Computer algorithm can match physicians' decisions about blood transfusions.
Ontology highlight
ABSTRACT: Background
Checking appropriateness of blood transfusion for quality assurance required enormous usage of time and human resources from the healthcare system. We report here a new machine learning algorithm for checking blood transfusion quality.Materials and methods
The multilayer perceptron neural network (MLPNN) was designed to learn an expert's judgement from 4946 clinical cases. The accuracy in predicting the blood transfusion was then reported.Results
We achieved a 96.8% overall accuracy rate, with a 99% match rate to the experts' judgement on those appropriate cases and 90.9% on the inappropriate cases.Conclusions
Machine learning algorithm can accurately match to human judgement by feeding in pre-surgical information and key laboratory variables.
SUBMITTER: Yao Y
PROVIDER: S-EPMC6785926 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BackgroundOne of the most important decisions that emergency department (ED) physicians make is patient disposition (admission vs discharge).ObjectivesTo determine how ED physicians perceive their discharge decisions for high-acuity patients and the impact on adverse events (adverse outcomes associated with healthcare management).MethodsWe conducted a real-time survey of staff ED physicians discharging consecutive patients from high-acuity areas of a tertiary care ED. We asked open-ended questions about rationale for discharge decisions and use of clinical judgement versus evidence. We searched for 30-day flagged outcomes (deaths, unscheduled admissions, ED or clinic visits). Three trained blinded ED physicians independently reviewed these for adverse events and preventability. We resolved disagreements by consensus. We used descriptive statistics and 95% CIs.ResultsWe interviewed 88.9% (32/36) of possible ED physicians for 366 discharge decisions. Respondents were mostly male (71.9%) and experienced (53.1% >10?years). ED physicians stated they used clinical judgement in 87.6% of decisions and evidence in 12.4%. There were 69 flagged outcomes (18.8%) and 10 adverse events (2.7%, 95% CI 1.1 to 4.5%). All adverse events were preventable (1 death, 4 admissions, 5 return ED visits). No significant associations occurred between decision-making rationale and adverse events.ConclusionsExperienced ED physicians most often relied on clinical acumen rather than evidence-based guidelines when discharging patients from ED high-acuity areas. Neither approach was associated with adverse events. In order to improve the safety of discharge decisions, further research should focus on decision support solutions and feedback interventions.
| S-EPMC4283689 | biostudies-literature
Project description:When searching for an object, we use a target template in memory that contains task-relevant information to guide visual attention to potential targets and to determine the identity of attended objects. These processes in visual search have typically been assumed to rely on a common source of template information. However, our recent work (Yu et al., 2022) argued that attentional guidance and target-match decisions rely on different information during search, with guidance using a "fuzzier" version of the template compared with target decisions. However, that work was based on the special case of search for a target amongst linearly separable distractors (e.g., search for an orange target amongst yellower distractors). Real-world search targets, however, are infrequently linearly separable from distractors, and it remains unclear whether the differences between the precision of template information used for guidance compared with target decisions also applies under more typical conditions. In four experiments, we tested this question by varying distractor similarity during visual search and measuring the likelihood of attentional guidance to distractors and target misidentifications. We found that early attentional guidance is indeed less precise than that of subsequent match decisions under varying exposure durations and distractor set sizes. These results suggest that attentional guidance operates on a coarser code than decisions, perhaps because guidance is constrained by lower acuity in peripheral vision or the need to rapidly explore a wide region of space while decisions about selected objects are more precise to optimize decision accuracy.
| S-EPMC9232460 | biostudies-literature
Project description:In the past 15 years, multiple clinical studies have identified a temporal association between red blood cell (RBC) transfusions and necrotizing enterocolitis (NEC). With some variability, most of these studies indicate that up to one-third of all cases of NEC involving very low-birth weight infants may occur within 24-48h after receiving a RBC transfusion. There is also evidence that the risk of such transfusion-associated NEC may be higher in infants transfused with the greatest severity of anemia. In this article, we summarize the clinical evidence pertaining to these issues; specifically, the contribution of RBC transfusions, and the contribution of severity of underlying anemia, to the pathogenesis of a type of NEC potentially termed, "transfusion/anemia-associated NEC."
| S-EPMC5776697 | biostudies-literature
Project description:Practical relevanceblood transfusions are a potentially life-saving procedure that are within the reach of most small animal practitioners. Only minimal equipment is required.Patient groupany cat with clinical signs attributable to a reduced red blood cell mass that is affecting oxygen transport (as a result of reduced packed cell volume or acute blood loss) is a potential candidate for a transfusion.Clinical challengesalthough the principles of transfusion medicine are not complicated, there can be fatal consequences if certain steps are omitted.Diagnosticsblood typing kits and blood filters are readily available from veterinary wholesalers, laboratories and blood banking services.Evidence baseover the past three decades, a substantial body of clinical research and reports has built up covering feline blood types and transfusion medicine. This article draws on that research to provide clinical guidance aimed at all veterinarians in feline or small animal practice who either currently practise transfusion medicine or plan to do so.
| S-EPMC7111214 | biostudies-literature
Project description:BACKGROUND:Transfusion of blood products is an essential component of the hematopoietic cell transplantation (HCT) process. Blood transfusion carries several risks including, but not limited to, lung injury. The effect of transfusions on lung complications after HCT has not been previously investigated. STUDY DESIGN AND METHODS:We retrospectively studied 215 adult allogeneic HCT recipients at the University of Minnesota and examined the association between transfusion of blood components and development of lung complications after HCT. Patients without lung complications were used as the control group. RESULTS:A total of 113 (58%) of the patients developed lung injury events before Day 180 after HCT. Six-month survival was significantly lower in the lung event group (52%) versus the controls (78%; p = 0.01). Patients who eventually developed lung events received more transfusion episodes per week in the first month after HCT (median, 4.3 vs. 2.7 for controls), platelet units per week (3.5 vs. 2.0), and RBC units per week (1.8 vs. 1.4; p < 0.01) for all. In a multivariable analysis, each additional transfusion before Day +30 was associated with a 2.7% higher risk of lung complication (95% confidence interval, 0.8-4.8; p = 0.01), adjusting for time to engraftment, conditioning intensity, and donor type. Blood utilization increased after the lung event and remained high for several months relative to controls. CONCLUSION:Our data suggest that transfusion of blood products is associated with and may further complicate lung complications after HCT. Cautious use of blood components in the post HCT period is recommended.
| S-EPMC5577937 | biostudies-literature
Project description:When searching for a target object, we engage in a continuous "look-identify" cycle in which we use known features of the target to guide attention toward potential targets and then to decide whether the selected object is indeed the target. Target information in memory (the target template or attentional template) is typically characterized as having a single, fixed source. However, debate has recently emerged over whether flexibility in the target template is relational or optimal. On the basis of evidence from two experiments using college students (Ns = 30 and 70, respectively), we propose that initial guidance of attention uses a coarse relational code, but subsequent decisions use an optimal code. Our results offer a novel perspective that the precision of template information differs when guiding sensory selection and when making identity decisions during visual search.
| S-EPMC8822198 | biostudies-literature
Project description: Not available
| S-EPMC10538875 | biostudies-literature
Project description:BackgroundExperienced and anticipated regret influence physicians' decision-making. In medicine, diagnostic decisions and diagnostic errors can have a severe impact on both patients and physicians. Little empirical research exists on regret experienced by physicians when they make diagnostic decisions in primary care that later prove inappropriate or incorrect. The aim of this study was to explore the experience of regret following diagnostic decisions in primary care.MethodsIn this qualitative study, we used an online questionnaire on a sample of German primary care physicians. We asked participants to report on cases in which the final diagnosis differed from their original opinion, and in which treatment was at the very least delayed, possibly resulting in harm to the patient. We asked about original and final diagnoses, illness trajectories, and the reactions of other physicians, patients and relatives. We used thematic analysis to assess the data, supported by MAXQDA 11 and Microsoft Excel 2016.Results29 GPs described one case each (14 female/15 male patients, aged 1.5-80 years, response rate < 1%). In 26 of 29 cases, the final diagnosis was more serious than the original diagnosis. In two cases, the diagnoses were equally serious, and in one case less serious. Clinical trajectories and the reactions of patients and relatives differed widely. Although only one third of cases involved preventable harm to patients, the vast majority (27 of 29) of physicians expressed deep feelings of regret.ConclusionEven if harm to patients is unavoidable, regret following diagnostic decisions can be devastating for clinicians, making them 'second victims'. Procedures and tools are needed to analyse cases involving undesirable diagnostic events, so that 'true' diagnostic errors, in which harm could have been prevented, can be distinguished from others. Further studies should also explore how physicians can be supported in dealing with such events in order to prevent them from practicing defensive medicine.
| S-EPMC7079478 | biostudies-literature
Project description:Stochasticity between independent LC-MS/MS runs is a challenging problem in the field of proteomics resulting in significant missing values (i.e., abundance measurements) among observed peptides. To address this issue, several approaches have been developed including computational methods such as MaxQuant’s Match-Between-Runs (MBR) algorithm. Often dozens of runs are all considered at once by MBR, transferring identifications from any one run to any of the others. To evaluate the error associated with these transfer events, we created a two-sample/two-proteome approach. In this way, samples containing no yeast lysates were assessed for false identification transfers from samples containing yeast. Overall, transfers from any of 20 runs with yeast were assessed to any of 20 runs without yeast were measured. While MBR increased the total number of spectral identifications by ~40%, we also found that 44% of all identified yeast proteins had identifications transferred to at least one sample without yeast. However, of these only 2.7% remained in the final dataset after applying the MaxQuant LFQ algorithm. We conclude that false transfers by MBR are plentiful, but likely have only a small impact on the final dataset due to downstream removal.
2019-09-25 | PXD014415 | Pride
Project description:OBJECTIVE: There are few data available about factors which influence physicians' decisions to discharge patients from their practices. To study general internists' and family medicine physicians' attitudes and experiences in discharging patients from their practices. DESIGN: A cross-sectional mailed survey was used. PARTICIPANTS: One thousand general internists and family medicine physicians participated in this study. MEASUREMENTS AND MAIN RESULTS: We studied the likelihood physicians would discharge 12 hypothetical patients from their practices, and whether they had actually discharged such patients. The effect of demographic data on the number of scenarios in which patients were likely to be discharged, and the number of patients actually discharged were analyzed via ANOVA and multiple logistic regression analysis. Of 977 surveys received by subjects, 526 (54%) were completed and returned. A majority of respondents were willing to discharge patients in 5 of 12 hypothetical scenarios. Eighty-five percent had actually discharged at least one patient from their practices. Most respondents (71%) had discharged 10 or fewer patients, but 14% had discharged 11 to 200 patients. Respondents who were in private practice (p < 0.000001) were more likely to discharge both hypothetical and actual patients from their practices. Older physicians (> or =48 years old) were more likely to discharge actual patients from their practices (p = 0.005) as were physicians practicing in rural settings (p = 0.003). CONCLUSIONS: Most physicians in our sample were willing to discharge actual and hypothetical patients from their practices. This tendency may have significant implications for the initiation of pay-for-performance programs. Physicians should be educated about the importance of the patient-physician relationship and their fiduciary obligations to the patient.
| S-EPMC2359473 | biostudies-other