Comparing methods for estimation of heterogeneous treatment effects using observational data from health care databases.

There is growing interest in using routinely collected data from health care databases to study the safety and effectiveness of therapies in "real-world" conditions, as it can provide complementary evidence to that of randomized controlled trials. Causal inference from health care databases is challenging because the data are typically noisy, high dimensional, and most importantly, observational. It requires methods that can estimate heterogeneous treatment effects while controlling for confounding in high dimensions. Bayesian additive regression trees, causal forests, causal boosting, and causal multivariate adaptive regression splines are off-the-shelf methods that have shown good performance for estimation of heterogeneous treatment effects in observational studies of continuous outcomes. However, it is not clear how these methods would perform in health care database studies where outcomes are often binary and rare and data structures are complex. In this study, we evaluate these methods in simulation studies that recapitulate key characteristics of comparative effectiveness studies. We focus on the conditional average effect of a binary treatment on a binary outcome using the conditional risk difference as an estimand. To emulate health care database studies, we propose a simulation design where real covariate and treatment assignment data are used and only outcomes are simulated based on nonparametric models of the real outcomes. We apply this design to 4 published observational studies that used records from 2 major health care databases in the United States. Our results suggest that Bayesian additive regression trees and causal boosting consistently provide low bias in conditional risk difference estimates in the context of health care database studies.

Performance and usability testing of an automated tool for detection of peripheral artery disease using electronic health records.

Peripheral artery disease (PAD) is a common cardiovascular disorder that is frequently underdiagnosed, which can lead to poorer outcomes due to lower rates of medical optimization. We aimed to develop an automated tool to identify undiagnosed PAD and evaluate physician acceptance of a dashboard representation of risk assessment. Data were derived from electronic health records (EHR). We developed and compared traditional risk score models to novel machine learning models. For usability testing, primary and specialty care physicians were recruited and interviewed until thematic saturation. Data from 3168 patients with PAD and 16,863 controls were utilized. Results showed a deep learning model that utilized time engineered features outperformed random forest and traditional logistic regression models (average AUCs 0.96, 0.91 and 0.81, respectively), P < 0.0001. Of interviewed physicians, 75% were receptive to an EHR-based automated PAD model. Feedback emphasized workflow optimization, including integrating risk assessments directly into the EHR, using dashboard designs that minimize clicks, and providing risk assessments for clinically complex patients. In conclusion, we demonstrate that EHR-based machine learning models can accurately detect risk of PAD and that physicians are receptive to automated risk detection for PAD. Future research aims to prospectively validate model performance and impact on patient outcomes.

Big Data and Adverse Drug Reaction Detection.

Big Data holds the promise of fundamentally transforming the manner in which adverse drug reactions can be identified and evaluated. This commentary discusses new data sources that are envisioned to form a Big Data-enabled pharmacovigilance system and the role of these data in powering the future of adverse drug reactions detection.

HyBrow: a prototype system for computer-aided hypothesis evaluation.

MOTIVATION: Experimental design, hypothesis-testing and model-building in the current data-rich environment require the biologists' to collect, evaluate and integrate large amounts of information of many disparate kinds. Developing a unified framework for the representation and conceptual integration of biological data and processes is a major challenge in bioinformatics because of the variety of available data and the different levels of detail at which biological processes can be considered. RESULTS: We have developed the HyBrow (Hypothesis Browser) system as a prototype bioinformatics tool for designing hypotheses and evaluating them for consistency with existing knowledge. HyBrow consists of a modeling framework with the ability to accommodate diverse biological information sources, an event-based ontology for representing biological processes at different levels of detail, a database to query information in the ontology and programs to perform hypothesis design and evaluation. We demonstrate the HyBrow prototype using the galactose gene network in Saccharomyces cerevisiae as our test system, and evaluate alternative hypotheses for consistency with stored information. AVAILABILITY: www.hybrow.org

Protection against haemorrhagic septicaemia induced by vaccination of buffalo calves with an improved oil adjuvant vaccine.

An experimental oil adjuvant vaccine was developed against haemorrhagic septicaemia, a disease of cattle and buffalo caused by Pasteurella multocida serotype B and E. Mineral oil, Mercol 52, was used as adjuvant together with Span 85 and Tween 85 as emulsifiers. The vaccine was evaluated by single dose intramuscular immunisation of 1-2 year old buffalo calves. IgG and IgM class antibodies were determined by ELISA. The group of animals immunised with the experimental oil adjuvant vaccine showed a high titre of the IgG class of antibodies measured at 300 days post vaccination. To compare the protective efficacy of the vaccine with the commonly used broth bacterin, another group of buffalo calves was immunised by broth bacterin. This group showed a low level of IgG antibodies. Protection was assessed by challenge with 10(9) viable bacteria of P. multocida type B:2,5 administered subcutaneously, 250 days post vaccination. Animals vaccinated with the experimental oil adjuvant vaccine were fully protected. The other groups of animals, vaccinated with broth bacterin or used as control (non-vaccinated), developed symptoms of haemorrhagic septicaemia. A strong relationship between IgG but not IgM class antibody level and resistance to challenge was observed. The experiment demonstrated that the experimental oil adjuvant vaccine was superior to broth bacterin in providing protection against experimental haemorrhagic septicaemia in young buffalo calves beyond 250 days.

Vacuolating cytotoxic activity of Pasteurella multocida causing haemorrhagic septicaemia in buffalo and cattle.

The toxic activity of Pasteurella multocida strains which cause haemorrhagic septicaemia (HS) in buffalo and cattle was examined in a mouse model. Mice were injected intraperitoneally with 10(2) cells of P. multocida serotype B:2,5. Electron microscopy of peritoneal macrophages obtained 6 h after injection revealed strong induction of cytoplasmic vacuolation, macrophage lysis and death. In vitro experiments with the mouse macrophage cell line RAW 264 incubated with cultures of various HS- and non-HS-associated strains of P. multocida or with culture supernatants revealed macrophage vacuolation when HS-associated strains were used. On pre-incubation of the strains with antiserum obtained from buffalo infected with P. multocida serotype B:2,5 no vacuolation was observed. These results are indicative of the presence of vacuolating cytotoxic activity in HS-associated strains of P. multocida.

Effect of formulation and process variables on porosity parameters and release rates from a multi unit erosion matrix of a poorly soluble drug.

The effect of drug loading, water required for granulation and spheronization time on porosity parameters (intrusion-extrusion isotherms, pore size distribution, total pore surface area, mean pore diameter, shape and morphology of pores) and drug release rates from pellets of a poorly soluble drug was investigated. Porosity parameters were determined by mercury intrusion porosimetry. The drug loading was found to have a profound effect on the porosity parameters. Pellets with low drug loading showed increased pore surface area with small mean pore diameters and an increased number of total pores. On the other hand, pellets with high drug loading had decreased pore surface areas with larger mean pore diameters and a reduction in the total number of pores. With high drug loading, the drug release rate decreased. Water required for granulation had a direct effect on the total porosity of the pellets. Spheronization time from 2 to 10 min had a pronounced effect on porosity parameters and release rates. No changes in porosity parameters and release rates were observed from 10 to 20 min of spheronization time. It was shown that each porosity parameter investigated was well correlated with drug release rates and thus it is important to study the effect of porosity parameters in evaluating the in vitro performance of the multi-unit erosion matrix for the controlled release of a poorly soluble drug.

Critical processing factors affecting rheological behavior of a wax based formulation.

The use of a wax-based vehicle is one approach to stabilize a drug which is susceptible to hydrolysis and/or oxidation. The drug used in the study, as a microfine powder, is dispersed in the wax mixture and encapsulated in a soft gelatin capsule. To ensure reproducibility of drug content uniformity and encapsulability of the soft gelatin capsule dosage form, optimal viscosity and lot to lot uniformity of the viscosity of the suspension are required. The objective of the study was to identify the critical processing factors which could affect the rheological behavior of the wax based vehicle. Rheological behavior of the vehicle at temperatures ranging from 15 to 90 degrees C was evaluated using a CSL Rheometer equipped with parallel plates and a shear rate sweep mode, unless otherwise specified. Viscosity vs. temperature profiles of the vehicle were determined using the same conditions at different cooling rates ranging from 1.3 to 20 degrees C per min. Three distinct regions of phase transition of the wax mixture can be seen in the Arrhenius plot: (i) a sol region at temperatures above 50 degrees C, (ii) a transition of gel to sol at temperatures ranging from 30 to 45 degrees C, and (iii) a gel region at temperatures below 30 degrees C. The vehicle in a sol region behaved as a Newtonian fluid, indicating minimal interactions between the hydrocarbon chains of the vehicle. The vehicle in a gel region behaved thixotropic in nature, as indicated by a hysteresis loop. The shear rate had a more pronounced effect on the area of thixotropy than the shear time. The cooling rate had a pronounced effect on the resultant viscosity. At the same applied shear rate, the vehicle which was cooled at a faster rate, may cause a recrystallization of the wax mixture in different crystalline forms, resulting in a higher viscosity than the vehicle cooled at a slower rate. This effect was more pronounced when the shear was applied at a lower rate. The results of this study indicate that shear rate and cooling rate are the critical processing factors in controlling the viscosity of the final product and must be well controlled in the manufacturing procedure.

Stability of a hydrophobic drug in presence of hydrous and anhydrous lactose.

The chemical stability of a hydrophobic Leukotriene receptor antagonist drug was investigated in the presence of lactose (both hydrous and anhydrous) under various humidity and temperature conditions. The effect of wet-granulation and direct-mixing on the stability of the drug was also studied. Mixtures of drug:lactose in the ratio 1:25, 1:50 and 1:100 were prepared and analyzed over a 6 week period after storage at 40, 83 and 97% RH (all at 25 degreesC) and 75% RH at 40 degreesC. The mixtures were subjected to LOD, Karl--Fischer titrimetry, HPLC and DSC analysis to evaluate the amount of moisture pickup, percent potency and presence of drug-moisture-lactose interaction. Mixtures containing lactose anhydrous picked up more moisture and exhibited greater drug degradation than those containing lactose hydrous. Also, mixtures stored under high temperature and humidity condition showed greater moisture uptake than those stored at high humidity alone. Lactose anhydrous becomes hydrated on exposure to high humidity/temperature and storage conditions. The transition state of lactose and not its stable state may be responsible for its greater interaction and subsequent degradation of the drug. Therefore, the normal belief that lactose anhydrous, which has less than 0.5% moisture, should provide greater stability as compared to lactose hydrous, needs to be properly evaluated.

Controlled release of drugs from injectable in situ formed biodegradable PLGA microspheres: effect of various formulation variables.

A novel in situ method for the preparation of injectable biodegradable poly(lactide-co-glycolide) (PLGA) microspheres for the controlled delivery of drugs is described here. A stable dispersion of PLGA microglobules ('premicrospheres' or 'embryonic microspheres') in a vehicle mixture on injection, comes in contact with water from aqueous buffer or physiological fluid, thereby hardening the microglobules into solid matrix type microparticles entrapping the drug (in situ formed microspheres). The drug is then released from these microspheres in a controlled fashion. The effect of the following formulation variables on the characteristics of the novel drug delivery system (NDDS) was investigated: (i) the concentrations of polyethylene glycol 400 (PEG 400), the encapsulated drug, and the hydrophilic excipient (mannitol); and (ii) the types of encapsulated drug (micromolecules and macromolecules such as protein) and vehicles (replacing triacetin and Miglyol 812 by triethyl citrate and soybean oil respectively). Also, the effect of formulation, process, and storage (15 days/4 degrees C) conditions on the physical stability of the encapsulated protein was evaluated. The in vitro drug release was enhanced with decrease in the PEG 400 concentration and increase in the drug and mannitol concentration. The drug release was retarded with increase in the molecular weight of the encapsulated drug. Substitution of triacetin by triethyl citrate and miglyol 812 by soybean oil resulted in variation in the release of the drug from the in situ formed microspheres. A preliminary investigation of the physical stability of the myoglobin revealed that the alpha-helical structure was unaffected by the formulation, process, and the storage conditions.

Evaluation of drug formulations using LD50 testing in mice.

The LD50 values were utilized to assess the relative rate of absorption of two very poorly soluble drugs. Formulations of these drugs were studied by micronization; addition of surfactant, alkaline or buffering agents, and/or bile salts; coprecipitation; melt or fusion techniques; or granulation with hydrophilic agents. Differences in toxicities were demonstrated from formulations compared to pure drugs by the LD50 method. This study shows that the LD50 is a practical, rapid method of achieving comparative evaluations of drug formulations.

Release performance of a poorly soluble drug from a novel, Eudragit-based multi-unit erosion matrix.

Mechanisms governing the release of drugs from controlled delivery systems are mainly diffusion, osmosis and erosion. For poorly soluble drugs, the existing mechanisms are limited to osmosis and matrix erosion, that are commonly observed in single unit matrix dosage forms. This study reports formulation and dissolution performance of Eudragit L 100 55 and Eudragit S 100 based multi-unit controlled release system of a poorly soluble thiazole based leukotriene D(4) antagonist, that was obtained by an extrusion/spheronization technique. Effect of triethyl citrate, that was incorporated in the matrix, on the dissolution performance of the drug was also evaluated. In vitro matrix erosion and drug release from the pellets were determined by the use of USP Dissolution Apparatus I, pH 6.8 phosphate buffer, gravimetry and UV spectrophotometry, respectively. Results obtained demonstrated that matrix erosion and drug release occurred simultaneously from the pellets. Pellets eroded with a consequent reduction in size without any change in the pellet geometry for over 12 h. Matrix erosion and drug release followed zero order kinetics. Data obtained strongly suggested a polymer controlled, surface erosion mechanism.

Safety and efficacy of an oil-adjuvant vaccine against haemorrhagic septicaemia in buffalo calves: cross-protection between the serotypes B:2,5 and E:2,5.

The safety, efficacy and duration of immunity of an improved oil-adjuvant vaccine against haemorrhagic septicaemia, containing inactivated cells of Pasteurella multocida serotype B:2,5, were tested in young buffalo calves in Pakistan. For safety testing, five buffalo calves were vaccinated intramuscularly with twice the normal dose, and six weeks later with a normal dose. Except for a transient rise in rectal temperature at six hours after the vaccinations, no systemic reactions were observed. The buffaloes remained in good condition and had a normal appetite. No local reactions were observed at the injection site. For efficacy testing two trials were carried out. In the first, buffalo calves were vaccinated intramuscularly either with two doses two-and-a-half months apart, or with a single dose, or left unvaccinated. They were challenged subcutaneously with virulent P multocida after eight, 13 or 15 months. After challenge at eight months the four buffaloes given two doses and the buffalo given one dose were protected, whereas the control animal developed the typical signs of the disease. After the challenges at 13 and 15 months, the vaccinated animals were still protected whereas the control animals died. In the second trial, buffalo calves were vaccinated intramuscularly either with two doses two months apart, or with a single dose at two months or left unvaccinated. The buffaloes were challenged after eight or 14 months. After challenge at eight months the four control animals died, whereas three of the four buffaloes given a single dose were protected. After challenge at 14 months, the three control animals died, whereas four of the five buffaloes given two doses and both the buffaloes given a single dose were protected. To test for cross-protection against the heterologous serotypes E:2,5 and B:3,4, groups of mice were vaccinated once or left unvaccinated. Four weeks later, the vaccinated and control groups were challenged with a dilution series of the different challenge cultures. The vaccine appeared to induce protection against challenge with different strains of serotypes B:2,5 and E:2,5 but not against strains of serotype B:3,4.

A modified bioassay for microbial mosquito larvicides.

The proposed bioassay-the grid assay-eliminates the influence of cannibalism while allowing recycling to take place. The LC50 values of the grid assay were intermediate to those of the group and the individual assay. The grid assay thus reflects the true toxicity values of the microbial pesticides being tested. Viable count studies of the microbial pesticides employed, viz., Bacillus thuringiensis H-14 and Bacillus sphaericus 2362, performed along with the assay support the results of the bioassay.

Percutaneous needle biopsy in exudative ascites.

Seventy-two patients with exudative ascites were subjected to percutaneous needle biopsy of peritoneum. The overall diagnostic yield was 64%, with the commonest finding being caseous granuloma (39%), followed by metastatic carcinoma (25%). The procedure was safe and easy to perform on the bed side. Peritoneal tissue was obtained in all cases, without any major complications.

Toward enhanced pharmacovigilance using patient-generated data on the internet.

The promise of augmenting pharmacovigilance with patient-generated data drawn from the Internet was called out by a scientific committee charged with conducting a review of the current and planned pharmacovigilance practices of the US Food and Drug Administration (FDA). To this end, we present a study on harnessing behavioral data drawn from Internet search logs to detect adverse drug reactions (ADRs). By analyzing search queries collected from 80 million consenting users and by using a widely recognized benchmark of ADRs, we found that the performance of ADR detection via search logs is comparable and complementary to detection based on the FDA's adverse event reporting system (AERS). We show that by jointly leveraging data from the AERS and search logs, the accuracy of ADR detection can be improved by 19% relative to the use of each data source independently. The results suggest that leveraging nontraditional sources such as online search logs could supplement existing pharmacovigilance approaches.

Tramadol inhibits the contractility of isolated human myometrium.

This study was conducted to determine whether the atypical opioid analgesic tramadol inhibits the contractility of isolated non-pregnant human myometrium. Ten strips of non-pregnant human myometrium stimulated with 55 mm potassium chloride (KCl) were treated with three concentrations (30, 100 and 300 µm) of tramadol to test for any inhibitory effect of tramadol. The effects of concurrent administration of the ß adrenoceptor antagonist propranolol (1 µm), the guanylyl cyclase and nitric oxide synthase inhibitor methylene blue (20 µm) and the opioid receptor antagonist naloxone (100 µm) with tramadol were also studied. Tramadol caused a concentration-dependent inhibition of KCl-induced myometrial contractility, which was statistically significant at all three concentrations of tramadol used. Propranolol significantly reversed the inhibitory effect of 100 µm tramadol on KCl-induced myometrial contractility but not that of 300 µm tramadol. Neither methylene blue nor naloxone reversed the inhibitory effect of tramadol on KCl-induced myometrial contractility. These results suggest that tramadol inhibits KCl-induced contractility of isolated human myometrium. They also suggest that tramadol relaxes the myometrium due to stimulation of ß1 adrenoceptors. However, the concentrations of tramadol required to relax the myometrium were high and likely to be attained at toxic doses, rather than therapeutic doses, of tramadol.

Performance of pharmacovigilance signal-detection algorithms for the FDA adverse event reporting system.

Signal-detection algorithms (SDAs) are recognized as vital tools in pharmacovigilance. However, their performance characteristics are generally unknown. By leveraging a unique gold standard recently made public by the Observational Medical Outcomes Partnership (OMOP) and by conducting a unique systematic evaluation, we provide new insights into the diagnostic potential and characteristics of SDAs that are routinely applied to the US Food and Drug Administration (FDA) Adverse Event Reporting System (AERS). We find that SDAs can attain reasonable predictive accuracy in signaling adverse events. Two performance classes emerge, indicating that the class of approaches that address confounding and masking effects benefits safety surveillance. Our study shows that not all events are equally detectable, suggesting that specific events might be monitored more effectively using other data sources. We provide performance guidelines for several operating scenarios to inform the trade-off between sensitivity and specificity for specific use cases. We also propose an approach and demonstrate its application in identifying optimal signaling thresholds, given specific misclassification tolerances.

Pharmacovigilance using clinical notes.

With increasing adoption of electronic health records (EHRs), there is an opportunity to use the free-text portion of EHRs for pharmacovigilance. We present novel methods that annotate the unstructured clinical notes and transform them into a deidentified patient-feature matrix encoded using medical terminologies. We demonstrate the use of the resulting high-throughput data for detecting drug-adverse event associations and adverse events associated with drug-drug interactions. We show that these methods flag adverse events early (in most cases before an official alert), allow filtering of spurious signals by adjusting for potential confounding, and compile prevalence information. We argue that analyzing large volumes of free-text clinical notes enables drug safety surveillance using a yet untapped data source. Such data mining can be used for hypothesis generation and for rapid analysis of suspected adverse event risk.