Skip to main content
Center for Healthcare Policy and Research

Center for Healthcare Policy and Research

2013 Outcomes Research Pilot Awards

puzzle pieces

The CHPR is proud to announce the recipients of the 2013 Outcomes Research Pilot Awards: 

Stephen G. Henry, M.D., M.Sc. (Department of Internal Medicine)
"Characterizing opioid dose escalation during early chronic opioid therapy for pain"

Decreasing the incidence of high-dose opioid therapy for chronic musculoskeletal pain requires a clear understanding of how patients on chronic opioid therapy progress to high-dose therapy in the first place. Knowledge of the patient factors and time periods associated with opioid dose escalation will enable re- searchers to design interventions for decreasing the incidence of opioid dose escalation early in the course of chronic opioid therapy that are targeted to the patients and time periods when they are likely to have the greatest effect. Preventing inappropriate opioid dose escalation is likely to be an effective strategy for de- creasing the public health risks associated with high-dose opioid therapy relative to tapering down patients already taking high-dose opioids.

The overall goal of this project is to identify patient factors associated with opioid dose escalation, the time period when dose escalation is likely to occur, and clinicians’ documented reasons for continuing versus escalating opioid doses during patients’ first year of opioid therapy for chronic musculoskeletal pain. This project comprises a case-control study nested within a retrospective cohort study identified with data from the UC Davis electronic health record. Opioid-naïve patients followed at UC Davis who received a new opioid prescription for musculoskeletal pain in 2011 and who remained on chronic opioid therapy 1 year later will be identified using the UC Davis electronic health record. Based on their daily opioid dose at 1 year, patients in this cohort will be classified as on either stable opioid therapy (≤25% increase over initial daily dose at 1 year) or escalated opioid therapy (>25% increase over initial daily dose).

First, patient factors associated with dose escalation vs. stable opioid therapy will be identified using multiple logistic regression. Next, detailed clinical data will be collected from a random sample of 50 patients each from both the stable and escalated groups (100 total). For each of these patients, trained chart abstractors will collect the prescribed daily opioid dose and the documented clinical reasoning related to opioid prescribing for each encounter during the patient’s first year on chronic opioid therapy. Data abstracted from these charts will be used to characterize the temporal pattern of opioid dose escalation during the first year of chronic opioid therapy and to identify a critical time period during which clinically significant dose escalation is likely to occur. Opioid prescriptions are handwritten (vs. being ordered through the electronic health record), so chart review is required to obtain the detailed clinical data necessary to analyze changes in opioid dosing over time. A 2-level random-effects linear regression with repeated measures will be used to model changes in daily opioid doses over time. Finally, clinicians’ documented reasons for continuing vs. escalating opioid doses at each visit will be analyzed using qualitative content analysis. A descriptive analysis comparing the documented clinical reasoning in the stable vs. escalated opioid therapy groups will be performed to generate hypotheses about whether and how clinicians’ reasoning differs for these two groups and to facilitate interpretation of the quantitative analyses.

This project advances the long-term goal of decreasing the incidence of high-dose opioid therapy by pre- venting inappropriate dose escalation during the first year of chronic opioid therapy for musculoskeletal pain. Data from this project will support an R21 application to the National Institute of Drug Abuse (NIDA) for a larger study investigating the natural history of new opioid prescriptions that will include data from multiple health systems to improve external validity and statistical power.


Caroline Chantry, M.D. (Department of Pediatrics)
"Delayed umbilical cord clamping for term infants: Evaluating the impact of adopting a hospital policy for vaginal delieveries and a pilot trial during Cesarean deliveries"

Multiple studies, including systematic reviews and meta-analysis of controlled trials have concluded that delayed cord clamping of the umbilical cord in full-term newborns has hematologic benefits for the newborn which extend into infancy, without significant increased risk of harm. Further, iron status at birth significantly predicts infant iron status and anemia later in infancy. Iron is essential for multiple aspects of neuronal development; iron deficiency anemia and even pre-anemic deficiency are associated with poorer neurodevelopmental outcomes. Hence, early cord clamping is placing infants at unnecessary increased risk for iron deficiency and sub-optimal neurodevelopmental outcomes. Despite expert recommendations to delay clamping the cord for at least 2 minutes, early cord clamping remains a common obstetrical practice. The World Health Organization has noted that data on cesarean deliveries are more limited, particularly with respect to long-term effects in infants. Recently, the American College of Obstetrics and Gynecology also noted in a committee statement that ” the ideal time for clamping of the umbilical cord in cesarean section vs. vaginal birth is a particularly important area for future research”.

The objectives of this study are three-fold:

1) The first objective is to evaluate the impact on patient outcomes that results from instituting a hospital policy on delayed cord clamping during vaginal deliveries of term infants. Outcomes will include: a) mean time to cord clamp and percent of term, vaginal deliveries with delayed cord clamping of at least 2 minutes; b) newborn hemoglobin levels; and c) incidence of phototherapy (initial hospitalization and readmissions for phototherapy);
2) The second objective is to conduct a pilot trial of delayed cord clamping during cesarean deliveries to determine if progressively longer intervals of cord clamping are feasible and not related to poorer maternal and neonatal outcomes as compared to outcomes from historical controls. To determine this we will measure: a) maternal blood loss as measured by: a.i) estimated maternal blood loss, a.ii) post-op hemoglobin levels and a.iii) pre-op/post-op difference in hemoglobin; b) incidence of neonatal hypothermia on admission; c) prevalence of newborn anemia and polycythemia.
3) The third objective is to plan a multi-center controlled trial of delayed cord clamping in cesarean deliveries and prepare a grant application for funding.


Garth Utter, M.D., M.Sc. (Department of Surgery)
"Anemia, transfusion and outcomes from traumatic brain injury"

Traumatic brain injury (TBI) represents a substantial public health burden in mortality and chronic disability, yet relatively little is known about optimal care for acute TBI. Patients with acute TBI frequently are anemic, but treatment of this anemia with blood transfusion must balance different risks: potential mortality risk from liberal transfusion on the one hand (particularly for young, otherwise healthy patients) versus secondary brain injury from insufficient oxygen delivery on the other hand. Different types of specialists who care for patients with TBI prioritize these risks differently and thus prefer different thresholds for transfusion.

Specific Aim: To determine whether anemia and/or transfusion during the early post-injury period are independently associated with worse neurologic outcome from severe TBI

Design and methods: A retrospective cohort study will be conducted using existing data on 2,538 patients from five prior TBI clinical trials. Anemia will be defined as ≥1 hematocrit <30%, and transfusion status will be categorized based on the amount of blood received. Multivariable logistic regression with adjustment for confounding factors will be used to determine the independent associations of anemia and transfusion with sliding dichotomy 6-month Glasgow Outcome Scale score.

As a parallel technique, propensity score analysis will also be performed to adjust for factors associated with transfusion. The validity of results will be checked with sensitivity analyses to assess the likelihood that unknown confounders and/or exposure misclassification could explain the results. For separate secondary analyses, anemia will be redefined as a function of anemia severity over time, and results of 6-month neuropsychological tests will be used as outcomes.

Significance: Evaluation of existing data will help guide clinical decisions as to whether patients with TBI should be transfused liberally or restrictively. In the absence of a well-designed randomized trial, the proposed study may provide the most definitive information on the independent influences of anemia and transfusion on TBI outcomes.