Joseph E. Cavanaugh

Unifying the derivations for the Akaike and corrected Akaike information criteria

The Akaike (1973, 1974) information criterion, AIC, and the corrected Akaike information criterion (Hurvich and Tsai, 1989), AICc, were both designed as estimators of the expected Kullback-Leibler discrepancy between the model generating the data and a fitted candidate model. AIC is justified in a very general framework, and as a result, offers a crude estimator of the expected discrepancy: one which exhibits a potentially high degree of negative bias in small-sample applications (Hurvich and Tsai, 1989). AICc corrects for this bias, but is less broadly applicable than AIC since its justification depends upon the form of the candidate model (Hurvich and Tsai, 1989, 1993; Hurvich et al., 1990; Bedrick and Tsai, 1994). Although AIC and AICc share the same objective, the derivations of the criteria proceed along very different lines, making it difficult to reconcile how AICc improves upon the approximations leading to AIC. To address this issue, we present a derivation which unifies the justifications of AIC and AICc in the linear regression framework.

A large-sample model selection criterion based on Kullback's symmetric divergence

The Akaike information criterion, AIC, is a widely known and extensively used tool for statistical model selection. AIC serves as an asymptotically unbiased estimator of a variant of Kullbacks directed divergence between the true model and a fitted approximating model. The directed divergence is an asymmetric measure of separation between two statistical models, meaning that an alternate directed divergence may be obtained by reversing the roles of the two models in the definition of the measure. The sum of the two directed divergences is Kullbacks symmetric divergence. Since the symmetric divergence combines the information in two related though distinct measures, it functions as a gauge of model disparity which is arguably more sensitive than either of its individual components. With this motivation, we propose a model selection criterion which serves as an asymptotically unbiased estimator of a variant of the symmetric divergence between the true model and a fitted approximating model. We examine the performance of the criterion relative to other well-known criteria in a simulation study.

Association of a Bundled Intervention With Surgical Site Infections Among Patients Undergoing Cardiac, Hip, or Knee Surgery

IMPORTANCE Previous studies suggested that a bundled intervention was associated with lower rates of Staphylococcus aureus surgical site infections (SSIs) among patients having cardiac or orthopedic operations. OBJECTIVE To evaluate whether the implementation of an evidence-based bundle is associated with a lower risk of S. aureus SSIs in patients undergoing cardiac operations or hip or knee arthroplasties. DESIGN, SETTING, AND PARTICIPANTS Twenty hospitals in 9 US states participated in this pragmatic study; rates of SSIs were collected for a median of 39 months (range, 39-43) during the preintervention period (March 1, 2009, to intervention) and a median of 21 months (range, 14-22) during the intervention period (from intervention start through March 31, 2014). INTERVENTIONS Patients whose preoperative nares screens were positive for methicillin-resistant S. aureus (MRSA) or methicillin-susceptible S. aureus (MSSA) were asked to apply mupirocin intranasally twice daily for up to 5 days and to bathe daily with chlorhexidine-gluconate (CHG) for up to 5 days before their operations. MRSA carriers received vancomycin and cefazolin or cefuroxime for perioperative prophylaxis; all others received cefazolin or cefuroxime. Patients who were MRSA-negative and MSSA-negative bathed with CHG the night before and morning of their operations. Patients were treated as MRSA-positive if screening results were unknown. MAIN OUTCOMES AND MEASURES The primary outcome was complex (deep incisional or organ space) S. aureus SSIs. Monthly SSI counts were analyzed using Poisson regression analysis. RESULTS After a 3-month phase-in period, bundle adherence was 83% (39% full adherence; 44% partial adherence). Overall, 101 complex S. aureus SSIs occurred after 28,218 operations during the preintervention period and 29 occurred after 14,316 operations during the intervention period (mean rate per 10,000 operations, 36 for preintervention period vs 21 for intervention period, difference, -15 [95% CI, -35 to -2]; rate ratio [RR], 0.58 [95% CI, 0.37 to 0.92]). The rates of complex S. aureus SSIs decreased for hip or knee arthroplasties (difference per 10,000 operations, -17 [95% CI, -39 to 0]; RR, 0.48 [95% CI, 0.29 to 0.80]) and for cardiac operations (difference per 10,000 operations, -6 [95% CI, -48 to 8]; RR, 0.86 [95% CI, 0.47 to 1.57]). CONCLUSIONS AND RELEVANCE In this multicenter study, a bundle comprising S. aureus screening, decolonization, and targeted prophylaxis was associated with a modest, statistically significant decrease in complex S. aureus SSIs.

Influence of smoking on gingival crevicular fluid cytokines in severe chronic periodontitis.

AIM The aim of this study was to compare the expression of 22 chemokines and cytokines in gingival crevicular fluid (GCF) from smokers and non-smokers with periodontitis and periodontally healthy control subjects. MATERIALS AND METHODS Forty subjects with generalized severe chronic periodontitis (20 smokers and 20 non-smokers) and 12 periodontally healthy control subjects participated in this study. Four diseased and two healthy sites were selected from each of the periodontitis subjects. GCF samples were collected and cytokines analysed utilizing a multiplexed immunoassay (Luminex(®) ). Statistical analyses employed non-parametric tests including the Mann-Whitney and Wilcoxon matched-pairs signed-rank tests. RESULTS Compared with healthy control subjects, GCF in subjects with chronic periodontitis contained significantly higher amounts of interleukin (IL)-1α, IL-1β, IL-6, IL-12(p40) (pro-inflammatory cytokines); IL-8, macrophage chemotactic protein (MCP)-1, macrophage inflammatory protein (MIP)-1α, regulated on activation normal T-cell expressed and secreted (RANTES) (chemokines); IL-2, IFN-γ, IL-3, IL-4 (Th1/Th2 cytokines); IL-15 [regulator of T-cells and natural killer (NK) cells]. Smokers displayed decreased amounts of pro-inflammatory cytokines [IL-1α, IL-6, IL-12(p40)], chemokines (IL-8, MCP-1, MIP-1, RANTES), and regulators of T-cells and NK cells (IL-7, IL-15). CONCLUSIONS Periodontitis subjects had significantly elevated cytokine and chemokine profiles. Smokers exhibited a decrease in several pro-inflammatory cytokines and chemokines and certain regulators of T-cells and NK-cells. This reflects the immunosuppressant effects of smoking which may contribute to an enhanced susceptibility to periodontitis.

A multiplex immunoassay demonstrates reductions in gingival crevicular fluid cytokines following initial periodontal therapy

BACKGROUND AND OBJECTIVE Cytokines and chemokines play an important role in the pathogenesis of periodontal diseases. The objective of this study was to quantitatively assess the effect of initial periodontal therapy on gingival crevicular fluid levels of a comprehensive panel of cytokines and chemokines, including several less extensively studied mediators. MATERIAL AND METHODS Clinical examinations were performed and gingival crevicular fluid samples obtained from six subjects with generalized severe chronic periodontitis prior to initial periodontal therapy and at re-evaluation (6-8 weeks). Four diseased and two healthy sites were sampled in each subject. Twenty-two gingival crevicular fluid mediators were examined using a multiplex antibody capture and detection platform. Statistical analyses were performed by fitting mixed effects linear models to log-transformed gingival crevicular fluid values. RESULTS Gingival crevicular fluid interleukin (IL)-1alpha and IL-1beta were the only cytokines to differ in initially diseased vs. initially healthy sites. Following initial therapy, 13 of the 16 detectable cytokines and chemokines decreased significantly in diseased sites, including IL-1alpha, IL-1beta, IL-2, IL-3, IL-6, IL-7, IL-8, IL-12 (p40), CCL5/regulated on activation, normally T cell expressed and secreted (RANTES), eotaxin, macrophage chemotactic protein-1, macrophage inflammatory protein-1alpha and interferon-gamma. At healthy sites, only three of the 16 mediators were significantly altered following therapy. CONCLUSION This is the first study, to our knowledge, to evaluate such an extensive panel of gingival crevicular fluid mediators within the same sample prior to and following initial therapy. The results confirm that periodontal therapy effectively reduces pro-inflammatory cytokines and chemokines, including less well-described mediators that may be important in initiation and progression of periodontitis. The multiplex assay will prove useful for future gingival crevicular fluid studies.

Relationship of Influenza Vaccination Declination Statements and Influenza Vaccination Rates for Healthcare Workers in 22 US Hospitals

The use of declination statements was associated with a mean increase of 11.6% in influenza vaccination rates among healthcare workers at 22 hospitals. In most hospitals, there were no negative consequences for healthcare workers who refused to sign the forms, and most policies were implemented along with other interventions designed to increase vaccination rates.

Regression and time series model selection using variants of the schwarz information criterion

The Schwarz (1978) information criterion, SIC, is a widely-used tool in model selection, largely due to its computational simplicity and effective performance in many modeling frameworks. The derivation of SIC (Schwarz, 1978) establishes the criterion as an asymptotic approximation to a transformation of the Bayesian posterior probability of a candidate model. In this paper, we investigate the derivation for the identification of terms which are discarded as being asymptotically negligible, but which may be significant in small to moderate sample-size applications. We suggest several SIC variants based on the inclusion of these terms. The results of a simulation study show that the variants improve upon the performance of SIC in two important areas of application:multiple linear regression and time series analysis.

Human β-defensin 3 binds to hemagglutinin B (rHagB), a non-fimbrial adhesin from Porphyromonas gingivalis, and attenuates a pro-inflammatory cytokine response

Regulatory mechanisms in mucosal secretions and tissues recognize antigens and attenuate pro‐inflammatory cytokine responses. Here, we asked whether human β‐defensin 3 (HBD3) serves as an upstream suppressor of cytokine signaling that binds and attenuates pro‐inflammatory cytokine responses to recombinant hemagglutinin B (rHagB), a non‐fimbrial adhesin from Porphyromonas gingivalis strain 381. We found that HBD3 binds to immobilized rHagB and produces a significantly higher resonance unit signal in surface plasmon resonance spectroscopic analysis, than HBD2 and HBD1 that are used as control defensins. Furthermore, we found that HBD3 significantly attenuates (P<0.05) the interleukin (IL)‐6, IL‐10, granulocyte macrophage colony stimulating factor (GM‐CSF) and tumor‐necrosis factor‐α (TNF‐α) responses induced by rHagB in human myeloid dendritic cell culture supernatants and the extracellular signal‐regulated kinases (ERK 1/2) response in human myeloid dendritic cell lysates. Thus, HBD3 binds rHagB and this interaction may be an important initial step to attenuate a pro‐inflammatory cytokine response and an ERK 1/2 response.

Generalizing the derivation of the schwarz information criterion

The Schwarz information criterion (SIC, BTC, SBC) is one of the most widely known and used tools in statistical model selection. The criterion was derived by Schwarz (1978) to serve as an asymptotic approximation to a transformation of the Bayesian posterior probability of a candidate model. Although the original derivation assumes that the observed data is independent, identically distributed, and arising from a probability distribution in the regular exponential family, SIC has traditionally been used in a much larger scope of model selection problems. To better justify the widespread applicability of SIC, we derive the criterion in a very general framework: one which does not assume any specific form for the likelihood function, but only requires that it satisfies certain non-restrictive regularity conditions.

An outbreak of severe Clostridium difficile-associated disease possibly related to inappropriate antimicrobial therapy for community-acquired pneumonia

We report a severe outbreak of Clostridium difficile infection. According to a chart review, half of the patients who received treatment for bacterial pneumonia before they developed C. difficile infection may not have had pneumonia. Excessive use of the hospitals new pneumonia care plan during the influenza season may have contributed to the intensity of this outbreak.