Evaluation of the revised Nipissing District Developmental

Evaluation Of The Revised Nipissing District Developmental-Free PDF

  • Date:15 Sep 2020
  • Views:3
  • Downloads:0
  • Pages:8
  • Size:687.81 KB

Share Pdf : Evaluation Of The Revised Nipissing District Developmental

Download and Preview : Evaluation Of The Revised Nipissing District Developmental


Report CopyRight/DMCA Form For : Evaluation Of The Revised Nipissing District Developmental


Transcription:

Cairney et al BMC Pediatrics 2016 16 42 Page 2 of 8. however is well below the estimated prevalence 10 n 500 and 100 in each of the remaining 3 age. Systematic screening provides a possible solution but bands 4 to 6 years of age group B n 300 for a total of. requires measures that are cost effective easily adminis 800 children across all 13 age bands Child age was. tered reliable and valid These requirements are exact adjusted for prematurity if the child was under 2 years. ing given the complexities of measuring development in and born 4 weeks or more prematurely. early childhood 11 While early screening and, surveillance is recommended by many professional Study design. organizations 5 10 and has been implemented in We randomly selected 111 14 participants to. many countries there is no consensus on the instru complete the NDDS a second time after an interval of. ments to be used 2 weeks and 55 7 to complete a qualitative inter. The Nipissing District Developmental Screening tool view Criterion measures were administered by research. NDDS is increasingly used for this purpose in Canada assistants all of whom had an undergraduate or Master s. 12 13 and the United States e g Early Head Start degree e g psychology health sciences RAs received a. Program http www nemcsa org headstart ECDHS A aspx minimum of 8 h of pre test administration training and. The NDDS was first developed in 1993 and its content and at least 10 h of supervised test administration experience. design were revised in 2011 It comprises 13 age prior to being able to conduct independent assessments. group specific parent completed checklists of developmental Assessment reports were monitored continuously for. milestones for children between 1 month and 6 years of age quality assurance throughout the study We received. In Ontario the NDDS is one of the recommended measures ethical approval from the McMaster University Research. to be used during the recently implemented enhanced Ethics Board and all parents provided informed written. 18 month well baby visit 14 15 a population wide consent. comprehensive developmental assessment and parenting. education session connected to the 18 month immunization. visit In Ontario the government has paid to provide free Parent completed measures. access to the NDDS to all parents Nipissing district developmental screen 2011. Despite its increasing use the psychometric properties The NDDS 2011 asks parents to indicate whether they. of the NDDS are largely unknown we could locate only have observed their child performing various motor cog. three reports two of them unpublished and all limited nitive or language tasks There are separate checklists. by small samples 16 18 Only Currie et al 16 evalu for each of 13 age groups The checklist for infants. ated the current version of the NDDS and this was a under 1 month old includes 4 items while others in. pilot study of 31 children only 4 of whom met criteria clude between 12 and 22 items Milestones not yet ob. for mild developmental delay The psychometric proper served by the caregiver are counted to produce a score. ties of the NDDS have not therefore been assessed with Current recommendations are for a health professional. an adequate sample to follow up with any scores of 1 or higher Before the. 2011 revision a cut point of 2 or higher was used 12 17. Methods As the proportion of children identified at the 1 thresh. Sample old may be too large for some situations we also explored. We recruited a sample of participants from community the performance of the NDDS at the 2 cut point. organizations who provide services to families in, Hamilton Ontario and surrounding areas and which Criterion measures. targeted sociodemographically diverse populations As there is no single gold standard for assessing. Organizations included Ontario Early Years Centres development in children we designed a protocol using. and Parent and Family Literacy Centres Staff of some widely used instruments with demonstrated reliability. organizations shared information about the study with and validity Given the broad age range covered by the. their clients and some referred families directly We NDDS it was not possible to use the same criterion. also used recruitment posters and notices on web measure for all children For children 3 years and under. sites and operated a booth at the Hamilton Baby and Group A we used the Bayley Scales of Infant Develop. Toddler Expo which is well attended by families from ment 3rd Edition BSID III 19 The BSID III produces. Hamilton and surrounding areas Families were re a set of raw and normal scores for each of five domains. cruited between May 2010 and October 2011 Parents Cognition receptive communication expressive com. were eligible if they could speak and read English munication fine motor and gross motor We identi. and were the child s primary caregiver and legal fied as mildly delayed those children who scored. guardian We aimed to recruit 50 children for each of below the borderline cut point in one or more do. the NDDS s 10 age bands up to 36 months group A mains and as severely delayed those with at least. Cairney et al BMC Pediatrics 2016 16 42 Page 3 of 8. one score below the extremely low cut point ac with exact binomial 95 confidence intervals We used. cording the manual 19 Stata 13 for all analyses 26. For children aged 4 to 6 Group B we selected three. separate measures assessing development in motor co Results. ordination cognition and language the Movement As We received initial referrals for 1012 parent child pairs. sessment Battery for Children 2nd Edition M ABC and have final data for 812 594 children aged 1 month. 20 the Kaufman Brief Intelligence Test 2nd Edition to 36 months Group A and 218 children aged 4 to. KBIT 2 20 and the Pre school Language Scale 4th 6 years Group B This represents an 80 2 response. edition PLS 4 21 22 respectively The M ABC 20 rate from the total sample of referrals and an 83 8 re. PLS 4 21 and KBIT 2 23 have all shown good agree sponse from eligible families Figure 1 shows the stages. ment with clinical evaluation and with other instru of recruitment participant exclusions and consent rate. ments Children were identified as having mild or Parent demographics are shown in Table 1 In 98 of. severe delay by using the 15th and 5th percentile cut cases the NDDS was completed by the child s biological. points on each instrument The M ABC does not mother and the 812 child parent pairings were drawn. provide a 15th percentile cut point instead the 16th from 572 families The number of children in each. percentile is recommended 20 The K BIT produces a NDDS age band varied from 41 to 98. standard score with a mean of 100 and an SD of 15 We. therefore used cut points of 84 5 and 75 which corres Test retest reliability. pond to the 15th and 5th percentiles Test retest reliability after a two week delay was moder. On the BSID III the published borderline cut points ate Spearman s rho 0 61 p 0 001 as was agreement. produced a prevalence of 27 in children under 1 and at specific cut points at the 1 cut point kappa 0 59. of only 5 in those aged 2 or 3 It is unlikely that this 2 kappa 0 57 86 of 111 78 retests produced the. reflects genuine variation within our sample as we drew same result as the initial screen of the remainder 15. on the same sources to recruit all participants Concerns 14 scores decreased and 10 9 increased The dif. over published BSID III norms have also been raised ference between the proportions increasing and decreas. previously 24 We therefore produced a second set of ing was not significant exact binomial p 0 42. classifications i e cut points to classify mild and severe. delay based on the distributions of raw scores We re Criterion validity. peated this process for the PLS 4 as the norms for this We fit models to identify distribution based cut points. instrument identified only a single case The K BIT for the BSID III and PLS 4 In both cases these resulted. and M ABC produced plausible prevalence s based on in higher prevalence than those derived using the pub. the literature that did not vary markedly with child age lished norms and in prevalences that did not vary sub. To produce distribution based indicators of caseness stantially with child age Results of this analysis are. we used quantile regression with the scale score as the illustrated in Fig 2 which shows borderline cases on. outcome and fractional polynomial transformations of the expressive communication subscale of the BSID III. age as the independent variables These models yield according to the published cut points crosses and ac. equations that can be solved at any child age to calculate cording to our distribution based model all those below. a cut point at the designated quantile For the BSID III the regression line Similar results were obtained for the. we fit two models for the raw score of each subscale other BSID III subscales and for the PLS 4. One corresponding to the borderline 1 33 SDs 9 2nd. percentile and one to the extremely low 2 SD Group A children 1 month to 3 years of age. 2 275th percentile cut point For the PLS 4 to be 103 of 594 children 17 3 scored in the borderline. consistent with other measures used for older children range in one or more BSID III domains At the recom. we estimated cut points at the 5th and 15th percen mended 1 cut point i e one or more no responses. tiles To do this analysis we used the xmfp Stata pro on the NDDS the sensitivity of the NDDS was 59. gram by Royston 25 and the specificity 67 17 children 2 9 scored in. the extremely low range in at least one domain and. Statistical analysis the sensitivity and specificity in this case were 65 and. We measured test retest reliability by calculating Spear 63 respectively see Table 2. man correlations for total scores and kappa statistics for Using distribution based cut points produced generally. agreement using scores of 1 and 2 as cut points poorer agreement 175 children 29 were below the. We compared the NDDS with the criterion measures borderline cut point in at least one domain For this. by calculating sensitivity specificity positive predictive outcome the sensitivity of the NDDS at the 1 cut. value PPV and negative predictive value NPV along point was 50 and the specificity 68 45 children. Cairney et al BMC Pediatrics 2016 16 42 Page 4 of 8. Fig 1 Participant flow diagram, 7 6 were below at least one extremely low cut Twelve children 5 7 met norms based criteria for. point The sensitivity and specificity in this case were severe delay The sensitivity of the NDDS was 67 and. 60 and 64 respectively see Table 2 the specificity 58 Using the adjusted measure pro. duced a prevalence of 8 1 17 of 211 a sensitivity of. Group B children 4 to 6 years of age 65 and a specificity of 59 at the 1 cut point on the. Seven children 3 2 had incomplete or invalid results NDDS see Table 3. on one or more instruments and were excluded from For severe delay all PPVs were under 20 implying a. the analysis Of the remaining 211 children 40 19 low probability that a child with a positive screen will. met norms based criteria for mild delay At the 1 cut meet reference criteria In keeping with the higher preva. point the NDDS had a sensitivity of 68 and a specifi lence PPVs for moderate delay were higher but still under. city of 63 For the adjusted outcome there were 57 50 Using the alternative 2 cut point raised specificities. cases 27 Sensitivity was 60 and specificity 63 to 81 84 but reduced sensitivities to 33 50. Cairney et al BMC Pediatrics 2016 16 42 Page 5 of 8. Table 1 Sample Description,Group A Group B Total,N 594 218 812. Sex of Person Most Knowledgeable,Female 586 99 213 98 799 99.
Male 8 1 4 2 12 1,Home ownership,Rent 137 23 42 20 179 22. Own 449 76 171 80 620 77,Other eg lives with family 3 1 1 1 4 1. Marital status,Never married 32 5 6 3 38 5, Married common law or living with a partner 545 92 202 94 747 93. Separated or divorced 14 2 8 4 22 3,Some secondary or less 25 4 5 2 30 4. Completed high school or GED 27 5 15 7 42 5,Some college or technical training 25 4 13 6 38 5.
Completed college or technical training 130 22 45 21 175 22. Some university 40 7 16 7 56 7, Completed a bachelor s degree BA BSc etc 212 36 81 37 293 36. Completed a graduate or professional degree MSc MD etc 135 23 42 19 177 22. Household income 2009,Under 35 000 75 14 28 14 103 14. 35 000 to 59 999 81 15 30 15 111 15,60 000 to 89 999 115 21 41 20 156 21. 90 000 to 129 999 162 29 60 30 222 29,130 000 or higher 118 21 43 21 161 21. Child s sex,Male 306 52 104 48 410 51,Female 288 48 113 52 401 49.
Number of siblings mean SD 0 9 0 9 1 3 0 8 1 0 0 9. Age of enrolled child in months mean SD 31 2 4 7 30 2 4 5 30 9 4 7. Discussion better accuracy sensitivity 83 specificity 95 but. For screening purposes it is generally recommended the criterion measure used in this study was also a. that sensitivity exceed 80 and specificity 90 27 parent reported instrument 18 Currie et al reported. Given the challenges of screening for developmental sensitivity and specificity at the 1 NDDS threshold to. delay lower thresholds sensitivity of 70 specificity of be 75 and 78 respectively and at the two flag rule. 80 have been suggested in this context 28 29 The 75 and 96 respectively 16 As noted previously. NDDS however did not meet either set of criteria On however the sample size for this study was very small. this basis we cannot recommend that the NDDS be n 31 with only 4 children identified with delay. Nipissing district developmental screen 2011 The NDDS 2011 asks parents to indicate whether they have observed their child performing various motor cog nitive or language tasks There are separate checklists for each of 13 age groups The checklist for infants under 1 month old includes 4 items while others in clude between 12 and 22 items

Related Books