From concept to measurement: a scoping review and development of a Culture of Care questionnaire for laboratory animal institutions

1 Introduction

A Culture of Care in laboratory animal science is commonly described in policy and guidance documents as a shared commitment to ethical responsibility, animal welfare and the wellbeing of staff (RSPCA & LASA, 2010; European Union, 2018). The concept was formally introduced in European legislation through Directive 2010/63/EU, which calls on institutions to foster a ‘climate of care’ for animals used in scientific procedures (European Union, 2010). The RSPCA and LASA, as well as the EU Working Document on Animal Welfare Bodies (RSPCA & LASA, 2010; European Union, 2018) further define this culture as one in which all staff, from researchers to animal care personnel, are encouraged and enabled to take responsibility for animal welfare and ethical standards, beyond minimum legal compliance.

Since the introduction of Directive 2010/63/EU, the Culture of Care has become an increasingly recognized topic within laboratory animal research, including initiatives such as the International Culture of Care Network, which was established in 2016 (Norecopa, 2025).

The Culture of Care builds on the field of animal welfare by emphasizing the role of organizational culture and staff wellbeing. Over the past decade, attention has shifted from focusing mainly on animal care, to also recognizing the wellbeing of the people involved in animal research (Ameli and Krämer, 2024a, 2024b; Morahan et al., 2024). Poor mental health among animal care staff, including symptoms of compassion fatigue (emotional exhaustion, due to ongoing exposure to animal suffering, ethical tension and a high workload) can reduce vigilance, empathy and professional judgment, potentially affecting animal welfare (Morahan et al., 2024). Therefore, a strong Culture of Care is defined by both having high welfare standards for animals, while also ensuring the wellbeing of the staff who take care of these animals.

Although the core principles of a Culture of Care are increasingly discussed in policy, guidance and scientific literature, these sources do not provide a clearly operationalized and literature-informed instrument for staff-level assessment within laboratory animal institutions. Existing publications often provide conceptual descriptions on the Culture of Care (e.g. Amarasekara et al., 2022; Robinson et al., 2019), reflective discussions or institutional case examples (e.g. Bertelsen and Øvlisen, 2021), and some studies have reported local survey approaches. However, the literature remains fragmented in how Culture of Care is translated into measurable domains and questionnaire items. The main gap is therefore the lack of a clearly described measurement approach that synthesizes the scientific literature and translates recurring themes into a practical questionnaire. In human healthcare, the NHS Culture of Care Barometer has been developed and validated to systematically assess organizational culture (Rafferty et al., 2017; Maassen et al., 2024). This shows the potential for developing similar tools within laboratory animal research.

Where Morahan et al. (2024) focused specifically on mental health in a scoping review of the literature on the Culture of Care in laboratory animal medicine, this present study maps the broader concept of the Culture of Care, including its relation to both animal welfare and staff wellbeing.

This study consisted of two phases. First, a scoping review was done to identify recurring themes, assessment approaches, and extractable survey content related to the Culture of Care in the scientific literature on laboratory animal research. A scoping review was chosen as the concept is heterogeneous and not consistently defined. Second, based on the insights from this scoping review and on existing instruments from human healthcare, we developed a questionnaire to assess key aspects of the Culture of Care in laboratory animal science.

2 Materials and methods

2.1 Scoping review

The scoping review followed the PRISMA-ScR guidelines (Tricco et al., 2018). The completed checklist can be found in Appendix A. The flow diagram of the scoping review can be found in Figure 1. A protocol was prepared before performing the review and can be found in Appendix B. The review question was structured using the Population, Concept, Context (PCC) framework, with the population defined as professionals, staff or organizational structures involved in laboratory animal science, the concept as the Culture of Care and closely related organizational culture concepts related to animal welfare and staff wellbeing and the context as laboratory animal settings.

2.1.1 Search strategy

The search strategy was designed to identify literature on the concept of a Culture of Care within the context of laboratory animal science. Multiple synonyms and related terms were included for both Culture of Care and the animal research context. Publications referring to the Culture of Care as well as closely related concepts were considered eligible when they addressed organizational culture, values, responsibilities or practices related to animal welfare and/or staff wellbeing in laboratory animal science. Boolean operators (AND/OR) were applied (see Table 1) and adapted for three databases: PubMed, Embase and Web of Science (see Appendix B for the full search string per database).

Table 1. Boolean search string used to identify literature on the Culture of Care in laboratory animal research.

For database-specific adaptations, see Appendix B.

The search was done on April 10, 2025. References were downloaded from PubMed, Embase and Web of Science. All citations were exported to EndNote 22 (Clarivate Analytics, Philadelphia, PA, USA) reference management software. Duplicate records were identified and removed using the DeDupEndnote tool (dedupendnote.nl, April 2025). A total of 155 unique records were uploaded to Rayyan (Rayyan Systems Incl, Doha, Qatar, May 2025) for title and abstract screening. Title and abstract screening was done independently and blinded by the first and last author using the eligibility criteria shown in Table 2. At full-text stage, exclusion reasons were categorized in broad groups, as reflected in the PRISMA flow diagram. Discrepancies were resolved through discussion until consensus was reached. Full-text screening was done by the first author. In cases of uncertainty or borderline eligibility, the second reviewer was consulted before a final inclusion decision was made.

Table 2. Eligibility criteria for inclusion and exclusion of sources at all screening stages of the scoping review.

Reference lists of included studies were screened for additional relevant publications, but this did not result in any additional studies meeting the inclusion criteria.

2.1.2 Data extraction

Although the protocol specified inclusion of gray literature as sources of evidence, in the final study grey literature was not included as part of the dataset. Instead, relevant policy and guidance documents (e.g. EU guidelines) were used to support interpretation and structuring of the themes identified in the included studies. This is reported as a deviation from the protocol.

Data from the included articles were extracted into an Excel sheet. Because the scoping review functioned as an input phase for questionnaire development, the data extraction form was iteratively refined after full text screening. The full data extraction table, including study characteristics, theme coding, and illustrative quotes for each included study, is provided in Appendix D.

Data were charted using four domains: (1) study characteristics (e.g. year, country, type of publication), (2) conceptualization of Culture of Care (definitions, terminology, and relation to human and animal wellbeing), (3) assessment approaches (e.g. whether CoC was assessed and by which method), and (4) extractable content for questionnaire development (e.g. presence of themes and survey items).

During the full-text screening, recurring themes describing the Culture of Care were identified across the included studies. Based on these patterns, an initial set of themes was developed. These themes were refined using relevant policy and guidance documents (e.g. EU legal frameworks). The resulting nine themes were: (1) compliance to organizational and legal standards, (2) beyond compliance: a proactive attitude, (3) strong and caring leadership, (4) sufficient resources and conditions, (5) empathy and respect towards colleagues and animals, (6) shared responsibility and accountability, (7) transparency and open communication, (8) supportive work environment and team collaboration, and (9) sharing knowledge and continuous reviewing. These themes were used as an analytical framework to systematically code the included studies (presence/absence) and to extract illustrative quotes.

In addition, for survey-based studies, a basic quality assessment was performed using predefined criteria covering sampling method, response rate, survey validation, reported reliability, and bias control strategies. These criteria and their scoring system (0–2) were simplified based on the SURGE and CHERRIES guidelines. The full data extraction framework, including all variables and the quality assessment criteria, is provided in Appendix C.

2.1.3 Data analysis

Extracted data were analyzed descriptively using R version 4.4.1. The variables from the extraction sheet were imported and cleaned (e.g. factor conversion, numeric formatting). For each of the nine Culture of Care themes, presence/absence was coded and relevant quotes were extracted. Study characteristics (e.g. year, country, literature type) and assessment-related variables (e.g. survey validation, type of method) were summarized using frequency tables. Quality items for survey studies were scored (0-2) and tabulated. Analyses were done using the dplyr, ggplot2, and knitr packages. No formal statistical tests were done, as the goal of the analysis was to describe the scope and content of the included literature rather than to test hypotheses.

In line with the aim of the scoping review, the analysis focused on mapping (1) recurring thematic domains used to describe the Culture of Care, (2) approaches used to assess these domains, and (3) extractable survey content relevant for questionnaire development.

2.2 Development of the questionnaire

Based on the scoping review, a questionnaire was developed using the nine thematic domains as an analytical framework. Each of the nine key themes identified in the scoping review was represented by multiple items. Four questions were taken directly from the NHS Culture of Care Barometer (Rafferty et al., 2017), using both the original English version and the validated Dutch translation published by Maassen et al. (2024). The complete 30-item Culture of Care Barometer questionnaire is available in Appendix E.

The first draft questionnaire consisted of 50 positively worded statements, written in the first person and present tense. This structure was chosen to make the items easier to understand, reduce cognitive load and to create consistency in the survey. All statements were rated on a 5-point Likert scale (Robinson, 2014), ranging from ‘Strongly disagree’ to ‘Strongly agree’. The idea behind this approach is that clear, positively phrased items help respondents give more accurate answers, by reducing the risk of satisficing or confusion (Lenzner et al., 2010). First-person, present-tense formulations encourage reflection on personal experiences. Fully labeled Likert scales are associated with higher response quality and interpretability (Menold et al., 2018). This style of wording the questions and using the Likert scale, is consistent with previously validated instruments, such as the NHS Culture of Care Barometer (Rafferty et al., 2017).

The initial draft questionnaire (50 items) was reviewed by staff of the Animal Welfare Body Utrecht to evaluate item relevance, clarity, and translational quality. Each item was rated on a 5-point Likert scale (‘highly irrelevant’ to ‘highly relevant’). Open-text fields were provided both at item level, allowing participants to give feedback on each individual question, and at the end of the questionnaire for general feedback on the questionnaire as a whole. Of the invited participants (n = 11), 8 completed the survey partially or fully. This expert review was used to assess content validity and face validity of the questionnaire. Quantitative ratings were used to identify items with lower perceived relevance, while open-text feedback provided detailed input on wording clarity, conceptual overlap, and translational accuracy.

Following this first review round, the results were discussed with three experts. Items were evaluated based on a combination of relevance scores and qualitative feedback. Items with lower relevance, redundancy or unclear wording were removed or merged, and several items were rephrased to improve clarity and consistency. In addition, translations were edited based on expert feedback. This resulted in a final set of 30 items.

A test-retest study was done to evaluate the stability of the questionnaire. The study started with a targeted invitation to 50 laboratory animal coordinators at Utrecht University and UMC Utrecht. These coordinators hold diverse roles, such as researchers, facility managers and laboratory animal technicians, and are responsible for implementing legal and institutional animal welfare policies within their departments. Of this group, 15 participants completed the first questionnaire (test), and 5 of them also completed the retest two weeks later.

To increase the sample size, an open call was posted in October 2025 on MyDALAS, the internal platform of the Dutch Association for Laboratory Animal Science (DALAS), which has approximately 700 active members across various Dutch research and education institutions. In the same week, an open call was posted on the International Culture of Care Network. In these calls, it was announced that a 30 euro gift card would be randomly awarded to one participant who completed both the test and retest. This was not offered to the initial group. Through MyDALAS and the International Culture of Care Network, 12 participants completed the test, and 11 of them also completed the retest.

In total, 16 participants completed both the test and the retest: 5 from the UU/UMCU group and 11 from the MyDALAS group and the International Culture of Care group. All participants received the same version of the questionnaire. Participants completed the same questionnaire twice, with a two-week interval in between completion of the first survey and the invite for the second survey. As the questionnaire was distributed through semi-open channels, the total number of individuals who viewed the invitation is unknown and response rates could not be calculated.

The test-retest survey was distributed via Qualtrics and optimized for both desktop and mobile devices. The average completion time was estimated at approximately 10 minutes. Participation was voluntary, and responses were collected confidentially without storing IP addresses or personal identifiers, apart from the email address. No demographic or role-specific data were collected in the test-retest study to reduce participant burden and preserve anonymity in this small sample. Participants all received the questions in the same order, grouped by theme. Ethical approval for the questionnaire was given September 4^th, 2025 by the Ethical Review Board of the Faculty of Social and Behavioral Sciences of Utrecht University and full privacy scan was reviewed and approved by the UU Privacy Officer. A Data Management Plan was reviewed by the UU data steward and is available on request.

2.2.1 Data analysis

Analysis of the data was done using R (version 4.4.1). The raw data originated from two separate Qualtrics exports (test and retest), which were merged into a single dataset. E-mail addresses were removed before analysis. The merged dataset included 30 Likert-scale items (Q18_1-Q18_30), one open-ended question and a unique login_id used to link participants across timepoints (1 = test, 2 = retest). Administrative variables from the Qualtrics export were removed manually before import.

The data were read using readxl and processed with janitor::clean_names() to standardize variable names. The first data row, which contained item labels, was removed. login_id was coerced to a character variable and timepoint to an integer (1 or 2). All Likert responses were converted to numeric scores (1-5) via a recoding function that standardized input to lowercase text before mapping to integers. A total score per participant was computed as the mean of the 30 items using rowMeans(…, na.rm = TRUE).

Internal consistency was estimated separately for test and retest using Cronbach’s alpha (psych::alpha), based on all 30 items. For test-retest reliability, only participants with both timepoints were included. The dataset was reshaped to wide format (pivot_wider()) to obtain paired values for each item (e.g., q18_1_t1, q18_1_t2). The stability of the total score was assessed using a two-way, absolute-agreement, single-measure intraclass correlation coefficient (ICC(2,1); irr::icc). The same ICC specification was applied to each of the 30 items individually. For each item, the number of complete pairs and the 95% confidence interval were reported.

To examine item distributions and potential ceiling or floor effects, standard deviations, score ranges, and the number of unique response categories were calculated for all items at both timepoints. In addition, the proportion of responses at the maximum scale point was computed to identify items with limited variance that could influence item-level ICCs.

The results (Cronbach’s alphas, ICCs) were exported using readr::write_csv.

3 Results

3.1 Results of scoping review

Of the 21 included articles, 6 focused on animal welfare, 4 on human welfare, and 11 explicitly addressed both. The relationship between animal welfare and human welfare was discussed in 12 articles. The articles were published between 2007 and 2024, with a median publication year of 2022. Studies were done across multiple geographical contexts, most frequently in the United Kingdom (n = 10), followed by the United States (n = 4) and Germany (n = 3). Several studies included multiple countries. A detailed overview of all included studies and their extracted data is provided in Appendix C.

Most publications described the Culture of Care at a conceptual level, while 11 articles included some form of assessment of the culture. 6 studies used survey-based approaches and 5 used qualitative methods such as interviews. 3 studies reported extractable survey items, resulting in a total of 118 items that informed questionnaire development.

Four survey-based studies (Bertelsen and Øvlisen, 2021; Fitzpatrick et al., 2023; Mason and Balch, 2024; Mehigan, 2020) were evaluated using predefined quality criteria (see Appendix C). Overall, methodological quality and reporting varied across studies. Survey designs ranged from more structured instrument development approaches to internally developed or practice-oriented questionnaires. For example, Fitzpatrick et al. (2023) developed a 12-item scale and evaluated its validity using principal component analysis, while Bertelsen and Øvlisen (2021) described pilot testing of a broader survey tool developed for use within an animal research setting. Mason and Balch (2024) reported the use of an anonymous internal wellbeing survey to inform local initiatives, but provided limited methodological detail on sampling, validation, and reliability. Mehigan (2020) described a questionnaire distributed through emails, forums, and mailing lists across three countries, but reported limited methodological detail beyond participant numbers, with no formal reporting of response rate, validation procedures, reliability statistics, or bias control measures. Across studies, formal reliability reporting remained limited, and bias control measures were generally modest.Based on the included articles, as well as the International Culture of Care Network (Norecopa, 2025) and EU guidelines (European Union, 2018, 2019), the following nine key topics were identified (Table 3). These themes represent an analytical framework used to structure recurring components described across the literature and formed the basis for questionnaire development (see Table 4).

Table 3. The nine identified key components of a Culture of Care and the number of included articles explicitly mentioning each component (n = 21).

Table 4. Final questionnaire items grouped by the nine Culture of Care themes and the literature sources informing each theme.

questions marked with * were directly adopted from Maassen et al. (2024).

3.1.1 Key part 1: compliance with organizational and legal standards

Included sources described compliance with legal and organizational standards as the basis of a Culture of Care. Although some authors argue that a true Culture of Care only exists when institutions go beyond compliance (Brown et al., 2018; Mehigan, 2020), legal and organizational standards are still essential. They define clear expectations and help prevent harm, both to animals and to staff.

Compliance includes following relevant legislation (in the European Union the EU Directive 2010/63/EU), institutional protocols and broader frameworks for ethical review and risk management. Included sources indicated that staff are expected to understand and apply core concepts such as the 3Rs (Replacement, Reduction, Refinement), be familiar with emergency procedures and know how to report welfare concerns.

Brown et al. (2018) state that ‘compliance with the relevant federal, state, and local regulatory requirements is just the starting point for a culture of care.’ Ameli and Krämer (2024a) identify the legal framework as one of the core components. von der Beck et al. (2024) emphasizes that ‘all legal requirements must be met,’ and the European Union (2010) mentions that institutions must understand and effect compliance. RSPCA & LASA (2010) argue for a shared set of values and standards that are ‘communicated, understood and implemented across all parts of the establishment.’ Mason and Balch (2024) include ‘compliance’ as one of the six building blocks of a Culture of Care.

Overall, compliance was described as an embedded component of daily practice within a Culture of Care.

3.1.2 Key part 2: beyond legal compliance: having a proactive attitude

Several sources described a Culture of Care as extending beyond legal compliance, emphasizing a proactive and reflective attitude towards continuous improvement. For instance, Mehigan (2020) notes that ‘a culture of care sets the stage for attitudes and behaviors that enhance animal welfare beyond the legal requirements.’ Similarly, Robinson et al. (2022) and Bertelsen and Øvlisen (2021) note that a Culture of Care involves a proactive and questioning attitude, not just passive adherence to the law. Amarasekara et al. (2022) specifically describe it as aiming ‘for more than just a culture of compliance.’ Brown et al. (2018) link this proactive mindset to the 3Rs, calling for active implementation by institutions and staff, not just theoretical awareness.

Across the included sources, this theme was characterized by continuous reflection, willingness to challenge existing practices and active efforts to improve animal welfare and working conditions.

3.1.3 Key part 3: strong and caring leadership

Leadership was frequently described as an important factor influencing the development and maintenance of a Culture of Care. Leaders influence norms, expectations and the climate of an organization. Included sources mention that leadership involves prioritizing both animal welfare and staff wellbeing as well as being approachable, knowledgeable and responsive to concerns of staff. Klein and Bayne (2007) state that leadership must provide ‘vision and direction’. Morahan et al. (2024) identify leadership in a literature review as one of the most frequently cited contributors to human wellbeing in laboratory animal science.

3.1.4 Key part 4: sufficient resources and conditions to do a proper job

Included sources described the availability of sufficient resources as an important condition for a Culture of Care. This includes time, staffing, facilities, training and financial resources. When these are lacking, even the most caring and professional staff will be unable to do their work properly (Williams, 2021). Several authors emphasize that providing adequate resources is an important practical part of a Culture of Care (Brown et al., 2018; Robinson et al., 2020). Bertelsen and Øvlisen (2021) distinguish between tangible (time, money) and intangible (competence, motivation) resources and mention how they are dependent of each other. Inadequate resources can directly lead to non-compliance (Klein and Bayne, 2007) and reduced animal welfare. Across sources, limited resources were associated with challenges in maintaining both compliance and animal welfare standards.

3.1.5 Key part 5: empathy and respect towards colleagues and animals

Many included sources emphasized empathy, compassion, and respect towards both animals and colleagues as central elements of a Culture of Care. As described in EU legislation, the intrinsic value of an animal should be respected (European Union, 2010). Brown et al. (2018) describe a culture that promotes ‘compassion and respect for laboratory animals and the people who work with them.’ Robinson et al. (2020) define a culture of care as ‘an organizational culture that supports and values caring and respectful behavior towards animals and co-workers.’ Williams (2021) adds that attentiveness is key: recognizing the needs of others, human or animal, is an important condition for effective care. Respect also includes openness to differing ethical perspectives (RSPCA & LASA, 2010). Empathy and respect are particularly relevant as compassion fatigue is a common issue among animal care staff. In their scoping review, Morahan et al. found that ‘the most common strategy suggested to mitigate WRS [work related stress] was acknowledgement and validation of LAP [laboratory animal personnel]s’ feelings and stress’ (Morahan et al., 2024). Overall, this theme focuses on recognition of both emotional and ethical dimensions of working with animals.

3.1.6 Key part 6: shared responsibility and accountability

Responsibility for animal welfare was described as distributed across roles within organizations. Everyone involved, whether researcher, technician, veterinarian, manager or supporting staff, has a role in ensuring a good quality of care. This does not mean responsibility becomes vague: roles should be clearly defined and all staff should understand what is expected of them and others (Hawkins and Bertelsen, 2019; Kitchenham, 2022).

Morahan et al. (2024) states: ‘all who care for, use, or produce animals for research, testing or teaching must assume responsibility for their well-being.’ Amarasekara et al. (2022) emphasize the importance of sharing responsibility for animal care. Kirk and Myelnikov (2022) describe responsibility within a Culture of Care as a shift: from individual responsibility of scientists towards a collective effort in daily practice. This collective approach can also be found in European guidance, which stresses the importance of shared responsibility without losing individual accountability (EU, 2018).

3.1.7 Key part 7: transparency and open communication

Open communication and transparency were frequently described as key components of a Culture of Care. It allows staff to raise concerns and communicate mistakes, and it can lead to pro-active improvement of animal welfare. Transparency includes not only internal communication between staff and leaders, but also openness to the public about the use of animals in science. Robinson et al. (2020) and Brown et al. (2018) both identify openness and transparency as core components of a Culture of Care. Amarasekara et al. (2022) stress the importance of reporting mechanisms and a no-blame culture, where incidents can be shared and learned from. von der Beck et al. (2024) mentions the need for ‘clear and transparent structures’ and ‘open and respectful communication with all parties.’ Compassion fatigue is common among staff working with laboratory animals (Morahan et al., 2024), and a supportive Culture of Care can help lower this risk by creating space for reflection, empathy and psychological safety.

3.1.8 Key part 8: supportive work environment and team collaboration

Several sources described a supportive and collaborative work environment as an important element of a Culture of Care. Williams (2021) notes that when staff feel seen and supported, they are better able to care for others, including animals. Robinson et al. (2020) and Gorman and Davies (2023) mention the need for emotional support systems, mentoring and recognition of staff contributions. Angela and Tremoleda (2019) note that reflection and support from colleagues and senior staff is especially important for junior staff. A supportive culture also includes psychological safety: a no-blame approach to mistakes and space for learning (von der Beck et al., 2024). Brown et al. (2018) suggest that programs recognizing excellence in care contribute to a Culture of Care.

3.1.9 Key part 9: sharing knowledge and continuous reviewing

Continuous learning, training and reflection were described as factors supporting a Culture of Care. Hawkins and Bertelsen (2019) and Brown et al. (2018) both point to continuous education as a key component. Amarasekara et al. (2022) stress the importance of stakeholder feedback and up-to-date knowledge. Included sources described the sharing of experiences, evaluation of practices, and ongoing training as contributing to continuous improvement within organizations.

3.2 Questionnaire development and expert review

The expert review provided structured input on the relevance and clarity of the initial item pool. The draft questionnaire consisted of 50 items derived from the scoping review and existing instruments, which were evaluated by staff of the Animal Welfare Body Utrecht (n = 11 invited; n = 8 respondents). Experts rated each item on relevance using a 5-point Likert scale (‘highly irrelevant’ to ‘highly relevant’). Across all items, mean relevance scores ranged from 3.8 to 4.8.

In addition to quantitative ratings, qualitative feedback was collected both per item and at the end of the questionnaire. Experts provided detailed comments on clarity, redundancy, specificity, and translation quality. Frequently mentioned issues included unclear or vague wording and overlap between items.

Translation-related feedback focused on improving consistency and nuance between the English and Dutch versions. For example, suggestions included replacing literal translations with more natural phrasing (e.g. ‘confident’ translated as ‘vol vertrouwen’ instead of ‘zeker’), clarifying terms such as ‘senior staff” and improving readability of longer sentences.

Based on quantitative scores and qualitative feedback, items with lower relevance scores, conceptual overlap or unclear wording were removed or revised. Additional refinements were made after discussion with three experts. This process resulted in a final set of 30 items.

The final questionnaire, including corresponding themes and source where applicable, is presented in Table 4.

3.3 Results test-retest

These nine themes formed the basis for the Culture of Care Barometer. After expert review and reduction to 30 items, the questionnaire was evaluated in a test–retest study.

3.3.1 Cronbach’s alpha

Internal consistency of the 30-item Likert-scale was very high at both timepoints, with Cronbach’s alpha coefficients of 0.917 (test) and 0.898 (retest). Commonly cited acceptable ranges fall between 0.70-0.95, although values above 0.90 may indicate some degree of item redundancy (Tavakol and Dennick, 2011). Overall, the instrument shows strong and stable internal consistency across the test-retest interval.

3.3.2 Intraclass correlation

The total-scale ICC(2,1) indicated moderately good test-retest reliability for the overall Culture of Care score (ICC = 0.73, 95% CI: 0.38-0.90; n = 16) (Koo and Li, 2016).

At the item level, ICC values showed substantial variability. Given the small retest sample (n=16), the wide confidence interval of the total ICC is expected. At item level, ICC values varied widely, ranging from poor (< 0.50) to good (0.75–0.90) agreement between the two timepoints. Confidence intervals for individual items were broad, again reflecting the small retest sample.

According to classical test theory, individual items contain more random error and therefore show lower temporal stability, whereas aggregated scores lead to higher reliability because random error cancels out across items (Zhang and Colvin, 2024). Thus, the more stable ICC for the total scale provides the most meaningful indication of the instrument’s reliability over time, whereas item-level ICCs should be interpreted cautiously.

In addition, for some items, the low ICC may not reflect true instability, but limited variance due to ceiling effects. For example, item 10 was scored almost exclusively as ‘4’ (agree) or ‘5’ (strongly agree) at both time points, with no responses of ‘1’ (strongly disagree) or ‘2’ (disagree). In such cases, between-subject variance is too low to detect meaningful test-retest. This emphasizes that item-level ICCs should be interpreted with caution in small samples.

4 Discussion

The scoping review showed that there is increasing attention for the Culture of Care within laboratory animal science. Across the included studies, the Culture of Care was primarily addressed at a conceptual level, although some studies described survey-based approaches to assess specific aspects or organisational practices. These approaches differed in scope and purpose and were often developed for use within specific institutional contexts, sometimes requiring local adaptation. This makes it difficult for institutions to evaluate their own practices in a systematic and comparable way.

Most included publications originated from the UK or European contexts, leaving uncertainty about the applicability in other regulatory or cultural settings. The nine topics identified across the literature provide a useful structure, but they are closely connected in practice. Elements such as transparency, leadership, resources, shared responsibility and proactive behavior influence each other and no single theme stands alone. Together, these themes provide an analytical framework for understanding how the Culture of Care is described across the literature and for structuring its assessment in practice. Although this framework was led by recurring themes in the included literature, its final structure was also informed by relevant policy and guidance documents, as described in the Methods.

Based on these findings, we developed and tested a 30-item questionnaire representing the nine themes. Internal consistency was very high at both timepoints, indicating that the items measure a coherent construct. The high Cronbach’s alphas suggest potential redundancy in some items. The test-retest results showed moderately good stability at the scale level, but substantial variability in item-level ICCs. This can be explained by ceiling effects, limited variance and the small sample of n = 16 participants who completed both timepoints.

It is important to consider that the test-retest data were collected in two separate rounds and populations. The first group consisted of laboratory animal coordinators from UU/UMCU, who were approached in September at the start of the academic year and received no gift card. The second group was recruited in October via MyDALAS and the International Culture of Care Network and informed that a €30 gift card would be randomly given to one participant completing both timepoints. This difference in timing and incentive may have influenced motivation, response quality and representativeness. In addition, the latter group had a more diverse professional background.

Response rates were low across both recruitment rounds, and the small number of participants completing both timepoints (n = 16) limits the reliability estimates. Additionally, as the total number of individuals reached through open calls was unknown, response rates could not be accurately determined, limiting assessment of representativeness.

Recruitment through a network that is already engaged with Culture of Care initiatives (i.e. Culture of Care Network) may have introduced selection bias, with participants potentially more positively inclined towards the concept. Combined with the small sample size and limited response rates, this likely contributed to ceiling effects and limited variability in responses.

Another consideration is the structure of the questionnaire itself. All items were positively worded and grouped by theme. Although this improves clarity for respondents and is based on literature, it may reduce response variability and make the underlying structure more transparent. Future studies could explore the effects of randomizing item order and including differently formulated items.

An important result of the scoping review was the lack of empirical assessments of the Culture of Care; few studies reported structured evaluation tools and the available approaches varied in scope, purpose, and level of development. The present questionnaire was developed to address this gap by translating recurring themes from the literature into measurable items. The instrument may support institutions in gathering structured feedback from staff and identifying areas for reflection and potential improvement. The questionnaire showed high internal consistency and moderately good test-retest reliability at the scale level, indicating that it captures a stable underlying construct. This study therefore presents a newly developed questionnaire with initial evidence of internal consistency and preliminary test-retest reliability. Further validation is required, including assessment of construct validity, factor structure, and performance in larger and more diverse samples. Future research could also explore its potential to detect differences between settings and its responsiveness to changes over time.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fetho.2026.1796584/full#supplementary-material

Completed PRISMA-ScR Checklist. Supplementary material at the source

Protocol for the scoping review on the Culture of Care within laboratory animal science. Supplementary material at the source

Data extraction variables for the scoping review. Supplementary material at the source

Data extraction table of included studies. Supplementary material at the source

Full questionnaire materials, including the 30-item Culture of Care questionnaire, Dutch translations, and additional demographic items for future use. Supplementary material at the source

Footnotes

Edited by: Cristiane Gonçalves Titto, University of São Paulo, Brazil

Reviewed by: Augusto Vitale, National Institute of Health (ISS), Italy

Heidi Morahan, The University of Sydney, Australia