Tuesday, March 30, 2010

Clinical Research Oriented Workshop (CROW) Meeting: March 26, 20010

Clinical Research Oriented Workshop (CROW) Meeting: Mar 26, 2010

Present: Matt Bovee, Connie van Eeghen

1. Start up: Book club – “The Checklist Manifesto”: based on lack of attendance, we decided to post-pone a “close out” conversation on our book, thinking that it would be more productive with more people. Key questions to lead us down the “so what” and “where do we go from here” path:
a. What is the book’s value/take-away according to CROW and individual members
b. What can or should be done based on the book’s message(s)? How? By whom? For what purpose?

2. Meeting Schedule: Connie to confirm with faculty whether meeting time (Fridays at 9:30) is a constraint on attendance. To be reviewed next meeting.

3. Roundtable, with update, short term goals, long term goals

a. Connie: Data advice on her small sample, pre-project survey data results for Case 1:
i. Data hygiene: critiquing her Excel spreadsheet data collection tool
1. Make the actual questions (verbatim) accessible for data interpretation
2. Note that data validation should check not only for alpha-numeric accuracy, but also for differences that Excel is blind to, e.g. Excel does not distinguish between the entry “0” and a blank (unfilled) cell. Since 0 is a legitimate value, and a blank represents a response that is left blank, validation should check to make sure that a 0 wasn’t entered for a blank and vice versa. (Matt came up with an impressive formula for this:
but I’m not explaining it here.)
ii. Data analysis: what statistics are relevant and how should they be executed
1. Currently planning to use simple descriptive statistics: sample size, mean, median, maximum, and minimum
2. Will consider using tertiles, quartiles, or quintiles for the collected sets of data across all pre-project and post-project surveys received on all cases (up to six cases planned)
iii. How to organize the post-project data for Case 1
1. Create a single Excel file for all of Case 1 data, pre and post. Data calculations are not expected to exceed the capacity of the Excel program
iv. How to organize the pre- and post-data for all other cases, so as to analyze the “pre” as a group and the “post” as a group
1. Extract sheets from each Case specific file and create separate Pre and Post files
v. How to analyze the collected “pre” and “post” data sets against each other
1. Consider using Odds Ratio to evaluate the responses to the independent variables vs. the independent variables
vi. Thank you Matt!

4. Next Fellows Meeting(s): April 2, 2010 from 9:30 – 11:00 a.m., at Given Courtyard Level 4
a. April 2: Bookclub (how does this apply to my work?) How to predict medical events effectively OR Mapping new NHANES data with mortality (Ben) – or, Connie’s data update
b. April 9: (no Connie or Matt)
c. April 16: (no Connie or Matt)
d. Future agenda to consider:
i. Future: Review of different types of journal articles (lit review, case study, original article, letter to editor…), when each is appropriate, tips on planning/writing (Abby)
ii. Future: Informed consent QI: Connie to follow up with Nancy Stalnaker, Alan Rubin will follow up with Alan Wortheimer or Rob McCauly

5. Fellows document – nothing this time

Recorder: Connie van Eeghen

1 comment:

  1. I circulated a spreadsheet w/a 'toy' example of the "blank = zero" problem in Excel and a formulaic solution. If you didn't get a copy, contact me.

    The workbook is suitable (w/correction of one typo!) for demo/explanation and contain brief description of the working formulas and why they don't or do work as the case may be. The formulas are adaptable to other spreadsheet workbooks used for comparing double data for validation. The logic is a bit tortuous; if I've missed an option setting in Excel 2003 or 2007 that someone knows about, PLEASE tell me. It would sure make life a lot easier for many to do that instead.


Note: Only a member of this blog may post a comment.