Cross-cultural adaptation and validation of Neonatal Eating Assessment Tool (NeoEAT)-Paladai/cup feeding in Tamil: a caregiver-reported assessment tool

Abstract

Background

Cup feeding is an alternative method to provide nutrition to infants exhibiting difficulties in breastfeeding. Paladai is a cup commonly used to feed milk to infants in several parts of India. Parental reports of infant feeding help address feeding concerns and establish successful oral feeding.

Method

This cross-sectional study was conducted in a quaternary care hospital in South India. The NeoEAT-Paladai/cup in Tamil was developed through a systematic process of cross-cultural adaptation and included 64 items. A total of 325 Tamil-speaking mothers of children younger than 7 months who were fed through the paladai exclusively or partially, along with breastfeeding, completed the NeoEAT-Paladai (Tamil).

Results

Factor analysis of the responses yielded five subscales with high internal consistency reliability across subscales (α=0.858 to 0.908) and the full scale (α=0.963). Test-retest reliability was high for the full scale (intraclass correlation=0.995). Infants with feeding concerns demonstrated significantly higher total and subscale scores than infants without feeding concerns (p<0.001).

Conclusion

The NeoEAT-Paladai (Tamil) allows clinicians and caregivers of Tamil-speaking populations to monitor the feeding behaviours of infants when being fed by paladai.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Paladai is used as an alternative method to feed milk to infants in South India.

• This mode of feeding leads to an early transition to complete oral feeding for those on gavage feeding and supplements breastfeeding.

• Many validated tools are available to observe breast and bottle feeding skills, but no tool is available for paladai/cup feeding.

WHAT THIS STUDY ADDS

• We have culturally adapted and translated the Neonatal Eating Assessment Tool (NeoEAT)-Bottle feeding tool in English to NeoEAT-Paladai/cup feeding in Tamil.

• This parent/caregiver-report measure was developed to observe the paladai feeding skills of neonates and infants.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This tool does not require specialised training to administer.

• The self-administered and scored tool is expected to help primary care clinicians and follow-up clinics identify infants with problematic paladai/cup feeding for further assessment and intervention.

Introduction

Feeding through a cup, spoon, bottle and other traditional feeding devices is an alternative way to provide oral nutrition to infants who cannot breastfeed.1,4 Paladai is a traditional feeding cup made of any metal, usually stainless steel, with a relatively long snout through which the milk or other fluids can be poured into the infant’s mouth.¹ It is familiar to many traditional communities, is economical and easy to clean. The term paladai or palady, emerging from South India, has become the commonly used term for this device over the years.⁴ This feeding cup is also known by various other terms such as suthi,⁴ bondla and jhinuk across parts of India. There is increasing evidence that paladai feeding can be initiated as early as 30 weeks of gestational age within the neonatal intensive care unit (NICU), leading to the early shift of infants from NICU to the mother’s bed.^{5 6} Further evidence suggests that using a paladai leads to an early transition to full oral feeding⁷ and supplements breastfeeding in infants.^{4 7} Paladai feeding is part of the standard care protocol for feeding preterm low birthweight infants in India.⁸

Successful feeding and provision of adequate nutrition remain the focus for the discharge of infants from hospitals. It is imperative to address any parental concerns related to feeding post-discharge and identify infants who need further assessment and monitoring for feeding and nutrition. Parent report tools or questionnaires may be the best for understanding early feeding difficulties among infants after discharge, as parents are best placed to report on their infants’ feeding behaviours over extended periods, in contrast to observations by clinicians in a novel environment for the infant.⁹

The current study aimed to develop a parent/caregiver-report measure of paladai feeding for neonates and infants. Specifically, the study aimed to develop the Neonatal Eating Assessment Tool (NeoEAT)-Paladai/cup in Tamil from the NeoEAT-bottle feeding version in English.¹⁰ The NeoEAT-Bottle feeding tool comprises 64 items divided into five subscales, including Infant Regulation (13 items), Energy and Physiologic Stability (12 items), Gastrointestinal Tract Function (28 items), Sensory Responsiveness (7 items) and Compelling Symptoms of Problematic Feeding (4 items). Parents are asked to rate how frequently each behaviour occurs on a scale from Never to Always. Each item is scored on a Likert scale from 0 to 5. Higher scores indicate more symptoms of problematic feeding.¹¹ In addition to the bottle-feeding version, the NeoEAT tools include the NeoEAT-Breastfeeding tool¹² and the NeoEAT-Mixed feeding tool¹³ (breastfeeding and bottle feeding) in English. The NeoEAT-Breastfeeding tool has been adapted and validated in Tamil.¹⁴

Methods

The cross-cultural adaptation and validation of the NeoEAT-Paladai/cup feeding tool in the current study followed a methodology similar to that used in the development of NeoEAT-Breastfeeding in Tamil.¹⁴

Development and adaptation of NeoEAT-Paladai (Tamil)

The study’s first phase involved the translation and cultural adaptation of the NeoEAT-Bottle feeding tool to paladai feeding following the tool developers’ procedures.¹⁵ Independent forward translations of the tool from English to Tamil, conducted by two fluent Tamil speakers with graduate-level education, were reconciled to form a single version. Back translation from Tamil to English was carried out. The back translation was reviewed by one of the original authors of the English version of the NeoEAT-Bottle feeding tool to ensure items maintained their original intention. The tool was reviewed by three Tamil-speaking speech-language pathologists (SLPs), two neonatologists and two mothers of infants for content validation. Further, in-person interviews were conducted with five mothers of infants below 7 months to verify that the mothers understood the items as intended.¹⁶

Psychometric evaluation of NeoEAT-Paladai (Tamil)

A cross-sectional study evaluated the psychometric properties of NeoEAT-Paladai feeding (Tamil).

Participants

The sample size of 320 was arrived at considering the minimal requirement of five times the tool’s item count for performing factor analysis.¹⁷ A total of 325 Tamil-speaking mothers with children younger than 7 months participated. Participants with infants who were entirely or partially fed through paladai for the past 7 days were included in the study. The investigator noted details of the infant’s birth history and other significant medical history through a review of medical records and caregiver interviews. Infants with congenital oral structural anomalies and genetic abnormalities were excluded from the study.

Procedure

Mothers were invited to participate in the study at the time of discharge from the NICU or during visits to the outpatient departments in the hospital. Mothers provided informed consent to participate in the study. The first author (GJ), an SLP, was the primary investigator for all assessments. Participants were assigned to two groups. The first group included typically feeding infants identified by the mothers and the SLP as having no concerns related to feeding. The second group included infants with feeding concerns whose mothers reported feeding concerns and were assessed to have feeding difficulties by an SLP. There were no instances of discordance between the mother’s report of feeding concerns and that of the SLP assessment regarding the feeding problems.

All mothers individually completed the NeoEAT-Paladai (Tamil) tool in the investigator’s presence. They were encouraged to ask any questions to the investigator if they had difficulty understanding any question or wanted any clarification. We asked participants if they were scheduled to visit the hospital for a follow-up visit within the next 2–3 weeks and if they would be interested in completing the tool again. Participants who indicated their interest were contacted on their scheduled follow-up day and invited to fill out the tool a second time. A subsample of 161 participants (50%) completed the tool for the second time. A minimum of 4–5 days elapsed between the two assessments to reduce the risk of recall of the participants’ previous responses and to limit the effect of feeding development over time.

Data analysis

The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity were analysed to determine whether the data were suitable for exploratory factor analysis. A KMO measure of 0.885 and a statistically significant result (p<0.05) on Bartlett’s test of sphericity confirmed the sample size’s adequacy for factor analysis.¹⁸ An exploratory factor analysis using principal components analysis with varimax rotation, consistent with accepted guidelines for health-related instrument development, was conducted.¹⁹ The factors were extracted based on an eigenvalue greater than one. Scree plots and possible factor solutions were examined for appropriate data representation. Each factor solution was evaluated for the total variance explained, the added variance of each factor, the number of cross-loading items and the conceptual clarity of factors. Items that cross-loaded on more than one factor at 0.3 or greater were evaluated for possible movement to a different factor based on conceptual fit. The factor solution with the best conceptual clarity of items within each assigned factor was chosen. Cronbach’s α was calculated to estimate the internal consistency reliability of each subscale. Mann-Whitney U tests were done with the presentation of median and IQR to assess the difference in total and subscale scores between infants with and without feeding concerns. Test-retest reliability was assessed by intraclass correlation (ICC) between the individual subscale scores and total scores obtained at two administration time points.

Patient and public involvement

Members of the public proficient in both Tamil and English were engaged in the translation and adaptation process of the tool. Mothers of infants under 7 months were involved in the content validation of the test tool items during the development phase.

Results

A sample of 325 mothers of infants, aged 1 week to 7 months, completed the NeoEAT-Paladai (Tamil). Tables1 2 describe 220 infants with typical feeding skills and 105 infants with feeding concerns. Both full-term and preterm infants were included in the study (table 2). Mothers of singleton or twin infants participated in the study.

Table 1. Summary of sex and age distribution of the infant groups.

	No feeding concerns (n=220)			Feeding concerns (n=105)
Corrected age (months)	Male	Female	Total	Male	Female	Total	Grand total
0–2	47	52	99	18	24	42	141
2–4	44	30	74	10	17	27	101
4–6	21	10	31	1	12	13	44
6–7	11	5	16	11	12	23	39
Total	123	97	220	40	65	105	325

Corrected age used for preterm infants.

Table 2. Description of infants and family types in the two groups.

Variable	No feeding concern, n (%)	Feeding concern, n (%)
Term	34 (15.45)	21 (20)
Preterm	186 (85.54)	84 (80)
Twins	49 (22.27)	25 (23.80)
Postnatal history
Respiratory support required	144 (65.45)	33 (31.42)
NICU	79 (35.90)	52 (49.52)
Family type
Joint (including grandparents)	106 (48.18)	40 (38.09)
Nuclear (only parents)	114 (51.81)	65 (61.90)

NICU, neonatal intensive care unit .

Cross-cultural adaptation of the English version to Tamil

The detailed reviews by professionals and mothers enabled the modification of the items into Tamil. A few items were modified by replacing them with culturally appropriate words. For example, ‘Eats enough to have at least five wet diapers per day’ was modified to ‘Drinks enough milk to pass urine at least five times per day’ because diapers are not used consistently across the region’s population. In the absence of an equivalent word for ‘roots’ in Tamil, the original item ‘roots when hungry (for example, sucks on fist, smacks lips, looks for breast/bottle)’ was rephrased as ‘indicates when hungry by sucking on fingers/fist, bites lips, searches for breast/paladai’. The use of a pacifier is not encouraged in the NICU; hence, the word pacifier was replaced with ‘clean finger’ in the item ‘is easy to console when upset (for example, stops crying when held or offered a pacifier)’.

Several semantic modifications were made to the items in the Tamil version by using words and descriptive phrases in Tamil instead of literal and poorly understood Tamil translations. The English word ‘stool/poop’ was replaced with ‘motion,’ a colloquial term used in India’s southern region. In the absence of suitable Tamil words, the words ‘upset’ and ‘enjoys’ were translated as ‘restless’ and ‘interest’, respectively, as these words conveyed the same meaning and were considered more appropriate to the context. The phrase ‘food getting stuck in the throat’ was used for the English word ‘choke,’ and ‘swallowing quickly in a hurry’ was used for the word ‘gulps’ because these phrases conveyed the same meaning. No significant modifications were required in the directions to participants to complete the tool. The directions were simple and easy for the participants to understand, and they completed the tool independently.

Psychometric analysis

Item 61 (‘Needs tube feeding’) was rejected from the factor analysis as all the participants scored zero on the item. The examination of eigenvalues and total variance explained by different factors in the four-factor solution (eigenvalue: 2.703; total variance: 46.67%) and five-factor solution (eigenvalue: 2.503; variance: 50.59%) favoured the five-factor solution. The examination of the scree plot also favoured the five-factor solution. A detailed examination of individual factor loadings of all items in the two solutions revealed a greater conceptual fit of items in the five-factor solution. As many as 54 items fit into different factors based on their highest factor loadings. The remaining 10 items that cross-loaded across multiple factors were reviewed and grouped into factors similar to those in the original tool. It was ensured that such factor loadings were higher than 0.30.

Table 3 represents the distribution of items and factor loadings for NeoEAT-Paladai (Tamil). Factor 1 (Infant Regulation—13 items) and Factor 2 (Energy and Physiological Stability—12 items) are loaded with items similar to the subscales 1 and 2 in the original NeoEAT-Bottle feeding. Items from the original tool in subscale 3 with 28 items (26–53) and subscale 4 with seven items (54–60) were distributed across three subscales in the NeoEAT-Paladai (Tamil) as Factor 3 (Oral, Pharyngeal and Gastro-oesophageal Function—16 items; 27–42), Factor 4 (Responsiveness to Feeding—6 items; Item 26, 43–47) and Factor 5 (Symptoms of Problematic Feeding—17 items; 48–64). The Cronbach’s alpha coefficient estimates of internal consistency reliability for the full scale (α=0.963) and the five factors/subscales of the NeoEAT-Paladai (Tamil) (α=0.858 to 0.908) (table 3) were high, suggesting good internal consistency. The test-retest reliability (table 4) for the full-scale scores (ICC=0.995) and the five subscales (ICC=0.989 to 0.997) was high.²⁰

Table 3. Distribution of items under each factor, respective English items and factor loading for items in the Tamil version.

Factor, number of items, Cronbach’s alpha	Original item number and items in the NeoEAT-Bottle feeding tool in English, listed as per the results of the factor analysis of NeoEat-Paladai (Tamil)	Item number Tamil version	Tamil version factor loading
Factor 1 Infant Regulation 13 items α=0.90	1. Eats enough to have at least five wet diapers per day (24 hours) (R)	1	0.45
	2. Enjoys eating (R)	2	0.63
	3. Is satisfied after eating (R)	0.79	0.79
	4. Sucks strong enough to get milk from the bottle (R)*	0.44	0.44
	5. Lets me know when he/she is hungry or thirsty (R)	5	0.78
	6. Is calm and relaxed while eating (R)	6	0.64
	7. Opens mouth to accept bottle (R)*	7	0.76
	8. Is easy to console when upset (eg, stops crying when held or offered a pacifier) (R)*	8	0.60
	9. Roots when hungry (eg, sucks on fist, smacks lips, looks for breast/bottle) (R)*	9	0.56
	10. Lets me know when he/she is done eating (R)	10	0.32
	11. Likes to put fingers or toys in mouth (R)	11	0.44
	12. Stools/poops at least once per day (24 hours) (R)	12	0.49
	13. Sleeps well lying flat on his/her back (R)	13	0.53
Factor 2 Energy and Physiologic Stability 12 items α=0.91	14. Gets exhausted during eating and is not able to finish	14	0.45
	15. Breathes faster and harder when eating	15	0.74
	16. Is exhausted after eating	16	0.54
	17. Needs to rest during eating to catch breath	17	0.68
	18. Can only suck a few times before needing to take a break	18	0.37
	19. Needs to be encouraged to keep eating (such as by touching or talking)	19	0.70
	20. Holds breath when eating	20	0.41
	21. Takes more than 30 min to eat (including rest/burping periods)	21	0.54
	22. Needs help latching on to bottle*	22	0.67
	23. Wants to eat again within an hour after feeding	23	0.40
	24. Eats more than 12 times/day (24 hours)	24	0.55
	25. Gulps when eating (swallows loudly)	25	0.73
Factor 3 Oral, Pharyngeal and Gastro-oesophageal Function 16 items α=0.92	27. Seems uncomfortable after feeding	26	0.52
	28. Throws up in between feedings	27	0.56
	29. Spits up during feeding	28	0.36
	30. Throws up during feeding	29	0.55
	31. Is uncomfortable if laid flat after eating	30	0.51
	32. Becomes upset during feeding (whines, cries, gets fussy)	31	0.57
	33. Becomes stiff/rigid during or after eating	32	0.62
	34. Chokes or coughs during eating	33	0.57
	35. Sounds gurgly or like they need to cough or clear their throat during or after eating	34	0.63
	36. Is very gassy	35	0.33
	37. Coughs or chokes on saliva/spit when not eating	36	0.45
	38. Coughs in between feedings	37	0.67
	39. Gets a bloated (big or hard) tummy after eating	38	0.64
	40. Needs to be burped more than once before the end of feeding	39	0.58
	41. Gags in between feedings when there is nothing in his/her mouth	40	0.66
Factor 4 Responsiveness to Feeding α=0.86	26. Spits up in between feedings	42	0.52
	43. Arches back during or after eating.	43	0.64
	44. Tilts head back during or after eating.	44	0.70
	45. Drools milk out of the side of the mouth when feeding.	45	0.64
	46. Gets watery eyes when eating.	46	0.37
	47. Gets a stuffy nose when eating.	47	0.73
Factor 5 Symptoms of Problematic Feeding 17 items α=0.89	48. Gets a red colour around eyes or face when eating.	48	0.44
	49. Sweats/gets clammy when eating.	49	0.46
	50. Gets the hiccups	50	0.38
	51. Gags on the bottle nipple*	51	0.47
	52. Gags on a pacifier or toys put in mouth*	52	0.44
	53. Has hard stools/poop.	53	0.51
	54. Will only eat if food (milk/formula/baby food) is a certain temperature.	54	0.42
	55. Will only take the bottle from specific people (such as by mom)*	55	0.69
	56. Will only eat from a specific kind of bottle/nipple*	56	0.42
	57. Will only eat if fed in a certain way (eg, in a certain chair or held upright)	57	0.46
	58. Needs a calm environment during feeding	58	0.62
	59. Eats best when very sleepy or asleep	59	0.62
	60. Refuses the bottle before having eaten enough (such as turns head, pushes bottle away, pushes nipple out of mouth with tongue)*	60	0.49
	61. Needs tube feedings	61	-a-
	62. Gets pale or blue colour around lips when eating	62	0.57
	63. Has blood or mucous in stool/poop	63	0.56
	64. Has milk come out of the nose when eating	64	0.53

Cronbach’s alpha for full scale=0.96.

-a- No factor loading because the item was not included in the factor analysis.

^*Items with bottle were replaced as paladai.

R, reverse scoring.

Table 4. Test-retest reliability for total scores for subscale and full scale.

Subscales	Interclass correlation	95% CI
Infant Regulation	0.997	(0.995 to 0.998)
Energy and Physiologic Stability	0.996	(0.994 to 0.998)
Oral, Pharyngeal and Gastro-oesophageal Function	0.989	(0.982 to 0.993)
Responsiveness to Feeding	0.990	(0.984 to 0.994)
Symptoms of Problematic Feeding	0.991	(0.985 to 0.994)
Total Scale	0.995	(0.991 to 0.997)

The Tamil paladai version differentiated infants with feeding concerns from those without. Parental ratings of infants with feeding concerns were higher (Median=65; IQR: 55–81) than those without concerns (Median=30; IQR=21–37), lending support for known-group validity (table 5). Similar results were obtained with the subscale scores.

Table 5. Comparison of infants with and without feeding concerns.

	Feeding concerns		No feeding concerns
Subscale (possible range of scores)	Mean (SD)	Median (25th −75th percentile)	Mean (SD)	Median (25th −75th percentile)	Mann-Whitney U (p)	Effect size
Infant Regulation (0–65)	14.57 (7.01)	13 (10–18)	5.77 (3.89)	6 (3–6)	3043.0 (0.000*)	0.61
Energy and Physiologic Stability (0–60)	15.14 (6.25)	13 (11–19)	5.43 (3.58)	5 (2–8)	1542.5 (0.000*)	0.69
Oral, Pharyngeal and Gastro-oesophageal Function (0–80)	18.00 (7.07)	17 (14–21)	7.50 (4.64)	7 (4–10)	1968.0 (0.000*)	0.66
Responsiveness to Feeding (0–30)	6.77 (3.66)	6 (5–8)	2.26 (1.93)	2 (1–4)	2289.5 (0.000*)	0.61
Symptoms of Problematic Feeding (0–85)	15.04 (6.02)	14 (12–18)	5.56 (4.29)	5 (2–8)	2052.5 (0.000*)	0.67
Total scores (0–320)	70.52 (21.44)	65 (55–81)	28.52 (10.56)	30 (21–37)	130.0 (0.000*)	0.78

^*p<0.001.

Discussion

A NeoEAT-Paladai (Tamil) was developed as a caregiver report of their infant’s feeding skills when fed with a paladai. The modifications during the cross-cultural adaptation of the NeoEAT-Paladai (Tamil) were similar to those made during the development of NeoEAT-Breastfeeding (Tamil).¹⁴ Several semantic modifications were necessary to achieve cross-cultural equivalence of the items. Such similarities in the NeoEAT-Paladai may be attributed to the linguistic and semantic constructs of the words in Tamil, as well as the cultural sensitivity to feeding in the region’s population.

The factor analysis resulted in five subscales in the NeoEAT-Paladai (Tamil), similar to the original NeoEAT-Bottle-feeding tool. While subscales 1 (Infant Regulation) and 2 (Energy and Physiologic Stability) remained identical in both the tools, items were considerably rearranged among the other three subscales in the NeoEAT-Paladai (Tamil). The distribution of the items under these subscales aligned more with the NeoEAT-Breastfeeding (Tamil)¹⁴ than the original bottle-feeding tool.^{10 11} Subscale 3 was termed ‘Oral, Pharyngeal and Gastro-oesophageal Function’ in NeoEAT-Paladai (Tamil), similar to the Breastfeeding tool, with a total of 16 items. A total of six items were loaded to subscale four in the Paladai (Tamil) and this new subscale was termed ‘Responsiveness to Feeding’ as a cover term for the diverse items in this subscale. The subscale 5 ‘Symptoms of Problematic Feeding’ had 17 items and included items from the subscales ‘Sensory Responsiveness’ and ‘Compelling Symptoms of Problematic Feeding’ of the NeoEAT-Bottle-feeding tool. This subscale, with 17 items in the Paladai (Tamil) tool, overlapped with a similar subscale with 16 items in the NeoEAT-Breastfeeding (Tamil) tool. This was in contrast to only four items clustered under ‘Compelling Symptoms of Problematic Feeding’ in the NeoEAT-Bottle-feeding tool.

Overall, the results supported adequate psychometric properties of the NeoEAT-Paladai (Tamil), including test-retest reliability, internal consistency reliability and construct validity (assessed by known-group validity). The internal consistency reliability estimate of the 64 items in the Tamil tool was consistent with that of the original tool.¹⁰ The test-retest reliability coefficient obtained in the current study for the total score was greater than that obtained in the original study. The mean total scores for the group of infants without feeding concerns in the current study were lower than those reported in the NeoEAT-Bottle-feeding tool and similar to those reported in the NeoEAT-Mixed feeding¹³ and Breastfeeding tools. Infants in this group were those with no parental concerns related to feeding and who were also assessed as having no feeding difficulties through a clinical examination by an SLP. This resulted in criteria for enrolment of infants with no feeding concerns related to paladai feeding.

The NeoEAT-Paladai (Tamil) version is intended to be completed by a parent or a caregiver and shared with a clinician who can use it with their clinical assessment to guide decision-making. This tool does not require specialised training on the part of the clinician, is inexpensive to administer and uses the primary caregiver as the expert on the infant’s feeding skills. It is helpful for primary care clinicians and neonatal follow-up clinics to identify infants with problematic feeding who need further assessment and intervention.

Conclusion

The availability of a culturally adapted caregiver-reported measure for evaluating feeding through paladai helps advance clinical practice and research, as paladai is extensively used for feeding infants in parts of India. The NeoEAT-Paladai (Tamil) offers a unique opportunity for clinicians and parents of Tamil-speaking populations to monitor the feeding behaviours of infants throughout their development.¹⁴ A researcher can analyse the effectiveness of various interventions to improve feeding and swallowing outcomes using this parent/caregiver-reported measure, which offers a valid measurement.

Although the tool comprises 64 items, it simply involves checking the responses. Future studies may use item-response theory to derive a shorter version of the tool with robust items. A tool with fewer items, requiring less time for completion by mothers, will be favoured for clinical practice in busy outpatient paediatric clinics.

Data availability statement

Data are available upon reasonable request.