostic system for risk-stratifying groups of intensive care unit (ICU) patients with cirrhosis and in predicting individual survival. We used data for 17,440 ICU admissions at 40 American hospitals to select 117 of the 537 patients with a history of cirrhosis who were ventilated on ICU day 1, a group known to have a high mortality rate. We then calculated each patient's probability of hospital death on ICU days 1 through 7, using seven previously validated multivariate equations. Hospital mortality was 63% for the 117 study patients. The most important determinants of risk for hospital death on ICU day 1 were the acute physiology score of APACHE III, ICU admission diagnosis, and operative status. Daily changes in the acute physiology score caused a rise or fall in the probability of hospital mortality and was useful in assessing individual response to therapy. APACHE III accurately risk stratifies critically ill patients with cirrhosis because it accounts for many of the factors known to influence prognosis. This capability can be used to assess severity of illness and risk-stratify patients with cirrhosis during clinical trials. Daily prognostic estimates based on physiological changes over time reflect patient response and can help physicians to assess the incremental benefit of therapy.

Intensive care unit admissions with cirrhosis: Risk-stratifying patient groups and predicting individual survival

Prognosis for acutely ill patients with cirrhosis is influenced by the severity of hepatic abnormalities and by dysfunction of other organ systems. The purpose of this study was to examine the usefulness of the Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) prognostic system for risk-stratifying groups of intensive care unit (ICU) patients with cirrhosis and in predicting individual survival. We used data for 17,440 ICU admissions at 40 American hospitals to select 117 of the 537 patients with a history of cirrhosis who were ventilated on ICU day 1, a group known to have a high mortality rate. We then calculated each patient's probability of hospital death on ICU days 1 through 7, using seven previously validated multivariate equations. Hospital mortality was 63% for the 117 study patients. The most important determinants of risk for hospital death on ICU day 1 were the acute physiology score of APACHE III, ICU admission diagnosis, and operative status. Daily changes in the acute physiology score caused a rise or fall in the probability of hospital mortality and was useful in assessing individual response to therapy. APACHE III accurately risk stratifies critically ill patients with cirrhosis because it accounts for many of the factors known to influence prognosis. This capability can be used to assess severity of illness and risk-stratify patients with cirrhosis during clinical trials. Daily prognostic estimates based on physiological changes over time reflect patient response and can help physicians to assess the incremental benefit of therapy. (Hepatology 1996 Jun;23(6):1393-401)

Epidemiology in Anaesthesia

Data on 13 043 operations were abstracted from 108 878 anaesthetic records in the Cardiff Anaesthetic Record System. Nine groups of surgical operations with high mortality in hospital were identified. These records were analysed using a logistic regression model to determine the probability of death in hospital. Variables included in the model were age, sex, elective or emergency operation and five intercurrent diseases. The significance of individual variables was assessed: age was the most important single variable in six of the nine operation groups. The model was used to calculate the risk of death in hospital for individuals with specific characteristics. With further development, this technique might be used by clinicians at other centres to predict the outcome for patients under their care.

Surgical results and protocols in the spectrum of tetralogy of Fallot.

Between 1967 and July 1982, 1103 operations were performed for the tetralogy of Fallot of all types with 116 (10.5%) hospital deaths. Eighty-eight hospital deaths (10.5%) occurred in the 836 patients undergoing repair. The incremental risk factors for hospital death after repair include pulmonary arterial problems (p = 0.0002), major associated cardiac anomalies (p less than 0.0001), small size (young age) (p less than 0.0001), and more than one previous operation (p = 0.0004). Absent pulmonary valve is a risk factor (p = 0.04). In patients with pulmonary stenosis, the hospital mortality has decreased with time (p = 0.08), but the incremental risk of a high hematocrit (p = 0.0003) and of transannular patching (p = 0.05) has persisted. In the current era, the risk of repair in patients with pulmonary stenosis is estimated to be 1.6% (70% confidence limits [CL] 0.7% to 3.5%) at age 5 years, and at age 12 months to be 4.1% (CL 2.7% to 6.3%) without a transannular patch and 7.7% (CL a5.3% to 11%) with one. When pulmonary atresia is present, the probability of hospital death after repair when a valved extracardiac conduit is used is estimated to be lowest (5%; CL 2% to 8%) between 5 1/2 and 16 years of age. No deaths occurred among 53 patients with pulmonary stenosis receiving a primary palliative Blalock-Taussig or Gore-Tex shunt, and six deaths (12%) occurred in 51 patients with pulmonary atresia. Serious interim complications (sudden death, brain abscess) after these shunts occurred in two (1.9%; CL 0.6% to 4.5%) of the patients operated on. No iatrogenic pulmonary arterial problems have been recognized. Protocols based on these results are presented. Many of these selective recommendations may become unnecessary if the damaging effects of cardiopulmonary bypass are overcome by future research.