Increased health-care utilisation in international adoptees

Most international adoptees appear to be well adjusted in their adoptive families in the recipient country [1]. However, many children arrive to the recipient country with developmental and growth delays, and with medical and mental health issues related to adversities and maltreatment in their countries of origin [2]. Infectious and parasite-induced diseases are the most common issues [3], but also cases of anaemia, hepatitis and tuberculosis are frequently seen [4]. Substantial post-adoption recovery from these initial disadvantages is documented, and they are most pronounced in the first years for physical parameters such as weight and height [5]. However, the effects of early institutional deprivation may persist in early adolescence; and for some children, psychological development follows a different pattern than that of non-adoptees [6].

Adoptees are reported to be overrepresented in psychiatric settings and use medical services at a higher rate than non-adoptees [7-9]. However, these studies investigate only very specific outcomes and do not give a comprehensive overview of health-care utilisation. Moreover, they have limited sample sizes and are not representative for international adoptees, or are based on self-reported data [8, 9]. The use of nationwide administrative databases, such as those established in the Nordic countries [10], allows us to analyse much larger, representative samples.

Increased morbidity and excessive use of health-care services is expected immediately after adoption, but it is uncertain if increased morbidity persists after a period of post-adoption recovery. The present study aims to assess health-care utilisation of international adoptees in a late post-adoption period.

METHODS

Study population

Data on international adoptees were obtained from the Danish Civil Registration System (CRS). In Denmark, adoption status is stated in the CRS along with country of origin [11]. The population in the present study was restricted to all Danish adoptions of children below 11 years of age processed from 1 January 1994 to 31 December 2005. For each included adoptee, the study population was augmented with all non-adopted children who, at the time of the adoption, were of the same sex and age, lived in the same municipality and in similar family constellation (e.g. nuclear family, single parent); this information was obtained from The Danish Family Relations Database which is based on kinship information from the CRS [11]. Next, national adoptees – primarily stepchild adoptions (Figure 1">Figure 1) – were excluded
so that the adoptees in the final study population are international adoptees who are biologically unrelated to their adoptants.

Outcomes

Outcomes in the present study are measures of health-care utilisation over a period of time after the adoption, or, in case of a non-adopted child, the time of adoption of the adoptee with whom the child was matched. An early two-year post-adoption period covers the first two full calendar years following the adoption year and represents the post-adoption recovery period. A late three-year post-adoption period spans the three calendar years following the early post-adoption period. Health-care utilisation is measured for selected primary care contacts or hospital admissions (see below), as the total number of these contacts in each of the post-adoption periods. Data on contacts to primary care up to 2005 were retrieved from the Danish National Health Services Register [12]. These data were registered under the child´s own identification number only as from 1997. Therefore, analyses on primary care outcomes for the early post-adoption period included only those adopted in the 1996-2003 period, and for the late post-adoption period only those adopted in the 1994-2000 period; these periods do not end by 2005 to ensure that all children have two and three years of follow-up, respectively. Data on diagnoses for in- and outpatient contacts to secondary care were obtained from the Danish national patient register (NPR) [13] and used the same inclusion years as the primary care analyses. If, for a certain person, register data were only available for part of a
period, the available outcome information (e.g. days in hospital, number) was divided by the fraction of the
period for which information was available.

Health-care utilisation in primary care (five outcomes) is defined as the number of:

– Daytime consultations to the general practitioner (GP)

– Medical specialist contacts

– Eye specialist contacts

– Ear specialist contacts

– Contacts to other medical specialists (e.g. surgery, child psychiatry, anaesthesiology, X-ray).

Health-care utilisation in hospitals (11 outcomes) follows the International Classification of Diseases (ICD)-10 classification and is defined as the number of contacts with specific diagnoses (hospital admissions or contacts to outpatient clinics):

– Hospital contacts (all entries in the NPR)

– Infectious and parasitic diseases (ICD-10: A00-B99)

– Diseases of the blood and blood-forming organs (ICD-10: D50-D89)

– Endocrine, nutritional and metabolic diseases (ICD-10: E00-E90)

– Mental and behavioural disorders (ICD-10: F00-F99)

– Diseases of the nervous system (ICD-10: G00-G99)

– Diseases of the respiratory system (ICD-10: J00-J99)

– Diseases of the digestive system (ICD-10: K00-K93)

– Congenital malformations, deformations and chromosomal abnormalities (ICD-10: Q00-Q99)

– Injury, poisoning and certain other consequences of external causes (ICD-10: S00-T98)

– Burns and corrosions (ICD-10: S00-T98).

Covariates

Sex, age, region, family constellation at the time of adoption and year of adoption were obtained from the CRS [11]. Family constellation indicates whether the child lives with both parents or with a single parent. Socioeconomic position of the household was measured both by the highest completed education obtained by
an adult in the household as obtained from the Danish Education Register, and by family income obtained from the Income Statistics Register.

Statistics

For both periods and for each of the 16 outcomes, a multivariable two-part model was used to analyse the influence of being adopted on health-care utilisation [14]. The first part of the model analyses the prevalence of the outcome in the corresponding period using a Poisson regression approach so that the regression parameters can be expressed as the relative risk (RR) of experiencing the outcome at all in the period [15]. This first part tentatively investigates health care seeking behaviour, such as contacting health-care professionals, is
often initiated by the parent(s). The second part of the model analyses the quantity of the outcome for the children who experienced the outcome in the corresponding period in a generalised linear model using a Gamma distribution and a logarithmic link function so that the regression parameters can be expressed as a multiplicative factor of how much more the outcome was experienced for the adopted children relative to the non-adopted children. The second part tentatively measures the behaviour of the health-care services because the quantity of use is often dependent on treatment protocols and medical decisions. Both parts of the model are adjusted for age at the time of adoption, sex, year of adoption, residence, income, education, family status and region. The analyses pertaining to the late post-adoption period also included the prevalence of hospital contact in the early post-adoption period to adjust for level of morbidity. A combined multiplicative effect of being adopted can be calculated by multiplying the RR from the first part and the factor from the second part.

Statistical significance was assessed at a 1% level. Statistical Analysis Software 9.2 (SAS Institute, Cary, North Carolina) was used to analyse the data.

Ethics

The study was approved by the Danish Data Protection Agency.

Trial registration: not relevant.

RESULTS

Demographic variables

The study population (n = 499,194) consisted of a group of non-adoptees (n = 492,374) and a group of international adoptees (n = 6,820) (Figure 1). Demographic data are shown in Table 1">Table 1. The international adoptees featured more girls (55.5%), while the non-adopted children had an approximately equal gender distribution (49.8% girls). Most of the adopted children were one year (42.4%) or 2-4 year old (40.5%) at the time of adoption. and most of the children arrived from the Far East (e.g. China, Vietnam, Korea) (40.9%), South America (24.0%), or Near East and the Indian subcontinent (e.g. Iraq, India) (18.7%).

Health-care utilisation

Table 2">Table 2 shows increased health-care use in both primary and secondary care among adopted children in the first two years after their arrival in Denmark.

The health-care use in years three, four and five after adoption (the late post-adoption period) is shown in Table 3">Table 3. After adjustment for increased morbidity in the early post-adoption period, the adopted children had an increased prevalence of contact in all surveyed primary care areas. Furthermore, when they had contact, the frequency was statistically significantly higher for contacts to the GP and medical specialist. Being adopted significantly increases the risk of hospital contacts in general and of diagnoses related to mental and behavioural disorders, malformations, deformations and abnormalities, burns and corrosions and other consequences of external causes.

DISCUSSION

Main findings

The results show that being adopted from abroad significantly increases the use of all services in primary care in the late post-adoption period. In secondary care, contacts related to malformations and consequences of external causes including burns and corrosions are more frequent in the late post-adoption period. Mental and behavioural disorders are also more frequent among adoptees. The analysis of health-care utilisation in the late post-adoption period is adjusted for hospitalisation in the early post-adoption period; a proxy for morbidity. In this way, the effect of the adoption itself is separated from the effect of a generally higher morbidity often seen in adoptees. The observed increase in the use of health care in the late post-adoption period may therefore be viewed as an effect of adoption per se and not just as a function of higher morbidity.

Relation to previous studies

Our results are consistent with a previous Danish study that describes higher rates of GP contacts and hospital admissions for adoptees that persist long after adoption; specialist contacts were even seen to be increasing compared with non-adoptees in the time after adoption [8]. An increased use of services in the early post-adoption period is expected, as the Danish Health and Medicines Authority advises Danish adoptive families to contact their GP after the adoption is completed to carry out standard physical examinations based on the child´s state of health [16]. Considering the children’s pre-adoption history, it is possible that the GP has a lower threshold for referring adoptees than non-adopted children to a medical specialist, which could explain the increased use of medical specialist services in the late post-adoption period. An increased risk for infectious and parasitic diseases, diseases of the respiratory and digestive system in the early post-adoption period is in accordance with previous results on clinical assessments on adopted children after their arrival, which show pathological findings in the abovementioned disease classifications [2]. These findings confirm insufficient conditions in the countries of origin, and these medical conditions show a tendency to disappear with proper treatment. Congenital malformations, deformations and abnormalities appear in the late post-adoption period, which indicates that a higher prevalence persists beyond post-adoption recovery [8, 17]. Previous studies have also established an increased risk for mental health problems among adopted children [9, 18], and another study has shown patterns with post-adoption syndrome that includes
attention deficit and hyperactivity (ADHD) [19]. These mental health problems could explain the increased risk of diagnoses related to consequences of external causes and burns and corrosions.

A higher psychiatric referral rate among adoptees is not necessarily related to the severity of their behavioural problems, but may be caused by a lower threshold for referral among GPs. Furthermore, the higher contact rate could be caused by the adoptive parents’ generally higher socioeconomic status and a presumed greater familiarity with the health-care system [20].

Limitations

In this register study we only had data on diagnoses for secondary care and this information is only indicative of the overall distribution of diagnoses since most mild diseases are diagnosed and treated in primary care only. No data were available on the circumstances of the adoptees prior to the adoption or the state of health of the adoptee in their country of origin. The effect of poor pre-adoption care could therefore not be analysed. Finally, it is uncertain if the increased morbidity in international adoptees, seen in the first part of the model, is evidence of the parent’s health care-seeking behaviour or the system’s treatment of the patient as seen in the second part of he model.

Implications/perspectives

The results indicate that the international adoptees arrive in their adoptive families with a varying degree of somatic and mental health problems. But most strikingly, our findings show that despite notable post-adoption recovery, some adoptees may experience long-term deficits in many domains and therefore need professional and medical assistance. These findings indicate the need for recognising the adoptees as a vulnerable group that needs special medical attention in the longer term – consisting of, for instance, medical evaluations and prolonged health assessment follow-ups. Future research should examine the clinical and practical implications of this for Danish health care. In particular, morbidity in adulthood among international adoptees should be studied.

CONCLUSION

International adoptees use medical services at a higher rate than non-adoptees. The increase is most pronounced in primary care but also present for some areas of secondary care, especially mental care.

Correspondence: Heidi J. Graff, Department of Public Health, Section of General Practice, Unit of Women and Gender Research in Medicine, University of Copenhagen, P.O. Box 2099, 1014 Copenhagen, Denmark.
E-mail: heidi.graff@sund.ku.dk/heidi.graff@supermail.dk

Accepted: 8 May 2015

Conflicts of interest:Disclosure forms provided by the authors are available with the full text of this article at www.danmedj.dk