Predictive Validity of the Test of Infant Motor Performance as Measured by the Bruininks-Oseretsky Test of Motor Proficiency at School Age,

Predictive Validity of the Test of Infant Motor Performance as Measured by the Bruininks-Oseretsky Test of Motor Proficiency at School Age, by J Flegel, THA Kolobe, Physical Therapy 2002;82:762-771. Purpose The study was designed to investigate the predictive validity of the Test of Infant Motor Performance (TIMP) for identifying school age children who are typically developing or demonstrating developmental delay when examined using the Bruininks-Oseretsky Test of Motor Proficiency (BOTMP). The authors indicate that the TIMP has been shown to be a reliable instrument for use with infants between the ages of 32 weeks postconception and 16 weeks postterm. The test has demonstrated concurrent validity with the Alberta Infant Motor Scale for three-month-old infants. The test can be used to discriminate between optimal and poor motor performance in infants who are born preterm and in very young infants, and is sensitive to developmental change. The TIMP is divided into elicited and observed behaviors and is comprised of 59 items that were taken from other neurologic and developmental tests or were developed specifically for this test. The items are considered to have ecological validity as they are similar to events that occur in infants daily care giving. The TIMP has been shown to be related to scores on the Problem Oriented Perinatal Risk Assessment System (POPRAS) and for that reason medical complications should be considered when examining predictive validity. The POPRAS is scored using data from a medical record with higher scores indicating more medical complications. The authors of the current study indicate that a secondary purpose of their work was to examine the relationship between perinatal risk using the POPRAS and performance on the BOMPT. Method Subjects. Thirty-five children, originally tested on the TIMP at the time of the test’s development comprised the sample. This group represented a subset of the 137 children who were originally tested on the TIMP. That original sample comprised children in seven different age groups with three categories of medical risk as determined by the POPRAS. After conducting a power analysis, the authors determined that 34 subjects would be needed to test associations at the 0.05 level of significance. Of the original sample 65 children were located and the subjects of the current study were randomly selected to include two children from each age group and risk category. However, seven age or risk categories had only one child available for follow-up. The stratified sample included four to six children in each of the seven age groups and a relatively even distribution of children who were at low risk (n=12) for developmental delay, at moderate risk (n = 10) and at high risk (n = 13). The children in the sample were between the ages of four years, nine months and seven years, three months at the time of follow up, with a mean age of five years, eight months. When originally tested on the TIMP they had an average gestational age of 32 weeks and were tested at an average postnatal age of 59 days. There were 19 boys and 16 girls in the sample. Forty-six percent were white, 26% were African American, and 29% were Latino. Three children had been diagnosed with cerebral palsy and one with Down syndrome. Data collection procedures. The complete BOTMP was used as the criterion test of motor outcome. The BOTMP composite scores for normal values for age are based on a T-score with a mean of 50 and a standard deviation of 10. Interrater reliability of the BOTMP was established prior to data collection on six children who were not subjects of the study. The ICC (2,1) for the composite BOTMP score was 0.97. A previous study reported acceptable interrater reliability and test retest reliability for the TIMP. The BOTMP was administered by one examiner who was blind to the original TIMP scores. All but three children were tested in their homes. Two children were considered uncooperative and as recommended in the BOTMP manual the testing was interrupted and completed on another day. Age at the time of testing was based on chronological age. At the end of testing the parent or guardian filled out a demographic questionnaire which including questions on the child’s medical and developmental history. Scores from the original administration of the TIMP and the POPRAS were obtained after the BOTMP was administered and scored. Data analysis. A TIMP cutoff score was identified by calculating sensitivity, specificity and positive and negative predictive values. This score was determined to be potentially of use in a clinical setting to identify infants with motor delays who were likely to continue to have long-term motor delay or poor performance on the BOTMP. Poor performance eon the BOTMP was defined as a z score of -1.5. This cutoff score represented the middle range of scores considered to represent low-performance on the BOTMP. Because the TIMP has not been normed, z scores were calculated for each of the seven age groups. The authors note that such scores should be considered preliminary because they are not based on random sampling. Because of the small sample size confidence intervals were calculated for sensitivity, specificity and positive and negative predictive values. This procedure was also employed for the POPRAS scores. The relationship between TIMP scores (corrected for age) and the BOTMP battery composite scores was analyzed using a Pearson product moment correlation. A Pearson product moment correlation was also calculated for POPRAS scores and the BOTMP composite scores. A hierarchical multiple regression analysis was then used to determine the amount of variance in the BOTMP score accounted for by the age-adjusted TIMP score beyond that accounted for by the POPRAS. Results The average BOTMP composite score for the sample was 43.9 with a standard deviation of 11.6. The mean TIMP for the sample was 82.7 with a standard deviation of 30.7. At school age, eight of the children in the sample were determined to have motor delays using the BOTMP. At birth the gestational age of six of these children was less than 29 weeks. The authors presented the sensitivity, specificity and positive and negative predictive values for predicting performance on the BOTMP at various cutoff scores for both the TIMP and the POPRAS. When the cutoff z score of -1.6 was used 89% of the children were correctly classified. Higher cutoff scores resulted in increased sensitivity but lower specificity. When the cutoff score of 80 was used on the POPRAS 77% of the children were correctly classified. The authors report that the partial correlation coefficient of TIMP Scores and BOTMP with age controlled was statistically significant although low (r = 0.36) The correlation coefficient between POPRAS and BOTMP composite scores was statistically significant and higher (r = −0.55). Thirty seven percent of the variance in BOTMP scores was explained by age, POPRAS and TIMP. The POPRAS was found to be a significant predictor. The TIMP did not reach significance when entered into the step-wise regression after the POPRAS. Discussion The authors address the unexpected high false negative classifications of the TIMP when the cutoff of -1.6 was used. These represent children who would not have received early intervention, but are later found to have developmental delay at school age. The authors examined each negative case to determine a specific age of testing with the TIMP would yield false negative classifications. They also examined the parents’ questionnaire to see if any events might have occurred following administration of the TIMP that would explain these false negatives. Each of the four children with false negative results were born pre-term and based on POPRAS scores were deemed to be a moderate risk for developmental delay. No evidence was found that a specific age of testing was responsible for the false negatives. The authors also report that there is evidence in the literature to suggest that neurological testing at term could lead to normal findings in infants who were previously considered abnormal and were later again considered abnormal. This was suggested by the authors to represent an area that would require further study. On the otherhand, the POPRAS cut off score of 80 may result in a more effective prediction of motor delays at school age. However, the POPRAS resulted in more false positives. The authors suggest that the association between POPRAS and BOTMP scores suggests that the child’s early medical complications influence motor performance at later ages. The authors discuss the clinical implications of their findings including the suggestion that the TIMP may not be the best test if it is important to identify the majority of children who are likely to have a poor motor outcome. However, the TIMP would be good to use if one wanted to predict a positive outcome. The limitations of the study are presented including the small number of subjects, the z-scores for the TIMP are not determined from a random sample of the population of infants at risk for developmental delay, and the lack of documentation of intervening variables that might have affected motor outcome following administration of the TIMP. The authors conclude that the TIMP is useful for identification of very young infants at risk for long-term poor motor performance. They caution the reader regarding the low sensitivity of the measure. Because the POPRAS had higher sensitivity they recommend use of both the TIMP and the POPRAS when making referrals for early intervention. Strength and Weaknesses of the Study The authors present a very clear purpose statement, although it is somewhat surprising that a secondary purpose statement appears in the background section of the article. This secondary purpose which ultimately becomes a very integral part of the study seems to deserve more than second rating in the presentation of the study. For example, the title might reflect the examination of the POPRAS as a key element of the study. The design and plan of the data collection from the original sample of children who were examined using the TIMP are clearly presented. The properties of each of the instruments are presented in sufficient detail to be able to understand each of the tests. Data collection procedures are well described as are the processes of data analysis. The authors present the results in a clearly understandable manner and the discussion is directed at those findings. The only element that detracts from the article is the conclusion that the TIMP would be useful for early identification of infants at risk for long-term motor performance problems, particularly when false negatives might result in children who need early intervention being missed. To the authors’ credit they do finally suggest using a combination of tests that would include the more sensitive POPRAS.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.