Preoperative Prediction of Postoperative Infections Using Machine Learning and Electronic Health Record Data.

OBJECTIVE: To estimate preoperative risk of postoperative infections using structured electronic health record (EHR) data. BACKGROUND: Surveillance and reporting of postoperative infections is primarily done through costly, labor-intensive manual chart reviews on a small sample of patients. Automated methods using statistical models applied to postoperative EHR data have shown promise to augment manual review as they can cover all operations in a timely manner. However, there are no specific models for risk-adjusting infectious complication rates using EHR data. METHODS: Preoperative EHR data from 30,639 patients (2013-2019) were linked to the American College of Surgeons National Surgical Quality Improvement Program preoperative data and postoperative infection outcomes data from 5 hospitals in the University of Colorado Health System. EHR data included diagnoses, procedures, operative variables, patient characteristics, and medications. Lasso and the knockoff filter were used to perform controlled variable selection. Outcomes included surgical site infection, urinary tract infection, sepsis/septic shock, and pneumonia up to 30 days postoperatively. RESULTS: Among >15,000 candidate predictors, 7 were chosen for the surgical site infection model and 6 for each of the urinary tract infection, sepsis, and pneumonia models. Important variables included preoperative presence of the specific outcome, wound classification, comorbidities, and American Society of Anesthesiologists physical status classification. The area under the receiver operating characteristic curve for each model ranged from 0.73 to 0.89. CONCLUSIONS: Parsimonious preoperative models for predicting postoperative infection risk using EHR data were developed and showed comparable performance to existing American College of Surgeons National Surgical Quality Improvement Program risk models that use manual chart review. These models can be used to estimate risk-adjusted postoperative infection rates applied to large volumes of EHR data in a timely manner.

The deimplementation of laboratory testing in low-risk patients as recommended by the American society of anesthesiologists: An ACS-NSQIP longitudinal analysis.

BACKGROUND: In 2012, the American Society of Anesthesiologists (ASA) published guidelines recommending against routine preoperative laboratory testing for low-risk patients to reduce unnecessary medical expenditures. The aim of this study was to assess the change in routine preoperative laboratory testing in low-risk versus higher-risk patients before and after release of these guidelines. METHODS: The ACS-NSQIP database, 2005-2018, was separated into low-risk versus higher-risk patients based upon a previously published stratification. The guideline implementation date was defined as January 2013. Changes in preoperative laboratory testing over time were compared between low- and higher-risk patients. A difference-in-differences model was applied. The primary outcome included any laboratory test obtained ≤90 days prior to surgery. RESULTS: Of 7,507,991 patients, 972,431 (13.0%) were defined as low-risk and 6,535,560 (87.0%) higher-risk. Use of any preoperative laboratory test declined in low-risk patients from 66.5% before to 59.6% after guidelines, a 6.9 percentage point reduction, versus 93.0%-91.9% in higher-risk patients, a 1.1 percentage point reduction (p < 0.0001, comparing percentage point reductions). After risk-adjustment, the adjusted odds ratio for having any preoperative laboratory test after versus before the guidelines was 0.77 (95% CI 0.76-0.78) in low-risk versus 0.93 (0.92-0.94) in higher-risk patients. In low-risk patients, lack of any preoperative testing was not associated with worse outcomes. CONCLUSIONS: While a majority of low-risk patients continue to receive preoperative laboratory testing not recommended by the ASA, there has been a decline after implementation of guidelines. Continued effort should be directed at the deimplementation of routine preoperative laboratory testing for low-risk patients.

Preoperative Prediction of Unplanned Reoperation in a Broad Surgical Population.

INTRODUCTION: Unplanned reoperation is an undesirable outcome with considerable risks and an increasingly assessed quality of care metric. There are no preoperative prediction models for reoperation after an index surgery in a broad surgical population in the literature. The Surgical Risk Preoperative Assessment System (SURPAS) preoperatively predicts 12 postoperative adverse events using 8 preoperative variables, but its ability to predict unplanned reoperation has not been assessed. This study's objective was to determine whether the SURPAS model could accurately predict unplanned reoperation. METHODS: This was a retrospective analysis of the American College of Surgeons' National Surgical Quality Improvement Program adult database, 2012-2018. An unplanned reoperation was defined as any unintended operation within 30 d of an initial scheduled operation. The 8-variable SURPAS model and a 29-variable "full" model, incorporating all available American College of Surgeons' National Surgical Quality Improvement Program nonlaboratory preoperative variables, were developed using multiple logistic regression and compared using discrimination and calibration metrics: C-indices (C), Hosmer-Lemeshow observed-to-expected plots, and Brier scores (BSs). The internal chronological validation of the SURPAS model was conducted using "training" (2012-2017) and "test" (2018) datasets. RESULTS: Of 5,777,108 patients, 162,387 (2.81%) underwent an unplanned reoperation. The SURPAS model's C-index of 0.748 was 99.20% of that for the full model (C = 0.754). Hosmer-Lemeshow plots showed good calibration for both models and BSs were similar (BS = 0.0264, full; BS = 0.0265, SURPAS). Internal chronological validation results were similar for the training (C = 0.749, BS = 0.0268) and test (C = 0.748, BS = 0.0250) datasets. CONCLUSIONS: The SURPAS model accurately predicted unplanned reoperation and was internally validated. Unplanned reoperation can be integrated into the SURPAS tool to provide preoperative risk assessment of this outcome, which could aid patient risk education.

Does Work Relative Value Unit Measure Surgical Complexity for Risk Adjustment of Surgical Outcomes?

INTRODUCTION: The purpose of this study was to determine whether the work relative value unit (workRVU) of a patient's operation can be useful as a measure of surgical complexity for the risk adjustment of surgical outcomes. METHODS: We retrospectively analyzed the American College of Surgeon's National Surgical Quality Improvement Program database (2005-2018). We examined the associations of workRVU of the patient's primary operation with preoperative patient characteristics and associations with postoperative complications. We performed forward selection multiple logistic regression analysis to determine the predictive importance of workRVU. We then generated prediction models using patient characteristics with and without workRVU and compared c-indexes to assess workRVU's additive predictive value. RESULTS: 7,507,991 operations were included. Patients who were underweight, functionally dependent, transferred from an acute care hospital, had higher American Society of Anesthesiologists class or who had medical comorbidities had operations with higher workRVU (all P < 0.0001). The subspecialties with the highest workRVU were neurosurgery (mean = 22.2), thoracic surgery (mean = 21.1), and vascular surgery (mean = 18.8) (P < 0.0001). For all postoperative complications, mean workRVU was higher for patients with the complication than those without (all P < 0.0001). For eight of 12 postoperative complications, workRVU entered the logistic regression models as a predictor variable in the 1st to 4th steps. Addition of workRVU as a preoperative predictive variable improved the c-index of the prediction models. CONCLUSIONS: WorkRVU was associated with sicker patients and patients experiencing postoperative complications and was an important predictor of postoperative complications. When added to a prediction model including patient characteristics, it only marginally improved prediction. This is possibly because workRVU is associated with patient characteristics.

Inpatient Versus Outpatient Surgery: A Comparison of Postoperative Mortality and Morbidity in Elective Operations.

BACKGROUND: Operations performed outpatient offer several benefits. The prevalence of outpatient operations is growing. Consequently, the proportion of patients with multiple comorbidities undergoing outpatient surgery is increasing. We compared 30-day mortality and overall morbidity between outpatient and inpatient elective operations. METHODS: Using the 2005-2018 ACS-NSQIP database, we evaluated trends in percent of hospital outpatient operations performed over time, and the percent of operations done outpatient versus inpatient by CPT code. Patient characteristics were compared for outpatient versus inpatient operations. We compared unadjusted and risk-adjusted 30-day mortality and morbidity for inpatient and outpatient operations. RESULTS: A total of 6,494,298 patients were included. The proportion of outpatient operations increased over time, from 37.8% in 2005 to 48.2% in 2018. We analyzed the 50 most frequent operations performed outpatient versus inpatient 25-75% of the time (n = 1,743,097). Patients having outpatient operations were younger (51.6 vs 54.6 years), female (70.3% vs 67.3%), had fewer comorbidities, and lower ASA class (I-II, 69.3% vs. 59.9%). On both unadjusted and risk-adjusted analysis, 30-day mortality and overall morbidity were less likely in outpatient versus inpatient operations. CONCLUSION: In this large multi-specialty analysis, we found that patients undergoing outpatient surgery had lower risk of 30-day morbidity and mortality than those undergoing the same inpatient operation. Patients having outpatient surgery were generally healthier, suggesting careful patient selection occurred even with increasing outpatient operation frequency. Patients and providers can feel reassured that outpatient operations are a safe, reasonable option for selected patients.

Using the Surgical Risk Preoperative Assessment System to Define the "High Risk" Surgical Patient.

BACKGROUND: Defining a "high risk" surgical population remains challenging. Using the Surgical Risk Preoperative Assessment System (SURPAS), we sought to define "high risk" groups for adverse postoperative outcomes. MATERIALS AND METHODS: We retrospectively analyzed the 2009-2018 American College of Surgeons National Surgical Quality Improvement Program database. SURPAS calculated probabilities of 12 postoperative adverse events. The Hosmer Lemeshow graphs of deciles of risk and maximum Youden index were compared to define "high risk." RESULTS: Hosmer-Lemeshow plots suggested the "high risk" patient could be defined by the 10th decile of risk. Maximum Youden index found lower cutoff points for defining "high risk" patients and included more patients with events. This resulted in more patients classified as "high risk" and higher number needed to treat to prevent one complication. Some specialties (thoracic, vascular, general) had more "high risk" patients, while others (otolaryngology, plastic) had lower proportions. CONCLUSIONS: SURPAS can define the "high risk" surgical population that may benefit from risk-mitigating interventions.

Associations Between Preoperative Risk, Postoperative Complications, and 30-Day Mortality.

BACKGROUND: Comorbidities and postoperative complications increase mortality, making early recognition and management critical. It is useful to understand how they are associated with one another. This study assesses associations between comorbidities, complications, and mortality. METHODS: We calculated associations between comorbidities, complications, and 30-day mortality using the 2012-2018 ACS-NSQIP database. We examined the association between mortality and number of complications which complications were most associated with mortality. RESULTS: 5,777,108 patients were included. 30-day mortality was 0.95%. For most comorbidities or postoperative complications, patients with these had higher mortality than patients without. Having ≥ 1 complication increased mortality risk by 32.5-fold (6.5% vs. 0.2%). Mortality rate significantly increased with increasing number of complications, particularly after two or more complications. Bleeding and sepsis were associated with the most deaths. CONCLUSION: The 30-day mortality rate was < 1% but was 32-fold higher in patients with complications and increased rapidly for patients with ≥ 2 complications. Bleeding and sepsis were the most prominent complications associated with mortality.

Postoperative pulmonary complications with adjuvant regional anesthesia versus general anesthesia alone: a sub-analysis of the Perioperative Research Network study.

BACKGROUND: Adjuvant regional anesthesia is often selected for patients or procedures with high risk of pulmonary complications after general anesthesia. The benefit of adjuvant regional anesthesia to reduce postoperative pulmonary complications remains uncertain. In a prospective observational multicenter study, patients scheduled for non-cardiothoracic surgery with at least one postoperative pulmonary complication surprisingly received adjuvant regional anesthesia more frequently than those with no complications. We hypothesized that, after adjusting for surgical and patient complexity variables, the incidence of postoperative pulmonary complications would not be associated with adjuvant regional anesthesia. METHODS: We performed a secondary analysis of a prospective observational multicenter study including 1202 American Society of Anesthesiologists physical status 3 patients undergoing non-cardiothoracic surgery. Patients were classified as receiving either adjuvant regional anesthesia or general anesthesia alone. Predefined pulmonary complications within the first seven postoperative days were prospectively identified. Groups were compared using bivariable and multivariable hierarchical logistic regression analyses for the outcome of at least one postoperative pulmonary complication. RESULTS: Adjuvant regional anesthesia was performed in 266 (22.1%) patients and not performed in 936 (77.9%). The incidence of postoperative pulmonary complications was greater in patients receiving adjuvant regional anesthesia (42.1%) than in patients without it (30.9%) (site adjusted p = 0.007), but this association was not confirmed after adjusting for covariates (adjusted OR 1.37; 95% CI, 0.83-2.25; p = 0.165). CONCLUSION: After adjusting for surgical and patient complexity, adjuvant regional anesthesia versus general anesthesia alone was not associated with a greater incidence of postoperative pulmonary complications in this multicenter cohort of non-cardiothoracic surgery patients.

A Pilot Study of Patient-Reported Outcome Measures Across a Broad Sample of Surgical Patients.

BACKGROUND: Patient-reported outcomes (PROs) have the potential to aid in surgical decision-making, predict surgical outcomes, assess recovery, and evaluate long-term success. We performed a pilot study testing the ability to use PROs in a broad surgical population in preparation for wide spread use. MATERIAL AND METHODS: Surgical patients completed five Patient-Reported Outcome Measurement Information System (PROMIS) measures during their preoperative encounter in the preanesthesia clinic and again postoperatively via emailed link. Preoperative to postoperative changes in PROMIS scores, factors related to completion of postoperative measures, intercorrelations between PROMIS measures, and numbers of patients with normal function, and mild, moderate, and severe deficits in PROMIS scores were analyzed. RESULTS: A total of 393 patients undergoing surgery in 8 specialties completed preoperative PROMIS measures; 239 (60.8%) completed them postoperatively. Physical function (P < 0.0001), pain (P < 0.0001), and cognitive function (P = 0.03) PROMIS scores significantly worsened after surgery but not mental PROMIS scores (P = 0.48). Hispanic and sicker patients had lower completion rates of postoperative measures. Intercorrelations were very high (>0.80) among the physical function and self-efficacy for activities of daily living PROMIS measures. Physical function and pain PROMIS measures had the largest number of patients in the "severe" range after surgery. CONCLUSIONS: Patients across a broad surgical population completed PROMIS measures successfully, both preoperatively and postoperatively, although the postoperative completion rate was lower than other studies reported in the literature. PROMIS scores were reflective of the effects of surgery. Some of the PROMIS measures were highly correlated suggesting that some measures could be eliminated or replaced with measures assessing other important effects of surgery. Consideration could be made to alert health care providers about patients having PROs in the "severe" range for potential intervention.

Preventative Root-Collar Excavation Reduces Peach Tree Mortality Caused By Armillaria Root Rot On Replant Sites.

In the southeastern United States, Armillaria root rot (ARR) is caused by Desarmillaria tabescens and has become the primary cause of premature mortality in peach orchards. Most rootstocks used in commercial orchards are susceptible and management options are limited. A postinfection practice known as root-collar excavation (RCE), which involves permanent removal of the soil from the base of the trunk, has been shown to improve yields and prolong the productive life of symptomatic trees. However, symptomatic trees already have an advanced infection at the base of the trunk. This study evaluated the efficacy of preventative RCE on the progression of tree mortality in two orchards that were planted in infested replant sites. To provide convincing data for growers, the study was carried out in a commercial orchard and an experimental orchard for 8 years. Furthermore, representative enterprise budgets and net present value (NPV) analysis were utilized to compare the profitability of the two approaches. Trees were planted shallow on berms (45 by 90 cm) to facilitate RCE with hoes and AirSpade 2 years later. Tree mortality in the RCE treatment of the experimental orchard was first observed in year 6 and increased 8% on average per year thereafter. In contrast, tree mortality in the "Grower Standard" treatment was first observed in year 4 and increased 12.7% on average per year thereafter. At the commercial orchard, tree mortality in the RCE treatment was first observed in year 7 and increased 1.9% on average thereafter, while tree mortality in the Grower Standard treatment first appeared in year 5 and increased 4.3% on average thereafter. The delayed onset of ARR-associated tree mortality and the lower annual tree mortality rate in the RCE treatment led to higher NPVs in both locations. There were no negative effects on yield or fruit quality. However, the new planting system can create horticultural challenges, including the formation of a proper berm, uneven ground around the tree interfering with tree care and harvest, increased erosion due to channeling of rainwater, and increased rootstock suckering. The RCE is a valid option for southeastern growers needing to manage high ARR disease pressure on replant sites or on sites only recently cleared from ARR-infected forest land.

First Report of the Root-Knot Nematode, Meloidogyne floridensis Infecting Guardian® Peach Rootstock in South Carolina, USA.

In 2018 to 2019, soil and root samples from some declining peach orchards were collected in Edgefield County, South Carolina, USA. Excavated roots of Guardian® peach (Prunus persica) rootstock showed strong gall symptoms. Extracted root-knot nematodes (RKN) were identified by both morphological and molecular methods as M. floridensis. This is the first detection of the peach RKN in South Carolina and the third state in the USA after Florida and California.

Temporal Trends, Determinants, and Outcomes of Inpatient versus Outpatient Arteriovenous Fistula Operations.

BACKGROUND: As high healthcare costs are increasing scrutinized, a movement toward reducing patient hospital admissions and lengths of stay has emerged, particularly for operations that may be performed safely in the outpatient setting. Our aim is to describe recent temporal trends in the proportion of dialysis access procedures performed on an inpatient versus outpatient basis and to determine the effects of these changes on perioperative morbidity and mortality. METHODS: The 2005-2008 American College of Surgeons National Surgical Quality Improvement Program database was queried for all primary arteriovenous fistula (AVF) procedures using current procedural terminology codes. Changes in the proportions of inpatient versus outpatient operations performed by year, as well as the associated 30-day postoperative morbidity and mortality, were analyzed using univariable statistics and multivariable logistic regression. RESULTS: Two thousand nine hundred fifty AVF procedures were performed over the study period. Overall, 71.7% (n = 2,114) were performed on an outpatient basis. Inpatient procedures were associated with higher 30-day morbidity (10.5% vs. 4.5%) and mortality (2.8% vs. 0.7%) than outpatient procedures (both, P < 0.001). There was a significant increase in the proportion of procedures performed on an outpatient basis over time (2005: 56% vs. 2008: 75%; P < 0.001). There were no changes in postoperative morbidity or mortality for inpatient or outpatient AVF over time (P ≥ 0.36). Independent determinants of having an inpatient procedure included younger age (OR 0.99), increasing ASA class (ASA IV OR 1.56), congestive heart failure (OR 3.32), recent ascites (OR 3.25), poor functional status (OR 3.22), the presence of an open wound (OR 1.91), and recent sepsis (OR 6.06) (all, P < 0.01). Acute renal failure (OR 2.60) and current dialysis (OR 1.44) were also predictive (P < 0.001). After correcting for baseline differences between groups, the adjusted OR for both morbidity (aOR 1.93, 95% CI 1.38-2.69) and mortality (aOR 2.85, 95% CI 1.36-5.95) remained significantly higher for inpatient versus outpatient AVF. CONCLUSIONS: Dialysis access operations are increasingly being performed on an outpatient basis, with stable perioperative outcomes. Inpatient procedures are associated with worse outcomes, likely because they are reserved for patients with acute illnesses, serious comorbidities, and poor functional status. Overall, for appropriately selected patients, the movement toward performing more elective dialysis access operations on an outpatient basis is associated with acceptable outcomes.

Operative variables are better predictors of postdischarge infections and unplanned readmissions in vascular surgery patients than patient characteristics.

OBJECTIVE: Although postoperative readmissions are frequent in vascular surgery patients, the reasons for these readmissions are not well characterized, and effective approaches to their reduction are unknown. Our aim was to analyze the reasons for vascular surgery readmissions and to report potential areas for focused efforts aimed at readmission reduction. METHODS: The 2012 to 2013 American College of Surgeons National Quality Improvement Program (ACS NSQIP) data set was queried for vascular surgery patients. Multivariable models were developed to analyze risk factors for postdischarge infections, the major drivers of unplanned 30-day readmissions. RESULTS: We identified 86,403 vascular surgery patients for analysis. Thirty-day readmission occurred in 8827 (10%), of which 8054 (91%) were unplanned. Of the unplanned readmissions, 61% (n = 4951) were related to the index vascular surgery procedure. Infectious complications were the most common reason for a surgery-related readmission (1940 [39%]), with surgical site infection being the most common type of infection related to unplanned readmission. Multivariable analysis showed the top five preoperative risk factors for postdischarge infections were the presence of a preoperative open wound, inpatient operation, obesity, work relative value unit, and insulin-dependent diabetes (but not diabetes managed with oral medications). Cigarette smoking was a weak predictor and came in tenth in the mode (overall C index, 0.657). When operative and postoperative factors were included in the model, total operative time was the strongest predictor of postdischarge infectious complications (odds ratio [OR] 1.2 for each 1-hour increase in operative time), followed by presence of a preoperative open wound (OR, 1.5), inpatient operation (OR, 2), obesity (OR, 1.8), and discharge to rehabilitation facility (OR, 1.7; P < .001 for all). Insulin-dependent diabetes, cigarette smoking, dialysis dependence, and female gender were also predictive, albeit with smaller effects (OR, 1.1-1.3 for all; P < .001). The overall fit of the multivariable model was fair (C statistic, 0.686). CONCLUSIONS: Infectious complications dominate the reasons for unplanned 30-day readmissions in vascular surgery patients. We have identified preoperative, operative, and postoperative risk factors for these infections with the goal of reducing these complications and thus readmissions. Expected patient risk factors, such as diabetes, obesity, renal insufficiency, and cigarette smoking, were less important in predicting infectious complications compared with operative time, presence of a preoperative open wound, and inpatient operation. Our findings suggest that careful operative planning and expeditious operations may be the most effective approaches to reducing infections and thus readmissions in vascular surgery patients.

Intermediate-Acting Nondepolarizing Neuromuscular Blocking Agents and Risk of Postoperative 30-Day Morbidity and Mortality, and Long-term Survival.

BACKGROUND: Nondepolarizing neuromuscular blocking drugs (NNMBDs) are commonly used as an adjunct to general anesthesia. Residual blockade is common, but its potential adverse effects are incompletely known. This study was designed to assess the association between NNMBD use with or without neostigmine reversal and postoperative morbidity and mortality. METHODS: This is a retrospective observational study of 11,355 adult patients undergoing general anesthesia for noncardiac surgery at 5 Veterans Health Administration (VA) hospitals. Of those, 8984 received NNMBDs, and 7047 received reversal with neostigmine. The primary outcome was a composite of respiratory complications (failure to wean from the ventilator, reintubation, or pneumonia), which was "yes" if a patient had any of the 3 component events and "no" if they had none. Secondary outcomes were nonrespiratory complications, 30-day and long-term all-cause mortality. We adjusted for differences in patient risk using propensity matched (PM) followed by assessment of the association of interest by logistic regression between the matched pairs as our primary analysis and multivariable logistic regression (MLR) as a sensitivity analysis. RESULTS: Our primary aim was to assess the adverse outcomes in the patients who had received NNMBDs with and without neostigmine. Administration of an NNMBD without neostigmine reversal compared with NNMBD with neostigmine reversal was associated with increased odds of respiratory complications (PM odds ratio [OR], 1.75 [95% confidence interval [CI], 1.23-2.50]; MLR OR, 1.71 [CI, 1.24-2.37]) and a marginal increase in 30-day mortality (PM OR, 1.83 [CI, 0.99-3.37]; MLR OR, 1.78 [CI, 1.02-3.13]). However, there were no statistically significant associations with nonrespiratory complications or long-term mortality. Patients who were administered an NNMBD followed by neostigmine had no differences in outcomes compared with patients who had general anesthesia without an NNMBD. CONCLUSIONS: The use of NNMBDs without neostigmine reversal was associated with increased odds of our composite respiratory outcome compared with patients reversed with neostigmine. Based on these data, we conclude that reversal of NNMBDs should become a standard practice if extubation is planned.

Surgical Risk Preoperative Assessment System (SURPAS): I. Parsimonious, Clinically Meaningful Groups of Postoperative Complications by Factor Analysis.

OBJECTIVE: To use factor analysis to cluster the 18 American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) perioperative complications into a reproducible, smaller number of clinically meaningful groups of postoperative complications, facilitating and streamlining future study and application in live clinical settings. BACKGROUND: The ACS NSQIP collects and reports on eighteen 30-day postoperative complications (excluding mortality), which are variably grouped in published analyses using ACS NSQIP data. This hinders comparison between studies of this widely used quality improvement dataset. METHODS: Factor analysis was used to develop a series of complication clusters, which were then analyzed to identify a parsimonious, clinically meaningful grouping, using 2,275,240 surgical cases in the ACS NSQIP Participant Use File (PUF), 2005 to 2012. The main outcome measures are reproducible, data-driven, clinically meaningful clusters of complications derived from factor solutions. RESULTS: Factor analysis solutions for 5 to 9 latent factors were examined for their percent of total variance, parsimony, and clinical interpretability. Applying the first 2 of these criteria, we identified the 7-factor solution, which included clusters of pulmonary, infectious, wound disruption, cardiac/transfusion, venous thromboembolic, renal, and neurological complications, as the best solution for parsimony and clinical meaningfulness. Applying the last (clinical interpretability), we combined the wound disruption with the infectious clusters resulting in 6 clusters for future clinical applications. CONCLUSIONS: Factor analysis of ACS NSQIP postoperative complication data provides 6 clinically meaningful complication clusters in lieu of 18 postoperative morbidities, which will facilitate comparisons and clinical implementation of studies of postoperative morbidities.

Surgical Risk Preoperative Assessment System (SURPAS): III. Accurate Preoperative Prediction of 8 Adverse Outcomes Using 8 Predictor Variables.

OBJECTIVE: To develop accurate preoperative risk prediction models for multiple adverse postoperative outcomes applicable to a broad surgical population using a parsimonious common set of risk variables and outcomes. SUMMARY BACKGROUND DATA: Currently, preoperative assessment of surgical risk is largely based on subjective clinician experience. We propose a paradigm shift from the current postoperative risk adjustment for cross-hospital comparison to patient-centered quantitative risk assessment during the preoperative evaluation. METHODS: We identify the most common and important predictor variables of postoperative mortality, overall morbidity, and 6 complication clusters from previously published prediction analyses that used forward selection stepwise logistic regression. We then refit the prediction models using only the 8 most common and important predictor variables, and compare the discrimination and calibration of these models to the original full-variable models using the c-index, Hosmer-Lemeshow analysis, and Brier scores. RESULTS: Accurate risk models for 30-day outcomes of mortality, overall morbidity, and 6 clusters of complications were developed using a set of 8 preoperative risk variables. C-indexes of the 8 variable models are between 97.9% and 99.2% of those of the full models containing up to 28 variables, indicating excellent discrimination using fewer predictor variables. Hosmer-Lemeshow analyses showed observed to expected event rates to be nearly identical between parsimonious models and full models, both showing good calibration. CONCLUSIONS: Accurate preoperative risk assessment of postoperative mortality, overall morbidity, and 6 complication clusters in a broad surgical population can be achieved with as few as 8 preoperative predictor variables, improving feasibility of routine preoperative risk assessment for surgical patients.

Surgical Risk Preoperative Assessment System (SURPAS): II. Parsimonious Risk Models for Postoperative Adverse Outcomes Addressing Need for Laboratory Variables and Surgeon Specialty-specific Models.

OBJECTIVE: To develop parsimonious prediction models for postoperative mortality, overall morbidity, and 6 complication clusters applicable to a broad range of surgical operations in adult patients. SUMMARY BACKGROUND DATA: Quantitative risk assessment tools are not routinely used for preoperative patient assessment, shared decision making, informed consent, and preoperative patient optimization, likely due in part to the burden of data collection and the complexity of incorporation into routine surgical practice. METHODS: Multivariable forward selection stepwise logistic regression analyses were used to develop predictive models for 30-day mortality, overall morbidity, and 6 postoperative complication clusters, using 40 preoperative variables from 2,275,240 surgical cases in the American College of Surgeons National Surgical Quality Improvement Program data set, 2005 to 2012. For the mortality and overall morbidity outcomes, prediction models were compared with and without preoperative laboratory variables, and generic models (based on all of the data from 9 surgical specialties) were compared with specialty-specific models. In each model, the cumulative c-index was used to examine the contribution of each added predictor variable. C-indexes, Hosmer-Lemeshow analyses, and Brier scores were used to compare discrimination and calibration between models. RESULTS: For the mortality and overall morbidity outcomes, the prediction models without the preoperative laboratory variables performed as well as the models with the laboratory variables, and the generic models performed as well as the specialty-specific models. The c-indexes were 0.938 for mortality, 0.810 for overall morbidity, and for the 6 complication clusters ranged from 0.757 for infectious to 0.897 for pulmonary complications. Across the 8 prediction models, the first 7 to 11 variables entered accounted for at least 99% of the c-index of the full model (using up to 28 nonlaboratory predictor variables). CONCLUSIONS: Our results suggest that it will be possible to develop parsimonious models to predict 8 important postoperative outcomes for a broad surgical population, without the need for surgeon specialty-specific models or inclusion of laboratory variables.

Drivers of readmissions in vascular surgery patients.

OBJECTIVE: Postoperative readmissions are frequent in vascular surgery patients, but it is not clear which factors are the main drivers of readmissions. Specifically, the relative contributions of patient comorbidities vs those of operative factors and postoperative complications are unknown. We sought to study the multiple potential drivers of readmission and to create a model for predicting the risk of readmission in vascular patients. METHODS: The 2012-2013 American College of Surgeons National Surgical Quality Improvement Program data set was queried for unplanned readmissions in 86,238 vascular patients. Multivariable forward selection logistic regression analysis was used to model the relative contributions of patient comorbidities, operative factors, and postoperative complications for readmission. RESULTS: The unplanned readmission rate was 9.3%. The preoperative model based on patient demographics and comorbidities predicted readmission risk with a low C index of .67; the top five predictors of readmission were American Society of Anesthesiologists class, preoperative open wound, inpatient operation, dialysis dependence, and diabetes mellitus. The postoperative model using operative factors and postoperative complications predicted readmission risk better (C index, .78); postoperative complications were the most significant predictor of readmission, overpowering patient comorbidities. Importantly, postoperative complications identified before discharge from the hospital were not a strong predictor of readmission as the model using predischarge postoperative complications had a similar C index to our preoperative model (.68). However, the inclusion of complications identified after discharge from the hospital appreciably improved the predictive power of the model (C index, .78). The top five predictors of readmission in the final model based on patient comorbidities and postoperative complications were postdischarge deep space infection, superficial surgical site infection, pneumonia, myocardial infection, and sepsis. CONCLUSIONS: Readmissions in vascular surgery patients are mainly driven by postoperative complications identified after discharge. Thus, efforts to reduce vascular readmissions focusing on inpatient hospital data may prove ineffective. Our study suggests that interventions to reduce vascular readmissions should focus on prompt identification of modifiable postdischarge complications.

Contributions of planned readmissions and patient comorbidities to high readmission rates in vascular surgery patients.

OBJECTIVE: Administrative data show that among surgical patients, readmission rates are highest in vascular surgery. Herein we analyze the contribution of planned readmissions and patient comorbidities to high readmission rates in vascular surgery. METHODS: The 2012 to 2014 American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) data set was analyzed for overall and unplanned readmissions. Bivariable and multivariable risk adjustment analyses were performed using patient comorbidities to compare risks of overall and unplanned readmissions in vascular surgery compared with other specialties. RESULTS: Among 1,164,421 surgical patients, 86,403 underwent a vascular operation (other specialties included general surgery, 587,829 [51%]; orthopedic surgery, 211,507 [18%]; gynecology, 82,771 [7%]; urology, 62,153 [5%]; neurosurgery, 55,030 [4.7%]; plastic surgery, 32,318 [3%]; otolaryngology, 31,070 [2.6%]; and thoracic surgery, 15,340 [1%]). Incidence of 30-day readmission was 10.2% for vascular and 5.5% for other specialties (P < .0001). Planned readmissions were more frequent for vascular than for other specialties (8.8% vs 5.4%; P < .0001). In unadjusted analysis, vascular patients had significantly greater risk for overall readmission (odds ratio [OR], 1.97; 95% confidence interval [CI], 1.93-2.02; P < .0001) and unplanned readmission (OR, 1.89; 95% CI, 1.84-1.93; P < .0001) compared with other specialties. In bivariable analysis, vascular patients were older (67 ± 13 vs 56 ± 17 years) and had more comorbidities such as diabetes (31% vs 14%), dialysis dependence (6.3% vs 0.9%), American Society of Anesthesiology class III/IV status (84% vs 41%), and many others (all P < .0001). After risk adjustment for baseline differences between groups, vascular patients had a marginally greater overall risk of readmission compared with other specialties (OR, 1.04; 95% CI, 1.01-1.07; P < .0001), but the risk of unplanned readmission was not significantly different (OR, 0.98; 95% CI, 0.95-1.01; P = .13). CONCLUSIONS: Incidence of 30-day readmission after vascular surgery appears high, but after account for planned readmissions and risk adjustment, the risk of unplanned readmission is similar to that in other surgical patients. Thus, the use of readmission rate as a quality measure must account for more frequent planned vascular readmissions and patient-specific differences between vascular surgery and other specialties.

A Community-Based Oral Health Intervention in Navajo Nation Head Start: Participation Factors and Contextual Challenges.

Successful interventions require consistent participation by intended recipients. We utilized mixed methods to describe participation of 518 parent-child dyads enrolled in a randomized cluster trial of a 2-year oral health intervention for Head Start (HS) families across Navajo Nation delivered by native Community Oral Health Specialists (COHS). We quantitatively assessed factors that contributed to participation and qualitatively examined barriers and strategies. The intervention offered fluoride varnish (FV) and oral health promotion (OHP) activities for two cohorts (enrolled in 2011, N = 286, or 2012, N = 232) of children in the HS classrooms and OHP for parents outside the classroom. Child participation was good: FV: 79.7 (Cohort 1) and 85.3 % (Cohort 2) received at least 3 of 4 applications; OHP: 74.5 (Cohort 1) and 78.4 % (Cohort 2) attended at least 3 of 5 events. Parent participation was low: 10.5 (Cohort 1) and 29.8 % (Cohort 2) attended at least three of four events. Analysis of survey data found significant effects on parent participation from fewer people in the household, Cohort 2 membership, greater external-locus of control, and a greater perception that barriers existed to following recommended oral health behaviors. Qualitative analysis of reports from native field staff, COHS, community members, and the research team identified barriers (e.g., geographic expanse, constraints of a research trial) and suggested strategies to improve parent participation (e.g., improve communication between COHS and parents/community). Many challenges to participation exist when conducting interventions in rural areas with underserved populations. Working with community partners to inform the development and delivery of interventions is critical.