A visual skills inventory for children with neurological impairments D L McCulloch* OD PhD FAAO; R T Mackie BSc PhD MCOptom, Vision Sciences, Glasgow Caledonian University; G N Dutton MD FRCOphth, Royal Hospital for Sick Children and Glasgow Caledonian University; M S Bradnam MA PhD CEng MIEE MIPEM, Royal Hospital for Sick Children and University of Glasgow; R E Day MBChB FRCP; G J McDaid DipCOT MBAOT; S Phillips DCH RCPS DPH, Royal Hospital for Sick Children; A Napier, Glasgow City Council, Glasgow, Scotland, UK. A M Herbert BSc MA PhD, Department of Psychology, Rochester Institute of Technology, Rochester, NY, USA. K J Saunders MCOptom PhD, School of Biomedical Sciences, University of Ulster, Coleraine, Northern Ireland; A J Shepherd, PhD BA(Hons) RGN, Department of Nursing and Midwifery, University of Stirling, Stirling, Scotland, UK. *Correspondence to first author at Vision Sciences, Glasgow Caledonian University, Cowcaddens Road, Glasgow, Scotland G4 0BA, UK. E-mail:
[email protected]
Children with neurological impairments often have visual deficits that are difficult to quantify. We have compared visual skills evaluated by carers with results of a comprehensive visual assessment. Participants were 76 children with mild to profound intellectual and/or motor impairment (33 males, 43 females; age range 7mo–16y; mean age 5y 1mo [SD 4y 2mo]) who completed a visual skills inventory before attending a special vision clinic. The inventory included 16 questions about visual skills and responses to familiar situations. Responses were augmented by taking a structured clinical history, compared with visual evoked potential (VEP) and/or acuity card measures of visual acuity, and examined using exploratory factor analysis. Acuity ranged from normal to no light perception, and was positively associated with responses to individual questions. After excluding four uninformative questions, an association between the remaining questions and two significant independent factors was found. Factor 1 was associated with questions about visual recognition (e.g. ‘Does your child see a small silent toy?’) and these items were correlated with both the VEP and acuity card thresholds. Factor 2 was associated primarily with questions about visually mediated social interactions (e.g. ‘Does he/she return your silent smile?’). Evaluation of visual skills in children with neurological impairment can provide valid information about the quality of children’s vision. Questions with the highest validity for predicting vision are identified.
In the Fraser of Allander Unit at the Royal Hospital for Sick Children in Glasgow, we hold a designated vision clinic for children with neurological impairments. This clinic caters for children with the full range of motor, learning, and developmental disabilities. In general, children are referred when the outcome of routine vision testing is equivocal, when visual behaviour is unusual, or when the neurological status places the child at significant risk of visual impairment. These children commonly have limited communication skills and inconsistent visual responses. As behaviour in a clinic may not be truly representative of the child’s abilities, caregivers are in a position to give important information about visual skills and behaviours. Neurological impairment in infancy and childhood is strongly related to visual deficits, with a high prevalence of large refractive errors, strabismus, nystagmus, and visual impairment.1,2 Impairment of visual function arises from a wide range of causes of damage to the brain. These include hydrocephalus, hypoxic–ischaemic encephalopathy, periventricular white matter pathology, neonatal hypoglycaemia, prolonged seizures, and congenital disorders of the brain, which can be structural, functional, static, or progressive.3 Such impairment can be difficult to elicit and can be missed, yet it is crucial to recognize the visual limitations in each child. Unrecognized visual impairment delays and alters motor and cognitive development in children both with and without other disabilities.4 Educational and habilitational strategies can be devised which take these limitations into account, ensuring that educational material can be seen and understood.5 Structured history-taking from carers using a general or a specific visual skills inventory can effectively identify visual dysfunction in children.6–8 We have developed a visual skills inventory that we send to the homes of all appropriate patients before their visit to collect consistent visual history information, which may be augmented or clarified with information gathered during the clinic visit. The questions applied in this inventory have been developed for infants with profound visual impairment, from experience of clinical history-taking, observing their behaviour, and measuring visual function. In the present study, we aimed to compare visual skills as evaluated by parents and carers using a visual skills inventory with the outcome of a comprehensive visual assessment that included acuity card tests, visual evoked potential (VEP), and clinical quantification of visual function. In addition to assessing the quality of individual items in the inventory, factor analysis was applied to determine whether sets of questions probe one or more related abilities. Finally, the present record review aimed to guide us in designing a more efficient inventory (i.e. to remove questions yielding redundant or uninterpretable answers). Method VISUAL SKILLS INVENTORY AND CLINICAL ASSESSMENT
A visual skills inventory consisting of 16 questions about visual skills and responses to familiar situations, and five questions concerning prior vision care (Appendix I), was sent to the homes of children with an appointment to attend the vision clinic . The questions were initially devised by the special needs teacher in the team (AN) and subsequently modified by the other members of the team. They were designed to address common situations encountered by children while minimizing the requirement for speech or fine motor responses, which
Developmental Medicine & Child Neurology 2007, 49: 757–763 757
may be impaired in this population. The skills assessed in the inventory were converted to appropriate ordinal scales; several questions had only two levels, yes or no. Other answers were assigned up to six levels (e.g. question 18 about reacting to a large object has six levels from no reaction to greater than 4ft [120cm]). When separate questions were asked for small and large objects (questions 14–19) the responses were combined to give three ranks, no reaction, large only, or large and small. Where the question was not relevant to the cognitive or motor skills of the child, the response was classified as not applicable and was not used in the analysis. Each child received a clinical examination that included optometric and orthoptic assessments and VEP testing, as well as a neurological assessment by a consultant neurologist (RED) and clinical eye examination by a consultant ophthalmologist (GND). Evaluation by the occupational therapist and the special needs teacher were integrated into the vision clinic when indicated. Although children spent several hours at the vision clinic most assessments were brief with breaks between them. Neurological and ophthalmological diagnoses were classified by type and severity (see Participants). During the examination a structured clinical history was taken. This included completing and clarifying information listed on the visual skills inventory. VA was assessed using Keeler acuity cards (Keeler, UK) and the Cardiff Acuity Test.9 Acuity card measures were converted to a 7-point ordinal scale using clinical and Snellen optotype acuity equivalents (calculated for the 6m standard) as follows: (1) blind (no light perception or brainstem visual function, ‘blindsight’ only10); (2) light perception or gross form perception; (3) severe impairment (6/100–6/400); (4) moderate impairment (6/50–6/100); (5) mild impairment (6/26–>6/50); (6) very mild impairment (6/13–6/26); and (7) no impairment (6/12 or better). When both tests gave a reliable threshold, the test giving the highest resolution was used as a behavioural measure of VA as these tests show good agreement in children who cooperate for both.11 VISUAL EVOKED POTENTIALS
Binocular, pattern-onset VEPs were recorded using a personal computer-based stimulus and recording system.12 Stimuli consisted of high contrast black and white vertical gratings and checkerboards, generated on a colour monitor. The mean luminance of the monitor screen was 90cd/m2, and the contrast between the black and white stripes and checks was greater than 90%. Each stimulus was presented from an isoluminant grey background for a period of 250ms. The children sat at either 43cm or 85cm from the monitor depending on their ages and abilities. Children were generally seated on the lap of one of their parents, in a wheelchair, or a special observation chair. These seating arrangements enabled the children to be held at the correct distance for testing and helped to stabilize both voluntary and involuntary head and body movements. Testing was performed in a dimly lit, quiet room to minimize distraction. Testing required two people; one person controlled the stimulus from a computer terminal and the other attracted the child’s attention and monitored fixation. The child’s attention was encouraged by talking, rattling keys, and/or singing. VEP recording was paused when the child was not fixating. Signals were recorded from three active occipital electrodes, Oz, O1, and O2, referenced and grounded to the ear
758
Developmental Medicine & Child Neurology 2007, 49: 757–763
lobes, A1 and A2.13 Occasionally limited cooperation led to the use of a single electrode at Oz. Signals were amplified and filtered through a band-pass filter of 0.8 to 30Hz. An artefact rejection procedure eliminated trials with amplitudes exceeding 50 or 100µV from baseline that may result from muscle artefact or excessive background electroencephalogram (EEG) activity, both of which were common in this group. The signals were averaged 50 times to produce a VEP waveform. The range of bar widths employed was 4˚ to 6 minutes of arc (’) when presented at 43cm, and 2˚ to 3’ when presented at 85cm. The bar widths available decreased in half octave steps (1˚, 30’, and 15’) except for the fine gratings that were available for 12’, 6’, and 3’ bars compatible with the monitor resolution. Grating stimuli were presented repetitively at 1.06Hz. These stimuli moved horizontally at either 3 cycles/s for small gratings and 14 cycles/s for large gratings.14 A reproducible response was declared if the waveform had a measurable positive peak (C1),15 which was reproducible to within ±20ms.16 The first stimulus presented was the onset of a stationary checkerboard (2˚ check width). If this generated a clearly reproducible VEP, then threshold estimation began with a grating onset stimulus of 1˚ bars; whereas if the VEP to 2˚ checks was poor or absent, then gratings larger than 2˚ were initially presented. Stimuli were decreased or increased in size depending on whether a VEP to the previous stimulus was reproducible or not. The stimulus with the minimum bar width that elicited a reproducible VEP was recorded as the threshold. If no reproducible response to any pattern-onset stimulus was obtained then a stroboscopic flash stimulus was presented (Grass PS33+, Intensity 8 Grass Instruments, West Warwick, RI). VEP thresholds were ranked on an ordinal scale similar to that used for acuity card thresholds as follows: (1) blind (no recordable VEP); (2) light perception or gross form perception (VEP recorded to flash only); (3) severe impairment (VEP threshold of 4˚ or 2˚ bars); (4) moderate impairment (VEP to 1˚ bars); (5) mild impairment (VEP to 30’ bars); (6) very mild impairment (VEP to 15’ or 12’ bars); and (7) no impairment (VEP to 6’ or 3’ bars). STATISTICAL ANALYSIS
Individual questions from the inventory were compared with visual outcomes using ranked comparisons (Kendall’s coefficient of concordance, tau). All results from the inventory and the vision testing were then evaluated using factor analysis (principal components analysis with an orthogonal rotation, SPSS, version 13). An exploratory factor analysis was undertaken using the complete questionnaires. After removing uninformative questions, the exploratory analysis was repeated using the remaining questions. The identification of factors was made using a scree plot (this is used to examine the contribution of each factor to the solution, and only includes the factors that appear to have the greatest contributions). PARTICIPANTS
Patients with neurological impairment had been referred to the vision clinic because they had no previous estimate of visual function or because they exhibited visual behaviour that could not be explained by the current understanding of their visual status. Of the 126 children for whom inventories were sent over a 3-year period, 77 (61%) returned the inventories and attended for full clinical evaluations. The youngest child,
a 3-month-old with marked optic atrophy, was excluded from further analysis because many of the questions were not relevant to such a young infant. The remaining 33 males and 43 females ranged in age from 7 months to 16 years as follows: 7 months to 2 years (n=18, 24%), 2 to 5 years (n=30, 39%), and 5 to 15 years (n=28, 37%). In typically developing children, a VEP is recordable by 3 months of age to the smallest stimulus category used16 so that children without visual impairment would fall within the ‘no impairment’ category across the full age range of this study. Acuity card thresholds are expected to improve throughout the first 4 years of life and are affected by developmental delay, as are visual skills. In the present study, actual visual thresholds and visual skills were categorized without correction for either chronological or developmental age. Fifteen children (20%) had motor function within normal limits by clinical examination. The others had neuromotor diagnoses including spastic hemiplegia (n=9, 12%), diplegia (n=12, 16%), quadriplegia (n=17, 22%), dyskinesia (n=6, 8%), ataxia (n=2, 3%), and other motor disabilities (n=15, 20%). The paediatric neurologist (RED) also classified the intellectual and motor function of the patients by severity into the categories of normal, mild, moderate, and severe impairment. In infant patients, functional categories occasionally could not be assigned for intellectual function (four unclassified) and motor function (one unclassified). Clinical data for all participants are summarized in Table I. In 65 of the children, a primary visual diagnosis could be specified. According to the site of the primary lesion, these included: ocular (retina and/or lens, n=11), optic nerve (n=14), cerebral visual dysfunction (including posterior visual pathways and visual field loss, n=32), and cognitive visual dysfunction (functional disorders sparing central acuity, n=8). In addition to the primary diagnosis, 57 children (75%) had strabismus and 44 (58%) had nystagmus. Visual field defects were found in 28 of the 57 children with sufficient visual attention for visual field screening; of these 12 had absolute hemianopic defects, 12 had relative hemianopic defects (inattention scotoma), and four had other defects (Table I). Results VEP testing was successful with a threshold measured in 61 cases (80%); acuity cards were successful in 65 patients (86%). The 53 children who completed both tests showed good agreement in the ranked categories for level of vision (tau=0.47, p0.1). However, levels of visual impairment were unevenly distributed across the age groups (χ2=13.2, p
