September 15, 2026
Page 1
Multicenter Perioperative Outcomes Group (MPOG) and Anesthesiology Performance
Improvement and Reporting Exchange (ASPIRE) Protocol
Project Summary
The Multicenter Perioperative Outcomes Group (MPOG) is a consortium of anesthesiology departments
from academic medical centers, community hospitals and ambulatory surgical centers with electronic
perioperative information systems. The purpose of MPOG is to allow multi-institutional collaboration
with the goal of accelerating quality improvement and outcomes research in perioperative medicine.
The Anesthesiology Performance Improvement and Reporting Exchange (ASPIRE) is the quality arm of
MPOG and is focused on using data to assess variation in practice, identify local/regional best practices,
measure process adherence and patient outcomes, create programs for quality improvement, and
enable collaboration among anesthesiologists, surgeons, CRNAs, and other allied health professionals.
ASPIRE will also develop research topics which will lead to quality assurance research projects; and
conversely, MPOG research will study the impact of quality improvement measures as incentives to
practice change, and associated outcomes.
MPOG was developed so that institutions across the globe can work together to pool their electronic
perioperative data into a common research database. The MPOG Coordinating Center, housed at the
University of Michigan, receives and merges the limited dataset into one centralized data repository.
These limited datasets (only date-of-service will be uploaded into the repository) are used for clinical
outcomes and quality assurance research purposes by the members of participating institutions. The
database will also include administrative information and outcomes data.
This protocol describes University of Michigan’s role as a performance site and Coordinating Center for
MPOG. As a performance site, the University of Michigan uploads a limited dataset from our anesthesia
electronic information systems to the MPOG Coordinating Center repository (HUM00024166). MPOG
has a Perioperative Clinical Research Committee (PCRC) which is comprised of members of active MPOG
institutions who are actively contributing data. The PCRC serves as the publication committee of MPOG
responsible for reviewing, refining, and modifying any research proposals and manuscripts created by
researchers at active/contributing institutions. All proposed research studies using data from the central
MPOG database must pass a peer-review process by the PCRC prior to submission for publication.
Patients included will be those who undergo surgical or procedural interventions requiring anesthesia
care, across all age groups and all medical conditions. There are no exclusions. All data are scrubbed for
any identifiable information prior to sending to MPOG central repository for merging into a MPOG
database. There are no patient identifiers stored in the MPOG central repository, and no members of
the research team will ever have access to identifiers. Automated database extraction process is
performed on secure UMHS servers. The only PHI element collected will be date-of-service.
September 15, 2026
Page 2
Background
The traditional prospective, randomized controlled trial (RCT) is a mainstay of medical research.
However, although RCTs are regarded as the gold standard of medical research, there are many clinical
conundrums that cannot be effectively addressed by a prospective RCT. Observational studies are often
the only option for research into
- Infrequent adverse events associated with a medication
- Emergency operations challenging conventional consent and randomization systems
- Generic medications without a reliable pharmaceutical funding source
- Safety studies of medical procedures
- Ethnic and racial groups that are not appropriately represented in prospective RCTs
- Rare perioperative or peri-procedural events
- Effectiveness and safety of a treatment once applied to a broad patient population as opposed
to specifically targeted group in an RCT
For these and many other clinical questions, the only practical research structure is a large observational
dataset based upon existing information system data elements. By combining point-of-care clinical
information systems with clinical registries, financial data, laboratory information systems, radiology
information systems, administrative datasets, and scheduling datasets, one can address previously
unanswerable clinical questions. Though limitations to causal inference exist, these hypothesis
generating studies are crucial to the advancement of clinical science.
Specifically, the field of perioperative medicine has witnessed a major improvement in safety over the
last few decades. As a result, many of the morbidity endpoints (death, myocardial infarction, pulmonary
complications, nerve injury) previously followed are now occurring with a frequency far too low (< 1%)
for prospective enrollment in controlled trials. However, the importance of these morbidity endpoints
cannot be diminished for the patients and families experiencing them. Alternative study techniques
must be used to continue the necessary improvements in patient safety and satisfaction.
Recent literature has demonstrated the value of large dataset research in altering clinical practice. The
widespread use of beta blockers for moderate risk patients and the routine use of aprotinin for cardiac
surgery have both been called into question as a result of large clinical dataset research. These landmark
hypothesis generating studies then led to large RCT that eventually identified an unacceptable risk
profile for certain patients, leading to major changes in the perioperative use of these drugs.
Objective
To evaluate variation in perioperative anesthesia care and its association with patient outcomes,
adverse events, and resource utilization using a large, multicenter clinical data repository. These
analyses support the development and implementation of evidence-based quality measures and enable
continuous quality improvement through provider feedback and benchmarking.
To identify patient health outcomes, and changes in resource utilization associated with variation in
anesthesia care. Create quality measures focusing on quality improvement, quality-focused research,
provider feedback reports, and adherence to standard quality metrics.
September 15, 2026
Page 3
Specific Aims
1. Data Aggregation and Scale
To aggregate standardized perioperative data across multiple institutions to enable investigation
of rare adverse events, uncommon clinical conditions, and low-frequency procedures.
2. Collaborative Research Infrastructure
To provide participating institutions access to a large, multicenter, de-identified dataset that
supports observational research and quality improvement initiatives in perioperative care.
3. Integration of Granular Clinical Data with Outcomes
To integrate detailed intraoperative physiologic data and anesthetic interventions with
longitudinal clinical, administrative, and financial outcomes to evaluate associations between
care processes and patient outcomes.
1. To aggregate patient data enabling investigation of infrequent adverse events, patient
conditions, and operations.
2. To allow our institution to enable access to a large, international limited dataset necessary for
observational and quality research
3. To combine detailed physiologic data and anesthesiology interventions with long term
outcomes recorded in surgical outcomes, administrative, and financial databases
Investigator Expertise
Dr. Mathis is an Associate Professor in the Department of Anesthesiology with expertise in perioperative
outcomes research using large observational health databases. He is a practicing cardiac
anesthesiologist who has board certification in clinical informatics and has served in MPOG research
leadership roles for over 10 years. His research has been supported through large-scale federally funded
grants, integrating anesthesia and surgical registry data to understand anesthesia practice variation and
associated outcomes. Under his leadership, he has guided and collaborated with MPOG investigators
who have generated over 100 research proposals, the majority of which have led to publications in high-
impact clinical journals.
Dr. Kheterpal is the chair of the Department of Anesthesiology and the Robert B. Sweet Endowed
Professor of Anesthesiology and associate dean for research information technology with expertise in
perioperative clinical outcomes research. He has served as a systems designer, database architect, and
data warehouse architect for over 15 years, creating a deep information technology and clinical
informatics knowledge base. He has previously used retrospective analysis of clinical documentation
databases to publish studies in leading peer-reviewed anesthesiology journals. In addition, he has served
as a representative to the anesthesia patient safety foundation.
Dr. Shah is a Clinical Professor and Associate Chair for Technology for the Department of Anesthesiology
with expertise in medical informatics and quality. He is the Director of Informatics and Systems
Integration for the Department of Anesthesiology and the University of Michigan Health System. He has
served as a systems designer, database architect both in industry and at academic institution for over 10
years. He brings a technical and medical background that will help to integrate the software and analytic
system.
September 15, 2026
Page 4
Inclusion/Exclusion Criteria
As a participating site, all adult and pediatric patients undergoing procedures requiring anesthesia care
at Michigan Medicine are included. There are no exclusion criteria.
Methodology1
The perioperative clinical information system, several UMHS surgical outcome databases, laboratory
data, administrative data, and financial data are merged together by data analysts at our institution. The
original data already exists in clinical or administrative information systems and was collected by clinical
providers and administrative processes as part of routine care. The clinical care delivered and
documentation will remain unchanged. The standard data extract and submission do not involve a
review of individual patient records for additional data items. Perioperative and peri-procedural data
include physiologic, medications, preoperative, intraoperative, and postoperative elements captured
from approximately four hours before anesthesia start through six hours after anesthesia end.
Data Acquisition and Standardization
After anesthetic case data are extracted from the source system, data are integrated with other data
sources including institutional research repositories, case data that may also be available from outcome
registries (such as an extract of data captured by that site as part of participation in the National Surgical
Quality Improvement Project [NSQIP], Society of Thoracic Surgeons General Thoracic Surgery Database
[STS-GTSD], Society of Thoracic Surgeons Adult Cardiac Surgical Database [STS-ACSD], Michigan Surgical
Quality Collaborative [MSQC]), and other clinical, laboratory, or administrative systems. Data are
matched at the participating site based on locally held unique identifiers (such as Medical Record
Number or Social Security Number). The unique identifiers are removed before transmission to a
centralized database at the Coordinating Center.
Once extracted from the local electronic health record (EHR), perioperative data are mapped to MPOG-
developed standardized, semantically interoperable concepts before submission to the central
repository. MPOG embraces standardized definitions where available, such as the use of ICD-10
diagnosis codes or Association of Anesthesia Clinical Directors (AACD) anesthesia events, but these are
supplemented by an MPOG-specific set of data elements.
Data Validation
Once the mapping process is completed—and before centralized database transmission—data from
participating sites are assessed for completeness and accuracy. Our Data Diagnostics tool facilitates the
assessment—identifying specific deficiencies across data category, institution, and time domains. MPOG
requires a clinically trained site representative to review and attest to data accuracy before each data
transmission to the central repository. At the Coordinating Center, MPOG research and quality
improvement leadership review the initial data upload (including the Data Diagnostics information)
before it is integrated into the main MPOG database. Participating MPOG sites perform a manual review
of a random sample of cases recorded within the local database before transmission to the centralized
database, because some errors may escape detection when assessed at an aggregate level.
In rare situations, research investigators using the centralized or performance site database may
observe a rare clinical event that requires additional data extraction. Every case in the centralized
September 15, 2026
Page 5
repository and at each performance site will have a distinct patient system number (that is randomly
generated) that is NOT a patient identifier. This system number is NOT related to or derived from any
PHI (ie, name, reg num, date of birth, etc). The centralized MPOG repository has no way of using this
random system number to link to any patient identifiers. If a rare clinical event that requires additional
data extraction is observed, this non-PHI system number will be provided to a performance site.
Technical staff at the performance site can access their own source databases in an attempt to link the
system number to PHI. This linkage exists in the source clinical information system, not in any research
database proposed by this project. If they choose to, the performance site may use this system number
to extract and provide additional de-identified clinical data to the Coordinating Center investigators. No
patient contact will occur. All additional data extracted would be from existing clinical and
administrative data sources.
Removal of Identifiers and Data Transmission
A limited data set is first created locally by removing selected PHI via a customized “scrubbing” tool
(leaving only dates and extremes of age) and then transmitted to a centralized MPOG database. The
scrubbing tool additionally removes common names that may be entered in the free text. Several
dictionaries are preloaded into the scrubbing application including the most common first and last
names from the US Census Bureau and the Systematized Nomenclature of Medicine (SNOMED)
dictionary to identify health care terminology that should remain with the transfer. Sites may add
additional information to be scrubbed such as names, initials, or internal identifiers assigned to
providers. All text is examined and passed through the scrubbing utility before upload.
Only after completion of validation procedures and the use of the scrubbing tool does the option of
transferring case-level data to the MPOG Coordinating Center become available. Data are transferred
into an encrypted repository, checked for validity, and integrated into the MPOG Coordinating Center
database. A database table containing patient identifiers and unique case-linking information remains
stored at the local site and is not transmitted to the MPOG Coordinating Center.
Automated Handling
Once data are transmitted and integrated into the MPOG Coordinating Center database, the data are
available for use within research and quality improvement projects. As specific to the needs of a project,
data are subject to focused examination to ensure appropriate values are included.
Data Use
All research projects using MPOG data sets must obtain project-specific IRB approval. Additionally, a
detailed proposal must be presented through the monthly MPOG Perioperative Research Committee
(PCRC), comprised of MPOG active site principal investigators, site chairs/heads of practice, statisticians,
and other interested research faculty with full appointments at active MPOG sites. The committee
critically reviews and amends the proposal, and subsequently votes to accept, require revisions, or
reject the proposal. Before accessing data, research project proposals are prospectively registered and
tracked on the MPOG website which remains accessible to members.
Quality improvement activities are governed by the MPOG Quality Committee, composed of active site
Quality champions and quality experts.
September 15, 2026
Page 6
Figure 1 provides an overview of the flow of information through the MPOG consortium, outlining the
process of data acquisition at the point of care, importing of data into local and central data registries,
and finally, curation of data for research and quality improvement measures.
Statistical Design
A variety of statistical techniques often used for large dataset research will be used. Each approved
research question may use distinct techniques, so it is beyond the scope of this document to
prospectively delineate a complete statistical plan. Logistic regression modeling (GEE and standard),
propensity score matching, cox proportional hazard modeling, and basic univariate comparisons are a
few of the many techniques that will be used. Descriptive and inferential statistical techniques will be
employed.
Risks
Clinical/Physical/psychological/social/reputation/financial Risk (No more than minimal risk)
There will be no care interventions, no process changes, no documentation changes, and no alterations
to a patient's clinical experience. Providers will not experience any changes in their roles,
responsibilities, or care. A limited dataset will be extracted months AFTER the clinical care episode is
complete. MPOG data are stored and analyzed within dedicated research environments, including
development and production research databases, which are separate from clinical production systems.
No changes to patient care systems or application performance will occur.
Privacy Risk (Rarely likelihood of risk)
Theoretically, since patient data is being extracted, there exists a non-zero privacy risk. However, since
the data extraction process is automated and the data structures involved separate patient identifiers
from the data being extracted, the likelihood is extremely rare. Patient identifiers (DOB, Name, MRN,
Insurance account numbers, SSN) are NOT stored or transmitted at any point during the data extraction
or transmittal process.
Data Security and Privacy
All data extraction processes are automated to eliminate the possibility of human error. No research
personnel manually review, match, or manipulate patient identifiers during the data extraction or
transmission process. Source data are extracted from existing clinical and administrative information
systems as part of routine operations and are processed using validated, automated workflows.
All database work is performed on University of Michigan Health System-approved, secured server
environments that comply with institutional information security policies. These systems are managed
within the University of Michigan computing infrastructure and are protected by role-based access
controls and multi-factor authentication. Patient identifiers are never stored on, downloaded to, or
transported via portable computing devices (e.g., laptops, USB drives).
Data is encrypted using standard relational database management system storage techniques.
Participating institutions submit limited datasets to the MPOG coordinating center on a regular basis,
following local validation. These data will not contain any patient identifiers. The only PHI included will
be date of service / surgery. The transmission will be using a secure-socket-layer / transport-layer-
September 15, 2026
Page 7
security encryption to ensure that data 'eavesdropping' is avoided during transmission of the data to the
Coordinating Center.
Access to identifiers is restricted to automated database processes operating within the secure source
systems. No members of the statistical teams, research investigators, manuscript authors, or MPOG
Coordinating Center personnel have access to patient identifiers. All identifiers are removed prior to any
data being made available for research or quality improvement use.
An internal perioperative clinical information system number for each operation that is completely
unrelated to the patient medical record number or name remains in the data extract. This system
number cannot be used to ascertain any PHI regarding the patient unless the perioperative clinical
system database is accessed by a database specialist. In rare cases, additional info about a patient may
be requested. In that case, a separate IRB application will be submitted to link the internal system
number to the perioperative clinical information system.
Figure 1.
Colquhoun DA, Shanks AM, Kapeles SR, et al. Anesth Analg. 2020.
September 15, 2026
Page 8
References
1. Colquhoun DA, Shanks AM, Kapeles SR, Shah N, Saager L, Vaughn MT, Buehler K, Burns ML, Tremper
KK, Freundlich RE, Aziz M, Kheterpal S and Mathis MR. Considerations for Integration of
Perioperative Electronic Health Records Across Institutions for Research and Quality Improvement:
The Approach Taken by the Multicenter Perioperative Outcomes Group. Anesth Analg.
2020;130:1133-1146.
