Urgent message: Although diagnosis of pediatric community-acquired pneumonia is not always straightforward, most cases can be diagnosed adequately by non-pediatricians and treated in the urgent care center, thereby reducing hospitalization for this condition.
By Deena R. Zimmerman, MD, MPH, IBCLC, Scott Fields, MD, Nahum Kovalski, BSc, MDCM
Introduction
Pediatric community-acquired pneumonia (CAP) is a common illness. Estimates are that the annual incidence of pneumonia per 1,000 children in North America ranges from 30 to 45 cases among those CAP is also a common diagnosis in urgent care. For example, 5.8% of all children under the age of 10 years who presented to Terem Emergency Medical Centers in Jerusalem, Israel, where the authors practice, were diagnosed with pneumonia in the 2007 calendar year.
As common as pneumonia is, there is not a clear consensus as to its diagnostic criteria. Some definitions rely on chest x-rays and require the presence of infiltrates.2 On the other hand, the World Health Organization has defined pneumonia, primarily for diagnosis in developing countries, solely on the basis of clinical findings obtained by visual inspection and timing of the respiratory rate.3
A recent Cochrane database has questioned the need for chest x-ray, claiming that such x-rays do not change the outcome in any way. In clinical practice, however, they are still widely performed as part of the evaluation of clinical syndromes suggestive of pneumonia. For the purposes of this article, pneumonia is defined as the presence of fever, acute respiratory symptoms (cough, dyspnea, tachypnea, abnormal lung examination), or both, plus evidence of infiltrates on chest radiography.
Background
CAP is almost always viral or bacterial in nature. On a practical level, no single feature has been shown to differentiate between these two etiologies. Several studies show that there are no radiologic features that can be reliably used to differentiate between these two major causes. In cases at either extreme of the spectrum, such as typical bronchiolitis with scattered infiltrates or dense lobar pneumonia with a large pleural effusion, the level of diagnostic certainty provided by findings on x-ray is high. In the majority of cases, though, attempts at differentiation based on chest x-rays alone is often incorrect.2 Laboratory tests are also of limited use in the separation of bacterial from viral causes.4
Due to the difficulty in proving the etiology, the decision to treat with antibiotics is one that requires careful judgment and the weighing of the data provided by the combination of clinical presentation and the findings of whatever radiologic and laboratory tests were performed.5
As mentioned, part of the clinical decision is often based on the reading of the chest x-ray. However, a number of studies have shown poor inter-rater reliability for chest x-rays in suspected cases of CAP, both in adults6,7 and children.8 Most of these studies compared radiologists with each other or with pediatricians. Since in most urgent care centers, pediatric cases are also managed by non- pediatricians, we felt it important to investigate differences in the reliability of x-ray interpretation between non-pediatricians and pediatricians, using one board-certified radiologist as the gold standard.
Once the decision has been made to treat, one is then left with the task of prescribing an appropriate antibiotic regimen. For most mildly ill patient, oral medication can be used.2,5,9 However, when the patient is moderately ill or vomiting and unable to tolerate oral medication, parenteral medication may be needed, at least for the first doses.
This interim group of patients – i.e. not ill enough to require hospitalization but too ill to be sent home with oral medication alone – are uniquely suited to urgent care. If a patient does not need supplemental oxygen, ongoing intravenous fluids, or inpatient monitoring, such a patient can benefit from ambulatory parenteral antibiotics.
This approach, of treating and keeping the patient in the community is a classic differentiator between urgent care and emergency medicine. Typically, it is not thought to be within the purview of an emergency department to provide for ambulatory repeat intravenous doses of any medication. As such, urgent care has a unique opportunity to reduce hospital visits and even a certain number of admissions. In light of this, we reviewed our experience with pediatric CAP patients who were appropriate for ambulatory IV antibiotic treatment.
Materials and Methods
Setting
Terem Emergency Medical Centers is a privately owned medical services company that establishes and manages freestanding urgent/emergent care clinics. The central clinic is open 24 hours per day, 365 days per year. Four other Terem clinics are open afternoon, evening, and weekend hours. All five clinics provide digitalized onsite radiology and laboratory testing that includes complete blood counts (CBCs) during all hours of operation. X-rays performed in the clinics are initially read by the treating physician. These x-rays are then reviewed by a board-certified radiologist within 24 hours. Discrepancies between the interpretation of the physician and the radiologist are referred to as “mismatches.” All such mismatches are further reviewed by a senior physician. This senior physician, at times in further consultation with other specialists, will decide if changes in the clinical management of the patient are needed.
Data source
Terem uses a proprietary electronic medical record (EMR) system called PARPAR to register, clinically manage, and administer all patient visits, laboratory tests, and radiological studies. Data captured by PARPAR include:
– the patient’s demographic information
– complaint on arrival
– diagnosis
– main procedures done
– referral decision made.
The digital radiological images are stored short term in an online picture archiving and communication system (PACS) and long term in a compressed JPEG format in an online, web-accessible archive. Printed materials, such as the handwritten doctor’s notes, are scanned into the system and attached to the matched patient EMR.
A unified “data warehouse” exists for housing the data from all Terem clinics. A set of web-based forms allows for the review of individual charts, as well as the extraction of summary statistics. For example, a reviewer can request a list of visits by patients of a certain age with a specified diagnosis. Each item on such a list includes a web-based link to the patient chart and x-ray reports associated with the given visit.
Furthermore, PARPAR allows for the integration of clinical decision support tools into the process of digital patient charting by the physician. Since the beginning of 2008, as part of a separate study, a clinical decision support “pop-up box” has appeared when children who presented with fever were seen by a Terem pediatrician. The pediatrician was required to digitally record the patient”s respiratory rate, oxygen saturation, and any specific respiratory signs. While some of these data would also appear on the handwritten chart, this tool was employed to assure more complete data collection. The data captured by this tool were used for the comparison of respiratory rate and saturation as reported later in this article.
Study population
We studied all patients
IV outpatient treatment
All patients who are perceived to need intravenous medication (e.g., due to ill appearance or poor tolerance of oral medication) but not supplemental oxygen are offered the possibility of IV treatment. Parents of patients who are interested in this treatment sign an informed consent form indicating their agreement to such care and follow-up. All patients with community-acquired pneumonia between the ages of 2 months and 10 years whose parents agreed to such treatment (n=129) between the dates of July 1, 2005 and June 30, 2006 were contacted at the phone number indicated on their medical records. The families were asked if the patient had needed hospitalization within one week of their discharge from, and if so, what intervention was done in the hospital. The patients’ families were also asked if the child had a history of asthma, as well as an open-ended question if they were pleased with their decision to attempt outpatient intravenous treatment.
Respiratory rate vs. saturation
The results from the clinical decision support tool described above are reported for 278 cases of children up to the age of 18 who were treated at Terem by pediatricians between January 2008 and June 2009 for a chief complaint of fever and for whom a chest x-ray was performed.
For the comparison of respiratory rate and saturation, tachypnea was defined, as per WHO definitions of pneumonia, as >50 breaths per minute for children under age 1 and >40 breaths per minute for children over age 1. Hypoxemia was defined as pulse oximetry reading of less than 95%. The association between tachypnea and hypoxemia was tested.
Analysis: Treatment by pediatricians vs. non-pediatricians
Pediatricians and non-pediatricians were compared as to percentage of cases where pneumonia was the final diagnosis, number of CBCs performed, number of x-rays performed, and percentage of mismatches. To minimize inter-radiologist variability in film interpretation, we looked at mismatch rates relative only to the most senior radiologist (who read 49% of all films done in the study population).
IV outpatient treatment
Frequency of hospitalization within one week, interventions given in hospital and length of stay were reported. An association with a past history of asthma was tested. For all analyses, categorical data were compared by chi square analysis. Statistical significance was set at p
Results
Treatment by pediatricians vs. non-pediatricians
During the study period, there were 24,080 visits by children up to age 10 with a chief complaint of fever. Of these, 17,728 (73.6%) were managed by a non-pediatrician.
As shown in Table 1, column B, the non-pediatricians were significantly more likely to diagnose pneumonia in the youngest age group (p
|
A |
B |
C |
D | |||||
|
Age group |
Cases with high fever
|
Pneumonia cases diagnosed/% of high fever cases |
Pneumonia cases with CBC (%) |
Pneumonia cases with x-ray (%) | ||||
|
Ped |
Non ped |
Ped |
Non ped |
Ped |
Non ped |
Ped |
Non ped | |
|
0-2 years |
3,390 |
9,711 |
284 (8.4) |
1,011 (10.4) |
184 (65) |
649 (64) |
271 (95) |
941 (93) |
|
2-5 years |
1,392 |
5,274 |
171 (12.3) |
758 (14.4) |
97 (57) |
474 (63) |
159 (93) |
709 (94) |
|
5-10 years |
570 |
2,743 |
60 (10.5) |
279 (10.2) |
28 (47) |
158 (57) |
54 (90) |
263 (94) |
IV outpatient treatment
The intravenous treatment group received either cefuroxime (60%) or ceftriaxone (40%). Most patients were treated over two to three days (minimum one day, maximum four days). In the follow-up, successful contact was made with 103 (80%) of these patients. Reasons for non contact included incorrect phone numbers on the chart or failure to answer the phone after multiple attempts. Only one patient contacted refused to answer the questions posed. There was no significant difference in the mean age or gender between those contacted and those not contacted.
In the group successfully contacted, only seven of 103 (7.6%) were admitted to a hospital within one week – five for three days, one for a week, and one for two weeks. None of these admitted patients required oxygen therapy during their hospital stay. There was a significant association with a history of asthma and subsequent hospital admission
Respiratory rate vs. saturation
There was a significant association between tachypnea and hypoxemia (p=.002). However, 40 of 279 children were hypoxemic without being tachypneic, and 60 of 279 were tachypneic without being hypoxemic.
Limitations
As the comparison of pediatricians and non-pediatricians was a chart-based retrospective study, we do not have the final outcome data for these patients. However, as the focus of that part of the study is the reliability of diagnosing pneumonia on chest x-ray and difference in rate of ordering of complete blood counts, we do not feel that this deters from the conclusions.
The inability to contact all patients treated intravenously means that there could have been a greater percentage of patients subsequently admitted than the 7.6% we found.
Treatment Discussion
The first step is to determine if the child does in fact suffer from pneumonia. While, as discussed previously, the x-ray findings are not a foolproof determinant of this condition, ability to correctly interpret pediatric chest x-rays is important in the diagnosis of this condition.
Our findings indicate that non-pediatricians are able to do this almost as well as pediatricians. The number of cases missed was small when compared to the “gold standard” of the radiologist. Having a back-up system of radiological review, along with senior physician follow-up, makes the clinical significance of these “mismatched” films minimal.
The most common bacterial cause of pneumonia in young children is Streptococcus pneumonia.10 Therefore, consensus statements recommend treating pediatric community-acquired pneumonia with amoxicillin. While beta-lactam resistance has been reported in S pneumonia, it is most often of an intermediate level. Therefore, high-dose amoxicillin (60 mg/kg – 80mg/kg) is the drug of choice in children under 5 years of age (excluding children under 2 months of age).
In children over 5 years of age, mycoplasm increases in frequency as a cause of CAP. Therefore, treatment with a macrolide 5 is recommended in these children.9,10
In children under age 5, only those who are penicillin-allergic should be treated with a macrolide. Using a macrolide in other children in this age group, even those that are more broad spectrum, is not recommended as there is a significance resistance to macrolides in most locations.
Determination of the respiratory status should include both measurement of hypoxemia by pulse oximetry and the actual counting of the respiratory rate. As seen by our findings, pulse oximetry alone can miss a large number of respiratory-distressed patients. The chest should be assessed for signs of distress such as subcostal retractions. Toxicity should also be determined with attention to skin color, hydration, and alertness of the child. If the child is not toxic appearing, oral medication is appropriate. For those children who are toxic-appearing but not in respiratory distress, intravenous medications can be given for the initial doses. The most appropriate medication is a second-generation cephalosporin such as cefuroxime. At times, ceftriaxone is chosen as it can be administered once daily. This convenience must be balanced against its broader spectrum and the desire to minimize the use of broad-spectrum antibiotics as part of the effort to reduce antibiotic resistance.
In our center, ceftriaxone was used primarily when the family would have been unable to return for multiple daily doses due to lack of transportation. Traditionally, this treatment has necessitated hospital admission. However, with the advent of urgent care centers, intravenous medication can often be administered during an outpatient visit, thus saving the inconvenience of hospital admission for a family while also reducing exposure to nosocomial infections. We have been routinely offering this option to parents whose children are hemodynamically stable and do not need supplemental oxygen for a number of years. As seen by our study of one year of such treatment, this approach is both responsible and safe.
Conclusions
In our study, pediatricians demonstrated superior skills in diagnosing pneumonias on chest x-ray; however, absolute differences in performance as compared to non-pediatricians were small. The clinical relevance of these differences was further minimized by a Terem-based mechanism for the follow-up (including treatment adjustment) on missed x-ray findings.
As such, we feel that it is clinically appropriate for non-pediatricians to assess such cases. Concurrently, we recommend (and provide) continued professional education to further improve non- pediatrician performance. The combination of review of x-rays by a radiologist and callback by a senior physician when needed reduces the possibility of incorrect care.
Provision of the option of outpatient intravenous medication can minimize hospital care. Significant monies can be saved, as management of pediatric CAP cases in an urgent care center is less expensive than direct ED care.
References
1. Jadavji T, Law B, Lebel MH, et al. A practical guide for the diagnosis and treatment of pediatric pneumonia. CMAJ. 1997;156:S703-S711.
2. McIntosh K. Community-acquired pneumonia in children. N Engl J Med. 2002;346:429-437.
3. Clinical management of acute respiratory infections in children: A WHO memorandum. Bull World Health Organ. 1981;59(5):707-716. Correction in: Bull World Health Organ. 1982;60(2):297.
4. Isaacs D. Problems in determining the etiology of community-acquired childhood pneumonia. Pediatr Infect Dis J. 1989;8:143-148.
5. Cincinnati Children’s Hospital Medical Center. Evidence-based care guideline for community acquired pneumonia in children 60 days through 17 years of age. Cincinnati Children’s Hospital Medical Center; 2006. Available at: www.guideline.gov.
6. Hopstaken RM, Witbraad T, van Engelshoven JM, et al. Inter-observer variation in the interpretation of chest radiographs for pneumonia in community acquired lower respiratory tract infections. Clinical Radiol. 2004;59:743-752.
7. Novack V, Avnon LS, Smolyakov A, et al. Disagreement in the interpretation of chest radiographs among specialists and clinical outcomes of patients hospitalized with suspected pneumonia. Eur J Intern Med. 2006;17:43-47.
8. Swingler GH. Observer variation in chest radiography of acute lower respiratory infections in children: a systematic review. BMC Med Imaging. 2001;1:1.
9. British Thoracic Society. Guidelines for the management of community-acquired pneumonia in childhood. Thorax. 2002;57:S1-S24.
10. Sinaniotis CA, Sinaniotis AC. Community-acquired pneumonia in children. Curr Opin Pulm Med. 2005;11:218-225.
