Article Text

Download PDFPDF

Continuing utilisation of specialised health services in extremely preterm infants
  1. T M Luu1,
  2. F Lefebvre1,
  3. P Riley2,
  4. C Infante-Rivard3
  1. 1Department of Pediatrics, Centre Hospitalier Universitaire Sainte-Justine, Montreal, Canada
  2. 2Department of Pediatrics, Montreal Children's Hospital, Montreal, Canada
  3. 3Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Canada
  1. Correspondence to Dr Thuy Mai Luu,Department of Paediatrics, CHU Sainte-Justine, 3175 Côte-Ste-Catherine, Montreal, Quebec H3T 1C5, Canada; thuy.mai.luu{at}


Objective To compare healthcare use from neonatal discharge to 18 months corrected age (CA) of two groups of extremely preterm neonates (23–25 vs 26–28 weeks).

Design Cohort study.

Patients Infants born at ≤28 weeks and admitted in three hospital centres in Quebec, Canada (n=254).

Main outcome measures Neurodevelopmental outcomes and healthcare use from neonatal discharge to 18 months CA.

Results Re-hospitalisation rates occurred in 57% of children born at 23–25 weeks and in 49% of those born at 26–28 weeks. In these two age groups, by 18 months, 61% vs 59% were followed in physical or occupational therapy, 29% vs 17% were enrolled in a long-term rehabilitation program, 38% vs 28% used prescribed medication, and 59% vs 33% required medical assistive devices, respectively. Risk of re-hospitalisation was related to bronchopulmonary dysplasia (BPD), severe brain injury, use of home oxygen or an apnoea monitor and older age at neonatal discharge. Multiple births, BPD, severe brain injury, older age at neonatal discharge and single parenthood were associated with risk of using out-patient health services above average (>2 services).

Conclusion Extremely preterm children are frequently re-hospitalised during infancy and use a substantial amount of healthcare resources. These results highlight the importance of resource allocation to preterm infants for medical and rehabilitation services after discharge from the neonatal intensive care unit.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Medical progress by modern obstetric and perinatal care has led to an increase in preterm birth along with improvement in survival of the sickest and most immature babies.1 The rising number of survivors of prematurity, often affected with neurodevelopmental disabilities, translates into a significant increase in health and educational costs.

In the USA, annual out-patient costs per extremely preterm survivor (<28 weeks of gestation) is estimated at $9356 for the first year of life compared to $1430 for a normal-term infant.2 Preterm infants are more often referred to medical specialists and for rehabilitation therapies, in particular physical and occupational therapy, as well as to special education programs.3,,5

In addition, acute care visits and re-hospitalisations in infancy among preterm children account for a significant burden on health resources. In the first year following initial discharge from the neonatal intensive care unit (NICU), from 40% to 70% of preterm infants require re-admission to hospital, mostly for respiratory illness or feeding problems.6 7

What is already known on this topic

Preterm children have increased rates of neurodevelopmental disabilities which likely translate into higher health resource use, the amount of which is not well documented.

What this study adds

  • ▶. Half of extremely preterm survivors are re-hospitalised during infancy.

  • ▶. The use of out-patient health services, including medical assistive devices at home, is substantial.

Characterising healthcare service use by preterm children with short-term or chronic conditions is important to guide planning of health resource allocation. Moreover, a better understanding of pattern of health resource use may facilitate care coordination after discharge from the NICU.

This study describes healthcare resource use by extremely preterm infants in the province of Quebec. The objective was to compare rates of (a) re-hospitalisation, (b) out-patient healthcare service use and (c) prescribed medication and assistive technology use from initial neonatal discharge to 18 months corrected age (CA) between children born at the limit of viability (23–25 weeks of gestation) and those born at 26–28 weeks.



All infants born alive at gestational age 286/7 weeks or less at Centre Hospitalier Universitaire Sainte-Justine (Montreal), McGill University Health Center (Montreal) and Centre Hospitalier Universitaire de Québec (Quebec City) from 1 January 2003 to 31 December 2004 who survived until discharge were eligible for follow-up as part of the Research Consortium on Children Born Extremely Preterm, a population-based cohort including extremely preterm infants followed in the five main perinatal centres in the province of Quebec (Canada). For the purpose of the current study, data from three out of the five centres were available for analysis. The three follow-up centres provide care to approximately 75% of extremely preterm children in Quebec (personal communication). No infants fulfilling the gestational age criterion were excluded. The study was approved by the institutional review boards of the three participating centres that provided data for the study.

Data collection

Perinatal data were abstracted from the medical chart. Gestational age status was based on the best obstetric estimate (early ultrasound or date of the last menstrual period). Other variables of interest included birth weight, small for gestational age status (birth weight below the tenth percentile according to Canadian norms),8 sex, multiple births, use of antenatal steroids, occurrence of bronchopulmonary dysplasia (BPD) defined as oxygen use at 36 weeks postconceptional age, severe brain injury (grade 3 and 4 intraventricular haemorrhage, periventricular leukomalacia and hydrocephalus on cranial ultrasounds, defined using Levene ventricular index >97% based on the ultrasound nearest 6 weeks of life), severe retinopathy of prematurity (ROP) (stage 4 and 5, plus disease, or requiring invasive ophthalmologic intervention) and culture-proven septicaemia.

At 18 months CA, children and families received a comprehensive assessment. A standardised questionnaire administered to parents by either nurses or physicians from neonatal follow-up clinics inquired about sociodemographic characteristics, hospitalisations and surgeries in the previous months, use of out-patient healthcare services, prescribed medication in the 2 months preceding the visit, and home assistive technologies. The child's medical chart was reviewed to check accuracy of the information provided by the parents if services were delivered in the hospital where neonatal follow-up took place. Out-patient healthcare services included any involvement still active or discontinued at the time of the 18-month visit with healthcare workers from NICU discharge until 18 months CA. New referrals made after the visit were recorded separately. We did not include visits for primary care, ophthalmologists, audiologists or neonatal follow-up clinics as all preterm infants were receiving routine care from these specialists. We recorded referral or active involvement in a long-term rehabilitation centre, which provides rehabilitation services (such as physical, occupational or speech therapy) to children with persisting motor or language deficits. Use of out-patient services above the average was defined as active involvement with two or more healthcare workers at any time between NICU discharge and the 18-month visit. This cut-off was chosen as it is routine practice for preterm children to be followed by either a physical or occupational therapist for the first months of life to monitor motor development.

Anthropometric measures were taken. Paediatricians experienced in neonatal follow-up performed a standard neurological exam with assessment of passive and active muscle tone, reflexes, gait and abnormal movements. Cerebral palsy was defined as a movement and posture abnormality causing activity limitation and secondary to a non-progressive central nervous system disorder.9 Certified psychologists trained to administer the Bayley Scales of Infant Development, 2nd edition (BSID-II) performed the developmental assessment based on CA.10 Mean scores for this test are set at 100 with a SD of 15. Assessors were aware of the neonatal course of the children they evaluated, as they were also involved with the families as regular caregivers.

Statistical analyses

Simple descriptive statistics were calculated by gestational age groups (23–25 weeks vs 26–28 weeks). We used OR and 95% CI to estimate the risk associated with use of specific healthcare services contrasting the two age groups. We adjusted for follow-up site because of its possible association with neonatal characteristics (because of different neonatal practices) and availability of health resources.

We used logistic regression analyses to explore perinatal and social factors associated with risk of (1) re-hospitalisation and (2) use of two or more out-patient healthcare services from NICU discharge to 18 months CA. The studied factors were selected based on the scientific literature. They were birth weight, gestational age, small for gestational age, male sex, antenatal steroids, multiple births, BPD, severe brain injury, severe ROP, length of neonatal stay, home oxygen, apnoea monitor, minority status, siblings, maternal education, household structure and source of income. Infants who were discharged from the NICU after 6 months CA (n=9) were excluded from the analysis for re-hospitalisation. Analyses were conducted with SAS 9.1.


In 2003–2004, 403 extremely preterm infants were admitted at either one of the three participating centres; 331 survived to 18 months CA (82%), and one of the 72 deaths occurred after discharge. Follow-up and consent was obtained for 254 infants (follow-up rates of 75% and 77% for the 23–256/7 group and the 26–286/7 group, respectively). Perinatal characteristics of participants and non-participants are shown in table 1.

Table 1

Comparison of neonatal and social characteristics of participants and non-participants

Among the children included in the analysis, mean gestational age was 27 (1) weeks, and mean birth weight was 962 (205) g. Almost one-fifth were born at the limit of viability (23–25 weeks). Children born at 23–25 weeks were discharged at an average age of 42 weeks postconception compared to 39 weeks in the 26–28-week group. Table 2 describes medical and developmental outcomes at 18 months CA by gestational age groups, showing some increased risks for the 23–25 weeks age group (in particular, growth <3rd percentile and any neurodevelopmental impairment) that did not reach statistical significance.

Table 2

Crude OR and 95% CI for medical developmental disorders at 18 months corrected age for preterm infants born at 23–256/7 weeks of gestation versus 26 to 286/7 weeks of gestation

Re-hospitalisation and surgeries

Table 3 compares hospital re-admission after discharge from the NICU until 18 months CA between children born at 23–25 weeks and those born at 26–28 weeks of gestation. Of 253 children with data, 117 required at least one re-hospitalisation for a total of 221 admissions, mostly for respiratory illness (63%), which included airway infections, breathing problems and BPD-related morbidities. Other reasons were surgery (16%), gastrointestinal/nutrition problems (11%), infections other than respiratory (4%) and neurological problems (2%). No difference in rates of re-hospitalisation was detected between the two gestational age groups.

Table 3

Re-hospitalisation from neonatal discharge to 18 months corrected age by gestational age groups

Out-patient healthcare service use

The majority of extremely preterm infants received out-patient health services besides the primary care physician, the neonatal follow-up paediatrician, the audiologist and the ophthalmologist (table 4).

Table 4

Use of out-patient services by gestational age groups: visits to medical specialists or rehabilitation therapists from neonatal discharge to 18 months corrected age

Sixty-two per cent of children in the lower gestational age group were followed by two or more healthcare practitioners compared to 46% of the higher gestational age cohort. Physical and occupational therapists provided developmental monitoring to 59% of the entire cohort. At 18 months CA, 13 (27%) infants born at 23–25 weeks and 36 (18%) infants born at 26–28 weeks were still receiving physical therapy, whereas 10 (20%) and 37 (18%) were still involved with occupational therapy for the 23–25-week and 26–28-week groups, respectively. For the entire cohort, the proportion of infants awaiting physical therapy was 9% and occupational therapy 16% (not shown). Only 19 of 254 children had been assessed by a speech therapist. However, following the 18-month visit, a consultation for speech therapy was requested for 40% of the cohort (not shown). Need for a long-term rehabilitation program was identified for 29% of the children born at 23–25 weeks compared to 17% of those born at 26–28 weeks. Overall, comparable use of all out-patient services was observed between the lower and higher gestational age groups.

Medication and medical assistive technology use

Regular use of prescribed medication over the 2 months preceding the visit was reported by 30% of parents (table 5). Almost one-quarter were taking bronchodilators and inhaled steroids for respiratory problems.

Table 5

Use of special healthcare resources by gestational age groups: use of prescribed medication and assistive devices from neonatal discharge to 18 months corrected age

Infants born at 23–25 weeks were almost four times more likely to require some medical assistive technology at home compared to those with a gestational age of 26–28 weeks (table 5). Among the most premature infants, 43% were sent home on oxygen, and 10% had an apnoea monitor.

Relationships between perinatal and social variables and risk of re-hospitalisation and use of out-patient services above average

Table 6 shows that the risk of re-hospitalisation was associated with BPD, severe brain injury, use of home oxygen or an apnoea monitor and older age at NICU discharge (for each additional week in the NICU, risk of re-admission increased by 10%). We did not find any association between unfavourable social environment and re-hospitalisation. The second model examined the risk of using out-patient health services above average in relation to various perinatal and social factors (table 6). Multiple births, BPD, severe brain injury, older age at NICU discharge and single-parent household (used as a proxy measure of unfavourable social environment) were associated with the outcome.

Table 6

Child and family characteristics associated with risk of re-hospitalisation and risk of using out-patient health services above average (two or more)


During infancy, extremely preterm children born in Quebec utilise a substantial amount of healthcare resources following discharge from the NICU: re-hospitalisation is frequent, numerous out-patient healthcare providers are involved, and use of prescribed medication or home assistive devices is common.

Our results show that 46% of children required hospital re-admission, primarily due to respiratory illness. This number is comparable to other studies across the globe. Gray et al found a re-admission rate of 53% in children <32 weeks in New Zealand.11 Similarly, among infants seen at 18–22 months of age by the Neonatal Research Network in the USA, 49% had been admitted overnight to a hospital ward.12 During infancy, respiratory disease is the leading cause of re-hospitalisation both in preterm and term infants.4 11,,13 Preterm infants with BPD, especially if on home oxygen, are particularly vulnerable.6 12 14,,16 This finding is not surprising given that BPD is characterised by an arrest of lung development leading to pulmonary dysfunction with increased susceptibility to severe respiratory infections and bronchospasm during early childhood.17 We also found that severe brain injury, prolonged neonatal stay and use of an apnoea monitor were associated with risk of re-hospitalisation. However, it is not possible from these data to determine whether these three factors reflected a more vulnerable population at higher risk of re-hospitalisation or referral bias. Infants with multiple neonatal morbidities are more likely to be admitted due to parental and physician concerns.

As expected, preterm children in our cohort were followed by a multidisciplinary team which often included either a physical or an occupational therapist. Half the children required services from two or more healthcare providers. This is probably an underestimate as we did not consider other medical specialists such as ear–nose–throat surgeons, pulmonologists and cardiologists, who are often consulted for prematurity-related problems. Moreover, the majority of preterm infants are seen on a routine basis by a neonatal follow-up team (paediatrician, developmental psychologist and nurse), an ophthalmologist and an audiologist. These services were not included in the analysis. In a similar study conducted through the Neonatal Research Network in the USA, Hintz et al found that 54.7% of infants <28 weeks used at least four services including neurodevelopmental follow-up and medical and surgical subspecialties from neonatal discharge to 18–22 months CA.5

Use of all out-patient services was not different between the two gestational age groups (23–25 weeks vs 26–28 weeks). The odds of requiring services above average were associated with neonatal morbidities such as BPD and severe brain injury, which are risk factors of poor neurodevelopmental outcomes.18 19 Because this study assessed use of healthcare resources, which may not reflect needs of healthcare resources accurately, confounding by indication could be a problem. Use of healthcare resources is influenced by parental perception of the child's needs as well as availability of the services in the community, which we could not assess. Furthermore, determination of the infant's needs is dependent on the physicians' and therapists' opinions, which may be influenced by awareness of neonatal history. Infants with significant neonatal complications are more likely to be referred for additional services because of their worse prognosis.20 Although this practice would cause confounding by indication, a methodological limitation for the interpretation of results, this will likely benefit the patient in the sense that given the risk of disabilities, preventive care is necessary. In their study, Hintz et al addressed the issue of perceived needs for services by parents and found that about 37% reported unmet service needs, speech and language therapy being most frequently mentioned as lacking.5

Multiple births were also associated with higher out-patient service use. The relationship between multiple births and higher health-related costs is thought to be mediated through the complications of preterm delivery, which subsequently lead to more pronounced developmental delays compared to singletons.2 3 21 Finally, living in a single-parent household increased the odds of requiring services above average. The reason underlying this finding is not clear. Nevertheless, it is reassuring to observe that access to healthcare resources in Quebec does not discriminate against children in less favourable environments.

Use of prescribed medication and medical devices was frequently reported by parents of preterm children, especially among the most immature group, who are more likely to develop long-term health problems. Apnoea monitors and home-based oxygen therapy were particularly common. Innovations in home medical technology now allow earlier discharge from the neonatal nursery for children who, although medically stable, still rely on devices to breathe or feed, or require extended cardiorespiratory monitoring. For a selected group of preterm infants, the home environment can be safe and, compared to the NICU, may be more appropriate. In addition to decreasing hospital cost, earlier home transition to home facilitates developmentally centred care (ie, which focuses on child neurodevelopment) and empowers parents to finally be the primary caregivers for their preterm children (as opposed to the nurses and doctors in the NICU).22 If a preterm child is sent home with special medical devices, major efforts are required to coordinate out-patient care, teach parents new technical competencies and ensure easy access to medical support.

This study was limited by the percentage of children lost to follow-up or who refused participation in research (23%). These children had a lower rate of severe brain injury compared to participants. With the knowledge that severe brain injury is associated with poorer motor and cognitive outcomes, we could infer that our findings overestimated the true amount of healthcare resource use by preterm children in Quebec.

Measurement errors related to outcomes were also expected. Information collected on use of healthcare services was dependent on parents' memory and thus susceptible to recall bias. However, most assessors knew the family since discharge from the neonatal nursery and, in addition, had access to the medical chart; therefore, the information provided by the parents could be verified.


Preterm children in Quebec, especially those born at the limit of viability, present with significant neurodevelopmental and medical issues that require use of multiple health-related resources throughout infancy. Re-hospitalisation is frequent in this vulnerable population. Most children are followed by several medical specialists and therapists. A significant proportion of children are sent home with special medical equipment which may necessitate higher technical skills.

With advances in sophisticated and costly techniques to save the most premature and ill newborns, access to healthcare services after discharge from the NICU must continue to be available in order to optimise development of these children. Families of preterm infants require ongoing guidance and support throughout their journey in raising a child with special healthcare needs.


Consortium de recherche sur les enfants nes extremement prematures au Quebec—Dr Rejean Tessier, Suzanne Ferland, Dr Sylvie Belanger, Dr Bruno Piedboeuf (Quebec City), Jocelyne Vallee (Montreal), Dr Elise Couture, Dr May Khairy (McGill); the Canadian Neonatal Network—Dr Shoo Lee and all children and their families who participated in the study.



  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the CHU Sainte-Justine, Montreal Children's Hospital, CHU Quebec, Canada.

  • Provenance and peer review Not commissioned; externally peer reviewed.