Cardiac intensivists frequently assess patient readiness to wean off mechanical ventilation with an extubation readiness trial despite it being no more effective than clinician judgement alone. We evaluated the utility of high-frequency physiologic data and machine learning for improving the prediction of extubation failure in children with cardiovascular disease.

This was a retrospective analysis of clinical registry data and streamed physiologic extubation readiness trial data from one paediatric cardiac ICU (12/2016-3/2018). We analysed patients’ final extubation readiness trial. Machine learning methods (classification and regression tree, Boosting, Random Forest) were performed using clinical/demographic data, physiologic data, and both datasets. Extubation failure was defined as reintubation within 48 hrs. Classifier performance was assessed on prediction accuracy and area under the receiver operating characteristic curve.

Of 178 episodes, 11.2% (N = 20) failed extubation. Using clinical/demographic data, our machine learning methods identified variables such as age, weight, height, and ventilation duration as being important in predicting extubation failure. Best classifier performance with this data was Boosting (prediction accuracy: 0.88; area under the receiver operating characteristic curve: 0.74). Using physiologic data, our machine learning methods found oxygen saturation extremes and descriptors of dynamic compliance, central venous pressure, and heart/respiratory rate to be of importance. The best classifier in this setting was Random Forest (prediction accuracy: 0.89; area under the receiver operating characteristic curve: 0.75). Combining both datasets produced classifiers highlighting the importance of physiologic variables in determining extubation failure, though predictive performance was not improved.

Physiologic variables not routinely scrutinised during extubation readiness trials were identified as potential extubation failure predictors. Larger analyses are necessary to investigate whether these markers can improve clinical decision-making.

Cohorting patients who are colonized or infected with multidrug-resistant organisms (MDROs) protects uncolonized patients from acquiring MDROs in healthcare settings. The potential for cross transmission within the cohort and the possibility of colonized patients acquiring secondary isolates with additional antibiotic resistance traits is often neglected. We searched for evidence of cross transmission of KPC+ Klebsiella pneumoniae (KPC-Kp) colonization among cohorted patients in a long-term acute-care hospital (LTACH), and we evaluated the impact of secondary acquisitions on resistance potential.

Genomic epidemiological investigation.

A high-prevalence LTACH during a bundled intervention that included cohorting KPC-Kp–positive patients.

Whole-genome sequencing (WGS) and location data were analyzed to identify potential cases of cross transmission between cohorted patients.

Secondary KPC-Kp isolates from 19 of 28 admission-positive patients were more closely related to another patient’s isolate than to their own admission isolate. Of these 19 cases, 14 showed strong genomic evidence for cross transmission (<10 single nucleotide variants or SNVs), and most of these patients occupied shared cohort floors (12 patients) or rooms (4 patients) at the same time. Of the 14 patients with strong genomic evidence of acquisition, 12 acquired antibiotic resistance genes not found in their primary isolates.

Acquisition of secondary KPC-Kp isolates carrying distinct antibiotic resistance genes was detected in nearly half of cohorted patients. These results highlight the importance of healthcare provider adherence to infection prevention protocols within cohort locations, and they indicate the need for future studies to assess whether multiple-strain acquisition increases risk of adverse patient outcomes.

Apolipoprotein E (APOE) E4 is the main genetic risk factor for Alzheimer’s disease (AD). Due to the consistent association, there is interest as to whether E4 influences the risk of other neurodegenerative diseases. Further, there is a constant search for other genetic biomarkers contributing to these phenotypes, such as microtubule-associated protein tau (MAPT) haplotypes. Here, participants from the Ontario Neurodegenerative Disease Research Initiative were genotyped to investigate whether the APOE E4 allele or MAPT H1 haplotype are associated with five neurodegenerative diseases: (1) AD and mild cognitive impairment (MCI), (2) amyotrophic lateral sclerosis, (3) frontotemporal dementia (FTD), (4) Parkinson’s disease, and (5) vascular cognitive impairment.

Genotypes were defined for their respective APOE allele and MAPT haplotype calls for each participant, and logistic regression analyses were performed to identify the associations with the presentations of neurodegenerative diseases.

Our work confirmed the association of the E4 allele with a dose-dependent increased presentation of AD, and an association between the E4 allele alone and MCI; however, the other four diseases were not associated with E4. Further, the APOE E2 allele was associated with decreased presentation of both AD and MCI. No associations were identified between MAPT haplotype and the neurodegenerative disease cohorts; but following subtyping of the FTD cohort, the H1 haplotype was significantly associated with progressive supranuclear palsy.

This is the first study to concurrently analyze the association of APOE isoforms and MAPT haplotypes with five neurodegenerative diseases using consistent enrollment criteria and broad phenotypic analysis.

The Comprehensive Assessment of Neurodegeneration and Dementia (COMPASS-ND) cohort study of the Canadian Consortium on Neurodegeneration in Aging (CCNA) is a national initiative to catalyze research on dementia, set up to support the research agendas of CCNA teams. This cross-country longitudinal cohort of 2310 deeply phenotyped subjects with various forms of dementia and mild memory loss or concerns, along with cognitively intact elderly subjects, will test hypotheses generated by these teams.

The COMPASS-ND protocol, initial grant proposal for funding, fifth semi-annual CCNA Progress Report submitted to the Canadian Institutes of Health Research December 2017, and other documents supplemented by modifications made and lessons learned after implementation were used by the authors to create the description of the study provided here.

The CCNA COMPASS-ND cohort includes participants from across Canada with various cognitive conditions associated with or at risk of neurodegenerative diseases. They will undergo a wide range of experimental, clinical, imaging, and genetic investigation to specifically address the causes, diagnosis, treatment, and prevention of these conditions in the aging population. Data derived from clinical and cognitive assessments, biospecimens, brain imaging, genetics, and brain donations will be used to test hypotheses generated by CCNA research teams and other Canadian researchers. The study is the most comprehensive and ambitious Canadian study of dementia. Initial data posting occurred in 2018, with the full cohort to be accrued by 2020.

Availability of data from the COMPASS-ND study will provide a major stimulus for dementia research in Canada in the coming years.

Although most hospitals report very high levels of hand hygiene compliance (HHC), the accuracy of these overtly observed rates is questionable due to the Hawthorne effect and other sources of bias. In the study, we aimed (1) to compare HHC rates estimated using the standard audit method of overt observation by a known observer and a new audit method that employed a rapid (<15 minutes) “secret shopper” method and (2) to pilot test a novel feedback tool.

Quality improvement project using a quasi-experimental stepped-wedge design.

This study was conducted in 5 acute-care hospitals (17 wards, 5 intensive care units) in the Midwestern United States.

Sites recruited a hand hygiene observer from outside the acute-care units to rapidly and covertly observe entry and exit HHC during the study period, October 2016–September 2017. After 3 months of observations, sites received a monthly feedback tool that communicated HHC information from the new audit method.

The absolute difference in HHC estimates between the standard and new audit methods was ~30%. No significant differences in HHC were detected between the baseline and feedback phases (OR, 0.92; 95% CI, 0.84–1.01), but the standard audit method had significantly higher estimates than the new audit method (OR, 9.83; 95% CI, 8.82–10.95).

HHC estimates obtained using the new audit method were substantially lower than estimates obtained using the standard audit method, suggesting that the rapid, secret-shopper method is less subject to bias. Providing feedback using HHC from the new audit method did not seem to impact HHC behaviors.

To determine the impact of recurrent Clostridium difficile infection (RCDI) on patient behaviors following illness.

Using a computer algorithm, we searched the electronic medical records of 7 Chicago-area hospitals to identify patients with RCDI (2 episodes of CDI within 15 to 56 days of each other). RCDI was validated by medical record review. Patients were asked to complete a telephone survey. The survey included questions regarding general health, social isolation, symptom severity, emotional distress, and prevention behaviors.

In total, 119 patients completed the survey (32%). On average, respondents were 57.4 years old (standard deviation, 16.8); 57% were white, and ~50% reported hospitalization for CDI. At the time of their most recent illness, patients rated their diarrhea as high severity (58.5%) and their exhaustion as extreme (30.7%). Respondents indicated that they were very worried about getting sick again (41.5%) and about infecting others (31%). Almost 50% said that they have washed their hands more frequently (47%) and have increased their use of soap and water (45%) since their illness. Some of these patients (22%–32%) reported eating out less, avoiding certain medications and public areas, and increasing probiotic use. Most behavioral changes were unrelated to disease severity.

Having had RCDI appears to increase prevention-related behaviors in some patients. While some behaviors are appropriate (eg, handwashing), others are not supported by evidence of decreased risk and may negatively impact patient quality of life. Providers should discuss appropriate prevention behaviors with their patients and should clarify that other behaviors (eg, eating out less) will not affect their risk of future illness.

Infect Control Hosp Epidemiol. 2017;38:1351–1357

Standard estimates of the impact of Clostridium difficile infections (CDI) on inpatient lengths of stay (LOS) may overstate inpatient care costs attributable to CDI. In this study, we used multistate modeling (MSM) of CDI timing to reduce bias in estimates of excess LOS.

A retrospective cohort study of all hospitalizations at any of 120 acute care facilities within the US Department of Veterans Affairs (VA) between 2005 and 2012 was conducted. We estimated the excess LOS attributable to CDI using an MSM to address time-dependent bias. Bootstrapping was used to generate 95% confidence intervals (CI). These estimates were compared to unadjusted differences in mean LOS for hospitalizations with and without CDI.

During the study period, there were 3.96 million hospitalizations and 43,540 CDIs. A comparison of unadjusted means suggested an excess LOS of 14.0 days (19.4 vs 5.4 days). In contrast, the MSM estimated an attributable LOS of only 2.27 days (95% CI, 2.14–2.40). The excess LOS for mild-to-moderate CDI was 0.75 days (95% CI, 0.59–0.89), and for severe CDI, it was 4.11 days (95% CI, 3.90–4.32). Substantial variation across the Veteran Integrated Services Networks (VISN) was observed.

CDI significantly contributes to LOS, but the magnitude of its estimated impact is smaller when methods are used that account for the time-varying nature of infection. The greatest impact on LOS occurred among patients with severe CDI. Significant geographic variability was observed. MSM is a useful tool for obtaining more accurate estimates of the inpatient care costs of CDI.

Infect. Control Hosp. Epidemiol. 2015;36(9):1024–1030