wisconsin spine outcome study–pilot: preliminary data · neurosurg focus / volume 33 / july 2012...

Neurosurg Focus / Volume 33 / July 2012

Neurosurg Focus 33 (1):E15, 2012

1

Low-back pain and associated lumbar spine disor-ders are a major social, political, economic, and medical problem in the US. Lumbar spine disor-

ders are a significant cause of compensable disability in North America, with estimated costs to American society

running in the hundreds of billions of dollars, depend-ing on the measures used.1–4,9,10 There is little high-quality clinical research guiding treatment of patients with de-generative lumbar disorders.5,11

Although RCTs represent the gold standard for eval-uating an intervention, the heterogeneity of spinal dis-orders and the desire for patients to participate in treat-ments considered effective based on established practice patterns make this form of evaluation ineffectual.12–15 An alternative to RCTs is prospectively obtained databases or patient registries. In April 2007, the Agency for Health-care Research and Quality published guidelines regard-ing the evaluation of established medical procedures;7

Wisconsin Spine Outcome Study–Pilot: preliminary data

Basheal M. agrawal, M.D., NathaNiel P. Brooks, M.D., aND DaNiel k. resNick, M.D., M.s.Department of Neurological Surgery, University of Wisconsin Hospital and Clinics, Madison, Wisconsin

Object. Given the pragmatic difficulties in developing randomized controlled trials in patients with disorders of the spine, the Wisconsin Spine Outcome Group has adopted the use of a prospective registry design to perform comparative effectiveness research on treatments for degenerative lumbar disorders. The goal of the Wisconsin Spine Outcome Study–Pilot (WISPOS-P) was to establish a Web-based, Health Insurance Portability and Accountability Act–compliant registry and to implement a patient registration paradigm that demonstrates at least 80% compliance in collecting pre- and posttreatment data in patients with lumbar disorders, regardless of the treatment they receive. The primary outcome measures were the percentage of patients with lumbar spine disorders who completed a Web-based survey preappointment, and at 1 and 3 months postappointment; the percentage of patients receiving a physi-cian-assigned diagnosis in the registry; and the success of electronic data transition from the Web-based interface to a locally controlled registry.

Methods. The WISPOS-P uses a prospective, diagnosis-based registry design. A universally accessible and secure Internet-based data management platform was created that accrues self-entered patient data on validated disability indi-ces, including the visual analog pain scale, Oswestry Disability Index (ODI), and the 36-Item Short Form Health Survey questionnaire. Data were obtained on patients, preappointment and at 1 and 3 months postappointment, regardless of the treatment rendered. A physician-entered diagnosis was assigned to each patient for data stratification.

Results. One hundred patients were invited into the WISPOS-P; 90 patients participated, and 10 withdrew for various reasons. Eighty-eight of 90 patients were assigned a diagnosis by the evaluating physician. Preliminary and qualitative assessment of the data shows that the major difference between patients who withdrew from the study and those who participated was the number of days between study invitation and clinic appointment (median 11 vs 20.5 days, respectively). In evaluating patients by mode of survey completion, the 2 largest groups were those who completed their intake forms electronically before their clinic appointment and those who used the paper format. The median age of patients electronically completing this survey was 14.34 years younger than those using the paper format. A significantly higher proportion of patients who completed their forms electronically had listed an email address. The 3 major diagnoses were disc disease (32 patients), stenosis (24 patients), and nonsurgical pain of spinal origin (14 patients). Patients with stenosis were older than those in the other 2 groups. Patients with nonsurgical pain of spinal origin had lower ODI scores compared with the other 2 groups.

Conclusions. A diagnosis-based registry design is effective in collecting pretreatment data for patients with lum-bar disorders. When stratified by diagnosis, comparative effectiveness analyses can be performed to identify optimum treatments for lumbar disorders given individual patient characteristics. The WISPOS-P has established a mechanism and proof of principle for the participation of patients in an outcomes registry.(http://thejns.org/doi/abs/10.3171/2012.4.FOCUS1297)

key worDs • degenerative lumbar disorder • outcomes registry • comparative effectiveness research

1

Abbreviations used in this paper: HIPAA = Health Insurance Por-tability and Accountability Act; IRB = institutional review board; MCS = Mental Component Summary; ODI = Oswestry Disability Index; PCS = Physical Component Summary; RCT = randomized controlled trial; SF-36 = 36-Item Short Form Health Survey; VAS = visual analog pain scale; WISPOG = Wisconsin Spine Outcome Group; WISPOS-P = Wisconsin Spine Outcome Study–Pilot.

Unauthenticated | Downloaded 05/26/21 04:52 AM UTC

B. M. Agrawal, N. P. Brooks, and D. K. Resnick

2 Neurosurg Focus / Volume 33 / July 2012

these guidelines describe the use of registries for the determination of “comparative effectiveness” for differ-ent treatments used for similar disorders. The guidelines specifically describe how the use of valid, responsive, and reliable outcomes measures in a broad patient population may provide high-quality information regarding the rela-tive efficacy of different procedures for similar diagno-ses, as well as describing differences in patient popula-tions with similar diagnoses who are selected for specific treatments.7

The WISPOG represents a collection of spine care providers seeking to establish a mechanism by which the comparative effectiveness of a variety of currently preva-lent treatments for low-back disorders may be assessed in the Wisconsin population. There is no doubt that de-termining a pathoanatomical diagnosis for patients with lumbar disorders can be challenging; for the purposes of the WISPOG, patients with low-back disorders and de-generative lumbar spine disease are defined as individu-als presenting with back and/or leg pain not attributable to trauma, infection, or neoplasm. The WISPOG’s overall objective is to use a prospective Web-based registry to accrue outcomes data for numerous lumbar spine disor-ders in a focused geographical region; that is, the state of Wisconsin. Currently the WISPOG consists of attending surgical spine physicians from the departments of neuro-surgery and orthopedic surgery at the University of Wis-consin (Madison, Wisconsin). When fully operational, the WISPOG will include a representative cross-section of spine care providers across the state, including surgical spine physicians, medical doctors, medical spine physi-cians, and associated midlevel providers. By acquiring data on a variety of patient types who receive care at mul-tiple sites and with multiple providers, we hope to make conclusions more generalizable.

In this article, we report preliminary results from an ongoing pilot study, the WISPOS-P. The WISPOS-P is a feasibility study occurring at a single major spine center in Madison, Wisconsin. The goal of the WISPOS-P is to establish a Web-based, HIPAA-compliant registry and to implement a patient registration paradigm that dem-onstrates at least 80% compliance in collecting pre- and posttreatment data for patients with lumbar disorders. Understanding the importance of diagnostic stratifica-tion, the registry also includes a physician-assigned di-agnosis for each enrolled patient. The primary end points for this pilot study are as follows: 1) the percentage of patients with lumbar spine disorders who completed a Web-based survey preappointment and at 1 and 3 months postappointment; 2) the percentage of patients receiving a physician-assigned diagnosis in the registry; and 3) suc-cess of electronic data transition from the Web-based in-terface to a locally controlled registry.

MethodsThe WISPOS-P uses a prospective, diagnosis-based

registry design.

Data InterfaceA universally accessible and secure Internet-based

data management platform was created that accrues self-entered patient data on validated disability indices, in-cluding the VAS, ODI, and SF-36 questionnaire. Scoring software for the SF-36 was purchased through Quality Metrics. Follow-up patient data can be acquired in regu-lar intervals. This information is automatically exported to a HIPAA-compliant, IRB-approved lumbar pathology registry.

Regarding security, the database is physically housed in a secure server room that can only be entered via a 2-fac-tor security protocol (keycode and card). The database it-self is protected via a secure login and secure socket layer (SSL).

The data interface contains 119 questions (Table 1) and serves to replace the standard intake form that pa-tients would complete prior to their consultation with a spine-care provider. The initial patient responses serve the dual purpose of obtaining pretreatment data in elec-tronic format, as well as providing the physician with standard “history of present illness” information, includ-ing medical history, allergies, medications, social history, and family history. The data interface has follow-up sur-veys focusing on treatments rendered, and again acquires VAS, ODI, and SF-36 scores.

The interface has a physician-accessible section where each patient can be assigned a diagnosis based on a modification of the Glassman diagnosis scheme.6 Diag-noses were subdivided into 3 general categories with 10 potential options: 1) primary surgical cases (disc disease, spondylolisthesis, instability, stenosis, scoliosis); 2) revi-sion cases (nonunion, adjacent-level degeneration, post-discectomy revision); and 3) nonsurgical cases (nonsurgi-cal pain of spinal origin, other).

Institutional Review BoardThe protocol for the WISPOS-P was submitted to the

minimal risk health sciences IRB in November 2010 and was approved in August 2011. Eight modifications to the protocol were required prior to approval. A work group consisting of hospital attorneys, privacy officers, and in-

TABLE 1: Questions included in the data interface for the WISPOS-P registry

No. of Questions Focus of Questions

10 identifying/demographic information25 symptomatology36 SF-36 questionnaire10 ODI1 VAS1 (15 options) previous treatments2 medical history

14 review of systems4 secondary gain2 allergies/medications5 family history9 social history

119 total




3

formation technology staff was required to approve the “legality of [this] study.” Consent to use the data interface is obtained electronically, and consent to transfer infor-mation from the data interface to the WISPOS-P registry is obtained in writing.

Patient RecruitmentPatients were recruited from the major spine center

associated with the University of Wisconsin (Research Park Spine Clinic, Madison, Wisconsin). Inclusion crite-ria included new patients to the spine center, ages 18–80 years, with symptomatic degenerative lumbar spine disor-ders. Exclusions included patients refusing to participate and pregnancy.

By focusing on patients seen at spine centers, the pa-tient population includes those individuals with “routine” back pain, as well as patients with more severe, chronic complaints. The vast majority of patients with degenera-tive low-back disorders seen at spine centers are seen in consultation and do not receive definitive care at the spine centers. For example, in fiscal year 2007, 5931 new-patient encounters occurred at the Research Park Spine Clinic, with 4040 encounters due to degenerative disease of the low back. Fewer than 500 of these patients under-went surgery, and the vast majority received an opinion and returned to their primary care physician for further care. Therefore, the majority of the data collected in this effort will reflect on the outcomes of patients who do not receive definitive care at specialty clinics, but who re-ceive care in the community.

Registry Protocol Clinic schedulers were asked to set aside new-patient

packets for participating spine care providers at the Re-search Park Spine Clinic. A research coordinator would screen these packets daily to identify patients meeting the study’s inclusionary criteria (Fig. 1). For those patients meeting criteria, a study introduction letter was mailed directing patients to the secure data interface. The study introduction letter also included the patient-intake survey in paper format. Of note, the questions in the data inter-face were identical to the paper intake survey.

The research coordinator contacted each patient via telephone to answer questions about the study and to help patients complete their survey. Patients were given the following options regarding the data interface: electronic completion prior to clinic, survey completion over the telephone, electronic completion during clinic visit, paper completion (research coordinator transfers data to regis-try), or refusal to participate (paper completion without transferring data to registry). Patients were given the op-tion to withdraw from the study at any time.

The research coordinator was present for each par-ticipant’s visit. The coordinator would print the informa-tion obtained from the data interface and add it into the patient’s chart prior to the patient’s consultation. The co-ordinator also obtained paper consent to transfer data into the WISPOS-P registry. After a physician evaluated the patient, the research coordinator obtained the single diag-nosis most likely accounting for the patient’s pain pattern. This diagnosis was added to the registry.

The research coordinator contacted patients at 1 and 3 months postappointment, either reminding them to com-plete the follow-up survey electronically or giving them the option to complete the survey over the telephone. These data were then transferred to the registry and correlated with the patient’s original responses.

ResultsPatients were recruited from a single spine center in

Madison, Wisconsin. Four neurosurgeons and 1 orthope-dic surgeon participated in the WISPOS-P. After IRB ap-proval, the first patient gave consent in September 2011. The goal sample size for this pilot project was 100 patients (Fig. 2). The sample size represents a procedural goal and will help to determine sample sizes more accurately for follow-up studies. The sample size was attained after a 4.5-month registration period. During this time period, 189 eligible patients were evaluated through the spine center. Of these patients, the clinic schedulers mailed appointment letters to 80 prior to screening by the research coordinator, and 9 patients were invited into the study but did not show up for their clinic appointment. These 89 patients were ex-cluded from the sample. One hundred patients were invited into the WISPOS-P and arrived at their clinic appointment. Ninety of these patients consented and participated in the preappointment survey, whereas 10 patients withdrew from the study. Reasons for withdrawal included time concerns, lack of interest in a study, not being offered surgery, secu-rity concerns, and other.

The 90 patients who participated were subdivided into groups based on method of completion of the preap-pointment survey. There were 4 subgroups, as follows: 1) electronic completion prior to clinic visit (41 patients); 2) telephone completion prior to clinic visit (3 patients); 3) completion using clinic computer just prior to appoint-ment (2 patients); and 4) paper completion, with the co-ordinator converting data into electronic format (44 pa-tients). The reasons for paper completion are listed in Table 2.

The median age of participating patients was 55 years, and the median age of patients who withdrew was 52.69 years. When evaluating the 2 largest subgroups of participating patients, those completing their intake forms prior to clinic and those completing the paper forms, the median ages were 44.72 and 59.06 years, respectively. The median time to complete the preappointment survey (Fig. 3) was 23.69 minutes (range 12.1–177.75 minutes). Patients who completed their survey electronically before their clinic visit had a median time of completion of 28.08 minutes. Note that the time for the research coordinator to convert paper forms into electronic data is included in the median time of all patients.

Patients who participated in the study were sent their study introduction packet a median of 20.5 days prior to their clinic appointment, whereas patients who withdrew were sent this material a median of 11 days prior to their appointment. Of the 90 participating patients, 63% listed an email address. Ninety percent of patients who com-pleted their intake form electronically prior to their clinic visit listed an email address, compared with 39% of pa-tients using the paper format.




Four standardized outcome indices were evaluated for each participating patient (Fig. 4): the SF-36 PCS score, the SF-36 MCS score, the ODI value, and the VAS score. The median score for all participating patients was a PCS of 29.41, an MCS of 47.15, an ODI of 45.5%, and

a VAS of 7.0. For those who completed their intake form electronically before clinic, the median PCS was 30.02, the MCS was 45.65, the ODI was 34.0%, and the VAS was 7.0. Those who completed the paper forms had a me-dian PCS of 27.92, an MCS of 48.13, an ODI of 53.0%, and a VAS of 7.0.

Of the 90 patients who participated in the study, a physician-assigned diagnosis was obtained in 88. Patients were evaluated using SF-36, ODI, and VAS scores, strati-fied by diagnosis (Fig. 5).

DiscussionAlthough advances in surgical treatment have dra-

matically advanced, there is a deficiency in clinical re-search guiding treatment of patients with lumbar dis-orders. Overall, this deficiency has caused skepticism

Fig. 1. Flowchart describing the WISPOS-P protocol.

TABLE 2: Reasons for completion of questionnaire on paper in 44 patients

Reason for Paper Form No. of Patients

ease of paper form 16no computer 14lack of computer skills 5technical difficulties w/ website 4other 3unknown 2




5

among patients, referring medical practitioners, and gov-ernment agencies.

The WISPOG was organized to help address these deficiencies. The overall objective of the WISPOG is to accrue outcomes data for numerous lumbar spine dis-orders and subsequently make recommendations for the management of these disorders in a patient-focused manner. Compared with other registries of patients with spine disorders, the WISPOS-P allows for the evaluation of both surgical and nonsurgical interventions, because patients are followed regardless of the treatment they re-ceive, thus allowing for true comparative effectiveness research.

Given the pragmatic difficulties in developing RCTs in patients with disorders of the spine, the WISPOG has

adopted the use of a prospective registry design. Al-though registries are becoming an increasingly common tool in medicine, they are relatively new in the evaluation of spinal disorders. Studying the feasibility of implement-ing a registry is extremely important, and is rarely done. Identifying the goals of a registry is critical to its suc-cess. The WISPOS-P is a feasibility study determining if a prospective registry for lumbar spine disorders will have high enough compliance among patients and physi-cians to be useful. The main purpose of this pilot project is to establish a mechanism and proof of principle for the participation of patients in an outcomes registry.

In addressing the primary end points, the WISPOS-P had a 90% success rate for patient participation in pre-appointment surveys. One-month and 3-month surveys are still occurring. Ninety-seven percent of patients re-ceived a physician-assigned diagnosis. The WISPOS-P also demonstrated the ease of data transition from a Web-based interface to a locally controlled registry. The benefits of having an appropriately designed Web-based interface include the following: minimization of extrane-ous information, standardization of data, convenience of data entry, and universal access to the data for extended statistical analyses.

Preliminary and qualitative assessment of the data shows that the major difference between patients who withdrew from the study and those who participated was the number of days between study invitation and clinic appointment (11 vs 20.5 days). In evaluating the partici-pating patients by mode of survey completion, the 2 larg-est groups were those who completed their intake forms electronically before their clinic visit and those who used

Fig. 2. Flowchart describing demographic characteristics of invited patients as a function of mode of completion of survey. Appt. = appointment.

Fig. 3. Bar graph showing median time for survey completion as a function of mode of completion of survey.




the paper format. The median age of patients who com-pleted this survey electronically was 14.34 years younger than those using the paper format. A significantly higher proportion of patients who completed their forms elec-tronically had listed an email address. Additionally, these patients had a lower score on the ODI. This may be a reflection of the general health of a younger individual.

Diagnostic specificity is critical in building an im-proved evidence base for lumbar disorders.8 Pain pattern-ing allows for data stratification in more clinically appli-cable groups. Of the participating patients, nearly 98% received a physician-assigned diagnosis. The 3 major diagnoses were disc disease (32 patients), stenosis (24 pa-tients), and nonsurgical pain of spinal origin (14 patients). Patients with stenosis were older than those in the other 2 groups. Patients with nonsurgical pain of spinal origin had lower ODI scores compared with the other 2 groups.

Although the original goal of the study was to have

physicians enter the diagnosis directly into the registry, this rarely occurred—the research coordinator would obtain the diagnosis for each patient and enter this information into the registry. Although there was a high compliance in assigning a diagnosis, physicians found the potential diag-noses to be confusing and artificial. Especially confusing for physicians, as well as for the purposes of performing comparative effectiveness research, was the breakdown of diagnoses into primary surgical, revision surgical, and nonsurgical cases. As an example, although a patient may present with disc disease (a primary surgical diagnosis), this patient may not have a surgically remediable patho-logical entity, may not be offered surgery given other con-founding factors, or may choose a nonsurgical treatment. Additionally, assigning diagnoses to patients with “non-surgical” pain was more challenging. The diagnoses of “nonsurgical pain of spinal etiology” versus “other” were meant to differentiate those patients presenting with unde-

Fig. 4. Flowchart describing variations in outcome indices as a function of mode of completion of survey.




7

fined pain of the back or leg from those with back or leg pain arising from hip disorders, sacroiliac joint pain, knee disorders, and so on. “Nonsurgical pain of spinal etiology” was used as a catchall for patients who could not be classi-fied with a more definitive diagnosis. Unfortunately, these diagnoses were imprecise, and thus would cause issues in comparative effectiveness research. Further efforts will use a diagnostic matrix in attempts to improve the accuracy of diagnosis and better describe patient characteristics.

Regarding the data interface, the original goal was to obtain a volume of information that could subsequently be analyzed for trends, while also providing the necessary components of a clinic history and physical. Therefore,

the preappointment survey consisted of 119 questions. It should be noted that a majority of this information was already collected from patients as the standard of care in their medical screening forms.

We originally intended this survey to have an average completion time of 15 minutes. The median time for indi-viduals completing this survey electronically prior to their visit was 28.08 minutes. Patients did comment that many questions seemed redundant, and there were notable com-plaints about the length and burden of the SF-36 survey. Accordingly, in the next iteration, we will probably change from the SF-36 to the 5-item EuroQol-5D health survey. We will remove the questions on symptomatology, because

Fig. 5. Flowchart describing variations in outcome indices as a function of diagnosis. Min = minutes; Pts. = patients.




these are better obtained during the clinic encounter, and their value is better expressed in the ODI. Additionally, we will condense portions of the medical history.

Regarding the efficiency of the study, a point of con-cern is the number of eligible patients who were missed by the clinic schedulers. For the registry to be successful, identifying potential participants is critical. We are work-ing with clinic managers to identify the reasons for this oversight. It is also important to identify the resources needed to perform such a study, especially when Inter-net access is available for patients. Because of constraints placed on the protocol by the IRB, every patient received the preappointment survey in paper format. In reviewing the reasons for paper completion (in 44 patients), 35 were associated with either the ease of paper completion or the lack of a computer and/or computer skills. We assume that eliminating this paper form would increase the num-ber of individuals who either complete the intake form electronically prior to the appointment or complete the intake form by using computers in the clinic with the re-search coordinator’s assistance. The caveat for in-clinic completion is that clinic flow is strongly dependent on the timeliness with which patients arrive to their appoint-ment.

Additionally, although having a research coordinator for this study was crucial, the question remains wheth-er this data interface could be managed solely by clinic staff. If the clinics were to adopt the data interface as the required method of data acquisition, establishing the in-clinic infrastructure needed for this registry, namely in the form of computers, would probably be more cost-effective than using a coordinator. Our experiences from the WISPOS-P, however, suggest that the accuracy and diligence with which data were collected could not be achieved without an individual dedicated to this data ac-quisition. Furthermore, we believe that compliance with follow-up surveys would be negatively affected without the efforts of a research coordinator.

Regarding the efficiency of starting such a study, in general we found the IRB and hospital approval process for essentially a quality-improvement project to be overly burdensome. The entire approval process took nearly 1 year; the protocol underwent 8 modifications and exten-sive reviews by hospital administration, attorneys, priva-cy officers, and information technology staff regarding the “legality of [the] study” in terms of the Department of Health and Human Services rules, including “HIPAA Privacy Rule Accounting of Disclosures Under the Health Information Technology for Economic and Clini-cal Health Act” [45 CFR Part 164].

There is a governmental push for registry-based com-parative effectiveness research, but independent IRBs are hampering the process by inaccurate interpretations of privacy rules, excessive requirements for consent, and in-efficient approval processes. In a recent communication with the Department of Health and Human Services, we were told that extensive IRB approval is only needed if “patient outreach/follow-up is not part of the existing or extended standard of care, but rather is initiated in part for data collection purposes for a clinical trial....” Further-more, “An enhancement or extension of an existing stan-

dard of care, including patient follow-up questions, would not fall within OHRP’s [Office for Human Research Pro-tections] regulations, if this change in care would have taken place regardless of any secondary purpose relating to the collection of information about that change.” Ad-ditionally, “regulations do not apply to the activities of … secondary analysis researchers [researchers receiving de-identified information], and there is no requirement for review and approval by an IRB….” (I. Pritchard [Senior Advisor to the Director, Office for Human Research Pro-tections], personal communication, December 2011).

The WISPOS-P represents the first cooperative project between multiple medical providers regarding the effective management of low-back pain and associated disorders in Wisconsin. Preliminary analysis of this feasibility study shows that using a Web-based platform is successful in obtaining pretreatment data. Barriers to participation have been identified, and will be eliminated in the next version of the WISPOS registry. One-month and 3-month follow-up data are being acquired in all participating patients, re-gardless of the treatment they are undergoing. In the future, we will expand the project to statewide private physician groups, to capture a better representative sample of patients with low-back disorders. This will allow us in addition to evaluate results stratified by the type of provider. We will also hopefully be able to address the proper use of referrals to spine centers and to make suggestions for treatment op-tions based on diagnosis.

ConclusionsThe establishment of a diagnosis-based prospective

registry infrastructure in which validated outcome indi-ces are used will have enormous value in the treatment of patients with lumbar spine disorders. A prospective reg-istry design is effective in obtaining baseline and follow-up data regardless of the prescribed treatment. Use of an Internet-based platform simplifies data acquisition and interpretation. The ease of data transfer from the Web-based platform to the registry minimizes human error and dispenses with unnecessary labor. Diagnostic speci-ficity facilitates data interpretation.

Disclosure

An AANS/CNS Joint Section on Disorders of the Spine and Peripheral Nerves Clinical Trials Grant was awarded to Dr. Agrawal in the amount of $50,000. Dr. Brooks is a consultant for Medtronic, Inc.

Author contributions to the study and manuscript prepara-tion include the following. Conception and design: all authors. Acquisition of data: all authors. Analysis and interpretation of data: Agrawal. Drafting the article: Agrawal. 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: Agrawal. Statistical analysis: Agrawal. Administrative/technical/material support: all authors. Study supervision: Agrawal, Resnick.

Acknowledgments

Special thanks to the research coordinator, Christine Agrawal, and the enrolling physicians, Gregory Trost, Paul Anderson, and Amgad Hanna.




9

References

1. Asche CV, Kirkness CS, McAdam-Marx C, Fritz JM: The so-cietal costs of low back pain: data published between 2001 and 2007. J Pain Palliat Care Pharmacother 21:25–33, 2007

2. Courtney TK, Matz S, Webster BS: Disabling occupational in-jury in the US construction industry, 1996. J Occup Environ Med 44:1161–1168, 2002

3. Dagenais S, Caro J, Haldeman S: A systematic review of low back pain cost of illness studies in the United States and inter-nationally. Spine J 8:8–20, 2008

4. Deyo RA, Mirza SK, Martin BI: Back pain prevalence and visit rates: estimates from U.S. national surveys, 2002. Spine (Phila Pa 1976) 31:2724–2727, 2006

5. Deyo RA, Nachemson A, Mirza SK: Spinal-fusion surgery—the case for restraint. N Engl J Med 350:722–726, 2004

6. Glassman SD, Carreon LY, Djurasovic M, Dimar JR, Johnson JR, Puno RM, et al: Lumbar fusion outcomes stratified by spe-cific diagnostic indication. Spine J 9:13–21, 2009

7. Gliklich RE, Dreyer NA (eds): Registries for Evaluating Patient Outcomes: A User’s Guide, ed 2. Rockville, MD: Agency for Healthcare Research and Quality, 2010 (http://www.effectivehealthcare.ahrq.gov/ehc/products/74/531/ Registries%202nd%20ed%20final%20to%20Eisenberg%209-15-10.pdf) [Accessed May 3, 2012]

8. Hall H, McIntosh G, Boyle C: Effectiveness of a low back pain classification system. Spine J 9:648–657, 2009

9. Halpern M: The costs of job accommodations for employees with low back pain. Work 21:271–278, 2003

10. Martin BI, Deyo RA, Mirza SK, Turner JA, Comstock BA, Hol-

lingworth W, et al: Expenditures and health status among adults with back and neck problems. JAMA 299:656–664, 2008

11. Resnick DK: Evidence-based spine surgery. Spine (Phila Pa 1976) 32 (11 Suppl):S15–S19, 2007

12. Weinstein JN, Lurie JD, Olson PR, Bronner KK, Fisher ES: United States’ trends and regional variations in lumbar spine surgery: 1992-2003. Spine (Phila Pa 1976) 31:2707–2714, 2006

13. Weinstein JN, Lurie JD, Tosteson TD, Hanscom B, Tosteson AN, Blood EA, et al: Surgical versus nonsurgical treatment for lumbar degenerative spondylolisthesis. N Engl J Med 356: 2257–2270, 2007

14. Weinstein JN, Tosteson TD, Lurie JD, Tosteson AN, Blood E, Hanscom B, et al: Surgical versus nonsurgical therapy for lumbar spinal stenosis. N Engl J Med 358:794–810, 2008

15. Weinstein JN, Tosteson TD, Lurie JD, Tosteson AN, Hanscom B, Skinner JS, et al: Surgical vs nonoperative treatment for lumbar disk herniation: the Spine Patient Outcomes Research Trial (SPORT): a randomized trial. JAMA 296:2441–2450, 2006

Manuscript submitted March 15, 2012.Accepted April 25, 2012.Please include this information when citing this paper: DOI:

10.3171/2012.4.FOCUS1297. Address correspondence to: Basheal M. Agrawal, M.D., Depart-

ment of Neurological Surgery, 600 Highland Avenue, K4/822, Mad-ison, Wisconsin 53792. email: [email protected].


wisconsin spine outcome study–pilot: preliminary data · neurosurg focus / volume 33 / july 2012...

Documents