ARTICLE Use of a Large National Database for Comparative Evaluation of the Effect of a /Viral Clinical Care Guideline on Patient Outcome and Resource Utilization

James Todd, MD; David Bertoch, MHA; Susan Dolan, RN, MSN

Objectives: To use a large national comparative data- ternal comparison with other children’s hospitals using base to measure the internal effect of a set of evidence- standardized data from the Pediatric Health Informa- based bronchiolitis/viral pneumonia clinical care guide- tion System database of the Child Health Corporation of lines on clinical practice at a children’s hospital, and to America (Shawnee Mission, Kan). compare these changes with those at other children’s hospitals. Results: Overall, 10636 bronchiolitis/viral pneumonia cases were studied: 1302 at the index hospital and 9334 Design: Prospective cohort study with retrospective and at the 7 comparison hospitals. Internally, the index hos- concurrent (other hospital) controls. pital’s residents and attending physicians responded fa- vorably to the bronchiolitis/viral pneumonia care guide- Setting: The Children’s Hospital, Denver, Colo. lines, resulting in decreases in targeted resource utilization. There were no fatalities, and the number of days in Participants: Hospitalized children with bronchiolitis the intensive care unit decreased even though the mean and/or viral pneumonia. severity of admitted cases increased significantly. Tar- geted utilization was favorably affected, whereas untar- Interventions: Our clinical guidelines focused on clear geted utilization was not. Nosocomial infections did not admission and discharge criteria, individualized transition- increase with a decreased use of respiratory syncytial vi- anticipating orders, and “prove it or don’t use it” crite- rus testing. The index hospital differed favorably from ria for the use of respiratory syncytial virus testing, bron- other children’s hospitals in several categories. chodilators, chest physiotherapy, and ribavirin. Conclusion: Evidence-based care guidelines can suc- Main Outcome Measures: The effect of guideline cessfully influence utilization and clinical outcome. implementation was determined by comparative mea- surement of internal changes in utilization and out- come (nosocomial infection rate) across time and by ex- Arch Pediatr Adolesc Med. 2002;156:1086-1090

T HAS BEEN widely assumed that new variables and labor-intensive, expen- clinical care guidelines can favor- sive data collection methods such as ques- ably influence utilization and tionnaires and medical record review.7,8 clinical outcome by agreeing on Attempts to use internal administra- and disseminating evidence- tive databases for the evaluation of guide- based, peer-recommended information for lines can be frustrated by the lack of exist- I 1,2 patient care. One common childhood ing data for risk adjustment and/or outcome disease that has been successfully im- measurement as well as the use of charge- From the Department of proved using this process is bronchiolitis/ based variables influenced by changing pric- Epidemiology, The Children’s viral pneumonia. Several investigators have ing practices. Many hospitals subscribe to Hospital, Denver, Colo reported improved compliance with rec- external comparative databases that may (Dr Todd and Ms Dolan); Child ommendations for laboratory utilization have the same limitations but can poten- Health Corporation of America, and medication use in bronchiolitis us- tially compare their practices with similar Shawnee Mission, Kan ing care guidelines.3-5 The overall value of institutions.9,10 This added comparative di- (Mr Bertoch); and the Department of Pediatrics, guidelines and whether observed reduc- mension is at the core of the Joint Com- Microbiology, and Preventive tions in utilization variation actually re- mission on Accreditation of Healthcare 6 11 Medicine/Biometrics, University sult in improved outcomes is debated. In Organizations’ ORYX initiative, which re- of Colorado School of Medicine addition, collaborative efforts to measure quires hospitals to select and measure pre- (Dr Todd), Denver. these effects may require the definition of defined clinical indicators on a popula-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 156, NOV 2002 WWW.ARCHPEDIATRICS.COM 1086

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 tion basis and compare them quarterly with a peer group lowing: standard demographic data, dates of service, Interna- of comparable hospitals. We employed a similar measure- tional Classification of Diseases, Ninth Revision, Clinical Modi- ment process in the evaluation of a set of bronchiolitis/ fication (ICD-9-CM)20 diagnostic and procedural codes, viral pneumonia clinical guidelines using a national col- utilization data for the date and unit level, and charges. Cases laborative, children’s hospital database. were grouped using the All Patient Refined Diagnosis-Related Groups version 12 software (3M Company, Salt Lake City, 10 METHODS Utah). This well-known software uses complex clinical algo- rithms based on age and ICD-9-CM diagnostic and procedural In 1994, we implemented a collaborative process for the de- codes to classify patients into diagnosis-related groups and as- velopment of evidence-based clinical care guidelines at The Chil- sign relative severity scores ranging from 1 (least severe) to 4 dren’s Hospital of Denver (TCH), Denver, Colo. This system (most severe). Infection control statistics were reviewed to de- included wide representation from specialists and community termine the rates of nosocomial infection with RSV from De- physicians; it focused on areas of agreement, rather than forc- cember through April of each study year, based on the Cen- ing resolution of strong disagreement, and a commitment to ters for Disease Control’s definition of symptoms beginning more develop practical means of measuring the effect of these guide- than 72 hours after admission; viral testing was encouraged in lines on an ongoing basis. The bronchiolitis/viral pneumonia such cases. guidelines were implemented in 1995. The RSV season in Colorado predictably occurs during the Bronchiolitis/viral pneumonia was operationally defined winter, and bronchiolitis and viral pneumonia cases at other as an acute lower respiratory illness in a previously healthy child times of the year may have different causes and outcomes. There- with no history of associated with tachypnea or retrac- fore, we restricted our analysis to cases from January through tions, diffuse wheezing and/or crackles, and no evidence of bac- March for 1994 to 1998. Because RSV can cause bronchiolitis terial infection. Admission criteria included a complicating con- and/or viral pneumonia, because we did not encourage etio- dition, an oxygen requirement that could not be accommodated logic testing of cases, and because new relevant codes were added at home, apnea, poorly responsive neurological status, diffi- during the study period, we selected patients from birth to age culty feeding, or suspicion of sepsis. Testing for respiratory syn- 21 years based on the following ICD-9-CM principal diagnos- cytial virus (RSV) was discouraged for bed placement; in- tic codes using data from the PHIS database: pneumonia, or- stead, patients with similar symptoms were grouped, and droplet ganism unspecified (486); acute bronchiolitis (4661); pneu- precautions (hand washing, gloves, gown, and mask) were monia due to RSV (4801); viral pneumonia of unspecified cause strictly enforced for direct patient contact.12 (4809); acute bronchiolitis due to RSV (46611); and acute bron- A “prove it or don’t use it” policy was encouraged, which chiolitis due to other organisms (46619). All patient and com- stated that laboratory and imaging studies and therapies should parison hospital identifiers were removed before analysis. We be performed or continued only if they were anticipated and/or compared 1 preguideline year (1994) with 3 postguideline sea- proved to make a difference in each patient’s clinical course.4,13,14 sons (1995-1997) for TCH as compared with 7 other chil- As a result, the continued use of bronchodilators and/or chest phys- dren’s hospitals that had complete data and consistent coding iotherapy required observed improvement as measured by a de- of key variables and services for those years. Several of these fined respiratory distress score that evaluated age-specific respi- hospitals had implemented bronchiolitis guidelines during the ratory rate, retractions, wheezing, and aeration.15,16 The use of study period. ribavirin was recommended only in patients with impending res- A hospital survey for price changes from 1994 to 1997 re- piratory failure.17 Respiratory syncytial virus antigen testing or vealed inconsistent practices both between and within indi- viral culture was advised only if it was anticipated that a positive vidual hospitals. Fiscal years varied. Some hospitals imple- test result would cause a change in antimicrobial therapy.14 No mented across-the-board increases, whereas others were more recommendations were made regarding chest imaging or antibi- selective; increases varied from 0% to 7%. Because of the re- otic use, although more recent studies suggest a lack of general sults of this survey, we elected not to use charges as a measure clinical usefulness in patients with bronchiolitis.13,18,19 These 2 vari- of utilization. Given that even the definition of drug units may ables were useful controls for the targeted variables, with the hy- differ between hospitals or within the same hospital across time, pothesis that only targeted variables would be altered with the we determined that the 2 most reliable measures of utilization implementation of clinical guidelines. A protocol was provided would be days of use and any use of a test or therapy during a for the weaning of oxygen based on the rate of clinical recovery, particular hospitalization. Then we created dummy variables including the ability to feed, decreasing respiratory distress, and for these items for each patient. The estimated length of stay absence of apnea and cyanosis. Discharge criteria included im- (LOS) in days was calculated by subtraction of the date of ad- proved work of , discontinuation of oxygen or a refer- mission from the date of discharge, not accounting for frac- ral for home oxygen therapy, adequate oral intake, parental edu- tional days. To account for LOS outliers that might bias con- cation regarding home oxygen therapy, and follow-up arranged tinuous variable comparisons, we excluded cases that exceeded with the primary care physician. the Health Care Financing Administration’s LOS outlier defi- We hypothesized that the use of an existing comparative nition for 1997: the geometric mean length of stay plus the lesser database could facilitate the evaluation of complex clinical care of 24 days or 3 SDs. This equaled 19.81 days for this data set, guidelines. Specifically, this meant that unique effects could be so all patients with an LOS greater than 19 were excluded.21 inferred if consistent postintervention intra-institutional ef- Calculations for LOS excluded any patient who did not have a fects on targeted variables were documented during sequen- discharge status of home or home health. tial years, with additional support if the index institution dif- All statistical analyses were conducted using SPSS version fered favorably in comparison with other institutions using 10 statistical software (SPSS Inc, Chicago, Ill). We assumed that standardized data from the Pediatric Health Information Sys- new clinical guidelines were effective if intra-institutionally, con- tem (PHIS) database of the Child Health Corporation of America sistent statistically significant trends could be documented dur- (Shawnee Mission, Kan). This database currently contains dis- ing sequential years and inter-institutionally, if the index hos- charge data from 29 children’s hospitals geographically dis- pital (TCH) differed significantly in a favorable direction from tributed throughout the United States. Available data ele- the comparison institutions. Given the assumption of index hos- ments abstracted and coded by the collaborating medical records pital admissions for bronchiolitis/viral pneumonia of 250 to 300 departments using PHIS data quality guidelines include the fol- cases per year, the power to detect a change in utilization from

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 156, NOV 2002 WWW.ARCHPEDIATRICS.COM 1087

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 1. Distribution of Bronchiolitis and Viral Pneumonia Cases by Year*

Year

Diagnosis ICD-9-CM 1994 1995 1996 1997 Total Cases Pneumonia, organism unspecified 486 597 775 695 586 2653 Acute bronchiolitis 4661 1662 2052 1600 . . . 5314 Pneumonia due to RSV 4801 145 203 97 91 536 Viral pneumonia, unspecified 4809 25 68 30 66 189 Acute bronchiolitis due to RSV 46 611 ...... 1162 1162 Acute bronchiolitis due to other organisms 46 619 ...... 782 782 Total Cases 2429 3098 2422 2687 10 636 Cases admitted to index hospital (TCH) 277 323 409 293 1302

*ICD-9-CM indicates International Classification of Diseases, Ninth Revision, Clinical Modification; RSV, respiratory syncytial virus; and TCH, The Children’s Hospital, Denver, Colo.

Table 2. Comparison of Index Hospital With Other Hospitals for Utilization Related to Variables Targeted and Untargeted by the Bronchiolitis/Viral Pneumonia Care Guidelines*

Change Between 1997 Index Hospital Comparison Hospitals and 1994 (Index vs Comparison Hospitals) 1994 1997 1994 1997 Indicator (n = 277) (n = 293) P Value† (n = 3963) (n = 3776)P Value† Difference 95% CI Chest physiotherapy‡ 69.0 43.9 Ͻ.001 22.3 26.9 Ͻ.001 −29.7 −21.6 to −37.8 Chest x-ray film 51.6 50.5 .79 74.3 75.6 .18 −2.4 6.0 to −10.8 Ribavirin‡ 5.8 1.0 .002 4.9 0.6 Ͻ.001 −0.5 2.6 to −3.6 Antibiotics 70.0 75.1 .18 73.0 78.5 Ͻ.001 −0.4 7.2 to −8.0 RSV testing‡ 56.7 35.8 Ͻ.001 56.9 42.6 Ͻ.001 −6.6 1.7 to −14.9 Bronchodilators‡ 75.1 76.1 .78 46.9 74.1 Ͻ.001 −26.2 −18.8 to −33.6 Bronchodilators (Ͼ1 d) 6.9 28.7 Ͻ.001 21.7 40.9 Ͻ.001 2.6 8.9 to −3.7 Ventilator 8.7 6.5 .32 3.3 5.3 Ͻ.001 −4.2 0.2 to −8.7 ICU 20.2 7.8 Ͻ.001 9.3 13.5 Ͻ.001 −16.6 −10.8 to −22.4 ICU severity score Ն3 32.1 (n = 56) 60.9 (n = 23) .02 37.1 (n = 367) 43.1 (n = 511) .07 22.8 47.1 to −1.5 Mean length of stay, d§ 3.85 3.67 Ͼ.99 4.02 3.76 Ͻ.001 0.08 1.23 to −1.39 Mean severity score 1.54 1.73 .02 1.48 1.59 Ͻ.001 0.08 0.60 to −0.76 (APR-DRG system)§

*Data are presented as percentage unless otherwise indicated. CI indicates confidence interval; RSV, respiratory syncytial virus; ICU; intensive care unit; and APR-DRG, All Patient Refined Diagnosis-Related Groups. †Probability of differences within the same hospital(s) between 1997 and 1994 determined using ␹2 test for proportionate data and Mann-Whitney test for continuous data (length of stay and severity). ‡Indicators targeted by index hospital guidelines. §Outlier-trimmed data.

50% to 40% was 0.65 and from 50% to 35% was 0.90. The ␹2 ICD-9-CM coding shift occurred in 1997, total case num- test was used to compare proportionate data, and because a nor- bers were relatively comparable from year to year. mal data distribution could not be assumed, the Mann- The effect of the bronchiolitis and viral pneumonia Whitney test was used to compare continuous and categorical guidelines at the index hospital was first estimated by com- variables (LOS and severity score). Single-entry logistic regres- paring the temporal influence on its targeted variables sion was used to evaluate internal changes (dependent variable: use or nonuse of targeted variable; independent variable: year) (Table 2). Between 1994 and 1997, there was a signifi- during the 4 study years. Confidence intervals were calculated cant (P=.02 using the Mann-Whitney test) increase in for the difference between 1994 and 1997 for the index hospital mean severity, with the mean severity score rising from as compared with the other institutions. All P values are re- 1.49 to 1.70 (outlier-trimmed data). However, there was ported as 2-tailed probabilities. no significant change in mean LOS. Despite the in- creased mean severity score, there was a significant Ͻ RESULTS (P .001) decrease (from 20.2% to 7.8%) in the propor- tion of patients admitted to the intensive care unit (ICU). A total of 10636 cases of bronchiolitis (68.2%) and viral This was accompanied by a significant increase in the per- pneumonia (31.8%) were identified using ICD-9-CM codes centage of ICU patients with bronchiolitis/viral pneu- (Table 1), with patients ranging in age from birth to 21 monia who had a severity score of 3 or higher (from 32.1% years (mean, 1.7 years). There were no associated fatali- to 60.9%). There were significant decreases in utiliza- ties at the index hospital during that time. Although an tion of the targeted variables of RSV testing, ribavirin treat-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 156, NOV 2002 WWW.ARCHPEDIATRICS.COM 1088

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 CPT RSV Test ICU Admission Ribavirin Table 3. Effect of Clinical Care Guidelines 80 on Seasonal Nosocomial RSV Infection Rate 70 OR = 0.67 (0.60-0.75) per 1000 Inpatient Days at Index Hospital* 60 50 Year 1994 1995 1996 1997 40 OR = 0.72 (0.64-0.80) Bronchiolitis and/or viral pneumonia 277 323 409 293 30 Positive RSV test results 293 293 257 149

Utilization, % 20 OR = 0.77 (0.66-0.90) Nosocomial RSV infections 15 18 4 12 10 OR = 0.43 (0.28-0.68) Total inpatient days 23 258 23 266 20 554 21 307 0 RSV nosocomial infection rate 0.64 0.77 0.19 0.56 1994 1995 1996 1997 per 1000 patient days Year *Data are presented as number of cases unless otherwise indicated. Figure 1. Association of bronchiolitis/viral pneumonia care guidelines with RSV indicates respiratory syncytial virus. individual patient use of targeted variables (chest physiotherapy [CPT], respiratory syncytial virus [RSV] testing, and ribavirin) and pediatric intensive care unit (ICU) admissions during sequential years. Guidelines were implemented in 1995. Odds ratios (ORs) and 95% confidence intervals index hospital compared with the other institutions (dif- indicate yearly decrease in utilization by single-variable entry logistic ference, −16.6%; 95% confidence interval, −10.8% to regression, with the dependent variable being use or nonuse of a listed test −22.4%). During this period, we were able to decrease our or therapy. use of chest physiotherapy relative to the comparison hos- pitals (difference, −29.7%; 95% confidence interval, −21.6% None 1 d ≥2 d to −37.8%), but our use of this treatment still exceeded the 80 comparison hospitals in 1997. Ribavirin use decreased in both groups, consistent with changing national guide- 60 lines. Respiratory syncytial virus testing, which was com- P<.001 parable in 1994, decreased in both groups by 1997. The 40 use of bronchodilators at any time during hospitalization

% of Cases actually increased at the comparison institutions to levels 20 observed during both periods at the index hospital. 0

1994 1995 1996 1997 COMMENT Year In this study, we used the Child Health Corporation of Figure 2. Association of bronchiolitis/viral pneumonia care guidelines with America’s existing PHIS database to evaluate the effect of bronchodilator use at the index hospital during sequential years. Guidelines were implemented in 1995. Overall, PϽ.001 using the Pearson ␹2 test. a set of bronchiolitis and viral pneumonia clinical care guide- lines on patient care practices. We applied the Joint Com- mission on Accreditation of Healthcare Organizations’ con- ment, and chest physiotherapy but no significant changes cept of performing both temporal intra-institutional and in the untargeted variables of antibiotic treatment and comparative inter-institutional analyses. Our results are chest imaging. comparable with other recent studies of bronchiolitis.3-5 As shown in Figure 1 with logistic regression, these Intra-institutionally, our residents and attending physi- changes in utilization were progressive during the 3 years cians appeared to respond favorably to the bronchiolitis/ following implementation of the guidelines. There was viral pneumonia care guidelines, resulting in targeted re- no overall decrease in the targeted use of bronchodila- source changes with no fatalities, a reduction in ICU tors; however, there was a significant decrease among pa- admissions despite increased severity, and no increase in tients treated for 1 day and an increase among those never nosocomial infection rates even though RSV testing was treated or treated for 2 or more days (Figure 2) that was not used for bed placement.12,22 The latter observation sug- consistent with the guidelines’“prove it or don’t use it” gests that cohorting based on symptoms rather than RSV recommendations. Despite the decrease in targeted RSV testing may be sufficient to prevent nosocomial infection. testing, there was no increase in the rate of nosocomial The untargeted reduction in ICU admissions appeared to RSV infections (Table 3). be due to a significant decrease in the number of patients As an additional method to evaluate the effect of our in the ICU with severity scores lower than 3. Besides the implementation of clinical care guidelines, the results of observed temporal changes, evidence supporting the ef- the comparison of targeted and untargeted variables from fectiveness of the guidelines included the fact that tar- our index hospital with those from 7 other children’s hos- geted utilization responded favorably, whereas untar- pitals for the 1994-1997 period are presented in Table 2. geted utilization (chest x-rays and antibiotics) did not. More Both the index and comparison hospitals demonstrated sig- recent publications suggest that these latter treatments may nificant increases in severity score, although this score was be considered targets for improvement as well.13,23,24 Inter- significantly higher for the index hospital than for the com- institutional comparison with 7 other hospitals in the da- parison hospitals in 1997 (1.73 vs 1.59; P=.04), with a lower tabase supported the index hospital’s observations. It is likely ICU admission rate (7.8% vs 13.5%; P=.006). The differ- that many of these children’s hospitals implemented simi- ence from 1994 to 1997 between the change in the per- lar clinical care guidelines during the study period, based centage of ICU patients was significantly decreased for the on evolving evidence-based trends in the care of bronchi-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 156, NOV 2002 WWW.ARCHPEDIATRICS.COM 1089

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 What This Study Adds REFERENCES

1. Christakis DA, Rivara FP. Pediatricians’ awareness of and attitudes about four Bronchiolitis and viral pneumonia are currently the most clinical practice guidelines. Pediatrics. 1998;101:825-830. common causes of pediatric hospital admission. Evidence- 2. Bergman DA. Evidence-based guidelines and critical pathways for quality im- based clinical care guidelines have been shown to influ- provement. Pediatrics. 1999;103:225-232. ence selected aspects of care. 3. Adcock PM, Sanders CL, Marshall GS. Standardizing the care of bronchiolitis. Arch Pediatr Adolesc Med. 1998;152:739-744. The association of clinical care guidelines with tem- 4. Perlstein PH, Kotagal UR, Bolling C, et al. Evaluation of an evidence-based guide- poral changes in resource utilization for bronchiolitis and line for bronchiolitis. Pediatrics. 1999;104:1334-1341. viral pneumonia is confirmed. Respiratory syncytial vi- 5. Perlstein PH, Kotagal UR, Schoettker PJ, et al. Sustaining the implementation of rus testing may not be needed to determine bed place- an evidence-based guideline for bronchiolitis. Arch Pediatr Adolesc Med. 2000; ment as long as droplet precautions are maintained for 154:1001-1007. 6. Merritt TA, Palmer D, Bergman DA, Shiono PH. Clinical practice guidelines in cohorts of patients with similar symptoms. Local and na- pediatric and newborn medicine: implications for their use in practice. Pediat- tional comparisons suggest that evidence-based changes rics. 1997;99:100-114. in practice will require ongoing reinforcement and mea- 7. Simmons D, Coppell K, Drury PL. The development of a national agreed mini- surement to be sustained. mum diabetes dataset for New Zealand. J Qual Clin Pract. 2000;20:44-50. 8. Singsen BH. Health status (arthritis impact) in children with chronic rheumatic diseases: current measurement issues and an approach to instrument design. Arthritis Care Res. 1991;4:87-101. 9. Roos NP, Black C, Roos LL, et al. Managing health services: how administrative data and population-based analyses can focus the agenda. Health Serv Manage olitis and viral pneumonia as reflected by comparable Res. 1998;11:49-67. 10. Muldoon JH. Structure and performance of different DRG classification sys- significant decreases in RSV testing and ribavirin use (al- tems for neonatal medicine. Pediatrics. 1999;103:302-318. though the use of antibiotics and bronchodilators re- 11. Joint Commission on Accreditation of Healthcare Organizations. A Guide to Per- mains high). Our study suggests that old practice patterns formance Measurement for Hospitals. Oakbrook Terrace, Ill: Joint Commission on Accreditation of Healthcare Organizations; 2000. (eg, chest physiotherapy and bronchodilators) may be slow 12. Krasinski K, LaCouture R, Holzman RS, Waithe E, Bonk S, Hanna B. Screening to change, validating the need for ongoing reinforcement, for respiratory syncytial virus and assignment to a cohort at admission to re- evaluation, and updating of guidelines. duce nosocomial transmission. J Pediatr. 1990;116:894-898. 13. Swingler GH, Hussey GD, Zwarenstein M. Randomised controlled trial of clini- In the process of this study, the strengths and limi- cal outcome after chest radiograph in ambulatory acute lower-respiratory infec- tations of working with a large comparative, administra- tion in children. Lancet. 1998;351:404-408. 14. Adcock PM, Stout GG, Hauck MA, Marshall GS. Effect of rapid viral diagnosis on tive database became apparent. Because ICD-9 codes and the management of children hospitalized with lower respiratory tract infection. coding practices change, it is imperative to check for in- Pediatr Infect Dis J. 1997;16:842-846. ternal and external coding consistency. Temporal trends 15. Dobson JV, Stephens-Groff SM, McMahon SR, Stemmler MM, Brallier SL, Bay C. The use of albuterol in hospitalized infants with bronchiolitis. Pediatrics. 1998; in charge data are confounded by variable price in- 101:361-368. creases (intra-institutionally and inter-institutionally), 16. Webb MS, Martin JA, Cartlidge PH, Ng YK, Wright NA. Chest physiotherapy in making charge comparisons unsatisfactory for most pur- acute bronchiolitis. Arch Dis Child. 1985;60:1078-1079. 17. Law BJ, Wang EE, MacDonald N, et al. Does ribavirin impact on the hospital course poses. Nonetheless, in large databases, dummy vari- of children with respiratory syncytial virus (RSV) infection? an analysis using ables such as days of use and/or any use of tests or thera- the pediatric investigators collaborative network on infections in Canada (PICNIC) RSV database. Pediatrics. 1997;99:E7. pies may have sufficient statistical power to show 18. Roback MG, Dreitlein DA. Chest radiograph in the evaluation of first time wheez- significant and clinically relevant differences. A notice- ing episodes: review of current clinical practice and efficacy. Pediatr Emerg Care. able deficiency in these databases is the lack of true out- 1998;14:181-184. 19. Antonow JA, Hansen K, McKinstry CA, Byington CL. Sepsis evaluations in hos- come data, so comparisons of utilization with evidence- pitalized infants with bronchiolitis. Pediatr Infect Dis J. 1998;17:231-236. based precedents become a necessary but incomplete 20. St Anthony’s Softbound ICD-9-CM Code Book, 1997. Reston, Va: St Anthony surrogate for true outcome measurement. Publishing; 1996. 21. 62 Federal Register 46040 (1997). 22. Mackie PL, Joannidis PA, Beattie J. Evaluation of an acute point-of-care system Accepted for publication June 7, 2002. screening for respiratory syncytial virus infection. J Hosp Infect. 2001;48:66- 71. Presented at the Pediatric Academic Societies Meet- 23. Liebelt EL, Qi K, Harvey K. Diagnostic testing for serious bacterial infections in ing, San Francisco, Calif, May 4, 1999. infants aged 90 days or younger with bronchiolitis. Arch Pediatr Adolesc Med. Corresponding author: James Todd, MD, Department 1999;153:525-530. 24. Davies HD, Wang EE, Manson D, Babyn P, Shuckett B. Reliability of the chest of Epidemiology, The Children’s Hospital, 1056 E 19th Ave, radiograph in the diagnosis of lower respiratory infections in young children. Pe- Denver, CO 80218 (e-mail: [email protected]). diatr Infect Dis J. 1996;15:600-604.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 156, NOV 2002 WWW.ARCHPEDIATRICS.COM 1090

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021