Systematic development and validation of a predictive model for major postoperative complications in the Peri-operative Quality Improvement Project (PQIP) dataset.

Complications are common following major surgery and are associated with increased use of healthcare resources, disability and mortality. Continued reliance on mortality estimates risks harming patients and health systems, but existing tools for predicting complications are unwieldy and inaccurate. We aimed to systematically construct an accurate pre-operative model for predicting major postoperative complications; compare its performance against existing tools; and identify sources of inaccuracy in predictive models more generally. Complete patient records from the UK Peri-operative Quality Improvement Programme dataset were analysed. Major complications were defined as Clavien-Dindo grade ≥ 2 for novel models. In a 75% train:25% test split cohort, we developed a pipeline of increasingly complex models, prioritising pre-operative predictors using Least Absolute Shrinkage and Selection Operators (LASSO). We defined the best model in the training cohort by the lowest Akaike's information criterion, balancing accuracy and simplicity. Of the 24,983 included cases, 6389 (25.6%) patients developed major complications. Potentially modifiable risk factors (pain, reduced mobility and smoking) were retained. The best-performing model was highly complex, specifying individual hospital complication rates and 11 patient covariates. This novel model showed substantially superior performance over generic and specific prediction models and scores. We have developed a novel complications model with good internal accuracy, re-prioritised predictor variables and identified hospital-level variation as an important, but overlooked, source of inaccuracy in existing tools. The complexity of the best-performing model does, however, highlight the need for a step-change in clinical risk prediction to automate the delivery of informative risk estimates in clinical systems.

Development and validation of a prognostic model for death 30 days after adult emergency laparotomy.

The probability of death after emergency laparotomy varies greatly between patients. Accurate pre-operative risk prediction is fundamental to planning care and improving outcomes. We aimed to develop a model limited to a few pre-operative factors that performed well irrespective of surgical indication: obstruction; sepsis; ischaemia; bleeding; and other. We derived a model with data from the National Emergency Laparotomy Audit for patients who had emergency laparotomy between December 2016 and November 2018. We tested the model on patients who underwent emergency laparotomy between December 2018 and November 2019. There were 4077/40,816 (10%) deaths 30 days after surgery in the derivation cohort. The final model had 13 pre-operative variables: surgical indication; age; blood pressure; heart rate; respiratory history; urgency; biochemical markers; anticipated malignancy; anticipated peritoneal soiling; and ASA physical status. The predicted mortality probability deciles ranged from 0.1% to 47%. There were 1888/11,187 deaths in the test cohort. The scaled Brier score, integrated calibration index and concordance for the model were 20%, 0.006 and 0.86, respectively. Model metrics were similar for the five surgical indications. In conclusion, we think that this prognostic model is suitable to support decision-making before emergency laparotomy as well as for risk adjustment for comparing organisations.

Predicting severe pain after major surgery: a secondary analysis of the Peri-operative Quality Improvement Programme (PQIP) dataset.

Acute postoperative pain is common, distressing and associated with increased morbidity. Targeted interventions can prevent its development. We aimed to develop and internally validate a predictive tool to pre-emptively identify patients at risk of severe pain following major surgery. We analysed data from the UK Peri-operative Quality Improvement Programme to develop and validate a logistic regression model to predict severe pain on the first postoperative day using pre-operative variables. Secondary analyses included the use of peri-operative variables. Data from 17,079 patients undergoing major surgery were included. Severe pain was reported by 3140 (18.4%) patients; this was more prevalent in females, patients with cancer or insulin-dependent diabetes, current smokers and in those taking baseline opioids. Our final model included 25 pre-operative predictors with an optimism-corrected c-statistic of 0.66 and good calibration (mean absolute error 0.005, p = 0.35). Decision-curve analysis suggested an optimal cut-off value of 20-30% predicted risk to identify high-risk individuals. Potentially modifiable risk factors included smoking status and patient-reported measures of psychological well-being. Non-modifiable factors included demographic and surgical factors. Discrimination was improved by the addition of intra-operative variables (likelihood ratio χ2 496.5, p < 0.001) but not by the addition of baseline opioid data. On internal validation, our pre-operative prediction model was well calibrated but discrimination was moderate. Performance was improved with the inclusion of peri-operative covariates suggesting pre-operative variables alone are not sufficient to adequately predict postoperative pain.

Development and internal validation of a model for postoperative morbidity in adults undergoing major elective colorectal surgery: the peri-operative quality improvement programme (PQIP) colorectal risk model.

Over 1.5 million major surgical procedures take place in the UK NHS each year and approximately 25% of patients develop at least one complication. The most widely used risk-adjustment model for postoperative morbidity in the UK is the physiological and operative severity score for the enumeration of mortality and morbidity. However, this model was derived more than 30 years ago and now overestimates the risk of morbidity. In addition, contemporary definitions of some model predictors are markedly different compared with when the tool was developed. A second model used in clinical practice is the American College of Surgeons National Surgical Quality Improvement Programme risk model; this provides a risk estimate for a range of postoperative complications. This model, widely used in North America, is not open source and therefore cannot be applied to patient populations in other settings. Data from a prospective multicentre clinical dataset of 118 NHS hospitals (the peri-operative quality improvement programme) were used to develop a bespoke risk-adjustment model for postoperative morbidity. Patients aged ≥ 18 years who underwent colorectal surgery were eligible for inclusion. Postoperative morbidity was defined using the postoperative morbidity survey at postoperative day 7. Thirty-one candidate variables were considered for inclusion in the model. Death or morbidity occurred by postoperative day 7 in 3098 out of 11,646 patients (26.6%). Twelve variables were incorporated into the final model, including (among others): Rockwood clinical frailty scale; body mass index; and index of multiple deprivation quintile. The C-statistic was 0.672 (95%CI 0.660-0.684), with a bootstrap optimism corrected C-statistic of 0.666 at internal validation. The model demonstrated good calibration across the range of morbidity estimates with a mean slope gradient of predicted risk of 0.959 (95%CI 0.894-1.024) with an index-corrected intercept of -0.038 (95%CI -0.112-0.036) at internal validation. Our model provides parsimonious case-mix adjustment to quantify risk of morbidity on postoperative day 7 for a UK population of patients undergoing major colorectal surgery. Despite the C-statistic of < 0.7, our model outperformed existing risk-models in widespread use. We therefore recommend application in case-mix adjustment, where incorporation into a continuous monitoring tool such as the variable life adjusted display or exponentially-weighted moving average-chart could support high-level monitoring and quality improvement of risk-adjusted outcome at the population level.

The association of pre-operative anaemia with morbidity and mortality after emergency laparotomy.

Pre-operative anaemia is associated with poor outcomes after elective surgery but its relationship with outcomes after emergency surgery is unclear. We analysed National Emergency Laparotomy Audit data from 1 December 2013 to 30 November 2017, excluding laparotomy for haemorrhage. Anaemia was classified as 'mild' 129-110 g.l-1; 'moderate' 109-80 g.l-1; or 'severe' ≤ 79 g.l-1. The primary outcome was 90-day mortality. Secondary outcomes were 30-day mortality, return to theatre and postoperative hospital stay. The primary outcome was available for 86,763 patients, of whom 45,306 (52%) were anaemic. There were 12,667 (15%) deaths at 90 postoperative days and 9246 (11%) deaths at 30 postoperative days. Anaemia was associated with increased 90-day and 30-day mortality, odds ratio (95%CI): mild, 1.15 (1.09-1.21); moderate, 1.44 (1.36-1.52); and severe, 1.42 (1.24-1.63), p < 0.001 for all; mild, 1.07 (1.00-1.12), p = 0.030; moderate, 1.30 (1.21-1.38), p < 0.001; and severe, 1.22 (1.05-1.43), p = 0.010, respectively. All categories of anaemia were associated with prolonged hospital stay, adjusted coefficient (95%CI): mild, 1.31 (1.01-1.62); moderate, 3.41 (3.04-3.77); severe, 2.80 (1.83-3.77), p < 0.001 for all. Moderate and severe anaemia were associated with increased risk of return to the operating theatre, odds ratio (95%CI): moderate 1.13 (1.06-1.21), p < 0.001; and severe 1.23 (1.06-1.43), p = 0.006. Pre-operative anaemia is common in patients undergoing emergency laparotomy and is associated with increased postoperative mortality and morbidity.

A survey of antenatal and peripartum provision of information on analgesia and anaesthesia.

Pregnant women should receive information about what they might expect to experience during their delivery. Despite this, research shows many women are inadequately prepared for anaesthetic interventions during labour. We surveyed 903 postnatal women across 28 Greater London hospitals about: the analgesic and anaesthetic information that they recalled receiving during pregnancy and delivery; their confidence to make decisions on analgesia; and their satisfaction with the analgesia used. Wide variation was observed between hospitals. Overall, 67 of 749 (9.0%) women recalled receiving antenatal information covering all aspects of labour analgesia, and 108 of 889 (12.1%) covering anaesthesia for caesarean section. Regarding intrapartum information, 256 of 415 (61.7%) respondents recalled receiving thorough information before epidural insertion for labour analgesia, and 102 of 370 (27.6%) before anaesthesia for caesarean section. We found that 620 of 903 (68.7%) women felt well enough informed to be confident in their analgesic choices, and 675 of 903 (74.8%) stated that their analgesia was as expected or better. Receiving information verbally, regardless of provider, was the factor most strongly associated with respondents recalling receiving full information: odds ratio (95%CI) for labour analgesia 20.66 (8.98-47.53; p < 0.0001); epidural top-up for caesarean section 5.93 (1.57-22.35; p = 0.01); and general anaesthesia for caesarean section 12.39 (2.18-70.42; p = 0.01). A large proportion of respondents did not recall being fully informed before an anaesthetic intervention. Collaboration with current antenatal service providers, both in promoting information delivery and providing resources to assist with delivery, could improve the quality of information offered and women's retention of that information.

Organisational factors and mortality after an emergency laparotomy: multilevel analysis of 39 903 National Emergency Laparotomy Audit patients.

BACKGROUND: Studies across healthcare systems have demonstrated between-hospital variation in survival after an emergency laparotomy. We postulate that this variation can be explained by differences in perioperative process delivery, underpinning organisational structures, and associated hospital characteristics. METHODS: We performed this nationwide, registry-based, prospective cohort study using data from the National Emergency Laparotomy Audit organisational and patient audit data sets. Outcome measures were all-cause 30- and 90-day postoperative mortality. We estimated adjusted odds ratios (ORs) for perioperative processes and organisational structures and characteristics by fitting multilevel logistic regression models. RESULTS: The cohort comprised 39 903 patients undergoing surgery at 185 hospitals. Controlling for case mix and clustering, a substantial proportion of between-hospital mortality variation was explained by differences in processes, infrastructure, and hospital characteristics. Perioperative care pathways [OR: 0.86; 95% confidence interval (CI): 0.76-0.96; and OR: 0.89; 95% CI: 0.81-0.99] and emergency surgical units (OR: 0.89; 95% CI: 0.80-0.99; and OR: 0.89; 95% CI: 0.81-0.98) were associated with reduced 30- and 90-day mortality, respectively. In contrast, infrequent consultant-delivered intraoperative care was associated with increased 30- and 90-day mortality (OR: 1.61; 95% CI: 1.01-2.56; and OR: 1.61; 95% CI: 1.08-2.39, respectively). Postoperative geriatric medicine review was associated with substantially lower mortality in older (≥70 yr) patients (OR: 0.35; 95% CI: 0.29-0.42; and OR: 0.64; 95% CI: 0.55-0.73, respectively). CONCLUSIONS: This multicentre study identified low-technology, readily implementable structures and processes that are associated with improved survival after an emergency laparotomy. Key components of pathways, perioperative medicine input, and specialist units require further investigation.

Development and internal validation of a novel risk adjustment model for adult patients undergoing emergency laparotomy surgery: the National Emergency Laparotomy Audit risk model.

BACKGROUND: Among patients undergoing emergency laparotomy, 30-day postoperative mortality is around 10-15%. The risk of death among these patients, however, varies greatly because of their clinical characteristics. We developed a risk prediction model for 30-day postoperative mortality to enable better comparison of outcomes between hospitals. METHODS: We analysed data from the National Emergency Laparotomy Audit (NELA) on patients having an emergency laparotomy between December 2013 and November 2015. A prediction model was developed using multivariable logistic regression, with potential risk factors identified from existing prediction models, national guidelines, and clinical experts. Continuous risk factors were transformed if necessary to reflect their non-linear relationship with 30-day mortality. The performance of the model was assessed in terms of its calibration and discrimination. Interval validation was conducted using bootstrap resampling. RESULTS: There were 4458 (11.5%) deaths within 30-days among the 38 830 patients undergoing emergency laparotomy. Variables associated with death included (among others): age, blood pressure, heart rate, physiological variables, malignancy, and ASA physical status classification. The predicted risk of death among patients ranged from 1% to 50%. The model demonstrated excellent calibration and discrimination, with a C-statistic of 0.863 (95% confidence interval, 0.858-0.867). The model retained its high discrimination during internal validation, with a bootstrap derived C-statistic of 0.861. CONCLUSIONS: The NELA risk prediction model for emergency laparotomies discriminates well between low- and high-risk patients and is suitable for producing risk-adjusted provider mortality statistics.

Perioperative risk.

The high-risk surgical patient is a growing challenge to modern health care. This cohort, although comprising only 10-15% of surgical procedures, accounts for approximately 80% of postoperative deaths and suffers a high rate of postoperative morbidity. Developing robust systems to help identify and better manage this patient group should be a priority. Risk stratification has become a valuable clinical tool for shared decision-making and the development of individualized care plans. Methods for stratifying individual risk include assessment tools, measures of functional capacity and plasma biomarker assays. Routine evaluation of perioperative risk is central to the delivery of high quality, appropriate surgical care.

Predicting postoperative morbidity in adult elective surgical patients using the Surgical Outcome Risk Tool (SORT).

BACKGROUND: The Surgical Outcome Risk Tool (SORT) is a risk stratification instrument used to predict perioperative mortality. We wanted to evaluate and refine SORT for better prediction of the risk of postoperative morbidity. METHODS: We analysed prospectively collected data from a single-centre cohort of adult patients undergoing major elective surgery. The data set was split randomly into derivation and validation samples. We used logistic regression to construct a model in the derivation sample to predict postoperative morbidity as defined using the validated Postoperative Morbidity Survey (POMS) assessed at 1 week after surgery. Performance of this 'SORT-morbidity' model was then tested in the validation sample and compared against the Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM). RESULTS: The SORT-morbidity model was constructed using a derivation sample of 1056 patients and validated in a further 527 patients. SORT-morbidity was well calibrated in the validation sample, as assessed using calibration plots and the Hosmer-Lemeshow test (χ 2 =4.87, P =0.77). It showed acceptable discrimination by receiver operating characteristic curve analysis [area under the receiver operating characteristic curve (AUROC)=0.72, 95% confidence interval: 0.67-0.77]. This compared favourably with POSSUM (AUROC=0.66, 95% confidence interval: 0.60-0.71), whilst being simpler to use. Linear shrinkage factors were estimated, which allow the SORT-morbidity model to predict a range of alternative morbidity outcomes with greater accuracy, including low- and high-grade morbidity, and POMS at later time points. CONCLUSIONS: SORT-morbidity can be used before surgery, with clinical judgement, to predict postoperative morbidity risk in major elective surgery.

Risk assessment tools validated for patients undergoing emergency laparotomy: a systematic review.

Emergency laparotomies are performed commonly throughout the world, but one in six patients die within a month of surgery. Current international initiatives to reduce the considerable associated morbidity and mortality are founded upon delivering individualised perioperative care. However, while the identification of high-risk patients requires the routine assessment of individual risk, no method of doing so has been demonstrated to be practical and reliable across the commonly encountered spectrum of presentations, co-morbidities and operative procedures. A systematic review of Embase and Medline identified 20 validation studies assessing 25 risk assessment tools in patients undergoing emergency laparotomy. The most frequently studied general tools were APACHE II, ASA-PS and P-POSSUM. Comparative, quantitative analysis of tool performance was not feasible due to the heterogeneity of study design, poor reporting and infrequent within-study statistical comparison of tool performance. Reporting of calibration was notably absent in many prognostic tool validation studies. APACHE II demonstrated the most consistent discrimination of individual outcome across a variety of patient groups undergoing emergency laparotomy when used either preoperatively or postoperatively (area under the curve 0.76-0.98). While APACHE systems were designed for use in critical care, the ability of APACHE II to generate individual risk estimates from objective, exclusively preoperative data items may lead to better-informed shared decisions, triage and perioperative management of patients undergoing emergency laparotomy. Future endeavours should include the recalibration of APACHE II and P-POSSUM in contemporary cohorts, modifications to enable prediction of morbidity and assessment of the impact of adoption of these tools on clinical practice and patient outcomes.

Enhanced recovery from surgery in the UK: an audit of the enhanced recovery partnership programme 2009-2012.

BACKGROUND: The UK Department of Health Enhanced Recovery Partnership Programme collected data on 24 513 surgical patients in the UK from 2009-2012. Enhanced Recovery is an approach to major elective surgery aimed at minimizing perioperative stress for the patient. Previous studies have shown Enhanced Recovery to be associated with reduced hospital length of stay and perioperative morbidity. METHODS: In this national clinical audit, National Health Service hospitals in the UK were invited to submit patient-level data. The data regarding length of stay and compliance with each element of Enhanced Recovery protocols for colorectal, orthopaedic, urological and gynaecological surgery patients were analysed. The relationship between Enhanced Recovery protocol compliance and length of stay was measured. RESULTS: From 16 267 patients from 61 hospital trusts, three out of four surgical specialties showed Enhanced Recovery, compliance being weakly associated with shorter length of stay (correlation coefficients -0.18, -0.14, -0.25 in colorectal, orthopaedics and gynaecology respectively). At a cut-off of 80% compliance, good compliance was associated with two, one and three day reductions in median length of stay respectively in colorectal, orthopaedic and urological surgeries, with no saving in gynaecology. CONCLUSIONS: This study is the largest assessment of the relationship between Enhanced Recovery protocol compliance and outcome in four surgical specialties. The data suggest that higher compliance with an Enhanced Recovery protocol has a weak association with shorter length of stay. This suggests that changes in process, resulting from highly protocolised pathways, may be as important in reducing perioperative length of stay as any individual element of Enhanced Recovery protocols in isolation.

Correlation between clinical success and apical dye penetration.

AIM: This study was undertaken to examine whether a correlation exists between apical dye penetration and the clinical performance of root fillings. METHODOLOGY: Apical dye penetration into 116 roots of human teeth that had been root-filled at least 6 months prior to extraction was tested in vitro using a vacuum technique and by measuring the length of dye penetration. Endodontic treatment was classified as clinically successful or unsuccessful and results for these groups were compared using analysis of variance and the Student's t-test. Positive and negative controls were used to test the experimental system. RESULTS: All controls performed as expected. Dye penetrated significantly further in unsuccessful cases although the raw data suggested little difference. Overall, dye penetrated 99.5% of the specimens, indicating that the presence of dye in the canal is a poor indicator of whether the technique or material will succeed. However, the extent of dye penetration may be related to the clinical outcome. CONCLUSIONS: Clinically placed root canal fillings do not provide an apical seal that prevents fluid penetration. The outcome of treatment cannot be predicted from the results of apical dye leakage studies.

An in vitro study of apical and coronal microleakage of laterally condensed gutta percha with Ketac-Endo and AH-26.

The purpose of this in vitro study was to compare both apical and coronal dye penetration when Ketac-Endo and AH-26 sealers were used with laterally condensed gutta percha. Crowns were removed from 28 teeth and the root canals were biomechanically prepared. The teeth were divided into two groups of 12-teeth each and a control group of 4 teeth. Root canals in the two experimental groups were filled with laterally condensed gutta percha and either Ketac-Endo or AH-26 sealer. The Ketac-Endo group had the coronal 3 mm of gutta percha and sealer removed and the resultant cavity was filled with Ketac-Endo alone. After the sealers had set, the root surfaces were coated with nail varnish except at the apex and at the coronal end. Positive controls had no root fillings and were coated with nail varnish in the same manner while the negative controls were sealed apically and coronally with Cavit prior to sealing the entire external root surface with nail varnish. Specimens were placed in 2% methylene blue dye in a vacuum of 660 mm of mercury for five minutes and then left immersed for a further two days. The roots were vertically sectioned to determine the following mean levels of dye penetration: Ketac-Endo, 1.08 mm apically and 6.29 mm coronally; AH-26, 0.75 mm apically and 6.67 mm coronally. Positive controls had total leakage and negative controls had no leakage. This study demonstrated that the apical and coronal seals obtained with Ketac-Endo and AH-26 were not significantly different although the apical seal obtained with each material was significantly better than the corresponding coronal seal.

Entrapped air and its effects on dye penetration of voids.

This study was conducted to determine the effect of entrapped air on the ability of methylene blue dye to penetrate known and measurable voids. Thirty six capillary tubes were sealed at one end and had standardized voids created within the lumen. Three equal groups of 12 test specimens and 4 controls were subjected to dye penetration by the methods of passive immersion for 7 days, centrifuging at 3,000 rpm for 5 minutes and immersion in a vacuum of 26 mm of mercury for 5 minutes. Measurement of linear dye penetration was made by removing shavings from the sealed end until dye could be observed. Negative controls showed no dye penetration and positive controls showed complete penetration. The vacuum method showed 100% penetration in all cases, while centrifuging had a mean of 91.7 /+ 8.7% penetration and passive immersion had a mean of 20.7 +/- 5.4% penetration. This study demonstrates that passive dye leakage studies are unreliable and very variable. The vacuum method whereby all entrapped air is removed is the most reliable method for dye penetration studies.

Gentle birth: its safety and its effect on neonatal behavior.

Leboyer describes a method of handling babies at birth designed to decrease initial stimulation and lessen trauma for the infant. Twenty clinic patients were delivered using a modification of this method. They were compared to 17 control patients delivered in the customary manner. Significant differences were noted between the two groups in five of 18 neonate behaviors observed in the delivery room. Among control subjects there was significantly more time spent with hand muscles tense during the first ten minutes after birth; more trembling, shuddering, and blinking during the first 15 minutes after deliver; and more trembling and shuddering overall. Experimental subjects spent significantly more time with hand muscles relaxed during the first ten minutes after birth and more time with open eyes overall. Results of this study indicate the Leboyer method is safe for mother and infant and promotes infant relaxation in the delivery room.