Development of a bedside tool to predict the probability of drug-resistant pathogens among hospitalized adult patients with gram-negative infections.

BACKGROUND: We developed a clinical bedside tool to simultaneously estimate the probabilities of third-generation cephalosporin-resistant Enterobacteriaceae (3GC-R), carbapenem-resistant Enterobacteriaceae (CRE), and multidrug-resistant Pseudomonas aeruginosa (MDRP) among hospitalized adult patients with Gram-negative infections. METHODS: Data were obtained from a retrospective observational study of the Premier Hospital that included hospitalized adult patients with a complicated urinary tract infection (cUTI), complicated intra-abdominal infection (cIAI), hospital-acquired/ventilator-associated pneumonia (HAP/VAP), or bloodstream infection (BSI) due to Gram-negative bacteria between 2011 and 2015. Risk factors for 3GC-R, CRE, and MDRP were ascertained by multivariate logistic regression, and separate models were developed for patients with community-acquired versus hospital-acquired infections for each resistance phenotype (N = 6). Models were converted to a singular user-friendly interface to estimate the probabilities of a patient having an infection due to 3GC-R, CRE, or MDRP when ≥ 1 risk factor was present. RESULTS: Overall, 124,068 patients contributed to the dataset. Percentages of patients admitted for cUTI, cIAI, HAP/VAP, and BSI were 61.6, 4.6, 16.5, and 26.4%, respectively (some patients contributed > 1 infection type). Resistant infection rates were 1.90% for CRE, 12.09% for 3GC-R, and 3.91% for MDRP. A greater percentage of the resistant infections were community-acquired relative to hospital-acquired (CRE, 1.30% vs 0.62% of 1.90%; 3GC-R, 9.27% vs 3.42% of 12.09%; MDRP, 2.39% vs 1.59% of 3.91%). The most important predictors of having an 3GC-R, CRE or MDRP infection were prior number of antibiotics; infection site; infection during the previous 3 months; and hospital prevalence of 3GC-R, CRE, or MDRP. To enable application of the six predictive multivariate logistic regression models to real-world clinical practice, we developed a user-friendly interface that estimates the risk of 3GC-R, CRE, and MDRP simultaneously in a given patient with a Gram-negative infection based on their risk (Additional file 1). CONCLUSIONS: We developed a clinical prediction tool to estimate the probabilities of 3GC-R, CRE, and MDRP among hospitalized adult patients with confirmed community- and hospital-acquired Gram-negative infections. Our predictive model has been implemented as a user-friendly bedside tool for use by clinicians/healthcare professionals to predict the probability of resistant infections in individual patients, to guide early appropriate therapy.

Delays in treatment intensification with oral antidiabetic drugs and risk of microvascular and macrovascular events in patients with poor glycaemic control: An individual patient simulation study.

AIMS: To use the Archimedes model to estimate the consequences of delays in oral antidiabetic drug (OAD) treatment intensification on glycaemic control and long-term outcomes at 5 and 20 years. MATERIALS AND METHODS: Using real-world data, we modelled a cohort of hypothetical patients with glycated haemoglobin (HbA1c) ≥8%, on metformin, with no history of insulin use. The cohort included 3 strata based on the number of OADs taken at baseline. The first add-on in the intensification sequence was a sulphonylurea, next was a dipeptidyl peptidase-4 inhibitor, and last, a thiazolidinedione. The scenarios included either no delay or delay, based on observed and extrapolated times to intensification. RESULTS: At 1 year, HbA1c was 6.8% for patients intensifying without delay, and 8.2% for those delaying intensification. For no delay vs delay, risks of major adverse cardiac events, myocardial infarction, heart failure and amputations were reduced by 18.0%, 25.0%, 13.7%, and 20.4%, respectively, at 5 years; severe hypoglycaemia risk, however, increased to 19% for the no delay scenario vs 12.5% for delay. At 20 years, the results showed similar trends to those at 5 years. CONCLUSIONS: Timing of intensification of OAD therapy according to guideline recommendations led to greater reductions in HbA1c and lower risks of complications, but higher risks of hypoglycaemia than delaying intensification. These results highlight the potential impact of timely treatment intensification on long-term outcomes.

Clinical- and cost-effectiveness of LDL particle-guided statin therapy: a simulation study.

We used the Archimedes Model, a mathematical simulation model (Model) to estimate the clinical- and cost-effectiveness of using LDL particle concentration (LDL-P) as an adjunct or alternative to LDL cholesterol (LDL-C) to guide statin therapy. LDL-P by NMR has been shown to be a better measure of cardiovascular disease (CVD) risk than LDL-C, and may therefore be a better gauge of the need for and response to statin treatment. Using the Model, we conducted a virtual clinical trial comparing the use of LDL-C alone, LDL-P alone, and LDL-C and LDL-P together to guide treatment in the general adult population, and in high-risk, dyslipidemic subpopulations. In the general population, the 5-year major adverse cardiovascular event (MACE) relative risk reduction (RRR) of LDL-P alone compared to the control arm (LDL-C alone) was 5.0% (95% CI, 4.7-5.3; p < .0001); using both LDL-C and LDL-P (dual markers) led to 3.0% RRR compared to the control arm (95% CI, 2.8-3.3; p < .0001). For individuals with diabetes, the RRR was 7.3% (95% CI, 6.4-8.2; p < .0001) for LDL-P alone and 6.9% for dual markers (95% CI, 6.1-7.8; both, p < .0001). In the general population, the costs per quality-adjusted life year (QALY) associated with the use of LDL-P alone were $76,052 at 5 years and $8913 at 20 years and $142,825 at 5 years and $25,505 at 20 years with the use of both markers. In high-risk subpopulations, the use of LDL-P alone was cost-saving at 5 years; whereas the cost per QALY for the use of both markers was $14,250 at 5 years and $859 at 20 years for high-risk dyslipidemics, $19,192 at 5 years and $649 at 20 years for diabetics, and $9030 at 5 years and $7268 at 20 years for patients with prior CHD. In conclusion, the model estimates that using LDL-P to guide statin therapy may reduce the risk of CVD events to a greater extent than does the use of LDL-C alone and maybe cost-effective or cost-saving for high-risk patients.

Cost-effectiveness of a genetic test for breast cancer risk.

Genetic testing of seven single-nucleotide polymorphisms (7SNP) can improve estimates of risk of breast cancer relative to the Gail risk test alone, for the purpose of recommending MRI screening for women at high risk. A simulation of breast cancer and health care processes was used to conduct a virtual trial comparing the use of the 7SNP test with the Gail risk test to categorize patients by risk. Average-risk patients received annual mammogram, whereas high-risk patients received annual MRI. Cancer incidence was based on Surveillance, Epidemiology, and End Results data and validated to Cancer Prevention Study II Nutrition Cohort data. Risk factor values were drawn from National Health and Nutrition Examination Survey (NHANES-4) and Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial data. Mammogram characteristics were derived from Breast Cancer Surveillance Consortium data. The test was most cost-effective when given to patients at an intermediate lifetime risk of breast cancer. For patients with a risk of 16% to 28%, it resulted in a 1.91% reduction in cancer deaths, saving 0.005 quality-adjusted life years per person at a cost of $163,264 per QALY. These results were sensitive to the age at which the test is given, the discount rate, and the costs of the genetic test and MRI. The cost effectiveness of using the 7SNP test for patients with intermediate Gail risk is similar to that of other recommended strategies, including annual MRI for patients with a lifetime risk greater than 20% or BRCA1/2 mutations.

Cooperation, competition, and coalitions in enzyme-producing microbes: social evolution and nutrient depolymerization rates.

Extracellular enzymes represent a public good for microbial communities, as they break down complex molecules into simple molecules that microbes can take up. These communities are vulnerable to cheating by microbes that do not produce enzymes, but benefit from those produced by others. However, extracellular enzymes are ubiquitous and play an important role in the depolymerization of nutrients. We developed a multi-genotype, multi-nutrient model of a community of exoenzyme-producing microbes, in order to investigate the relationship between diversity, social interactions, and nutrient depolymerization. We focused on coalitions between complementary types of microbes and their implications for spatial pattern formation and nutrient depolymerization. The model included polymers containing carbon, nitrogen, or phosphorus, and eight genotypes of bacteria, which produced different subsets of the three enzymes responsible for hydrolyzing these polymers. We allowed social dynamics to emerge from a mechanistic model of enzyme production, action, and diffusion. We found that diversity was maximized at high rates of either diffusion or enzyme production (but not both). Conditions favoring cheating also favored the emergence of coalitions. We characterized the spatial patterns formed by different interactions, showing that same-type cooperation leads to aggregation, but between-type cooperation leads to an interwoven, filamentous pattern. Contrary to expectations based on niche complementarity, we found that nutrient depolymerization declined with increasing diversity due to a negative competitive effect of coalitions on generalist producers, leading to less overall enzyme production. This decline in depolymerization was stronger for non-limiting nutrients in the system. This study shows that social interactions among microbes foraging for complementary resources can influence microbial diversity, microbial spatial distributions, and rates of nutrient depolymerization.

Direct benefits of genetic mosaicism and intraorganismal selection: modeling coevolution between a long-lived tree and a short-lived herbivore.

We model direct fitness benefits of genetic mosaicism for a long-lived tree in coevolution with a short-lived herbivore to test four hypotheses: that mosaicism reduces selection on the herbivore for resistance to plant defenses; that module-level selection allows the individual tree to adapt to its herbivore; and that this benefits the tree population, increasing average tree fitness and reducing local adaptation of the herbivore. We show that: mosaicism does not sufficiently reduce selection for resistance in the herbivore to benefit the tree; that individual trees do benefit from module-level selection when somatic mutation introduces new defenses; and that mosaicism does reduce local adaptation in the herbivore, which increases average tree fitness. These results are robust to varying genetic assumptions of dominance and the somatic mutation rate, but only hold for sufficiently long-lived trees with relatively strong selection. We also show that a mixed reproductive strategy of primarily asexual reproduction interspersed with occasional sexual reproduction is effective in coevolving with the herbivore, as it maintains beneficial allele combinations. Finally, we argue that intraorganismal genetic heterogeneity need not threaten the integrity of the individual and may be adaptive when selection acts concordantly between levels.