Development and testing of a self-report measure of preparing to parent in the context of a fetal anomaly diagnosis

Abstract

Objective:

To generate a self-report instrument to capture clinically relevant variations in expectant parents’ caregiving development, specified by how they are preparing to parent an infant with a major congenital anomaly.

Methods:

Recent literature structured domains to guide item generation. Evaluations by experts and expectant parents led to a refined instrument for field testing. Psychometric testing included exploratory factor analysis, internal consistency, and test-retest reliability.

Results:

Samples included expert evaluators (n = 9), and expectant parent evaluators (n = 20) and expectant mother field testers (n = 67) with fetal anomaly diagnoses. Preparing to Parent-Act, Relate, Engage (PreP-ARE) resulted from a three factor solution that explained 71.8 % of the total variance, with global Cronbach’s α = 0.72, and sub-scales 0.81, 0.65, 0.72 respectively. Cohen’s weighted kappa indicated all items were acceptably reliable, with 14 of 19 items showing moderate (≥ 0.41) or good (≥ 0.61) reliability. Convergent validity was found between the maternal antenatal attachment and Act scales (r = 0.39, p = 0.001).

Conclusion:

This empirically-based instrument was demonstrated to be valid and reliable, and has potential for studying this transitional time.

Practice Implications:

PreP-ARE could be used to understand patient responses to the diagnosis, level of engagement, readiness to make decisions, and ability to form collaborative partnerships to manage healthcare.

1. Introduction

Parental caregiving development begins during pregnancy [1–3], and is influenced by biological, psychological and social factors [4]. For expectant parents who continue pregnancy in the face of a major or life-threatening fetal anomaly, however, the process of parental caregiving development is complicated by uncertainty and psychological distress [5–7]. This development is further shaped by their own emotional and cognitive responses to the fetus and the diagnosis, and their interactions with family members and others regarding fetal and infant health [6]. Caregiving development specific to this context requires collaborative relationships with clinicians [8,9]. Thus, clinicians become integral to a parent’s caregiving development.

Knowledge of the construct of prenatal parental caregiving has been limited to what is known about preparation for becoming a mother, and parent-infant attachment in a healthy context [3,4]. Empirical study of variations in parental caregiving development for purposes of a more systematic approach towards prenatal clinical assessment and counseling, or outcomes-focused research is lacking. No systematic assessment tools are available to identify variations, strengths, or deficits in caregiving development that might indicate a need for intervention. The aim of this study was to generate and evaluate a self-report measure for capturing variations in expectant parents’ caregiving development before the birth of an infant with a major anomaly.

2. Methods

A stepwise design was used to develop and test Preparing to Parent - Act, Relate, Engage (PreP-ARE), using the developmental and judgment-quantification stages described by Lynn [10] and the guidelines of DeVellis [11] and Sapnas and Zeller [12]. The first step, the developmental stage, focused on identifying the conceptual domains and sub-domains (Table 1) using the contemporary, relevant literature [5–7,9,13–20]. A pool of 98 items was generated to correspond to these domains and sub-domains [11,21–22]. Instructions for a 4-point Likert-type scale were drafted: Please mark your response to each of the following statements from 1=“strongly disagree” to 4=“strongly agree.” Please respond to these statements based on what you think or feel, not based on other’s views about how you should think or feel.

Table 1.

Domains and sub-domains for item generation to capture the phenomenon of preparing to parent after a fetal anomaly diagnosis.

Domain	Sub-Domains
Mental outlook for “getting through it”
	Managing one day at a time
	Choosing a scenario
	Cultivating mental flexibility
	Having hope
Health information
	Engaging in healthcare
	Recognizing and supporting the child
	Protecting the child
	Doing what is necessary
	Focusing on logistics
	Considering what could be known
	Seeking normative information
Interactions with others
	Determining the meaning of the diagnosis
	Using variety of resources
	Accepting low level of control
	Compartmentalizing diagnosis
	Carrying on as usual
	Censoring conversations
Experiences of loss
	Documenting the pregnancy/child
	Being with the child in utero
	Distancing oneselffrom the pregnancy/fetus
	Promoting acceptance of the pregnancy/ future child
Motivations
	Wanting healthcare providers to offer holistic care
	Interacting with healthcare providers in certain ways
	Developing agency as a parent
	Reconciling illness and non-illness related care

The second step, the judgement-quantification stage, involved quantitative and qualitative evaluations of items for content validity. Nine evaluators in clinical and/or research roles across the United States in nursing, medicine, social work, and psychology, and with expert knowledge on expectant parents’ experiences following fetal anomaly diagnoses, provided evaluation data through an online survey [23]. Complying with Lynn’s approach [10], data were collected on all 98 items using a 4-point scale for relevance (4=Completely Relevant, 3=Somewhat Relevant, 2=Somewhat Irrelevant, 1=Completely Irrelevant) and clarity (4=Completely Clear, 3=Somewhat Clear, 2=Somewhat Unclear, 1=Completely Unclear), to determine content validity index (CVI) ratings. Endorsement of an item by at least 7 of the 9 experts, or a CVI ≥ 0.78 for both relevance and clarity was required to be acceptably content valid. Ratings of relevance ranged from 0.50 to 1.00 with a mean of 0.94, and clarity ratings ranged from 0.38 to 1.00 with a mean of 0.93. Additionally, experts had the opportunity to provide qualitative feedback regarding each item. Based on this analysis and research team considerations of conceptual sound-ness, redundancy, and item complexity, 29 items were eliminated. The remaining 69 items were verified for ease of readability [11].

The third step of this process was cognitive interviewing of experiential experts, which is expected for a comprehensive approach to instrument development [23–25]. Following institutional review board approvals, purposive sampling [26] was used to recruit a sample of 20 eligible expectant parents who reasonably reflected this patient population at three regional care centers in two Upper-Midwestern states. Fetal anomaly diagnoses included those affecting the circulatory (e.g., complex congenital heart disease), digestive (e.g., omphalocele, gastroschisis), central nervous (e.g., spina bifida), and respiratory systems (e.g., congenital pulmonary airway malformation), as well as chromosomal syndromes (e.g., Down syndrome). Using a “think aloud” approach [24,24,25], items were read to the expectant parent, who was then asked to restate the item in their own words, and provide feedback on the relevance and meaning of the item. After analysis, 43 items were retained and refined for field-testing.

The fourth step was field-testing the instrument PreP-ARE. Institutional review board approvals for field testing were obtained at the same three sites used for cognitive interviews. The 43 retained items (using the same response options) were tested with a new convenience sample of expectant parents via an online survey between July 2017-July 2019. The survey included demographic and health questions, and the maternal antenatal attachment scale (MAAS) [27] for testing convergent validity. MAAS is a 19-item 5-point Likert scale designed to measure a pregnant woman’s frequency, extent and type of attention toward the fetus and imagined infant [27]. PreP-ARE items were designed to capture more than antenatal attachment, including the extent to which expectant parents seek information about fetal health and development, and whether they perceived themselves as taking a lead or collaborative role in fetal care and preparing for infant care. Expectant parents generated an anonymous code used for de-identification, and for linking test-retest data [28].

Analysis of data collected from field testing was conducted with SAS version 9.4. Descriptive statistics were calculated for sample characteristics. Although the sample size was smaller than 10 participants per item [21], smaller sample sizes have been shown to be adequate for psychometric testing [12]. Inter-item correlations and the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy were examined at the item level. Exploratory factor analysis (EFA) with a varimax rotation was used to extract the factors. Pearson’s correlations between the factors were calculated to discern whether separate dimensions of the phenomenon were captured. Average inter-item and corrected item-total correlations, and Cronbach’s coefficient α were calculated to determine internal consistency. Cohen’s weighted kappa was then used to examine test-retest reliability.

3. Results

The sample for field-testing PreP-ARE included 67 pregnant women between 18–43 (mean = 29.7; SD = 5.8) years of age, completing the survey between 24–38 weeks of pregnancy (Table 2). The major, non-lethal fetal anomaly diagnoses in the cognitive interview and field-testing samples were qualitatively similar.

Table 2.

Sample characteristics.

Sample characteristics for cognitive interviews.
Expectant parents		Total n = 20
Reported gender
Women		13
Men		7
Relationship status
Married		15
Partnered		4
Single		1
Race
Black		2
White		18
Highest level of education
High school degree		2
Vocational degree		5
Undergraduate degree		8
Graduate degree		5
Annual household income (USD)
Less than $9999		2
$50,000–69,999		6
$70,000–89,999		3
$90,000–109,999		2
$150,000 or more		5
Declined to answer		2
First child
Yes		9
Sex of fetus (singletons)
Male		12
Female		8
Sample characteristics for field-testing.
Expectant mothers	Total n = 67	%
Relationship status
Married	48	72
Partnered	16	24
Single	3	4
Race
Black	4	6
White	56	83
Asian	5	8
Other	2	3
Ethnicity
Hispanic	2	3
Highest level of education
High school degree	21	31
Vocational degree	3	5
Undergraduate degree	30	45
Graduate degree	13	19
Annual household income (USD)
Less than $9999	6	9
$10,000–29,999	7	10
$30,000–49,999	7	10
$50,000–69,999	8	12
$70,000–89,999	6	9
$90,000–109,999	13	19
$110,000–129,999	7	10
$130,000–149,999	3	4
$150,000 or more	9	13
Prefer not to answer	1	1
First child
Yes	29	43
Sex of fetus (singletons)
Male	34	51
Female	28	42
Unknown	5	7

Analyses of the 43 items led to identifying items with extreme responses (e.g., only “agree” or “strongly agree”), which prompted elimination of 2 items. An iterative process of examining inter-item correlations and the KMO resulted in removal of 19 items with a KMO < 0.50, resulting in 22 items remaining. Adequacy of the sample size, KMO = 0.68, p = 0.001, was established for proceeding with EFA with these 22 items. Results indicated 4 factors with eigenvalues greater than one, and the Scree plot supported a structure interpretation of 2, 3, or 4 factors.

The 3-factor specified solution was most robust (Table 4). Factors I, II, and III explained 71.8 % of the total variance. Within the 3-factor solution, items that did not meet the recommended 0.35 threshold for minimum load strength and were eliminated [21]. Factor I was the largest, with one item removed. Due to a low factor loading as well as tight cross-loading with factor II, the following item was removed, “I like having extra time to look at the baby on the ultrasound screen.” The resultant 9-item Act scale explained 32.4 % of total variance. Factor II had two items removed. One item, “I want to set aside time to be with the baby during pregnancy (e.g. feel movement, talk or read to baby, or play music for baby),” was removed due to a low factor loading. A second item, “There is nothing I can do about the baby’s health,” was removed due to a low factor loading as well as tight cross-loading with factor I. The resultant 5-item Relate scale explained 17.1 % of total variance. Factor III retained all items, resulting in a 5 item Engage scale, and explained 22.3 % of total variance. The resulting 19 items constituted the final version of PreP-ARE.

Table 4.

Means, standard deviations, mean response ranges, Pearson’s correlation coefficients between each factor, and totals, N = 67.

Factor	Mean	SD	Rangea	1.Act	2.Relate	3.Engage
1.Act	3.5	0.4	1.4 – 4.0	1.000
2.Relate	2.8	0.4	1.8 – 4.0	0.011	1.000
3.Engage	3.1	0.4	1.7 – 3.8	0.070	−0.006	1.000
Total	3.3	0.3	2.3 – 3.9	0.765	0.461	0.493

^aPreP-ARE response options range from 1=”strongly disagree” to 4=”strongly agree.”.

Pearson’s correlations indicate that the factors capture unique dimensions of preparing to parent in this context (Table 4). The average inter-item correlations for Act = 0.54, Relate = 0.49, and Engage = 0.65, and the corrected item total correlations were all positive and more than 0.36, which support reliability. The Cronbach’s coefficient α for each scale was Act = 0.81, Relate = 0.65, Engage = 0.72, and the global scale = 0.72. All but one scale (Relate), exceeded the minimum acceptable α=0.70, [21,29]. No benefit was found in eliminating any item based on the “alpha when item deleted.”

Test-retest of PreP-ARE was completed 11–26 days (median = 15 days) after the first administration by a sub-set of 46 women [22]. Cohen’s weighted kappa coefficients suggested that all items were acceptably reliable, with 14 of 19 items demonstrating moderate (≥ 0.41) or good (≥ 0.61) reliability for this initial field testing [30].

Regarding convergent validity, moderately strong and significant correlation between the global MAAS scale and Act scale, r = 0.39, p = 0.001. Moderate r>0.32, significant p > 0.01 correlations were found between the MAAS scale and three Act items (Table 3).

Table 3.

Results of the Preparing to Parent - Act, Relate, Engage (PreP–ARE)^a item analysis and exploratory principal components factor analysis with varimax rotation (N = 67).

	Factor I	Factor II	Factor III
	ACTb	RELATEc	ENGAGEd
What I say during clinic visits is heard and respected.	0.72
I need to be hopeful about the baby’s condition, no matter what.e	0.62
Now, I need to focus only on the best outcome for my baby.	0.60
I trust the healthcare team.	0.59
I find out what I need to know from the healthcare team.	0.57
It is my job to make sure the right decisions are made for my baby.f	0.55
I need to be ready for anything, from the best to the worst that could happen with my baby.	0.54
I am working to understand the details of the baby’s health condition.g	0.53
I can enjoy being a parent while also taking care of my baby’s health condition.	0.42
I would like the healthcare team to learn more about what I think is important.		0.72
I would like the healthcare team to find out from me what I already know about the health condition.		0.62
I would like the healthcare team to help me think in the short term, instead of the long term.		0.50
The new things I find out about the baby always seem to make me feel bad.		0.45
I would like someone from the healthcare team to help me if I am very sad, worried, or afraid.		0.43
My family members do not need to know about the baby’s health condition.h			0.80
I have many family members and friends who I will need to keep informed about the pregnancy and baby.			0.80
My friends and coworkers do not need to know about the baby’s health issue.i			0.55
It is important to talk about what the baby’s health condition could mean for the family.			0.44
I do things like keep a journal, or share pictures and videos, about the pregnancy and my baby.			0.41
Eigenvalues	3.52	1.85	2.42
% Explained Variance	32.40	17.10	22.30

^aThe Preparing to Parent–Act, Relate, Engage (PreP–ARE) is copyright property of the first author (ACM).

^bFactor I, ACT describes parents’ positive experiences and efforts as they act as a partner in healthcare.

^cFactor II, RELATE describes parents’ held expectations of healthcare providers and consequences of clinic appointments.

^dFactor III, ENGAGE describes parents’ efforts to reflect on and communicate with others about maternal-fetal and anticipated infant health.

^eCorrelation with global MAAS scale r = 0.37, p = 0.002.

^fCorrelation with global MAAS scale r = 0.32, p = 0.01.

^gCorrelation with global MAAS scale r = 0.34, p = 0.005.

^hReverse scored.

ⁱReverse scored.

4. Discussion and conclusion

4.1. Discussion

Results from the EFA supported the conceptual basis for PreP-ARE as a multidimensional scale with three separately scored factors. Items woven throughout the three scales underscored the salient influence of an expectant parent’s efforts to interact in a medically-informed manner on the transition to parenthood. The most robust first scale, Act, included items related to the expectant parents’ ratings of abilities and knowledge as an emerging care provider for a fetus, and future infant who will not be born healthy. The Relate scale represents preferences regarding how clinicians should interact with expectant parents to support collaboration and tailor care. These items also highlight the positionality of the expectant parent as a novice, yet with the potential for emotional turmoil that could undermine caregiving development. The Engage scale is focused on disclosure juxtaposed with efforts to make meaning of having an infant who needs specialized care with others.

PreP-ARE also captures emotional states. Studies have highlighted the severity and potential chronicity of psychological distress in this population [31–33]. Future testing of PreP-ARE for correlation with depression and traumatic stress could provide opportunities for key early intervention. Further, PreP-ARE was based on study findings that included a range of major fetal anomalies, suggesting that the phenomenon is commonly experienced. Additional field-testing is needed, perhaps in sub-groups by diagnosis.

There are two main limitations to recognize. First, factor II, Relate, did not reach the acceptable alpha = 0.70 [29]. Sample size alone could account for this result. It should be noted, however, that four of the five items in this factor specify preferences regarding how the healthcare team could relate to the expectant mother. Participant characteristics (e.g., personality, cultural background) as well as previous experiences with healthcare may influence these preferences. It is also possible that this scale requires additional items to represent this dimension sufficiently. Second, while the sample size was sufficient for EFA, and offered a beginning point for examining the validity and reliability of this new instrument [12], additional field testing is needed. For example, a larger sample size would allow sub-group analyses of women with different maternal age, risk factors, and gestational weeks. A more diverse sample is also needed to address health inequities reflected in inadequate prenatal care for Black and Hispanic women [34].

4.2. Conclusion

Initial field testing of PreP-ARE demonstrates validity and reliability as a measure of variation in parental prenatal caregiving development in the context of expecting a medically fragile infant. With an expanded understanding of expectant parents’ efforts to prepare, tailored counseling can better support collaborative relationships between families and clinicians to reach the goal of optimal health outcomes. Further research could provide evidence that this instrument is predictive of family and infant outcomes.

4.3. Practice implications

Clinicians caring for patients and families with complex, and life-threatening fetal anomaly diagnoses need empirically-developed and tested measures to inform counseling strategies. PreP-ARE can be administered in less than 5 min, and could assess the extent to which an expectant parent is able to take on a caregiving role, is ready to engage in collaborative care, and requires support for mental health.