Introduction
A valid and reliable diagnostic instrument is the foundation of clinical practice since it gives accurate information about the patient’s health and clinical conditions [
1]. Validity is an essential property since assessing validity answers the question of whether the instrument measures exactly what it proposes to measure [
2].
In the neonatal intensive care unit (NICU), taking care of infants is complicated due to the comorbidities associated with maternal stress [
3], labor and delivery complications [
4,
5], and preterm births [
6]. Infants can show several pathologies that threaten their survival [
6], and many routine-care procedures can be both stressful and painful for the infants and, hence, influence their neurological development in the short- and long-term [
7,
8]. Besides, NICUs are growing in complexity and professionals need an ever more throughout knowledge about infants’ physio-pathology [
9‐
11].
Therefore, all neonatologists and NICU professionals need reliable and valid diagnostic tools that provide measurements about the infant’s development, growth, and clinical conditions to manage the infant optimally [
12].
In the NICU babies are handled hundreds of times per day [
13]. Despite this, touch remains largely non-specific and effects on infants understudied [
8]. Moreover, many commonly used clinical assessment procedures rarely use touch as a diagnostic tool [
12,
14,
15]. Therefore, it would be beneficial to define and develop a structured touch-based approach to assess infants.
However, assessing the validity of manual approaches has always been complicated: a previous attempt to define an evaluative procedure for the preterm newborns [
16] showed several problems indeed. One limitation was the manual nature of the procedure, which made it difficult to extend it to other NICU professionals. Another limitation was the lack of standardized manual procedures in neonatology: the procedures were somehow adapted from the ones on adults even though the anatomy of newborns, especially if preterm, and adults are different [
16]. Besides, typical weaknesses of manual approaches are the lack of reliability among different operators and the sensitivity of the procedures themselves, which can be strongly biased by the operator’s subjective experience [
17].
The Neonatal Assessment Manual scorE (NAME) model – a new assessment procedure in the neonatology field – was developed for overcoming these difficulties [
18]. The NAME aims to evaluate the infant’s general clinical conditions through the assessment of how the body tissues respond and adapt to manual stimulation, i.e., static light touch. It is designed to produce a score that could correlate with the infant’s clinical condition and therefore improve their clinical management [
18]. Moreover, it is designed to be used by every NICU professional with enough experience in the neonatal ward [
18]. Since the rationale underlying the NAME model was described elsewhere (see [
18] for more details of the NAME model significance), the present paper aims to begin the NAME validation process, investigating the model’s face, content and, as a preliminary analysis, construct validity.
Discussion
This study investigated the NAME model face, content, and, as preliminary analysis, construct validity. Results demonstrate that the NAME model seems to have good validity. The expert panel gathered to assess face validity viewed the two items of compliance and homogeneity as highly relevant to evaluate infants and to communicate about infants in the neonatal ward. The same expert panel judged the NAME model also to have excellent content validity measured through the S-CVI/Ave [
39].
However, a diagnostic tool can be considered valid only if it shows construct validity, which directly connects to the theory [
1]. Using a representative sample of hospitalized babies, we found that both gestational age and bodyweight correlated positively with the NAME score, whether categorical or numerical. In particular, we found that both gestational age and birth weight correlated stronger with the NAME score than their counterparts at the time of assessment. From a clinical standpoint, this result is relevant as both low gestational age and low birth weight represent severe risk factors for the infant’s growth and neurological development [
6,
43,
44].
The complexity index we calculated to differentiate healthy and complicated babies correlated negatively with the NAME score – the more the pathologies, the less the capacity of the baby’s body to adapt to an external stressor. This finding supports our hypothesis that the NAME model can discriminate between healthy and complicated infants. The two scores, categorical and numerical, seemed to behave similarly, even though, at first, the numerical scale could be viewed as more sensitive due to being larger than the categorical scale. This result supports the hypothesis that the categorical scale made of the scores Bad, Marginal, and Good can efficiently categorize the infants according to their conditions. Therefore, the NAME could become part of the neonatology ward routine-care to assess the infant’s general clinical condition.
Regarding other widely used procedures, such as the Alberta Infant Motor Scale [
15] or the Assessment of Preterm Infants’ Behavior [
14], the NAME can be performed in the crib/incubator, even when the infant is asleep, thus reducing the risk of distressful position or maneuver. For its intrinsic easiness, the NAME can also be applied in partially stable infants. Based on the literature, it took less time to be performed compared to other procedures. Except for pain assessment [
45], those procedures take from 30 min – for assessing newborns, to several hours – for writing the clinical report [
14,
46]. Concerning a previous attempt [
16] to codify a neonatal manual procedure, the NAME is more straightforward since it involves static or gentle touch that can be performed by every professional using touch-based procedures. Moreover, it is far less stressful since the previous proposed assessment procedure lasted 10 min.
Future research should further test the NAME model to make sure it could be efficiently introduced in clinical practice and help professionals to define good therapeutic plans – in particular, its advantages over the existing assessment procedures need to be consistently and specifically tested. A new test should indeed be easier to use, take less time to be performed, show better predictive capacity, or reduce the costs [
1].
Content validity might require an additional study. One key question we might pose is: Does the measurement scale include every aspect related to the underlying construct – the rationale – and exclude what is irrelevant? [
2] At this stage, it is difficult to answer in a definitive way to such a question, especially in the medical field, where researchers and professionals deal with complex systems [
2]. However, future studies might shed light on additional elements to take into account. Following the expert panel’s judgment, the NAME model might also be enhanced by integrating physiological parameters like heart rate, respiratory rate, blood pressure, SpO
2, temperature (vital signs already monitored in NICUs) [
47,
48], or heart rate variability – a measurement that is emerging as reliable and valid to predict and monitor the infants’ clinical progression [
49].
In the same way, the assessment of construct validity requires further evidence. Despite the positive results, the significant correlations between the NAME score and the infants’ characteristics (age, weight, and complexity index) were not perfect – the Kendall’s τs were far from the score of 1 (perfect agreement). This result was expectable: weight and gestational age are just two characteristics that may give important information about the babies’ health, but they certainly do not depict the whole clinical picture. Concerning the complexity index, a limitation is that it was calculated evaluating the quantity, not the severity of complications: in fact, it is conceivable that infants with fewer but more severe complications might receive a worse score than infants with more complications that are, however, less severe. Indeed, the present paper represents just the first step of the NAME validation process: further studies should assess the correlation between the NAME score and specific clinical outcomes, and even whether changes in the NAME score could recognize acute changes in clinical conditions, e.g., in case of sepsis or other emergencies. Besides, could ROIs on the infant’s body indeed be found? And could they correlate with specific clinical conditions? (e.g., could the ROIs in the upper thorax correlate with respiratory pathologies?).
In the present paper, we assessed only discriminant validity, which is whether the test can discriminate among people with different conditions (e.g., healthy vs. diseased) [
1]. Still, other forms of construct validity can be assessed. Examples are convergent validity – if the test correlates with the measurements of other instruments that can assess some elements of the underlying construct – and divergent validity – if the test fails to correlate with tools that relate to different constructs [
1].
We could not assess criterion validity, the fourth facet of validity, due to a lack of a gold standard test that evaluates how the infant’s body mechanically adapts to external stressors. Indeed, criterion validity compares the new test being assessed and an existing criterion that is established as valid, and it can be evaluated as concurrent validity and predictive validity. Concurrent validity is assessed when the new test is compared in the present with an existing criterion that measures the same construct. Predictive validity is assessed when the new test is compared with a criterion in the future. When possible, the criterion used for comparison should be a gold standard [
1,
2]. Future studies could evaluate criterion validity through correlating the NAME score with other scales that measure the infant’s development or conditions.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.