What patients know

Patients with lumbar disk herniation and radiculopathy showed substantial improvement with both surgical and nonoperative treatment in the randomized trial component of the Spine Patient Outcome Research Trial (SPORT) [17]. The trial compared outcomes for 501 patients randomized to standard open diskectomy or nonoperative treatment individualized to the patient. With intent-to-treat analysis, outcomes were equivalent for all primary and secondary outcomes, with the exception of sciatica symptoms, which showed greater improvement with surgery. Between-group differences in improvement favored surgery, but the differences were small and not statistically significant for the 3 primary outcomes, the Short-Form 36 Bodily Pain and Physical Function Scales and the Oswestry Disability Index.

The SPORT trial has been the subject of much negative criticism from the surgical community. Prominent surgeons expressed concerns about this study even during its inception phase [10]. At the time, a New York Times article quoted a relevant comment from the principal investigator of the study, Dr James Weinstein: “I've met with 2 groups who said they fear the results will take away their practice,” Dr Weinstein said. “I don't know how to deal with that. I don't know what the results will be [6].”

Since publication of 2-year outcome results, varied arguments have been advanced to interpret the findings as favorable for surgical treatment [1], [2], [3], [7], [14], [19]. The SPORT has been criticized for (1) asking the wrong question because we already know surgery is effective [14]; (2) studying the wrong diagnosis because lumbar disk herniation is a heterogeneous diagnosis in which clinical details determine preferred treatment [19]; (3) studying the wrong patients because enrolled subjects did not have adequate prior treatment [3]; (4) lacking standardized nonoperative treatment [5]; (5) lacking sham surgery as control [4]; (6) performing the wrong analysis [1]; or (7) getting all of these and many other things wrong [8]. These varied and protracted arguments cannot diminish the crucial knowledge reinforced by the SPORT trial—informed patients know which treatment will lead to a successful outcome.

Let us say that a middle-aged woman presents to a spine surgeon with radicular pain radiating below the knee associated with lumbar intervertebral disk herniation at the L5-S1 level. The symptoms have been persistent despite 6 weeks of nonoperative care, which included education/counseling, physical therapy, epidural injections, chiropractic therapy, antiinflammatory medications, and opioid analgesics. Physical examination shows evidence of nerve root irritation; straight leg raise is positive between 30° and 70°. She has an asymmetrical ankle reflex, depressed on the symptomatic side. On the symptomatic side, she also has decreased sensation in the S1 dermatomal distribution and weakness in the S1 myotomal distribution. She has undergone magnetic resonance imaging that shows a protruded, extruded, or sequestered fragment at the level and side corresponding to the clinical symptoms. She has not had prior lumbar surgery, and she does not have cauda equina syndrome, scoliosis, segmental instability, vertebral fracture, spine infection or tumor, or inflammatory spondyloarthropathy. She is seeking a surgical opinion. She is not pregnant, she has no comorbid conditions contraindicating surgery, and she is willing to have surgery within the next 6 months. Should she be advised to undergo lumbar diskectomy?

Among the questions that need to be addressed before making such a recommendation are the following [18]:

These are questions that SPORT aimed to answer [16], [17].

The clinical description of our patient is an almost verbatim reiteration of the inclusion and exclusion criteria published for subjects enrolled in SPORT [16], [17]. Such a patient, who has persistent symptoms after initial 6 weeks of nonoperative treatment and focal neurologic findings that include weakness, is often considered an ideal surgical candidate. Yet, these were the patients who entered the SPORT study. These potentially ideal surgical candidates were randomized to either receive surgery or continued nonoperative treatment.

In contrast to many clinical trials, patients enrolled in SPORT had a unique process for informed consent that incorporated principles of shared decision making [13], [15]. Patients met a research coordinator to receive detailed information about their treatment options. They watched a previously validated educational video that described genuine scientific uncertainty regarding the preferred further treatment at this stage of the patient's symptoms. The SPORT showed that informed consent is achievable, and many patients are willing to accept random allocation between surgical and nonsurgical care when they understand clinical equipoise.

The informed consent process in SPORT also ensured that patients understood that their right of autonomy was not relinquished in becoming research subjects. This type of informed consent empowered the enrolled research subjects to not be afraid to opt out of the assigned treatment if their symptoms changed. Those whose symptoms improved before assigned surgery decided against surgery; patients with worsening symptoms decided against assigned nonoperative treatment. Thus, by the 2-year end point, 40% in the surgery group and 45% on the nonoperative group did not receive the assigned treatment.

The high rate of crossover for both treatment arms is the most common criticism of SPORT, and often, it is portrayed as a fatal flaw that invalidates the study [1], [2], [3], [4], [7], [8], [19]. However, high crossover is an expected consequence of informed research subjects exercising autonomy of their role as patients. Patients understandably value optimal care more than compliance with a research protocol. The decision by many patients to not continue their assigned treatment did not violate their responsibility as research subjects because the change in their symptoms represented a change from the conditions under which they agreed to enroll in the study. If symptoms improved after their decision to enroll, some chose not to proceed with their random assignment to surgery; if symptoms worsened, many chose surgery despite random allocation to nonoperative care. The SPORT researchers showed respect for patient autonomy by allowing crossovers.

The SPORT involved more than 50 surgeons practicing in 13 widely dispersed clinical sites across 11 US states; the research goal was clearly to study clinical practice in a real setting and not to conduct a scientific experiment under artificial laboratory conditions. Although rigidly forcing patients to follow the randomly allocated treatment by shrinking the time interval between treatment allocation and surgery or by using a sham surgery control arm may provide a more scientific comparison of surgery and nonoperative treatment, the results would not be as applicable to real-world conditions of spine care, where time passes between evaluation and treatment, symptoms of patients change, and consequently, preferences of patient-research-subjects change. Such a study may be better science, but it may also be irrelevant to patient care.

The SPORT shows “use effectiveness” for surgery for informed patients [9]. It depicts outcomes for real-world spine care. Suppose the random treatment assignment were the advice a patient received during a consultation from a confident, assured, or dogmatic surgeon; the crossover patients would then represent those patients who failed to heed the advice of the surgeon. Are the patients who do not follow the recommendations of this consultation doomed to poor outcome?

The intent-to-treat analysis allows us the opportunity to see the consequences of following or not following the surgeon's advice. The SPORT showed that patients can safely choose treatment based on how their symptoms change further, after they have a surgical consultation for persistent symptoms after initial nonoperative treatment. If patients whose symptoms are not progressing receive a strong recommendation for surgery, they should feel free to not follow that advice. They can safely choose to monitor their symptoms and only have surgery if their condition worsens further. Continued nonoperative care is safe and does not burn any bridges. Avoiding surgery does not risk catastrophe or irreversible harm [17]. Patients who choose to wait will not become paralyzed, incontinent, wheelchair bound, or bedridden. A separate study showed delayed surgery does not lead to worse outcome [11], [12].

The surgeons participating in SPORT were experienced, respected, and in many ways elite surgeons. It would be hard to argue that more skilled surgeons may have yet better patient selection or surgical technique and even better outcomes in their hands at their specific sites of practice. Regarding diskectomy, the randomized trial portion of SPORT showed diskectomy was relatively safe, with 4% frequency of dural laceration, 4% reoperation within 1 year [17], median blood loss of 49.5 mL, median surgery time of 75 minutes, and 2% rate of blood transfusion. In the observational cohort portion of SPORT, a 1-level lumbar diskectomy procedure had an interquartile range of 0 mL to 1500 mL for blood loss and 15 minutes to 333 minutes for surgery duration [16]—that means 25% of cases exceeded these ranges for blood loss and surgery duration for a 1-level lumbar diskectomy. Portraying lumbar disk surgery as a trivial operation is a fallacy. Perhaps the surgeons who required longer than 333 minutes should desist; we do not know what to say about the 25% of surgeons who performed a diskectomy in less than 15 minutes.

The SPORT is an important, well-designed, and well-reported research project. Results of SPORT show that it would be wrong to tell our patient that she must have surgery soon for a successful outcome. Alternatively, it would also be wrong to tell her she must delay or avoid surgery. The crucial determinant of outcome from this point forward is how the patient's symptoms change further. Also, because the enrollment criteria did not include them, additional considerations beyond the inclusion and exclusion specifications, such as more detailed understanding of the symptom patterns, more nuanced physical examination findings, more sophisticated interpretations of the imaging findings, or more weighty pronouncements of expert or reputable surgeons, are not necessary to prognosticate treatment outcomes. Informed patients know which treatment is right for them.

The SPORT teaches us humility. It emphasizes the importance of informed consent and shared decision making. In this process, the surgeon's role is to adequately evaluate and understand the patient's condition, know which research evidence applies to the specific circumstances at hand, and inform patients of the relevant treatment options. Within the clinical conditions described by the SPORT enrollment criteria, rigidly advocating surgery will not necessarily lead to better outcomes. Although for some it seems comfortable to find fault with the method used in SPORT and thereby dismiss the findings, this approach will not nullify the most important lesson taught by SPORT—informed patients know which treatment is best for them.