Introduction: La réinstitution de l’alimentation entérale en deçà de 24h après une chirurgie digestive semble a priori conférer une diminution du risque d’infections de plaie, de pneumonies et de la durée de séjour. Le but de cette étude est de vérifier l’effet de la reprise précoce de l’alimentation entérale sur la durée de séjour hospitalier suite à une chirurgie colique. Méthodes: Il s’agit d’une étude prospective randomisée dans laquelle 95 patients ont été divisés aléatoirement en deux groupes. Dans le groupe contrôle, la diète est réintroduite lorsque le patient passe des gaz ou des selles per rectum, et qu’en plus il n’est ni nauséeux ni ballonné. Les patients du groupe expérimental reçoivent pour leur part une diète liquide dans les 12 heures suivant la chirurgie, puis une diète normale aux repas subséquents. L’objectif primaire de cette étude est de déterminer si la réinstitution précoce de l'alimentation entérale post chirurgie colique diminue la durée de séjour hospitalier lorsque comparée au régime traditionnel de réintroduction de l’alimentation. Les objectifs secondaires sont de quantifier l’effet de la réintroduction précoce de la diète sur les morbidités periopératoires et sur la reprise du transit digestif. Résultats: La durée de séjour hospitalier a semblé être légèrement diminuée dans le groupe expérimental (8,78±3,85 versus 9,41±5,22), mais cette difference n’était pas statistiquement significative. Des nausées ou des vomissements furent rapportés chez 24 (51%) patients du bras experimental et chez 30 (62.5%) patients du groupe contrôle. Un tube nasogastrique a du être installé chez un seul patient du groupe experimental. La morbidité périopératoire fut faible dans les deux groupes. Conclusion: Il semble sécuritaire de nourrir précocément les patients suite à une chirurgie colique. Cependant cette étude n’a pu démontrer un impact significatif de la reintroduction précoce de l alimentation per os sur la durée de séjour hospitalier. / Introduction: of early feeding within 24 hours of intestinal surgery seems advantageous in terms of reduction of wound infection, pneumonia and length of hospital stay. The aim of the study is to evaluate the impact of early enteral nutrition in length of hospital stay in comparison to traditional postoperative feeding regimen. Method: This prospective study enrolled 95 patients randomized in two groups: control group patients receive enteral feeding in absence of nausea or vomiting, abdominal distension and after passage of flatus or stools, while patients in experimental group were fed a liquid diet within 12 hours of surgery, followed by a regular diet at the next meal. The primary endpoint was the impact of early oral feeding on hospital length of stay. The secondary endpoint was to measure the impact of the diet reintroduction modality on the incidence of early postoperative morbidity and return of bowel function. Result: Length of hospital stay was slightly diminished in the experimental group compared to control (8,78±3,85 versus 9,41±5,22), but the difference was not statistically significant. Postoperative nausea and vomiting were reported in 24 (51,0%) patients in experimental group and 30 (62,5%) in control group. Only one patient required nasogastric tube insertion. The majority of patients did not demonstrate any postoperative morbidity in both groups. Conclusion: Early enteral nutrition is safe after intestinal surgery. However we did not demonstrate that early enteral feeding diminished length of hospital stay or hastened the return of bowel function.

Estudo retrospectivo de 119 pacientes com diagnóstico de adenocarcinoma ductal invasivo no estádio clínico III tratadas com quimioterapia neoadjuvante (FEC), mastectomia e adjuvância. Destas, 85 optaram por reconstrução mamária imediata (RMI) com retalho transverso músculo-cutâneo de reto-abdominal e 34, não. Com seguimento médio de 52,7 meses avaliou-se o tempo de hospitalização, a inter-relação com a adjuvância, recidiva local, o tempo livre de doença e o tempo total de sobrevida, concluindo-se que, apesar de aumentar o tempo de hospitalização, a RMI não interfere com os demais ítens, podendo ser indicada para pacientes portadoras de carcinoma mamário em estádio clínico avançado / A retrospective study with 119 patients diagnosed with invasive ductal adenocarcinoma of the breast treated with neoadjuvant chemotherapy (FEC), mastectomy and adjuvant therapy. Eight-five patients chose immediate breast reconstruction (IBR) with transverse rectus abdominis myocutaneous flap and, 34 did not do it. The mean follow-up was 52.7 months. Length of stay, adjuvant therapy interrelation, local recurrence, disease-free survival and overall survival were evaluated. It was concluded that despite a longer stay, IBR did not interfere with any of the other factors analyzed and may be indicated for patients with advanced breast disease

A identificação de reações adversas a medicamentos (RAM) nos hospitais constitui uma importante medida da morbidade associada a medicamentos e de seu ônus sobre o sistema de saúde. Este estudo observacional não intervencionista teve por objetivo avaliar a incidência de RAM em pacientes hospitalizados, as características clínicas das reações e fatores de risco associados. Foram avaliados 472 pacientes de cinco clínicas do Hospital das Clínicas da FMUSP (Clínica Médica, Cirurgia Geral, Neurologia, Geriatria, Alergia e Imunologia Clínica), com formação de coorte prospectiva, analisando as características demográficas, comorbidades, número de medicações utilizadas antes e durante a hospitalização e tempo de internação. A prevalência das RAM foi de 1,7% e a incidência geral de RAM foi 16,2%, variando conforme a clínica avaliada, sendo maior na Clínica Médica (30%). As reações mais frequentes foram as do tipo A, predominando as manifestações gastrointestinais. A maior parte das reações foi classificada de gravidade moderada. O maior número de medicações utilizadas por paciente, insuficiência renal crônica e tempo de internação foram fatores de risco para RAM, porém não houve associação das reações com idade avançada. Antecedente de RAM anterior à internação foi identificado como fator de proteção. A incidência de reações de hipersensibilidade a medicamentos (RHM) foi de 3,2%, com maior número de medicações utilizadas por paciente como único fator de risco isolado, sem associação com as clínicas avaliadas ou gênero dos pacientes. As medicações mais associadas às RAM e RHM foram os antibióticos, opióides e contrastes iodados. Os medicamentos mais prescritos foram os sintomáticos. O estudo concluiu que as RAM são frequentes e potencialmente evitáveis. O conhecimento da incidência e dos fatores associados pode estimular a prevenção. A prescrição de medicações para pacientes internados deve ser mais criteriosa, especialmente para os mais susceptíveis, evitando a polifarmácia / The detection of adverse drug reactions (ADRs) in hospitalized patients is an important measure of morbidity associated with drugs and its burden on the health system. The objective of this non-interventionist observational study was to assess the incidence of ADRs in hospitalized patients, the clinical characteristics of reactions and associated risk factors. We evaluated 472 patients from five medical specialties of the Hospital das Clínicas-FMUSP (Internal Medicine, General Surgery, Neurology, Geriatrics, Clinical Immunology and Allergy). We performed a prospective cohort, analyzing the demographics features, comorbidities, number of medications used before and during hospitalization and length of stay in the hospital. The prevalence of ADRs was 1.7% and the overall incidence of ADRs was 16.2%, varying according to the specialty assessed, higher in the Internal Medicine (30%). The most frequent reactions were type A, with gastrointestinal manifestations being the most frequent. Most of the reactions were classified as moderate in severity. The greater number of drugs used, chronic renal failure and longer hospital stays were risk factors for ADRs, but there was no association between reactions and age. History of previous ADRs to admission was identified as a protective factor. The incidence of hypersensitivity drug reactions (HDRs) was 3.2%, with the greater number of medications used per patient as the sole isolated risk factor, without association with specialty or patient\'s gender. The main medications associated with ADRs and HDRs were antibiotics, opioids and iodinated contrast media. The most commonly prescribed medications were symptomatic ones. The study concluded that the ADRs are frequent and potentially preventable. Knowledge of the incidence and associated factors can stimulate prevention. The pharmacotherapy of in-patients should be more careful, especially for the more susceptible patients, avoiding polypharmacy

INTRODUÇÃO: A despeito de todos os avanços no tratamento cirúrgico do câncer, ainda existem fatores complicadores para a adequada evolução no pós-operatório. Realizou-se estudo retrospectivo em uma população de pacientes submetidos a tratamento cirúrgico oncológico eletivo para o câncer retal, a fim de caracterizá-los epidemiologicamente e determinar quais variáveis estão associadas a maior morbimortalidade tais como aumento de permanência hospitalar, utilização de unidade de terapia intensiva, reoperações, óbito em 30 dias e custos hospitalares totais. MÉTODO: Foram selecionados todos os pacientes submetidos a cirurgias retais eletivas para o tratamento oncológico no período de 01 de outubro de 2008 a 31 de dezembro de 2012, resultando numa amostra de 405 pacientes de idade adulta e ambos os sexos. Realizou-se análise univariada para cada desfecho e análise multivariada através de regressão logística com ajustamento para as seguintes variáveis: sexo, idade, índice massa corpórea, estadiamento TNM, procedimento realizado, escore ASA, via de acesso, realização de QRT neoadjuvante, intervalo entre a neoadjuvância e a cirurgia e tempo cirúrgico, com validação interna através da técnica de bootstrap. RESULTADOS: Foram identificados como fatores de risco para aumento de permanência hospitalar a obesidade classe II e III (OR 15,44; IC95% 1,05 - 227,52; p = 0,05) e anestesia combinada - regional associada à geral (OR 5,38; IC95% 1,08 - 29,95; p = 0,04); amputação abdominoperineal foi fator de risco para reoperação em 30 dias (OR 4,54; IC95% 1,15 - 17,90; p = 0,03); amputação abdominoperineal (OR 5,38; IC95% 1,21 - 23,73; p = 0,03) e exenteração pélvica (OR 19,98; IC95% 0,99 - 401,32; p = 0,05) foram fatores de risco para utilização de UTI; idade acima de 79 anos (OR 13,99; IC95% 1,51 - 128,95; p = 0,02) foi fator de risco para complicação pós-operatória; idade acima de 79 anos (OR 0,07; IC95% 0,01 - 0,39; p = 0,01) foi fator de proteção para tempo cirúrgico elevado, obesidade classe II e III (OR 12,87; IC95% 1,54 - 107,67; p = 0,02) e via de acesso laparoscópica com transição para aberta (OR 8,7; IC95% 2,67 - 28,36; p < 0,001) foram fatores de risco para tempo cirúrgico elevado. CONCLUSÕES: Obesidade classe II e III e anestesia combinada são fatores de risco para tempo de permanência hospitalar prolongado para pacientes submetidos a cirurgia para tratamento de câncer retal. Outros estudos se fazem necessários para entender quais são os mecanismos que levam a anestesia combinada, eventualmente, a este aumento de permanência hospitalar / INTRODUCTION: Despite all the developments in the surgical treatment of cancer, there are still complicating factors for a correct postoperative evolution. There were conducted a retrospective study in a population of patients submitted to surgical rectal cancer treatment to characterize them epidemiologically and determine which variables are associated with increased morbimortality such as increased hospital length of stay, use of intensive care unit, reoperations, 30-day mortality and total hospital costs. METHOD: All patients submitted to colorectal surgery treatment for cancer during the period October 1st, 2008 to December 31th, 2012, resulting in a sample of 405 patients of both genders. Univariate analysis was conducted for each outcome and multivariate analysis through logistic regression with adjustment for the following variables: sex, age, body mass index, TNM stage, procedure performed, ASA score, laparoscopic or open surgery, neoadjuvant treatment, interval between the neoadjuvant therapy and the surgery, and operative time, with internal validation by the bootstrap technique. RESULTS: there were identified as risk factors for increased hospital stay, the obesity class II and III (OR 15.44; 95% CI 1.05-227.52; p = 0.05) and combined anesthesia - regional and general (OR 5.38; 95% CI 1.08-29.95; p = 0.04); abdominoperineal amputation was a risk factor for 30-day reoperation (OR 4.54; 1.15 95% CI-17.90; p = 0.03); abdominoperineal amputation (OR 5.38; 95% CI 1.21-23.73; p = 0.03) and pelvic exenteration (OR 19.98; 95% CI 0.99-401.32; p = 0.05) were risk factors for ICU use; age over 79 years (OR 13.99; CI 1.51-128.95; p = 0.02) was a risk factor for postoperative complication; age over 79 years (OR 0.07; CI 0.01-0.39; p = 0.01) was protective factor to prolonged operative time , obesity class II and III (OR 12.87; CI 1.54-107.67; p = 0.02) and laparoscopic approach with transition to open (OR 8.7; CI 2.67-28.36; p 0.001) were risk factors to prolonged operative time. CONCLUSIONS: Obese class II and III and combined anesthesia are risk factors for prolonged hospital stay for patients undergoing surgery for rectal cancer treatment. Further studies are needed to understand the mechanisms that lead combined anesthesia to increase hospital stay

Foram analisados os resultados retrospectivos do transplante de medula óssea alogênico em 100 pacientes portadores de leucemia mielóide crônica, divididos em dois grupos: Grupo I, de 49 pacientes que recebeu alta hospitalar após a pega medular, e grupo II, de 51 pacientes, que recebeu alta precocemente, antes da pega medular, para acompanhamento ambulatorial. Foram comparados os dias de ocupação de leitos hospitalares, a morbidade e a mortalidade entre os grupos de alta até o dia 100 pós transplante. Verificou-se que, no grupo de alta precoce, houve, significativamente, menos dias de ocupação de leitos hospitalares, sem aumento da morbidade e da mortalidade / We analyzed the results of a retrospective study of 100 patients with chronic myelogenous leukemia submitted to allogeneic stem cell transplantation in one of two settings: Group I, with 49 patients, transplanted in the traditional inpatient and group II, with 51 patients, in partial outpatient. We compared the median number of days spent in hospital, morbidity and mortality within 100 after bone marrow transplantation. We concluded that there was a significant reduction in the median of hospital length of stay in the partial outpatient group, without increasing morbidity and mortality

<p>Veno-arterijski gradijent ugljen-dioksida (Pv-aCO2) se smatra pokazateljem adekvatnosti microcirculatornog venskog protoka. U stanjima usporenog protoka dolazi do povećavanja Pv-aCO2 zbog fenomena zadržavanja CO2. Vrednost Pv-aCO2 predložena je kao dodatni cilj rane usmerene terapije kod pacijenata sa septičnim &scaron;okom. Cilj rada bilo je utvrditi postojanje korelacije promene Pv-aCO2 s promenom SOFA (&ldquo;Sequential Organ Failure Assessment&rdquo;) skora (delta SOFA) nakon primene rane ciljane terapije, kao i korelacije vrednosti različitih pokazatelja krvnog protoka unutar prvih 12 sati od početka lečenja pacijenata sa sepsom. Sekundarni cilj bilo je utvrditi postojanje korelacije Pv-aCO2 6 sati nakon početka rane ciljane terapije (T6) s dužinom boravka u intenzivnoj jedinici i ishodom lečenja. Prospektivnim, neintervencijskim ispitivanjem obuhvaćeno je 150 pacijenata sa sepsom ili septičnim &scaron;okom. Merenja serumskog laktata, saturacije kiseonikom me&scaron;ane venske krvi (ScvO2) i Pv-aCO2 vr&scaron;ena su na početku rane ciljane terapije (T0), posle 6 i 12 sati (T6, T12). Pv-aCO2 se računao kao razlika između parcijalnog pritiska ugljen dioksida arterijske i me&scaron;ane venske krvi. Vrednost SOFA skora određivana je u vremenu T0 i nakon 48 časova (T48). Pacijenti su za potrebe analize podeljeni u dve grupe na osnovu promene SOFA skora [(1) pacijenti kod kojih je do&scaron;lo do smanjenja SOFA skora (delta SOFA &lt; 0); (2) pacijenti kod kojih je smanjenje SOFA skora izostalo (delta SOFA &ge; 0)] i na osnovu vrednosti Pv-aCO2 u vremenu T6 [(1) pacijenti sa visokim Pv-aCO2 (&ge; 0.8 kPa); (2) pacijenti sa normalnim Pv-aCO2 (&lt; 0.8 kPa)]. Između dve grupe pacijenata, sa normalnim i visokim Pv-aCO2, statistički značajne razlike uočene su samo u odnosu na najvi&scaron;u vrednost respiratorne komponente SOFA skora (p=0.01). Uočena je statistički značajna korelacija između vrednosti Pv-aCO2 i laktata u vremenu T6 (r=0.2), Pv-aCO2 i ScvO2 u vremenu T0 (r=-0.4) i T12 (r=-0.24) kao i laktata i ScvO2 u vremenu T0 (r=-0.26) i T12 (r=-0.18). Analizom ponavljanih merenja nije utvrđena statistički značajna korelacija između promene vrednosti Pv-aCO2 unutar prvih 6 sati s promenom SOFA skora unutar prvih 48 sati nakon početka rane ciljane terapije (p=0.12). Utvrđeno je da su vrednosti Pv-aCO2 u vremenu T6 bile lo&scaron; prediktor smrtnog ishoda. Nisu utvrđene statistički značajne razlike u dužini boravka u intenzivnoj jedinici i ishodu lečenja u zavisnosti od vrednosti Pv-aCO2.</p> / <p>Central venous-arterial CO2 difference (Pv-aCO2) reflects adequacy of microcirculatory venous flow. Widening of Pv-aCO2 due to CO2-stagnant phenomenon is described in the low flow states. Pv-aCO2 was proposed as an additional resuscitation target for patients with septic shock.The aim of this study was to examine correlation between changes in Pv-aCO2 and SOFA score as well as different blood flow indices (lactate, mixed venous oxygen saturation) 12 hours after onset of resuscitation in patients with sepsis or septic shock. Secondary aim was to evaluate association of delta CO2 6 hours after onset of resuscitation and patient outcomes (length of stay in the ICU, mortality). Prospective observational study included 150 patients with sepsis. Simultaneous measurements of lactate, mixed venous oxygen saturation (ScvO2) and delta PCO2 were performed at onset of resuscitation (T0) and after 6 hours (T6). Delta PCO2 was calculated as a difference between arterial PCO2 and PCO2 from mixed venous blood. Organ dysfunction was evaluated with the Sequential Organ Failure Assessment (SOFA) score at T0 and after 48 hours (T48). Mortality was assessed after 28 days. For data analysis purposes two groups were created based on delta SOFA [(1) patients with SOFA score decrease (delta SOFA &lt;0); (2) patients without SOFA score decrease (delta SOFA &ge; 0)] and based on Pv-aCO2 [(1) patients with high Pv-aCO2 (&ge;0.8 kPa); (2) patients with normal Pv-aCO2 (&lt;0.8 kPa). Patients with high and normal Pv-aCO2 differed only with respect to highest respiratory SOFA score (p=0.01) Change in Pv-aCO2 between T0 and T6 was not in correlation with change in SOFA score between T0 and T48 (p=0.12). Moderate statistically significant correlation was found between Pv-aCO2 and lactate at T6 (r=0.2), and moderate inverse correlation between Pv-aCO2 and ScvO2 at T0 (r=-0.4) and T12 (r=-0.25) and ScvO2 and lactate at T0 (r=-0.27) and T12 (r=-0.18). Pv-aCO2 at T6 was not associated with 28-day mortality and length of stay in the ICU.</p>

Indiana University-Purdue University Indianapolis (IUPUI) / With the increasing availability of Electronic Health Records (EHRs) and advances in deep learning techniques, developing deep predictive models that use EHR data to solve healthcare problems has gained momentum in recent years. The majority of clinical predictive models benefit from structured data in EHR (e.g., lab measurements and medications). Still, learning clinical outcomes from all possible information sources is one of the main challenges when building predictive models. This work focuses mainly on two sources of information that have been underused by researchers; unstructured data (e.g., clinical notes) and a patient network. We propose a novel hybrid deep learning model, DeepNote-GNN, that integrates clinical notes information and patient network topological structure to improve 30-day hospital readmission prediction. DeepNote-GNN is a robust deep learning framework consisting of two modules: DeepNote and patient network. DeepNote extracts deep representations of clinical notes using a feature aggregation unit on top of a state-of-the-art Natural Language Processing (NLP) technique - BERT. By exploiting these deep representations, a patient network is built, and Graph Neural Network (GNN) is used to train the network for hospital readmission predictions. Performance evaluation on the MIMIC-III dataset demonstrates that DeepNote-GNN achieves superior results compared to the state-of-the-art baselines on the 30-day hospital readmission task. We extensively analyze the DeepNote-GNN model to illustrate the effectiveness and contribution of each component of it. The model analysis shows that patient network has a significant contribution to the overall performance, and DeepNote-GNN is robust and can consistently perform well on the 30-day readmission prediction task. To evaluate the generalization of DeepNote and patient network modules on new prediction tasks, we create a multimodal model and train it on structured and unstructured data of MIMIC-III dataset to predict patient mortality and Length of Stay (LOS). Our proposed multimodal model consists of four components: DeepNote, patient network, DeepTemporal, and score aggregation. While DeepNote keeps its functionality and extracts representations of clinical notes, we build a DeepTemporal module using a fully connected layer stacked on top of a one-layer Gated Recurrent Unit (GRU) to extract the deep representations of temporal signals. Independent to DeepTemporal, we extract feature vectors of temporal signals and use them to build a patient network. Finally, the DeepNote, DeepTemporal, and patient network scores are linearly aggregated to fit the multimodal model on downstream prediction tasks. Our results are very competitive to the baseline model. The multimodal model analysis reveals that unstructured text data better help to estimate predictions than temporal signals. Moreover, there is no limitation in applying a patient network on structured data. In comparison to other modules, the patient network makes a more significant contribution to prediction tasks. We believe that our efforts in this work have opened up a new study area that can be used to enhance the performance of clinical predictive models.

No description available.

Recent advancements in machine learning and the widespread adoption of electronic healthrecords have enabled breakthroughs for several predictive modelling tasks in health care. One such task that has seen considerable improvements brought by deep neural networks is length of stay (LOS) prediction, in which research has mainly focused on adult patients in the intensive care unit. This thesis uses multivariate time series extracted from the publicly available Medical Information Mart for Intensive Care III database to explore the potential of deep learning for classifying the remaining LOS of newborns in the neonatal intensive care unit (NICU) at each hour of the stay. To investigate this, this thesis describes experiments conducted with various deep learning models, including long short-term memory cells, gated recurrentunits, fully-convolutional networks and several composite networks. This work demonstrates that modelling the remaining LOS of newborns in the NICU as a multivariate time series classification problem naturally facilitates repeated predictions over time as the stay progresses and enables advanced deep learning models to outperform a multinomial logistic regression baseline trained on hand-crafted features. Moreover, it shows the importance of the newborn’s gestational age and binary masks indicating missing values as variables for predicting the remaining LOS. / Framstegen inom maskininlärning och det utbredda införandet av elektroniska hälsoregister har möjliggjort genombrott för flera prediktiva modelleringsuppgifter inom sjukvården. En sådan uppgift som har sett betydande förbättringar förknippade med djupa neurala nätverk är förutsägelsens av vistelsetid på sjukhus, men forskningen har främst inriktats på vuxna patienter i intensivvården. Den här avhandlingen använder multivariata tidsserier extraherade från den offentligt tillgängliga databasen Medical Information Mart for Intensive Care III för att undersöka potentialen för djup inlärning att klassificera återstående vistelsetid för nyfödda i den neonatala intensivvårdsavdelningen (neonatal-IVA) vid varje timme av vistelsen. Denna avhandling beskriver experiment genomförda med olika djupinlärningsmodeller, inklusive longshort-term memory, gated recurrent units, fully-convolutional networks och flera sammansatta nätverk. Detta arbete visar att modellering av återstående vistelsetid för nyfödda i neonatal-IVA som ett multivariat tidsserieklassificeringsproblem på ett naturligt sätt underlättar upprepade förutsägelser över tid och gör det möjligt för avancerade djupa inlärningsmodeller att överträffaen multinomial logistisk regressionsbaslinje tränad på handgjorda funktioner. Dessutom visar det vikten av den nyfödda graviditetsåldern och binära masker som indikerar saknade värden som variabler för att förutsäga den återstående vistelsetiden.

No / There are few data on the associations between anaemia, allogeneic blood transfusion (ABT), patient blood management, and outcome after arthroplasty in the UK. National agencies nevertheless instruct NHS Trusts to implement blood conservation measures including preoperative anaemia management. Internationally, blood management programmes show encouraging results. METHODS: We retrospectively audited 717 primary hip or knee arthroplasties in a UK general hospital and conducted regression analyses to identify outcome predictors. We used these data to modify previously published algorithms for UK practice and audited its introduction prospectively. The retrospective audit group served as a control. RESULTS: Preoperative haemoglobin (Hb) concentration predicted ABT (odds ratio 0.25 per 1 g dl(-1), P<0.001). It also predicted the length of stay (LOS, effect size -0.7 days per 1 g dl(-1), P=0.004) independently of ABT, including in non-anaemic patients. Patient blood management implementation was associated with lower ABT rates for hip (23-7%, P<0.001) and knee (7-0%, P=0.001) arthroplasty. LOS for total hip replacement and total knee replacement decreased from 6 (5-8) days to 5 (3-7) and 4 (3-6) days, respectively, after algorithm implementation (P<0.001). The all-cause re-admission rate within 90 days decreased from 13.5% (97/717) before to 8.2% (23/281) after algorithm implementation (P=0.02). CONCLUSIONS: We conclude that preoperative Hb predicts markers of arthroplasty outcome in UK practice. A systematic approach to optimize Hb mass before arthroplasty and limit Hb loss perioperatively was associated with improved outcome up to 90 days after discharge.