Machine learning model from a Spanish cohort for prediction of SARS-COV-2 mortality risk and critical patients.

Patients affected by SARS-COV-2 have collapsed healthcare systems around the world. Consequently, different challenges arise regarding the prediction of hospital needs, optimization of resources, diagnostic triage tools and patient evolution, as well as tools that allow us to analyze which are the factors that determine the severity of patients. Currently, it is widely accepted that one of the problems since the pandemic appeared was to detect (i) who patients were about to need Intensive Care Unit (ICU) and (ii) who ones were about not overcome the disease. These critical patients collapsed Hospitals to the point that many surgeries around the world had to be cancelled. Therefore, the aim of this paper is to provide a Machine Learning (ML) model that helps us to prevent when a patient is about to be critical. Although we are in the era of data, regarding the SARS-COV-2 patients, there are currently few tools and solutions that help medical professionals to predict the evolution of patients in order to improve their treatment and the needs of critical resources at hospitals. Moreover, most of these tools have been created from small populations and/or Chinese populations, which carries a high risk of bias. In this paper, we present a model, based on ML techniques, based on 5378 Spanish patients' data from which a quality cohort of 1201 was extracted to train the model. Our model is capable of predicting the probability of death of patients with SARS-COV-2 based on age, sex and comorbidities of the patient. It also allows what-if analysis, with the inclusion of comorbidities that the patient may develop during the SARS-COV-2 infection. For the training of the model, we have followed an agnostic approach. We explored all the active comorbidities during the SARS-COV-2 infection of the patients with the objective that the model weights the effect of each comorbidity on the patient's evolution according to the data available. The model has been validated by using stratified cross-validation with k = 5 to prevent class imbalance. We obtained robust results, presenting a high hit rate, with 84.16% accuracy, 83.33% sensitivity, and an Area Under the Curve (AUC) of 0.871. The main advantage of our model, in addition to its high success rate, is that it can be used with medical records in order to predict their diagnosis, allowing the critical population to be identified in advance. Furthermore, it uses the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD 9-CM) standard. In this sense, we should also emphasize that those hospitals using other encodings can add an intermediate layer business to business (B2B) with the aim of making transformations to the same international format.

The Maritime Declaration of Health (MDH) as a tool to detect maritime traffic-related health risks: analysis of MDH forms submitted to Spanish ports, October 2014 to March 2015.

The international maritime traffic of people and goods has often contributed to the spread of pathogens affecting public health. The Maritime Declaration of Health (MDH), according to the International Health Regulations (IHR) (2005), is a document containing data related to the state of health on board a ship during passage and on arrival at port. It is a useful tool for early detection of public health risks. The main objective of our study was to evaluate compliance with the model provided in the IHR, focusing on the format and degree of completion of MDH forms received at Spanish ports. We reviewed the content of 802 MDH forms submitted to nine Spanish ports between October 2014 and March 2015. Study results show that 22% of MDH forms presented did not comply with the recommended model and 39% were incomplete. The proportion of cargo ships with correct and complete MDH forms was lower than passenger ships; thus, the nine health questions were answered less frequently by cargo ships than passenger ships (63% vs 90%, p value < 0.001). The appropriate demand and usage of MDH forms by competent authorities should improve the quality of the document as a tool and improve risk assessment.

Repeat head computed tomography after minimal brain injury identifies the need for craniotomy in the absence of neurologic change.

BACKGROUND: In this era of cost containment, the value of routine repeat head computed tomography (CT) in patients with mild TBI (mTBI) and no interval neurologic change has been challenged. The purpose of this study was to test the hypothesis that routine repeat head CT provides critical information after mTBI even with no neurologic change. METHODS: From January 1996 to May 2010, records from all patients admitted to our Level I trauma center with an arrival Glasgow Coma Scale (GCS) score of 13 to 15 and at least one head CT were retrospectively reviewed. RESULTS: In 360 patients with mTBI and positive initial head CT finding, the most common abnormalities were subarachnoid hemorrhage (64%), intraparenchymal hemorrhage (57%), and subdural hemorrhage (40%). Scans were repeated in 8 ± 6 hours; 11% were recalled, 59% remained stable, but 30% showed injury progression. Those patients with worsening repeat head CT finding had higher Injury Severity Score (ISS), were more likely to be intubated and require craniotomy, had longer stay, and had higher mortality (all p < 0.001). On multiple logistic regression, altered GCS score (odds ratio, 3.1-4.0), ISS (odds ratio, 1.1), and presence of mass effect (odds ratio, 2.0) were independently associated with worsening repeat head CT finding. In patients receiving a neurosurgical operative intervention, 32% to 59% had no clinical decline before the worsening repeat CT finding. CONCLUSION: After mTBI, worsening of repeat head CT finding is seen in a third of patients and is associated with worse outcomes. A substantial fraction of patients who require operative intervention will have no clinical changes in the first 8 hours, supporting the value of repeat head CT within this time frame. LEVEL OF EVIDENCE: Care management study, level III.

Admission hematocrit and transfusion requirements after trauma.

BACKGROUND: The decision to transfuse packed RBCs (PRBC) during initial resuscitation of trauma patients is based on physiologic state, evidence for blood loss, and potential for ongoing hemorrhage. Initial hematocrit (Hct) is not considered an accurate marker of blood loss. This study tests the hypothesis that admission Hct is associated with transfusion requirements after trauma. METHODS: From June to December 2008, data from 1,492 consecutive admissions at a Level I trauma center were retrospectively reviewed to determine whether initial Hct was associated with PRBC transfusions. From October 2009 through October 2011, data from 463 consecutive transfused patients were retrospectively reviewed to determine whether Hct correlated with number of PRBC units received. RESULTS: Packed RBC transfusion was not correlated with heart rate and was more highly correlated with Hct (r = -0.45) than with systolic blood pressure or base deficit (r = -0.32 or r = -0.26). Hematocrit was a better overall predictor than systolic blood pressure (sensitivity 45% vs 29%, specificity 94% vs 98%, area under receiver operator characteristic curve 0.71 vs 0.64). Lower Hct was associated with hypotension, more advanced shock, higher blood loss, and increased transfusion of PRBC, plasma, platelets, or cryoprecipitate (all, p < 0.01). CONCLUSION: Admission Hct is more strongly associated with the PRBC transfusion than either tachycardia, hypotension, or acidosis. Admission Hct is also correlated with 24-hour blood product requirements in those receiving early transfusions. These findings challenge current thinking and suggest that fluid shifts are rapid after trauma and that Hct can be important in initial trauma assessment.

Venous thromboembolism after trauma: a never event?*.

OBJECTIVE: Rates of venous thromboembolism as high as 58% have been reported after trauma, but there is no widely accepted screening protocol. If Medicare adds venous thromboembolism to the list of "preventable complications," they will no longer reimburse for treatment, which could have devastating effects on many urban centers. We hypothesized that prescreening with a risk assessment profile followed by routine surveillance with venous duplex ultrasound that could identify asymptomatic venous thromboembolism in trauma patients. DESIGN: Prospective, observational trial with waiver of consent. SETTING: Level I trauma center intensive care unit. PATIENTS: At admission, 534 patients were prescreened with a risk assessment profile. INTERVENTIONS: Patients (n = 106) with risk assessment profile scores >10 were considered high risk and received routine screening venous duplex ultrasound within 24 hrs and weekly thereafter. RESULTS: In prescreened high-risk patients, 20 asymptomatic deep vein thrombosis were detected with venous duplex ultrasound (19%). An additional ten venous thromboembolisms occurred, including six symptomatic deep vein thrombosis and four pulmonary emboli, resulting in an overall venous thromboembolism rate of 28%. The most common risk factors discriminating venous thromboembolism vs. no venous thromboembolism were femoral central venous catheter (23% vs. 8%), operative intervention >2 hrs (77% vs. 46%), complex lower extremity fracture (53% vs. 32%), and pelvic fracture (70% vs. 47%), respectively (all p < .05). Risk assessment profile scores were higher in patients with venous thromboembolism (19 ± 6 vs. 14 ± 4, p = .001). Risk assessment profile score (odds ratio 1.14) and the combination of pelvic fracture requiring operative intervention >2 hrs (odds ratio 5.75) were independent predictors for development of venous thromboembolism. The rates of venous thromboembolism for no chemical prophylaxis (33%), unfractionated heparin (29%), dalteparin (40%), or inferior vena cava filters (20%) were not statistically different (p = .764). CONCLUSIONS: Medicare's inclusion of venous thromboembolism after trauma as a "never event" should be questioned. In trauma patients, high-risk assessment profile score and pelvic fracture with prolonged operative intervention are independent predictors for venous thromboembolism development, despite thromboprophylaxis. Although routine venous duplex ultrasound screening may not be cost-effective for all trauma patients, prescreening using risk assessment profile yielded a cohort of patients with a high prevalence of venous thromboembolism. In such high-risk patients, routine venous duplex ultrasound and/or more aggressive prophylactic regimens may be beneficial.

[Use of complementary and alternative medicine in patients with chronic hepatitis C]. / Utilización de las medicinas alternativas y complementarias por los pacientes con hepatitis C crónica.

BACKGROUND AND OBJECTIVE: Complementary and alternative medicines (CAM) have achieved a great development in western countries. However, their use among patients simultaneously treated by the mainstream medicine is largely unknown. Our goal was to assess how many patients with chronic hepatitis C treated in a tertiary hospital use or have used CAM. PATIENTS AND METHOD: Analysis of the answers of 319 patients to a self-administered questionnaire. RESULTS: 113 (37%) patients had used or were using CAM, 63 (20%) because of chronic hepatitis and 50 (17%) for other reasons. Women, those with higher education, divorced and widows were those who more frequently used CAM. More than half of patients felt some subjective improvement, yet none of them normalized their serum transaminase activities. CONCLUSIONS: CAM are used by a high proportion of patients who are simultaneously attended by 'official' physicians. The perceived efficacy of these practices is high but no changes in the hepatic disease could be seen in any of the patients who answered the questionnaire.