operative and consultative proportions of neurosurgical

CLINICAL ARTICLEJ Neurosurg 130:1098–1106, 2019

ABBREVIATIONS HIC = high-income country; IQR = interquartile range; LIC = low-income country; MIC = middle-income country; TBI = traumatic brain injury; TSI = trau-matic spinal injury; WB = World Bank; WFNS = World Federation of Neurosurgical Societies.SUBMITTED February 8, 2017. ACCEPTED October 18, 2017.INCLUDE WHEN CITING Published online May 11, 2018; DOI: 10.3171/2017.10.JNS17347.

Operative and consultative proportions of neurosurgical disease worldwide: estimation from the surgeon perspective Michael C. Dewan, MD, MSCI,1,2 Abbas Rattani, MBe,1,3 Ronnie E. Baticulon, MD,4 Serena Faruque, MS,5 Walter D. Johnson, MD, MBA, MPH,6 Robert J. Dempsey, MD,7 Michael M. Haglund, MD, PhD, MACM,8 Blake C. Alkire, MD, MPH,1,9 Kee B. Park, MD,1 Benjamin C. Warf, MD,1,10 and Mark G. Shrime, MD, MPH, PhD1,11

1Global Neurosurgery Initiative, Program in Global Surgery and Social Change, Department of Global Health and Social Medicine; 10Department of Neurological Surgery, Boston Children’s Hospital, Harvard Medical School; 9Office of Global Surgery and Health; 11Department of Otolaryngology, Massachusetts Eye and Ear Infirmary; 5Department of Neurosurgery, Brigham and Women’s Hospital, Boston, Massachusetts; 2Department of Neurological Surgery, Vanderbilt University Medical Center, Nashville; 3Meharry Medical College, School of Medicine, Nashville, Tennessee; 4University of the Philippines College of Medicine–Philippine General Hospital, Manila, Republic of the Philippines; 6Emergency & Essential Surgical Care Programme Lead, World Health Organization, Geneva, Switzerland; 7Department of Neurological Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin; and 8Division of Global Neurosurgery and Neurology, Department of Neurosurgery and Duke Global Health Institute, Duke University, Durham, North Carolina

OBJECTIVE The global magnitude of neurosurgical disease is unknown. The authors sought to estimate the surgical and consultative proportion of diseases commonly encountered by neurosurgeons, as well as surgeon case volume and perceived workload.METHODS An electronic survey was sent to 193 neurosurgeons previously identified via a global surgeon mapping initiative. The survey consisted of three sections aimed at quantifying surgical incidence of neurological disease, con-sultation incidence, and surgeon demographic data. Surgeons were asked to estimate the proportion of 11 neurological disorders that, in an ideal world, would indicate either neurosurgical operation or neurosurgical consultation. Respondent surgeons indicated their confidence level in each estimate. Demographic and surgical practice characteristics—including case volume and perceived workload—were also captured.RESULTS Eighty-five neurosurgeons from 57 countries, representing all WHO regions and World Bank income levels, completed the survey. Neurological conditions estimated to warrant neurosurgical consultation with the highest frequen-cy were brain tumors (96%), spinal tumors (95%), hydrocephalus (94%), and neural tube defects (92%), whereas stroke (54%), central nervous system infection (58%), and epilepsy (40%) carried the lowest frequency. Similarly, surgery was deemed necessary for an average of 88% cases of hydrocephalus, 82% of spinal tumors and neural tube defects, and 78% of brain tumors. Degenerative spine disease (42%), stroke (31%), and epilepsy (24%) were found to warrant surgi-cal intervention less frequently. Confidence levels were consistently high among respondents (lower quartile > 70/100 for 90% of questions), and estimates did not vary significantly across WHO regions or among income levels. Surgeons reported performing a mean of 245 cases annually (median 190). On a 100-point scale indicating a surgeon’s perceived workload (0—not busy, 100—overworked), respondents selected a mean workload of 75 (median 79).CONCLUSIONS With a high level of confidence and strong concordance, neurosurgeons estimated that the vast majority of patients with central nervous system tumors, hydrocephalus, or neural tube defects mandate neurosurgical involvement. A significant proportion of other common neurological diseases, such as traumatic brain and spinal injury,

J Neurosurg Volume 130 • April 20191098 ©AANS 2019, except where prohibited by US copyright law

Unauthenticated | Downloaded 03/13/22 07:00 PM UTC

J Neurosurg Volume 130 • April 2019 1099

M. C. Dewan et al.

In 2015, the Lancet Commission on Global Surgery outlined the current estimate of the global surgical burden and the status of access to surgical care world-

wide.8 While the worldwide public health and medical communities have made headway in addressing common infectious, maternal, and neonatal illnesses, surgical care has been largely overlooked. Yet, an estimated 28%–32% of the total global burden of disease requires the expertise of a surgeon.15 Five billion people lack access to safe sur-gical care, and as a result more than 140 million surgical cases are being left untreated.1 Within this total surgical burden resides the contribution of neurosurgical disease, whose figure remains ill-defined. The burden of neurosur-gical diseases is likely great, and there is a tremendous need to train and redistribute neurosurgeons to meet this burden.11 While understanding the incidence and volume of neurological diseases is important, it is insufficient to ascertain the neurosurgical workforce and infrastructure necessary. We must also understand the proportion of dis-ease requiring surgical expertise. We sought to inform the estimation of neurosurgical case volume by quantifying the rate at which surgical intervention and surgical consul-tation are indicated for neurological diseases.

MethodsSurvey Administration

We conducted a survey of neurosurgical providers us-ing the Research Electronic Database Capture (REDCap) platform.6 Survey structure and question format were in accordance with the methodology previously outlined by Shrime et al.15 Surgeons had been identified via a global mapping project facilitated by the World Federation of Neurosurgical Societies (WFNS) and the Global Initia-tive Emergency and Essential Surgical Care (GIEESC) at the WHO.13,14 Briefly, officers of the 130 societies reg-istered with the WFNS or the members of the GIEESC were initially contacted via email to invite participation. For some countries without WFNS/WHO neurosurgery representation, an Internet search including “neurosurg* + country name” was conducted to identify the contact in-formation for a single practicing neurosurgeon willing and able to complete the questionnaire. Surgeons registered with the open access Global Neurosurgery forum (https://globalneurosurgery.org/members) were also included. In countries wherein a neurosurgeon could not be identi-fied by one of the means described above, the Ministry of Health, or an equivalent official body, was contacted and questioned regarding the presence and identity of any resi-dent neurosurgeons. Excluding small island nations in the Caribbean Sea, nearly 20 countries without a single neuro-surgeon were identified; therefore, their neurosurgical pro-

portion was not directly represented by means of survey administration. Finally, 12 North American neurosurgeons known to participate in international outreach missions, and with whom we had a personal relationship, were solic-ited. This final means of surgeon inclusion did not follow a systematic process but was believed to fairly capture the perspective from the North American continent.

On November 7, 2016, an electronic survey invitation was sent to 193 neurosurgeons in 89 countries. The survey consisted of three sections aimed at quantifying surgical incidence of neurological disease, consultation incidence, and surgeon demographic data (see Supplemental Mate-rials for full survey instrument). For 11 common, broad neurological categories (for example, stroke, brain tumor, neural tube defect), respondents were asked what percent-age of patients (0%–100%) with a given disease would, in an ideal world, require neurosurgical intervention. For the same neurological diseases, surgeons were asked what proportion of patients would require consultation by a neu-rosurgeon. For each estimate, a supplementary question asked survey respondents how confident they were in the answer provided (0—not at all confident, 100—extremely confident). A series of demographic questions explored the duration and setting of the practice, the patient population, and the most commonly treated diseases, among other items. Finally, neurosurgeons were asked to estimate their annual case volume, as well as their perceived workload on a scale from 0 (not at all busy) to 100 (extremely busy, overworked). The annual case volume corresponding to the mean surgeon workload was considered to represent the expected case volume of an average neurosurgeon. After the initial invitation and two reminder emails, the survey closed on November 23, 2016, at which time data were tabulated across responders.

Data Aggregation and Statistical AnalysisDiscrete demographic data were aggregated and report-

ed as absolute values and proportions. A weighted average was used to incorporate confidence levels of the expert esti-mates. Consider a fictitious example wherein two surgeons are estimating the proportion of a disease requiring surgi-cal intervention, as well as their degree of confidence for that estimate. Surgeon 1 estimates that a certain percent-age, pa, of the disease warrants surgery, and he describes his confidence in this estimate with percent certainty, wa. Meanwhile surgeon 2 estimates that pb of the disease war-rants surgery, with wb confidence. The weighted average can be calculated as follows: p̄ = (wapa + wbpb)/(wa + wb).

The median and interquartile range were described for both percentages of consultation and percentages of op-eration. The surgeon confidence for each disease was de-

vascular anomalies, and degenerative spine disease, demand the attention of a neurosurgeon—whether via operative intervention or expert counsel. These estimates facilitate measurement of the expected annual volume of neurosurgical disease globally.https://thejns.org/doi/abs/10.3171/2017.10.JNS17347KEYWORDS burden; global; neurosurgery; surgeon; survey


https://thejns.org/doi/suppl/10.3171/2017.10.JNS17347


M. C. Dewan et al.

J Neurosurg Volume 130 • April 20191100

scribed in the same fashion. Analysis of variance was used to evaluate differences in responses between WHO re-gions, World Bank (WB) income groups, and patient payer status. Similarly, differences in case volume and perceived workload were compared across the same metrics using ANOVA. A p value < 0.05 was used a priori to indicate significant differences between respondent groups. To ac-count for multiple comparisons across the 11 disease enti-ties, the Bonferroni correction was applied. The R statisti-cal package (version 3.3.2, The R Project for Statistical Computing) was used for weighting analysis, while Stata (version 14, StataCorp LLC) was used for ANOVA and graphing case volume data.

To illustrate the results of the survey, two heat matrices were generated using Tableau Public (version 10.2.0). For every disease category, aggregated data from each WHO region and WB income group were represented by a circle whose diameter and color varied depending on the per-centage of cases and the mean confidence of respondents, respectively.

ResultsDemographics

Among 193 neurosurgeons queried, 91 submitted par-tial responses and 85 completed the entire questionnaire (44% response rate). Survey responses were received from a total of 57 countries on 6 continents. Neurosur-geons reported being in surgical practice for a mean of 15 years (median 13 years). Twenty-five surgeons from high-income countries (HICs) responded, compared with 43 from middle-income countries (MICs) and 15 from low-income countries (LICs; Table 1). A diverse spread of respondents across WHO regions was observed: 26 in the African Region (AFR), 11 in the Region of the Americas–Latin America (AMR-L), 14 in the Region of the Ameri-cas–United States and Canada (AMR-US/Can), 5 in the Eastern Mediterranean Region (EMR), 6 in the European Region (EUR), 6 in the Southeast Asia Region (SEAR), and 15 in the Western Pacific Region (WPR).

The majority (82%) of respondents were based out of a national or regional referral hospital. Most (74%) engaged primarily in clinical responsibilities with some adminis-trative duties, while 23% had strictly clinical duties (Table 1). Public patients represented the primary payer status for 62% of respondents, relative to private patients (16%); 21% of respondents treated an equal proportion of private and public patients. Fifty-three percent of surveyed neurosur-geons treated primarily adult patients, while 16% treated primarily children, and 31% treated both age groups about equally.

Consultation IncidenceConsultation by a trained neurosurgeon was reported

to be indicated in almost all cases of brain and spinal tu-mors (96% and 95%, respectively; Table 2). Hydrocepha-lus (94%) and neural tube defects (92%) were among the next highest conditions to warrant neurosurgical evalu-ation, while stroke (54%) and epilepsy (40%) were the lowest. Conditions with a lower overall consultation rate (stroke, CNS infection, epilepsy) demonstrated a wider

TABLE 1. Demographics and practice of neurosurgeon respondents

ParameterNo. of

Respondents (%)

WHO region (n = 85) African 26 (31) Latin American 11 (13) North American 14 (16) Eastern Mediterranean 5 (6) European 6 (7) Southeast Asian 6 (7) Western Pacific 15 (18) Unspecified 2 (2)WB income level (n = 85) HIC 25 (29) MIC 43 (51) LIC 15 (18) Unspecified 2 (2)Hospital setting (n = 85) Tertiary or quaternary referral hospital 70 (82) Other hospital 11 (13) Hospital administration (nonclinical) 2 (2) Other 2 (2)Time distribution (n = 84) Entirely clinical 19 (23) Primarily clinical w/ some administrative duties 62 (74) Entirely administrative or research 1 (1) Retired, other 2 (2)Patient payer status, primary (n = 85) Public 53 (62) Private 14 (16) Both about equal 18 (21)Patient age epoch, primary (n = 85) Adult 45 (53) Pediatric 14 (16) Both about equal 26 (31)Most commonly treated conditions* (n = 87) Brain tumors 57 (66) Traumatic brain injury 51 (59) Hydrocephalus 41 (47) Degenerative spine disease 38 (44) Congenital malformations 21 (24) Traumatic spinal injury 17 (20) Vascular anomalies 16 (18) Stroke 10 (11) Spinal tumors 8 (9) Epilepsy 6 (7) Pain disorders 5 (6) Movement disorders 2 (2) Other 3 (3)* Respondents were permitted to select up to three conditions. Though some surveys were incomplete, the response to this question was completed on 87 surveys.



M. C. Dewan et al.

interquartile range (IQR) for consultation: 30%–78% for stroke, 30%–80% for CNS infection, and 46%–90% for degenerative spine disease. On the other hand, the respon-dents’ self-reported confidences of estimates were consis-tently high. Even for conditions with a wide consultation rate spread, the lowest confidence quartiles were 72 (epi-lepsy) and 74 (stroke) out of 100 (Fig. 1).

Surgical IncidenceHydrocephalus was reported to warrant operative inter-

vention (88% of patients with hydrocephalus) more than any other condition (Table 3). Neural tube defects, brain tumors, and spinal tumors were all perceived to merit sur-gery in more than 75% of cases. Meanwhile, patients with stroke (31%) and epilepsy (24%) garnered a surgical indi-cation at rates lower than those for other conditions. As with the consultation proportion, diseases with lower rates of surgery carried a wider range of answers and a wider range of confidence levels. Traumatic brain injury (TBI) and traumatic spinal injury (TSI) had a weighted surgical average of 47% (median 45%, IQR 30%–63%) and 52% (50%, 30%–70%), respectively. The interquartile confi-dence estimates for TBI and TSI, however, were both high at 78–93 and 72–91, respectively. There was no difference in surgical estimates or confidence estimates between re-spondents of differing WHO regions or primary patient payer statuses (Supplemental Table S1). There were also no differences in estimates between different income lev-els, with the exception of spinal tumors, wherein surgical estimates were higher for MIC (83%) than for LIC (65%; p = 0.0045; Fig. 2).

Case Volume and WorkloadSurgeons worldwide reported performing a median of

190 cases (IQR 120–300) and a mean of 245 cases (range 15–1500). The average perceived workload was 75 (on a 100-point scale; range 24–100), and the median response was 79 (IQR 65–87). The regression line in Fig. 3 demon-

strates the interaction between case volume and perceived surgeon workload. At the mean workload described by surgeons (75 on a 100-point scale), 223 cases are per-formed each year. The mean case volume for respondents did not differ by WHO region, WB income group, or payer status (Supplemental Table S2). Similarly, there were no differences in perceived workload across these three cat-egories (Supplemental Table S3).

Surgeons were also asked about the most common con-ditions they manage and were permitted to select up to 3 conditions from a broad list of neurological diseases. Six-ty-six percent selected brain tumors, 59% selected TBI, and 47% selected hydrocephalus (Table 1). Degenerative spine disease (44%), congenital malformations (24%), and TSI (20%) were the next most frequently treated condi-tions. Movement disorders (2%), pain disorders (6%), and epilepsy (7%) were the least commonly selected condi-tions among respondents.

DiscussionWe report the results of an international survey of neu-

rosurgeons describing surgical and consultative practices related to common neurological conditions. Eighty-five surgeons from 57 countries performed an average of 245 cases annually and reported a mean perceived workload of 75 on a 100-point scale. At the average workload, neuro-surgeons can be expected to perform 223 cases each year. Hydrocephalus, neural tube defects, and brain and spinal tumors were reported to warrant surgical intervention at a higher rate (78%–88%) than other neurological condi-tions. Meanwhile, stroke (31%) and epilepsy (24%) were reported to merit surgery less frequently. Despite nonsur-gical management for a large proportion of some diseases, neurosurgical consultation was deemed appropriate for more than half of all cases of stroke, CNS infection, and degenerative spine disease. Overall, neurosurgeons re-ported surgical and consultation incidence figures with a high degree of confidence; no median confidence estimate

TABLE 2. Estimated proportion of disease requiring neurosurgical consultation

Condition Mean Percent Requiring Consultation* Median Consultation % [IQR] Median Confidence of Estimate [IQR]†

Brain tumor 96% 100 [94–100] 98 [88–100]Spinal tumor 95% 99 [90–100] 97 [85–100]Traumatic brain injury 84% 88 [73–100] 95 [85–100]Traumatic spinal injury 84% 87 [70–100] 93 [83–99]Hydrocephalus 94% 98 [91–100] 98 [89–100]Neural tube defect 92% 98 [85–100] 97 [88–100]CNS vascular anomaly 88% 95 [75–100] 94 [81–100]Stroke 54% 50 [30–78] 88 [74–97]CNS infection 58% 60 [30–80] 87 [80–98]Degenerative spine disease 68% 71 [46–90] 89 [76–98]Epilepsy 40% 35 [20–51] 86 [72–96]

Respondents were asked to “indicate whether, to the best of your knowledge, the management of each condition would, in an ideal world, warrant consultation by a neurosurgeon.” Answers scaled from 0 (0% of patients require consultation) to 100 (100% of patients require consultation). * Proportions represent the mean values for each respondent after adjusting for the confidence level.† Confidence is measured on a continuous scale from 0 (not at all confident) to 100 (extremely confident).





M. C. Dewan et al.


fell below 80%. Similar to prior reports, the most common neurosurgical conditions treated by this multinational co-hort were hydrocephalus, tumors, and TBI.5

The volume and burden of neurosurgical disease is unknown. Though studies of global estimates of the neu-rological disease burden have been conducted,4,7,9 a study on the proportion requiring surgery has not. Without understanding the proportion of neurological diseases needing the attention of a neurosurgeon, quantifying the global demand for neurosurgery—and the workforce of neurosurgeons needed to meet this demand—is not pos-sible.16 The present study is important because it provides an essential component of the global neurosurgical bur-den calculation. The practical value of these figures will be realized when they are combined with known disease incidence, so that a neurosurgical case volume can be calculated. For this reason, broad disease categories (for example, TBI, CNS infection) were deliberately used to

promote compatibility of our data with epidemiological statistics gathered by government agencies and develop-ment organizations.

Determining a reliable estimate of the proportion of a given disease requiring neurosurgery is challenging. Comparing operating room logs with hospital admis-sion data on each disease necessarily omits the sample of patients not presenting to a health care facility with that disease.2 Even with an accurate sample, the proportion of patients receiving surgical care would differ across mul-tiple facilities—particularly between different countries, income levels, and WHO regions. Even country-wide data from billing registries fail to capture the ideal surgical rate and have limited generalizability outside the country of origin.10,12 Thus, we have followed the precedent estab-lished by the Lancet Commission on Global Surgery, to seek expert opinion from a diverse group of providers to estimate surgical incidence figures for common neurosur-

FIG. 1. Estimated proportion of neurosurgical conditions requiring consultation. For all disease categories, each circle repre-sents aggregated data from every WHO region and WB income group. The diameter of each circle is directly proportional to the percentage of neurosurgical cases that respondents believe requires consultation. On the other hand, color represents the mean confidence of the respondents. The light-to-dark spectrum corresponds to a confidence range from 60% to 100%. AFR = African Region; AMR-L = Region of the Americas–Latin America; AMR-US/Can = Region of the Americas–United States and Canada; CNSI = central nervous system infection; DSD = degenerative spine disease; EMR = Eastern Mediterranean Region; EUR = Eu-ropean Region; HC = hydrocephalus; HIC = high-income country; LIC = low-income country; MIC = middle-income country; NTD = neural tube defect; SEAR = Southeast Asia Region; TBI = traumatic brain injury; TSI = traumatic spinal injury; WHO = World Health Organization; WPR = Western Pacific Region. Figure is available in color online only.



M. C. Dewan et al.

FIG. 2. Estimated proportion of cases requiring neurosurgical operation. For all disease categories, each circle represents aggre-gated data from every WHO region and WB income group. The diameter of each circle is directly proportional to the percentage of cases that respondents believe requires neurosurgical operation. On the other hand, color represents the mean confidence of the respondents. The light-to-dark spectrum corresponds to a confidence range from 60% to 100%. See the legend for Fig. 1 for definitions of abbreviations. Figure is available in color online only.

TABLE 3. Estimated proportion of disease requiring neurosurgical operation

Condition Mean Percent Requiring Surgery* Median Surgical % [IQR] Median Confidence of Estimate [IQR]†

Brain tumor 78% 80 [70–90] 88 [75–96]Spinal tumor 82% 82 [70–90] 86 [79–96]Traumatic brain injury 47% 45 [30–63] 87 [78–93]Traumatic spinal injury 52% 50 [30–70] 82 [72–91]Hydrocephalus 88% 90 [81–95] 93 [85–99]Neural tube defect 82% 85 [70–94] 89 [80–97]CNS vascular anomaly 61% 62 [40–79] 84 [72–93]Stroke 31% 28 [16–43] 80 [65–90]CNS infection 44% 37 [20–60] 84 [72–92]Degenerative spine disease 42% 35 [20–61] 82 [72–95]Epilepsy 24% 20 [10–30] 82 [66–93]

Respondents were asked to “indicate whether, to the best of your knowledge, the management of each condition would, in an ideal world, warrant operation by a neuro-surgeon.” Answers scaled from 0 (0% of patients require consultation) to 100 (100% of patients require consultation). * Proportions represent the mean values for each respondent after adjusting for the confidence level.† Confidence is measured on a continuous scale from 0 (not at all confident) to 100 (extremely confident).


M. C. Dewan et al.


gical conditions.15 Though each estimate necessarily car-ries an error margin, as a group they represent evidence from the best, most feasible means of obtaining otherwise elusive and difficult-to-surmise surgical epidemiological figures.

Neurological disorders sampled in population-based studies are infrequently categorized into clean, discrete surgical or nonsurgical entities; without estimates like those captured in our survey, quantifying the operative burden is not possible. Hydrocephalus may seem like an obligatory neurosurgical entity, but consider the 1-day-old with intraventricular hemorrhage of prematurity and ven-triculomegaly, in whom a bowel perforation is detected. If comfort care is elected, is a neurosurgical consultation necessary? Similarly, a subpopulation of patients with neural tube defect (mild dysraphism) or brain tumors (dif-fuse systemic metastases and poor performance status) exists that may never need neurosurgical operation or per-haps even referral—even in an ideal world. The ethical is-sues surrounding these scenarios are beyond the scope of this paper, but they serve to illustrate the point that not all patients with classic neurosurgical diseases mandate the attention of a neurosurgeon. Ignoring this point and con-structing burden models without practical data risks set-ting up low-resourced health care systems for failure.

The findings of this study should be interpreted in light of several limitations. First, the method of respon-dent identification renders the results susceptible to se-lection bias. The survey data also carry the potential for responder bias and experiential bias—surgeons who treat a certain condition with frequency are likely to overesti-mate its incidence within a population. The overwhelm-ing representation of neurosurgeons operating in national or regional referral hospitals may have also biased these results. Similarly, a neurosurgeon wishing to emphasize—consciously or subconsciously—the importance of their trade may overstate the role that a surgeon plays in the care of neurological disease. This limitation could be ad-dressed by surveying nonneurosurgeons as well (for ex-ample, neurologists, emergency physicians, etc.); similar estimates would help to validate the figures reported in Tables 2 and 3. Second, the definition of what constitutes surgery may vary from one surgeon to the next, influenc-ing results. This was addressed, in part, by averaging sur-vey responses and weighting by degree of certainty. Third, while disease entities were intentionally kept broad to re-flect designations by the Institute for Health Metrics and Evaluation (IHME), this rendered estimates vulnerable to differing interpretations of disease states by the respond-ing surgeons. For example, when presented with the term

FIG. 3. Number of neurosurgical operations per annum and surgeon workload. The horizontal axis represents the estimated number of cases for each surgeon, relative to the surgeons’ perceived workload on the vertical axis. Workload was rated on a continuous scale from 0 (not at all busy) to 100 (extremely busy, overworked). The hollow diamond represents the intersection between the median number of cases (190) and the median workload (79). The solid triangle represents the intersection of the mean number of cases (245) and the mean workload (75 on a 100-point scale). The dotted line represents the regression line of fit [y = 65.7 + 0.042(x)]. The vertical dashed arrow indicates the number of annual operations (223) that correlates with the mean perceived workload by respondent neurosurgeons (75/100). Figure is available in color online only.



M. C. Dewan et al.

“traumatic brain injury,” a surgeon may be less inclined to incorporate an entity such as “concussion” into the disease denominator. In so doing, the estimate for the proportion requiring surgery would be inflated. In the case of TBI, the surgical estimate (47%) more likely reflects a TBI case admitted to the hospital or perhaps even TBI referred to a neurosurgeon. The use of broad disease terms (for exam-ple, hydrocephalus, TBI) represents an inherent weakness manifested during results interpretation. However, this deliberate strategy was intended to maximize the capture of neurosurgical disease. Moreover, our incorporation of confidence estimates provides a quality modifier not pre-viously employed. Also, by estimating the proportion of disease requiring neurosurgical consultation, we have ac-counted for a significant population that may not need an operation but whose health and well-being are enhanced by the presence and expertise of a neurosurgeon. Finally, a relatively small proportion of the worldwide population of neurosurgeons was surveyed to obtain results for globally generalized proportions. While even more respondents would have been advantageous, our numbers compare very favorably with those of peer investigators in the glob-al surgery community,8,15 and our results maintain their validity after testing for regional heterogeneity. Moreover, a 44% response rate for an email-based questionnaire ex-ceeded our study design expectations.

There are several strengths to this methodology and the results. First, each WHO region and each WB income level are represented by respondents. On the surface, the 15 surgeons (18%) from LICs may suggest underrepre-sentation of an income group. However, the proportion of neurosurgeons polled from LICs was much greater than that from HICs, a reconcilable feature of this study when considering the greatest unmet burden of neurosurgical disease likely resides in LICs. Second, the vast majority of clinicians described themselves as clinically busy and conduct their practice in a referral hospital, suggesting that they encounter the diverse range of pathology about which they were queried. Rather than polling policy makers, hospital administrators, and nonclinical neurosurgeons, we have captured answers from surgeons on the ground and in the operating theaters treating patients. Next, we asked not what surgeons do, but rather what they believe should be done if traditional barriers to optimal care were lifted. The result is an estimation of the optimized health care environment we seek, void of the pitfalls and inef-ficiencies that may be present in one’s own hospital, city, or country. Admission and surgical logs, country-wide hospital surveys, and even national billing data cannot capture this entity. Finally, we supplemented each discrete estimate with a surgeon confidence level, thereby add-ing an additional measure of data validation.3 There was remarkable concordance between surgeons from every background with regard to each disease entity. Regardless of their practice setting and training background, neuro-surgeons generally share similar sentiments regarding the degree to which neurological conditions demand neuro-surgical expertise.

The strong concordance among neurosurgeons from varying geographic regions and income levels lends cre-dence to our results. Furthermore, by obtaining proportion

figures not only for operative intervention but also for neu-rosurgical consultation, our results capture a more com-plete appraisal of the essential role played by neurosur-geons in health care systems worldwide. The importance of these results resides in their implication for research-ers, policy makers, and capacity strengthening personnel to extrapolate these figures into an overall neurosurgical disease burden estimate. To this end, this work represents an essential step in the epidemiological quest to accurately define the global neurosurgical volume. Until the scope of the problem is reliably quantified, public health measures to improve worldwide access to neurosurgical care cannot be efficiently and responsibly directed.

ConclusionsMore than 80 surgeons from around the world esti-

mated with confidence and strong concordance that surgi-cal intervention is indicated in more than 78% of cases of CNS tumors, hydrocephalus, and neural tube defects. While frequently mandating expert neurosurgical con-sultation, cases of epilepsy, stroke, TBI, and degenerative spine disease require surgical intervention in far fewer in-stances. These estimates can be used by researchers and health care officials worldwide to quantify the burden of neurosurgical disease and the workforce required to ad-dress this need.

AcknowledgmentsWe thank Charles Howard for his assistance in identifying

potential study participants. We acknowledge the Vanderbilt Medi-cal Scholars Program for providing Abbas Rattani with support on this project.

Grant funding support from NCATS/NIH UL1 TR000445 for Vanderbilt REDCap (M.C.D.). Grant funding support from GE Foundation Safe Surgery 2020 and the Steven and Carmella Kletjian Foundation (M.G.S.).

References 1. Alkire BC, Raykar NP, Shrime MG, Weiser TG, Bickler SW,

Rose JA, et al: Global access to surgical care: a modelling study. Lancet Glob Health 3:e316–e323, 2015

2. Anderson JE, Erickson A, Funzamo C, Bendix P, Assane A, Rose J, et al: Surgical conditions account for the majority of admissions to three primary referral hospitals in rural Mo-zambique. World J Surg 38:823–829, 2014

3. Aspinall W: A route to more tractable expert advice. Nature 463:294–295, 2010

4. Feigin VL, Forouzanfar MH, Krishnamurthi R, Mensah GA, Connor M, Bennett DA, et al: Global and regional burden of stroke during 1990–2010: findings from the Global Burden of Disease Study 2010. Lancet 383:245–254, 2014

5. Haglund MM, Kiryabwire J, Parker S, Zomorodi A, Ma-cLeod D, Schroeder R, et al: Surgical capacity building in Uganda through twinning, technology, and training camps. World J Surg 35:1175–1182, 2011

6. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG: Research electronic data capture (REDCap)—a metada-ta-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 42:377–381, 2009

7. Lo A, Polšek D, Sidhu S: Estimating the burden of neural tube defects in low- and middle-income countries. J Glob Health 4:010402, 2014


M. C. Dewan et al.


8. Meara JG, Leather AJM, Hagander L, Alkire BC, Alonso N, Ameh EA, et al: Global Surgery 2030: evidence and solutions for achieving health, welfare, and economic development. Lancet 386:569–624, 2015

9. Miranda-Filho A, Piñeros M, Soerjomataram I, Deltour I, Bray F: Cancers of the brain and CNS: global patterns and trends in incidence. Neuro Oncol 19:270–280, 2017

10. Nayar V, Kennedy A, Pappas J, Atchley KD, Field C, Smathers S, et al: Improving cardiac surgical site infection reporting and prevention by using registry data for case as-certainment. Ann Thorac Surg 101:190–199, 2016

11. Park KB, Johnson WD, Dempsey RJ: Global neurosurgery: the unmet need. World Neurosurg 88:32–35, 2016

12. Pugely AJ, Martin CT, Harwood J, Ong KL, Bozic KJ, Cal-laghan JJ: Database and registry research in orthopaedic surgery: part I: claims-based data. J Bone Joint Surg Am 97:1278–1287, 2015

13. Punchak M, Hung YC, Rattani A, Sachdev S, Dewan M, Park K: Current unmet neurosurgical workforce need in Africa and Latin America. Neurosci Colomb 23:294, 2016 (Ab-stract)

14. Punchak M, Mukhopadhyay S, Sachdev S, Hung YC, Peeters S, Rattani A, et al: Neurosurgical care: availability and ac-cess in low- and middle-income countries. World Neurosurg 112:e240–e254, 2018

15. Shrime MG, Bickler SW, Alkire BC, Mock C: Global burden of surgical disease: an estimation from the provider perspec-tive. Lancet Glob Health 3 (Suppl 2):S8–S9, 2015

16. Warf BC: “Who is my neighbor?” Global neurosurgery in a non-zero-sum world. World Neurosurg 84:1547–1549, 2015

DisclosuresThe results outlined in this manuscript do not necessarily reflect the official opinions or policies of the WHO. Dr. Haglund has received support from NuVasive for non–study-related clinical or research effort.

Author ContributionsConception and design: Dewan, Rattani, Shrime. Acquisition of data: Dewan, Rattani. Analysis and interpretation of data: Dewan, Rattani, Baticulon, Faruque, Haglund, Shrime. Drafting the arti-cle: Dewan. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Dewan. Statis-tical analysis: Dewan, Faruque. Administrative/technical/mate-rial support: Rattani, Baticulon. Study supervision: Park, Warf, Shrime.

Supplemental Information Online-Only ContentSupplemental material is available with the online version of the article.

Supplemental Materials. https://thejns.org/doi/suppl/10.3171/ 2017.10.JNS17347.

CorrespondenceMichael C. Dewan: Vanderbilt University Medical Center, Nash-ville, TN. [email protected].




operative and consultative proportions of neurosurgical

Documents