Author Interviews, Hospital Readmissions, UCSD / 24.11.2014

MedicalResearch.com Interview with: Ralitza P. Parina, MPH, Senior medical student John Rose, MD MPH Department of Surgery at University of California San Diego Medical Research: What is the background for this study? What are the main findings? Response: This study looked at the association between hospital 30-day readmission rates and 30-day mortality rates. While readmission rates are coming into increasing focus with CMS reimbursement cuts for hospitals with higher than expected rates, they remain a poorly studied metric of quality. High readmission rates have been unequivocally tied to increased costs, but it remains unclear whether they actually represent poor quality of care and worse outcomes for patients. We chose to compare readmission rates as a quality metric to the well-established “gold standard” of mortality. We found that 85% of hospitals did not show a correlation between readmission and mortality, i.e. their rates were not both high or both low. Furthermore, among hospitals that were outliers in at least one of the measures, almost a third were in the category of low or normal readmission rates with higher than expected mortality. The implications are twofold: first, readmission and mortality rates are not strongly correlated. Second, focusing on readmission rates as an outcome will miss a large number of poorly performing hospitals with higher than expected mortality rates but low or expected readmissions.
Author Interviews, Electronic Records, Outcomes & Safety, Surgical Research, UCSD / 09.11.2014

MedicalResearch.com Interview with: Jamie Anderson MD MPH Department of Surgery University of California, San Diego Medical Research: What is the background for this study? What are the main findings? Dr. Anderson: Risk adjustment is an important component of outcomes and quality analysis in surgical healthcare. To compare two hospitals fairly, you must take into account the “risk profile” of their patients. For example, a hospital operating on predominately very sick patients with multiple co-morbidities would be expected to have different outcomes to a hospital operating on relatively healthier patients with fewer co-morbidities. Somewhat counter-intuitively, it is possible that a hospital with a 10% mortality rate may be better than a hospital with 5% mortality rate when you adjust for the risk of the patient population. Currently, the “gold standard” database to evaluate surgical outcomes is the American College of Surgeons National Surgical Quality Improvement Program (NSQIP), which includes a number of variables on each patient to perform risk adjustment. However, collecting these variables is costly and time consuming. There is also concern that risk adjusted benchmarking systems can be “gamed” because they include data elements that require subjective interpretation by hospital personnel. With the widespread adoption of electronic health records, the aim of this study was to determine whether a number of objective data elements already used for patient care could perform as well as a traditional, full risk adjustment model that includes other provider-assessed and provider-recorded data elements. We tested this hypothesis with an analysis of the NSQIP database from 2005-2010, comparing models that adjusted for all 66 pre-operative risk variables captured by NSQIP to models that only included 25 objective variables. These results suggest that rigorous risk adjusted surgical quality assessment can be performed relying solely on objective variables already captured in electronic health records.
Blood Clots, General Medicine, UCSD / 19.05.2014

MedicalResearch.com Interview withTimothy Fernandes, M.D., M.P.H. University of California, San Diego - La Jolla, CA Timothy Fernandes, M.D., M.P.H. University of California, San Diego La Jolla, CA MedicalResearch: What are the main findings of this study? Dr. Fernandes: The fibrinopeptides are cleaved off of fibrinogen by thrombin during the generation of a new clot. These small molecules are excreted into the urine and we have developed a urine assay to measure the level of FPB. We examined the performance of urine FPB as a screening test for acute pulmonary embolism, blood clots that travel to the lungs. The study group consisted of 344 patients: 61 (18%) with pulmonary embolism and 283 (83%) without. At a threshold of 2.5 ng/ml, urine FPB demonstrated sensitivity of 75.4% (95% CI: 62.4-85.2%), specificity of 28.9% (95% CI: 23.8-34.7%), and negative likelihood ratio of 0.18 (0.11-0.29), weighted by prevalence in the sample population. However, the thresholds of 5 ng/ml and 7.5 ng/ml had sensitivities of only 55.7% (95% CI: 42.5-68.2%), and 42.6% (30.3-55.9%), respectively. The urine fibrinopeptide B assay at a cut-off of 2.5 ng/ml had a sensitivity of 75.4% for detecting pulmonary embolism. For diagnosis of PE, this sensitivity is comparable to previously published values for the first generation plasma latex and whole blood D-dimer assays (not as well and the D dimer ELISA assay).
Author Interviews, PLoS, Toxin Research, UCSD / 30.03.2014

Beatrice A. Golomb MD, PhD Professor of Medicine Family and Preventive Medicine University of California, San DiegoMedicalResearch.com Interview with: Beatrice A. Golomb MD, PhD Professor of Medicine Family and Preventive Medicine University of California, San Diego MedicalResearch.com: What are the main findings of the study? Dr. Golomb: The main finding is that veterans with Gulf War illness have bioenergetic defects -- dysfunction of mitochondria, the energy producing elements of cells -- that is evident in comparing affected veterans to matched healthy controls. An estimated 1/4 to 1/3 of the ~700,000 US veterans from the 1990-1 Gulf War developed chronic multisymptom health problems that entail fatigue, cognitive and other CNS problems, muscle pain, weakness and exercise intolerance, with high rates of gastrointestinal (especially diarrhea) and neurological problems, and other symptoms - as well as autonomic dysfunction. Evidence suggests these problems have not abated with time. Veterans from other nations that have conducted epidemiological studies, including the UK, Canada, and Australia, also show elevated rates of problems.
Autism, Genetic Research, NEJM, UCSD / 26.03.2014

MedicalResearch.com Interview with: Dr. Erik Courchesne PhD Professor, Department of Neurosciences UC San Diego School of Medicine MedicalResearch.com: What are the main findings of the study? Dr. Courchesne: “Building a baby’s brain during pregnancy involves creating a cortex that contains six layers,” Courchesne said. “We discovered focal patches of disrupted development of these cortical layers in the majority of children with autism.” The authors created the first three-dimensional model visualizing brain locations where patches of cortex had failed to develop the normal cell-layering pattern. The study found that in the brains of children with autism key genetic markers were absent in brain cells in multiple layers. “This defect,” Courchesne said, “indicates that the crucial early developmental step of creating six distinct layers with specific types of brain cells – something that begins in prenatal life – had been disrupted.”  The study gives clear and direct new evidence that autism begins during pregnancy.