Evaluating the performance of Australian and New Zealand intensive care units in 2009 and 2010.

The Australian and New Zealand Intensive Care Society Adult Patient Database (ANZICS APD) is one of the largest databases of its kind in the world and collects individual admissions' data from intensive care units (ICUs) around Australia and New Zealand. Use of this database for monitoring and comparing the performance of ICUs, quantified by the standardised mortality ratio, poses several theoretical and computational challenges, which are addressed in this paper. In particular, the expected number of deaths must be appropriately estimated, the ICU casemix adjustment must be adequate, statistical variation must be fully accounted for, and appropriate adjustment for multiple comparisons must be made. Typically, one or more of these issues have been neglected in ICU comparison studies. Our approach to the analysis proceeds by fitting a random coefficient hierarchical logistic regression model for the inhospital death of each patient, with patients clustered within ICUs. We anticipate the majority of ICUs will be estimated as performing 'usually' after adjusting for important clinical covariates. We take as a starting point the ideas in Ohlssen et al and estimate an appropriate null model that we expect these ICUs to follow, taking a frequentist rather than a Bayesian approach. This methodology allows us to rigorously account for the aforementioned statistical issues and to determine if there are any ICUs contributing to the Australian and New Zealand Intensive Care Society database that have comparatively unusual performance. In addition to investigating the yearly performance of the ICUs, we also estimate changes in individual ICU performance between 2009 and 2010 by adjusting for regression-to-the-mean.

Mortality and cost outcomes of elderly trauma patients admitted to intensive care and the general wards of an Australian tertiary referral hospital.

Mortality and cost outcomes of elderly intensive care unit (ICU) trauma patients were characterised in a retrospective cohort study from an Australian tertiary ICU Trauma patients admitted between January 2000 and December 2005 were grouped into three major age categories: aged > or =65 years admitted into ICU (n = 272); aged -65 years admitted into general ward (n = 610) and aged < 65 years admitted into ICU (n = 1617). Hospital mortality predictors were characterised as odds ratios (OR) using logistic regression. The impact of predictor variables on (log) total hospital-stay costs was determined using least squares regression. An alternate treatment-effects regression model estimated the mortality cost-effect as an endogenous variable. Mortality predictors (P < or = 0.0001, comparator: ICU > or = 65 years, ventilated) were: ICU < 65 not-ventilated (OR 0.014); ICU < 65 ventilated (OR 0.090); ICU age > or = 65 not-ventilated (OR 0.061) and ward > or = 65 (OR 0.086); increasing injury severity score and increased Charlson comorbidity index of 1 and 2, compared with zero (OR 2.21 [1.40 to 3.48] and OR 2.57 [1.45 to 4.55]). The raw mean daily ICU and hospital costs in A$ 2005 (US$) for age < 65 and > or = 65 to ICU, and > or = 65 to the ward were; for year 2000: ICU, $2717 (1462) and $2777 (1494); hospital, $1837 (988) and $1590 (855); ward $933 (502); for year 2005: ICU, $3202 (2393) and $3086 (2307); hospital, $1938 (1449) and $1914 (1431); ward $1180 (882). Cost increments were predicted by age < or = 65 and ICU admission, increasing injury severity score, mechanical ventilation, Charlson comorbidity index increments and hospital survival. Mortality cost-effect was estimated at -63% by least squares regression and -82% by treatment-effects regression model. Patient demographic factors, injury severity and its consequences predict both cost and survival in trauma. The cost mortality effect was biased upwards by conventional least squares regression estimation.

Gene profiling for determining pluripotent genes in a time course microarray experiment.

In microarray experiments, it is often of interest to identify genes which have a prespecified gene expression profile with respect to time. Methods available in the literature are, however, typically not stringent enough in identifying such genes, particularly when the profile requires equivalence of gene expression levels at certain time points. In this paper, the authors introduce a new methodology, called gene profiling, that uses simultaneous differential and equivalent gene expression level testing to rank genes according to a prespecified gene expression profile. Gene profiling treats the vector of true gene expression levels as a linear combination of appropriate vectors, for example, vectors that give the required criteria for the profile. This gene profile model is fitted to the data, and the resulting parameter estimates are summarized in a single test statistic that is then used to rank the genes. The theoretical underpinnings of gene profiling (equivalence testing, intersection-union tests) are discussed in this paper, and the gene profiling methodology is applied to our motivating stem-cell experiment.

Assessment of outcome over a 10-year period of patients admitted to a multidisciplinary adult intensive care unit with haematological and solid tumours.

The risk factors for time to mortality, censored at 30 days, of patients admitted to an adult teaching hospital ICU with haematological and solid malignancies were assessed in a retrospective cohort study. Patients, demographics and daily ICU patient data, from admission to day 8, were identified from a prospective computerized database and casenote review in consecutive admissions to ICU with haematological and solid tumours over a 10-year period (1989-99). The cohort, 108 ICU admissions in 89 patients was of mean age (+/-SD) 55+/-14 years; 43% were female. Patient diagnoses were leukaemia (35%), lymphoma (38%) and solid tumours (27%). Median time from hospital to ICU admission was five days (range 0-67). On ICU admission, 50% had septic shock and first day APACHE II score was 28+/-9. Forty-six per cent of patients were ventilated. ICU and 30-day mortality were 39% and 54% respectively. Multivariate Cox model predictors (P<0.05), using only ICU admission day data were: Charlson comorbidity index (CCI), time to ICU admission (days) and mechanical ventilation. For daily data (admission through day 8), predictors were: cohort effect (2nd vs 1st five-year period); CCI; time to ICU admission (days); APACHE II score and mechanical ventilation. Outcomes were considered appropriate for severity of illness and demonstrated improvement over time. Ventilation was an independent outcome determinant. Controlling for other factors, mortality has improved over time (1st vs 2nd five year period). Analysis restricted to admission data alone may be insensitive to particular covariate effects.

Expression profiling reveals functionally important genes and coordinately regulated signaling pathway genes during in vitro angiogenesis.

Angiogenesis is a complex multicellular process requiring the orchestration of many events including migration, alignment, proliferation, lumen formation, remodeling, and maturation. Such complexity indicates that not only individual genes but also entire signaling pathways will be crucial in angiogenesis. To define an angiogenic blueprint of regulated genes, we utilized our well-characterized three-dimensional collagen gel model of in vitro angiogenesis, in which the majority of cells synchronously progress through defined morphological stages culminating in the formation of capillary tubes. We developed a comprehensive three-tiered approach using microarray analysis, which allowed us to identify genes known to be involved in angiogenesis and genes hitherto unlinked to angiogenesis as well as novel genes and has proven especially useful for genes where the magnitude of change is small. Of interest is the ability to recognize complete signaling pathways that are regulated and genes clustering into ontological groups implicating the functional importance of particular processes. We have shown that consecutive members of the mitogen-activated protein kinase and leukemia inhibitory factor signaling pathways are altered at the mRNA level during in vitro angiogenesis. Thus, at least for the mitogen-activated protein kinase pathway, mRNA changes as well as the phosphorylation changes of these gene products may be important in the control of blood vessel morphogenesis. Furthermore, in this study, we demonstrated the power of virtual Northern blot analysis, as an alternative to quantitative RT-PCR, for measuring the magnitudes of differential gene expression.

Modelling thirty-day mortality in the Acute Respiratory Distress Syndrome (ARDS) in an adult ICU.

Variables predicting thirty-day outcome from Acute Respiratory Distress Syndrome (ARDS) were analysed using Cox regression structured for time-varying covariates. Over a three-year period, 1996-1998, consecutive patients with ARDS (bilateral chest X-ray opacities, PaO2/FiO2 ratio of <200 and an acute precipitating event) were identified using a prospective computerized data base in a university teaching hospital ICU. The cohort, 106 mechanically ventilated patients, was of mean (SD) age 63.5 (15.5) years and 37% were female. Primary lung injury occurred in 45% and 24% were postoperative. ICU-admission day APACHE II score was 25 (8); ARDS onset time from ICU admission was 1 day (median: range 0-16) and 30 day mortality was 41% (95% CI: 33%-51%). At ARDS onset, PaO2/FiO2 ratio was 92 (31), 81% had four-quadrant chest X-ray opacification and lung injury score was 2.75 (0.45). Average mechanical ventilator tidal volume was 10.3 ml/predicted kg weight. Cox model mortality predictors (hazard ratio, 95% CI) were: APACHE II score, 1.15 (1.09-1.21); ARDS lag time (days), 0.72 (0.58-0.89); direct versus indirect injury, 2.89 (1.45-5.76); PaO2/FiO2 ratio, 0.98 (0.97-0.99); operative versus non-operative category, 0.24 (0.09-0.63). Time-varying effects were evident for PaO2/FiO2 ratio, operative versus non-operative category and ventilator tidal volume assessed as a categorical predictor with a cut-point of 8 ml/kg predicted weight (mean tidal volumes, 7.1 (1.9) vs 10.7 (1.6) ml/kg predicted weight). Thirty-day survival was improved for patients ventilated with lower tidal volumes. Survival predictors in ARDS were multifactorial and related to patient-injury-time interaction and level of mechanical ventilator tidal volume.

Factorial and time course designs for cDNA microarray experiments.

Microarrays are powerful tools for surveying the expression levels of many thousands of genes simultaneously. They belong to the new genomics technologies which have important applications in the biological, agricultural and pharmaceutical sciences. There are myriad sources of uncertainty in microarray experiments, and rigorous experimental design is essential for fully realizing the potential of these valuable resources. Two questions frequently asked by biologists on the brink of conducting cDNA or two-colour, spotted microarray experiments are 'Which mRNA samples should be competitively hybridized together on the same slide?' and 'How many times should each slide be replicated?' Early experience has shown that whilst the field of classical experimental design has much to offer this emerging multi-disciplinary area, new approaches which accommodate features specific to the microarray context are needed. In this paper, we propose optimal designs for factorial and time course experiments, which are special designs arising quite frequently in microarray experimentation. Our criterion for optimality is statistical efficiency based on a new notion of admissible designs; our approach enables efficient designs to be selected subject to the information available on the effects of most interest to biologists, the number of arrays available for the experiment, and other resource or practical constraints, including limitations on the amount of mRNA probe. We show that our designs are superior to both the popular reference designs, which are highly inefficient, and to designs incorporating all possible direct pairwise comparisons. Moreover, our proposed designs represent a substantial practical improvement over classical experimental designs which work in terms of standard interactions and main effects. The latter do not provide a basis for meaningful inference on the effects of most interest to biologists, nor make the most efficient use of valuable and limited resources.

Cost calculation and prediction in adult intensive care: a ground-up utilization study.

The ability of various proxy cost measures, including therapeutic activity scores (TISS and Omega) and cumulative daily severity of illness scores, to predict individual ICU patient costs was assessed in a prospective "ground-up" utilization costing study over a six month period in 1991. Daily activity (TISS and Omega scores) and utilization in consecutive admissions to three adult university associated ICUs was recorded by dedicated data collectors. Cost prediction used linear regression with determination (80%) and validation (20%) data sets. The cohort, 1333 patients, had a mean (SD) age 57.5 (19.4) years, (41% female) and admission APACHE III score of 58 (27). ICU length of stay and mortality were 3.9 (6.1) days and 17.6% respectively. Mean total TISS and Omega scores were 117 (157) and 72 (113) respectively. Mean patient costs per ICU episode (1991 dollar AUS) were dollar 6801 (dollar 10311), with median costs of dollar 2534, range dollar 106 to dollar 95,602. Dominant cost fractions were nursing 43.3% and overheads 16.9%. Inflation adjusted year 2002 (mean) costs were dollar 9343 (dollar AUS). Total costs in survivors were predicted by Omega score, summed APACHE III score and ICU length of stay; determination R2, 0.91; validation 0.88. Omega was the preferred activity score. Without the Omega score, predictors were age, summed APACHE III score and ICU length of stay; determination R2, 0.73; validation 0.73. In non-survivors, predictors were age and ICU length of stay (plus interaction), and Omega score (determination R2, 0.97; validation 0.91). Patient costs may be predicted by a combination of ICU activity indices and severity scores.

Phosphate metabolism in intensive care patients with acute respiratory failure.

OBJECTIVE: To determine the variables predicting the change of plasma phosphate over the first 24 hr period in intensive care in patients with acute respiratory failure. METHODS: Fifty-seven patients were studied prospectively in a university teaching hospital intensive care unit (ICU). Thirty two patients were classified as having acute respiratory failure and a primary respiratory system diagnosis (group I), 10 were classified as having acute cardiogenic pulmonary oedema (group II) and 15 were general ICU patients (group III). Arterial blood specimens at intensive care unit admission (T0) and at 24 hr post-admission (T24) were assayed for multiple plasma biochemical parameters including phosphate (PO4) and red blood cell 2,3-diphosphoglycerate (2,3-DPG). Timed urine collections were used to determine 24 hr urine phosphate loss and renal phosphate threshold concentration (RTP). During the measurement period glucose-free fluids only were infused. RESULTS: Fifty seven patients had a mean (+/- SD) age of 67 +/- 12 years and Apache II score of 22 +/- 6. The plasma PO4 at T0 was 1.55 +/- 0.71 mmol/L and showed a significant 24 hr decrease of 0.55 mmol/L (p < 0.0001) at T24. Hypophosphataemia at T0 was observed in 26% of patients. Red blood cell 2,3-DPG was not elevated at T0 (13.5 +/- 3.3 umol/gHb) and showed a non-significant increment over 24 hr. Urine phosphate loss over the 24 hr period was 21.8 +/- 14.0 mmol with RTP being reduced below the lower reference range limit in groups I (0.65 +/- 0.29 mmol/L) and II (0.57 +/- 0.29 mmol/L). The naive form of phosphate change (PO4T24-PO4T0) was significantly related to initial plasma PO4 and was subject to regression to the mean, which was estimated to have inflated the relationship by 25%. The appropriate form of phosphate change was found to be log ratio T24/T0 phosphate. Independent predictors of log ratio T24/T0 phosphate were 24 hr change (T24-T0) in both 2,3-DPG and arterial pH, RTP, prescription of aminophylline (categorical factor) and the interaction of aminophylline and RTP (R2 = 0.65, ordinary least squares regression). CONCLUSIONS: Twenty-four hour plasma phosphate decrement in intensive care unit patients was multi-factorial and was attended by a lowered renal threshold phosphate concentration.

Disease surveillance and data collection issues in epidemic modelling.

This paper is founded on a tutorial session given to the School on Modern Statistical Methods in Medical Research which was held at the International Centre for Theoretical Physics, Trieste in September 1999. We review the aims, scope and purposes of infectious disease surveillance including determining transmission information to underpin model structure and parameterization in epidemic modelling. The practical problems inherent in collecting surveillance data are illustrated by a study of HIV/AIDS in Cambodia. We also review the basic elements of mathematical models developed to represent the transmission dynamics of infectious diseases, and discuss reasons for the gap between mathematical epidemic models and available data.