3. ANÁLISIS ECONÓMICO FINANCIERO DEL SEGURO OBLIGATORIO DE
3.2. ANÁLISIS ECONÓMICO FINANCIERO
3.2.2. MÉTODOS DE ANÁLISIS FINANCIERO
3.2.2.1. ANÁLISIS HORIZONTAL
The capacity of the injured brain to recover function is understandably modified by a range of factors such as age at onset of injury and size and location of the lesion (what Teuber, 1975, refers to as the "when" and "where" facets of recovery), and the presence of seizures. An explanation for the frequently contradictory results in the anti- equipotentiality research was offered by Aram and Ekelman (1986). They proposed that there has been a tendency for researchers to study heterogeneous groups of brain damaged children, so including cases with bilateral pathology and seizures alongside those with pure unilateral damage and no seizure disorder. The studies by Woods and his colleagues. Riva and Cazzaniga (1986) and Vargha-Khadem, O ’Gorman and Watters (1985) may be criticised on these grounds. In the more recent studies, including those of Aram, greater care has been taken to exclude cases where there is bilateral and/or seizure involvement; this has been made all the more possible by advances in, and the wider availability of, neuroradiological scanning techniques, including CT and Magnetic Resonance Imaging (MRI) scans. However, Aram’s work may itself be criticised for failing to take due account of age at onset of injury, with the result that her group is a mix of pre-/peri-natal and acquired cases. The same may be said of Dennis’s hemidecorticated sample. In general, studies carried out during the last five to ten years have recognised the powerful influence of these factors, and have either controlled for them or in several instances attempted to assess and quantify their contribution.
The effects of lesion size and of seizures on IQ have been studied by Levine, Huttenlocher, Banich and Duda (1987). They assessed 41 hemiplegic children, most of whom had unilateral lesions of pre- or peri-natal origin. Half the sample suffered from seizures. EEG records were taken along with CT scans. Larger lesion size tended to be associated with lower IQs. Indeed, lesion size accounted for 21 percent of the between subjects variance in IQ. However, there was no effect of EEG abnormality or of seizures on IQ when lesion size was entered as a covariate. Levine et al. concluded that "seizures are rarely a factor in the depressed IQ of hemiplegic children" (p. 33). A later study by Banich, Levine, Kim and Huttenlocher (1990) supported the finding of an association between lesion size and IQ (at least for hemiplegia acquired after the age of one month).
However, studies conducted by Sussova, Seidl and Faber (1990) and by Vargha-Khadem, Isaacs, Van der Werff, Robb and Wilson (1992) have produced contradictory findings to those of Levine et al. (1987) and Banich et al. (1990). Sussova et al. collected IQ, handedness, CT scan and EEG data from 51 hemiplegic children, of whom 19 suffered from clinical paroxysms. They found that the children with paroxysms had significantly lower IQs than those who did not. If the children did not have paroxysms, even if their EEGs showed some focal or epileptic changes, their IQs were comparatively unaffected. Similarly, the study by Vargha-Khadem et al. (1992) demonstrated that the incidence and degree of cognitive deficit was highly related to the presence of seizures and/or severe EEG abnormality. Unilateral brain damage, even if it was extensive (as verified on CT or MRI scans), resulted in few and mild deficits, provided the damage was not accompanied by seizure activity or severe EEG abnormality. Whereas lesions uncomplicated by seizures affected only non-verbal functions, lesions accompanied by seizures adversely affected performance on nearly all the measures, verbal and non verbal.
A complicating factor in any study that includes patients with seizures must be the influence of medication on cognitive functioning. Rodin, Schmaltz and Twitty (1986) evaluated the effects of medication on WISC performance in 64 epileptic patients. The subjects were initially evaluated at between ages five to 16 years, and re-evaluated after a period of at least five years. All but one patient was taking at least one anti-convulsant
drug. The WISC IQ scores showed a significant decrease over time. Phenobarbitol levels were found to be inversely correlated with IQ. These authors concluded that the upper limit of the therapeutic range of phenobarbitol may be toxic with respect to learning abilities. Ideally, studies of cognitive performance that include subjects with seizure conditions should monitor and, if necessary, control for drug levels.
There is very little reliable information concerning the relationship between lesion location and cognitive functioning. In CT studies conducted by Cohen and Duffner (1981) and Kotlarek, Rodewig, Brull and Zeumer (1981), cortical lesions and lesions extending from the surface of the cortex to the lateral ventricle were found to be associated with lower intellectual level than were lesions confined to subcortical white matter and basal ganglia. Aram and Ekelman (1986) attempted to relate characteristics of their hemiplegic samples's WISC-R profiles to the site of lesion. The children were grouped by CT scan findings as presenting prerolandic or retrorolandic involvement, and also as having either cortical or subcortical involvement. Unfortunately, the numbers of children in the resultant groups were so small that statistical analysis was precluded. There was, however, a trend for left-lesioned children with retrorolandic involvement to have relatively lower IQs, and for children with subcortical lesions only to have lower Verbal than Performance IQs. In Levine et al’s. (1987) study, no relationship was found between lesion location and IQ. When the lesions were classified according to Cohen and Duffner's criteria, any differential effects on IQ disappeared after lesion size was entered as a covariate.
Filipek, Kennedy and Caviness (1992) discuss the methodological difficulties encountered by studies that attempt to link localisation of function with anatomical structures. First, studies of very large numbers of children are required before the effects of lesion size and specific location on cognitive functioning can be disentangled; given the low incidence of focal lesions in children, this is a difficult criterion to meet. Second, most studies to date have employed CT scanning procedures that are known to give relatively poor grey-white matter resolution; they are consequently restricted to gross anatomical localisation, comparing extent of lesion, left versus right, anterior versus posterior. The outcome of CT studies are affected by procedural variations that are not always
standardised within, let alone across, studies e.g. thickness, orientation and position of slice. Finally, it may not be appropriate to employ adult norms for paediatric imaging data. Filipek et al. propose that MRI scans offer more scope for precise levels of anatomical localisation and should, therefore, be the imaging technique of choice in brain-behaviour studies.
The effects of cerebral lesion on language function appears to be related to the age at which the lesion is acquired. In the Levine et al. (1987) study, children with acquired lesions (after one month of age) performed significantly worse than those with congenital lesions on several WISC-R verbal subtests, as well as on the Peabody Picture Vocabulary Test (Dunn and Dunn, 1981). Aram had by 1988 become aware of her failure to take sufficient account of age at onset of injury in the earlier studies of her mixed pre-/peri- natal and acquired sample. Aram (1988) re-appraised the data from her extensive hemiplegic series and reported, first, that left-lesioned injury, with onset prior to one year of age, was associated with homogeneous Verbal, Performance and Full Scale IQs. Second, when the onset of the hemiplegia was after one year, the Verbal IQ was significantly depressed relative to the Performance IQ. Finally, the complementary pattern i.e. Performance IQ lower than Verbal IQ was demonstrated as a non-significant trend in right-lesioned cases with onset after one year. However, age at onset bore no relation to outcome on specific language measures, including tests of educational attainment (Aram and Ekelman, 1988). Earlier studies, including those of Annett (1973), Woods and Carey (1979) and Vargha-Khadem, O ’Gorman and Watters (1985) have reported a higher incidence of language difficulties among children with late lesion onset.
4.8 Evidence from Studies of Childhood Hemiplegia and of Hemidecortication Re