Quality indicators in gynecologic oncology

Highlights

• Quality assessments must consist of patient-centric outcomes.

• Multiple intrinsic challenges exist to reliably assess quality in the American healthcare system.

• The guiding principle behind efforts to assess and improve quality is to improve the human condition.

Introduction

In the mid-nineteenth century, an abolitionist and preacher named Theodore Parker popularized three concepts championed as uniquely American at the time: 1. All people are created equal, 2. All possess unalienable rights, and 3. All should have the opportunity to develop and enjoy those rights. In the 21st century the concept of universal healthcare coverage as a fundamental right remains controversial. Should all citizens have the opportunity to enjoy healthcare coverage, or should wealthy or otherwise privileged individuals have better access to care, and by extension, higher quality of care? The cognitive dissonance created by the absence of universal healthcare coverage exists in the context of these American ideals, but even more acutely in juxtaposition to the internal ideals shared by many healthcare providers. Regardless of the ultimate decision on universal coverage in this country, the quality of the healthcare we provide will be of central and growing consequence in coming years, and has direct relevance to the concept of equal opportunities. Are we committed to healthcare equity (justice and freedom from bias to promote health for all) or just equality (likeness of assistance regardless of need)? Or neither? If we cannot accurately define and measure high quality care, we cannot safeguard consistent, high quality healthcare delivery to the American population. Despite the importance of the Affordable Care Act, the turmoil surrounding its debate has distracted many from the seismic shift towards a reimbursement model based on performance rather than services provided [10]. This focus on quality will drive practice changes to ideally improve patient outcomes and lower cost. Quality indicators are useful tools to help create expectations around specific medical conditions and objectively assess effectiveness. However, quality indicators also have the potential to create consternation or outright indignation within a profession with a history of relatively little oversight on matters pertaining to performance. If properly executed, the development of quality indicators will improve care, better inform patients, and foster the delivery of consistently exceptional healthcare for all, irrespective of region of residence, race, ethnicity, or gender. If poorly executed, quality indicators may result in economic penalties for high performers, paradoxically worsen outcomes and access to those in greatest need, and disillusion providers who make the greatest sacrifices for patients. Indeed, healthcare delivery is complex, and any process to improve upon its current state will be necessarily iterative with an eye towards long-term success. This review aims to define quality, examine challenges to creating indicators reflective of quality for the purpose of outcomes improvement and benchmarking, and provide examples of useful tools and metrics, which may be applied to surgical gynecologic oncology.

What is quality healthcare?

The Institute of Medicine defines health care quality as “the degree to which health care services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge [23].” This definition emphasizes population- and evidence-based considerations, and goes on to highlight six dimensions of quality:

• Effectiveness (use of evidence-based medicine to avoid both over- and under-treatment)

• Efficiency (responsible use of resources)

• Equity (justice in healthcare)

• Patient centeredness (autonomy)

• Safety (To err is human) [11]

• Timeliness (access)

In surgical specialties, quality has been distilled to a simple equation consisting of outcomes divided by complications: excellent outcomes paired with low complication rates equate with high quality. More recently, other facets of quality have been recognized, such as patient satisfaction and patient reported outcomes. But it is difficult to discuss quality without also acknowledging the importance of value, as noted in the efficiency category of the IOM's six quality dimensions. High value care is excellent quality at low cost. The importance of value is illustrated by the patient who undergoes hysterectomy, is dismissed from the hospital same day without complications, and returns to work within a week. While this may be considered high quality care, if there was no indication for hysterectomy in the first place, the intervention was of very low value. The critical nature of understanding and measuring value to drive improvement has been championed by Dr. Michael Porter and his colleagues at Harvard Business School and provides an important framework for developing rational improvements not only to individual hospitals, but to the entire American healthcare system [28]. Some of the concepts presented here are heavily influenced by this work and the reader is invited to read these primary sources [29]. Furthermore, an excellent review on value-based healthcare by Cohn, et al., has recently appeared in this journal and will not be repeated here [10].

In order to measure quality care in the field of gynecologic oncology, we must first agree on which outcomes are most important. The outcomes measure hierarchy is an objective, multidimensional construct divided into three tiers [28] (Fig. 1). Tier 1 outcomes are most important and include long-term measurements of survival and functional status. Overall survival of patients diagnosed with ovarian cancer or pregnancy rates after fertility-sparing surgery for cervical cancer are two examples. Patient reported outcomes are also critical Tier 1 endpoints. Tier 2 outcomes include short-term measures related to the process of recovery. Time to return to work, occurrence of surgical complications such as surgical site infection, and the length of time required for the cycle of diagnosis and treatment for a patient with endometrial cancer are Tier 2 outcomes. The sustainability of recovery and longer term effects of treatment are included in Tier 3. Examples include cancer recurrence following surgery for ovarian cancer or occurrence of neuropathic pain after treatment with intraoperative radiation therapy. Tier 1 outcomes carry greater weight than Tier 3, but one commonality among all these endpoints is that they are patient-centric. This realignment of reimbursement according to patient outcomes (rather than provider or payer endpoints) is a fundamental shift that cannot be overemphasized and is a critical consideration for any assessment of quality. Importantly, these outcomes measurements are built around broad disease states (ovarian cancer, dysfunctional uterine bleeding), not specific interventions (lymphadenectomy, hysterectomy). The use of medical conditions rather than procedures permits a meaningful assessment of larger populations and solves some (but not all, as will be discussed) of the challenges of risk adjustment necessary to compare practices with disparate case mixes. The disease-based nature of these measurements is an important distinction that is lacking in our healthcare system. Thus, a complex and expertly performed surgery for a patient with ovarian cancer may be offset by a delayed or inappropriate chemotherapy regimen if survival is the ultimate endpoint. A less than aggressive cytoreduction leaving residual disease will be associated with improved short term outcomes, but perhaps at the expense of long-term survival. In this way, surgical outcomes cannot be divorced from patient outcomes. In fact, higher quality care may be associated with medical rather than surgical treatment in many instances, or even no intervention at all.

The outcomes hierarchy (Fig. 1) provides general guidance for the types of outcomes which are reflective of quality, but further questions are in need of refinement, including the magnitude of survival improvements considered important, which instruments are most reliable to measure patient reported outcomes, and which complications should be most heavily weighted. For instance, composite measures of morbidity may include complications as disparate as urinary tract infection, sepsis, and death. Outcomes measures are also less meaningful without benchmarking, which sets context and expectations, and focuses improvement efforts. The American College of Surgeons National Surgical Quality Improvement Project provides benchmarking for some Tier 2 outcomes and represents the gold standard for risk adjustment according to initial patient conditions and procedures. But measured outcomes are short-term, based on the use of an operative intervention rather than around specific medical conditions, and functional outcomes are not widely available. Nevertheless, transparent sharing of complication rates made available through institutional and societal registries or national databases such as NSQIP have provided an opportunity to drive improvement within individual departments, within health systems, and across the country.

Fig. 2 shows two examples of anonymized NSQIP results. As competitive surgeons working towards perfection, our eye is automatically drawn towards the decile ranking of 8 on the left, indicating performance in the lower 20% of the hospitals sampled. However, the tight distribution of odds ratios means that one greater or fewer complication in this category may shift the results by several deciles. So while this institution is in the lowest 20% of those sampled the odds ratio of 1.06 suggests other areas may offer greater opportunities for improvement. In contrast, outcomes are widely distributed on the right and this particular hospital is a clear outlier requiring intervention. The number of events and historic performance also help direct resources to improve the lowest performing areas rather than attempting to elevate the highest performers.

Ideal assessments of quality would therefore measure tier 1, 2, and 3 outcomes that are important to patients, not just providers. All six domains of quality outlined by the IOM would be measured individually, within each local health system, and nationally, with a focus on the well-being of the entire population. Outcomes would be measured long-term and based on disease state rather than linked to specific interventions. Transparent reporting would encourage individual practices to optimize workflows, reduce costs, and standardize medical decision-making towards the singular goal of achieving the best outcomes for patients rather than doing what is most profitable. Quality measures would be used to drive standardization and dissemination of best practices throughout the country. Healthcare systems with poor outcomes in specific disciplines or medical conditions would be compelled to improve or develop relationships with higher performing hospitals to serve those particular patients. This would centralize services, concentrate volumes and expertise, and improve care across the population. Realizing these ideals could restore faith in the American medical system and elevate national outcomes to that seen in other first world countries, while simultaneously cutting costs by reducing incentives to deliver profitable, yet ineffective care.

In 2018, the quality assessment elements described in the preceding paragraph are both necessary and hopelessly naïve. Effective implementation would necessitate massive changes to our healthcare system that are unlikely to take shape for many years. Achieving meaningful improvements to a $3 trillion industry of unimaginable complexity will occur over many decades, not months or years. However, in our ongoing quest to understand human pathophysiology and cure disease, physicians have faced even larger challenges in the past. Above all, we must reaffirm our commitment to the guiding principle most succinctly stated by Dr. Porter, namely, to improve the human condition. A discussion of the flaws in the American healthcare system is beyond the scope of this review. Rather we will next focus on challenges to quality assessment faced by providers and physician leaders, and strategies to work towards a more ideal future state. Creating concrete measures of quality will be an important aspect of that future state.