Tuesday, March 26, 2013

Clinical Research Oriented Workshop (CROW) Meeting: March 21, 2013



Present:  Marianne Burke, Abby Crocker, Kairn Kelley, Amanda Kennedy, Ben Littenberg, Connie van Eeghen
Guest:     Molly Schmidt
1.                  Start Up: Mollie Schmidt has a PhD in School Psychology and is certified in Clinical Psychology.  She is visiting from Rome, Maine and trying to figure out what an Assistant Professor does with all her time (that would be her daughter, Connie).  She was warmly welcomed.

2.                  Presentation: Kairn Kelley – F31
a.       Kairn has a goal of submitting her F31 proposal in April, and has her foot hard on the accelerator.   She shared the significance and specific aims.  She wants feedback on whether her story makes sense and what measures she should use to capture her results.
                                                  i.      The reason for the study is well-defined: the problem is common, morbid, and treatable but the diagnostic tests are not well-supported.  The issue Kairn is addressing the diagnostic value of the tests and which tests to use, and how to evaluate the results.
                                                ii.      Set the stage: we don’t know these tests, even at the level of dichotic and non-dichotic tests.
                                              iii.      The message, which has to be readable by looking at the first sentence of each paragraph, is something like: “The tests are mysterious, we’ll fix the tests, and patient care will be better.”  Use subheadings to lead the reader through the story: Scope of the Problem (what do you want to fix), Tools (how to diagnose the condition is based on tests not subjected to rigorous evaluation of the psychometric properties (or text characteristics or reliability) to date.  Dichotic tests (describe) in particular, have not been well…
                                              iv.      In addition, the Overview section needs to be shorter and less dense, even though the advice is also to add more detail.  (Good luck with this.)
b.      Audience: Institute on Deafness and Hearing Disorders, with a high proportion of applications accepted.  The study section is unique to F proposals; it is helpful to identify them to help plan how to write the story.  Know what their priorities are in research and be ready to focus definitions on their likely level of understanding.  Study section titles include: Brain Disorders and Language Communications; Behavioral Neuroscience; Sensory Motor Science Cognition and Perception; and Neuro-Degeneration.  Be clear about what you are going to accomplish. 
                                                  i.      The study section for Brain Disorders includes radiologists, critical care specialists, psychologists, neurologists, pharmacologist… no audiologists.
                                                ii.      Sensory motor science: has fMRI studies; list is not available.
                                              iii.      Contact the Project Officer – find out if these are the right study sections and if one can request a study section.
c.       Specific Aims: this is the most commonly read page across the entire study section.
                                                  i.      Should it include the training and mentoring plan?  Per the Project Officer, everything should demonstrate how it fits into your training plan with the result that you will be an “independent, successful investigator and author of RO1s”
                                                ii.      Look at previous F applications (narratives and aims).  Gabe’s was strictly focused on the research study, not the training plan.  Consider stating the broad goal: lay the foundation of diagnostic testing and improve accuracy.  Specifically for this project, the focus is test/retest reliability.  We will achieve this by meeting the following specific aims…
                                              iii.      Start the story here: children are being harmed due to lack of accurate testing and follow up.  Consider using the “Objectives” as the Aims statement; move the specificity of the Aims to the Approach section.  Create an elevator speech with the four essential sentences: relevance, question, action, future.
                                              iv.      The third Aim is expressed as a hypothesis test; do this formally with an H0 and a null.  Otherwise, remove the language.  If so, use “compare” or “evaluate” and avoid being boxed into an assumption of the reader’s.  Leave “registry” for the future work planned or deliverables.
d.      Measures:
                                                  i.      Subjects: originally 7-10 yr. old, due to guidelines (for <7) and consent requirements (>10). Recommendation: go as high as clinically relevant to improve power calculations (up to 14 yrs.).  Training effectiveness is highest from 7-12.  Check clinical guidelines. 
                                                ii.      Inclusion: clinical practice guideline eligibility and other subject characteristics.  Parents will be asked if children have received special services but these children will be included. 
                                              iii.      Exclude: febrile, coughing, otitis media, and other acute problems (pain, rash,…).  IQ is not an exclusion criteria
                                              iv.      Goal: 125 complete analyzable cases
e.       Limitations: one practice, one location; clinical relevance required small, focused sample; conducted under field conditions.
f.       Good luck!  Not that Kairn will need luck, but at least may Good Fortune go with all emerging PhD students – and this would include Abby and her upcoming dissertation defense (see below).

3.                  Next Workshop Meeting(s): Thursday, 2:00 p.m. – 3:30 p.m., at Given Courtyard South Level 4. 
a.       Mar 28: Abby: Dissertation Defense Dry Run (will start at 2p, on schedule – new location?) in HSRF 300
b.      April 4:
c.       April 11:
d.      April 18:
e.       April 25:
f.       May 2:
g.      May 9: Charlie: Exploration of analytical plan for Natural History of Acute Opiate Use (and perhaps more)
h.      Future agenda to consider:
                                                  i.      Christina Cruz, 3rd year FM resident with questionnaire for mild serotonin withdrawal syndrome?
                                                ii.      Peter Callas or other faculty on multi-level modeling
                                              iii.      Charlie MacLean: demonstration of Tableau
                                              iv.      Journal article: Gomes, 2013, Opioid Dose and MVA in Canada (Charlie)

Recorder: Connie van Eeghen

Wednesday, March 20, 2013

Clinical Research Oriented Workshop (CROW) Meeting: March 14, 2013



Present: Abby Crocker, Kairn Kelley, Amanda Kennedy, Rodger Kessler, Charlie MacLean, Ben Littenberg, Connie van Eeghen (updated in a conversation with Peter Kreiner 3/15/13)
1.                  Start Up: New data sources out on health behavior and geo-locations available (driver license data – 18 states); Ben has the details.

2.                  Presentation: Charlie – Research Question
a.       Charlie reviewed the priorities that he, Liz, and Neil have:
                                                  i.      Opiates Pharmaco-epi
                                                ii.      Workforce snapshot: estimating efforts, locations, and distributions of providers throughout the state, based on claims.  To be used as a forecasting tool for AHEC, Blueprint, Green Mountain Care Board
1.      Licensure data has data problems: low response rate for survey data
2.      This clean-up process is a necessary precursor to using VHCURES for other analytical projects
                                              iii.      (Atypical antipsychotics – included in the request for opiate data, see i above)
                                              iv.      Back pain: Sharon Henry (RMS) for PT visits and costs
b.      Current operational
                                                  i.      Clean up provider data: next 4 weeks (plan for end of May)
                                                ii.      Patient data need to be cleaned?
c.       Opiate data:
                                                  i.      2 – 3 years of data, all VT insured patients including part D Medicare (no cash)
                                                ii.      Based on a claim for at least one opiate
                                              iii.      Data: drug claim, patient demographics, prescription and prescriber data
d.      Outcomes we could study:
                                                  i.      Natural history of acute opiate use converting to chronic – descriptive study
                                                ii.      Survival analysis: first fill of prescription to last
1.      What proportion of acutes become chronic users?
2.      What variables describe chronic users (age, gender, increase/decreased use, co-interventions: psych meds, nicotine replacements, heart disease…)
3.      May be able to develop a prevalence indicator, over years and seasons, compared to census data (zipcode, age, and sex adjusted) including median income
                                              iii.      Opiate intensity over time: morphine equivalence based on prescribed dose. 
Added after discussion with Peter Kreiner
                                              iv.      Cross-validate against EHR/Claims
                                                v.      Multiple interrupted time series by state of health outcomes (e.g. adverse events [deaths, overdoses…], questionable activity, red flags…)
                                              vi.      Multiple interrupted time series by state of prescribing changes (red flags)
                                            vii.      Get use data (FL?) as predictor and provider or patient level red flag data (future request)
e.       Future data request:
                                                  i.      Entire med list
                                                ii.      Diagnosis – from office visits or procedures; messier data
f.       Next steps:
                                                  i.      Workplan
1.      IRB approval – CM/pending
2.      Literature review: restrict to epidemiology – Amanda & Kairn/mid-May, to include the Introduction section to the manuscript
a.       Sources: CDC – MMWR, VPMS annual report
b.      Check MeSH headings for opioids, acute, chronic, longitudinal
c.       Develop a search strategy
d.      Original research and reviews… for the reviewers to decide
e.       Abby and Charlie have supplemental Endnote libraries
3.      Protocol and Methods, one for each research area (to define variables & statistical process)
a.       Charlie to provide data dictionary and field values
b.      Natural history: Charlie, Ben, Abby, & Rodger/Sylvie to set up 1st meeting on or after April 8 for an hour; protocol outline done and methods section sketched by 3rd week of April
                                                                                                                          i.      Consider developing the protocol for #3 at the same time (opiate dose/intensity)
c.       Prevalence of Opiate use by zip/census (consider clustering the data by region/primary care service areas)
4.      Data cleaning and creating variables – Charlie/mid-April
5.      Merging census (not needed for research areas 1 and 3) -
6.      Analysis – To be started at CROW May 9
a.       Intro and Methods sections need to be complete by this date
7.      Report writing – completed by end of May 2013
a.       Project coordinator/communicator and Lead Author: Abby Crocker
b.      Last author: Charlie (remainder of us will listed be alphabetically)
                                                ii.      NAPCRG annual meeting in Fall 2013 in Ottawa (North America Primary Care Research Group); deadline for abstracts is in April 17th  
1.      Charlie to submit FAHC data
2.      Amanda to submit caffeine study – maybe?
3.      Connie to submit Opiate Rx Management QI study results
g.      Peter Kreiner update meeting (3/15/13)
                                                  i.      His data bases include 3 states: CA, FL, and ME (more history), going back several years (at least 2011 on)
                                                ii.      HID is the vendor for two of these states (same as VT)
                                              iii.      Data are in different formats but all are SAS readable; could generate a SAS transport data set.  CDC has asked for an on-line portal to access the data.
                                              iv.      A request for data would go through an oversight committee of federal and state partners (they meet monthly by phone) who would review the protocol.  The data could be transferred through the UVM FTP site.  Brandeis IRB protocol already in place.
                                                v.      All records have patient identifier, year of birth, zip, NDC, provider identifier, provider zipcode, quantity, refills, pharmacy identifier, pharmacy zipcode, payment source
1.      CDC has provided a red book (2011); to be replaced with a broader set of NDC codes for 2012 (subscription is ~$15,000/year) to translate the NDC code into drug and drug category
2.      DAWN data set is another possible source with NDC data
                                              vi.      Peter will share the elements of the standard reports with Charlie and Ben
1.      Multi-state studies would require HID support, as the DEA number is not available in the data set
                                            vii.      Possible research topics discussed from CROW meeting above.  Possible work plan:
1.      Conduct study using VHCURES all payor data set on the Natural History of Acute Opiate Use (Peter indicated that this is high in the Novel part of FINER.  Peter has seen only one such research paper [from Maine] which involved the loss of the PDMP for a short period of time.)
2.      Request for similar data from Brandeis.  No cost estimate at this time; may be part of collaborative work done jointly.  May result in a fundable grant submitted jointly.
3.      Follow up plans could include cross-state comparisons, selection of likely interventions resulting in RO1 proposal.  Peter was interested.  Some research funding is currently available, depending on the proposal.  Other states are coming along: Delaware, TX, …

3.                  Next Workshop Meeting(s): Thursday, 2:00 p.m. – 3:30 p.m., at Given Courtyard South Level 4. 
a.       Mar 21: Kairn: F31
b.      Mar 28: Abby: Dissertation Dry Run (will start at 2p, on schedule – new location?)
c.       May 9: Charlie: Exploration of analytical plan for Natural History of Acute Opiate Use (and perhaps more)
d.      Future agenda to consider:
                                                  i.      Christina Cruz, 3rd year FM resident with questionnaire for mild serotonin withdrawal syndrome?
                                                ii.      Peter Callas or other faculty on multi-level modeling
                                              iii.      Charlie MacLean: demonstration of Tableau
                                              iv.      Journal article: Gomes, 2013, Opioid Dose and MVA in Canada (Charlie)

Recorder: Connie van Eeghen