A New Scale of Readiness for Health Care Students to Encounter Partner Abuse

Abstract

Partner abuse (PA) is associated with significant morbidity and mortality worldwide. Health care practitioners regularly encounter patients experiencing PA and require comprehensive education on how to respond. This study describes the creation and validation of a new measure of readiness to encounter patients experiencing PA for health care practitioners and students.

Initial item development and content validation were informed by expert feedback. Psychometric properties were assessed using data collected from Australian health care students, using Principal Components Analysis (PCA) and Confirmatory Factor Analysis (CFA). Internal consistency, inter-scale correlations, and test–retest analysis were performed.

An initial pool of 67 items was reduced to 48 following content validation by 5 experts as a measure of construct validity. A total of n = 926 responses were collected, which were randomly split into two groups to perform a PCA and CFA. The PCA resulted in a 31-item version, which was further reduced to a 27-item version following the CFA, containing four factors. Internal consistency and test–retest analyses demonstrated good reliability.

The produced scale is a 27-item measure of readiness to encounter patients experiencing PA, which has demonstrated good psychometric properties with a sample of Australian health care students. Results indicate that self-efficacy and Emotional-readiness are a large component of readiness. The scale may be used to measure the readiness of a cohort, or as a pre and post-intervention measure, and results may provide insight into the educational needs of a cohort.

Keywords

domestic violence intimate partner violence psychometrics scale development scale validation health care allied health

Introduction

Partner abuse (PA; also called intimate partner violence) is a major health issue associated with significant morbidity and mortality worldwide (García-Moreno, 2013). Health care practitioners are believed to regularly encounter patients with injuries and illnesses caused by or contributed to by PA (Black, 2011). Calls have been made for increased education for health care students and practitioners, which aims to increase their readiness to recognize and refer patients experiencing PA to specialist care and advocacy (World Health Organization [WHO], 2013). Despite these calls, current evidence indicates health care students and practitioners rarely receive any meaningful PA education (Bravo et al., 2019; Sawyer et al., 2016), and evaluations of educational interventions often focus on knowledge and do not measure impacts on confidence and skills (Sawyer et al., 2016). A key barrier to reporting on the outcomes of educational interventions in many health care professions is the lack of validated outcome measures designed for use specifically with their practitioners and students. To assist with identification of effective educational interventions it would be useful to develop a scale that is designed for use with any health care practitioner or student from any discipline.

There are several scales with reported psychometric properties that have been used to measure PA educational intervention outcomes in some health care cohorts. The Physician REadiness to Manage Intimate partner violence patients Scale (PREMIS; Short et al., 2006), and the scale created by Maiuro et al. (2000), demonstrated good psychometric properties when used with U.S. physician groups (Maiuro et al.’s with nurses as well). The psychometric properties of a modified version of the PREMIS were measured with a cohort of U.S. health care students including dental, nursing, social work, and medical, with the authors identifying six of the original eight subscales and varied internal consistency subscales (Cronbach’s alpha between .46–.91; Connor et al., 2011). A recent study of the PREMIS when delivered to Australian paramedic students found the scale did not demonstrate robust properties, only identifying five of the original eight subscales, most of which demonstrated low internal consistency (Cronbach’s alpha between .47 and .80; Sawyer et al., 2019).

The above-mentioned scales were designed for use primarily with practicing physicians, rather than for any health care practitioner or student group. Additionally, they were created before the introduction of the first major clinical guidelines for health care practitioners to respond to PA in 2014 (WHO, 2013), meaning items were generated before the current conceptual underpinning for the health care sector’s response to PA was produced. Likewise, it has been noted that while these scales are measures of practitioner readiness they were not published along with a theoretical foundation for the construct of “readiness” they were measuring (Leung et al., 2017a).

The authors of the more recently developed General practitioners’ Perceived Readiness to identify and respond to Intimate Partner abuse Scale (GRIPS; Leung et al., 2017a) noted that self-reporting knowledge, attitudes, and behaviors framework may not encompass aspects of readiness such as Emotional- and Motivational-readiness. The authors cited qualitative work with GPs which indicated that these aspects of readiness are central to a practitioner’s willingness to act on their suspicions of PA, and therefore critical to assess and measure when attempting to define the “readiness” of a person or workforce (Leung et al., 2017a). Additionally, it is known from qualitative research with health care practitioners that a major barrier to discussing PA with patients is feeling underprepared or lacking confidence (Sprague et al., 2012), and therefore these aspects of readiness should be considered in an outcome measures of overall readiness.

There is a clear need for increased PA education for all health care students and practitioners (García-Moreno et al., 2015) and the accurate measurement of the efficacy of educational interventions should be prioritized. It is timely to create a new scale designed to be applicable to all health care students and practitioners, and which is informed by a robust conceptual underpinning of the concept of readiness, and the health care sector’s framework for responding to PA. This study aims to create a psychometrically sound scale, which can be used with all health care student and practitioner populations, to assess their readiness to encounter patients experiencing PA.

Methods

Study Design

Initial items were developed and tested for construct validity, following which initial psychometric testing was performed with a sample of Australian health care students.

Initial Item Development

Before creation of initial items, the concept of “readiness” required clear definition. The creators of the GRIPS performed an initial study which explored this concept with regards to the health care response to PA to inform their own scale development (Leung et al., 2017b). In this study, the authors used qualitative data from General Practitioners to help define the term “readiness,” finding that the concept could be defined as a “…psychological state that indicates the extent to which an individual is cognitively, motivationally, and emotionally inclined to embrace and respond to a phenomenon” (Leung et al., 2017b). The authors found that terms “readiness” and “preparedness” were distinct, and they identified six individual constructs that made up the terms (refer to Table 1). Preparedness was defined as holding the required knowledge, resources, and skills to respond when encountering a patient experiencing PA (Leung et al., 2017b). Items in the present study were generated with this definition as the theoretical underpinning for “readiness” in the context of a health care practitioner responding to PA.

The World Health Organization’s (WHO, 2013) clinical guidelines were used to provide the theoretical basis for the knowledge, attitudes, and skills which a health care practitioner should hold in regards to PA. This allowed us to generate items that would theoretically measure an individual’s readiness for the actual tasks that a health care practitioner would be expected to undertake when encountering a patient experiencing PA.

When developing items terms were chosen which would be answerable by both students and practitioners, meaning items were worded without reference to current practice behaviors or previous clinical patient contact. Items were designed to be understandable and relevant to international cohorts, by ensuring use of internationally recognizable terms and concepts.

The term “partner abuse” was chosen over “intimate partner violence,” as anecdotal evidence from previous education delivered to Australian health care practitioners and students indicated that few people understood the meaning of the term intimate partner violence, while the term partner abuse was generally understood. Preference was therefore given to the term partner abuse, as the term that the intended users of the scale would be more likely to understand.

An initial pool of 67 items was developed and then presented to a panel of five family violence experts for face and content validation. Experts were chosen from prominent Australian family violence researchers working across several different universities and health care sectors. The experts made comments on the appropriateness and clarity of items and suggested changes, based on which a final version of 48 items was derived which included items from each of the six constructs previously identified by Leung et al. (2017a).

Table 1.

Readiness and Preparedness Constructs Identified by Leung et al. (2017b) (11).

Domain	Construct	Description
Readiness	Self-efficacy	The respondent’s belief and confidence in their abilities and skills in responding to family violence
	Attitudinal-readiness	Having an open mind and considering family violence as a potential factor in the patient’s presentation; being non-judgmental towards the patient
	Emotional-readiness	Feeling comfortable discussing family violence and responding to disclosures
	Motivational-readiness	Belief that it is the role of health care providers to respond to family violence and help patients
Preparedness	Knowledge	Understanding of the nature of family violence, health care outcomes, legal aspects, and how family violence can be managed by health care workers
Preparedness	Communication	Belief in personal communication skills; understanding of how to communicate with a patient experiencing family violence

Initial Scale Development and Psychometric Testing

Instrumentation.

The pool of 48 items was delivered to participants with data used for psychometric evaluation. Items were both positively and negatively scored and were answered on a 7-point-Likert scale, anchored by strongly agree (7) and strongly disagree (1), with neutral (4) as the central point. Higher scores were intended to align with higher readiness (after reverse scoring negatively scored items).

In addition to scale items, students also input their age, sex, the name of their degree, and their current year level, and they were asked to estimate the number of hours of previous education and training they had received concerning family violence within or outside of their course. The term “family violence” was used instead of “partner abuse” when estimating hours of training as students would generally receive education on the entire concept of family violence, rather than just PA, and it would be difficult for respondents to estimate the hours dedicated to PA as opposed to other family violence presentations (such as child abuse and neglect or elder abuse).

Participants.

Participants were health care students undertaking a range of undergraduate and graduate degrees at a large Australian university. Recruitment took place between March and October 2018. Students were recruited in person immediately before or after normally scheduled lectures or workshops. Students were asked to respond to the items online using internet-capable devices. As an incentive participants were offered the opportunity to be entered into a random draw to win one of four $50 gift cards. Participation was voluntary and anonymous.

Statistical analyses.

SPSS version 24 was used for all analyses.

Tests of validity.

The dataset was randomly divided into two equal-sized groups, with group one being used for a PCA (Sample 1) and group two being used for a CFA (Sample 2). Given no a priori assumptions were established regarding the association between items and factors a PCA was undertaken over a traditional Exploratory Factor Analysis (DeCoster, 1998). As per Rubin (1976), where missing data were encountered surveys with items missing completely at random were retained, and questionnaires with non-random or greater than 10 missing items were excluded for the PCA. All data with missing values were subsequently deleted for the CFA. Reverse scored items were recoded so that higher scores on all items referred to preferred responses. The Kaiser–Meyer–Olkin (KMO) measure was used to define sampling adequacy, with a score above .70 considered adequate (Kaiser, 1974). Bartlett’s test for sphericity was estimated to test the null hypothesis that no variables are significantly correlated. The variable to case ratio was measured, with acceptance set at a minimum of 10:1 (Nunnally, 1994). The minimum r value for correlations between items was set at .3, which is considered the minimum correlational size to indicate an effect (Pett et al., 2003).

To investigate the latent variable structure generated from the PCA, a maximum likelihood CFA using AMOS (version 22.0; IBM, AMOS Development, Meadville, PA, United States) was performed to test the model-to-data fit. During the analysis, all factors were allowed to co-vary. Missing answers to any items in the SPSS dataset were replaced by estimated means (Meyers et al., 2016). Non-distributed data were also assessed to examine univariate outliers. Influential multivariate outliers were identified by calculation of the Mahalanobis distance and were accordingly deleted from the dataset. For a model to have an acceptable fit, a number of indices are required including the comparative fit index (CFI) ≥ .95, adjusted goodness-of-fit index (AGFI) ≥ .95, root mean square error of approximation (RMSEA) ≤ 0.06, Tucker–Lewis Index ≥ .95, and standardized root mean residual (SRMR) ≤ 0.06 (Hair et al., 1998).

Tests of reliability.

Cronbach’s alpha coefficient was used to assess internal consistency within identified subscales as a measure of reliability. Spearman’s rank coefficients were calculated between subscales and hours of training to determine the association between length of training and subscale results. Pearson’s correlation coefficient r was computed to assess the test–retest reliability based on the results of the 36 participants who were re-administrated the questionnaire after a two-week period.

Ethical Considerations

Ethical approval was granted by the respective University Human Research Ethics Committee.

Results

Demographics

A total of N = 944 responses were received from across eight different health care professions. Eighteen records were removed due to incomplete responses, leaving n = 926 records. The data were randomly split into two groups, with sample 1 (n = 464) used in the PCA and sample 2 (n = 462) in the CFA. Table 2 presents demographic data of participants.

Table 2.

Participant Demographics.

		n	%
	Total Records	926
Student group	Medicine Ug	62	6.7
	Nursing Ug	105	11.3
	Occupational Therapy Ug	71	7.7
	Occupational Therapy Me	26	2.8
	Paramedicine Ug	315	34.0
	Physiotherapy Ug	67	7.2
	Psychology M	27	2.9
	Psychology PhD or D	16	1.7
	Radiography	198	21.4
	Social Work M	39	4.2
Year level	1	304	32.83
	2	174	18.79
	3	260	28.08
	4	188	20.30
Sex	Male	247	26.7
Sex	Female	679	73.3
Hours of family violence education	None	264	28.5
	<1 hour	214	23.1
	1–2 hours	170	18.4
	2–3 hours	94	10.2
	3–4 hours	69	7.5
	4–5 hours	34	3.7
	5–6 hours	19	2.1
	6–7 hours	15	1.6
	7–8 hours	11	1.2
	>8 hours	36	3.9
Age range	18–20	374	40.4
	21–30	471	50.9
	31–40	46	5.0
	41–50	26	2.8
	51–60	6	.6
	Missing	3	.3

Item Reduction

PCA.

Using sample 1 (n = 464) a PCA was conducted on the 48 scale items using varimax rotation, as this assumes no correlation between the variables (Pett et al., 2003). The KMO measure verified the sampling adequacy for the analysis, the result was a KMO = .936, described as “marvelous” according to Kaiser, and above the acceptable limit to conduct factor analysis (Kaiser, 1974). Bartlett’s test of sphericity indicated that correlations between items were sufficiently large for PCA, x²(465) = 7562.618, p < .001. Variable to case ratio was above 9:1 but below the ideal sample size of 10:1, though this is still eligible for PCA (Pett et al., 2003).

An initial analysis was run to obtain eigenvalues for each construct in the data. Four constructs had eigenvalues over Kaiser’s criterion of 1. The scree plot demonstrated inflections that would justify retaining four factors (Tabachnick & Fidell, 2007). A Parallel Analysis was performed which showed only three factors with eigenvalues exceeding the corresponding criterion values for a randomly generated data matrix of the same size (48 variables × 464 respondents; Ledesma & Valero-Mora, 2007). Given the large sample size, and the convergence of the scree plot, four factors were retained as these theoretically and practically made clinical sense. Seventeen items not loading on the four factors were removed from the final solution resulting in 31 items being retained. The four-factor solution explained a total of 55.67% of the variance, component contributions are presented in Table 3. The rotated solution demonstrated the presence of multi-dimensionality. Intercorrelations between items are presented in Appendix 1.

Table 3.

Principal Component Analysis Matrix (Sample 1).

Item	Self-efficacy	Emotional-readiness	Motivational-readiness	Partner Abuse Knowledge	H ²	r_it	Mean	SD
I know what questions to ask patients who I suspect might be experiencing partner abuse	.769	.197	–.049	.036	.635	.645	3.23	1.63
I understand the legislation relevant to my health care profession regarding partner abuse	.761	.029	–.031	–.091	.590	.538	3.27	1.68
I have been adequately trained to respond to patients experiencing partner abuse	.749	.106	.004	–.037	.573	.573	2.96	1.61
I am confident I know what to do if I am concerned about a patient’s safety	.738	.130	.015	–.070	.566	.594	4.20	1.57
I know how to respond if a patient discloses sexual abuse, sexual assault or rape to me	.725	.158	–.019	–.004	.550	.575	3.47	1.61
I am able to properly document partner abuse in my notes	.722	.211	.019	–.084	.573	.616	3.47	1.68
I am able to refer a patient experiencing partner abuse to the right services to get help and support	.709	.172	.021	.032	.534	.599	3.80	1.58
I have the skills to talk to patients about partner abuse	.708	.363	.114	.038	.647	.729	3.73	1.52
I can perform a safety risk assessment with a patient who is experiencing partner abuse	.704	.123	.033	–.129	.528	.555	3.58	1.68
I would be able to respond appropriately if a patient disclosed abuse in their relationship	.703	.328	.046	.053	.607	.689	4.32	1.53
I have a good understanding of the different presentations of partner abuse	.675	.151	.044	.075	.486	.584	4.09	1.59
I know what is meant by “forensic evidence collection” and how to do this appropriately within my health care profession	.616	.110	–.029	–.173	.422	.449	2.48	1.48
I am able to make effective validating statements to patients who disclose partner abuse	.587	.162	.188	.056	.410	.561	4.67	1.40
I know where to access help for myself if I feel stressed after talking to patients about partner abuse	.571	.214	.001	.160	.398	.531	4.73	1.75
If a patient disclosed partner abuse to me I would not know how to respond (R)	.567	.307	.093	.132	.441	.595	3.89	1.50
I am comfortable talking about partner abuse with elderly patients	.278	.868	.015	–.045	.834	.600	4.21	1.69
I am comfortable talking about partner abuse with lesbian, gay, bisexual, transgender and/or intersex patients (LGBTI) patients	.294	.863	.047	.033	.834	.642	4.29	1.69
I am comfortable talking about partner abuse with patients from different cultural backgrounds	.300	.849	.044	–.077	.819	.613	4.03	1.68
I am comfortable talking about partner abuse with male patients	.253	.827	.017	–.066	.752	.554	3.94	1.64
I am comfortable talking about partner abuse with female patients	.370	.821	.071	.075	.821	.689	4.47	1.59
The thought of having to ask patients about partner abuse makes me feel stressed (R)	.176	.593	–.048	.230	.438	.413	3.70	1.60
I will do my best to support patients experiencing partner abuse	.032	.088	.810	.156	.688	.291	6.56	0.72
Everyone has the right to live without violence or abuse	–.035	–.010	.780	.059	.613	.177	6.86	0.49
I can use active and empathic listening	.075	.061	.760	.197	.626	.319	6.48	0.75
It is every health care professional’s responsibility to help and support patients living with partner abuse	.131	.007	.635	.142	.440	.298	6.02	0.99
I am concerned about the welfare of patients	–.011	–.031	.607	.192	.407	.172	6.51	0.72
Partner abuse usually only happens to certain types of people (R)	–.077	.018	.225	.632	.456	.111	6.07	1.15
It is easy for most people to leave an abusive relationship (R)	.042	–.064	.078	.625	.402	.136	6.34	1.00
Leaving an abusive relationship means the abuse will stop (R)	–.059	.027	.254	.608	.438	.124	5.87	1.01
Most health care practitioners only rarely encounter patients experiencing partner abuse (R)	.087	.095	.044	.586	.362	.221	4.81	1.20
If a patient wants my help with partner abuse they will usually initiate the conversation (R)	–.093	.025	.147	.578	.365	.089	5.06	1.31
Eigenvalues	9.78	3.56	2.51	1.41
Explained Variance %	31.54	11.47	8.09	4.56
Cronbach’s Alpha	.93	.93	.77	.63
Cronbach’s Alpha (Overall Scale): .92

Note. h²: Communality; r_it: Corrected item-total correlations; SD: Standard deviation; (R): Signifies reverse coded item.

Factor 1 included items pertaining to Self-efficacy, which is the individual’s confidence in their abilities;

Factor 2 included items pertaining to Emotional-readiness, which is the individual’s comfort in discussing PA and responding to disclosures;

Factor 3 included items pertaining to Motivational-readiness, which is the individual’s belief that responding to PA is the role of a health care provider; and

Factor 4 included items pertaining to PA knowledge, including understanding of the nature of PA, belief in common myths, and how a patient may present to a health care service.

CFA.

Using sample 2 data (n = 462) a CFA was conducted using the 4-factor 31-item scale identified in the PCA.

A maximum-likelihood CFA was performed to test the underlying latent variable structure of the scale. Standardized regression weights on each item were >.40, and modification indices were used to guide the iterative process. Following the PCA results the 31 items of the scale that loaded on the four factors were tested as a 4-factor model. These data did not fit the proposed 4-factor model. The x² value for the overall model was (709.01) = 366 (p < .0001) demonstrating a poor fit between the hypothesized model and model data (p > .05). Examination of other suggested model-fit indices suggested by Brown (2014) still demonstrated a poor model fit, with CFI = .926, AGFI = .854, RMSEA (90% confidence interval [CI]) = .0523 (.048, .057), SRMR = .097, and TLI = .920.

A review of the modification indices and a total of six iterations suggested allowing covariance between four items: item 2 “I know what questions to ask patients who I suspect might be experiencing partner abuse,” item 24 “I am concerned about the welfare of patients,” item 36 “If a patient wants my help with partner abuse they will usually initiate the conversation,” and item 48 “Everyone has the right to live without violence or abuse”. With these paths being constrained improved model fit was produced in each index except AGFI, which failed to reach .095: CFI = .955, AGFI = .892, RMSEA (90% CI) = .044 (.039, .049), and SRMR = .087 (refer to Figure 1). Based on these data the four items were removed with the resulting 4-factor 27-item version representing the most psychometrically sound version, which is named the READIness to encounter PA patients Scale (The READI Scale) and is the version which should be used in subsequent research.

Figure 1.

Notes. e = error.

Test–retest Analysis

Thirty-six students provided data for test–retest analysis. As can be seen from Table 4, using the final 27-item version, the scale demonstrated good test–retest reliability (r = .777, p < .0). The Self-efficacy and Emotional-readiness factors demonstrated strong positive test–retest reliability (r = .774 and r = .736 respectively, both significant at p < .01). Both the Motivational-readiness and PA knowledge subscales demonstrated acceptable positive test–retest reliability (r = .516 and r = .582 respectively, both significant at p < .01).

Table 4.

Test–retest Results.

Subscale	Baseline Mean	Baseline SD	Retest Mean	Retest SD	Mean Difference	Pearson’s Correlation
Self-efficacy	4.04	0.91	4.18	1.10	0.14	.774*
Emotional-readiness	4.33	1.27	4.58	1.25	0.26	.736*
Motivational-readiness	6.58	0.31	6.64	0.37	0.06	.516*
Partner abuse knowledge	5.84	0.59	5.82	0.56	–0.02	.582*
Total READI Score	4.90	0.61	5.01	0.66	0.12	.777*

Note. *p < .01.

Correlations

As can be seen from Table 5, using the final 27-item version, the total mean score had a strong correlation with Self-efficacy (r = .918, p < .01), and moderate correlation with hours of training (r = .502, p < .01) and Emotional-readiness (r = .788, p < .01). Hours of training had a moderate correlation with Self-efficacy (r = .525, p < .01). There was a moderate correlation between Self-efficacy and Emotional-readiness (r = .608, p < .01), and the Motivational-readiness and PA knowledge subscales had a low correlation (r = .349, p < .01).

Table 5.

Correlation Matrix (Sample 2).

	Hours	Self-efficacy	Emotional-readiness	Motivational-readiness	Partner Abuse Knowledge	Total Mean Score
Hours	1.000	.525**	.306**	.135**	.106*	0.502**
Self-efficacy		1.000	.608**	.214**	.017	0.918**
Emotional-readiness			1.000	.194**	–.020	0.788**
Motivational-readiness				1.000	.349**	0.390**
Partner abuse knowledge					1.000	0.209**
Total mean score						1.000

Notes. *Correlation is significant at the .05 level (2-tailed).

**Correlation is significant at the .01 level (2-tailed).

Discussion

This study aimed to create a new scale to measure the readiness of health care practitioners and students to encounter patients experiencing PA. The resultant scale, the READI Scale, comprises 27 items and 4 subscales and takes approximately 7 minutes to complete. Results indicate that the READI Scale has demonstrated good reliability and validity when delivered to a cohort of Australian health care students. Following will be a discussion of key results and implications for future practice.

Factor Analysis

While items were created which were intended to measure readiness over six constructs, the factor analysis delivered only a four-factor solution. Items created to measure the “Communication” and “Attitudinal-readiness” domains were not included in the final four-factor solution. Some of the items intended to measure these two domains did load into the Self-efficacy and Motivational-readiness subscales and were therefore included in those subscales. Neither of the Communication or Attitudinal-readiness domains was found when different items were developed and presented to a GP practitioner cohort as part of the psychometric validation of the GRIPS (Leung et al., 2017a), though again some items were retained as part of other subscales. These two domains were identified as important through research with practicing GPs (Leung et al., 2017b), and appear to be important aspects of an individual’s ability to respond to PA, and so it is unclear why they were not derived. Potentially health care students and practitioners view these constructs as important, but more aligned with self-efficacy and Motivational-readiness, which is why they were incorporated into these subscales.

The Self-efficacy subscale had a very strong correlation with overall readiness (r = .918), and in the PCA it accounted for 31.55% of the variance of the scale. Similar results were found by the authors of the GRIPS (Leung et al., 2017a). Likewise, the Emotional-readiness subscale was also high (r = .788, 11.47% of variance). Conversely, the subscales pertaining to Motivational-readiness and PA knowledge both had moderate correlations with overall readiness, and only accounted for a minor percentage of the scale. This indicates that students view their own self-efficacy and comfort in responding to PA as far greater components of their overall readiness than their knowledge or attitudes around their role. This may indicate that students feel readiness as less about the requirements of their role, and more about their own feelings and perceptions of their abilities. This gives further credence to research which has theorized that improving the readiness of practitioner cohorts should emphasize skills based training, where participants can practice their approach to the patient and gain feedback on their performance (Sawyer et al., 2016).

Furthermore, low self-efficacy has been identified as a major barrier for health care practitioners to discuss PA with patients (Sprague et al., 2012). An emphasis on improving self-efficacy in educational interventions, rather than knowledge or role clarity, maybe the key step in creating confident and competent future practitioners. When examining the correlation between hours of training and the identified subscales it can be seen that the largest correlation is with the Self-efficacy subscale, demonstrating that as the hours of training increases so does an individual’s self-efficacy. This indicates that educational interventions can have a positive impact on self-efficacy, and further research may benefit from examining the impact of different methods of educational intervention to determine the highest impact on overall self-efficacy.

Validity and reliability.

The developed scale has demonstrated content validity, as measured through feedback from family violence experts. Hours of training show good correlations with scale total score, as well as the subscales Self-efficacy and Emotional-readiness. It is expected that as hours of training increase, provided the training is effective, a student’s belief and confidence in their abilities (Self-efficacy and Emotional-readiness) should increase. Therefore, this demonstrates validity of the scale total score and these two subscales.

Correlations of hours of education with Motivational-readiness and Partner-abuse-knowledge, while significant, were weak, which may indicate poor validity of these subscales. However, while it would also be expected that both Motivational-readiness and partner-abuse-knowledge should increase with hours of education, both these subscales returned high results regardless of total hours of education (i.e., students with no education still reported high results). Therefore, there was little possibility of statistical improvement with increased hours of training. It is possible that the health care students studied began their training with a strong belief that they should be responding to PA as part of their role (Motivational-readiness), and felt they had good overall knowledge (partner-abuse-knowledge), and therefore any education was unlikely to significantly impact this, though this cannot be derived from the data we collected. This may suggest these subscales lack the capacity to differentiate between high and low scorers, or that the education had little impact on knowledge, or that students did not accurately report their knowledge. Further research with a wider group of health care students may provide greater insight into the validity of these two subscales.

The developed scale has shown good reliability. The internal consistency, as measured through Cronbach’s alpha, demonstrated close relations between the items within the scale and subscales. The test–retest analysis demonstrates that the scale shows good reliability over time, though further analysis over longer periods would assist in this measure. Therefore, we conclude that in this study, the scale has demonstrated good reliability.

Implications for future practice.

The developed scale, as created for use with all student and practicing health care practitioners, has shown good psychometric properties when delivered to a cohort of Australian health care students. This scale can be used by educators to ensure that cohorts have been appropriately educated and trained to respond to patients experiencing PA. Results from the delivery of this scale may be used to identify training needs, and the impact of educational interventions. Results may also inform the enhancement of educational interventions. The scale has not been validated for use with a practitioner cohort, and we suggest a psychometric evaluation should be performed prior to its use in such populations.

Limitations.

This study was limited in that it did not measure readiness in all health care student populations, and future research to validate the scale in a wider range of cohorts would be useful, in particular nutrition and dietetics, dentistry, and midwifery. This study did include practitioners, and should therefore it should undergo psychometric evaluation in practitioner populations before use outside of student cohorts. Likewise, the scale was developed for use in international populations, however, this validation only included students studying in Australia. The scale has not be validated against actual performance, and so it is not yet clear if readiness, as measured through this scale, will be an accurate measure of actual performance on a task.

Conclusions

This research has created a new 27 item scale, named the READI Scale, which can measure the readiness of health care cohorts to encounter patients experiencing PA. The scale has shown good psychometric properties with a cohort of Australian health care students and is underpinned by a sound conceptual basis for both the concept of readiness, and the health care sector’s expected response to PA. Results demonstrate that four factors makeup readiness in student cohorts and that self-efficacy is a major factor, indicating that educational interventions that emphasize improvements in self-efficacy may lead to the greatest gains in readiness. The scale can be used to measure the readiness of a cohort, and may be useful as a pre and post measure to determine the effect of an intervention. The READI Scale will allow evaluation of PA educational interventions in the health care sector, which may assist with building an effective health care workforce capable of early identification of PA, leading to increased referrals and a reduction in the overall duration and impact of PA in society.

Supplemental Material

Supplemental material for this article is available online It is also available at https://www.acu.edu.au/about-acu/faculties-directorates-and-staff/faculty-of-health-sciences/school-of-nursing-midwifery-and-paramedicine/research/readiscale

Supplemental Material for A New Scale of Readiness for Health Care Students to Encounter Partner Abuse by Simon Sawyer, Glenn Melvin, Angela Williams, Brett Williams, in Journal of Interpersonal Violence

Footnotes

Authors’ Note

Glenn Melvin is also affiliated with Deakin University, Burwood, Victoria, Australia. The READI Scale is free to access and use, please email Dr Simon Sawyer (simon.sawyer@acu.edu.au) for a copy of the READI Scale.

Declaration of Conflicting Interests

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Simon Sawyer

Author Biographies

Simon Sawyer, PhD, is an Advanced Life Support Paramedic and Lecturer of Paramedicine at Australian Catholic University in Melbourne Australia.

Glenn Melvin, PhD, is from School of Psychology, Deakin University.

Angela Williams, MBBS, is from Victorian Institute of Forensic Medicine, Monash University.

Brett Williams, PhD, is from Department of Paramedicine, Monash University.

References

Black

M. C.

(2011). Intimate partner violence and adverse health consequences implications for clinicians. American Journal of Lifestyle Medicine , 5(5), 428–439.

Bravo

R. G.

, Lygidakis

, Feder

, Reuter

R. A.

, & Vögele

(2019). Family Violence Curricula in Europe (FAVICUE): A cross-sectional descriptive study protocol. BMJ Open , 9(2), e024519.

Brown

T. A.

(2014). Confirmatory factor analysis for applied research . Guilford Publications.

Connor

P. D.

, Nouer

S. S.

, Mackey

S. T. N.

, Tipton

N. G.

, & Lloyd

A. K.

(2011). Psychometric properties of an intimate partner violence tool for health care students. Journal of Interpersonal Violence , 26(5), 1012–1035.

DeCoster

(1998). Overview of factor analysis . http://www.stat-help.com/factor.pdf

García-Moreno

(2013). Global and regional estimates of violence against women: Prevalence and health effects of intimate partner violence and non-partner sexual violence . World Health Organization.

García-Moreno

, Hegarty

, d’Oliveira, A. F. L., Koziol-McLain

, Colombini

, & Feder

(2015). The health-systems response to violence against women. The Lancet , 385(9977), 1567–1579.

Hair

, Anderson

, Tatham

, & Black

W. C.

(1998). Multivariate data analysis with readings . Macmillan Publishing Company.

Kaiser

H. F.

(1974). An index of factorial simplicity. Psychometrika , 39(1), 31–36.

10.

Ledesma

R. D.

, & Valero-Mora

(2007). Determining the number of factors to retain in EFA: An easy-to-use computer program for carrying out parallel analysis. Practical Assessment, Research & Evaluation , 12(2), 1–11.

11.

Leung

T. P. Y.

, Bryant

, Phillips

, & Hegarty

(2017a). GPs’ perceived readiness to identify and respond to intimate partner abuse: Development and preliminary validation of a multidimensional scale. Australian and New Zealand Journal of Public Health , 41(5), 512–517.

12.

Maiuro

R. D.

, Vitaliano

P. P.

, Sugg

N. K.

, Thompson

D. C.

, Rivara

F. P.

, & Thompson

R. S.

(2000). Development of a health care provider survey for domestic violence: Psychometric properties. American Journal of Preventive Medicine , 19(4), 245–252.

13.

Meyers

L. S.

, Gamst

, & Guarino

A. J.

(2016). Applied multivariate research: Design and interpretation . SAGE Publications.

14.

Nunnally

J. C.

(1994). Psychometric theory . McGraw-Hill.

15.

Pett

M. A.

, Lackey

N. R.

, & Sullivan

J. J.

(2003). Making sense of factor analysis: The use of factor analysis for instrument development in health care research . SAGE Publications.

16.

Leung

T. P. Y.

, Phillips

, Bryant

, & Hegarty

(2017b). How family doctors perceived their “readiness” and “preparedness” to identify and respond to intimate partner abuse: A qualitative study. Family Practice .

17.

Rubin

D. B.

(1976). Inference and missing data. Biometrika , 63(3), 581–592.

18.

Sawyer

, Coles

, Williams

, & Williams

(2016). A systematic review of intimate partner violence educational interventions delivered to allied health care practitioners. Medical Education , 50(11), 1107–1121.

19.

Sawyer

, Coles

, Williams

, & Williams

(2019). The psychometric properties of an intimate partner violence education outcome measure. Nurse Education Today , 76, 148–153.

20.

Short

L. M.

, Alpert

, Harris

J. M.

, & Surprenant

Z. J.

(2006). A tool for measuring physician readiness to manage intimate partner violence. American Journal of Preventive Medicine , 30(2), 173–180.

21.

Sprague

, Madden

, Simunovic

, Godin

, Pham

N. K.

, Bhandari

, & Goslings

(2012). Barriers to screening for intimate partner violence. Women & Health , 52(6), 587–605.

22.

Tabachnick

, & Fidell

(2007). Using multivariate statistics . Pearson Education.

23.

World Health Organization (WHO). (2013). Responding to intimate partner violence and sexual violence against women: WHO clinical and policy guidelines . https://www.who.int/reproductivehealth/publications/violence/9789241548595/en/

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB