2016 Department of Biostatistics Faculty & Staff Retreat, 3 June 2016

Hachland Hill Countryside Retreat
Joelton, TN 37080
Friday 2016-06-03
8am-8pm

At the retreat we will have three meals, snacks, joint meetings, and breakout sessions for staff biostatisticians and IT personnel, staff, and faculty. There will also be time for recreation.

Please edit this page to indicate your attendance

Pictures

Attendees

Frank Harrell (Beef) Shawn Garbett (Beef Tenderloin)
Ashlee Bartley (Tuna) Dandan Liu (Tuna)
Shi Huang (Tuna) Matt Shotwell
Thomas Stewart (Beef Tenderloin) Fei Ye
Xulei Liu (Tuna) Heidi Chen (no dinner)
Hakmook Kang(beef) Mario Davidson (Beef)
Meridith Blevins (Tuna) Robert Greevy (no dinner)
Amy Perkins (Beef Tenderloin) Hank Domenico (Beef)
Tan Ding Jennifer Morse (Beef)
Guanhua Chen (no dinner) Chris Slaughter (Beef Tenderloin)
Zhengzheng Tang (no dinner) Ron Jannetta (Beef)
Lei Luo (Beef) Run Fan (beef)
Jennifer Thompson (beef) Jeffrey Blume (Beef)
Amber Hackstadt (no dinner) Dale Plummer (Beef Tenderloin)
Jackie Pennings (Veggie) Jeremy Stephens (no dinner)
Cathy Jenkins (beef) Charles Dupont (Beef Tenderloin)
Hui Nian (Tuna) Jeffrey Horner
Bill Dupont (Beef) Tatsuki Koyama (no dinner)
Dan Ayers (Tuna) Leena Choi (Tuna)
Sharon Phillips (Tuna) JoAnn Alvarez no dinner)
Meng Xu (Tuna) Samuel Nwosu (no dinner)
Yaomin Xu (Beef Tenderloin) Chris Fonnesbeck (no dinner)
Yuwei Zhu (Beef Tenderloin) Aihua Bian (no dinner)
Zhouwen Liu (Tuna) Jonathan Schildcrout (no dinner)
Cindy Chen (Beef)  

Agenda

  • Breakfast 8:00a
  • Opening 9:00a
  • Strategic planning breakout sessions 9:10a-10:30a
    • Faculty: Leaders: Jeffrey Blume and Jonathan Schildcrout
      • Graduate program: Meeting intial goals? Adjustments needed to better prepare students for their careers? How to bring in new faculty to teach / rotate faculty? How and what to add for specialty electives? How to have more short courses for students?
      • Statistical Education at VUMC: Does the department need to increase is education offerings such as courses, programming clinics, and summer short-courses?
      • National Reputation of the Department: How do others view us? What is our national reputation for science, biomedical collaboration, statistical methodology, data science and faculty? Where would we like to see our national reputation grow? Are there areas that need to be nurtured? Do we need to work to change our reputation in others? How could be do that?
      • Access to data: do we want to have EHR expertise so that we can access data on our own?
      • Mentoring junior faculty
    • Staff biostatisticians: Leaders: Cathy Jenkins & Jennifer Thompson
      • Communication - ways to improve with supervisor and PIs
      • Working with collaborators
        • Personnel/effort issues - how to negotiate these discussions
        • Technical/practical issues - tools beyond email? Practical approaches?
      • Technology
        • Explore tools like Alfresco/Slack, Google Docs, Box for collaborative use
        • Would love education on using data.vanderbilt.edu
      • How to improve transitions for new hires (first-week logistics, department culture...)
        • List of resources/education for initial down time
          • JHU data science specialization
          • Terri's R lectures
          • knitr/markdown/latex tutorials
        • Orientation to administrative stuff (a la tech orientation)
          • dbconnect
          • effort reporting
          • travel/professional development funds - how much and how to spend
          • keys to conference rooms
          • who to talk to for each admin question
        • "Buddy" on staff for first 3(?) months to ask questions about knitr, latex, etc
    • IT: Leader: Dale Plummer
      • Working with collaborators
      • Changes in type of work in next 5 years
      • Improving work flow/processes for security, efficiency, safety, reproduciblity
      • New skills to be added to remain relevant/competitive
      • Networking/firewall changes
      • What tools can application developers create to improve statistician productivity?
      • hiring/interviewing/job descriptions
      • Better use of shared resources (e.g. ACCRE, virtual servers)
  • Reassemble 10:30a
    • Breakout group reports
    • Planned new center to be strongly affiliated with the Department and VIMPH
  • Skills enhancement breakout sessions 11:15a
    • Faculty and staff biostatisticians
    • IT
      • Define skills enhancement
      • Areas where our skills are lacking
      • Resources for skills enhancement
      • Time for skills enhancement
      • Resume regular show & tell meetings?
  • Lightning round: efficiency tips, tricks, etc. 11:45a
  • Lunch noon
  • Games 1:00p
  • Administrative overview: Ron Jannetta 2:00p
    • Structure
    • Effects of the VU/VUMC administrative/financial split
    • BCC
    • Department budget: Will we have a budget surplus next year? Should we be providing more support to teaching faculty, research advisors and thesis advisors?
  • Department Size: How big do we want to get?
  • Working groups 2:30p
    • Collaborative web services
    • Percent efforts on grants policy and implementation
    • Pre-specified statistical analysis plan policy and implementation
    • Statistical computing skills - continuous enhancement
    • Operating system (updating, shortcuts, command scripts) and file editing - improving skills
    • Optimally working with collaborators
    • Enhancing inter-biostatistics faculty collaboration
    • Re-writing the department brochure and other media to develop
    • Clinics: nail down how to schedule PhD and MS biostatistician leaders for a given day; making clinics more effective
  • Working group reports 3:15p
  • Games 4:00p
  • Mixer 6:00p
  • Dinner 6:30p

Summaries and Follow-up Discussion

Education

  1. Facilitate student internships to increase our reputation and student learning/hiring opportunities.
    • Have an informational point person.
    • List the contact person on medschool.vanderbilt.edu/biostatistics and biostat.mc.vanderbilt.edu/wiki.
    • Go beyond mass emails. Have something searchable and with memory like a web or wiki page.
  2. Facilitate greater faculty interaction with students.
    • Focus on win-win opportunities for faculty already stretched too thin.
    • Encourage faculty to keep a list of projects / papers they'd like to do if they only had the time ... or a student to help them get it done.
    • Use chalk talks as a way for faculty to pitch papers they'd like to work with someone on.
    • Be careful about any web or electronic sharing of the ideas to prevent any true or perceived cases of poaching.
  3. Encourage teaching of short courses at national meetings and internally.
    • Talk to Robert Greevy if you'd like to present a series of HSR seminars to help you develop your short course at a reasonable pace (just one 1hr talk per month).
  4. Data science is the 800 pound gorilla that wants to eat our lunch.
    • Consider encouraging our students to take BMIF classes that teach working with large datasets.

Service Activities

There was a consensus that these activities are both very important and difficult to fulfill given the demands on our time from funded co-investigators and our own needs and interest in methods research and other activities that will build our national reputation. Suggestions as to how to motivate more and better service work are as follows:

  • Faculty should receive some % effort to cover service activities.
  • More recognition for service activities would be helpful. This might include public posting of lists of who was doing what service work. It was felt that not all faculty should be required to engage in all types of service work. For example, someone who is serving on a study section should be largely exempt from other service activities. Faculty should receive certificates acknowledging commendable levels of service.
  • The department should pay for travel costs for faculty who were organizing sessions at national meetings or doing other extramural service work for our professional activities.
  • Tenure-track faculty should be encouraged to serve on study sections and referee papers as this enhances their national reputation, which will help them get promoted. Faculty who decline requests for study section service should suggest the names of junior faculty who would benefit from this service activity.
Other recommendations:
  • There should be more meetings of the Middle-Tennesse chapter of the ASA. Some faculty with family obligations find it difficult to attend such meetings in the evening.
  • The department should maintain a server to develop and host web-based software that can perform interactive analysis, graphics and data management. The PLOTLY package was given as an example of such software. An open-source tool along these lines might be very helpful for web-based supplements to published papers.
  • Concerns and recommendations were made about clinics that are given below.

Pre-specifying analysis plans

  1. The objective is to help with:
    • improving scientific integrity
    • discouraging "data fishing"
    • better managing workflow
  2. This is not a problem for every research team/project, but it is for some teams/projects.
  3. One solution is to make the first few pages of the analysis report the analysis plan.
    • State the aims, data collection summary, and analysis plan.
    • Footnote noteworthy changes with dates and explanations.
    • Write the analysis report with the expectation that it will be submitted as a supporting document to be published as on online supplement. And/or publish it on your wiki yourself.

Clinics

Clinics are often the first interaction that faculty from other departments have with us and it is important that this interaction go well. It was felt that the current system of having clients broadcast needs to the clinic listserve does not work well, as no one is responsible for answering or following through on these requests. The following suggestions were made:
  • Clients should be able to post their needs and preferred time directly on the clinic web page. The client will also complete a checklist that includes their name, title, department, and purpose (VICTR, study design, analysis for manuscript, etc.) Someone on staff should be assigned to monitor these requests, to make sure that no more than two people sign up for a given day, that the day chosen is appropriate, to communicate with clients to confirm dates and send calendar notifications, to respond to email inquiries, and to make sure Biostatistics faculty/staff have signed up to cover each clinic.
  • A clinic facilitator needs to take responsibility for each day. This person will agree to come to clinic on their day, to lead the clinic, and to bring in other experts when needed. It was felt that there are currently enough faculty that this would not be too burdensome (each faculty member would need to lead about 8 clinics a year). Similarly, staff statisticians would also be asked to sign up to attend about 14 clinics a year so that each clinic would be guaranteed to have at least one faculty and staff member in attendance. A VICTR-associated person should attend, when requested, to provide a brief introduction to the application process. All departmental members would be encouraged to attend more clinics as time permits. This this would prevent the embarrassing situation in which a client comes to clinic and no one is there to help.
  • Consider asking senior students lead at least one clinic a year.
  • We need a brochure that we can give to clients and post on the web that clearly states what services we provide in clinic, and who we provide them to. Clients who need services that we do not provide should be encouraged to take this brochure to their department/division chiefs and complain. A list of active collaborations and what they pay for statistical services will be included. It should be made clear that we do not analyze data or review lengthy proposals in advance of the clinic. There will be designated space to write the URL address for the client's notes. The brochure will also be posted on the web page.

Transitions for new hires

We discussed what does and does not work with regards to integrating new hires for staff into the department. Below is a summary:
  • Cole's individual meetings with new hires to introduce them to their Linux workstation received positive comments. This should be continued.
  • It was suggested that another staff member should be assigned to a new hire to help them in the orientation process -- possibly provide them templates to knitr files, be a go-to person for questions in R, etc, as it can be intimidating just walking up to a random staff member's cubicle and asking them questions.
  • It was suggested that we also have someone from the admin team sit down with each new hire to give them a brief introduction to things such as how to request travel, how to order books or other supplies, etc.
  • All new hires should receive a copy of the MCN D-2221 key.
  • Concrete suggestions on R tutorials would be helpful for those new to the language.
  • Faculty supervisors should meet early with new hires assigned to their collaborations. Setting up expectations early would be helpful.
Topic revision: r92 - 05 Jul 2016, JenniferThompson
 

This site is powered by FoswikiCopyright © 2013-2017 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback