PEDIATRICS Vol. 104 No. 5 November 1999, pp. 1124-1125

COMMENTARY:
The Pitfalls of Using Managed Care Databases in Cost-Effectiveness Research

The article by Joffe and colleagues¹ that appeared in the September 1999 issue of Pediatrics discusses cost-effectiveness of respiratory syncytial virus (RSV) prophylaxis among preterm infants. This is based on a retrospective analysis of hospitalization rates from RSV lower respiratory (tract) infection (LRI) in children with a history of prematurity (36 weeks' gestation) from July 1, 1991 through March 31, 1996. This analysis² was published in the October 1999 issue of Pediatrics. The population, obtained from a large managed care database (Northern California Kaiser Permanente), is similar to that evaluated in clinical prevention trials of a hyperimmune globulin (RespiGam) and of a humanized monoclonal antibody (palivizumab, Synagis). In this article, the authors describe an overall RSV hospitalization rate of 3.2%. This rate is much lower than those reported both in prospective immunoprophylaxis studies and in prospective epidemiologic evaluations of preterm populations. In these studies, hospitalization rates range from 10.6 to 45%.^3-6

The discrepancy between the RSV hospitalization rate described by Joffe et al² and those observed in other published reports raises a real concern that RSV hospitalization rates in this study were significantly underestimated. Methodologic and systematic limitations in this database include the following:

1. The authors cite a significant attrition rate (26.5%) in their insurance membership population after neonatal intensive care unit discharge. The infants lost to follow-up were disproportionately skewed toward African-Americans and Latinos who may be at higher risk for RSV LRI hospitalization.⁷
2. The authors state that only 80% of level II neonatal intensive care unit admissions were accounted for in this study.
3. Laboratory-proven RSV hospitalization was dependent on RSV rapid antigen testing, which may not have been uniformly applied to all patients.
4. Duplicate health insurance coverage may have allowed for RSV hospitalizations to occur outside of the Kaiser database.
5. The tails of the RSV season (November and April) were excluded from study analysis.
6. Multiple RSV hospitalizations of a single infant were considered only as a single outcome.
7. Nosocomial RSV infections were excluded from analysis.
8. RSV LRI hospitalizations were defined using International Classification of Diseases, 9th Revision codes. The coding is dependent on uniformity in physician recording and in interpretation by coding clerks. In addition, the specific International Classification of Diseases, 9th Revision code for RSV, 079.6, was only created in 1996.

Recently presented Centers for Disease Control and Prevention data⁸ have demonstrated that the national incidence of hospitalization for bronchiolitis alone in the United States in 1995 was 34/1000 cases. The finding by Joffe et al² that preterm children had a lower RSV hospitalization rate than healthy infants <1 year of age (3.2% vs 3.4%) also suggests that RSV hospitalization rates for high-risk infants in Dr Joffe's analysis were seriously underestimated.

The results of this study underscore the limitations of using a retrospective managed care database to conduct epidemiologic studies. Clinical practitioners, administrators, and managed care organizations should view the cost-effectiveness conclusions derived from this database with great caution.

Jessie R. Groothuis, MD
Abbott Laboratories, International Division
Abbott Park, IL 60064-6188

FOOTNOTES

Received for publication Jul 26, 1999; accepted Jul 26, 1999.

Reprint requests to (J.R.G.) Medical Director of Immunology, Scientific Affairs, Abbott International, 200 Abbott Park Rd, Department 6LK, Building AP34, Abbott Park, IL 60064-6188. E-mail: jessie.groothuis{at}ln.ssw.abbott.com

ABBREVIATIONS

RSV, respiratory syncytial virus; LRI, lower respiratory (tract) infection.

REFERENCES

1. Joffe S, Ray GT, Escobar GJ, Black SB, Lieu TA Cost-effectiveness of respiratory syncytial virus prophylaxis among preterm infants. Pediatrics. 1999; 104:419-427 [Abstract/Free Full Text]
2. Joffe S, Escobar GJ, Black SB, Armstrong MA, Lieu TA Rehospitalization for respiratory syncytial virus among premature infants. Pediatrics 1999; 104:894-899 [Abstract/Free Full Text]
3. The PREVENT Study Group Reduction of respiratory syncytial virus hospitalization among premature infants and infants with bronchopulmonary dysplasia using respiratory syncytial virus immune globulin prophylaxis. Pediatrics. 1997; 99:93-99 [Abstract/Free Full Text]
4. The IMpact-RSV Study Group Palivizumab, a humanized respiratory syncytial virus monoclonal antibody, reduces hospitalization from respiratory syncytial virus infection in high-risk infants. Pediatrics. 1998; 102:531-537 [Abstract/Free Full Text]
5. Yuksel B, Greenough A Birth weight and hospital readmission of infants born prematurely. Arch Pediatr Adolesc Med. 1994; 148:384-388 [Abstract]
6. Cunningham CK, McMillan JA, Gross SJ Rehospitalization for respiratory illness in infants of less than 32 weeks' gestation. Pediatrics. 1991; 88:527-532 [Abstract/Free Full Text]
7. Glezen WP, Paredes A, Allison JE, Risk of respiratory syncytial virus infection for infants from low income families in relationship to age, sex, ethnic group, and maternal antibody levels. J Pediatr. 1981; 98:708 [Medline]
8. Shay DK, Holman RC, Anderson LJ. Bronchiolitis-associated hospitalizations among United States children from 1980 through 1995. IDSA 36th Annual Meeting; November 12-15, 1998; Denver, CO. (Abstract 44).