Net cost of palivizumab for respiratory syncytial virus prophylaxis during the 1998/99 season in northern Alberta

Abstract

OBJECTIVE:

Palivizumab has been shown to decrease respiratory syncytial virus (RSV) hospitalization rates in preterm infants and infants with chronic lung disease. The objective of the present study was to determine whether the use of palivizumab during the 1998/99 RSV season would have resulted in a cost-saving in infants discharged from Edmonton hospitals.

DESIGN:

A retrospective study of RSV hospitalizations was performed by contacting parents and reviewing hospital lists. The net cost of using palivizumab was determined by comparing the cost of giving the drug from November 1, 1998 to April 1, 1999 with the cost of potentially averted medical transports and hospitalizations.

POPULATION:

One hundred fifty-nine infants discharged from Edmonton hospitals who met the Canadian Paediatric Society’s criteria for receiving palivizumab during the 1998/99 RSV season were studied.

RESULTS:

The cost of using palivizumab in these 159 study infants would have been $753,300. The infants had 21 RSV hospitalizations and required four medical transports. The estimated cost of RSV hospital-based care for these infants was $168,888. Assuming a drug efficacy of 39% in infants with chronic lung disease and 78% in infants born before 33 weeks’ gestation with no chronic lung disease, $121,147 of these costs could have been averted if palivizumab had been used.

CONCLUSIONS:

The net cost to the health care system of using palivizumab, as recommended in the Canadian Paediatric Society guidelines, in study infants in northern Alberta during the 1998/99 RSV season would have been $632,153.

Respiratory syncytial virus (RSV) infects almost all children by two years of age. The risk of hospitalization is highest for those who acquire RSV before six weeks of age or who have cyanotic heart disease, prematurity, chronic lung disease or immunodeficiency. Prematurity probably increases the severity of RSV because oxygen, ventilation and infection have already damaged the airways. Also, preterm infants have lower levels of transplacentally acquired RSV antibody, and consequently, reduced protection against RSV lower respiratory tract disease.

Because RSV causes significant morbidity and some mortality, immunization of neonates in-utero or soon after birth would be ideal. Progress is being made on the development of an effective vaccine, but, in the meantime, other strategies must be employed. Simple measures, such as the promotion of breastfeeding and handwashing in the home and in childcare centres, have been advocated, but are of unproven efficacy (1). The use of intravenous immune globulin to prevent RSV was only successful after a hyperimmune product was developed. This product, RSV immune globulin intravenous (RSV-IGIV [RespiGam; MedImmune, Inc, USA]), was shown to decrease RSV hospitalization rates by 20% in infants born at 35 weeks’ gestation or less with no chronic lung disease (CLD) and who were younger than six months of age at study entry, and by 49% in children with CLD (diagnosed by a neonatologist or pulmonologist and requiring oxygen during the past six months but otherwise not defined) aged 24 months or younger at study entry (2). Problems with the product include the requirement for monthly intravenous administration of a large volume of fluid over several hours, the need to delay live viral vaccines for nine months after the last dose and the small risk of the acquisition of infectious agents from a human blood product.

Therefore, a monoclonal antibody that could be administered by a monthly intramuscular injection (palivizumab [Synagis; MedImmune, Inc, USA]) was developed. This product was shown to decrease RSV hospitalization rates by 78% in preterm infants without CLD born before 36 weeks’ gestation and who were younger than six months of age at study entry, and by 39% in infants with a clinical diagnosis of CLD (defined as the use of supplemental oxygen, bronchodilators, steroids or diuretics six months before study entry) and younger than 24 months of age at study entry (3). There were no statistically significant differences between palivizumab and placebo groups in the need for mechanical ventilation or in mortality. Palivizumab did not result in any significant adverse events. One of the 1998 recommendations of the American Academy of Pediatrics (AAP) was that the product be considered for monthly administration throughout the RSV season to all infants born before 29 weeks’ gestation and younger than 12 months of age at the start of the RSV season, infants born at 29 to 32 weeks’ gestation and less than six months of age at the start of the RSV season, and all children less than 24 months of age who required medical therapy for CLD in the previous six months (4). The 1999 recommendation of the Canadian Paediatric Society (CPS) mirrors the AAP’s recommendation, except that it does not recommend that the product be given to infants older than six months of age at the start of RSV season unless they have CLD, and the use of oxygen in the previous six months is required in the CPS definition of CLD (1).

Determination of the start and finish of the RSV season may be difficult because the peak months vary from year to year in any location; however, most children would require at least five injections of palivizumab for the prophylaxis to be effective. The cost of this drug is $1,012.50 per dose (assuming that the child weighs 5 kg and palivizumab 75 mg is administered). Therefore, the cost effectiveness of this intervention has been questioned (5,6). Data from the 1998/99 RSV season in the authors’ region were examined to determine the net cost of the use of palivizumab according to the CPS and AAP guidelines. It was also questioned whether any modification to these criteria would have decreased the net cost of using the product.

METHODS

Study population

The Children’s Health Centre (CHC) includes five hospitals with paediatric beds in the Edmonton area. The referral area for these five hospitals includes all of northern Alberta and portions of western Saskatchewan, eastern British Columbia and the Northwest Territories. The population served in these areas is approximately 1.6 million. Four of the five hospitals admit preterm newborns. By analyzing the gestational age of all newborns admitted to each of these four hospitals and by reviewing their databases to identify infants with CLD, a list of study infants was derived. These four hospitals have databases that are maintained by dedicated encoders who enter patient data from charts and use the International Classification of Disease, 9th edition (7) code for discharge diagnoses. The data for the two tertiary care neonatal intensive care unit (NICU) databases were collected from patient charts by clinical nurse specialists or research nurses by using a standardized data collection sheet that was common to both sites. All infants born before 33 weeks’ gestation who were younger than six months of age on November 1, 1998, and those born before 33 weeks’ gestation who were discharged between November 1, 1998 and April 1, 1999, were included. The definition of CLD was those infants requiring oxygen supplementation at 36 weeks’ corrected age. Eight infants who were given palivizumab as part of a phase IV study of the drug were excluded. No other infants in the authors’ region received palivizumab during the 1998/99 RSV season. Approval was obtained from the Ethics Review Board of the CHC to conduct this study.

Study design

A list of all children admitted to the CHC with RSV between July 1, 1998 and June 30, 1999 was obtained, and this list was matched with the list of study children. It is standard practice at the CHC to do a nasopharyngeal aspirate for RSV on all children admitted with compatible respiratory symptoms. RSV was diagnosed by using direct immunofluorescent antigen detection (Merifluor; Meridian Diagnostics, Inc, USA) or TestPack RSV EIA (Enzyme Immunoassay; Abbott, Canada) in combination with culture.

Because children could have been admitted to a hospital outside the authors’ region, parents of study children were contacted by telephone to inquire whether their child had any hospital admissions for respiratory complaints between July 1, 1998 and June 30, 1999. If so, their permission was obtained to contact the hospital where the child was admitted to determine whether RSV infection was possible and if a test for RSV was performed. A large percentage of families in the Northwest Territories do not have telephones. Therefore, a list of all children admitted to the Stanton Hospital in Yellowknife with RSV during the 1998/99 season was obtained. Patients were included in data analysis only if their parents were contacted by telephone, or if hospital records showed that they were admitted to a CHC site or the Stanton hospital with RSV infection. If the child was admitted to hospital with a respiratory illness, but was not tested for RSV infection, the authors considered that to be an RSV hospitalization.

Calculation of days of RSV hospitalization

A child was considered to have RSV infection if either a rapid antigen detection test or culture from nasopharyngeal secretions was positive for RSV. Hospital records of all RSV-infected children were searched to determine whether they required air or ambulance transport to hospital, intensive care unit (ICU) admission or ventilation, and the duration of hospital stay that was attributable to RSV. If the child’s stay was prolonged while they were waiting transport home, these days were considered to be due to RSV, but if the child remained in hospital because of medical problems unrelated to RSV, the days were only counted until the child was back on his/her usual rate of oxygen flow and feeding well with no apnea.

Calculation of direct medical costs

There is no evidence that palivizumab decreases the severity of RSV infection in children who fail prophylaxis and require hospital admission. In fact, in the original study, there was a trend toward longer hospital stay and longer duration of increased oxygen requirement in palivizumab-treated children (5), and the only death attributable to RSV occurred in the palivizumab group (3). Therefore, the benefit of palivizumab may be that it decreases the number of hospitalizations rather than the duration or complexity of hospitalizations. It is not known whether the prevention of RSV hospitalization decreases the incidence of reactive airways disease or confers other long term benefits on the patient, so no costs were assigned to possible long term sequelae of RSV infection.

The cost of an individual transport is not determined in the authors’ centre, so the transport cost used in calculations was the average cost per paediatric ICU (PICU) transport ($3,150). The per diem cost for hospitalization used in calculations was the daily charge at the University of Alberta Hospital to out-of-province patients ($1,062). This per diem cost was extrapolated to other hospitals because costs were not available for many of the hospitals to which the children were admitted. Physician costs used in calculations were $1,000/transport, $110/admission and $20/diem for hospital visits. When calculating the cost of palivizumab, it was assumed that the average child weighed 5 kg and, therefore, would receive palivizumab 75 mg. Based on the assumption that no wastage would occur, a cost of $1,012.50/dose was used. The cost of acquisition, storage and administration of the drug was ignored in calculations because this is a very small proportion of the total cost of the drug.

It was assumed that if palivizumab was to have been administered, it would have been started on November 1, 1998 and given monthly until April 1, 1999 for a total of six doses because the RSV season usually starts in December or January and ends in March or April. It was assumed that palivizumab would have been used only in outpatients, and that it would have been started during the first month after discharge for infants discharged during the RSV season. Based on the fact that the product offers protection for about one month, it was assumed that for RSV admissions occurring between November 1, 1998 and May 1, 1999, 39% of the cases could have been prevented in infants with CLD and 78% of cases could have been prevented in infants born before 33 weeks’ gestation with no CLD (3).

The cost of following the AAP guidelines for the use of palivizumab was calculated. The authors also looked at the cost if palivizumab had been only offered to the four subgroups of patients that were thought to be at the highest risk for RSV hospitalization rather than the entire cohort included in the CPS guidelines. The groups selected were as follows: infants with CLD; children born before 33 weeks’ gestation who required oxygen beyond day 28 of life and were discharged within three months before the time that palivizumab was started (6); infants born before 33 weeks’ gestation or with CLD who were 44 weeks’ postconceptual age or younger at the time that palivizumab would have been given (8); and infants who were part of a multiple gestation (9).

If a physician waited until RSV was identified in the community to start palivizumab, it might be possible to give only two or three doses and still provide prophylaxis to the majority of infected patients. Therefore, the cost of giving palivizumab was also calculated only during the two months when the most RSV hospitalizations occurred in the study population.

Comparison with other RSV seasons

Because RSV seems to produce a more significant disease in some years compared with other years (possibly because of differences in the predominant subtype circulating that year), the number of RSV admissions to the ward and to the ICU for the CHC over the previous five years were examined.

Statistical analysis

The comparison of values among groups was performed by using the two-tailed Student’s t test and the Mann-Whitney U test (rank sum) for cases in which the test of normality failed. Comparisons of proportions for categorical measure were analyzed using the χ² test with Yates correction when the expected number of occurrences was less than five. Means and standard deviations for selected variables were compared among NICU infant groups. Statistical analyses were performed using computer software (SigmaSTAT for Windows version 2.03; SPSS Inc, USA). P<0.05 was considered to be significant. Values are expressed as mean ± SD.

RESULTS

RSV hospitalizations in study patients

Two hundred six infants met the CPS criteria to receive palivizumab, and 159 infants (77.2%) were included in the data analysis because their RSV hospitalization status was definitely established. Table 1 shows that the infants included in the study were comparable in gestational age, birth weight and age at the end of the study with those who were excluded because their parents could not be contacted. None of the excluded infants had CLD; however, 8.1% of the study infants had CLD. Figure 1 shows the RSV hospitalization rates in study patients, with 20 patients being hospitalized with RSV and one being hospitalized with respiratory illness where no testing for RSV was performed. These 21 patients accounted for 139 days of hospitalization, including 16 days of ICU care. None of these hospitalizations occurred outside the six months when RSV prophylaxis would have been offered. The median age at hospitalization was 4.1 months. Table 2 shows the risk of RSV hospitalization for the study population. For all infants meeting the CPS criteria to receive palivizumab, the hospitalization rate was 13.2%, with 19.0% of the hospitalized infants being admitted to ICU and 14.3% of the hospitalized infants being ventilated. The highest risk group was infants with CLD who were discharged in September, October or November 1998 who had a 37.5% RSV hospitalization rate.

TABLE 1:

Comparison of 159 infants who met the Canadian Paediatric Society criteria to receive palivizumab with 47 infants excluded from the study because follow-up data could not be obtained

	Included (n = 159)	Excluded (n = 47)	P value
Mean gestational age (weeks)	29.9±2.5	30.3±1.9	0.16
Mean body weight (g)	1505±400	1512±408	0.09
Chronic lung disease	13 (8.1%)	0 (0%)	0.09
Mean chronological age at the end of the respiratory syncytial virus season (months)	7.8±2.9	8.2±2.6	0.47

Figure 1).

Respiratory syncytial virus (RSV) hospitalizations from November 1,1998 to April 30, 1999 in infants discharged from Edmonton hospitals who would have met the Canadian Paediatric Society (CPS) criteria to receive palivizumab prophylaxis during that time period. All infants were born before 29 weeks’ gestation. *One infant went home on oxygen. CLD Chronic lung disease; ICU Intensive care unit

TABLE 2:

Rate of respiratory syncytial virus (RSV) hospitalizations from November 1, 1998 to April 30, 1999 for infants discharged from Edmonton hospitals who would have met the Canadian Paediatric Society criteria to receive palivizumab

Risk factor	Hospitalized with RSV
32 weeks’ gestation or less with or without CLD and discharged before or during RSV season (n = 159)	13.2%
32 weeks’ gestation or less with or without CLD and discharged before or during RSV season, or 28 weeks’ gestation or less with or without CLD at age six to 12 months at onset of RSV season according to American Academy of Pediatrics guidelines (n = 192)	12.5%
Infants with CLD (n = 13)	23.1%
Infants born before 33 weeks’ gestation, placed on oxygen on day 28, and discharged September, October or November 1998 (n = 8)	37.5%
Infants who were at 44 weeks’ postconception or less at time first dose is given (n = 102)	14.7%
Infants who are part of a multiple gestation (n = 52)	5.8%

Only infants who had their RSV status definitely established are included. Infants with chronic lung disease (CLD) are included in more than one group

Calculation of costs

Table 3 shows the calculation of the net cost of using palivizumab. The cost of administering palivizumab from November 1, 1998 to April 1,1999 to the 159 children in the study would have been $753,300. The estimated cost of inpatient treatment of RSV and transports in study patients was $168,888.

TABLE 3:

Calculation of the net cost of using palivizumab from November 1, 1998 to April 30, 1999 for infants discharged from Edmonton hospitals who would have met the Canadian Paediatric Society criteria to receive palivizumab

Cost of paliviumab	Cost of RSV hospitalizations
91 patients × 6 doses = $552,825 16 patients × 5 doses = $81,000 10 patients × 4 doses = $40,500 10 patients × 3 doses = $30,375 16 patients × 2 doses = $32,400 16 patients × 1 dose = $16,200	Transport costs    Four transports × $3,150/transport = $12,600    Four physican fees × $1,000/transport = $4,000
	Hospital costs    139 days × $1,062/day = $147,618
	Physician fees    First hospital day: 21 × $110 = $2,310    Subsequent days: 118 × $20 = $2,360
Total drug costs = $753,300	Total hospitalization costs = $168,888

Only infants who had their respiratory syncytial virus (RSV) status definitely established are included. The cost per dose of palivizumab is $1,012.50. Palivizumab could have prevented 39% of the cost of one transport in a child with chronic lung disease (CLD) and 78% of the three transports in children born before 33 weeks’ gestation with no CLD. Palivizumab could have prevented 39% of the 21 hospital days of the three children with CLD and 78% of the 118 hospital days of the 17 children without CLD. Therefore, palivizumab could have reduced hospitalization, physician and transport costs to $47,741 from $168,888. The net cost of using palivizumab was $632,153.

Therefore, if only direct costs are considered and the assumption is made that palivizumab would have prevented 78% of RSV hospitalization costs in infants with prematurity alone and 39% of costs in infants with CLD, then the net cost of following the CPS guidelines on the use of palivizumab would have been $632,153.

The net cost of following the AAP guidelines would have been $825,771 because 34 more infants would have required prophylaxis and only 16 hospitalization days occurred in infants who fulfilled the AAP criteria, but not the CPS criteria for palivizumab prophylaxis.

Table 4 shows the cost calculations if palivizumab was offered to different subgroups. For all infants who met the CPS criteria to receive palivizumab, the net cost to prevent one RSV admission was $41,746 and the cost to prevent one day of hospitalization was $6,307. In total, 10.5 infants would have to receive prophylaxis to prevent one RSV admission. Even in the highest risk group (infants with CLD discharged during the RSV season), 6.8 infants would have to receive prophylaxis to prevent one hospitalization. The net cost of using the drug always exceeds the cost of preventable hospitalizations.

TABLE 4:

Net cost of administering palivizumab to infants discharged from Edmonton hospitals assuming the drug would have been given monthly only to a selected group between November 1, 1998 and April 1, 1999 starting at the time of discharge

Group selected to receive palivizumab	Number needed to treat to prevent one hospitalization	Cost/day of hospitalization avoided ($)	Net cost of palivizumab use ($)
Infants who fulfill CPS guidelines (n = 159)	10.5	6307	632,153
Infants who fulfill AAP guidelines (n = 192)	11.7	7792	825,771
Infants with CLD (n = 13)	11.1	6124	50,165
Infants born before 33 weeks’ gestation, on oxygen on day 28, and discharged September, October or November 1998 (n = 8)	6.8	4670	36,412
Infants who were 44 weeks’ postconception or less at time first dose is given (n = 102)	9.0	4710	414,113
Infants who are part of a multiple gestation (n = 52)	22.2	20,100	235,193

AAP American Academy of Pediatrics; CLD Chronic lung disease; CPS Canadian Paediatric Society

Thirteen of the 21 RSV hospitalizations occurred in December or January. If each infant received palivizumab only during those two months, then the net cost of using the drug would have been $158,123.

Comparison with other RSV seasons

Table 5 shows the number of RSV hospitalizations to the ward and to the ICU for the CHC over the past five years. The number of hospitalizations during the year studied was higher than that in the previous four years, and the number of ICU admissions was similar.

TABLE 5:

Total respiratory syncytial virus hospitalizations at the Children’s Health Centre from July 1 to June 30 of the following years

Year	Admissions to ward	Admissions to intensive care unit
1998/99	600	46
1997/98	265	12
1996/97	469	49
1995/96	324	21
1994/95	452	48

DISCUSSION

The net cost of using palivizumab during the 1998/99 RSV season, as per the CPS guidelines, in the 159 study infants discharged from CHC sites would have been $632,153. The net cost to prevent an RSV admission was $41,746 and the cost to prevent one day of hospitalization was $6,307.

Limitations of the study methodology

Because costs are not calculated for the individual patient in the Canadian health care system, it was not possible to obtain the true cost of hospitalization or transport for any of the patients. It is not possible at the authors’ centre to separate out per diem costs on the ward compared with in the PICU. Even taking into account inflation, the daily hospitalization rate of $1,062 is higher than the rate of $564/day (US$412/day times the 1993 exchange rate of 1.37) calculated in the Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) study evaluating the economic costs of paediatric RSV infection in different Canadian centres (10). If we had used the hospitalization costs of $16,500 (US$11,000) per RSV hospitalization, as calculated in a study done in Rochester, New York (11) (where costs are almost certainly higher than in the Canadian system), the hospitalization costs saved by the use of palivizumab would have increased to $250,965, but this cost-saving still would have been only 33.3% of the cost of palivizumab.

RSV hospitalization rates can vary considerably from year to year. The number of RSV hospitalizations in the authors’ region during the year studied was higher than that in the previous four years, and the number of ICU admissions was similar; therefore, it is unlikely that the net cost of administering palivizumab would have been lower in any of the previous four years.

Ideally, a physician would also look at the cost to parents of having a child hospitalized with RSV or of having to make time to take their child for palivizumab injections. However, these indirect costs would be less than 10% of the direct costs of hospital admission and would not have significantly altered the results of the study.

The fact that the hospitalizations for respiratory illness in children who were not tested for RSV (a total of six days of hospitalization) were included may have overestimated RSV hospitalization costs. The present calculations may have also overestimated the efficacy of palivizumab. In the original study (3), palivizumab prevented 78% of RSV admissions in infants born before 36 weeks’ gestation with no CLD, and the authors used an efficacy rate of 78% because there are no published data on efficacy in infants born before 33 weeks’ gestation without CLD. Furthermore, the assumption that less than one full 100 mg vial of palivizumab would be used for each dose underestimates true drug costs, because an opened vial must be used within 6 h, so, the assumption can be made that the drug would often be wasted because it is difficult to arrange for multiple infants to be given the drug on the same day. The recent addition of 50 mg vials will decrease this waste to some degree. The authors’ definition of CLD (requirement for oxygen supplementation at 36 weeks’ corrected age) differs from the definition used in the original study (medical therapy for CLD within six months of study entry).

Comparison with previous studies

Two other studies of the potential cost effectiveness of palivizumab in patients from the Kaiser Permanente-Northern California health maintenance organization (HMO) (6) and the Rochester area (11) also concluded that palivizumab is far from cost effective when used according to the AAP guidelines. It has been suggested that the Kaiser Permanente study could have missed RSV hospitalizations in the study population because identification of RSV hospitalizations depended on accurate coding by medical records, RSV hospitalizations were only included if they occurred between December and March, nosocomial RSV or second RSV hospitalizations in the same RSV season were excluded, patients could have been insured by more than one HMO and admitted elsewhere, and not all children admitted with compatible symptoms would have been tested for RSV (12). The Rochester study (11) extrapolated data from one hospital to estimate the total number of RSV hospitalizations.

The present study’s methodology overcomes the above difficulties. By telephoning parents, more complete follow-up was achieved. The authors looked at RSV admissions year-round and would have identified nosocomial cases or repeat hospitalizations. All children admitted to hospitals in Edmonton with compatible symptoms are tested for RSV because it allows infected children to be placed into cohort groups, and it was establised that for admissions for respiratory illness in hospitals outside Edmonton, testing for RSV occurred in all but one case. It seems unlikely that children whose parents could not be contacted would have had a higher RSV hospitalization rate than the study patients. However, even if all of the 47 infants excluded from the study were hospitalized with RSV outside the CHC, and the assumption is made that their hospital costs mirrored that of the 21 hospitalized infants included in the study, then the costs of hospital-based care are still less than the cost of palivizumab.

A study of RSV-IGIV found the product to be cost effective, despite a cost similar to palivizumab (13). This conclusion was reached by comparing RSV costs in the years the product was used with previous years. However, the burden of RSV illness can vary considerably from year to year, and seems to be decreasing in preterm infants (14). RSV-IGIV was used in infants whom the attending physicians perceived to be high risk rather than in all infants who fulfilled the AAP criteria for use of RSV-IGIV. The main reason the product was cost effective is that it appeared to have a markedly higher efficacy than the original trials, and the average number of doses given was only 2.35/patient. The bulk of RSV admissions in our region usually occur over a two-month period, but this two-month period varies from year to year and seems to occur later in northern communities in the authors’ region. Therefore, it seems unlikely that prophylaxis could be limited to two or three doses in the authors’ setting. The present study showed that even if the two months during which the most RSV hospitalizations occurred could have been predicted, the net cost of using the drug would still have been substantial.

Reasons why palivizumab was not cost-saving

The only way that palivizumab could be cost-saving in the authors’ centre for the population for which it is recommended is if the cost of the drug were to be about 16.1% of the current cost. This may be because there has been a trend toward a decrease in RSV admissions in more recent studies of premature infants with or without CLD compared with older studies (14). This decrease in admissions could be due to the improved education of parents about prevention of RSV and increased use of oximeters and home oxygen, obviating the need for hospitalization. Also, the use of a surfactant and improved methods of ventilating preterm infants may result in infants being discharged with less lung damage. Our RSV hospitalization rate of 13.2% in infants born before 33 weeks’ gestation or with CLD is comparable with recent studies (14,15), but is considerably lower than the 20% RSV hospitalization rate in a study done in 1990/91 (14). Palivizumab may have been cost-saving a decade ago when RSV morbidity was more severe in preterm infants.

Patient populations where palivizumab may be cost-saving

Infants with CLD consistently have higher RSV admission rates than infants with prematurity alone, and are more likely to require ventilation and a prolonged hospital stay (14,16). There is no uniform definition of CLD, and it is possible that the efficacy of palivizumab in the original study (3) would have been more impressive if a definition was chosen that selected infants with more severe pulmonary disease, rather than specifying clinical CLD with medical treatment in the preceding six months. One study of the cost effectiveness of palivizumab determined that if a physician treated only infants with CLD (as defined by the need for oxygen on day 28 of life) born before 33 weeks’ gestation who were discharged in September through November, the number of patients that needed to be treated to prevent one RSV hospitalization was only 7.4, and the cost of preventing that hospitalization fell to $18,000 (6). In the present study, only eight infants fulfilled these criteria and three were readmitted with RSV (Table 4). Assuming that the drug has 39% efficacy in infants with CLD, it would have cost $4,670 to prevent one day of hospitalization in this high-risk group.

A Canadian study (16) and a study of Alaskan natives (17) found Aboriginal race to be a risk factor for severe RSV disease, even in the absence of prematurity. A previous study (18) and the authors’ review of local ICU RSV admissions suggest that children with tracheoesophageal fistula sometimes have prolonged RSV hospitalizations. Another high risk group is undoubtedly preterm infants younger than six months of age who are in child care centres or are members of large sibling groups. Further studies should be done on the incidence of RSV hospitalization in these groups to see whether palivizumab would be cost-saving in any of these patient populations.

Comparison with other disease prevention methods

Economic evaluations of preventive measures in other diseases have been performed. A comparison of many of these studies with the present study is difficult given the differences in how costs were calculated. However, a study considering the vaccination of influenza in preschool children showed a marginal cost savings when comparing direct costs of disease treatment to the cost of vaccination (19). Likewise, a Canadian study of the two-dose measles immunization in children showed benefit to cost ratios of 1.30 to 1 to 2.95 to 1 under a variety of sensitivity analyses comparing direct benefits to vaccine cost (20). The difference in net cost between these studies (19,20) and the present study is most likely due to the high cost of palivizumab prophylaxis of $4,738/person, relative to the cost of vaccination for which the cost of the influenza vaccine in the first study was estimated to be $21.80 (US$14.53) per person (19).

CONCLUSIONS

Palivizumab would not have been cost-saving to the health care system in the authors’ region in 1998/99 for the patients included in the study according to the CPS or AAP guidelines. A group of patients could not be selected where palivizumab would have been cost-saving. In the future, it may be shown that palivizumab can decrease the incidence of reactive airways disease or other possible long term sequelae of severe RSV infection. In the meantime, the decision to use the drug should be based on whether it is worth spending $6,307 to prevent the discomfort of the patient and cause stress to a family that is associated with one day of hospitalization for the management of RSV infection.