Abstract
OBJECTIVE: We assessed the effects of very preterm birth (gestational age <32 weeks or birth weight <1501 g) and prematurity-related morbidities on health care costs during the fifth year of life.
METHODS: The study population consisted of 588 very preterm children and 176 term control subjects born in 2001–2002. Costs of hospitalizations, visits to health care professionals and therapists, and the use of other social welfare services were assessed during the fifth year of life. Hospital visits were derived from register data and other health care contacts, and the use of social welfare services were derived from parental reports. The effects of 6 prematurity-related morbidities (cerebral palsy [CP], seizure disorder, obstructive airway disease, hearing loss, visual disturbances or blindness, and other ophthalmologic problems) on the costs of health care were studied.
RESULTS: The average health care costs during the fifth year of life were 749€ in the term control subjects, 1023€ in the very preterm children without morbidities, and 3265€ in those with morbidities. The costs of social welfare services and therapies exceeded the hospitalization costs in all groups. Among children who were born preterm, CP was associated with 5125€ higher costs, whereas later obstructive airway diseases increased the costs by 819€ compared with individuals without these morbidities.
CONCLUSIONS: The health care costs during the fifth year of life in very preterm children with morbidities were 4.4-fold and in those without morbidities 1.4-fold compared with those of term control subjects. This emphasizes the importance of prevention of morbidities, especially CP, to reduce the long-term costs of prematurity.
WHAT'S KNOWN ON THIS SUBJECT:
The hospitalization costs for very preterm infants are high during the first year of life and have been shown to decrease markedly with age. Few studies, however, have evaluated other health care–related costs in children who were born very preterm.
WHAT THIS STUDY ADDS:
This study demonstrates that although the costs of hospitalizations decrease with age in very preterm children, other health care–related costs, such as costs for social welfare services and therapies, increase and thereby become more considerable.
The costs of the initial hospitalization for very preterm infants are high,1,–,3 and the costs increase with decreasing gestational age (GA) and birth weight.2,–,5 In addition, very preterm infants continue to need more health care services than term infants later in childhood.6,–,9 During the first 3 years of life, the need for hospital care clearly decreases more in very preterm children without prematurity-related morbidities than in very preterm children with these morbidities.10
Only a few studies have explored the costs of hospitalizations of very preterm children after the first year of life, and other health care costs have barely been evaluated. In the study by Petrou et al,8 the hospitalization costs significantly decreased after the first year. The fifth-year costs amounted to 1% to 2% of the costs of the first year of life8; however, the costs at 11 years of age were still higher in extremely preterm children compared with the control children when including also educational and social welfare costs.11 Our study on the total hospitalization costs during the first 4 years of life for children who were born very preterm showed similar results: the initial hospital costs composed 79% of the total costs, and thereafter the costs decreased each year.12 In addition, individuals with prematurity-associated morbidities used significantly more hospital resources during the first 4 years of life than those without these morbidities.
Mangham et al13 recently modeled the costs of preterm birth to the public sector during the first 18 years of life and concluded that the initial hospital stay composes 92% of the incremental costs per preterm survivor. Because prematurity is associated with increased chronic morbidity and disability,14,–,16 costs other than hospitalization should be taken into account. The aim of our study was to assess the costs of inpatient and outpatient hospital care, primary care, social welfare services, and therapies during the fifth year of life for very preterm infants and compare them with the costs of healthy control subjects. We hypothesized that the need and costs for all health care resources are significantly higher for children who were born very preterm with prematurity-associated morbidities compared with children without these morbidities.
METHODS
The study protocol was approved by the ethics committee of the National Institute for Health and Welfare.
Study Population
We included all surviving children who were born at <32 weeks' GA or had a birth weight <1501 g during 2001–2002 in Finnish hospitals that have level 2 or 3 NICUs as defined by the American Academy of Pediatrics Committee on Fetus and Newborn.17 Healthy gender-matched term infants (GA 38–42 weeks) who were born next after every third study infant in the same delivery hospital were selected for control subjects. Those with incomplete Medical Birth Register or Hospital Discharge Register data (n = 181) and those who were born at a hospital with <3 very preterm deliveries in 2001–2002 (n = 4) were excluded from the analysis. A total of 23 children who were born very preterm and 13 control subjects were excluded because they were living abroad or had a missing address. Infants who were admitted to the NICU during the first 7 days of life were excluded from the control group. The final study population consisted of 901 very preterm children and 368 control subjects.
Data Collection
Register data used in this study were collected from the National Medical Birth Register and the Hospital Discharge Register, both maintained by the National Institute for Health and Welfare. The Hospital Discharge Register contains data on all inpatient and outpatient hospital visits in Finland. A parental questionnaire was sent 0.5 to 1.5 months before the child's fifth birthday, and 2 reminders were sent 1.5 and 2.5 months thereafter if needed. The questionnaire enquired about any long-term diagnoses of the child, the number of visits to various health care professionals during the last 12 months, the family structure, and parental education and current employment. The parents of 588 (65% of all surviving) children who were born very preterm and 176 (46%) control subjects returned the questionnaire. A dropout analysis (Table 1) showed no systematic bias in the study population.18
>Background Variables of Responders and Nonresponders in the Very Preterm Children
The study individuals were defined as having a prematurity-related morbidity when ≥1 of the studied morbidities, which have been shown to be overrepresented in preterm populations, were reported at least once to the Hospital Discharge Register by the end of 2006. The morbidities and the International Statistical Classification of Diseases and Related Health Problems, 10th Revision codes of the morbidities are presented in Table 2.
Prematurity-Related Morbidities and Their ICD-10 Codes
Diagnoses for both inpatient and outpatient visits in hospitals were recorded. On the basis of the health care system in Finland, diagnoses of these disease groups can be reliably derived from hospital registers, because these diseases are diagnosed and treated in specialized care in public hospitals. There are no private children's hospitals in Finland. The diagnoses from the register data were in accordance with the diagnoses reported by the parents, which support the validity of the register data.18
Cost data for outpatient visits and hospital care after the initial hospitalization were collected from 4 hospitals: 1 level 3 hospital and 3 level 2 hospitals. The costs for hospital stays were estimated for those with missing cost data on the basis of the type of visits (emergency outpatient visit, nonemergency outpatient visit, and inpatient hospital visit) and the child's age at the time of the visit. The costs for municipal health and social welfare services were calculated according to reference costs determined by the National Institute for Health and Welfare.19
Statistical Analyses
We analyzed the number of visits during the fifth year of life to a (1) physician, (2) nurse practitioner, (3) physiotherapist (PT) or occupational therapist (OT), (4) psychologist, (5) speech therapist, (6) dietician, and (7) other services (family support clinic, family adaptation courses, training in sign language, home visits, and use of communal transfer services) on the basis of the parental questionnaire responses. The number of inpatient days was received from the register data, as well as the number of nonemergency and emergency outpatient visits to specialized health care. The register data on hospital visits and the parental reports on the contacts with other health and social care professionals and use of other municipal resources were combined with the cost data to obtain a net cost per child during the fifth year of life. The patient-level data were linked by means of unique encrypted identification codes.
We analyzed the costs and the number of visits according to GA groups (23 weeks, 24–25 weeks, 26–27 weeks, 28–29 weeks, 30–31 weeks, ≥32 weeks, or term) and according to the presence of the morbidities in the very preterm population (no morbidities, ≥1 prematurity-related morbidity). Analysis of variance was used to detect cost differences between these groups. P < .05 was considered significant. In addition, we used a generalized linear model with γ distribution to analyze the effect of morbidities on the costs of health care use during the fifth year of life. The model was adjusted for gender, GA at birth, intrauterine growth (small, appropriate, or large for GA), and multiple pregnancy. The marginal effect describes how the presence of the studied morbidities affects the costs during the fifth year of life. The costs are presented in 2008 prices, and the discount rate was 3% per year. The analyses were performed with SAS 9.1 (SAS Institute, Cary, NC) and Stata 9 (Stata Corp., College Station, TX).
RESULTS
Of the very preterm children, 68% did not have any of the studied morbidities. The number of visits to health care specialists and the number of hospital days are presented in Table 3. The most common health care contact in all study individuals was with a physician in either the public or the private sector. Very preterm children with prematurity-related morbidities had more contacts with all health care professionals except nurse practitioners compared with children without these morbidities.
Number of Visits to Hospitals and Health Care Professionals During the Fifth Year of Life According to Morbidities and GA (weeks) in the Very Preterm Population and the Term Control Population
The average total health care costs during the fifth year of life amounted to 749€ in control subjects, 1023€ in very preterm children without prematurity-related morbidities, and 3265€ in those with ≥1 of these morbidities (Table 4). Thus, the costs of preterm children without prematurity-related morbidities were 1.4-fold higher than the costs of term control subjects; however, the costs of the preterm children with morbidities were 4.4-fold higher compared with those of the term control subjects. They composed 61% of the total costs in the very preterm population. The mean costs of all cost categories except the visits to the nurse practitioner were lower for very preterm children without prematurity-related morbidities compared with those with prematurity-related morbidities.
Costs of Visits to Hospitals and Health Care Professionals During the Fifth Year of Life According to Morbidities in the Very Preterm Population and the term Control Population
The costs of hospital care composed only 33% of the total costs in the very preterm population. The costs for therapies (PT/OT, psychologist, speech therapy) composed 44% of the total costs in the very preterm children with morbidities, 27% in those who were born very preterm without morbidities, and 30% in the control subjects. The hospital inpatient and outpatient costs composed 37% of the total costs in very preterm infants with morbidities, 34% in those without morbidities, and 17% in the control subjects. The costs decreased with increasing GA (Fig 1).
Mean health care–related costs during the fifth year of life according to GA. (Those born at 23 weeks' GA were excluded from the figure because of the small number [5] of individuals.)
Among children who were born very preterm, cerebral palsy (CP) was associated with 5127€ higher costs; later obstructive airway diseases increased the costs by 819€ compared with individuals without these morbidities. The cost increases that were attributed to each disease group according to the generalized linear model are presented in Table 5. In addition, the average cost for boys was 450€ higher than for girls (P = .019).
Generalized Linear Model on the Effect of Morbidities on the Total Costs of Care During the Fifth Year of Life in Very Preterm Infants (N = 588)
DISCUSSION
The two-thirds of very preterm infants who survive without prematurity-related long-term morbidities incurred only a little more health care–related cost than their term peers; however, prematurity-related morbidities were still associated with a significant cost burden during the fifth year of life. These costs were created especially by PT/OT visits, nonemergency outpatient hospital visits, and inpatient hospital care.
The yearly hospitalization costs, which were 623€ per child for very preterm children, were lower during the fifth year of life compared with third and fourth years of life (1179€ and 776€, respectively12); however, the costs for therapies, primary care, and visits to private practitioners during the fifth year of life exceeded the costs for hospitalizations in all of the 3 study groups. It seems, thus, that although the costs of hospitalizations decrease with age in children who are born very preterm, other health care–related costs, such as costs for social welfare services and therapies, increase and thereby become more considerable; therefore, measuring only hospitalization costs will significantly underestimate the later costs of prematurity.
Morbidities, especially CP, increased greatly the costs of the very preterm population during the fifth year of life. Parallel to our results, Stevenson et al20 showed in a small regional cohort of low birth weight children that children with disabilities accounted for a disproportionately high amount of the total expenditure (hospital inpatient and outpatient care, visits to general practitioner, special education services) up to age 8 to 9 years. Similarly, Petrou et al8 showed that several different morbidity groups were associated with higher costs during the first year of life, but the study focused merely on the costs of hospitalization. Thus, calculating the cost-effectiveness of care of all very preterm infants as 1 group can be misleading, because there is now evidence that children who are born very preterm without prematurity-related morbidities do not cause significant additional costs for public services after the initial hospitalization compared with infants who are born healthy at term. In contrast, individuals with prematurity-related morbidities not only consume more health care resources but also induce more social and productivity costs. For example, in a recent Danish study21 on the lifetime costs of CP, two-thirds of individuals with CP never entered the labor market, and the average lifetime cost of CP was calculated to be 860 000€ for men and 800 000€ for women, the social costs accounting for the majority of the costs. These facts underline the importance of effective perinatal care to prevent CP in the preterm population. The cost-effectiveness of primary prevention of CP in perinatal care should be evaluated carefully to attain optimal use of resources.
The major strength of this study is the use of data on true hospital visits for the whole population, because all hospital care in Finland is comprehensively registered in the national registers.22 By using a national cohort of preterm infants as a study population, the selection bias in studies that are based on a smaller geographic region or a single hospital is avoided. We defined the 6 most common prematurity-related morbidity groups. Very preterm infants may also have other morbidities; however, because the costs in the very preterm group without morbidities were very similar to the costs in the term group, it can be assumed that these 6 groups include the diseases and conditions that entail the most resource use.
A limitation of this study could be that the data on the use of therapies and social welfare services obtained from the questionnaires rely on parental recall; however, because the study period was 1 year only, we assume that the parents remembered the health care and therapy contacts well, especially when there were only a few contacts in most cases. In addition, the morbidities documented by the parents were in accordance with the morbidities from the registers, which supports the validity of the data from the parental reports. Another potential limitation is that the hospitalization costs were calculated according to data from 4 hospitals, so we did not have actual cost data for all of the hospitalizations; however, it is unlikely that the costs would differ significantly between hospitals because all hospitalizations occurred in the public sector, because there are no private children's hospitals in Finland. Because Finnish children start school at 7 years of age, the special education costs do not exist yet at 5 years of age. Although all other early intervention is included, it is likely that special education will add the cost burden in later childhood.
Despite that costs for social welfare services and therapies exceeded the hospitalization costs in our study, the total costs during the fifth year of life were still low compared with the initial hospitalization costs, which were 54 000€ in the same very preterm population.12 This is in accordance with previous studies, which showed that the use of health care resources declines with increasing age in individuals who were born preterm.13,23 Thus, the first-year hospitalization costs and, in particular, the initial hospital stay compose the great majority of the total costs of care of very preterm infants.
CONCLUSIONS
The health care–related costs during the fifth year of life in very preterm children without prematurity-related morbidities did not differ greatly from the costs of children who were born term; however, the costs of the very preterm children with morbidities, especially CP, were 4.4-fold higher compared with those of the very preterm children without prematurity-related morbidities. Additional prevention of morbidities such as CP would thus significantly reduce the long-term costs of prematurity. In addition, our study indicates that when estimating the costs of prematurity after the first year of life, one should calculate not only the hospitalization costs but also other costs for social welfare services, primary care, and therapies, because these exceed the hospitalization costs in very preterm infants during the fifth year of life.
Footnotes
- Accepted January 6, 2010.
- Address correspondence to Emmi Korvenranta, MD, MSc, Turku University Hospital, Department of Pediatrics, Kiinamyllynkatu 4-8, 20520 Turku, Finland. E-mail: emmi.korvenranta{at}utu.fi
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- GA =
- gestational age •
- PT =
- physiotherapist •
- OT =
- occupational therapist •
- CP =
- cerebral palsy
REFERENCES
- Copyright © 2010 by the American Academy of Pediatrics