Faculty Retreat 2015-11-13
Department of Biostatistics
2525 West End Avenue| Archive

Vision for the Department

• Where do we want to be in 5 years? What do we want to be known for?
  • Having one of the top graduate programs
  • Recruit faculty with subject matter expertise, possibly with joint appointments
  • Bring efficiency to problem solving
  • Ability to do state-of-the-art analysis of highly complex datasets (single cell imaging, brain imaging linked to EEG, high dimensional sequencing, etc.)
• We need to be known for using (AND DEVELOPING) best practices of biostatistics to improve health outcomes and in helping Vanderbilt to become a learning healthcare system.
• We need to define 10, 5-year goals that are SMART to accomplish by 1/1/2021, for example:.
  • Goal #1. At least 90% of the members of this department rate this department as a great place to work and would like to continue working here. (Have an external consultant conduct exit interviews to learn about retention issues. Also, implement the use of Stay interviews.)
  • Goal #2. Our department helped increase Vanderbilt's NIH research funding to >$400M per year. We need to each faculty member to volunteer to be a PCORI reviewer and help Vanderbilt increase the PCORI funding to >$100M Increase deparmtent profile by participating in more service roles: PCORI, NIH, reviewers for jounrals.
  • Goal #3. We demonstrated that by using modern robust statistical methods and predictive modeling techniques, Vanderbilt was able to accomplish >5 of the Pillar Goals. We take the lead on important translational research projects and decrease the amount of time that we spend giving physicians P Values for less impactful projects. We need to focus on the big questions in healthcare and biomedical research.
    • Goal #3a. Say 'no' more often. When? When the projects are trivial or investigators are poor collaborators.
  • Goal #4. Our department has >70 highly-skilled biostatisticians who are effective collaborators. How big to get?
  • Goal #5. The statistics courses taught by our faculty have student evaluations that rank in the upper quartile and we create a high-quality online degree program with a series of teaching videos to expand our teaching reach. We become much for evidence-based about evaluating our own teaching, service, and impact. We integrate best practices into our training programs.
  • Goal #6. We create and use a comprehensive incentive program for raises and bonuses based on factors that are aligned with the Vanderbilt's leaders, for example participating in Studios and study sections. More than 90% of the department approves of the incentive program. [Bryan: I like the incentives, but given the great diversity of roles in the department, it will be tricky to define the right incentives.] Agree. It can be tricky but others have done this and we can also. The alternative - a department without incentives - will be problematic.
  • Goal #7. Graduates from our program are skilled at all aspects of biostatistics and are prepared to teach and work in a collaborative environment.
  • Goal #8. We create a culture of collegiality in which all members of the department are treated with respect and we ban the pejorative terms, such as "MS biostatistician and MS-level". We create a low-stress, fun, but productive, healthy work environment, with healthy food choices at events and a healthy work-life balance. We encourage and reward positive, solution-oriented, can-do attitudes. Work on using precise position titles more. Focus on using all terms respectfully. Investiagte and discuss to resolve any issues.
  • Goal #9. Each and every member of the department leads a Continuous Quality Improvement project, which factors into raises and promotions. The faculty must lead by example with 100% participation before asking any staff to lead CQI projects. A key to the success of this goal is that each person must be the project leader of their own CQI. People can work in teams but the team must have a leader responsible for the success of the project.
  • Goal #10. We provide more administrative and project management support to faculty, especially when they are submitting grants and papers. Each faculty member is assigned 0.20 FTE of administrative/project management support. Part of the $8500 scientific resource fee is used to fund this.
  • Goal #11. Grow methodological research. For example, we have 10 methods grants, an average of 20 papers per year in top statistics/biostatistics journals, and are respected across the nation not just for our high quality collaborative research, but as leaders in developing new and important statistical methods. Start with a bulletin board with recent methods publications. Graduate students are furthering this cause. Make our methodology contributions more obvious to biomedical researchers. Grow our reputation by our students.
• I imagine we all would say that we want to be good at teaching, methods research, collaboration, and computing -- but with limited bandwidth/funding, what do we want to prioritize. A good portion of this relates to individual prioritization.
• If methods research is a priority, then how do we make it a reality?
• Is teaching undergraduates in our department vision?
• How to align our vision for the department with that of the Medical Center? Both the Medical Center's vision for itself and the Medical Center's vision of our department. We need to be strong in showing leadership the value of our academic output including our graduates. Maximize the clinical and biomedical use of our methodology papers. Show the medical center that if you want to have the capability of analyzing challenging datasets you need to have talented biostatisticians here who can be heavy hitters methodologically.
• How do we maximize our impact on biomedical research?
  • In addition to how we market ourselves and whom we teach, by being true to our principles and being rigorous
• How do we want to meet new challenges in the era of big data analystics to improve our impact? What can we do to maximize our contribution to further Vanderbilt's sucess in personalized medicine?
• Find and encourage the activities that benefit both the department and the individual, e.g. what activities above also are great for building your CV?

State of the Department

• Health of budget
• Scientific resource fee
  • 1 yr goal: try to expand the conferences that are eligible while maintaining tie to collaborative needs, since collaborations fill the fund
  • What are good things to spend on early in the fiscal year, e.g. ENAR?
  • ... late in the fiscal year, e.g. computers, servers, etc.?
  • Can the resource fee help buy more ACRE time?
  • Can we use it to buy our own gateways at ~$20-25K each? Can grad program funds be found?
• Faculty external funding performance
• Summary of 2015 SOM faculty survey for Biostatistics
• Summary of 2015 custom department survey and publication output

Institutional Academic Support for Faculty

• How do we formulate % effort to cover with and without teaching commitments?
• Coming back to this to discuss specifics.

Collaboration

Policies

• What is the minimum % effort for a senior biostatistician to accept on a grant?
• What is the minimum % effort for the junior biostatistician?
• How to limit risks to grant non-participation
  • • one way is to offer to review the grant but not be included in the proposal * How to get the right percent efforts and to pay attention to changing needs for different years of the grant? * Can accept lower % effort if the same statisticians are working on a group of related grants whose efforts total an adequate amount * Need an accepted department policy that statisticians can reference to PIs * One policy is to require a (enforced) minimum of 10% if you are the sole statistician on the grant * can be appropriate if the grant covers very adequate computer systems analyst support * But recognize that majority of grants need a statistician team * For grants unable to support sufficient effort, what about putting a junior biostatistician on the grant by herself and letting the PI take the risk that the study section won't like the absence of a senior biostatistician?
• Working group to develop a policy: Jonathan, Day A, Jeffrey B, Dan B, Amber, Robert G
  • Those who want more help asking for a higher percent effort need tools for that, e.g. an official policy that they have to formally ask for an exception for a percent effort of <10% and a senior:junior ratio <2.
    • Who reviews the budget to catch this if it gets past the senior statistician? * Add these question to Linda Wilson's monthly email. She can ask, "Please indicate 1) if the percent efforts are accurate, 2) if any of the projects have a percent effort that is too small for the effort required and what effort is needed, and 3) if any of the projects have percent effort funding available that is not being utilized." * She essentially is asking what efforts are okay, what efforts are not okay, and what's wrong with them. She can then push those concerns to the Allocations Committee to help. * The allocations committee would need to be able to help with Biostatistician I through Professor allocations.
  • A standing item on the monthly faculty meeting and executive committee meeting should be "Percent effort concerns: undermet needs, untapped funding, and grants in development". This would create a steady flow of information and a constant reminder of our need to steadily push for greater efforts. It would also create an opportunity for two faculty to work something out without the allocations committee being the mediator of every step.
  • Have rubrics for negotiating times, e.g. start at 20% senior and 50% junior and negotiate up or down from there. Groups will need rubrics for the types of grants they see. The initial values shouldn't be ridiculuous, rather they should be a good default and the impetus to discuss the work that needs doing and who would be doing it. For example, who will be pulling the data? Do you have programmer support for that? Who will be cleaning the data? Do they have experience? What tools do they use? Oh, your RA has 10 years of experience in data management and is an expert in REDCap - great, we won't need much effort on our end for that. Oh, you have a new RA who knows MS Excel - we're going to need more junior percent effort because the RA is going to need help.

Collaboration Plan

• What is working?
• What is not working?
• Should the plan evolve?
• How can we quantify the long-term payoff to other departments/divisions for taking part in the collaboration plan?

Possible Funding Options

• Recruit ahead: one PhD faculty member, one staff biostatistician, instant response to NOGA
• Have institutional academic support for faculty that has the right target average but will
  • allow protected time to shrink below the target upon NOGA
  • allow protected time to grow to above the target as grants end or are transferred to another faculty member
• How to collect information on and respond to under-met needs, e.g. grants that 20% funding available and we're only able to take 15%?
• Need to make it clear to the SOM that constant expansion requires increasing IAS
• How to keep this out of conflict with the new collaboration plan?
• Easier to sell this to SOM if we allocate a significant portion of the IAS to research development for other departments and centers
• Separate IAS for tenure track (academics for our own discipline)

Possible New Model for VICTR Biostatistics

• Expand VICTR to merge with collaboration plan
• Require divisions/departments to fund part of PhD statistician "retainer" and all of the needed staff biostatistician support to qualify for VICTR assistance
• K award support and protocol review done by divisional or department-specific biostat team
• Studio reimbursements will have to be curtailed most likely
• Need to ensure that groups enter the new plan with the right spirit
• Need to increase outreach in other ways: clinics, workshops

Graduate Program

• Funding of faculty effort for teaching
  • Choice: more IAS for those participating in the educational mission? Or same IAS for everyone in the same faculty track and those not teaching must engage in collaborative research development? Sentiment is for option 1.
• Recognition of faculty time requirements for mentoring/heading committees
• Require or strongly encourage non-teaching faculty to mentor MS students
  • Faculty can be encouraged to be thinking of ideas they would like to have a student explore with them. Mentoring a student should ultimately be win-win. You get to explore something you are mutually interested in and wouldn't have time to do on your own.
• TIPS, Data Science, and future directions
• New classes, electives?
• Get non-teaching faculty more involved with students (academic year and summer activities).
• Increase learning clinic / workshops/Research in progress sesssions with faculty and students.
• MS of Statistical Analytics / Data Science with an emphasis in statistics
  • Call a "Masters of Statistical Analytics and Data Science" ?
  • Is a Masters in Statistical Analytics in our department vision?
  • Is our department vision exclusive to analysis or do we want it to include a large data management component?
  • How do we keep data scientists from marginalizing the role of our profession?
  • The money follows the data. Can we have more expertise in getting and housing data?
• Fourth year medical students

Other Educational Programs

• MPH and MSCI
• Intro service courses

Department Outreach

• Clinics
  • Increase faculty involvement (2x/mo)
  • Consider moving Monday clinic to 2525; leave 3d/w in MCN for now

Department Resources

• Office, student, and seminar space
• Alfresco Server
• Computing Servers

Department Relations

• How to create more mixing opportunities for faculty and staff?
• Seminars and seminar space

Public Relations

• How do we better market our value?

Department Long-Term Strategy

• Faculty mentoring: 2 internal (possibly one of these junior) 1 external; we now have sufficient senior faculty
  • Committees for all positions other than Prof?
• Once/year for some, twice/year for tenure track assistant professors; perhaps twice per year in first year, once thereafter; mentor responsible for informal contact outside of the scheduled meetings
• Mentee can suggest committee members
• Areas of expertise and specialization
• Faculty recruiting
  • generalists vs. specialists
• Appropriate mix and roles of 3 PhD faculty tracks, who should teach
• Do we have a target faculty size?
• Process for career development of MS biostatisticians, and criteria for Assistant in Biostatistics
• How best to be ready for PCORI and VHAN
• How do we get the SOM to understand that for us to expand to meet real needs the total cost of % effort coverage must necessarily rise?
• Increase visability by having a research day with posters

Strategy Related to VIMPH

• PH grand rounds 4x/yr
• Undergrad program
• Make VIMPH "real"; get outside silos; consider seminars, education; consider SPH