 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Sections of Neurology
b Medical Research
c Department of Radiology, Children's Mercy Hospitals and Clinics, Kansas City, Missouri
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS. Retrospective, cross-sectional analysis was conducted of neuroimaging rates for 725 children and adolescents who were aged 3 to 18 years with nonacute headache and normal neurologic examination and were evaluated in a single pediatric neurology clinic during study years 1992, 1996, 2000, and 2004. Following recommendations of current practice parameters, patients with conditions that justify consideration for neuroimaging (eg, progressive headache, abnormal neurologic examination) were excluded from this analysis. We recorded the origin of any neuroimaging request at the time of the clinic visit and any abnormal neuroimaging findings that led to major clinical consequences.
RESULTS. Overall, the mean rate of neuroimaging for patients with nonacute headache was 45%. Use rates remained steady during the 13-year study period (range: 41%–47%). The majority of neuroimaging studies were ordered originally by primary care providers. The proportion of neuroimaging studies that were ordered by primary care providers increased significantly from 1992 to 2004.
CONCLUSIONS. In the evaluation of patients who had nonacute pediatric headache and were referred to a child neurology clinic, neuroimaging-use rates remained stable during the past decade. An increasing proportion of neuroimaging studies are ordered by primary care providers. The influence of evidence-based medicine on medical decision-making may be partly responsible for curbing increases in neuroimaging overuse. The perceived value of neuroimaging by physicians and consumers deserves ongoing study.
Key Words: Diagnostic procedures • headache • neuroimaging • practice parameters/guidelines • practice-based research
Abbreviations: CT—computed tomography
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Headache is a relatively common symptom in children and adolescents. Headache pathogenesis is complex with many potential causes, but clinicians can generally distinguish primary headache (ie, pain without evidence of a serious underlying illness) from secondary headache (ie, pain resulting from a serious pathologic condition). Although most pediatric patients who complain of headache do not have a serious underlying disorder, neuroimaging can be a valuable diagnostic tool for conditions for which headache is coupled with other worrisome signs and symptoms.4 Practice parameters for the evaluation of children and adolescents with recurrent headaches, published in 2002, recommend that diagnostic neuroimaging be considered for children with an abnormal neurologic examination or other historical features that suggest neurologic dysfunction.5 These recommendations agreed with the previous practice parameters for the evaluation of headache in adults, published in 1994 and 2000, but emphasized that obtaining a neuroimaging study on a routine basis is not indicated for children with recurrent headache and a normal neurologic examination.4,6
Nevertheless, both overuse and underuse of neuroimaging have been topics of debate during the past 2 decades, when health care costs, malpractice liability, consumer expectations, and the number of neuroimaging facilities all have dramatically increased. The aim of this study was to assess diagnostic neuroimaging-use trends in a nonacute pediatric headache population before and after the publication of clinical practice parameters. We hypothesized that neuroimaging-use rates increased from 1992 to 2004.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The study was performed in a single-institution, regional children's hospital that serves a large metropolitan area. During the study period, the child neurology clinic was composed of 4 to 6 child neurologists, a nurse practitioner, and a pediatrician with special interest in pediatric headache. Two of the child neurologists worked continuously in the clinic throughout the entire study period. The clinic was the main resource for child neurology in the study metropolitan region, and there were no other child neurologists or pediatric headache specialists in private practice in the area. All caregivers in the clinic were salaried by the institution. The patient population comprised a mixed combination of Midwest urban, suburban, and rural children divided along all socioeconomic, racial, and health insurance groups. This study was approved by the institutional review board.
Inclusion Criteria
Patient charts were included in the study when the referral was made for outpatient evaluation of recurrent, nonprogressive headache with no recorded evidence of serious underlying medical conditions or remarkable abnormalities on physical examination.
Exclusion Criteria
According the recommendations of the current clinical practice parameter, we identified a priori conditions and risk factors for which "neuroimaging should be considered."5 Specifically, patients were excluded from the study when they had historical evidence of progressive headache, meningitis, stroke, vascular malformation, neurocutaneous syndrome, hydrocephalus, traumatic brain injury, seizures, a significant neurodevelopmental disorder, brain tumor, other cancer, or major illness. Patients were excluded when abnormal neurologic examination findings were documented (eg, focal deficits, signs of increased intracranial pressure, alteration of consciousness, macrocephaly [head circumference >98th percentile], and microcephaly [head circumference <2nd percentile]).
Data Collection
No patients were contacted or interviewed by data gatherers. We recorded the origin of the neuroimaging request at the time of the clinic visit. Historical, physical examination and neuroimaging data were abstracted from the medical chart. Categorically recorded (normal/abnormal) historical data included height, weight, blood pressure, and diet, as well as significant past medical, family, school and social histories; review of systems; and general physical and neurologic examinations. Other recorded data included age, gender, current medications, and family history of migraine, as well as a synopsis of headache chronology, location, severity, and associated symptoms.
In a separate data file, diagnostic neuroimaging data were recorded as CT, MRI, both, or neither. We further analyzed the frequency and types of neuroimaging findings but focused on study years 2000 and 2004 because of easier access to electronic medical charts and picture archive and communication systems neuroimaging. Neuroimaging reports by the neuroradiologist or pediatric radiologist were recorded as "normal," "abnormal," or "remarkable." Abnormal neuroimaging results were defined as pathologic neuroanatomic findings that resulted in a new primary clinical diagnosis, prognosis, or significant change in the patient's clinical management. Remarkable neuroimaging findings were any comments in the summary of the neuroradiology report of unexpected, probable incidental findings that the study child neurologists and pediatric neuroradiologist believed to be unrelated to the presenting headache symptoms.
Charts were abstracted by a medical student and child neurology fellow with close supervision by the senior child neurologist. Abstracted charts of all patients for whom there was a historical question of progressive headache or possible physical examination abnormality were independently reviewed by 2 child neurologists who were blinded to any neuroimaging data. Any abnormal or remarkable neuroimaging findings were reviewed by the pediatric neuroradiologist.
Major Outcome Variables
We assessed the diagnostic neuroimaging-use rates from the 4 study years during the 13-year period and the proportion of normal, remarkable, and abnormal neuroimaging results.
Statistical Analysis
Statistical analyses were conducted by using SPSS 12 (SPSS Inc, Chicago, IL). Median values and associated ranges were reported for nonnormal data. We used 95% confidence intervals where appropriate to summarize the uncertainty associated with sampling error.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Trends in Neuroimaging Use
Of 725 total charts of patients with nonacute headache analyzed, 325 (45%) had at least 1 neuroimaging study. Neuroimaging-use rates ranged between 41% and 47% without discernible trends throughout the 4 main study years of the 13-year period. As in previous studies of neuroimaging for nonacute headache, no patient had abnormal neuroimaging findings that led to a neurosurgical procedure or other major change in diagnosis or clinical outcome. In this study, the majority (75.7%) of neuroimaging tests were ordered by primary care providers. The proportion of test orders that originated from primary care providers increased significantly during the course of this study (odds ratio: 1.40 [95% confidence interval: 1.10–1.78]; P = .006; Table 1).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Although the design of this study used the current practice parameters as a benchmark control, it could not be determined from these observational data whether referring physicians were aware or unaware of the 2002 practice guidelines or which "aware" physicians either followed or ignored the guidelines for any number of plausible reasons (Table 2). Nevertheless, this retrospective study reveals compelling data that deserves additional practice-based comparative effectiveness research. Furthermore, this study did not assess the outcomes of patients with progressive headache or those with abnormal neurologic examination findings for which neuroimaging should be strongly considered. According to the published practice guidelines and our exclusion criteria, all patients who were assessed in this study lacked apparent clinical indications for neuroimaging. As in previous neuroimaging studies for nonacute pediatric headache without additional abnormal signs, this study found no clear pathologic results that led to a new primary clinical diagnosis, surgery, or other major change in patient treatment or prognosis.9–15 Thus, the findings in this study corroborate the conclusions and recommendations of the current clinical practice parameter.5 It is known that primary headache disorders, such as migraine and tension-type headaches, account for the vast majority of pediatric headaches. Neuroimaging is helpful in the diagnosis of secondary headache as a result of serious underlying illness that is reliably accompanied by a shorter symptom interval and other abnormal clinical features by the time of presentation or reevaluation.16,17 It is possible that evidence-based clinical practice parameters have played an important role in educating physicians and in helping to curtail an otherwise increasing trend to use neuroimaging. Future studies will be needed to demonstrate clearly the influences of such practice guidelines.
|
Limitations to this study are mostly related to its retrospective, cross-sectional study design. Multiple variables may influence individual physician judgment and decision analysis around a point of time in a study. These study data were gathered in the sole venue for child neurology consultation in a diverse community, but there was no analysis of cost, patient income, insurance status, patient expenditure, educational status, geographic origin, clinic access time, or appointment time; however, overuse of neuroimaging has been reported in child neurology clinics in other geographic regions, and it is unlikely that the findings reported here are unique to the study hospital in the Midwestern United States.21
These study results cannot project an estimate of neuroimaging-use rates in the larger population of patients who have nonacute headache in primary care practice and are not referred to a local child neurology clinic. Nevertheless, this study shows the odds that primary care provider–ordered neuroimaging for nonacute headache rose from 
2-to-1 in 1992 to 7-to-1 in 2004 (Table 1). These findings strongly suggest a trend of increasing neuroimaging use in the primary care setting. Additional study is needed to evaluate the awareness of practice parameters among primary care and specialist providers as well as the utility of such practice guidelines in various clinical practice settings.
There are numerous concerns about the overuse of neuroimaging. Although the risk of ionizing radiation that are associated with CT scanning is generally considered rather small when compared with the benefits of accurate diagnosis and treatment, recent studies suggested that physicians are not sufficiently aware of the potential long-term health risks that are associated with the radiation from CT imaging, which is highest in children.22–25 Sixty-two percent of the neuroimaging patients in this study had CT alone or CT plus MRI.
As in the application of many technologies, the value of neuroimaging cannot be judged solely on how it is used to diagnose disease in populations of patients. From formal evidence for nonacute headache, the practice of obtaining a neuroimaging study in an otherwise neurologically healthy individual should now seem to have little value; however, clinical studies that are used as a basis for practice parameters generally do not evaluate variables such as patient preference, physician perception of complexity, or medicolegal perspectives of health care standards.26,27 Furthermore, access to medical technology such as diagnostic imaging may have become a proxy for health care quality as measured by patient-perceived value. Reassurance for consumers and physicians might be evolving from the value of examination and clinical judgment into "imaging for problems imagined."28,29 Future studies of neuroimaging use could attempt to measure the perceived value of imaging procedures for various symptoms from the perspective of various health care stakeholders.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank Gloria Womelduff, Amy Wolf, and Benjamin C. Williams for assistance with data retrieval.
|
FOOTNOTES |
|---|
Address correspondence to William D. Graf, MD, Children's Mercy Hospitals and Clinics, 2401 Gillham Rd, Kansas City, MO 64108. E-mail: wdgraf{at}cmh.edu
Statistical analysis was performed by Dr Simon.
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject Neuroimaging recommendations in the evaluation of children and adolescents with recurrent headache have been established in clinical practice guidelines. Little is known about how these guidelines influence neuroimaging use in contemporary clinical practice.
|
| What This Study Adds We found stable neuroimaging-use rates, rampant neuroimaging overuse, and a shift toward primary care neuroimaging in the evaluation of nonacute pediatric headache during 1992–2004. Neuroimaging practice patterns seem to be influenced by evidence-based clinical practice guidelines and other value-based perceptions.
|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2008 American Academy of Pediatrics. All rights reserved. |
||
|
||
