Catheter related bloodstream infection as a function of time following a collaborative intervention in 103 intensive care units in Michigan. Circles represent mean infection per quarter; thick blue line represents estimated mean rate of infection; thin red lines represent changes in observed infection rates over time within a random sample of 50 intensive care units 
12.3 Collaborative Quality Improvement in Surgery
There is a long history of successful CQI initiatives in surgery as well as in medical disciplines but the focus is somewhat different. In medicine, there exist numerous evidence-based processes, which are often compiled into consensus guidelines, designed to improve the care of patients with diabetes, asthma, acute MI and other acute and chronic illnesses [29–31]. Improving adherence to published guidelines is a primary objective of many medicine-based collaboratives. The same is not true with surgical collaboratives, as evidence-based guidelines are generally lacking in surgery . Therefore, surgical collaboratives often focus on determining the drivers of patient outcomes and identifying best practices to optimize those outcomes [13, 32].
12.3.1 Northern New England Cardiovascular Disease Study Group
The first major surgical QI collaborative, and the one on which most subsequent CQI efforts have been based, was the Northern New England Cardiovascular Disease Study Group (NNECDSG). Founded in 1987 as a response to government-mandated public reporting with coronary artery bypass graft (CABG) surgery, the NNECDSG is a voluntary consortium of clinicians, administrators and researchers that represents all hospitals performing CABG procedures in Maine, New Hampshire and Vermont. Their goal upon launch was to foster continuous improvement in the quality, safety, effectiveness and cost involved with the management of cardiovascular disease. Their approach to reducing mortality after CABG surgery provides a beautiful illustration of the way in which regional collaboration can dramatically affect patient care across multiple institutions and settings [33–35].
In 1991, the NNECDSG examined in-hospital mortality following CABG surgery in the region and found substantial variation among hospitals and surgeons that could not be explained by patient factors alone . They reported a 2.5-fold difference in adjusted mortality between the best and worst hospitals and a 4.2-fold difference among surgeons. They concluded that these differences in mortality most likely represented variation in unmeasured aspects of patient care. This important finding led to the group’s first major intervention to reduce CABG mortality across the region.
The intervention had three components. The first component was to provide continuous performance feedback to the participating centers. This step allowed for ongoing self-assessment and benchmarking at each participating center. The second part involved extensive training courses in the techniques of continuous QI for both the collaborative leadership and the general members. The third component was a series of round-robin site visits to all centers with visiting teams consisting of industrial engineers, surgeons, nurses and perfusion staff. These benchmarking visits allowed the clinical teams from each hospital to learn from each other and ultimately resulted in practice changes that involved technical and organizational aspects of patient care, as well as methods used for evaluating patients. These changes led to a 24 % reduction in in-hospital mortality in the post-intervention period, with significant improvement at all of the participating institutions and across all patient subgroups [33, 35].
From there the group launched an effort to identify the factors that led to mortality after CABG surgery through an analysis of all deaths in the region over a 2-year period. They found that low-output cardiac failure was not only the most common mode of death throughout the region but also accounted for 80 % of the difference in mortality between low-risk and high-risk surgeons . This discovery then led to an in-depth investigation into processes that could bear upon low-output failure. With further site visits and the inclusion of additional perioperative variables, the group identified four process variables that were associated with a reduced risk for mortality from low-output failure: continuation of preoperative aspirin , use of the left internal mammary artery as a bypass conduit , avoidance of anemia while on cardiopulmonary bypass , and adequacy of beta-blockade-induced heart rate control before induction of anesthesia . Individualized care protocols, based on a patient’s predicted risk for low-output failure, were instituted and over the following 3 years, the incidence of fatal low-output failure declined across the region from 1.24 to 0.72 % .
12.3.2 Surgical Care and Outcomes Assessment Program
CQI projects are no longer limited to cardiovascular surgery. The Surgical Care and Outcomes Assessment Program (SCOAP) was developed in Washington State in 2006 as a consortium of surgeons, hospital QI leaders and health services researchers focused on improving outcomes with general surgical procedures. SCOAP has since broadened its mission to other disciplines (e.g. vascular surgery and interventional radiology) and currently collects patient data from 60 of 65 hospitals that perform at least 2 colon resections per year [42, 43].
As with other CQI programs, SCOAP has a standardized clinical data collection platform that contains information on patient characteristics, process measures and procedure-specific outcomes for all patients undergoing the selected procedures at participating hospitals. A strong emphasis was placed on tracking optimal processes of care and SCOAP now tracks over 50 different process measures, some linked to evidence and others determined by consensus, as well as several measures considered to be under evaluation. Over time, these processes are then evaluated in the context of the collaborative.
These quality metrics include processes such as avoiding transfusion if the hemoglobin is ≥7, continuing beta-blocker use in the perioperative period, routine intraoperative leak testing after colorectal resection, using diagnostic imaging in patients with presumed appendicitis, employing nutritional supplements for malnourished patients scheduled for elective operations, obtaining glycemic control during colorectal operations, and using appropriate neoadjuvant therapy for patients with rectal cancer. Adherence to these surgeon-determined process measures is reinforced by means of operating room checklists, preprinted order sets, educational interventions, e-newsletters and regional meetings. Efforts to correct under-performance occur largely through education and peer support/pressure, often with peer-led interventions focusing on sharing best practices and creating behavior change around quality metrics.
Since its launch, SCOAP has registered a number of achievements. Surgeons in all participating centers now use standardized order sets and a SCOAP OR checklist that addresses several areas of under-performance . The rate of negative appendectomies has steadily declined through efforts to encourage the use of preoperative imaging among high-risk patients (Fig. 12.2) . Adverse events have declined with elective colorectal resection (Fig. 12.3), coincident with increased adherence to processes such as leak testing and glycemic control measures. Finally, when compared to non-SCOAP institutions, hospitals participating in SCOAP have significantly reduced the costs associated with appendectomy, as well as colorectal and bariatric operations (Fig. 12.4) .
Negative appendectomy rates, by calendar quarters, among hospitals participating in Washington State’s Surgical Care and Outcomes Assessment Program 
Rates of operative complications in elective colorectal operations in sites (n = 6) that eventually joined the Surgical Care and Outcomes Assessment Program 
12.3.3 Partnering with Payers – The Michigan Plan
One of the major challenges for CQI efforts is funding. There are significant costs associated with participation in CQI programs, particularly with regard to data collection, which can create a substantial financial burden, especially for smaller hospitals. At the same time, complications are very expensive and their cost is borne largely by payers . In states with at least one dominant payer, therefore, there is a strong business case to be made for payer-supported CQI programs, since even a small reduction in complications can result in substantial cost savings for the payer . This model of quality improvement has been in place in Michigan for nearly a decade.
Since 2004, following earlier success with a pilot CQI program in percutaneous coronary interventions, Blue Cross and Blue Shield of Michigan/Blue Care Network (BCBSM/BCN) has partnered with Michigan providers and hospitals to support statewide registry-based CQI programs in a number of different disciplines. The insurer currently invests over $30 million annually to fund 16 programs, which collectively encompass the care of well over 200,000 patients each year. The programs focus on clinical conditions and procedures that are common and associated with high episode costs. The targeted procedures tend to be technically complex, rapidly evolving and associated with wide variation in hospital practice and outcomes. Current collaboratives include general and vascular surgery, thoracic and cardiovascular surgery, bariatric surgery, trauma, urology, breast oncology, interventional cardiology and others .
In this Pay for Participation model, most of the costs for administering the collaborative programs are in the form of payments to hospitals, based on a fixed percentage of each hospital’s total payments from BCBSM/BCN. In 2007, annual payments to hospitals participating in at least 1 regional collaborative ranged from $11,000 to over $1 million. In exchange for these supplemental payments, hospitals are expected to submit timely, accurate data to the coordinating center and to allow regular site visits from data auditors. Each hospital is also required to send a physician champion and program coordinator to the quarterly meetings held by each collaborative and is expected to participate actively in regional quality improvement interventions .
The coordinating center for each collaborative maintains a clinical registry containing high quality clinical outcomes data, including information on patient characteristics necessary for risk adjustment, procedure-specific processes of care and relevant outcomes. The data are prospectively collected by trained abstractors using standardized definitions and are externally audited annually to ensure accuracy and completeness. Hospitals and surgeons are provided with timely feedback on their performance, benchmarked against the other providers in the collaborative. That performance data is not publicly reported and is not released to the payer. Rather, these data are used to drive QI initiatives that are implemented at all participating hospitals under the direction of local program coordinators. The interventions are then evaluated and discussed at quarterly meetings and then further refined .
Over the last several years, the Michigan CQI programs have resulted in improvements across a wide range of clinical conditions and have led to reduced costs in a number of important areas. One example is the Michigan Surgery Quality Collaborative (MSQC), the largest of the programs and one that focuses on general and vascular surgery procedures. Given the broad range of procedures, the QI activities of the MSQC tend to focus on aspects of perioperative care, including specific practices designed to reduce venous thromboembolism and surgical site infections. In a study designed to evaluate the added value of the CQI model, hospitals participating in the MSQC were compared to non-Michigan hospitals participating in the American College of Surgeons’ National Surgical Quality Improvement Program (NSQIP). In the period between 2005 and 2009, risk-adjusted complication rates at MSQC hospitals fell from 13.1 to 10.5 % (p < 0.001), while the complication rate at non-Michigan NSQIP hospitals remained relatively flat between 2005 and 2008 with a modest decline in 2009 (Fig. 12.5). The 2.6 % decline in morbidity observed in the MSQC hospitals represents approximately 2,500 fewer patients with surgical complications annually, or an annual cost savings of roughly $20 million, far more than the $5 million annual cost of administering the MSQC .
Risk-adjusted morbidity with general and vascular surgery: Hospitals in Michigan versus hospitals outside of Michigan, 2005–2009 (Source: Michigan Surgical Quality Collaborative and National Surgical Quality Improvement Program registries, 2005–2009 )
In some instances, the cost savings have come not only from reducing rates of adverse outcomes but also from reducing unnecessary procedures. The Michigan Bariatric Surgery Collaborative (MBSC) launched in 2006 and now collects data on over 95 % of patients undergoing bariatric surgery in the state. As they began to collect data, it became apparent that almost 10 % of patients undergoing gastric bypass in Michigan hospitals had IVC filters placed preoperatively to prevent postoperative venous thromboembolism. IVC filter use varied widely, from 0 to 35 % across the 20 hospitals participating in the MBSC at that time.
Analysis of data from the MBSC revealed that IVC filter use did not reduce risk for VTE or other complications and was itself a source of complications, such as filter migration . Feedback of this data prompted a QI initiative leading to implementation of statewide guidelines for risk-stratified VTE prophylaxis. Within a year of implementation, IVC filter use had dropped to less than 2 % of patients. Given an estimated average cost for IVC filter placement of $13,000, the decline in this unnecessary procedure resulted in an estimated annual cost savings of $4 million. At the same time, implementation of VTE guidelines was associated with a decline in VTE-related mortality. Furthermore, between 2007 and 2009, 30-day mortality rates after bariatric surgery in Michigan hospitals declined at a faster rate than in non-Michigan hospitals participating in the NSQIP program (p = .045) (Fig. 12.6) .
Thirty-day mortality after bariatric surgery: Hospitals in Michigan versus hospitals outside of Michigan, 2007–2009 (Source: Michigan Bariatric Surgery Collaborative and National Surgical Quality Improvement Program registries, 2007–2009 )
Quality improvement interventions resulting in improved care and/or reduced cost have occurred across the spectrum of the Michigan collaboratives. An intervention focused around implementation of practice guidelines and the use of bedside tools for risk assessment with percutaneous coronary interventions led to reductions in contrast-induced nephropathy, transfusions, strokes and vascular complications associated with this procedure . Furthermore, a series of specific, focused interventions in cardiac surgery resulted in an increase in the use of internal mammary grafts for coronary artery bypass and a reduction in the use of two expensive therapies: intra-aortic balloon pumps and prolonged mechanical ventilation . Finally, after implementation of a QI initiative centered on comparative performance feedback and dissemination of practice guidelines, Michigan urologists improved adherence with recommended staging practices, resulting in a decline in the use of expensive bone and computerized tomography scans for surveillance of prostate cancer in low and intermediate risk tumors .
As these examples illustrate, partnerships between surgeons, hospitals and payers can be a win for all. Payers have better access to capital resources and can provide infrastructure to support collaboration among surgeons and hospitals, as well as the resources necessary for interventions involving large numbers of patients and hospitals. Furthermore, payers may have the political influence with hospitals to ensure broad participation in CQI efforts [13, 51]. The payers then see returns in terms of a reduction in costs from adverse events and unnecessary tests and procedures. Hospitals receive compensation for their participation, as well as the assurance that their outcomes will not be publicly reported. Surgeons and other health care providers benefit from the professional satisfaction that comes from collaborative learning and interaction with their colleagues. Most important of all, patients receive better care.
12.4 Challenges to Regional Collaborative Quality Improvement
CQI efforts face a number of challenges. Regional centers often compete with one another and this can pose a problem when trying to create the sense of community that is required in order to build a successful collaborative. The process of engagement and development of mutual trust can take time, considerable effort and strong, effective clinical leadership. Furthermore, the financial incentives may be poorly aligned, as the time required to attend meetings and implement QI initiatives takes away from a provider’s practice and there is often no financial compensation for these activities. Finally, all CQI efforts rely on having a clinically rich and accurate patient registry housed on one data platform. The process of data abstraction can be expensive and without a funding source, such as a state or federal agency or payer, the cost for data entry falls to the hospitals, which may have limited ability to pay for it .
Given these constraints, studies of CQI projects across healthcare settings demonstrate varying degrees of success. Some CQI reviews suggest that only about 50–60 % of participating centers ever fully implement the recommended changes or achieve the desired outcome [52, 53]. Other studies have shown that even among centers that successfully implement recommended changes, only about 60 % will exhibit sustained improvements over time . While it is not always known where the breakdown occurs, the variability in success with QI initiatives often reflects differences in the degree of implementation, a measure influenced by local leadership, attitudes, culture, commitment and resources .
12.5 Keys to Success with Collaborative Quality Improvement
Although the hallmark of QI collaboratives is a bottom-up, participatory approach to decision making, a major contributor to success with CQI efforts is strong central leadership. The collaboratives represent a clinical community, whose members are often drawn from diverse professional backgrounds. It is the leader’s responsibility to ensure the cohesiveness of the community and coordinate the group’s efforts . The leader must also be able to interpret the collected data and use it to identify targets for improvement, while ensuring that the chosen interventions are based on current best evidence [25, 54]. And she must be able to communicate a clear, uniting vision of where the collaborative is headed and what can be achieved by individual interventions .