The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP^®) is the only nationally benchmarked, clinical, risk-adjusted, outcomes based program in the USA that is designed to measure and improve care across the surgical specialties [3, 11]. ACS NSQIP is a nationally benchmarked, peer-controlled database that allows hospitals to compare 30-day patient outcomes to hospitals of all sizes and types across the country. ACS NSQIP uses data that are:

The STS National Database was established in 1989 as an initiative to enhance the quality and safety of cardiothoracic surgery and to provide an accurate and valid basis for measuring performance in our specialty [4, 12, 13]. The STS National Database has thus far had five chairs: Richard E. Clark (1989–1997), Frederick L. Grover (1997–2004), Fred H. Edwards (2004–2010), David M. Shahian (2010–2015), and Jeffrey P. Jacobs (2015– ). The STS National Database has three major component databases, each focusing on a different area of cardiothoracic surgery: the STS Adult Cardiac Surgery Database (ACSD) , the STS Congenital Heart Surgery Database (CHSD) , and the STS General Thoracic Surgery Database (GTSD) (Fig. 44.5) [4, 12, 13]. Table 44.1 documents the size and penetration of the three major component databases of the STS National Database. STS-ACDS is the largest adult cardiac surgical database in the world and contains data from over 90 % of the hospitals that perform adult cardiac surgery in the USA. STS-CHSD is the largest pediatric cardiac surgical database in the world and contains data from over 95 % of the hospitals that perform pediatric cardiac surgery in the USA. STS-GTSD is the largest clinical registry of general thoracic operations in the world. All three component database of STS National Database function as platforms for outcomes analysis, quality improvement, and research.

The first step in creating a surgical registry is developing a standardized nomenclature so that all diagnoses and procedures are coded uniformly across centers. Ample data exists demonstrating the limitations of administrative systems of nomenclature that were designed for billing and not for the analysis of outcomes [14–18]. A universal clinical system of nomenclature is the foundation of any surgical registry.

Once a system of nomenclature is established, the next step is creating a platform of data collection with a shared minimal dataset and standardized definitions for fields of data.

In ACS NSQIP [3], each hospital assigns a trained Surgical Clinical Reviewer (SCR) to collect preoperative data through 30-day postoperative data on randomly assigned patients. The number and types of variables collected differs from hospital to hospital, depending on the size of the hospital and the population of its patients, and its quality improvement focus. The ACS provides SCR training, ongoing educational opportunities, and auditing, to ensure data reliability. Data are entered online in a HIPAA-compliant, secure, Web-based platform that can be accessed 24 h a day. A surgeon champion assigned by each hospital leads and oversees program implementation and quality initiatives. Blinded, risk-adjusted information is shared with all hospitals, allowing them to nationally benchmark their rates of complications and surgical outcomes. ACS also provides monthly conference calls, best practice guidelines, and many other resources to help hospitals target problem areas and improve surgical outcomes.

In each of the three STS National Databases [4], data are collected regarding patient demographics, preoperative factors that may impact the outcomes of surgery, details of the specific disease process that led to the surgery (e.g., degree of coronary artery stenosis in each vessel [19], etiology and severity of valvar lesions, type of thoracic aortic pathology, stage of lung cancer, or esophageal cancer, type of congenital cardiac lesion); technical details of the conduct of the operation that was performed; detailed clinical outcomes; and disposition of the patient (e.g., home, rehabilitation facility, or dead). Data from the STS National Database are reported back to participants in Feedback Reports that include the types of procedures performed; demographics and risk factors of the patients; details about the conduct of the surgical procedure; and outcomes. In each database, individual institutional outcomes are benchmarked against aggregate data from all programs in the given database. Data in each of the STS National Database are either entered by a trained abstractor (database managers) or entered by caregivers and carefully reviewed by the database manager. These database managers work with surgeons, physician assistants, nurse practitioners, and others to ensure that that data entered into the STS National Database adhere to the definitions established by STS and that they are supported by documentation in the patient’s medical record. These data managers have many resources available to them including:

Standardization of definitions of all fields in the database is essential [19]. For example, Operative Mortality is defined in all STS databases as (1) all deaths, regardless of cause, occurring during the hospitalization in which the operation was performed, even if after 30 days (including patients transferred to other acute care facilities); and (2) all deaths, regardless of cause, occurring after discharge from the hospital, but before the end of the 30th postoperative day [20, 21].

After standardizing nomenclature and establishing a database with defined fields of data, the next step is the incorporation of a mechanism to evaluate and account for case complexity. Case mix can vary between surgeons and hospitals. Risk adjustment is essential when assessing and comparing healthcare performance among programs and surgeons, as this adjusts for differences in the complexity and severity of patients they treat. Reliably accounting for the risk of adverse outcomes mitigates the possibility that providers caring for sicker patients will be unfairly penalized because their unadjusted results may be worse simply because of case mix. A variety of strategies exist to adjust for variations in case mix [22]. Risk models can adjust for variations in the preoperative status of patients and the overall case mix of a given provider.

Three fundamental issues in health care performance measurement must be addressed when comparing the performance of providers and hospitals: selection of a homogeneous target population, risk adjustment, and assignment of quality rating categories [22]. Differences in provider classification may result from these methodologic decisions [22–25]. Multi-domain composite performance metrics may be utilized that combine the outcome domains of mortality and morbidity [26]; this strategy is important because of progressively decreasing mortality rates and because survival is only one measure of the quality of care. For example, consider two patients who undergo the same surgical repair of an abdominal aortic aneurysm. Patient one recovers with no complications. Patient two survives but has a postoperative stroke, develops dialysis dependent renal failure, and needs a gastrostomy because of an inability to swallow after the stroke. These two patients will both count as survivors in a model that only measures mortality; however, a multi-domain composite that includes postoperative morbidity will differentiate the outcomes of these two patents. Such composite measures provide more end points and also a much more comprehensive assessment of quality of care, because such composites include both risk-adjusted mortality and risk-adjusted morbidity.

Once one has a developed a standardized nomenclature, a core database, and a system to adjust for variations in case mix, the next step is to assure the completeness and the accuracy of the data. Three potential strategies may ultimately allow for optimal verification of data:

Data quality in all STS databases is evaluated through intrinsic data verification by DCRI (including identification and correction of missing/out of range values and inconsistencies across fields). In addition to intrinsic data verification by DCRI, each year, approximately 10 % of participants in all STS databases are randomly selected for audits of their center. The audit is designed to complement the internal quality controls, with an overall objective of maximizing the integrity of the data in all STS databases by examining the completeness and accuracy of the data. STS has selected Telligen (http://​www.​telligen.​com/​) to perform these independent, external audits. As the state of Iowa’s Medicare Quality Improvement Organization (QIO), Telligen partners with health care professionals to assure high quality, cost effective health care. As a QIO, Telligen is HIPAA compliant and performs audits adhering to strict security policies.

It is often stated that caring for surgical patients is a “team endeavor,” bringing together a variety of professionals to maximize the outcomes [27, 28]. The harmonization of nomenclature and database standards between medical and surgical databases can enhance the science of outcomes analysis and quality improvement and benefit our patients [29]. Medical and surgical databases can be linked through a variety of strategies including linkage based on indirect identifiers using probabilistic matching [8, 30, 31] and linkage with direct identifiers using deterministic matching [8, 32, 33].

One weakness of most surgical registries is their inability to provide longitudinal outcomes. The transformation of a surgical registry into a platform for longitudinal follow-up will ultimately result in higher quality of care for all surgical patients by facilitating longitudinal comparative effectiveness research. Several potential strategies will allow longitudinal follow-up with the surgical registries, including the development of clinical longitudinal follow-up modules within the surgical registry itself, and linking the surgical registry to other clinical registries, administrative databases , and national death registries: