2016 INFORMS Annual Meeting Program

129 / 561

INFORMS Nashville – 2016
129
MA23
108-MCC
Human Resource Analytics in Anesthesia
Sponsored: Health Applications
Sponsored Session
Chair: Franklin Dexter, University of Iowa, 200 Hawkins Drive, 6-JCP,
Iowa City, IA, 52242, United States,
1 - Comparing Providers Based On An “obvious” Short-term Clinical
Outcome: Pain Scores
Jonathan Porter Wanderer, Vanderbilt University, Nashville, TN,
37204, United States,
In this presentation, we will review a “obvious” metric for comparing supervising
anesthesiologists: their patients’ pain scores on admission to the recovery room.
Using modeling, we will evaluate the appropriateness of this metric and analyze
the degree to which these scores are useful in evaluating the performance of
supervising anesthesiologists. Spoiler alert: this is not as useful a metric as it may
first appear.
2 - Comparing Anesthesia Providers Based On Discrepancies In
Their Controlled Substance Reconciliation Logs
Richard H. Epstein, University of Miami,
Accurate accounting of controlled drugs is required in the US under the
Controlled Substances Act. However, discrepancy rates of 15% between
automated drug cart transactions and administration records in the anesthesia
information management systems (AIMS) have been reported. We describe how
we reduced our discrepancy rate from 15% to 3.5% through a sequential process
involving daily, individualized email reports, a near real-time tool publicly
showing the transaction balance compared to the AIMS, and involvement of the
pharmacy to contact individuals with discrepancies. Although there were a few
extreme outliers, discrepancies were widely distributed, requiring a system based
approach.
3 - Comparing Faculty Anesthesiologists Based On Daily Evaluations
Of Quality Of Supervision By The Resident Physicians
Franklin Dexter, University of Iowa,
Since July 2013, resident physicians (and nurse practitioners) evaluate daily the
anesthesiologists in our department based on quality of supervision. The uni-
dimensional 9-item scale is valid and psychometrically dependable. Graduate
students evaluate the faculty daily. In 9 papers, we have learned: what is being
measured, achievable response rates, lack of important covariates, appropriate
Bernoulli CUSUM analysis, role of feedback, minimum perceived acceptable
scores, lack of correlation with individual productivity, lack of advantage to
faculty specialization, and improvement over time. Textual analysis of resident
comments showed behaviors associated with better scores.
4 - Privacy Policies For Protecting Human Resources
Liam O’Neill, University of North Texas,
The dominant paradigm of information privacy is based on de-identification. We
calculated the percent uniqueness for hospital surgical patients from a database of
2.8 million hospital records. Only three attributes were considered: hospital
name, fiscal quarter, and surgical procedures. For patients undergoing two or
more procedures, 64.4% of records were found to be unique. For a patient
selected at random, the probability of matching this record to state database was
42.8% (SE < 0.1%). Data sharing plans, as required by some journals and funding
agencies, may place patient and clinician privacy at risk. Compliance with HIPAA
alone is insufficient to protect health data from re-identification.
MA24
109-MCC
Organizational Structures and Innovation
Invited: Strategy Science
Invited Session
Chair: Todd Zenger, University of Utah, Eccles School of Business,
Salt Lake CIty, UT, 9, United States,
1 - Organizational Structure, Attention Allocation Andinnovation In
The Multidivisional Firm
John Joseph, University of California, Irvine, CA, 9, United States,
, Alex Wilson
We propose that in order to understand how organizations develop inventions
with greater impact, we must not only consider the formal divisionalization of the
firm, but also its attention structure. We find that unit attention which is either
highly specialized (distinct from other units) or highly integrated (similar to other
units) yields higher quality inventions. We suggest that this U-shape relationship
reflects the need for attentional coherence to product quality inventions. We test
our theory using novel organizational structure and topic modeling of patent data
from Motorola. Our findings have implications for theories of organizational
design and innovation.
2 - Stage-gate Processes And Selection In Innovation
Ronald Klingebiel, Frankfurt School of Finance and Management,
1, Frankfurt, Germany, , Peter Esser
We examine resource allocation at stage gates of the innovation process at former
handset manufacturer Sony Ericsson, illuminating project discontinuation
behavior. We find that with financials unavailable early on, escalation can be
unknowable as well as a bias. Further, the propensity to search and update
financials declines as projects near completion, accompanying known self-
justification and goal-substitution effects. Finally, projects, for which negative
information does get translated into business case updates, are more likely to be
continued than projects with stable or improving financials. This strong form of
escalation suggests a counter-intuitive effect of attention.
3 - Intra-firm Collaboration Boundaries And The Efficacy
Of Innovation
Andy Wu, Harvard Business School, Soldiers Field, Boston, MA,
02163, United States, David Hsu, Vikas Aggarwal
What are the implications of alternate approaches to organizing the inventive
human capital of a knowledge-intensive firm? We empirically examine how the
locus of technical experience diversity within a firm influences firm-level
innovation output. Using a sample of biotechnology start-ups we find that
organizing human capital so that there is greater indirect diversity (i.e., where
technical experience diversity lies outside the locus of production) yields greater
firm-level innovation benefits compared to organizing human capital so that there
is greater direct diversity (i.e., where technical experience diversity lies inside
production units themselves).
4 - How Do Pilots Work? Examining Pilot Use In New
Practice Transfer?
Megan Lawrence, Vanderbilt University, Owen Graduate School of
Management, Nashville, TN, 3, United States,
This paper examines how the use of pilots influences new practice
implementation. Using data from a Fortune 100 retail chain that implemented a
new restocking process, I compare implementation performance in 280 stores. I
propose and test that pilots enable learning from outside experiences. I find that
stores only learn from the pilot store in their own district, even when closer to
another district’s pilot. Stores’ prior performances relative to their pilot influence
the extent to which they learn from their own experiences. Lastly, I provide
evidence that the most likely alternative mechanism - peer effects - may also be
present but that these effects do not eliminate the learning findings.
MA25
110A-MCC
Modern Project Management
Invited: Project Management and Scheduling
Invited Session
Chair: Nicholas G Hall, Ohio State University, Columbus, OH,
United States,
1 - Opportunities For Behavioral Research In Project Management
Enno Siemsen, University of Wisconsin,
Behavioral research in operations management is on the rise. The field of Project
Management seems ripe for more applications of behavioral research. How do
team members react to deadlines and reminders? How are projects selected and
abandoned? What is the impact of multi-tasking, and how does assigning new
resources change team dynamics? These are all important research topics, and I
will both highlight past and current work in these areas, as well as opportunities
for future research.
2 - Information Protection In New Product Development Projects
Xiaoqin Wen, Shanghai University,
Nicholas G Hall
Motivated by the threat of industrial espionage during new product development,
a project company minimizes information leakage by using decoy work. The
competitor validates the information it observes in a cost-effective way, based on
three alternative prior assumptions about its validity. We model this problem as a
two stage Stackelberg game, identify an equilibrium solution, and obtain
managerial insights. Coordinated setting of the project deadline and budget is
needed to protect the project. Counterintuitively, it benefits the project company
to announce that it uses decoy work.
3 - Research Challenges In Project Management
Ted Klastorin, University of Washington,
A great deal of previous project management research has focused on scheduling
problems. But the research issues faced by project managers are far more complex
and just as challenging as those faced by supply chain managers. In this talk, I will
explore many of these issues and directions for future research studies, with an
emphasis on problems associated with decentralized projects.
MA25