Quality of the development of self-report instruments assessing women’s antepartum expectations of motherhood: a systematic review

ABSTRACT

Background

The selection of valid and reliable instruments for research and clinical purposes is needed to increase our understanding of the role that expectations of motherhood have in early identification and treatment of perinatal mental health difficulties. We aimed to identify and assess the psychometric properties of self-report measures of expectations of motherhood.

Method

A systematic review was conducted and articles describing the development or validation of quantitative self-report measures of expectations of motherhood were included. The developmental quality (risk of bias) and psychometric properties of each measure were assessed according to the 2018 COSMIN guidelines.

Results

Twenty-five articles describing 16 measures were identified and included. Eleven measures were designed to assess a broad range of expectations, while the remaining measures mainly assessed expectations of childbirth. The quality of measure development was mostly poor and rated as adequate or better for only three measures. The most suitable measures with the strongest evidence of good psychometric properties are the Slade-Pais Expectations About Childbirth Scale and the Rigidity of Maternal Beliefs Scale.

Conclusions

This systematic review of measures of expectations has identified relatively few measures with adequate psychometric properties. Further development of existing measures is needed for use in both research and clinical settings.

KEY POINTS

What is already known about this topic:

1. Maternal mental health difficulties can have significant and long-term adverse consequences on mothers, their children, and families.

2. Previous evidence suggests that holding unrealistic maternal expectations may contribute to postpartum mental health difficulties.

3. Early identification of the potentially modifiable risk factor of maternal expectations may contribute to improving postpartum mental health outcomes.

What this topic adds:

1. This is the first systematic review to assess the developmental quality and psychometric properties of 16 measures of expectations of motherhood.

2. The Slade-Pais Expectations About Childbirth Scale had the strongest evidence for good psychometric properties.

3. For the majority of the measures, adequate methods for ensuring good content validity during measure development and validation were not used.

KEYWORDS:

• Perinatal
• mental health
• measurement
• quality assessment
• expectations of motherhood

The perinatal period, from pregnancy through to the early postpartum, is a time of psychological and physical transition that is often accompanied by elevated levels of stress and increased risk of mental health problems (Austin & Highet, 2017). There is a high estimated prevalence of self-reported anxiety symptoms during pregnancy (23%) and early postpartum (17.8%; Dennis et al., 2017), while the prevalence of depressive disorders appeared to be stable across the perinatal period (12%; Woody et al., 2017). Symptoms of perinatal depression and anxiety are also often comorbid with other psychological disorders (e.g., post-traumatic stress disorder, eating disorders, personality disorders) (Howard et al., 2014). The effects of maternal mental health difficulties during the perinatal period are of particular concern, as there are serious adverse consequences for women, their children, and families (Dowse et al., 2020; Wachs et al., 2009). For example, in a systematic review, Stein et al. (2014) reported significant associations between maternal perinatal depression and anxiety with their child’s emotional, behavioural, cognitive, and developmental difficulties.

During the perinatal period, women often have increased contact with a range of health professionals, providing an opportunity for early identification and intervention with women at risk of developing or presenting with perinatal mental health difficulties. A history of psychopathology and psychosocial difficulties are predictors of perinatal mental health difficulties (Howard et al., 2014). For instance, a previous diagnosis of a depressive disorder increases a woman’s risk for developing postpartum depression (PPD) more than 20 times (Silverman et al., 2017) and a history of either depressive or anxiety disorders prior to pregnancy are robust risk factors for perinatal anxiety (Martini et al., 2015). Current methods used to identify women at increased risk of developing postpartum symptoms of depression and anxiety are brief screening instruments aimed at identifying the presence of symptoms during pregnancy (Austin & Highet, 2017). However, antepartum screening is limited in its predictive power, as many significant risk factors occur during delivery and in the postpartum period (Austin & Lumley, 2003). For example, Wisner et al. (2013) suggested that 40% of depressive episodes started in the postpartum compared to 33% during pregnancy and 26% before pregnancy. In addition, Heron et al. (2004) reported that in a large community sample 44% of women with symptoms of PPD and 35% of women with symptoms of postpartum anxiety did not exhibit depression or anxiety in the antepartum. Thus, the identification of future risk factors occurring onwards from late pregnancy is needed. Researchers have identified several future pregnancy-related risk factors including gestational diabetes, preterm birth (Silverman et al., 2017), problems with breastfeeding, lack of social support, relationship problems (Robertson et al., 2004), and mode of delivery (Martini et al., 2015).

An area of research where attempts have been made to identify other potentially modifiable risk factors related to mental health focuses on women’s expectations of what the future holds across a broad range of domains. Results from two qualitative meta-syntheses of women’s experiences with perinatal depression identified the incongruity between expectations and the reality of motherhood (Beck, 2002) and disconfirmed maternal role expectations (Mollard, 2014) as overarching themes. Additionally, in the qualitative literature, researchers have consistently identified that the discrepancy between expectations and experiences of motherhood is associated with difficulties with PPD (O’Mahen et al., 2012), anxiety (Harrison et al., 2020), perceptions of birth experience (Hauck et al., 2007), and adjustment to motherhood (Choi et al., 2005; Read et al., 2012; Young, 2008).

Several avenues of research have developed measures to quantify the construct of expectations in a perinatal context by assessing areas such as maternal role fulfilment (e.g., Henshaw et al., 2014), parenthood (Harwood et al., 2007), postpartum support (e.g., Dennis, & Ross, 2006), the infant (Muscat et al., 2014), societal influence (e.g., Thomason et al., 2015), breastfeeding (Gregory et al., 2015) and obstetric events (e.g., Houston et al., 2015). However, results are often inconsistent. For instance, some studies have found that most women have positive expectations which align with their postpartum experiences (e.g., Robakis et al., 2015), while others have found significant associations between expectations and poor postpartum emotional adjustment (e.g., Henshaw et al., 2014). These inconsistencies may be explained by the quality of existing measures of expectations, as an appropriate interpretation of results depends on the degree to which measures accurately assess the construct of interest (Vogt et al., 2004).

Upon initial review of the literature, several overarching problems with measurement instruments were observed. First, there is an absence of a gold standard measure of expectations and many of the instruments used are novel (i.e. lack of validation studies) and study specific (i.e. assess a narrow range of domains). Second, the construct of perinatal expectations is often defined inconsistently across studies and used interchangeably with other concepts (e.g., beliefs and attitudes). A precise definition of the construct of interest and its function within a theoretical framework is necessary to ensure that a measurement instrument is relevant and representative of the target construct (Vogt et al., 2004). Third, inconsistent methods in instrument development coupled with limited reporting of procedures provide inadequate information about the validity and reliability of many instruments. Thus, the limitations of existing instruments of expectations may prevent appropriate interpretation of results and impede the development of a theoretical framework encompassing expectations of motherhood. To date, no reviews of instruments measuring expectations of motherhood have been identified. To provide researchers and clinicians with a comprehensive, systematic, and standardised review of instruments, the evidence-based COnsensus-based Standards for the selection of health Measurement INstruments guidelines (COSMIN; Mokkink et al., 2018; Prinsen et al., 2018) were selected to evaluate both the psychometric properties and the developmental quality of the instruments.

The aim of the current systematic review is to provide researchers and clinicians with a critical systematic analysis of available quantitative instruments intended to measure antepartum expectations of motherhood. The selection of valid and reliable instruments for research and clinical purposes has the potential to increase our understanding of the role that expectations of motherhood have in early identification and treatment of perinatal mental health difficulties. Specifically, the aims of this review are to 1) identify measures of women’s antepartum expectations of motherhood; 2) evaluate the quality of development studies, 3) evaluate the psychometric properties of instruments and 4) offer recommendations for use and further development of instruments.

Methods

The general methodology for this systematic review was conducted in accordance with the Preferred Reporting of Items for Systematic Reviews and Meta-Analyses (PRISMA; Moher et al., 2009) guidelines. The search strategy was guided by recommendations from the Joanna Briggs Institute (Aromataris & Munn, 2017). Data extraction, quality (risk of bias) assessment, summary of results, and recommendations were conducted according to the COSMIN guidelines (Mokkink et al., 2018; Prinsen et al., 2018). The protocol for this review was registered with the PROSPERO International prospective register of systematic reviews on 26 May 2016 (Registration number: CRD42016039534).

Eligibility criteria

To ensure that the predefined eligibility criteria for this review would allow for appropriate differentiation between expectations and other constructs, it was necessary to define the common terms which are often used interchangeably in this area of research. For the purposes of this review, expectations are conceptualised as mental representations of specific future outcomes and may be about an individual’s internal responses or external events (Atlas & Wager, 2013). Beliefs are defined as ideas held to be true by an individual about their world and themselves and are a result of complex cognitive processes (for detailed review see Connors & Halligan, 2014). Belief systems provide individuals with a framework for interpreting and processing information and generating expectations (Connors & Halligan, 2014). Attitudes can be defined as an evaluation of an object or concept based on an individual’s beliefs, feelings, and experiences with the attitude object (for detailed review see Maio & Haddock, 2004). Taken together, outcome expectations related to motherhood are based in belief systems, may be influenced by attitudes towards specific future outcomes, and are differentiated from these concepts as they are specifically future-oriented estimations about the likelihood of an event occurring.

Inclusion and exclusion criteria

Inclusion criteria were (1) studies describing the development or validation of quantitative self-report measures of antepartum expectations of motherhood, (2) in English, (3) peer-reviewed primary research studies where full text was available, (4) systematic reviews and meta-analyses of relevant measures and (5) theses describing the development of published measures. Exclusion criteria were (1) articles citing measures used to assess constructs other than expectations, (2) non-research articles (i.e. literature reviews, editorials, commentaries) and (3) qualitative instruments (i.e., structured interview).

Search strategy

The Cochrane library and PROSPERO register were searched to ensure that no other systematic reviews have been published on this topic. The final search was conducted on 22 February 2021. To identify articles related to expectations of motherhood a 3-step search strategy was used recommended by Aromataris and Munn (2017). First, an initial limited search of PsycINFO and EMBASE was conducted to analyse the keywords contained in the titles and abstracts of the index terms used to describe target articles. A logic grid was created to guide the search strategy including key words for the (a) domain being studied (e.g., expectations), (b) population of interest (e.g., mothers), (c) timing (e.g., antepartum), and (d) type of instrument (e.g., scale). Second, a search of databases using the identified keywords and index terms was conducted (see Appendix A for a sample search). The following electronic databases were searched: PsycINFO, PsycTESTS, Mental Measurements Yearbook and Tests in Print, EMBASE, Maternity & Infant Care Database (MIDIRS), PubMed, Scopus, and CINAHL Plus. Third, the reference lists of identified systematic reviews and a hand search of one journal devoted to the population of interest (Archives of Women’s Mental Health) were reviewed for additional studies. To identify development and validation studies that may have been missed in the original search, an additional step was added. This final step included a database search of PsycINFO and Google Scholar for each identified self-report measure and a review of reference list of all included articles.

Study selection

EndNote X9 reference management software was used to review search results. After the database searches, duplicate references were removed, and the remaining titles and abstracts were screened for relevance. Next, a full-text review was conducted of the remaining articles and of the identified self-report measures. Studies that did not meet eligibility criteria were excluded with reasons. The screening and eligibility review were conducted by the first author (IB). A random 10% sample of full-text articles was assessed for inclusion by an independent researcher and the inter-rater reliability analysis using Cohen’s Kappa statistic was conducted (κ = .83) and interpreted as “almost perfect” agreement (Landis & Koch, 1977). If the methods of the development study of a measure referred to another article or thesis, where possible this was obtained and included in the analysis. Study authors were contacted to obtain any further relevant information about the measures. Data collection was conducted independently by IB, RR, LB, RK, and KC using standardised forms.

Evaluation of measurement properties of included measures

This review used the COSMIN definitions of measurement properties (see Table 1) (Prinsen et al., 2018). The procedure for evaluating the measurement properties of included instruments consists of three steps: Step 1) Quality assessment of studies using the COSMIN Risk of Bias (RoB) checklist (Mokkink et al., 2018); Step 2) Rating of study results against the criteria for good measurement properties (Prinsen et al., 2018); and Step 3) Summary of pooled evidence was rated against the criteria for good measurement properties (Prinsen et al., 2018) and the quality of the evidence was graded using the modified Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach (Mokkink et al., 2018). The 3-step procedure was applied to the evaluation of each of the following measurement properties: 1) content validity, 2) structural validity, 3) internal consistency, 4) reliability, and 5) hypothesis testing (for construct validity). Criterion validity was not evaluated as a “gold standard” instrument in this area of research is not available. None of the included studies reported measurement error, cross-cultural validity, or responsiveness, and these measurement properties were excluded from the analysis. A detailed description of each step is provided below.

Table 1. COSMIN Definitions of measurement properties.

Measurement property		Definition
Validity
	Content validity	The degree to which the content of a measure is an adequate reflection of the construct to be measured. Three aspects of content validity include relevance (all items are relevant to construct within a specific population and context of use), comprehensiveness (no key aspects of the construct are missing), comprehensibility (items/instructions should be understood by patients as intended).
	Structural validity	The degree to which the scores on a measure are an adequate reflection of the dimensionality of the construct to be measured
	Hypotheses testing for construct validity	The degree to which the scores of a measure are consistent with hypotheses (e.g., internal relationships, relationships to scores of other instruments, or differences between groups) based on the assumption that the instrument validly measures the construct to be measured
*	Cross-cultural validity	The degree to which the performance of the items on a translated or culturally adapted measure are an adequate reflection of the performance of the items of the original version
*	Criterion validity	The degree to which the scores of a measure are an adequate reflection of a ‘gold standard’
Reliability
	Internal consistency	The degree of interrelatedness among items
	Reliability	The proportion of the total variance in the measurements which is due to ‘true’ differences between individuals
*	Measurement error	The systematic and random error of an individuals’ score that is not attributed to true changes in the construct to be measured
*	Responsiveness	The ability of a measure to detect change over time in the construct to be measured

Note. Adapted from Prinsen et al. (2018). * Measurement property not assessed in any of the included studies.

Step 1: Quality assessment of studies

The COSMIN RoB checklist (Mokkink et al., 2018) was used to assess the methodological quality of each single study on a measurement property (see Appendix B for order of boxes). The purpose of assessing the quality of studies is to determine whether results based on the study are trustworthy (Mokkink et al., 2018). The quality of a study on each measurement property was rated on a “worst score counts” procedure and yielded overall ratings of “very good”, “adequate”, “doubtful”, or “inadequate” quality. As per COSMIN guidelines, if the content validity or the internal structure of a study was evaluated as “inadequate”, that instrument was excluded from further evaluation at that point (Prinsen et al., 2018).

Step 2: Assessment of study results

The results of each single study on a measurement property (see Table 2) were evaluated against the updated criteria for good measurement properties (Prinsen et al., 2018) and each result was rated as “sufficient” [+], “insufficient” [-], “indeterminate” [?], or “inconsistent” [±]. Content validity is considered to be the most important measurement property of self-report instruments (Mokkink et al., 2018). The content of an instrument should reflect the experiences of members of the target population and as such their involvement is essential in both content development and validity studies (Terwee et al., 2018). An evaluation of content validity requires a subjective judgement by the reviewers based on a summary of multiple available evidence including: (1) the quality of the instrument development study, (2) the quality of additional studies on content validity, (3) reviewers’ subjective ratings of content, and (4) an assessment of the quality of available evidence using the modified GRADE approach (Mokkink et al., 2018).

Table 2. Updated criteria for good measurement properties.

Measurement property	Rating	Criteria
Content validity^1	+	Relevance is [+], comprehensiveness is [+], and comprehensibility is [+]
	−	Relevance is [−], comprehensiveness is [−], and comprehensibility is [−]
	±	At least one rating is [+] and at least one is [−]
	?	Two or more ratings are [?]
Structural validity	+	CTT: CFA (CFI or TLI or comparable measure > 0.95 OR RMSEA < 0.06 OR SRMR < 0.08) IRT/Rasch: No violation if unidimensionality (CFI or TLI or comparable measure > 0.95 OR RMSEA < 0.06 OR SRMR < 0.08) AND no violation of local independence (residual correlations among items after controlling for the dominant factor <0.20 OR Q3’s <0.37) AND no violation of monotonicity (adequate looking graphs OR item scalability >0.30) AND adequate model fit (IRT: χ^2>0.001; Rasch: infit and outfit mean squares ≥ 0.5 and ≤ 1.5OR Z-standardized values > 2 and < 2)
	?	CTT: not all information for [+] reported IRT/Rasch: model fit not reported
	−	Criteria for [+] not met
Internal consistency	+	At least low evidence (GRADE) for sufficient structural validity AND Cronbach’s alpha(s) ≥ 0.70 for each unidimensional scale or subscale
	?	Criteria for “at least low evidence (GRADE) for sufficient structural validity” not met
	−	At least low evidence (GRADE) for sufficient structural validity AND Cronbach’s alpha(s) < 0.70 for each unidimensional scale or subscale
Reliability	+	ICC OR weighted Kappa OR Pearson OR Spearman correlation coefficient ≥ 0.70
	?	ICC, weighted Kappa, Pearson or Spearman correlation coefficient not reported
	−	ICC, weighted Kappa, Pearson or Spearman correlation coefficient < 0.70
Hypotheses testing (construct validity)	+	75% of the results are in accordance with the hypothesis
	?	No hypothesis defined (by the review team)
	−	The result is not in accordance with the hypothesis

Note. The criteria are based on Terwee et al. (2007) and Prinsen et al. (2016).

Sufficient [+]; insufficient [−]; inconsistent [±]; indeterminate [?]. CTT classical test theory, CFA confirmatory factor analysis, CFI comparative fit index, ICC intraclass correlation coefficient, RMSEA root mean square error of approximation, SRMR standardized root mean residuals, TLI Tucker-Lewis index.

Criteria for cross-cultural validity, measurement error, criterion validity, and responsiveness were not evaluated as none of the included studies assessed these properties.

¹Definitions of facets of content validity in Table 1.

An evaluation of internal structure refers to an assessment of item-level relationships and their quality, including an assessment of structural validity (unidimensionality) and internal consistency (Prinsen et al., 2018). Structural validity is usually assessed by factor or IRT/Rasch analysis and is considered to be a prerequisite for the interpretation of an internal consistency statistic (i.e. Cronbach’s alpha). The criteria for structural validity were amended to include an evaluation of Exploratory Factor Analysis (EFA) as proposed in a similar recent systematic review by Wittkowski et al. (2020). Thus, EFA was rated as “sufficient” if ≥50% of the variance was explained by the proposed factor structure.

An evaluation of the remaining reported measurement properties (reliability and hypothesis testing) refers to an assessment of the quality of a scale or sub-scale (Prinsen et al., 2018). Although the preferred statistics for reporting reliability are the intraclass correlation (ICC) or weighted Kappa (Streiner & Norman, 2008), in practice other correlation coefficients (e.g., Pearson) are commonly reported. Thus, the criteria for reliability were amended to include an evaluation of Pearson or Spearman correlation coefficients and rated as “sufficient” if coefficient is ≥0.70 (Wittkowski et al., 2020).

Step 3: Summary of evidence and recommendations

The ratings of all available evidence from steps 1 and 2 were qualitatively summarised or quantitatively pooled and the results on each measurement property were rated against the criteria for good measurement properties (Prinsen et al., 2018). The quality of the evidence was graded using the modified GRADE approach (Mokkink et al., 2018). The GRADE ratings are presented in Table 3 and further information about this approach can be found in the COSMIN manuals (Mokkink et al., 2018; Prinsen et al., 2018; Terwee et al., 2018). The summarised results yield an overall rating of “sufficient” [+], “insufficient” [-], “inconsistent” [±], or “indeterminate” [?]. Recommendations are based on guidance by COSMIN (Table 4) (Prinsen et al., 2018).

Table 3. COSMIN Definitions of quality levels using the modified GRADE approach.

Quality level	Definition
High	We are very confident that the true measurement property lies close to that of the estimate of the measurement property
Moderate	We are moderately confident in the measurement property estimate: the true measurement property is likely to be close to the estimate of the measurement property, but there is a possibility that it is substantially different
Low	Our confidence in the measurement property estimate is limited: the true measurement property may be substantially different from the estimate of the measurement properly
Very Low	We have very little confidence in the measurement property estimate: the true measurement property is likely to be substantially different from the estimate of the measurement property

Note. The estimate of the measurement property refers to the pooled or summarised result of the measurement property of an instrument. These definitions were adapted from the GRADE approach.

Table 4. Recommendations for future use and development of instruments per category.

Category	Criteria	Recommendation
A	Evidence (any level) for sufficient [+] content validity and internal consistency	Recommended for use in research; results can be trusted
B	Evidence for inconsistent [±] content validity and/or internal consistency	Potential to be recommended for use in research; instruments require further research to assess quality
C	Evidence for an insufficient [-] measurement property	Not recommended for use; instruments require substantial further development

Note. Adapted from Prinsen et al. (2018).

Results

Systematic literature search and study identification

Figure 1 shows a flow-chart of the study selection process and a list of exclusion reasons. The database search identified 4,342 records and another 278 were identified by a search of relevant systematic review references and one journal devoted to women’s mental health. After duplicates were removed, 3,361 articles remained which were screened for relevance based on the title and abstract. The full texts of the remaining 240 articles were assessed for eligibility and 217 studies were excluded with reasons. At this stage, 23 articles describing 14 instruments were identified and included. Next, a database search of each instrument and a reference check was conducted identifying an additional 214 studies. After duplicates were removed, the remaining 141 articles were screened for relevance. The full texts of the remaining 35 articles were assessed for eligibility and 33 were excluded with reasons. In total, 25 articles describing the development or validation of 16 instruments were identified and included (see reference list for included articles).

Figure 1. PRISMA flow diagram showing study selection process and exclusion criteria.

Characteristics of self-report instruments

The characteristics of the instruments are presented in Table 5. Twenty-five articles were identified, describing the development and/or validation of 16 self-report measures of expectations of motherhood. Eleven instruments were designed to address general expectations of motherhood: Maternal Expectations Scale (MES; Flynn et al., 2010; Henshaw et al., 2014; O’Mahen et al., 2012), Attitudes Towards Motherhood Scale (AToM; Sockol et al., 2014; Sockol & Battle, 2015), Childbearing Attitudes Questionnaire (CAQ; Ruble et al., 1990), Rigidity of Maternal Beliefs Scale (RMBS; O’Mahen et al., 2013; Thomason et al., 2015), Prenatal Experiences for Mexican Americans (PESMA; Gress-Smith et al., 2013), Parenting Expectations Measure (PEM; Harwood et al., 2007; Harwood, 2004), Prenatal Maternal Expectations Scale (PMES; Coleman et al., 1999), Maternal Attitudes Questionnaire (MAQ; Warner et al., 1997), Mother’s Expectations of Parenthood Scale (MEP; Lazarus & Rossouw, 2015), Pregnancy Related Beliefs Questionnaire (PRBQ; Leach et al., 2018; Moorhead et al., 2003), and Measure of Maternal Optimism (MMO; Robakis et al., 2015). Four instruments were designed to measure the domain of expectations of childbirth: Slade-Pais Expectations About Childbirth Scale (SPECS; Slade et al., 2016), Expectations About Childbirth Scale (EC1; DeLuca, 1999; DeLuca & Lobel, 2014), Childbirth Expectations Questionnaire (CEQ; Gupton et al., 1991), and Expectations of the Childbirth Experience (ECBE; Oweis, & Abushaikha, 2004). One instrument was designed to measure infant expectations: Prenatal Expectations (compared to postnatal experience) Questionnaire (PEQ; Evans et al., 2012).

Table 5. Characteristics of included self-report instruments.

Domain	Self-report instrumen (First author, year)	Construct(s)	Target population** (Administration time)	Sub/scale(s) (number of items)	Number of items (Response options)	Range of scores (Scoring)	Original language (Country)
Motherhood	Rigidity of Maternal Beliefs Scale (RMBS; Thomason, 2015)	General expectations (Beliefs about future expectations)	Primipara/multipara (ante/postpartum)	1) Perceptions of societal expectations of mothers (9) 2) Role identity (7) 3) Maternal confidence (4) 4) Maternal dichotomy (4)	24 (1–7)	1 - 7 (Average of items)	English (USA)
	Parenting Expectations Scale (PEM; Harwood, 2004*, 2007)	Parenting expectations across a broad range of domains	Primipara (ante/postpartum)	1) Infant (11) 2) Partner (11) 3) Self (20) 4) Social (13)	55 (1–7)	1) 11–77, 2) 11–77, 3) 20–140, 4) 13 - 91 (Sum score per subscale)	English (AUS)
	Childbearing Attitudes Questionnaire (CAQ; Ruble, 1990)	General expectations/attitudes towards, pregnancy, birth, self-esteem, and relationships	Primipara/multipara (ante/postpartum)	1a) Maternal worries (6) 1b) Maternal self-confidence (4) 2) Relationship (husband) (6) 3) Relationship (mother) (4) 4) Body image (4) 5) Identify with pregnancy (6) 6) Feelings about children (4) 7) Negative self-image (4) 8) Attitude to breastfeeding (3) 9) Pain tolerance (2) 10) Interest in sex (3) 11) Denial (3) 12) Feelings of dependency (3) 13) Negatives of caretaking (3) 14) Social boredom (2) 15) Information seeking (3)	60 (1–7)	60 – 420 (Total sum score)	English (USA, CAN)
	Attitudes Toward Motherhood Scale (AToM; Sockol, 2014)	General expectations/attitudes about motherhood	Primipara (ante/postpartum)	1) Others’ judgements (4) 2) Maternal responsibility (4) 3) Maternal role idealisation (4)	12 (1–6)	12 - 72 (Total sum score)	English (USA)
	Prenatal Experiences Scale for Mexican Americans (PESMA; Gress-Smith, 2013)	General expectations about motherhood	Mexican American, primipara/multipara (antepartum)	1) Paternal support (6) 2) Family support (6) 3) Maternal role fulfillment (5)	17 (1–5)	17 – 85 (Total sum score)	English; Spanish (USA)
	The Maternal Expectations Scale (MES; Henshaw, 2014)	Common cultural expectations about transition to motherhood	Primipara (immediate postpartum)	1) Natural fulfilment (6) 2) Self-sacrifice (3) 3) Infant reflects mothering (5)	14 (1–7)	14 - 98 (Total sum score)	English (USA)
	Prenatal Maternal Expectations Scale (PMES; Coleman, 1999)	Expectations of maternal role and infant in early postpartum	Primipara (antepartum)	1) Baby (10) 2) Enjoy (11) 3) Friends (9) 4) Life (8) 5) Image (10)	46 (1–5)	46 - 230 (Total sum score)	English (USA)
	Maternal Attitudes Questionnaire (MAQ; Warner, 1997)	General expectations/attitudes about motherhood (in depressed women)	Primipara/multipara (6 weeks postpartum)	-	14 (1–4)	0 – 28 (Total sum score)	English (UK)
	Mother’s Expectations of Parenthood (MEP; Lazarus, 2015)	General expectations about motherhood	Primipara/multipara (postpartum)	1) Infant (20) 2) Support (14) 3) Self (12)	46 (1–4)	46 – 184 (Sum score per subscale)	English (AUS)
	Pregnancy Related Beliefs Questionnaire (PRBQ; Moorhead, 2003)	General expectations/attitudes about motherhood	Primipara/multipara (antepartum)	-	54 (1–7)	54 – 378 (Total sum score)	English (UK)
	Measure of Maternal Optimism (MMO; Robakis, 2015)	General expectations about motherhood (optimism)	Primipara/multipara (ante/postpartum)	-	10 (1–3)	−10 to +10 (Total sum score)	English (USA)
Childbirth	Slade-Pais Expectations of Childbirth Scale (SPECS; Slade, 2016)	Expectations of childbirth	Primipara/multipara (antepartum)	1) Coping & robustness to pain (8) 2) Staff responsive to needs (11) 3) Fear (10) 4) Out of control & embarrassed (10) 5) Perceptions of partner’s coping (6) 6) Positive anticipation of birth (5)	50 (1–5)	50 – 250 (Total sum score)	English (UK)
	Childbirth Expectations Questionnaire (CEQ; Gupton, 1991)	Expectations of childbirth	Primipara/multipara (ante/postpartum)	1) Coping with pain (11) 2) Support by partner/coach (7) 3) Nursing support (8) 4) Intervention (9)	35 (1–5)	1) 11–55, 2) 7 – 35 3) 8–40, 4) 9 – 45 (Sum score per subscale OR total sum score)	English (CAN)
	Expectations About Childbirth Scale (EC1; DeLuca, 1999*, 2014)	Expectations of childbirth (control)	Primipara/multipara (EC1 -antepartum; EC2 -postpartum)	-	11 (1–5)	5 - 55 (Total sum score AND pre/post difference score)	English (USA)
	The Expectations of the Childbirth Experience (ECBE; Oweis, 2004)	Expectations of childbirth	Primipara (antepartum)	1) Overall childbirth experience 2) Feelings during labour 3) Birth experience 4) Related to baby 5) Related to own behaviour	37 (1–4)	37-148 (Total sum score)	Arabic (Jordan) English translation
Infant	Prenatal Expectations (compared to postnatal experience) Questionnaire (PEQ; Evans, 2012)	Expectations of infant following preterm birth	Mothers of preterm infants (postpartum)	-	10 (1–5)	− 40 to 40 (Difference score pre- and post-birth)	English (AUS)

* Thesis, ** Primipara – woman who has/will give birth for first time; Multipara – woman who has/will give birth more than once.

Characteristics of samples in development studies

The sample characteristics representing the target population are presented in Appendix E. A substantial lack of consistency of reporting demographic characteristics between studies was observed. One measure was designed for a specific target population: Mexican American women of low income (<25,000 USD) and was excluded from the following summary. The parity of the sample was reported for 15 measures with six measures developed in a primiparous sample and nine measures in a combination of primiparous and multiparous participants. Mean maternal age was reported for 15 measures and ranged from 25 to 32.5 years of age. Ten measures were developed in a sample of pregnant women, three measures in a postpartum sample, and one measure was developed in a sample of pre-pregnant, pregnant, and postpartum women. The mean gestation age was reported in nine studies and ranged from 23 to 34 weeks of pregnancy. The ethnicity of the sample was reported for 12 measures with 10 measures developed in a predominantly Caucasian/White sample (range of 78–100%), while the remaining two measures reported the nationality of the sample. Relationship status was reported for 15 measures and the majority of the women were partnered (78–100%). Some information about educational attainment of the sample was reported for 12 measures. For 10 studies, a range of 0 to 22% of participants did not attain a high school education, while 13 studies reported a range of 21 to 70% of participants with a college/university education, and 10 studies reported a range of 23 to 58% of participants with a postgraduate degree. Information regarding income was reported for eight measures, with the majority of participants described as “middle class”.

Results of step 1: Quality assessment of studies

Quality of instrument development

Instrument design

For three studies concept elicitation was based on a widely recognised or well justified qualitative method in a sample of the target population and provided sufficient information about the methods and data analysis (MES, CEQ, SPECS). Another five reported using some form of qualitative methods, but did not provide sufficient information (PMES, PRBQ, AToM, CAQ, MEP). Three studies incorporated open-ended questions regarding relevance, comprehensiveness, and comprehensibility as part of a quantitative survey (AToM, PEM, MAQ). For two studies, concept elicitation was based on quantitative methods only (RMBS, PESMA). Four studies did not report any involvement from the target population (qualitative or quantitative) during concept elicitation (MMO, PEQ, EC1, ECBE). Out of the 12 studies that involved members of the target population in the concept elicitation stage, four provided sufficient information (SPECS, CEQ, MES, PESMA), while eight provided limited, unclear, or no information regarding data collection and analysis procedures which contributed to reduced ratings (PEM, PMES, RMBS, AToM, PRBQ, CAQ, MAQ, MEP).

Pilot testing

An attempt to assess the comprehensibility and/or comprehensiveness of the instrument during the pilot test phase was reported in six studies (SPECS, CEQ, AToM, PEM, MAQ, CAQ). The CEQ used brief interviews with the initial ten participants, the SPECS compared outcome of empirical procedures to initial interviews with target population, and the CAQ used follow-up interviews with a small pilot, which resulted in very good ratings. The AToM, PEM, and MAQ used open-ended questions incorporated into the survey which resulted in an adequate rating. The authors of one study (PRBQ) explicitly reported that participants were not asked about the comprehensibility of the instrument which resulted in an inadequate rating. A pilot test was conducted for the remaining nine studies, and none reported whether participants were asked about the comprehensibility and/or comprehensiveness of the draft instrument which resulted in doubtful ratings (PMES, MES, RMBS, MMO, PESMA, PEQ, EC1, MEP, ECBE).

Summary

Detailed ratings of the quality of studies on instrument development are presented in Appendix C and overall quality of development is provided in Table 6. The total quality of instrument development studies was assessed as very good for two (SPECS, CEQ), adequate for one (CAQ), doubtful for eight (PEM, PMES, MES, RMBS, PESMA, AToM, MAQ, MEP), and inadequate for five (MMO, PEQ, PRBQ, EC1, ECBE). The studies rated as having inadequate quality were excluded from subsequent data extraction. No qualitative content validity studies were identified by this review.

Table 6. Synthesis of quality of studies on a measurement property, pooled ratings of results, and the quality of evidence (GRADE).

		Quality of development studies ^1		Content validity: Combined instrument rating ^2 & quality of evidence^3
Domain	Measure	Development study	Content validity study	Content validity			Overall rating	Quality of evidence
				Relevance	Comprehen-siveness	Comprehen-sibility
Motherhood	RMBS	Doubtful	None	[+]	[+]	[+]	[+]	Low
	PEM	Doubtful	None	[+]	[+]	[+]	[+]	Low
	CAQ	Adequate	None	[+]	[-]	[+]	[±]	Moderate
	AToM	Doubtful	None	[+]	[-]	[+]	[±]	Low
	PESMA	Doubtful	None	[+]	[-]	[+]	[±]	Low
	MES	Doubtful	None	[+]	[-]	[+]	[±]	Low
	PMES	Doubtful	None	[+]	[-]	[+]	[±]	Low
	MAQ	Doubtful	None	[+]	[-]	[+]	[±]	Low
	MEP	Doubtful	None	[+]	[-]	[+]	[±]	Low
	PRBQ	Inadequate	None	NA	NA	NA	NA	NA
	MMO	Inadequate	None	NA	NA	NA	NA	NA
Childbirth	SPECS	Very good	None	[+]	[+]	[+]	[+]	Moderate
	CEQ	Very good	None	[+]	[+]	[+]	[+]	Moderate
	EC1	Inadequate	None	NA	NA	NA	NA	NA
	ECBE	Inadequate	None	NA	NA	NA	NA	NA
Infant	PEQ	Inadequate	None	NA	NA	NA	NA	NA
		Internal structure: Quality^1, pooled rating^2, & quality of evidence^3				Remaining measurement properties: Quality^1, pooled rating^2, & quality of evidence^3
Domain	Domain	Structural validity quality [rating]	Quality of evidence	Internal consistency quality [rating]	Quality of evidence	Reliability quality [rating]	Quality of evidence	Hypothesis testing/ construct validity quality [rating]	Quality of evidence
Motherhood	RMBS	Adequate [+]	Moderate	Very good [+]	High	Inadequate [-]	Very low	Very good [+]	High
	PEM	Inadequate [-]	Low	Inadequate [?]	Low	NA	NA	NA	NA
	CAQ	Very good [+]	High	Very good [+]	High	NR	NR	NR	NR
	AToM	Very good [+]	High	Very good [+]	High	NR	NR	Very good [+]	High
	PESMA	Very good [+]	High	Very good [+]	High	NR	NR	NR	NR
	MES	Very good [+]	High	Very good [+]	High	NR	NR	NR	NR
	PMES	Inadequate [-]	Very low	Inadequate [?]	Very Low	NA	NA	NA	NA
	MAQ	Inadequate [-]	Very Low	Inadequate [?]	Very Low	NA	NA	NA	NA
	MEP	Inadequate [-]	Very Low	Inadequate [?]	Very Low	NA	NA	NA	NA
	PRBQ	NA	NA	NA	NA	NA	NA	NA	NA
	MMO	NA	NA	NA	NA	NA	NA	NA	NA
Childbirth	SPECS	Adequate [+]	Moderate	Very good [+]	High	NR	NR	Very good [+]	High
	CEQ	Very good [+]	High	Very good [+]	High	Adequate [+]	Moderate	NR	NR
	EC1	NA	NA	NA	NA	NA	NA	NA	NA
	ECBE	NA	NA	NA	NA	NA	NA	NA	NA
Infant	PEQ	NA	NA	NA	NA	NA	NA	NA	NA

^aQuality of studies on a measurement property: very good; adequate; doubtful; inadequate; NA (not assessed, measure excluded); NR (not reported.

^bRating of measurement property: sufficient [+]; insufficient [-]; inconsistent [±]; indeterminate [?].

^cGRADE quality of available evidence: high; moderate; low; very low.

Quality of internal structure (Table 6). Of the remaining 11 development studies, the quality of studies on structural validity were assessed as very good for four (PESMA, AToM, CAQ, CEQ), adequate for three (MES, RMBS, SPECS), and inadequate for four (PEM, PMES, MAQ, MEP). The quality of studies on internal consistency were assessed as very good for seven (MES, RMBS, PESMA, AToM, CAQ, CEQ, SPECS) and inadequate for the four measures that did not demonstrate sufficient structural validity (PEM, PMES, MAQ, MEP). As evidence of structural validity is considered a prerequisite for the interpretation of an internal consistency statistic, these studies were excluded from further analyses. No studies described cross cultural validity.

Quality of remaining measurement properties (Table 6)

The quality of studies on reliability (test-retest) was assessed as adequate for one (CEQ) and inadequate for one (RMBS). The quality of studies on hypothesis testing for construct validity was assessed as very good for five studies (MES, RMBS, PESMA, AToM, SPECS). No studies reported measurement error, cross-cultural validity, responsiveness, or criterion validity.

Results of step 2: Assessment of study results

The results of each single study on a measurement property were summarised and overall ratings presented in Table 6. Content validity was rated as sufficient for four instruments (SPECS, CEQ, RMBS, PEM) and inconsistent for seven (CAQ, AToM, PESMA, MES, PMES, MAQ, MEP). Structural validity and internal consistency were rated as sufficient for seven instruments (SPECS, CEQ, RMBS, CAQ, AToM, PESMA, MES). Four instruments were rated as having insufficient structural validity which resulted in an indeterminate rating of internal consistency (PEM, PMES, MAQ, MEP). Two studies reported reliability (test-retest) of instruments: one was rated as sufficient (CEQ) and one insufficient (RMBS). All three studies that reported hypothesis testing for construct validity were rated as sufficient (SPECS, RMBS, AToM).

Results of step 3: Overall summary of the evidence and recommendations

A synthesis of the evidence per instrument is provided in Table 6. Results and recommendations are presented according to the intended measurement domain, and additional detailed information is provided for instrument characteristics in Table 5 and target population characteristics in Appendix D. Instruments were categorised into three categories of recommendations and suggestions for ongoing development of these instruments are provided in Table 7. A narrative overview of development procedures and recommendations is provided per instrument in each recommendation category in the next section.

Table 7. Recommendations per instrument.

Domain	Measure	Recommendation category	Clinical use	Research use	Required further development	Suggested further development
Motherhood	RMBS	A	No	Yes	Not required	Assessment of content validity; reliability study; construct validity study
	PEM	C	No	No	Structural validity study	Reliability study; construct validity study
	CAQ	B	No	No	Updated assessment of content validity	Reliability study; construct validity study
	AToM	B	No	No	Assessment of content validity (comprehensibility and comprehensiveness)	Reliability study; construct validity study
	PESMA	B	No	No	Content validity study; structural validity study	Reliability study; construct validity study
	MES	B	No	No	Assessment of content validity (comprehensibility and comprehensiveness)	Reliability study; construct validity study
	PMES	C	No	No	Updated assessment of content validity; structural validity study	Reliability study; construct validity study
	MAQ	C	No	No	Updated assessment of content validity; structural validity study	Reliability study; construct validity study
	MEP	C	No	No	Content validity study; structural validity study	Reliability study; construct validity study
	PRBQ	C	No	No	Content validity study; structural validity study	Reliability study; construct validity study
	MMO	C	No	No	Content validity study; structural validity study	Reliability study; construct validity study
Childbirth	SPECS	A	No	Yes	Not required	Clinical utility; structural validity study in larger sample; reliability study
	CEQ	A	No	Yes	Updated assessment of content validity	Reliability study; construct validity study
	EC1	C	No	No	Content validity study; structural validity study	Reliability study; construct validity study
	ECBE	C	No	No	Content validity study; structural validity study	Reliability study; construct validity study
Infant	PEQ	C	No	No	Content validity study; structural validity study	Reliability study; construct validity study

Category A: Recommended for research

There are three instruments (SPECS, RMBS, CEQ) in category A with evidence of any level for sufficient content validity and internal consistency that can be recommended for use in research and results obtained with these instruments can be trusted.

The Slade-Pais Expectations About Childbirth Scale (SPECS; Slade et al., 2016) was developed to assess pregnant women’s expectations of childbirth. Concept elicitation and item selection were based on a literature review, semi-structured interviews with members of the target population (n = 18), and an expert review. Comprehensibility was assessed in a small sample (n = 5) of the target population and comprehensiveness was reviewed by comparing the domains of the final measure with the initial qualitative interviews. The quality of measure development was assessed as very good and further development is recommended to establish clinical utility (see Slade et al. (2016) for suggestions).

The Childbirth Expectations Questionnaire (Beaton & Gupton, 1990; Bramadat, 1990; Gupton et al., 1991) was developed to assess childbirth expectations. Concept elicitation and item selection were based on a literature review and semi-structured interviews with members of the target population (n = 11). Draft 1 of the CEQ was administered antepartum to a sample of the target population (n = 202) and included open-ended questions regarding relevance, comprehensiveness, and comprehensibility. Ten participants were briefly interviewed regarding comprehensibility and comprehensiveness of the scale. The quality of measure development was assessed as very good. As the CEQ was developed 30 years ago, thus an updated assessment of content validity is suggested.

The Rigidity of Maternal Beliefs Scale (RMBS; O’Mahen et al., 2012, 2013; Thomason et al., 2015) was developed to assess the rigidity of maternal parenting-specific beliefs. Concept elicitation was based on an expert consensus approach (i.e., expert researchers and clinicians), a literature review, a review of a qualitative study (O’Mahen et al., 2012), and a small pilot test (n = 39) in an ongoing intervention study (O’Mahen et al., 2013). No details were provided describing the pilot test study design, data collection, or analysis procedures, and it was not reported whether either the ante- or postpartum administrations of the draft were used to inform comprehensibility and comprehensiveness. Content validity was rated as sufficient; however, the quality of measure development was assessed as doubtful, thus a content validity assessment is suggested.

Category B: Potential to be recommended for research

Four instruments in category B have the potential to be recommended for use with some further development required (CAQ, AToM, MES, PESMA). Category B instruments have inconsistent content validity ratings and require additional qualitative assessment of comprehensibility and comprehensiveness or a content validity study to increase confidence in this measurement property.

The Childbearing Attitudes Questionnaire (CAQ; Ruble et al., 1990) was developed to assess attitudes towards pregnancy, birth, self-esteem, and interpersonal relationships. Concept elicitation and item selection were based on previous research, a small pilot and discussions with pregnant women, and interviews with pregnant and postpartum women. No details were provided describing qualitative data collection and analysis procedures. Comprehensibility and comprehensiveness were assessed with follow-up interviews (n = not reported). The quality of measure development was assessed as adequate. Internal structure was rated as sufficient; however, content validity was rated as inconsistent as the measure reflected the experiences of participants from 30 years ago which may differ significantly from experiences of the target population today. Further required development includes an updated qualitative content validity assessment.

The Attitudes Towards Motherhood Scale (AToM; Sockol et al., 2014) was developed to assess first-time mothers’ (perinatal) attitudes about motherhood. Concept elicitation and item selection were based on previous research, modification of items from scales of general cognitive bias and from an expectations measure (PEM; Harwood et al., 2007), a manual for cognitive-behavioural therapy for postpartum depression, interviews with pregnant women (details of study not cited), an expert review, and a pilot test (n = not reported) with women who provided qualitative feedback. No details were provided describing qualitative data collection and analysis procedures, and it was not reported whether the pilot was used to inform comprehensibility and comprehensiveness. The quality of measure development was assessed as doubtful. Two follow-up studies were conducted (Sockol et al., 2014; Sockol & Battle, 2015) which were not considered to be content validity studies as it was not reported if participants were asked about comprehensiveness or comprehensibility. Internal structure was rated as sufficient; however, content validity was rated as inconsistent as aspects of maternal attitudes appear to be missing. Further required development includes a qualitative content validity assessment to verify comprehensiveness.

The Maternal Expectations Scale (MES; Flynn et al., 2010; Henshaw et al., 2014; O’Mahen et al., 2012) was developed to assess the rigidity of maternal expectations about common cultural beliefs regarding motherhood in the immediate postpartum. Concept elicitation and item selection were based on a literature review, clinical experience, an expert review, and qualitative work (Flynn et al., 2010; O’Mahen et al., 2012). It was not reported whether the pilot test was used to inform comprehensibility and comprehensiveness. The quality of measure development was assessed as doubtful. Internal structure was rated as sufficient; however, content validity was rated as inconsistent as some aspects of maternal expectations appear to be missing. Further required development includes a qualitative content validity assessment to verify comprehensiveness.

The Prenatal Experiences for Mexican Americans (PESMA; Gress-Smith et al., 2013) was developed to assess specific cultural aspects of prenatal expectations. Concept elicitation and item selection were based on previous research (Coleman et al., 1999), theory (not cited), prior empirical research (not cited), and preliminary statistical analyses (EFA) which contributed to significant modification of the scale. It was unclear whether the pilot test was used to inform comprehensibility and comprehensiveness. The quality of measure development was assessed as doubtful. Internal structure was rated as sufficient; however, content validity was rated as inconsistent based on a lack of qualitative input from the target population. Further required development includes a content validity study in a sample of the target population, followed by a structural validity study.

Category C: Not recommended for use

The remaining nine instruments in category C have evidence of an insufficient measurement property and are not recommended for use at this time. Four instruments (PEM, PMES, MAQ, MEP) have insufficient internal structure as factor analysis (FA) was not performed. Five instruments (MMO, EC1, ECBE, PRBQ, PEQ) were inadequately developed as target population involvement was not reported. Substantial further development is required before drawing conclusions for research purposes for these measures.

The Parenting Expectations Measure (PEM; Harwood et al., 2007; Harwood, 2004) was developed to assess first time mothers’ parenting expectations during pregnancy and the early postpartum period. Concept elicitation and item identification were based on a literature review and a quantitative survey of a pool of items that were adapted from previous measures and written for the study. Relevance and comprehensiveness of items were assessed in a sample of the target population (n = 87) using survey methods that included open-ended questions. Limited information was provided about qualitative data analysis and the quality of measure development was assessed as doubtful. A follow-up study was conducted (Harwood et al., 2007) which was not considered to be a content validity study as it was not reported whether participants were asked about comprehensiveness or comprehensibility. FA was not performed and the study on structural validity was rated as insufficient which excluded the PEM from further assessment. Further required development includes a structural validity study.

The Prenatal Maternal Expectations Scale (PMES; Coleman et al., 1999) was developed to assess first time mothers’ expectations of the maternal role and their infant in the early postpartum period. Concept elicitation and item identification were based on a literature review, interviews, and an expert review. However, no details were provided describing the qualitative data collection and analysis procedures. It was not reported whether the pilot test was used to inform comprehensibility and comprehensiveness. The quality of measure development was assessed as doubtful. Content validity was rated as inconsistent as the measure reflected the experiences of participants from 21 years ago which may differ significantly from experiences of the target population today. FA was not performed and the study on structural validity was rated as insufficient which excluded the PMES from further assessment. Further required development includes an updated content validity assessment, followed by a structural validity study.

The Maternal Attitudes Questionnaire (MAQ; Warner et al., 1997) was developed to assess expectations and attitudes towards motherhood in relation to postnatal depressive illness in the postpartum period. Concept elicitation was based on a literature review, clinical experience, and items were written for the study or identified from an existing expectations measure (Kumar et al., 1984). A survey pilot test (n = 50) with open-ended questions was used to assess relevance and comprehensiveness, although it was not reported whether results informed comprehensibility. No details were provided describing the qualitative data collection and analysis procedure. The quality of measure development was assessed as doubtful. Content validity was rated as inconsistent as the measure reflected the experiences of participants from 23 years ago which may differ significantly from experiences of the target population today. FA was not performed and the study on structural validity was rated as insufficient which excluded the MAQ from further assessment. Further required development includes an updated content validity assessment, followed by a structural validity study.

Mother’s Expectations of Parenthood Scale (MEP; Lazarus & Rossouw, 2015) was developed to assess mothers’ expectations (antepartum) and experiences (postpartum) about her infant, support, and herself. Concept elicitation was based on a literature review and items were identified from an existing parenting expectations measure (Harwood et al., 2007), an infant characteristics measure (Bates et al., 1979), and written specifically for this study based on personal experience and feedback from a panel of mothers. No details were provided describing the qualitative data collection and analysis procedure and it was not reported whether the pilot was used to inform comprehensibility and comprehensiveness. The quality of measure development was assessed as doubtful. Content validity was rated as inconsistent as some aspects of mother’s expectations appear to be missing. FA was not performed and the study on structural validity was rated as insufficient which excluded the MEP from further assessment. Further required development includes an updated content validity study, followed by a structural validity study.

The remaining five measures in category C were excluded from assessment based on an inadequate quality of the development study. According to COSMIN standards, evidence of target population involvement in concept elicitation is required to ensure good content validity. However, no details were provided describing target population involvement for the MMO, PEQ, EC1, PRBQ, and ECBE. For these measures, substantial further development is required before drawing conclusions for research purposes including a content validity study in a sample of the target population, followed by a structural validity study.

The Measure of Maternal Optimism (MMO; Robakis et al., 2015) was developed to assess expectations about experiences related to motherhood. Concept elicitation was based on discussions with a group of clinical providers in women’s mental health (i.e., obstetricians, psychiatrists) and clinical experience. The Prenatal Expectations (compared to postnatal experience) Questionnaire (PEQ; Evans et al., 2012) was developed to assess the difference between women’s antepartum expectations of the experience with their baby to their actual experience, when the delivery was preterm. The general design requirements of the PEQ were assessed as inadequate as a limited description of the construct was provided. The Expectations About Childbirth Scale (EC1; DeLuca, 1999; DeLuca & Lobel, 2014) was developed to assess expectations about the level of general control that women will have during labour and delivery. Concept elicitation and item selection for the EC1 were based on a literature review, previous control research (Krantz et al., 1980), and previous qualitative studies of control during caesarean deliveries (not referenced). The Expectations of the Childbirth Experience (ECBE; Oweis & Abushaikha, 2004) was developed to assess Jordanian women’s expectations of their first childbirth experience. The psychometric properties of the ECBE were assessed in a recent review of women’s childbirth experiences (Nilvér et al., 2017). However, an appraisal of the quality of instrument development was not conducted by the authors and the ECBE was included in the current review. Concept elicitation and item selection were based on a literature review and the measure was designed by the researchers. The Pregnancy Related Beliefs Questionnaire (PRBQ; Moorhead et al., 2003) was developed to assess underlying beliefs associated with postpartum depression. Concept elicitation and item selection were based on a literature review, a book of interviews with women with PPD experience (Aiken, 2001), an expert review, and interviews with two women with PPD; however, no details were provided describing the qualitative data collection or analysis procedure for the interviews and the authors reported that the pilot test was not used to inform the comprehensibility of the measure. A published validation study (Leach et al., 2018) was identified; however, no information was provided regarding further content validation procedures; thus, this study was not assessed.

Discussion

In this systematic review, we aimed to identify and evaluate instruments intended to measure women’s antepartum expectations of motherhood based on available development and validation studies. To appropriately investigate the association of women’s expectations about the future and postpartum adjustment difficulties, such as symptoms of depression and anxiety, it is important to use valid and reliable measurement instruments to assess expectations. As a result of a comprehensive 4-step search strategy, we identified 16 measures described in 25 articles for inclusion. We evaluated the methodological quality of instrument development studies and of studies on psychometric properties and rated the psychometric properties of included instruments according to a rigorous multi-step systematic process using the updated COSMIN guidelines (Mokkink et al., 2018; Prinsen et al., 2018). The findings of this review highlight the importance of evaluating the methodological quality of instrument development studies as an integral first step in the appropriate interpretation of psychometric properties of self-report measures and of the results based on that measure. In particular, assessing the extent of the involvement of the target population in concept elicitation is crucial to evaluating the content validity of a measure based on self-report. This review identified that several of the measures used in this area of research were not developed with input from the target population or had insufficient measurement properties. As such, the use of unvalidated measures may be contributing to the inconsistent results observed in this area of research. However, by providing an overview of the gaps in instrument development studies, we offered future researchers evidence-based guidance for further development of these instruments.

Implications for research and practice

Based on the results of this systematic review and in line with COSMIN guidelines (Prinsen et al., 2018), we advise on the most suitable instruments for each of the two main measurement domains: motherhood and childbirth. Overall, none of the identified instruments can be recommended for clinical use and more evidence on clinical utility (e.g., responsiveness, predictive validity) is needed for all of the instruments. To establish clinical utility, future investigations exploring associations between measure outcomes and clinical symptoms (e.g., depression, anxiety, adjustment) are needed. In addition, it is unclear why the two domains of expectations of motherhood and childbirth appear to be consistently considered separately by researchers. A possible future direction for research may be to combine these two domains in a single measure that reflects a comprehensive view of the perinatal experience and transition to motherhood.

To measure self-reported expectations of childbirth for research purposes, we recommend the use of the SPECS. We recommend that the development of new questionnaires of expectations of childbirth is not necessary, and that future research should focus on further development of the SPECS for possible use in clinical populations. To measure general expectations of motherhood for research purposes, we recommend the use of the RMBS. However, we suggest further development of this scale to increase confidence in all measurement properties. In this domain, we have identified an opportunity for researchers to develop new measures designed to assess a comprehensive range of sub-domains of expectations of motherhood including childbirth. Alternatively, researchers may wish to focus on further development of comprehensive instruments, such as the PEM or PRBQ.

Methodological considerations

Updated COSMIN guidelines

Until recently, common reporting procedures by authors of systematic reviews of measures did not often include an evaluation of methodological quality of studies (Mokkink et al., 2009). The first COSMIN guidelines (Mokkink et al., 2010; Terwee et al., 2012) proposed criteria for the evaluation of content validity but recommended that the results of all measurement properties are presented regardless of inadequate ratings. One of the changes in the latest COSMIN guidelines is the new order of evaluating measurement properties whereas an instrument is excluded from further evaluation if evidence of an insufficient measurement property is identified (Prinsen et al., 2018). Specifically, an instrument is not further evaluated if there is evidence that the quality of the instrument development study is inadequate, next if content validity is insufficient, and last if structural validity is insufficient (Prinsen et al., 2018). In a recent systematic review of measures using the latest COSMIN guidelines, Wittkowski et al. (2020) amended this procedure and evaluated the remaining measurement properties despite inadequate quality of development or insufficient content validity. The authors suggested that this was done to present a comprehensive overview of a particular measurement property. However, based on the results of the current review, we propose that reporting the results of measurement properties that are based on instruments with a serious risk of bias or insufficient content validity may encourage the use of instruments that are untrustworthy. In addition, the use of inadequately developed measures may be unethical or misleading. For example, we identified two instruments (PEM, PMES) with good internal consistency (Cronbach’s alpha) per sub-scale. However, both of these instruments had inadequate structural validity as factor analysis was not conducted and therefore there was no evidence of unidimensionality of the proposed sub-scales. Thus, the interpretation of an internal consistency statistic would not be appropriate. We suggest that in future systematic reviews, the new order of evaluating measurement properties is observed by review teams. Consequently, this practice may encourage greater consideration of methodological quality and content validity procedures during measure development.

Risk of bias checklist

The RoB checklist aims to provide a systematic approach to evaluating the trustworthiness of the results based on the methodological quality of a study (Mokkink et al., 2018). Although we have attempted to document all alterations to the checklist, we observed difficulties when attempting to apply the standards and a consequent need to make several subjective judgements. For example, some studies used quantitative methods, but incorporated qualitative questions during concept elicitation and pilot testing. Although this method is not considered to be a “widely recognised qualitative method”, such as focus groups or interviews, the review team agreed that the inclusion of a qualitative component differentiates this method from studies that used purely quantitative methods. Thus, our interpretations of some checklist items may differ to those of another review team. In addition, despite the extensive written guidance provided in COSMIN manuals, we found that many of the checklist items included vague wording that may increase the subjectivity of the quality assessment. For example, the RoB checklist (Box 1b.18) requires reviewers to assess whether “an appropriate qualitative method” was “widely recognised”, “well justified”, “assumable”, “doubtful”, or “not appropriate”. To maintain the systemic nature of the checklist, when evaluating methods that deviated from best practice, the review team resolved issues through discussion and consensus and documented our approach.

Criteria for good measurement properties

Based on the results of this review, the decision to include criteria for results of several additional statistical procedures for structural validity and reliability was made by the review team. The first alteration is the addition of EFA as sufficient when examining the structural validity in a development study. This decision was made by the review team as often EFA is necessary when a new construct is being explored and there are no clear hypotheses about underlying constructs. Second, we included correlation coefficients (Pearson and Spearman) as an acceptable statistic for reporting test-retest reliability. Third, when considering hypothesis testing for construct validity, we tested hypotheses proposed by the authors of development studies rather than formulating reviewer hypotheses. This criterion was relaxed, and construct validity was considered sufficient if the hypotheses proposed by authors were supported. In cases where no hypotheses about relationships were proposed, the review team did not suggest hypotheses.

Lack of adequate reporting

First, we found that the rigorous method recommended by COSMIN for ensuring good quality of content validity was rarely fully implemented or reported in most of the identified studies. This is unsurprising, as there is a paucity of empirical guidance concerning content validity procedures for psychological scale development. Out of the 12 development studies that involved members of the target population, reduced ratings were most often observed as a result of a lack of or insufficient reporting of qualitative research during both instrument design (for six studies) and pilot testing (for nine studies). Second, there is a lack of sufficient information and consistency when reporting results to allow for a full assessment of study results against the stringent COSMIN criteria. For example, authors most commonly reported correlation coefficients for reliability or omitted reporting detailed statistical procedures (i.e., percentage of variance explained for EFA). Third, none of the included articles reported assessment of measurement error, cross-cultural validity, criterion validity, or responsiveness.

Lack of content validity studies

The findings of this review did not identify any additional content validity studies for any of the included measures and this affected the evaluation process in several ways. First, the rating of content validity is based in part on the evaluation of the development study, which includes several subjective judgements, and in part by a “reviewer rating”. However, this was found to be beneficial for most studies. For example, most studies did not ask about comprehensibility of instructions and items, and therefore the total comprehensibility score was ultimately based on the reviewer’s rating of instructions and items which improved the rating for all of the instruments. Thus, the review team observed that the lack of additional content validity studies increased the subjective nature of the rating of content validity, and that this rating should be interpreted with caution. Second, with a lack of content validity studies, the grading of the quality of evidence (GRADE) was based on the quality of the development study only. Specifically, an assessment of a very good or adequate development study equalled a “moderate” level of quality, doubtful equalled a low level, and inadequate equalled a very low level, each time. Although the grading was somewhat redundant, it was included in the review as the semantic description of the quality of evidence helped to describe confidence of the assessments.

Development sample characteristics

The findings of this review identified substantial homogeneity in the characteristics of the sample populations across the majority of the studies. Specifically, the majority of the measures were developed in samples of Caucasian/white, partnered, well educated, middle class, and cis gender women. Future research may wish to investigate whether significant differences in expectations can be identified across various demographic characteristics.

Strengths and limitations

There are several strengths of this review resulting from the comprehensive approach to the search strategy and data extraction. The 4-step search strategy was designed according to evidence-based guidance and expanded to include an additional second search using the names of all identified measures identified in the initial search. Eligibility criteria were predefined and registered. At all points in the data extraction procedure, multiple independent reviewers were involved providing assurance of reliability. The systematic application of the current COSMIN guidelines provides users with a multi-layered evaluation of each instrument. Unpublished scales were included if referred to in published work. We implemented COSMIN advice to exclude measures from subsequent review if the quality of development, content validity, or structural validity studies was inadequate. A final strength of this review is the narrative synthesis of evidence and recommendations per instrument provided to simplify the process of interpreting results and choosing a measure.

Several limitations were identified. Although a comprehensive search strategy was employed, it is possible that not all measures of expectations were captured in our search. Specifically, our search may have missed measures developed in languages other than English. In addition, the modification of some standards and criteria may have increased the risk of bias in this review. However, we note that every attempt was made to apply these modifications strictly to all measures so as to maintain the systematic approach of this process. Lastly, we deviated from our original protocol (Best et al., 2016) on several occasions. First, after the original three-step search procedure that was detailed in the protocol was conducted, it became apparent that this was not sufficient to capture all relevant studies, and an additional step was added to conduct a second database search for each identified instrument. Second, our original aim included the evaluation of feasibility for the use of measures in clinical contexts. The primary reason for not including a feasibility appraisal is that none of the measures identified in this review were recommended for clinical populations. Although we did not present a formal appraisal of feasibility, we reported on many of the characteristics that we had specified in the protocol (i.e. administration considerations, target population) that researchers may wish to consider when choosing a measure. Finally, after the publication of the protocol, the COSMIN guidelines were updated, and we decided to utilise these in our review in place of the previous procedure.

Conclusion

This review provides an overview of existing measures of expectations of motherhood to help researchers in identifying the most suitable measure for their purposes. To select an instrument that is most appropriate for their use, we recommend that researchers first review the relevant characteristics of each of the identified instruments, with a primary focus on: 1) the measurement construct, 2) characteristics of the target population, and 3) time of administration. Following this, researchers can consider our recommendations on the use and further development of instruments. At this stage, none of the instruments are recommended for clinical use. We advise on the use of the SPECS to measure expectations of childbirth and the RMBS to measure general expectations of motherhood for research purposes.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/13284207.2022.2108694.