In their reviews of the study of cervical radiofrequency neurotomy (RFN) by MacVicar, Borowczyk et al. , some of the referees raised questions that were not critical of the internal validity of the study but which pertained to its external validity. In essence, these questions concerned the context of the study and differences between New Zealand and the United States with respect to patients and medical practices. Conventionally, such questions would be answered in a covering letter to the satisfaction of the Editor. However, in the present case, the questions and answers contained information that would be of wider interest beyond the peer-review process. That information serves not only to elaborate and explain the study of MacVicar, Borowczyk et al. , but also to underscore the wider relevance of stringent standards of practice. For that reason, we have composed this letter to serve as a compendium to the published article. It provides answers to several of the key questions raised in the review process.
Why Is New Zealand Specified in the Title?
At the time that the study was commenced, New Zealand was in a unique position. RFN was not practiced, was not available to patients, and was not reimbursed. In one sense, New Zealand was behind the rest of the world, but in another sense it had not been polluted with disparate methods of RFN of unknown efficacy. When a small group of practitioners wanted to make RFN available, the opportunity arose to do it well, and avoid mistakes made elsewhere in the world.
The senior medical advisor of the single insurance company that covers accident compensation in New Zealand faced a dilemma. There was no universally accepted technique for cervical RFN, and no consistent protocol. Various Dutch, German, and other protocols had been found wanting , and their outcomes were not impressive. Consequently, the medical advisor insisted that practitioners would be recognized and reimbursed only if they undertook formal training and were accredited to the standards of the authors of the seminal, controlled trials of the treatment [3,4]. Only by such means might the New Zealand public be protected from a flood of invalid and ineffective procedures that resembled the original only in name. Implicitly, if implementation of RFN in New Zealand produced the same outcomes for patients as in the seminal literature, it would be worth paying for; but carte blanche approval without accredited training invited clinical and financial chaos.
Under these conditions, seven medical practitioners were trained (for a population of 4 million). They underwent preliminary training to International Spine Invention Society (ISIS) standards in medial branch blocks , by cadaver courses and clinical practice. Subsequently, they underwent a 2-day cadaver course on cervical RFN, followed later by a 2-day period of practice supervised by one of the authors of the original study , again according to ISIS standards .
Following this training, it became relevant to document the experience acquired in New Zealand, both as a quality-assurance exercise, and to see if the plan of the medical advisor had been vindicated. Specifically, New Zealand data were also required to guard against opposition to RFN from hostile regulatory authorities who would rely on poor results reported in other countries, but by authors who used less stringent techniques for patient selection and for performing RFN.
The results reported by MacVicar, Borowczyk et al.  served both these local purposes, but they also serve a wider, international purpose. They set a benchmark of what outcomes can be achieved, by second-generation practitioners, and implicitly by adhering to stringent standards of practice [5,6]. A particular reason for stipulating “New Zealand” in the title of the article is to highlight that results obtained pertain explicitly to how RFN was adopted in New Zealand. Those results surpass any reported from Europe or the United States, but equal or surpass the original results from Australia, and consequently vindicate the use of RFN in New Zealand under the conditions under which it was introduced. Conversely, the lesser results from elsewhere cannot be used as evidence against RFN to prevent its continuation in New Zealand.
It Is Striking that Most of the Studies of Cervical RFN Have Been Produced by Authors Associated with Those Who Developed the Procedure. Unfortunately, Again the Authors Are Associated with the Developers!
It is somewhat strange that association with the developers of a technique is viewed as a flaw, sufficient to attract an exclamation mark. The implication is that such association must somehow bias the results of a study, as might corporate sponsorship. This concern seems to apply only in pain medicine. In other fields of science, such as physics, biology, or neuroscience, the usual response is for interested parties to take a sabbatical to work with the original authors, and learn a new technique. This is what the New Zealanders did: they went to the source in order to learn how to perform RFN properly. Yet, they are rendered suspect for having done so. It seems that pundits would prefer that they should have been self-taught, without the opportunity to have mistakes corrected; or to have learned from some other self-appointed expert who practiced their own version of RFN, not in accordance with either the original  or ISIS standards [5,6]. Yet, this preference strikes at the heart of the current confusion about RFN. While so long as other so-called experts are respected, and both practice and teach their own versions of RFN but without measuring and publishing their outcomes, patients and those who pay for treatment cannot know if the outcomes being obtained are the best possible. By training with the original developers, the New Zealanders have shown that they can reproduce the best possible outcomes, and their patients can therefore expect those outcomes.
How Long Was the Period of Training? How Steep Is the Learning Curve?
Initially, the trainees undertook a 2-day cadaver course devoted solely to cervical medial branch blocks and intra-articular injections. They practiced radiography to obtain and correct the required views. They practiced needle placement. They were drilled in radiographic interpretation of normal and distorted views of the cervical spine. They were finally assessed for accuracy of needle placement in cadavers. Following the cadaver course, the trainees spent 5 working days observing established practitioners [3,7] performing procedures. Thereafter, they returned to their own clinical practices to offer medial branch blocks to their own patients.
Within a calendar year, delayed largely by the logistics involved for the second phase of training, they undertook a 2-day cadaver course devoted exclusively to the practice of cervical RFN. Each candidate twice had to emulate performing an entire 2-hour procedure, identifying the target points for each nerve, placing the electrode into each of three positions along each of a sagittal and an oblique pass, and holding the electrode in place for the full duration of each mimicked lesion. During each procedure, fellow candidates and the mentor provided comments and corrections if required.
On a later occasion, over a 2-day period, each trainee demonstrated their technique on each of two of their own patients, observed by their fellow trainees, and supervised by an original author. Errors observed by their fellows or the mentor were corrected and later discussed. After this training, the trainees returned to their own practices to offer cervical RFN to their own patients. Several months later, they were visited and observed, for reinforcement and consolidation, by another one of the original authors .
During this training, the trainees were rapidly competent at placing the electrode and making lesions. The key learning, however, was always to obtain the correct radiographic view and to remain alert to maintaining that view.
Why Did Most Patients Have a History of Injury, and Why Were Not More Patients with Idiopathic Neck Pain Included?
For various reasons, both in Australia and in New Zealand, virtually all patients in the reported literature who undergo cervical RFN have had a history of injury, typically whiplash. An epidemiologic reason is that cervical zygapophysial joint pain is common in patients with chronic neck pain after whiplash injury. No studies have shown how common cervical zygapophysial joint pain is explicitly in patients with no history of injury. None of the seminal literature on cervical RFN [3,5,7] described patients with idiopathic neck pain. An administrative reason in New Zealand is that the only means by which to access RFN and have it reimbursed is through the Accident Compensation Corporation. Therefore, the majority of patients treated will have had a history of injury.
Is There More Information about the Patients with Injury, Such as Lesions Evident; and Was MRI Mandatory?
Diagnostic practices in Australia and New Zealand appear to differ from those in the United States. Australian and New Zealand practices more closely adhere to the evidence. The Canadian C-Spine rule  stipulates that imaging is not indicated for patients with simply a whiplash injury, or after an injury that does not preclude walking and turning the head. The patients in the study of MacVicar, Borowczyk et al.  satisfied these provisions and so, it would have been contrary to evidence-based guidelines to have undertaken imaging.
Why Were the Majority of Treatments Directed at Two Levels: C2,3 and C5,6?
The available epidemiologic data indicate that cervical zygapophysial joint pain is not randomly distributed throughout the cervical spine; it is distinctly more common at C2-3 and C5-6 [9–12]. Therefore, it is only natural to expect that cervical RFN would be more commonly performed at these segmental levels. A potential reason for cognitive dissonance in this regard is that, both in the study of MacVicar, Borowczyk et al.  and in the original studies of cervical RFN [3,4,7], individual joints or pairs of joints that were symptomatic were sought and identified. What was not countenanced was a practice common elsewhere: to perform medial branch blocks simultaneously at several segmental levels, and if the response was positive, to perform cervical RFN at all levels. This latter practice defies the guidelines [5,6] and is a disservice both to patients and to the craft of cervical RFN. It is both unnecessary and specious to denervate structures that have not been shown to be symptomatic.
What Percentage of Patients Had Complete Relief from Medial Branch Blocks?
Exact records upon which to answer this question were lost during the earthquakes that devastated Christchurch, New Zealand, in 2010 and 2011; but the experience of the authors was not detectably dissonant with the reported literature. Some 60% of patients with chronic neck pain prove positive to cervical medial branch blocks at one or more segmental levels .
When Performing Diagnostic Blocks, How Did You Find the Correct Level? Did You Need Several Attempts to Find the Correct Level?
The segments to be tested with medial branch blocks were selected by comparing the distribution of pain reported by patients with the pain maps of Cooper et al. . Doing so proved to be reasonably efficient. A post hoc analysis of the records of the patients in the study of MacVicar, Borowczyk et al.  showed that blocks were positive at the initially selected level in 48% of cases; a second level had to be tested in 20% of cases; a third level in 27%; and more than three levels in only 5%. Most of the second and third levels tested were in patients with pain either at a segment adjacent to the one initially tested, or from consecutive segments embracing the one initially tested. Under the latter conditions, one block was required to find the upper of two levels, a second for the lower of two levels, and a third to confirm that both levels simultaneously had to be blocked to achieve complete relief of pain (visual analog scale [VAS] = 0).
The Use of “Complete Relief” Is Misleading; It Would Be More Accurate to State “at Least 80% Relief”
The paradigm of cervical RFN is that if medial branch blocks provide complete relief of pain, RFN should also provide complete relief of pain. Therefore, a priori, the criteria for successful RFN required complete relief of pain. However, when all outcomes had been collected, and were being assessed, five patients were identified who posed a reality challenge. Variously, they reported only 80%, 85%, or 90% relief of pain, but nevertheless, they restored all of their desired activities of daily living, and required no health care for their neck pain. Although these five patients failed the a priori criterion for complete relief of pain, they had outstanding responses for functional outcomes. The reality was that these patients looked and behaved as if they were successes. Therefore, as described in the text of MacVicar, Borowczyk et al. , for these five patients the criteria for successful outcome were relaxed post hoc. However, at no time was the expectation to entertain a criterion of “at least 80% relief.” There were other patients who did achieve 80% relief, or 60% relief, but they did not restore their activities, or did not cease health care for their neck pain. Consequently, they were classified as not successful outcomes.