Impact of Missing Physiologic Data on Performance of the Simplified Acute Physiology Score 3 Risk-Prediction Model.

OBJECTIVES: The Simplified Acute Physiology 3 outcome prediction model has a narrow time window for recording physiologic measurements. Our objective was to examine the prevalence and impact of missing physiologic data on the Simplified Acute Physiology 3 model's performance. DESIGN: Retrospective analysis of prospectively collected data. SETTING: Sixty-three ICUs in the Swedish Intensive Care Registry. PATIENTS: Patients admitted during 2011-2014 (n = 107,310). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Model performance was analyzed using the area under the receiver operating curve, scaled Brier's score, and standardized mortality rate. We used a recalibrated Simplified Acute Physiology 3 model and examined model performance in the original dataset and in a dataset of complete records where missing data were generated (simulated dataset). One or more data were missing in 40.9% of the admissions, more common in survivors and low-risk admissions than in nonsurvivors and high-risk admissions. Discrimination did not decrease with one to two missing variables, but accuracy was highest with no missing data. Calibration was best in the original dataset with a mix of full records and records with some missing values (area under the receiver operating curve was 0.85, scaled Brier 27%, and standardized mortality rate 0.99). With zero, one, and two data missing, the scaled Brier was 31%, 26%, and 21%; area under the receiver operating curve was 0.84, 0.87, and 0.89; and standardized mortality rate was 0.92, 1.05 and 1.10, respectively. Datasets where the missing data were simulated for oxygenation or oxygenation and hydrogen ion concentration together performed worse than datasets with these data originally missing. CONCLUSIONS: There is a coupling between missing physiologic data, admission type, low risk, and survival. Increased loss of physiologic data reduced model performance and will deflate mortality risk, resulting in falsely high standardized mortality rates.

Comparing Time-Fixed Mortality Prediction Models and Their Effect on ICU Performance Metrics Using the Simplified Acute Physiology Score 3.

OBJECTIVES: To examine ICU performance based on the Simplified Acute Physiology Score 3 using 30-day, 90-day, or 180-day mortality as outcome measures and compare results with 30-day mortality as reference. DESIGN: Retrospective cohort study of ICU admissions from 2010 to 2014. SETTING: Sixty-three Swedish ICUs that submitted data to the Swedish Intensive Care Registry. PATIENTS: The development cohort was first admissions to ICU during 2011-2012 (n = 53,546), and the validation cohort was first admissions to ICU during 2013-2014 (n = 57,729). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Logistic regression was used to develop predictive models based on a first level recalibration of the original Simplified Acute Physiology Score 3 model but with 30-day, 90-day, or 180-day mortality as measures of outcome. Discrimination and calibration were excellent for the development dataset. Validation in the more recent 2013-2014 database showed good discrimination (C-statistic: 0.85, 0.84, and 0.83 for the 30-, 90-, and 180-d models, respectively), and good calibration (standardized mortality ratio: 0.99, 0.99, and 1.00; Hosmer-Lemeshow goodness of fit H-statistic: 66.4, 63.7, and 81.4 for the 30-, 90-, and 180-d models, respectively). There were modest changes in an ICU's standardized mortality ratio grouping (< 1.00, not significant, > 1.00) when follow-up was extended from 30 to 90 days and 180 days, respectively; about 11-13% of all ICUs. CONCLUSIONS: The recalibrated Simplified Acute Physiology Score 3 hospital outcome prediction model performed well on long-term outcomes. Evaluation of ICU performance using standardized mortality ratio was only modestly sensitive to the follow-up time. Our results suggest that 30-day mortality may be a good benchmark of ICU performance. However, the duration of follow-up must balance between what is most relevant for patients, most affected by ICU care, least affected by administrative policies and practically feasible for caregivers.

Gender differences in outcome and use of resources do exist in Swedish intensive care, but to no advantage for women of premenopausal age.

INTRODUCTION: Preclinical data indicate that oestrogen appears to play a beneficial role in the pathophysiology of and recovery from critical illness. In few previous epidemiologic studies, however, have researchers analysed premenopausal women as a separate group when addressing potential gender differences in critical care outcome. Our aim was to see if women of premenopausal age have a better outcome following critical care and to investigate the association between gender and use of intensive care unit (ICU) resources. METHODS: On the basis of our analysis of 127,254 consecutive Simplified Acute Physiology Score III-scored Swedish Intensive Care Registry ICU admissions from 2008 through 2012, we determined the risk-adjusted 30-day mortality, accumulated nurse workload score and ICU length of stay. To investigate associations with sex, we used logistic regression and multivariate analyses on the entire cohort as well as on two subgroups stratified by median age for menopause (up to and including 45 years and older than 45 years) and six selected diagnostic subgroups (sepsis, multiple trauma, chronic obstructive pulmonary disease, acute respiratory distress syndrome, pneumonia and cardiac arrest). RESULTS: There was no sex difference in risk-adjusted mortality for the cohort as a whole, and there was no sex difference in risk-adjusted mortality in the group 45 years of age and younger. For the group of patients older than 45 years of age, we found a reduced risk-adjusted mortality in men admitted for cardiac arrest. For the cohort as a whole, and for those admitted with multiple trauma, male sex was associated with a higher nurse workload score and a longer ICU stay. CONCLUSIONS: Using information derived from a large multiple ICU register database, we found that premenopausal female sex was not associated with a survival advantage following intensive care in Sweden. When the data were adjusted for age and severity of illness, we found that men used more ICU resources per admission than women did.

Evolution of chronic renal impairment and long-term mortality after de novo acute kidney injury in the critically ill; a Swedish multi-centre cohort study.

INTRODUCTION: Acute Kidney Injury (AKI) is common in critical ill populations and its association with high short-term mortality is well established. However, long-term risks of death and renal dysfunction are poorly understood and few studies exclude patients with pre-existing renal disease, meaning outcome for de novo AKI has been difficult to elicit. We aimed to compare the long-term risk of Chronic Kidney Disease (CKD), End Stage Renal Disease (ESRD) and mortality in critically ill patients with and without severe de novo AKI. METHOD: This cohort study was conducted between 2005 and 2011 in Swedish intensive care units (ICU). Data from 130134 adult patients listed on the Swedish intensive care register-database was linked with other national registries. Patients with pre-existing CKD (4192) and ESRD (1389) were excluded, as were cases (26771) with incomplete data. Patients were classified according to AKI exposure during ICU admission. Outcome in the de novo AKI group was compared to the non-exposed (no-AKI) intensive care control group. Primary outcome was all-cause mortality. Follow-up ranged from one to seven years (median 2.1 years). Secondary outcomes were incidence of CKD and ESRD and median follow-up was 1.3 years. RESULTS: Of 97 782 patients, 5273 (5.4%) had de novo AKI. These patients had significantly higher crude mortality at one (48.4% vs. 24.6%) and five years (61.8% vs. 39.1%) compared to the control group. The first 30% of deaths in AKI patients occurred within 11 days of ICU admission whilst the 30-centile in the no-AKI group died by 748 days. CKD was significantly more common in AKI survivors at one year (6.0% vs. 0.44%) than in no-AKI group (adjusted incidence rate ratio (IRR) 7.6). AKI patients also had significantly higher rates of ESRD at one (2.0% vs. 0.08%) and at five years (3.9% vs. 0.3%) than those in the comparison group (adjusted IRR 22.5). CONCLUSION: This large cohort study demonstrated that de novo AKI is associated with increased short and long-term risk of death. AKI is independently associated with increased risk of CKD and ESRD as compared to an ICU control population. Severe de novo AKI survivors should be routinely followed-up and their renal function monitored.

Reducing night-time discharge from intensive care. A nationwide improvement project with public display of ICU outcomes.

PURPOSE: Discharge from an intensive care unit (ICU) during the night is an independent risk factor for adverse outcomes. A quality improvement project was conducted with the aim of reducing the incidence and the associated mortality after night-time discharge. MATERIALS AND METHODS: ICUs that submitted data to the Swedish Intensive Care Registry (SIR) agreed to appoint night-time discharge as a national quality indicator with detailed public display on the internet of various discharge proportions and outcomes. The registry was then examined for trends during a 10-year period with use of multilevel mixed-effects models. RESULTS: We analysed 163,371 patients who were discharged alive from 70 ICUs to a general ward within the same hospital during 2006-2015. The prevalence of night-time discharge fell from 7.0% (95% CI: 5.2 to 8.7%) in 2006 to 4.9% (95% CI: 4.3 to 5.5%) in 2015 (P = .035 for trend). The original increased risk of death within 30 days after night-time discharge in 2006-2010, OR 1.20 (95% CI: 1.01 to 1.42), disappeared in 2011-2015, OR 1.06 (95% CI: 0.96 to 1.17). CONCLUSIONS: During the 10-year period of the quality improvement project, the annual prevalence and risk of death within 30-days after night-time discharge were reduced. The public display and feedback of audit data could have helped in achieving this.