Improving Outcomes With Subspecialization and Regionalization

Introduction

The provision of high-quality, safe, affordable neurosurgical care is a goal sought after by the surgeon, patient, hospital, and insurers. Evaluation of hospital and individual neurosurgical performance has become more elaborate due to increased costs, rising patient expectations, demands for greater transparency and concerns over quality, efficiency, and safety. In neurosurgery, it is known that procedures and techniques involve learning curves with improvement as experience is accumulated, up to a point. At the hospital level, increased case volume has been strongly associated with improved outcomes. Subspecialization and regionalization for management of specific complex patients within neurosurgery has been an ongoing process aimed at improving care and lowering costs. There is less tolerance for suboptimal results due to learning curves. Also, there is decreased working and training hours available to learn such techniques, creating a demand for methods of accelerating the surgical learning curve via surgical simulation.

Most hospital quality improvement projects concentrate on the process and protocols of health care delivery. In surgery, there is an additional component that greatly impacts the quality of care, namely, the technical skills of the surgeons and the teams that play a role in the delivery of care to neurosurgical patients. To evaluate outcomes one must have measures that are examined which are uniform. Part of the difficulty in addressing quality of surgical care is the absence of universally agreed upon outcome metrics. There are several surrogates for quality which are used with mostly population derived databases including rates of mortality, rates of major postoperative complications, rates of type of discharge disposition, cost, as well as length of stay. These outcomes have been used in the examination of volume-outcome relationships within neurosurgery.

To improve those metrics, there has been a move toward focusing expertise and experience of both surgeons and care teams, and evidence exists that both can improve quality of care. Here we review the available evidence supporting subspecialization, experience, volume-outcome relationships, efficiency of training and simulation, patient satisfaction, and regionalization as methods to improve quality and safety of delivered neurosurgical care while improving patient satisfaction and cutting costs. We also discuss the difficulties in measuring these outcomes and limitations of currently available data.

Subspecialization

Despite an earlier effort to increase primary care delivery in an attempt to defragment medical care and cut costs, there continues to be a growing trend toward subspecialization within most medical specialties and neurosurgery is certainly no exception. Specialization within the medical system has naturally followed medical progress and increases in scientific and medical knowledge for nearly 200 years. The Flexner Report, published in 1910, called for more research-based medical education and led to the growth of physiology, anatomy, and biochemistry-based approaches to medical education and practice. This has logically led to increased subspecialization in the pursuit of greater knowledge and ultimately improved patient care. The first medical specialty to create its own assessment board was ophthalmology in 1917.

Subspecialization and focus of practice is a natural consequence in neurosurgery. As knowledge regarding disease processes and management increases, it is not possible for a general neurosurgeon to remain up to date on relevant literature, and maintain what is thought to be an acceptable volume to achieve high-quality safe patient care in all aspects of neurosurgery. There is evidence that subspecialization through either fellowship training or focus of one's practice can improve outcomes. Long before neurosurgical fellowships existed, neurosurgeons began to focus their practices and subspecialize in order to provide higher-quality care to their patients. This happened as neurosurgical care became more complex.

The Committee of Advanced Subspecialty Training (CAST) of the Society of Neurological Surgeons (SNS) is a non-Accreditation Council for Graduate Medical Education (ACGME) mechanism for recognizing subspecialty training. The breakdown of subspecialty and the number of accredited programs are as follows: neurocritical care: 17 programs, neurotrauma: 1 program, cerebrovascular: 7 programs, endovascular: 21 programs, spinal neurosurgery: 40 programs, neurosurgical oncology: 12 programs, peripheral nerve: 2 programs, stereotactic and functional neurosurgery: 24 programs. Pediatric neurosurgery fellowships can have accreditation from CAST and the Accreditation Council for Pediatric Neurosurgery Fellowship (ACPNF); there are currently 28 ACPNF accredited programs. The breakdown of fellowships illustrates that not all graduating neurosurgeons can do fellowships, and nor is there an expectation to do so. Not all fellowships are accredited and the apprenticeship model does exist and is still accepted for postresidency training. Enfolded fellowships which take place during residency are also not eligible to be accredited by CAST.

Subspecialty training is thought to lead to improved outcomes. For example, a large metaanalysis of the literature on the effect of procedural volume or surgeon specialty on outcome of lung resections for cancer was undertaken. General surgeons had significantly higher mortality risk than general thoracic (OR 0.78) or cardiothoracic surgeons (OR 0.82), supporting the notion that subspecialty training and perhaps a threshold number of cases are what matters.

Fellowship training within neurosurgery is not only thought to lead to safer and higher-quality patient care, but many residency graduates feel they are more marketable and able to get better jobs with fellowship training. Numerous neurosurgery job postings advertise for “fellowship trained” neurosurgeons. Also, the nature of neurosurgical training has changed, especially since the reduction in resident work hours in 2003. This, along with the dramatic increase in neurosurgical knowledge, has made it difficult for residents to feel comfortable performing the full spectrum of neurosurgical cases by the end of their training. Consolidation and regionalization of health care services has also impacted this current feeling of a need for further training after residency, with many highly specialized services such as cerebrovascular and functional neurosurgery concentrated in high-volume centers. Moreover, there is evidence of a correlation between clinical volumes of surgeons, hospitals, complication rates, and mortality. As will be discussed later, many studies suggest that high-volume surgeons and medical centers deliver higher quality care for complex patients and procedures. Which procedures have improved outcomes with increased volume, and what is the threshold volume required to achieve superior outcomes are questions that remain to be answered. This will likely vary depending on procedure complexity. Some residency programs, such as those at the University of Texas, Houston, have begun creating divergent tracts within residency, so that residents may begin to focus their career prior to graduation from their training program. Other programs have chosen to include away rotations when their own centers are not hospitals where they can provide the high-volume experience for their trainees. A survey of neurosurgery training programs found that 20% of programs have mandatory away rotations .

The trend toward subspecialization and focus of practice has led to recent changes in the American Board of Neurological Surgery (ABNS), directed at more relevant exams that are aimed at improving patient safety and fulfilling the broad mission of the ABNS, to encourage the study, improve the practice, elevate the standards, and advance the science of neurological surgery, and thereby to serve the cause of public health. Beginning in 2017, the ABNS certifying oral examination was modified to the following: one session aimed to test knowledge and competency in general neurosurgery, a second session testing the candidate's competency on elements of his or her focused surgical practice, and a third session consisting of 5–8 cases selected from the 150 cases submitted during the application process.

Experience

While subspecialty training is associated with clinical interest and expertise, it is different from experience. Experience comes with practice and is typically measured in terms of case volume or years of practice. Experience is not necessarily associated with subspecialty training or interest. For example, a general neurosurgeon with significant experience clipping aneurysms may have better outcomes than a fellowship trained cerebrovascular surgeon fresh out of fellowship, or a fellowship trained cerebrovascular surgeon whose practice environment provides an inadequate volume of open vascular cases. Experience is recognized as a key driver of individual surgical performance, giving rise to a learning curve. In the largest metaanalysis done at the cross-specialty level, Maruthappu and colleagues systematically reviewed studies evaluating the influence of surgical experience on individual performance by searching MEDLINE, EMBASE, PsycINFO, AMED, and the Cochrane Database from inception to 2013, identifying 57 eligible studies from 6950 citations, including 1,061,913 cases and 35 procedure types performed by 17,912 surgeons. They showed that increased case volume was associated with significantly improved outcomes. A plateau phase or maturation in the surgical learning curve that ranged from 25 to 750 procedures depending on the procedure was observed. Increased years of experience was also associated with improved outcomes.

There are stages of surgical learning: an initial learning phase, a plateau phase, and occasionally a deterioration phase. This has been studied primarily in the field of general surgery. For both years of practice and case volume, initial increases in experience lead to improvements in a range of outcomes such as operative time, mortality, and costs. Variation exists between individual surgeons in the time taken to master a procedure and average learning curves can be derived from a statistically modeled curve. After an initial learning phase, a plateau in performance is frequently seen whereby increases in experience do not result in improvements in outcomes. Reaching this plateau varies depending on the surgeon and procedure. In general surgery, for example, a single-port laparoscopic cholecystectomy procedure can be successfully learned in 8 cases, whereas even a well-trained surgeon requires at least 5 years to learn septomeatoplasty. Importantly, in certain cases, there exists a deterioration in performance where increasing experience, either case volume or years of practice, results in worsened performance. A cause for this is not clear. Possibilities include mental fatigue, burnout, poorer compliance with evidence based medicine, and reduced stress tolerance in older surgeons, or greater administrative research and leadership responsibility reducing case frequency and volume.