Joseph R. Ruiz

Ongoing provision of individual clinician performance data improves practice behavior

BACKGROUND: Clinical practice guidelines summarize evidence from science and attempt to translate those findings into clinical practice. Pervasive and consistent adoption of these guidelines into daily provider practice has proven slow. METHODS: Using postoperative nausea and vomiting (PONV) prophylaxis guideline compliance as our metric, we compared the effects of continuing medical education (CME) alone (I), CME with a single snapshot of provider compliance (II), and ongoing reporting of provider compliance data without further CME (III). We retrospectively analyzed guideline compliance of 23,279 anesthetics at the University of Texas M.D. Anderson Cancer Center. Compliance was defined as a patient with 1 risk factor for PONV receiving at least 1 antiemetic, 2 risk factors receiving at least 2 antiemetics, and 3 risk factors receiving at least 3 antiemetics. Drugs of the same class were counted as single antiemetic administration. Propofol-based anesthetic techniques were counted as receiving 1 antiemetic. Patients with 0 risk factors for PONV were not included. We estimated the compliance rates for each of the 4 time periods of the study adjusting for multiple observations on the same clinician. Individual performance feedback was given once at 6 months after intervention I coincident with a refresher presentation on PONV (start of intervention II) and on an ongoing quarterly basis during intervention III. RESULTS: Compliance rates were not significantly influenced with CME (intervention I) compared with baseline behavior (54.5% vs 54.4%, P = 0.9140). Significant improvement occurred during the time period when CME was paired with performance data (intervention II) compared with intervention I (59.2% vs 54.4%, P = 0.0002). Further significant improvement occurred when data alone were presented (intervention III) compared with intervention II (65.1% vs 59.2%, P < 0.0001). For patients with 3 risk factors, we saw significant improvement in compliance rates during intervention III (P = 0.0002). In post hoc analysis of overtreatment, the percentage differences between the baseline and time period III decreased as the number of risk factors increased. CONCLUSIONS: We observed the greatest improvement in guideline compliance with ongoing personal performance feedback. Provider feedback can be an effective tool to modify clinical practice but can have unanticipated consequences.

The effect of an anatomically classified procedure on antiemetic administration in the postanesthesia care unit.

BACKGROUND: The effect of the type of surgical procedure on postoperative nausea and vomiting (PONV) rate has been debated in the literature. Our goal in this retrospective database study was to investigate the effect the type of surgical procedure (categorized and compared anatomically) has on antiemetic therapy within 2 h of admission to the postanesthesia care unit (PACU). METHODS: We retrospectively analyzed data for oncology surgeries (n = 18,109), from our automated anesthesia information system database. We classified the types of surgical procedures anatomically into seven categories, with the integumentary musculoskeletal and the superficial surgeries chosen as the referent group. Our analysis included nine other risk factors for each patient, such as gender, smoking status, history of PONV or motion sickness, duration of anesthesia, number of prophylactic antiemetics administered, intraoperative opioids, ketorolac, epidural use, and postoperative opioids. Multivariate logistic regression was used to assess the effect of the type of surgery on antiemetic administration within the first 2 h of PACU admission, while adjusting for the other risk factors. RESULTS: Compared with integumentary musculoskeletal and superficial surgeries, patients undergoing neurological (P < 0.0001), head or neck (P < 0.0001), and abdominal (P < 0.0001) surgeries were administered PACU antiemetic significantly more often, whereas patients undergoing thoracic surgeries were administered PACU antiemetic significantly less often (P = 0.02). Breast or axilla (P = 0.74) and endoscopic (P = 0.28) procedures did not differ from the referent category. Female, nonsmoker, history of PONV or motion sickness, anesthesia duration, and intraoperative and postoperative opioid administration were significantly associated with antiemetic administration during early PACU admission. CONCLUSIONS: Using our automated anesthesia information system database, we found that the type of surgery, when categorized anatomically, was associated with an increased frequency of early PACU antiemetic administration in our population.

Does deep sedation with propofol affect adenoma detection rates in average risk screening colonoscopy exams

AIM To determine the effect of sedation with propofol on adenoma detection rate (ADR) and cecal intubation rates (CIR) in average risk screening colonoscopies compared to moderate sedation. METHODS We conducted a retrospective chart review of 2604 first-time average risk screening colonoscopies performed at MD Anderson Cancer Center from 2010-2013. ADR and CIR were calculated in each sedation group. Multivariable regression analysis was performed to adjust for potential confounders of age and body mass index (BMI). RESULTS One-third of the exams were done with propofol (n = 874). Overall ADR in the propofol group was significantly higher than moderate sedation (46.3% vs 41.2%, P = 0.01). After adjustment for age and BMI differences, ADR was similar between the groups. CIR was 99% for all exams. The mean cecal insertion time was shorter among propofol patients (6.9 min vs 8.2 min; P < 0.0001). CONCLUSION Deep sedation with propofol for screening colonoscopy did not significantly improve ADR or CIR in our population of average risk patients. While propofol may allow for safer sedation in certain patients (e.g., with sleep apnea), the overall effect on colonoscopy quality metrics is not significant. Given its increased cost, propofol should be used judiciously and without the implicit expectation of a higher quality screening exam.

Abstract PR622: A Comparison of Mortality Predictors in Cancer Surgery Patients

Materials & Methods: We retrospectively apply four mortality predictors to 62,763 adult surgical cancer patients from the University of Texas MD Anderson Cancer Center during January 2007 March 2014. We use the first surgery for each patient that is over 60 minutes in duration and compare the following indexes: Charlson Comorbidity Index1 as implemented by Deyo et al.2; Dalton’s Risk Quantification Index (RQI)3; Sessler’s Risk Stratification Index (RSI)4 and the Surgical Apgar Score5.

Do Anesthesiologists Know What They Are Doing? Mining a Surgical Time-Series Database to Correlate Expert Assessment with Outcomes

Anesthesiologists are taught to carefully manage patient vital signs during surgery. Unfortunately, there is little empirical evidence that vital sign management, as currently practiced, is correlated with patient outcomes. We seek to validate or repudiate current clinical practice and determine whether or not clinician evaluation of surgical vital signs correlate with outcomes. Using a database of over 90,000 cases, we attempt to determine whether those cases that anesthesiologists would subjectively decide are “low quality” are more likely to result in negative outcomes. The problem reduces to one of multi-dimensional time-series classification. Our approach is to have a set of expert anesthesiologists independently label a small number of training cases, from which we build classifiers and label all 90,000 cases. We then use the labeling to search for correlation with outcomes and compare the prevalence of important 30-day outcomes between providers. To mimic the providers’ quality labels, we consider several standard classification methods, such as dynamic time warping in conjunction with a kNN classifier, as well as complexity invariant distance, and a regression based upon the feature extraction methods outlined by Mao et al. 2012 (using features such as time-series mean, standard deviation, skew, etc.). We also propose a new feature selection mechanism that learns a hidden Markov model to segment the time series; the fraction of time that each series spends in each state is used to label the series using a regression-based classifier. In the end, we obtain strong, empirical evidence that current best practice is correlated with reduced negative patient outcomes. We also learn that all of the experts were able to significantly separate cases by outcome, with higher prevalence of negative 30-day outcomes in the cases labeled as “low quality” for almost all of the outcomes investigated.

The association between opioid administration and response to therapy in patients with acute lymphoblastic leukemia

Acute lymphoblastic leukemia (ALL) is a malignant clonal proliferation of precursor lymphocytes. Symptoms at presentation may include musculoskeletal pain, which is widely treated with opioids. In addition to their analgesic properties, opioids have been shown to elicit a wide variety of biological effects that may lead to immunomodulation and promote either apoptosis or proliferation of certain cancer cell lines [1–4]. Inhibition of spontaneous and cytokine-enhanced natural-killer cell cytotoxicity by morphine is an example of an immunomodulatory effect that may inhibit the ability to eliminate cancer cells [5]. Clinical studies examining the association between opioid administration and cancer recurrence in patients with “solid” tumors have reported mixed results [6–9]. There are, however, no clinical studies examining the association between opioid administration and response to treatment or survival of patients with ALL. The objective of our study was to determine whether opioid administration during induction therapy of patients with ALL is associated with a poorer response to treatment and worse survival. This was a retrospective study of newly diagnosed patients with ALL undergoing treatment per frontline protocol regimens between 1 January 2006 and 31 January 2013. The study was approved by the Institutional Review Board of The University of Texas M. D. Anderson Cancer Center. Pre-treatment variables including: demographics, disease immunophenotype, cytogenetics, bone marrow blast % (BMA), white blood cell count (WBC) and lactate dehydrogenase levels (LDH) were recorded. The chemotherapy regimen, as well as post-treatment variables including: post-induction minimal residual disease (MRD yes/no), quantitative MRD (MRD %), overall response to treatment and vital status at the time of the study (i.e. alive versus deceased) were also recorded. Anesthesia and pharmacy databases were queried for opioids administered or prescribed to this group of patients up to 1 month after initial admission for induction chemotherapy of ALL. Patients with incomplete medical records and those who were primarily treated at outside hospitals were excluded due to lack of data. Partial responders to treatment (1.7% of the patient population) and patients who died before completion of their first chemotherapy regimen (2.7% of the patient population) were excluded from the study. Descriptive statistics were used to summarize demographics, response to treatment and survival of patients. Comparisons of patient characteristics were made between patients who received opioids (group O) and those who did not (group B). Kaplan–Meier survival estimates were provided for select patient characteristics. To visualize the survival experience by levels of the patient characteristics of interest, Kaplan–Meier plots were created, and Cox regression was used to assess the association between survival and patient characteristics of interest. After we conducted univariable analysis, a multivariable model was constructed using those variables found to be statistically significant or marginally significant (p  0.10) in both Table I and in the univariable analysis presented in Table II. The mean age of the patient population was 40.9 years (range 15–80 years), 55% were men, and the most common immunophenotype was “precursor B” (85.3%). Of the study population 14.3% had “precursor T” lymphoblastic leukemia/lymphoma, and one case had mixed precursor B- and T-ALL. A total of 142 patients (49%) received opioids. The mean age was lower for patients receiving opioids (37.9 vs. 44.0 years; p  0.004). With the exception of age, there were no statistically significant differences in demographics between the two groups. There were 77 patients with high-risk disease (Philadelphia positive ALL) in the study population. Forty high-risk patients (51.9%) received opioids, whereas 37 (48.1%) did not receive opioids. There were also no significant differences between the groups with regard to the pretreatment variables of initial bone marrow blast % (p  0.689), WBC count (p  0.490) and LDH level (p  0.270) or the posttreatment variables of MRD (p  0.850), MRD% (p  0.543), treatment response (p  0.419) and overall survival (p  0.478). For this assessment, the log-rank test was used to compare

Beat to Beat: A Measured Look at the History of Pulse Oximetry

It can be argued that pulse oximetry is the most important technological advancement ever made in monitoring the well-being and safety of patients undergoing anesthesia. Before its development, the physical appearance of the patient and blood gas analysis were the only methods of assessing hypoxemia in patients. The disadvantages of blood gas analysis are that it is not without pain, complications, and most importantly does not provide continuous, real-time data. Although it has become de rigueur to use pulse oximetry for every anesthetic, the road leading to pulse oximetry began long ago.