Publications

This study investigated the correspondence among four groups of raters on adherence to STAGE-12, a manualized 12-step facilitation (TSF) group and individual treatment targeting stimulant abuse. The four rater groups included the study therapists, supervisors, study-related ("TSF expert") raters, and non-project related ("external") raters. Results indicated that external raters rated most critically mean adherence - the mean of all the adherence items - and global performance. External raters also demonstrated the highest degree of reliability with the designated expert. Therapists rated their own adherence lower, on average, than did supervisors and TSF expert raters, but therapist ratings also had the poorest reliability. Findings highlight the challenges in developing practical, but effective methods of fidelity monitoring for evidence based practice in clinical settings. Recommendations based on study findings are provided.

Colonization of the nasopharynx by Streptococcus pneumoniae is a necessary precursor to pneumococcal diseases that result in morbidity and mortality worldwide. The nasopharynx is also host to other bacterial species, including the common pathogens Staphylococcus aureus, Haemophilus influenzae, and Moraxella catarrhalis. To better understand how these bacteria change in relation to pneumococcal colonization, we used species-specific quantitative PCR to examine bacterial densities in 52 subjects 7 days before, and 2, 7, and 14 days after controlled inoculation of healthy human adults with S. pneumoniae serotype 6B. Overall, 33 (63%) of subjects carried S. pneumoniae post-inoculation. The baseline presence and density of S. aureus, H. influenzae, and M. catarrhalis were not statistically associated with likelihood of successful pneumococcal colonization at this study's sample size, although a lower rate of pneumococcal colonization in the presence of S. aureus (7/14) was seen compared to that in the presence of H. influenzae (12/16). Among subjects colonized with pneumococci, the number also carrying either H. influenzae or S. aureus fell during the study and at 14 days post-inoculation, the proportion carrying S. aureus was significantly lower among those who were colonized with S. pneumoniae (p = 0.008) compared to non-colonized subjects. These data on bacterial associations are the first to be reported surrounding experimental human pneumococcal colonization and show that co-colonizing effects are likely subtle rather than absolute.

Online physician reviews are a massive and potentially rich source of information capturing patient sentiment regarding healthcare. We analyze a corpus comprising nearly 60,000 such reviews with a state-of-the-art probabilistic model of text. We describe a probabilistic generative model that captures latent sentiment across aspects of care (eg, interpersonal manner). We target specific aspects by leveraging a small set of manually annotated reviews. We perform regression analysis to assess whether model output improves correlation with state-level measures of healthcare. We report both qualitative and quantitative results. Model output correlates with state-level measures of quality healthcare, including patient likelihood of visiting their primary care physician within 14 days of discharge (p=0.03), and using the proposed model better predicts this outcome (p=0.10). We find similar results for healthcare expenditure. Generative models of text can recover important information from online physician reviews, facilitating large-scale analyses of such reviews.