BACKGROUND: Ventricular shunt complications in children can be severe and life-threatening if not identified and treated in a timely manner. Evaluation for shunt obstruction is not without risk, including lifetime cumulative radiation as patients routinely receive computed tomography (CT) scans of the brain and shunt series (multiple radiographs of the skull, neck, chest, and abdomen).
METHODS: A multidisciplinary team collaborated to develop a clinical pathway with the goal of standardizing the evaluation and management of patients with suspected shunt complication. The team implemented a low-dose CT scan, specifically tailored for the detection of hydrocephalus and discouraged routine use of shunt series with single-view radiographs used only when specifically indicated.
RESULTS: There was a reduction in the average CT effective dose (millisievert) per emergency department (ED) encounter of 50.6% (confidence interval, 46.0–54.9; P ≤ .001) during the intervention period. There was a significant reduction in the number of shunt surveys obtained per ED encounter, from 62.4% to 5.32% (P < .01). There was no significant change in the 72-hour ED revisit rate or CT scan utilization rate after hospital admission. There were no reports of inadequate patient evaluations or serious medical events.
CONCLUSIONS: A new clinical pathway has rapidly reduced radiation exposure, both by reducing the radiation dose of CT scans and eliminating or reducing the number of radiographs obtained in the evaluation of patients with ventricular shunts without compromising clinical care.
- CT —
- computed tomography
- ED —
- emergency department
- ICD —
- International Classification of Diseases
- VP —
Evaluation of a child with a ventricular shunt presenting to the emergency department (ED) with symptoms of potential shunt malfunction presents a diagnostic challenge for the clinician. Ventricular shunt complications such as obstruction can be life-threatening, and they require diagnosis and treatment in a timely manner. However, differentiating shunt complications from other benign clinical conditions is challenging because signs and symptoms such as vomiting and headache are nonspecific. Evaluation for ventricular shunt malfunction in the ED typically includes brain computed tomography (CT) scanning and shunt series (multiple radiographs of the skull, neck, chest, and abdomen). Previous studies suggest that 15% to 25% of children evaluated radiographically for shunt malfunction in a pediatric ED go on to require surgical revision.1,2 Repeated evaluations may lead to substantial radiation exposure over a child’s lifetime.3,4
Existing literature has explored several methods to reduce radiation exposure for children with suspected ventricular shunt malfunction. Rapid cranial MRIs have been described as an alternative to CT scans.5–8 However, MRIs may not be readily available in all settings and may have other disadvantages, including cost and differences in visualization of the ventricles and catheter(s) in preparation for surgery. Another approach has been to reduce the radiation dose and/or coverage with CT scanning to limit evaluation to the ventricles.9,10 For the radiographic shunt series, the available literature demonstrates low yield of routine performance, with opportunities to reduce unnecessary radiation in this area.1,2
We developed a multidisciplinary quality improvement group to implement an institutional clinical pathway for evaluation of shunt complications with the goal of reducing radiation exposure. Interventions included a reduced-dose CT scan protocol specifically tailored for the evaluation of ventricular size and surgical navigation should operative revision be required. Specific indications for shunt radiographs were developed to discourage routine use of a shunt series.
The aim of the project was to standardize care and reduce radiation exposure for children and young adults requiring evaluation in the ED for ventricular shunt complications.
Our project focused on the care of patients with a ventricular shunt who present to the ED with signs or symptoms concerning for a shunt complication. The setting (ie, Children’s Hospital of Philadelphia) is an urban, tertiary care, pediatric academic ED with an annual volume of >90 000 patients, including an average of 330 patients presenting with “VP (ventriculoperitoneal) shunt” as part of their chief complaint.
A multidisciplinary team consisting of physicians and nurses across care teams representing all relevant stakeholders (emergency medicine, neurosurgery, critical care, neuroradiology, and general pediatrics) collaborated over the course of 6 months to develop a Web-based clinical pathway with the goal of standardizing the evaluation and management of patients with suspected shunt complication. The overall goal was to reduce unnecessary imaging and associated radiation exposure for evaluation of ventricular shunts. The team performed a detailed analysis of the current process and identified key drivers for improvement to be ordering practices of front line clinicians and radiation dosage of imaging studies.
Although a number of pediatric institutions have moved to using MRIs for the evaluation of shunted hydrocephalus,5–8 our team believed that the rapid and widespread availability of CT scans, ability to provide a volumetric data set for surgical navigation purposes in <1 minute of scanning time, and superior depiction of the intracranial and extracranial portions of the catheter favored implementing a lowered dose CT protocol. Our standard-dose brain CT scans already used a pediatric-specific protocol with a reference mA level of 250; lowering the reference mAs level to 150 was determined by consensus with the Division of Neuroradiology as reaching an acceptable balance of reduced dose and image quality. Given the substantial inherent difference between the CT density of the ventricular system and the brain parenchyma, this reduced dose would not adversely affect the ability to assess ventricular size. Sensitivity and specificity for brain parenchymal lesions (ie, acute ischemia) might be decreased; however, encouragingly, reconstruction methods have shown that even a dose reduction of 90% permits detection of hemorrhage.11
In addition, informal review of the value of the radiographic shunt series by the neurosurgeons determined that the rare contribution of the series to patient management did not warrant their routine use. Specific indications for shunt radiographs were developed to discourage routine use of the “shunt series.” Scout image(s) obtained to perform the brain CT scan provide at a minimum the lateral view similar to skull films of the catheter location in the brain. Instead of obtaining a full series of radiographs, we provided specific recommendations for single-view radiographs of the chest and/or abdomen, which included: (1) localized swelling or pain along the shunt tubing; (2) distal erosions (rare) with shunt tubing present outside the body; or (3) to assist with shunt revision at request of the neurosurgery team.
During development and at completion, the clinical pathway was presented at divisional meetings across subspecialties to advertise and seek feedback. With expert consensus and agreement of all important stakeholders, the pathway was implemented on a specific “Go Live” date and made easily accessible for care providers in our ED and inpatient units (Fig 1 [live pathway available at http://www.chop.edu/pathways]).
Clinical decision support was provided in the form of an electronic order set that included a specific order for the lowered dose CT imaging protocol in addition to the recommended single-view chest/abdomen radiograph. This tool made it easy to follow the recommendations of the pathway, but it did not inhibit clinicians from making different decisions if necessary based on a unique presentation.
In addition, several months after pathway implementation, metrics were displayed on computer screensavers, as well as larger screen monitors, throughout the ED to remind and encourage providers of the clinical pathway.
Study of the Intervention
The quality improvement team met twice monthly for 6 months, and then monthly thereafter, to monitor pathway implementation by reviewing data, to measure outcomes including radiation exposure, and to check for adverse outcomes.
The primary outcome measure was the reduction in radiation exposure per ED patient encounter. CT dose data were directly recorded for each scan, including the CT dose index per volume and dose–length product in milligrays per centimeter and converted to millisieverts. The average total number of radiographs per ED encounter was measured before pathway implementation and for 6 months after pathway implementation. The rate of shunt series was monitored throughout.
Balance measures included rate of CT scan utilization per ED encounter, 72-hour ED revisits, 72-hour ED revisits requiring shunt revision, and inpatient CT scan utilization.
To capture all patients with suspected shunt complication, we designed an analytic platform to pull data from the electronic health record and display the data on a weekly basis throughout the intervention. Criteria for inclusion in the database included:
ED chief complaint of “VP shunt”;
Encounters in which the “VP shunt order set” was ordered;
Encounters in which “CT brain for hydrocephalus” was ordered;
VP shunt in problem list and ED chief complaint concerning for a complication; and
VP shunt in problem list and VP shunt or the procedure identified by using International Classification of Diseases (ICD), Ninth Revision, and ICD-10 Revision, codes.
A time series design was used to study the impact of the interventions. Comparative data were available for 12 months before the start of the intervention and for 20 months after pathway implementation. Patient characteristics for prepathway and postpathway groups were reported, including the rate of ED visits associated with a shunt complication. Patient encounters with a shunt complication were identified by corresponding ICD-9, and ICD-10, diagnosis and procedural codes.
Our measures were analyzed continuously by using control charts with pulled data from our analytical platform. Special attention was paid to data points falling outside the upper control limit (3 σ) or the lower control limit (−3 σ). We also monitored for serial data points (≥7) above or below the mean. No interventions were made for anticipated, beneficial changes to the measured outcomes. Although none was identified, our team was prepared to investigate adverse changes to our balance measures that were believed to be a special cause of variability.
The reporting of the aggregated outcomes of this quality improvement project were reviewed by the Office of Clinical Quality Improvement and the Institutional Review Board at Children’s Hospital of Philadelphia and classified as nonhuman subjects research.
The prepathway and postpathway groups of patients analyzed were similar (Table 1), including the rate of ED visits associated with a shunt revision (32% prepathway and 35% postpathway; P = .33).
There was no significant change in the mean CT scan utilization rate before (59.7%) and after (66.6%) pathway implementation (Fig 2). There was a rapid increase in the use of lowered dose CT scanning for hydrocephalus over the first 2 months, after which the mean rate plateaued at 91.5%. The average CT effective dose with the lowered dose CT scan during the intervention period was 1.17 mSV, compared with 2.37 mSV (standard CT scan) before pathway implementation, resulting in a 50.6% (confidence interval, 46.0–54.9; P ≤ .001) mean reduction in effective radiation from CT scanning (Fig 3).
There was a significant reduction in the number of shunt surveys obtained per ED encounter, from 62.4% to 5.32% before and after pathway implementation (P < .01), respectively (Fig 4).
During the first 6 months after pathway implementation, patients received 2.1 (range, 0–8) radiographs per visit, a significant reduction from the routine shunt series, resulting in a 64.6% (confidence interval, 55.6–73.6; P ≤ .0001) mean reduction in effective radiation from radiographs per visit. No radiographs were obtained in 45 (46.9%) of 96 visits.
There was no significant change in 72-hour revisit rate to the ED (prepathway, 4%; postpathway, 5%; P < .13) (Fig 5).
The rate of patients returning to the ED within 72 hours of initial presentation and requiring operative shunt revision did not change after pathway implementation: prepathway, 1.35%; postpathway, 1.74% (P = .66).
There was no significant change in CT scan utilization after a patient was admitted to the hospital (prepathway, 21.5%; postpathway, 26.5%; P = .21). Of 596 patients who had a CT scan for hydrocephalus in the ED postpathway, 20 patients (3.3%) were admitted and had a routine CT scan performed during hospitalization. There have been no reports of inadequate patient evaluations or serious medical events.
Evaluation for shunt obstruction is not without risk, including lifetime cumulative radiation as patients routinely receive CT scans of the brain and shunt series.12,13 Initial experience with our new clinical pathway has significantly reduced radiation exposure without compromising patient care, both by reducing the dose of brain CT scans and significantly reducing the number of radiographs obtained in the evaluation of children and young adults with complications related to ventricular shunts.
Alternative neuroimaging techniques such as low-dose CT scans and rapid-sequence MRIs5–10 have been suggested as modalities for radiation reduction in this high-risk population. An MRI requires screening for contraindications, which can be time-consuming, potentially delaying diagnosis and care. In addition to being more expensive and less readily available, concerns about the accuracy of MRIs have been raised. Specifically, the ventricular catheter can be more difficult to visualize and may require the performance of additional MRI sequences, thereby increasing scan time and likelihood for motion.14,15 In addition, at our institution, neurosurgeons are using surgical navigation CT scans (incorporated into the initial low-dose CT scanning protocol) during operative procedures for improved operative safety. Furthermore, a novel application has been recently reported in which the fusion function of a neuronavigation system can be used to compare ventricular size, calculate ventricular volumes, and display a color-coded overlay of change (increases or decreases) between a previous and a current scan.16 Although this innovative method requires validation, it holds great promise.
On average, patients presenting to the ED with a concern for a shunt complication received 1.2 mSV less ionizing radiation from CT scanning after implementation of our clinical pathway. This level is equivalent to the radiation of ∼30 chest radiographs or one-half of a routine brain CT scan. Further reductions in radiation exposure were achieved by eliminating or substantially decreasing the number of radiographs obtained to assess shunt tubing. Because cumulative lifetime radiation exposure may be significant for patients with ventricular shunts, these reductions can lead to clinically significant outcomes over a patient’s life.3,4 Because lowered dose CT scanning is specifically used to assess ventricular size and not intraparenchymal pathology, we would not expect 100% of patients to have a lowered dose CT scan, as some may require routine brain CT scanning.
Our quality improvement project resulted in an immediate and significant improvement in patient care with no measured adverse outcomes. This result was accomplished by first defining a new process for evaluating patients with ventricular shunts, then by “hard-wiring” the changes with an electronic order set for clinical decision support. Unlike most quality improvement projects, this one did not require significant educational resources or change in behavior of care providers, which are often high hurdles. There was great value in using a multidisciplinary team in approaching our quality improvement project. Innovation occurred at the overlap of varying expertise and led to a unified goal of providing best solutions and processes for the patient.
This project was limited to patients evaluated in a pediatric ED at a large, urban, tertiary care center. We anticipate this research could be used in other similar clinical settings; however, it may not be universally generalizable. There are limitations and difficulty in accurately defining and capturing our patient population of interest. We monitored for adverse clinical outcomes mainly by using the ED CT scan utilization rate, 72-hour ED revisit rates, inpatient CT scan utilization rate, and intermittent review of the data with the goal of identifying inadequate imaging compared with routine CT scanning and shunt series before our interventions. Imaging quality is a challenging and subjective outcome to monitor, and we recognize this limitation.
This new clinical pathway rapidly lowered radiation exposure, both by reducing the radiation dose of CT scans and eliminating or reducing the number of radiographs obtained in the evaluation of patients with ventricular shunts without compromising clinical care. Further monitoring is needed to confirm the trends seen in our data. Future gains are expected with increased pathway compliance and further CT dose reduction strategies, particularly for more radiosensitive younger patients.
The authors acknowledge Arastoo Vossough, MD, PhD, and Winnie Zhu, PhD, in the Department of Radiology at the Children’s Hospital of Philadelphia for their assistance with CT dose implementation.
- Accepted January 12, 2017.
- Address correspondence to Ronald F. Marchese, MD, Department of Pediatrics, Children’s Hospital of Philadelphia, 3501 Civic Center Blvd, CTRB, 9th Floor, Philadelphia, PA 19104. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Kim I,
- Torrey SB,
- Milla SS,
- Torch MC,
- Tunik MG,
- Foltin JC
- Udayasankar UK,
- Braithwaite K,
- Arvaniti M, et al
- Gabriel S,
- Eckel LJ,
- DeLone DR, et al
- Miller JH,
- Walkiewicz T,
- Towbin RB,
- Curran JG
- Copyright © 2017 by the American Academy of Pediatrics