This approach offers the advantage of dealing, at least in part, with fairly concrete and accessible information. Programs run by organizations such as the ACS Committee on Trauma (COT) focus on structures. For example, the ACS COT defines and publishes Optimal Resources for the Care of the Injured Patient [7], a manual which serves as the basis for standardized review of the resource availability (and processes) of trauma centers. The review of structure has the major limitation that the relationship between structure and process or structure and outcome, is often not well established.

The analysis of processes includes the surgeon’s processes in delivering care – such as whether “good” care has been delivered. As Donabedian [6] enumerated, judgments may be based on considerations such as the appropriateness, completeness or redundancy of information obtained through clinical history, physical examination and diagnostic testing; justification of diagnosis and treatment plans; technical competence in the performance of diagnostic and therapeutic procedures; application of preventive care; coordination and continuity of care; and acceptability of care to the patient. The SCIP, which has focused on adherence to process measures as a means of reducing surgical morbidity and mortality, has mandated collection and public reporting of data for the past several years. While compliance has steadily improved over time, it has become apparent that simple adherence to the prescribed measures does not necessarily result in improved outcomes [8]. As Merkow and colleagues [9] note, there are a number of potential explanations for this: (1) there is no relationship between the processes and the outcomes; (2) a level of performance (compliance) has been reached that is high enough – and resulted in outcomes that are good enough – that further improvements are not measurable; (3) relationships between processes and outcomes are not measurable; (4) the most appropriate or sensitive outcomes have not been identified; or (5) the relationships are more complex, with co-factors that may or may not be measured. There is no indication that SCIP is going to end anytime soon, so surgeons are obliged to continue to follow the recommended process measures. Quality improvement science, meanwhile, continues to investigate this area.

The question of whether medicine is “properly practiced” is also a focus of process-driven QI efforts. The current emphasis on delivery of evidence-based care follows from this concept. It may take years for a new care guideline to be inculcated into mainstream practice; a large body of literature has been devoted to the challenges and barriers to implementation of evidence-based practices [10]. On the other hand, just as with SCIP, it is unclear how far “evidence-based” practices can be extrapolated beyond the study populations on whom the evidence is based. Another issue to be cautious about is the fact that many practices make good sense, yet risk being discarded in favor of practices that are supported by a single well-designed study [11].

The patient is an active participant in this dimension, as his or her compliance with prescribed care is also measured. In the performance of surgical QI, it is important to note the patient’s behavior. If a surgeon recommended treatment in good faith and the patient failed to follow up, the surgeon should not take the “credit” for an adverse outcome.

The outcome of surgical care – in terms of recovery, functional restoration, or survival – is commonly used as an indicator of quality. There are clear advantages to the assessment of outcomes as a measure of quality: the validity of outcome as a reflection of quality is generally not questioned, nor is the value of outcomes such as survival and functional recovery. But as the science of surgical quality measurement has evolved, many questions have arisen [9].

First, what are the most relevant outcomes to measure? Return to premorbid functional status, including job performance, is a reasonable expectation after hernia surgery, while simply survival with some level of consciousness may be all that can be asked after emergency craniotomy for trauma. Thus, measures of pain control and time away from work are relevant to one, but not the other. A second consideration is the source of data. It is well-recognized that one would have markedly disparate results from datasets built on physician self-reports, concurrent chart reviews, and hospital billing records. In addition to the accuracy of the data, different systems use different definitions and exclusion/exclusion criteria. An example given by Wick and colleagues is as follows [12]: Consider that there are currently two programs available for colon surgical site infection (SSI) outcome monitoring: the Centers for Disease Control and Prevention (CDC) National Healthcare Safety Network (NHSN) and the ACS National Surgical Quality Improvement Program (NSQIP). The NSQIP defines case inclusion by Current Procedural Terminology (CPT) coding (surgeon professional fee coding), whereas NHSN inclusion is based on hospital billing coding by International Classification of Diseases, Ninth Revision (ICD-9). Exclusion criteria vary as well: NSQIP excludes surgical procedures where the wound was not closed, while NHSN does not. Although both systems employ the same CDC definitions to identify SSIs, the approach to follow-up diverges as NSQIP clearly outlines the process for obtaining 30-day follow-up on all patients (medical record review of index admission and all subsequent readmissions to the index hospital as well as other hospitals, clinic notes, and finally telephone calls to patients) but NHSN mandates review of inpatient records within 30 days of the procedure, with additional follow-up at the discretion of the reviewer. A third consideration is whether the outcomes are valid. As Donabedian [6] pointed out, many factors other than the surgical care may influence outcome. Proper risk adjustment of outcomes is currently being debated in many areas. It is difficult to reach consensus on what constitutes valid risk adjustment; moreover, the more risks are adjusted for, the more patients are required to have a sufficient number on which to conclude anything definitive. With all of these considerations, outcomes remain the ultimate measure of quality of medical care.