The children in the four sub-samples were very similar in age and gender and the differences found are likely to be as a result of either their HIV status or their residential setting. Very few children had the experience of living with their natural parents and families before institutionalisation or fostering. A large percentage of children in foster care were cared for by older members of the various communities. The limited number of people prepared to foster children, the increase in the number of single parent households due to the impact of the migrant labour system and parents leaving to work or parents having difficulty coping with raising their children [16
] might explain this finding. The phenomenon of elderly persons caring for young children is common in South Africa where the progression and spread of the HIV/AIDS epidemic has led to an increase in the grandparent headed households [17
] and this is predicted to rise.
As anticipated, children with HIV performed at a lower motor level than their peers in both settings, a finding consistent with the literature [6
]. Although many of the children reported serious illnesses at some stage of their lives prior to testing, none of the children experienced any serious illness at the time of testing and none of them were ill as a result of opportunistic infections for the duration of the study. The improved health status of the infected children is likely to have been due to the administration of ARVT to all the children and none of them had been on treatment for less than five months. A study in Uganda, reported that treatment on ARVT for six months or longer resulted in increased longevity and an improved health status [20
]. The fact that the health status of all the children was good implies that the decreased scores for their development obtained on the PDMS II were unlikely to be due to current illness and/or recent hospital admissions.
FMQ was advanced in all institutionalised children. Fine motor activity depends on the ability to perform precise movement co-ordination and includes elements of hand eye co-ordination and that of manual dexterity which required increased concentration [21
]. As fine motor function is related to practice and opportunities to use the upper limb in skilled tasks, it might be that the children in institutions were subjected to more play stimulation than children raised by foster parents, who were often fostering several children. The presence of skilled personnel and a large number of volunteers working at the institutions might have afforded these children greater attention and stimulation.
In a study on the dual impact of institutionalisation and HIV on 64 children in the Ukraine, the increased opportunities for play and use of toys within a well resourced institutional setting compared to foster care were also noted [6
]. In contrast to the current study, children who were in institutions, whatever their HIV status, performed worse than those in family care. However, in the Ukraine study, the children who were institutionalised seemed to have suffered from a greater degree of early deprivation than those staying with families and the families were their own biological families and this represents an important confounding variable. The result of a study comparing 94 children in institutions with 48 in foster care in Iraq Kurdistan was somewhat more equivocal. It concluded that there were more similarities than differences between the two samples, although the fostered children did show more improvement in activity scale, externalizing problem scores and posttraumatic stress disorder-related symptoms [8
]. A major difference between the Iraq study and the current research is that the 84% of the fostered children had a relative as a formal caregiver, whereas in our study no children were cared for by family members. It might be that the relationship of the foster parents to the child may impact on the quality of care and degree of stimulation that the child receives.
A striking finding was that a significant improvement was noted in the performance of all participants from baseline to six months. Although the researcher did give advice to care-givers due to ethical considerations, it is unlikely that a single session of advice on infant stimulation could have resulted in such a large change. The improvement was not a function of maturation as no correlation between age and TMQ was found and scores on fine motor development have been found to remain stable over time, from two to five years of age [22
]. Exposure to the test at baseline might have led to a learning effect which resulted in improved performance on the second test. However this is not likely as the tests were six months apart. The improvement could have been as a result of the initiatives described in the research setting to empower care-givers to provide more developmental stimulation to their children.
It needs to be stressed that the current study did not examine the emotional well being of the children or their cognitive development. The results are not incompatible with the suggestion that long term fostering offers children security, a loving family environment and a close substitute of parental relationships while institutional care is thought to provide unstable caregiver relationships, lack of a sense of belonging and children may be missing out on the family unit [24
]. However, this is not necessarily so in every case. There has been concern expressed over 'voluntourism' in which the impact of having international volunteers working in institutions for relatively short periods of time is being questioned in terms of the rejection that children may feel when the volunteer returns home [25
]. There is a clear need for increased stimulation of these children both within the residential care and the foster care setting and it would be a pity if all resources were not mobilised. If it is the case that children in residential care are disadvantaged, well meaning and suitably trained volunteers could also interact with foster parents and provide additional support to children within the community.
The sample size was smaller than planned but there were still significant differences noted between groups and settings. Attrition is one of the greatest problems associated with longitudinal studies and the attrition in this study was about 20%, which is high, but not as high as the 58% attrition at six months reported in a similar study in Cape Town on younger children with HIV cared for by their mothers [13