Development of a clinical feeding assessment instrument to identify oropharyngeal dysphagia in high-risk neonates

Abstract

There is a dearth of validated neonatal feeding assessment instruments available for use in clinical practice in resource-constrained developing contexts. The Neonatal Feeding Assessment Scale (NFAS) was developed to identify and diagnose oropharyngeal dysphagia (OPD) in neonates. The main aim of the study was to develop and test the psychometric performance of a clinical assessment scale for the early identification and diagnosis of OPD in the high-risk neonatal population in South Africa. To meet the main aim, the research project was divided into three separate studies. The research design across the three studies was an exploratory sequential mixedmethod design. The NFAS was developed using the Delphi method in the first study. Two international and three South African speech-language therapists (SLTs) formed the expert panel that participated in two rounds of electronic questioning to develop the instrument. For the second and third studies, a comparative crosssectional within-subject design was used. In the second study the participants were 20 neonates with a median age of 35.0 weeks gestational age (GA) in a 29-bed neonatal intensive care unit (NICU). In the third study 48 participants with a median age of 35.5 weeks GA were included. During the second study the preliminary psychometric performance of the NFAS was determined and in the third study, the final psychometric properties of the NFAS were determined to describe the validity and reliability of the NFAS. The NFAS was developed and approved, using expert collaboration through the Delphi method in the first study. All participants agreed on the need for the development of a valid clinical feeding assessment instrument to use with the highrisk neonatal population. The initial NFAS consisted of 240 items across 8 sections; after the Delphi process was implemented, the final format was reduced to 211 items across 6 sections. The final format of the NFAS is scored using a binary scoring system guiding the clinician to identify the presence or absence of OPD. All members agreed on the format, the scoring system and the feeding constructs addressed in the final format of the NFAS. The second study showed that 9 out of 20 participants presented with OPD on the NFAS. Comparison of NFAS results with modified barium swallow studies (MBSS). indicated that all participants with OPD were correctly identified (100% specificity). The sensitivity was 78.6%, indicating that three participants were falsely identified with OPD on the NFAS. The instrument took approximately 30 minutes to complete during observation of a habitual feeding session with the mother. Inter-rater reliability was determined on 50% (n=10) of the study sample. Substantial agreement (80%) was obtained between two raters in five of the six sections of the NFAS and on the diagnostic outcome of the scale. The preliminary performance of the NFAS appeared to be promising. The formal validation process of the NFAS then followed on a larger sample in the third study. Results of the third study indicated that 15 of the 48 participants were identified with OPD on the NFAS, whereas 14 of these 15 infants were diagnosed on MBSS. A sensitivity score of 78.6% was obtained, with specificity determined to be 88.2% for the newly developed NFAS. The subsequent accuracy of the NFAS to identify OPD correctly was 85.4% when compared with the MBSS outcome. Inter-rater reliability was determined using 35% of the sample. The agreement on overall instrument outcome between the two raters was considered substantial beyond chance, with Cohen's Kappa at 0.598, with an asymptotic standard error of 0.211. The scale may be of use to SLTs working without MBSS equipment and to reach underserved preterm neonates. Inexperienced SLTs may benefit from observational prompts provided by the NFAS. The NFAS may be suitable for use in South Africa and similar developing contexts to identify and diagnose high-risk neonates with OPD.