Skip to main content
Download PDF

Development and validation of the Health Activation Scale for Children (HAS-C): an important intermediate outcome measure for health promotion initiatives

BMC Health Services Research volume 24, Article number: 1120 (2024) Cite this article

Abstract

Background

Valid and reliable measures for assessing health activation in school-aged children are currently lacking. This study aimed to develop a scale to measure health activation and evaluate its psychometric properties among English-speaking primary school children in Singapore.

Methods

The development of the Health Activation Scale for Children (HAS-C) involved an extensive literature review, expert consultations, cognitive interviews with primary school children, and thorough discussions for dimension and item refinement. A cross-sectional study was conducted with 597 children aged 8 to 12 years, recruited from four mainstream primary schools, comprising 50.1% boys and 64.8% Chinese students. The potential scale, along with other measures, was independently completed by the children. Descriptive statistics were provided for individual scale items. Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were performed to assess factorial validity. Convergent validity was examined by correlating the scale scores with scores of health confidence and self-efficacy measures. Hypothesis-testing validity was evaluated by assessing the scale’s correlation with self-reported health behaviours, including daily consumption of vegetables and fruits, time spent on sedentary activities and physical activities. Internal consistency reliability was measured using Cronbach’s alpha.

Results

EFA suggested a three-factor structure for the 12-item HAS-C, which was subsequently confirmed by CFA with a good model fit. All three HAS-C dimension scores demonstrated moderate correlations (rho = 0.34–0.52) with health confidence and self-efficacy measures, indicating good convergent validity. They were positively correlated with more vegetable and fruit intakes, more time spent on exercises, and negatively correlated with time spent on sedentary activities, supporting hypothesis-testing validity. Internal consistency reliability for individual HAS-C dimensions was generally acceptable, with Cronbach’s alpha values of 0.70 or above.

Conclusion

The 12-item multi-dimensional HAS-C exhibited good validity and reliability, making it a valuable tool for assessing health activation in primary school-aged children.

Peer Review reports

Background

Existing evidence underscores the potential long-term negative impact of poor dietary habits [1], insufficient physical activity [2], and excessive sedentary behaviours on adult health [3]. Recognising the pivotal phase of childhood in shaping lifelong behaviours with positive implications for long-term health outcomes [4], efforts have emerged to foster healthy lifestyles and behaviours among children and adolescents. Four key themes have gained prominence in this regard: physical activity, healthy eating, screen time management, and adequate sleep [5, 6]. However, adopting these health-promoting behaviours is intricately influenced by various factors, including beliefs, attitudes, knowledge, skills, social influence, environmental factors, and perceived barriers [7]. Understanding an individual’s knowledge, confidence, and activation level - essentially, their willingness and ability to take charge of their own health - is essential for effective health promotion.

The concept of activation, introduced by Judith Hibbard in 2004, encapsulates an individual’s knowledge, skills, and confidence in managing their health and health care [8]. It assesses whether individuals comprehend their role in maintaining their health and possess the willingness and self-efficacy to take proactive actions. Extensively used in earlier literature to evaluate adult patients’ engagement in health management, health activation has consistently been found to correlate with better adherence to treatment plans, improved self-management of chronic conditions, and healthier lifestyle choices [8, 9]. Despite its demonstrated importance in adults, health activation in children remains relatively under-explored. Hence, clearly defining and further developing the health activation concept for children is essential due to their unique developmental stages and cognitive capacities.

The primary school years represent a pivotal developmental state that offers a unique and time-sensitive opportunity for children to begin developing autonomy and self-management skills, albeit to a lesser extent compared to adolescents and adults [10, 11]. Prior studies have shown that these children possess considerable knowledge about health, illness, and disease risks [12,13,14]. They tend to be health-conscious, hold positive attitudes toward health and health behaviour adoption [14], and begin to learn health self-management, drawing from personal experiences and gaining confidence in decision-making [15]. While not to the same extent as adults, primary school-aged children do apply their health knowledge to make daily choices, such as dietary decisions, physical activity participation, and sleep habits, which can impact their overall health and well-being [16, 17].

Health activation is a critical concept in health promotion, mediating the effect of health literacy on the adoption of healthy behaviours [18]. Evidence suggests that individuals with higher health literacy are more likely to engage in healthy behaviours due to greater activation and confidence in managing their health [19]. Given the ongoing health education programmes in primary schools, it becomes crucial not only to assess the effectiveness of these programmes in increasing health literacy but also to understand their impacts on students’ health activation – a key factor influencing the adoption of healthy behaviours.

Despite the importance of health activation, there is a notable gap in the availability of instruments specifically designed for primary school-aged children. Existing instruments, such as the Patient Activation Measure [20, 21], the Patient Health Engagement Scale [22], the Consumer Health Activation Index [23], and the Altarum Consumer Engagement Measure [24, 25], primarily measure patient activation among adults in the context of chronic disease management. These instruments are designed for situations where patients need to actively participate in decision-making, self-manage their condition, and collaborate with their healthcare providers. However, applying these adult-focused, disease-management-related measures to primary school-aged children is inappropriate, as this age group generally experiences fewer health issues, has limited autonomy in health decisions, and exhibits lower self-efficacy in health management compared to adults. Consequently, developing a contextualised instrument of health activation for primary school children is crucial to gain valuable insights into their readiness for engaging in healthy behaviours.

De Civita and colleagues [26] have underscored the importance of developing assessment instruments specifically validated for young children, emphasising the need to adapt items to their developmental stage, emerging self-concept, cognitive capacity, and emotional awareness to ensure reliability and validity. Given the unique characteristics and developmental needs of primary school-aged children, there is a pressing need for tailored instruments to measure health activation in this population. Such contextualised instruments can provide valuable insights into how health education programmes impact children’s readiness to engage in and take charge of their health, informing strategies to promote lifelong healthy behaviours.

Recognising the pivotal role of health activation in fostering healthy behaviours and the lack of appropriate measures tailored for children in health promotion contexts, we undertook this study. Our objectives were twofold: firstly, to develop a measure specifically designed to evaluate health activation toward a healthy lifestyle and behaviours among primary school children in Singapore; and secondly, to examine the psychometric properties of this newly developed measure, including its factorial validity, convergent and hypothesis-testing validity, and internal consistency reliability.

Materials and methods

The development of the Health Activation Scale for Children (HAS-C) followed the scale development protocol proposed by Boateng and colleagues [27] and took place in three phases: dimension identification and item development, scale development and factor extraction, scale evaluation and validation.

Phase 1: dimension identification and item development

Step 1: dimension identification and initial item development

We conducted a literature review to achieve the following objectives: (1) identify various published definitions of health activation and health engagement and their underlying constructs, including those not specific to children; (2) determine the potential dimensions or constructs of health activation; and (3) identify existing tools measuring these specific dimensions or constructs. Other than “health activation” and “health engagement”, other terms related to “locus of control”, “self-efficacy for healthy diet”, “self-efficacy for physical activities”, “perceived health knowledge”, and “motivation” were also used for literature search. Any relevant existing measures specific to children were identified.

Based on an in-depth review of the identified literature, we proposed an operational definition of health activation relevant to primary school students: “a child’s health beliefs, knowledge, confidence and intention to adopt and/or maintain healthy lifestyles and healthy behaviours.”

To operationalise this definition and its constructs, we generated a list of potential health activation dimensions based on a thorough review of the scientific literature, existing health activation and engagement measures, and dimension-specific measures. Six theoretical dimensions emerged: (1) perceived knowledge, (2) confidence, (3) health beliefs, (4) motivation / intention for action, (5) self-efficacy, and (6) internal locus of control.

Initial items were generated by referring to existing tools that measure the above-mentioned dimensions or constructs. Through an interactive process involving multiple rounds of discussions and refinements, the preliminary HAS-C (version S1) was developed, comprising 55 items spanning the aforementioned six dimensions, with 3 to 16 items proposed for each dimension.

Step 2: content validity

A set of documents were sent to identified external experts for their assessment. These documents included the assessment form for step 1 HAS-C dimensions with descriptions (Supplementary Table S1a), the assessment form for step 1 HAS-C items (Supplementary Table S1b), and instructions. The panel comprised seven specialists from various related fields: one each in education, speech pathology, and communication, two in health / developmental psychology / behavioural medicine, and two in family medicine and nutrition. They were tasked with evaluating whether each potential dimension was crucial for the HAS-C and proposed additional relevant dimensions. Furthermore, experts assessed the relevance and importance of each item in measuring health activation, assigned each item to appropriate dimension, suggested wording revisions, and provided additional comments.

Subsequently, the research team compiled the experts’ feedback. A dimension was retained if 80% or more of experts affirmed its status as an essential HAS-C component. Items underwent rigorous content validity assessment based on the following criteria:

  1. 1)

    Items with less than 75% expert agreement on relevance to health activation were excluded.

  2. 2)

    The Content Validity Index (CVI) was calculated by dividing the number of experts rating an item as “3-relevant” or “4-very relevant” by the total number of experts. Items with a CVI of 0.75 or higher were retained [28].

  3. 3)

    The Importance Index (II) was computed by dividing the number of experts who rated an item as “3-important” or “4-very important” by the total number of experts. Items with an II of 0.75 or higher were retained (see Supplementary Table S2 for CVI and II values).

  4. 4)

    The study team reviewed experts’ additional comments and suggestions, revising item wording to enhance clarity. After rounds of intensive discussions, 20 items measuring four dimensions of health activation were retained, namely “perceived knowledge”, “health beliefs”, “behavioural confidence”, and “intention for action” in the step 2 HAS-C (Supplementary Table S3). Each dimension contained five items.

The items were then structured and formatted into a questionnaire. This included a proposed 4-point Likert rating scale, an option to indicate difficulty in understanding or responding, and detailed administration instructions, creating the step 2 HAS-C for use in cognitive interviews.

Phase 2: scale development and factor extraction

Step 3: pre-testing the step 2 HAS-C via cognitive interviews

To test if students correctly interpreted the intended meaning of the 20 items in the step 2 HAS-C, cognitive interviews were conducted with a convenience sample of 24 students. These students were capable of independent reading and communicating in English and were attending primary 3, 4, or 5 at any of the Ministry of Education (MOE) mainstream schools in Singapore.

Each participant was asked to independently respond to the 20-item HAS-C implemented through FormSG, a secured platform that allows public officers to create and fill digital forms for classified and sensitive data collection, replacing the use of paper forms. They were encouraged to mark items that they found challenging to understand and / or answer. Following survey completion, a cognitive interviewing session was conducted using a pre-developed interview guide to assess their comprehension of specific terms used, the interpretation of items within the step 2 HAS-C, and their views about the four response options. If a participant indicated any item as difficult to understand or answer, further questions were prompted to uncover the underlying reasons for their difficulty.

The cognitive interviewing sessions were recorded, and audio files were transcribed into text. Responses related to specific terms and items were analysed to determine whether they had been correctly interpreted, assess if there were any potential ambiguities, and evaluate the suitability of the survey response categories. Each item was reviewed and discussed further to decide whether the item should be retained, deleted, or modified and how it should be modified. At the end of the process, four items remained unchanged, 18 items were reworded, and two item were added (of which, HAS-C16 was added for validity checking only), resulting in 22 items in the step 3 HAS-C (see Table 1), with four dimensions “perceived knowledge” (4 items), “health beliefs” (5 items), “behavioural confidence” (6 items), and “intention for action” (6 items). The instruction and the four response options (“Strongly Agree”, “Agree”, “Disagree”, and “Strongly Disagree”) remained the same as the step 2 HAS-C used for cognitive interviews.

Step 4: Survey administration and sample size

Setting and participant recruitment for HAS-C validation

This is a cross-sectional survey involving Primary 3, 4, or 5 school students from four selected MOE mainstream primary schools participating in the Living Well @ School pilot project. This initiative, a collaborative effort between a public healthcare institution and selected schools, aims to cultivate healthy lifestyle habits in youths to support their transition into healthy adults. Students were recruited if they met the following eligibility criteria:

  1. 1)

    They were attending primary 3, 4, or 5 at these schools,

  2. 2)

    Their parents had consented for their participation in the study,

  3. 3)

    They gave assent to take participate in the study, and

  4. 4)

    They were able to complete the survey independently

The research team distributed the invitation letter and study information sheet to the parents of identified students through Parent Gateway, a platform used by MOE schools for parental communication, with the assistance of the schools. This communication, inclusive of the research team’s contact details, was dispatched at least two weeks prior to the scheduled survey sessions at each school. Additionally, a reminder was sent one week prior to the survey date.

Parents were encouraged to reach out to the research team with any questions or concerns regarding the study. Those expressing willingness to support their child’s participation were guided to provide consent via Parent Gateway. Subsequently, school representatives coordinated sessions within the curriculum timetable of each class to facilitate the research team’s introduction of the study to eligible students, obtaining their assent, and conducting the survey within individual classrooms during regular hours. In instances where parental consent was not obtained or students did not provide assent, teachers provided supervision and engaged with these students. The recruitment sessions took place between mid-August 2023 and mid-October 2023 across the four participating schools.

The survey was conducted immediately after each student gave their written assent. Each student was assigned a unique study ID which was pre-printed on the Child Assent forms. The students who returned fully signed assent forms were instructed to scan the survey Quick Response (QR) code, enter their study ID in the survey form, and complete the online survey independently, either electronically through FormSG or via the traditional pen-and-paper method in a classroom setting. Research team members provided instructions and assisted with technical difficulties.

In addition to the 22-item step 3 HAS-C (re-ordered), the survey collected basic demographic information, including age, gender, ethnicity, and primary school grade. Data for other measures intended for validity testing (described under Step 8) were also gathered. The full survey questionnaire is available in Supplementary File 1.

Sample size

The minimum necessary sample size for exploratory factor analysis (EFA) is typically determined by factors such as the number of variables/items, the number of factors, the number of variables/items per factor, and the size of the communalities [29]. A commonly recommended minimum sample size for EFA is at least 10 participants per item, with 15 participants per item being ideal; a larger sample size is generally considered more desirable [30, 31]. The necessary sample size recommended for confirmatory factor analysis (CFA) is 300 participants for the population model [32]. Hence, the minimum sample size should be 520 participants.

We received 597 valid responses at the end of the survey period. Hence, the number of responses were sufficient for the research purpose. As it is recommended that EFA and CFA should be conducted using two separate datasets, the study dataset was randomly split into two datasets. One dataset (n = 299) was used for EFA and internal consistency test and another one (n = 298) was used for CFA.

Step 5: item reduction and factor extraction

The EFA with Promax rotation, an oblique rotation method, was conducted to reduce items and extract factors for the HAS-C (excluding HAS-C16). The determination of the optimal number of factors was guided by considering eigenvalues greater than 1, scree plots, as well as conceptual considerations related to the underlying constructs, while a factor loading threshold of 0.40 was employed to assess the significance of item loadings [33]. Whenever there was any change in the number of items included for exploration, the EFA was reconducted and item loading coefficients were reproduced. Once the optimum number of factors was determined, we fixed the number of factors for the final EFA.

Phase 3: scale evaluation and validation

Step 6: test of dimensionality

Following the identification of the factor structure of HAS-C through the final EFA, CFA was conducted to confirm the factor structure using the separate dataset. The items loaded onto the same factor in EFA were allowed to load onto the same latent factor in the model, with the variance of each latent factor being set at 1.0. Standardised factor loading values were generated for each item. The model’s goodness of fit was evaluated using the following criteria: the chi-square to degrees of freedom ratio (χ2/df), with a cutoff of 3 or lower; the Root Mean Square Error of Approximation (RMSEA), with a cutoff of 0.06 or lower; the Comparative Fit Index (CFI), with the cutoff of 0.95 or higher; and the Tucker-Lewis Fit Index (TLI) with the cutoff of 0.95 or higher [34].

Step 7: test of reliability

The internal consistency reliability of the HAS-C was evaluated using Cronbach’s alpha coefficient, calculated on the same dataset after conducting the EFA. A Cronbach’s alpha coefficient of 0.70 or higher was considered acceptable, as recommended by the literature [35].

Step 8: tests of construct validity

Given the cross-sectional nature of this study and the absence of a “gold standard” measure for health activation in children, the focus was on examining the construct validity of the HAS-C using the measures described below.

Health confidence score for convergent validity

The Health Confidence Score (HCS) is a unidimensional tool which was developed to assess a child’s confidence in looking after their own health [36]. The original HCS comprises four items: “I know enough about my heath”, “I can look after my heath”, “I can get the right help if I need it”, and “I am involved in decisions about me”. Each item is accompanied by four response options: “Strongly Agree”, “Agree”, “Neutral”, and “Disagree”. Based on feedback from a pilot study, an additional option, “Strongly Disagree”, was added to capture a wider range of participant opinions. The mean score of the four items was calculated, with a higher score representing higher health confidence. The validation study demonstrated that the HCS had good readability (reading age 8), construct validity (single dimension), hypothesis-testing validity (moderate correlation with the “My Health Confidence” measure), and internal consistency reliability (Cronbach’s alpha = 0.82) [36]. The unidimensionality of the HCS was corroborated in the present study, although the Cronbach’s alpha was slightly lower at 0.68.

Healthy eating and physical activity self-efficacy questionnaire for children for convergent validity

The Healthy Eating and Physical Activity Self-Efficacy Questionnaire for Children (HEPASEQ-C) was a 9-item questionnaire developed by engaging children through focus group discussions to measure children’s self-efficacy related to healthy eating (e.g., “I will eat healthy food even when my friends eat food that is not”) and physical activity (e.g., “I will be physically active even when my friends choose to sit still and hang out”) [37]. Each item has three response options: “1 = There is no way I can do this.”, “2 = This could be hard for me.”, and “3 = I believe I can do this.” The total score of the HEPASEQ-C was calculated by summing the scores of its nine items, with the possible range being 9 to 27, where a higher score indicated greater self-efficacy. The validation study demonstrated that the HEPASEQ-C exhibited a unidimensional factor structure, accompanied by satisfactory content validity (evidenced by CVI: 0.80 -1.00) and acceptable internal consistency reliability with a Cronbach’s alpha coefficient of 0.75 [37].

Health behaviours for hypothesis-testing validity

The survey collected information on students’ health behaviours, which included daily vegetable and fruit intake (categorised as 0, 1 serving, and 2 or more servings); the number of hours spent per day on homework, private tuition, screen time, and exercise, respectively (options: None, 1 h or less, 2, 3, 4, and 5 or more hours); and the number of days engaging in moderate-to-vigorous physical activity (MVPA) for at least 60 min in the past week, with response options of 0, 1, 2, 3, and 4 or more days per week. These items were culturally adapted from the 2017 Questionnaire of the New South Wales School Students Health Behaviours Survey [38].

For data analysis, the hours spent on homework, private tuition, screen time were summed up and then categorised into three groups based on tertiles. The hours spent on exercise were divided into three groups: 0 h,1 h, and ≥ 2 h and number of days engaged in MVPA was categorised into two groups: ≤ 2 days and > 2 days.

To evaluate the convergent validity of the HAS-C, we calculated dimension scores using an unweighted approach. This involved summing the individual item scores for each dimension and then dividing by the number of items within that dimension for all participants. The HAS-C total score was calculated by dividing the sum of all remaining HAS-C items in the CFA confirmed model by the number of remaining items. To assess the association between HAS-C dimension and total scores with the HCS and HEPASEQ-C scores, Spearman’s rank correlation coefficients (rho) were calculated, given the ordinal nature of the response options. Our hypothesis was that there would be at least moderate (rho > 0.30) and positive correlations between HAS-C scores and HCS as well as HEPASEQ-C scores.

To assess the hypothesis-testing validity of the HAS-C, we calculated the mean scores of HAS-C dimensions for the categories of individual health behaviours, assuming each item has an equal weight to the concept of respective domains. Subsequently, we conducted Kruskal-Wallis tests to examine the significance of the correlation between HAS-C scores and individual health behaviours among all participants. We hypothesised that individuals with higher HAS-C scores eat more vegetables, engaged in more frequent physical activity, and spend less time in sedentary activities.

Additionally, participants’ characteristics were described using frequency and percentage. The item-level descriptive analyses for the 22 items in the step 3 HAS-C (HAS-C16 was excluded from other analyses) were conducted for all 597 responses. Mean, standard deviation (SD), median, and the distribution of responses were presented for individual items. Floor and ceiling effects were considered present if more than 15% of the participants achieved the lowest (“Strongly Disagree”) or highest (“Strongly Agree”) possible score [39].

Results

The age of the participants in the study ranged from 8 to 12 years old, with the majority being 9-year-old (37.0%), 10-year-old (34.8%) and 11-year-old (23.1%). Approximately half of the participants (50.1%) were boys. More than two-thirds (64.8%) identified as Chinese, 14.7% were Malays, 9.2% were Indians, and the remaining 11.2% were of other ethnicity groups. Approximately one-third of the participants (34.3%) were from primary 3, 38.7% were from primary 4, and the remaining 27.0% were from primary 5.

Item-level descriptive statistics

Table 1 presents the descriptive statistics of individuals items of the HAS-C. The distribution of the four response options (“Strongly Disagree”, “Agree”, “Disagree, and “Strongly Disagree”) for each item showed that they were heavily skewed. The “Strongly Agree” option was the most frequently selected response (range: 21.9–83.9%), indicating ceiling effects for all items. The “Strongly Disagree” was the least selected response for all items (range: 0–6.9%), indicating no floor effects (Table 1).

Table 1 Item-level descriptive statistics
Full size table

Item reduction and factor extraction

The EFA of the 21-item HAS-C (excluding HAS-C16) yielded four factors with eigenvalues greater than 1 for each factor (see Supplementary Table S4). However, the item arrangement based on the primary factor loadings did not result in a coherent and easily interpretable structure consistent with the theoretical dimensions of health activation. Three factors loaded onto more than one factor with a factor loading above 0.30 and some items, particularly those proposed for “perceived knowledge” and “behavioural confidence” loaded onto other factors, making interpretation challenging.

Hence, we conducted a thorough iterative review of item content and wording, identifying and removing items that did not align with the theoretical dimensions, were not theoretically essential to the construct, or had lower factor loadings. Given that perceived knowledge reflects an individual’s beliefs about their knowledge and confidence in understanding of specific health topics, it was likely that some items proposed for “perceived knowledge” dimension loaded onto factors related to “health beliefs” or “behavioural confidence” dimensions. Thus, three-factor structure excluding “perceived knowledge” as a separate dimension was theoretically acceptable.

This iterative process resulted in the removal of nine items. A final EFA with Promax rotation on the remaining 12 items, specifying three factors to be extracted, yielded a three-factor solution with four unique items loading onto each factor. The factor loadings of individual items for the three-factor solution are presented in Table 2.

Table 2 Component loadings for the individual items of the three rotated components
Full size table

Internal consistency reliability

The internal consistency reliability for “intention for action” and “health beliefs” dimensions was found to be acceptable, with a Cronbach’s alpha coefficients of 0.753 and 0.732, respectively. The “behavioural confidence” dimension showed slightly lower reliability, with a Cronbach’s alpha of 0.698, which is close to the acceptable threshold of 0.700. Overall, the 12-item HAS-C demonstrated acceptable internal consistency reliability, with Cronbach’s alpha coefficient of 0.844.

Dimensionality examined using confirmatory factor analysis

The CFA confirmed the three-factor structure of the 12-item HAS-C. Figure 1 presents the standardised factor loading for each item onto their respective latent construct (“health beliefs”, “behavioural confidence”, and “intention for action”). Each item under these constructs had a factor loading higher than 0.35. The χ2/df was 1.583 (χ2 (46) = 72.8, p = 0.007) and the goodness fit indexes (RMSEA = 0.044, CFI = 0.968, TLI = 0.954) suggested a good fit between the three-factor model and the observed data.

Fig. 1
figure 1

The path diagram for the three-factor CFA model for the 12-item HAS-C: standardised estimates. All p-values for loadings were < 0.001

Full size image

Construct validity

Convergent validity

The three HAS-C dimension scores, as well as the total HAS-C score, were moderately correlated with both the HCS and the HEPASEQ-C (Table 3).

Table 3 The Spearman correlation coefficients (rho) between HAS-C dimension scores and HCS and HEPASEQ-C scores (N = 597)
Full size table

Hypothesis-testing validity

As detailed in Table 4, the “health beliefs” dimension was positively correlated with daily vegetable servings (p < 0.001) and number of days engaging in MVPA for at least 60 min per day in the past week (p = 0.031). The “behavioural confidence” and “intention for action” dimensions exhibited positive correlations with daily servings of vegetables and fruits, hours spent on exercise, and days of MVPA (all p < 0.001). Additionally, these two dimensions were negatively correlated with hours spent on sedentary activities (including homework, tuitions, and screen time) after school (both p < 0.05).

Table 4 The mean (SD) of HAS-C total and dimensional scores by categories of various health behaviours
Full size table

Discussion

Discussion on findings

This study developed and validated the HAS-C, tailored for primary school children. The EFA and CFA supported a three-factor structure encompassing “health beliefs”, “behavioural confidence”, and “intention for action”. This multidimensional structure aligns with established constructs of health/patient activation in adults, such as knowledge, beliefs, confidence/self-efficacy, and readiness to take action [20, 40].

The determination of the optimal number of factors was guided by both statistical methods (e.g., eigenvalues) and conceptual/theoretical expectations related to health activation. Although the initial EFA conducted on the 21 items suggested a four-factor solution, this structure lacked coherence and interpretability in the context of the conceptual dimensions of health activation. Notably, “perceived knowledge” did not emerge as a separate factor, with items loading onto factors meant for “health beliefs” or “behavioural confidence”. Given that perceived knowledge reflects confidence in understanding or knowledge [41], a three-factor solution was deemed more appropriate. Consequently, we specified the extraction of three factors in the final EFA, despite the eigenvalue for the third factor being slightly below the conventional threshold of 1. CFA corroborated the three-factor structure of the 12-item HAS-C, demonstrating good model fit indices and providing further evidence of factorial validity.

The three-factor HAS-C demonstrated good construct validity, evidenced by moderate correlations with conceptually related measures like the HCS and the HEPASEQ-C. Those associations provide convergent validity, supporting the notion that health activation shares commonalities with health confidence and self-efficacy [8]. Furthermore, higher HAS-C scores corresponded to healthier behaviours, including increased vegetables and fruits intake, physical activity levels, and reduced sedentary time, establishing hypothesis-testing validity.

Despite no observed floor effects, ceiling effects were present across all HAS-C items and two dimensions. Ceiling effects, where scores cluster at or near the highest possible value on a scale [42], are commonly reported in self-reported measures of health confidence [36], self-efficacy [37], and activation [21], potentially due to social desirability bias or lack of challenging items [43]. The convenience sampling from schools participating in a health promotion programme might have contributed to these effects. Administering the HAS-C to a broader, more representative sample and revising the measure by adding more challenging items or response options could mitigate this issue.

Limitations

The study is subject to several limitations. Firstly, convenience sampling from schools involved in the Living Well @ School pilot project may have introduced selection bias, as participants’ responses could be influenced by the pilot project or their parents’ higher health consciousness. Recruiting from a more diverse set of schools, including those not involved in the pilot, could enhance generalisability [44]. Secondly, the lack of an established gold standard measure of health activation for primary school-aged children limits our capacity to determine criterion validity of the HAS-C. Thirdly, while the HCS and HEPASEQ-C were used for convergent validity testing, these measures were developed and validated in non-Singaporean populations, potentially limiting their applicability. Lastly, the cross-sectional design precludes assessment of predictive validity and causal relationships between HAS-C scores and future health behaviours and outcomes.

Practice implications and future studies

Understanding children’s health beliefs, confidence, and readiness to take an active role in managing their health is crucial for effective health promotion. The concept of health activation, adapted from the adult-focused patient activation concept, is highly relevant in this context. A measure like the HAS-C can serve as a screening tool to identify children needing targeted interventions and as an intermediary outcome for health promotion initiatives, including health education, weight management programmes, and school health activation programmes.

To bolster the credibility of the HAS-C, future research should prioritise a detailed investigation of its psychometric properties. This includes establishing discriminant and predictive validity, test-retest reliability, responsiveness or sensitivity to change, and appropriate scoring approaches and thresholds [45, 46]. If ceiling effect persists, revising the measure by adding more challenging items or an additional response option (e.g., “Totally Agree”) may be considered, followed by further validation analyses.

Conclusions

The development of the HAS-C addresses a gap in assessing health activation in primary school-aged children. The three-factor HAS-C has demonstrated good factorial validity, convergent validity, hypothesis-testing validity, and satisfactory internal consistency reliability. These findings collectively support the three-factor HAS-C being a reliable and validated tool for educators and researchers to better understand primary school students’ health activation levels, identify areas where children may require support or intervention, customise programmes, and examine programmes’ intermediate effects.

Data availability

The dataset used and/or analysed during the current study is available from the corresponding author on reasonable request.

Abbreviations

CFA:

Confirmatory factor analysis

CFI:

Comparative Fit Index

CVI:

Content Validity Index

EFA:

Exploratory factor analysis

HAS-C:

Health Activation Scale for Children

HCS:

Health Confidence Score

HEPASEQ-C:

Healthy Eating and Physical Activity Self-Efficacy Questionnaire for Children

II:

Importance Index

MOE:

Ministry of Education

MVPA:

Moderate-to-vigorous physical activity

QR code:

Quick Response code

RMSEA:

Root Mean Square Error of Approximation

SD:

Standard deviation

References

  1. Mikkilä V, Räsänen L, Raitakari OT, Pietinen P, Viikari J. Consistent dietary patterns identified from childhood to adulthood: the cardiovascular risk in Young finns Study. Br J Nutr. 2005;93:923–31.

    Article  PubMed  Google Scholar 

  2. Telama R. Tracking of physical activity from childhood to adulthood: a review. Obes Facts. 2009;2:187–95.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Thorp AA, Owen N, Neuhaus M, Dunstan DW. Sedentary behaviors and subsequent health outcomes in adults a systematic review of longitudinal studies, 1996–2011. Am J Prev Med. 2011;41:207–15.

    Article  PubMed  Google Scholar 

  4. Hilliard ME, Riekert KA, Ockene JK, Pbert L. The handbook of health behavior change. Fifth. Springer Publishing Company; 2018.

  5. Chaput J-P, Gray CE, Poitras VJ, Carson V, Gruber R, Olds T, et al. Systematic review of the relationships between sleep duration and health indicators in school-aged children and youth. Appl Physiol Nutr Metab. 2016;41(6 Suppl 3):S266–282.

    Article  PubMed  Google Scholar 

  6. Public Health England. How healthy behaviour supports children’s wellbeing. England: Public Health England; 2013.

    Google Scholar 

  7. Glanz K, Rimer BK, Viswanath K, Vish, editors. Health behavior: Theory, research, and practice. 5th edition. Hoboken, NJ, US: Jossey-Bass/Wiley; 2015.

  8. Hibbard JH, Greene J. What the evidence shows about patient activation: better health outcomes and care experiences; fewer data on costs. Health Aff. 2013;32:207–14.

    Article  Google Scholar 

  9. Hibbard JH, Mahoney ER, Stockard J, Tusler M. Development and testing of a short form of the patient activation measure. Health Serv Res. 2005;40(6 Pt 1):1918–30.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Buford TA. Transfer of asthma management responsibility from parents to their school-age children. J Pediatr Nurs. 2004;19:3–12.

    Article  PubMed  Google Scholar 

  11. Ramdzan SN, Khoo EM, Liew SM, Cunningham S, Kendall M, Sukri N, et al. How young children learn independent asthma self-management: a qualitative study in Malaysia. Arch Dis Child. 2020;105:819–24.

    Article  PubMed  Google Scholar 

  12. Goldman SL, Whitney-Saltiel D, Granger J, Rodin J. Children’s representations of everyday aspects of health and illness. J Pediatr Psychol. 1991;16:747–66.

    Article  CAS  PubMed  Google Scholar 

  13. Natapoff JN. A developmental analysis of children’s ideas of health. Health Educ Q. 1982;9:34–45.

    Article  Google Scholar 

  14. Piko BF, Bak J. Children’s perceptions of health and illness: images and lay concepts in preadolescence. Health Educ Res. 2006;21:643–53.

    Article  PubMed  Google Scholar 

  15. Catarino M, Charepe Z, Festas C. Promotion of Self-Management of Chronic Disease in Children and teenagers. Scoping Rev Healthc (Basel). 2021;9:1642.

    Google Scholar 

  16. Borzekowski DLG. Considering children and health literacy: a theoretical approach. Pediatrics. 2009;124 Supplement:S282–8.

    Article  Google Scholar 

  17. DeWalt DA, Hink A. Health literacy and child health outcomes: a systematic review of the literature. Pediatrics. 2009;124(Suppl 3):S265–274.

    Article  PubMed  Google Scholar 

  18. Osborn CY, Paasche-Orlow MK, Bailey SC, Wolf MS. The mechanisms linking health literacy to behavior and health status. Am J Health Behav. 2011;35:118–28.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Hibbard JH. Patient activation and the use of information to support informed health decisions. Patient Educ Couns. 2017;100:5–7.

    Article  PubMed  Google Scholar 

  20. Hibbard JH, Stockard J, Mahoney ER, Tusler M. Development of the patient activation measure (PAM): conceptualizing and measuring activation in patients and consumers. Health Serv Res. 2004;39(4 Pt 1):1005–26.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Ge L, Kaur P, Yap CW, Heng BH. Psychometric properties of the patient activation measure in community-dwelling adults in Singapore. Inquiry. 2022;59:469580221100781.

    PubMed  Google Scholar 

  22. Graffigna G, Barello S. Spotlight on the Patient Health Engagement model (PHE model): a psychosocial theory to understand people’s meaningful engagement in their own health care. Patient Prefer Adherence. 2018;12:1261–71.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Wolf MS, Smith SG, Pandit AU, Condon DM, Curtis LM, Griffith J, et al. Development and validation of the Consumer Health Activation Index. Med Decis Mak. 2018;38:334–43.

    Article  Google Scholar 

  24. Castellon-Lopez Y, Skrine Jeffers K, Duru OK, Moreno G, Moin T, Grotts J, et al. Psychometric properties of the Altarum Consumer Engagement (ACE) measure of activation in patients with prediabetes. J Gen Intern Med. 2020;35:3159–65.

    Article  PubMed  PubMed Central  Google Scholar 

  25. Ge L, Yap CW, Kaur P, Ong R, Heng BH. Psychometric evaluation of the 8-item Altarum Consumer Engagement (ACE) Measure™ in community-dwelling adults in Singapore. BMC Health Serv Res. 2021;21:1347.

    Article  PubMed  PubMed Central  Google Scholar 

  26. De Civita M, Regier D, Alamgir AH, Anis AH, Fitzgerald MJ, Marra CA. Evaluating health-related quality-of-life studies in paediatric populations: some conceptual, methodological and developmental considerations and recent applications. PharmacoEconomics. 2005;23:659–85.

    Article  PubMed  Google Scholar 

  27. Boateng GO, Neilands TB, Frongillo EA, Melgar-Quiñonez HR, Young SL. Best practices for developing and Validating Scales for Health, Social, and behavioral research: a primer. Front Public Health. 2018;6:149.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Polit DF, Beck CT, Owen SV. Is the CVI an acceptable indicator of content validity? Appraisal and recommendations. Res Nurs Health. 2007;30:459–67.

    Article  PubMed  Google Scholar 

  29. Mundfrom DJ, Shaw DG, Ke TL. Minimum sample size recommendations for conducting factor analyses. Int J Test. 2005;5:159–68.

    Article  Google Scholar 

  30. Clark LA, Watson D. Constructing validity: basic issues in objective scale development. Psychol Assess. 1995;7:309–19.

    Article  Google Scholar 

  31. DeVellis RF. Scale development: theory and applications. 2nd edition. Newbury Park: Sage Publications; 2003.

  32. Myers N. Sample size and power estimates for a confirmatory factor analytic model in exercise and sport: a Monte Carlo approach. rqes. 2011;82.

  33. Costello AB, Osborne JW. Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis. Practical Assess Res Evaluation. 2005;10:9.

    Google Scholar 

  34. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equation Modeling: Multidisciplinary J. 1999;6:1–55.

    Article  Google Scholar 

  35. Bland JM, Altman DG. Statistics notes: Cronbach’s alpha. BMJ. 1997;314:572.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Benson T, Potts HWW, Bark P, Bowman C. Development and initial testing of a Health confidence score (HCS). BMJ Open Qual. 2019;8:e000411.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Lassetter J, Macintosh C, Williams M, Driessnack M, Ray G, Wisco J. Psychometric testing of the healthy eating and physical activity self-efficacy questionnaire and the healthy eating and physical activity behavior recall questionnaire for children. J Spec Pediatr Nurs. 2018;23:e12207.

    Article  PubMed  PubMed Central  Google Scholar 

  38. NSW Health. The New South Wales School Students Health Behaviours Survey – 2017 Questionnaire. NSW Health. 2017. https://www.health.nsw.gov.au/surveys/student/Documents/student-health-survey-2017-quest.pdf. Accessed 22 Jan 2024.

  39. McHorney CA, Tarlov AR. Individual-patient monitoring in clinical practice: are available health status surveys adequate? Qual Life Res. 1995;4:293–307.

    Article  CAS  PubMed  Google Scholar 

  40. Greene J, Hibbard JH. Why does patient activation matter? An examination of the relationships between patient activation and health-related outcomes. J Gen Intern Med. 2012;27:520–6.

    Article  PubMed  Google Scholar 

  41. Pinkosova Z, McGeown WJ, Moshfeghi Y. Moderating effects of self-perceived knowledge in a relevance assessment task: an EEG study. Computers Hum Behav Rep. 2023;11:100295.

    Article  Google Scholar 

  42. Feng Y, Hancock GR, Harring JR. Latent growth models with floors, ceilings, and random knots. Multivar Behav Res. 2019;54:751–70.

    Article  Google Scholar 

  43. Badejo MA, Ramtin S, Rossano A, Ring D, Koenig K, Crijns TJ. Does adjusting for social desirability reduce ceiling effects and increase variation of patient-reported experience measures? J Patient Experience. 2022;9:23743735221079144.

    Article  Google Scholar 

  44. Bornstein MH, Jager J, Putnick DL. Sampling in developmental science: situations, shortcomings, solutions, and standards. Dev Rev. 2013;33:357–70.

    Article  PubMed  PubMed Central  Google Scholar 

  45. Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol. 2008;61:102–9.

    Article  PubMed  Google Scholar 

  46. Streiner DL, Norman GR, Cairney J. Health measurement scales: a practical guide to their development and use. 5th ed. New York, NY, US: Oxford University Press; 2015.

    Book  Google Scholar 

Download references

Acknowledgements

We acknowledge the inputs given by Ms Sabrina Ho Wei Ling, Ms Palvinder Kaur D/O Satwant Singh, Ms Natalie Lim Wai Harn, and Dr Loke Hsi-Yen during the planning and item development phase, the valuable assessment and feedback on the dimensions and items provided by the experts, namely, Assistant Professor Andrew Yee Z H, Mr Brian So Jian Wei, Dr Goh Tze Jui, Associate Professor Konstadina Griva, Mdm Lua Wee Suan, Dr Mary Chong Foong-Fong, and Dr Tan Shu Yun, as well as the project support by Mr Derick Ng Kok Shern and Mr Umar Abdul Hamid and the four selected primary schools, and the cooperation and support of the participants and their parents.

Funding

This study did not receive any funding from any external body.

Author information

Authors and Affiliations

  1. Health Services & Outcomes Research, National Healthcare Group, Annex@National Skin Centre Level 4, 1 Mandalay Road, Singapore, 308205, Singapore

    Lixia Ge & Joseph Molina

  2. Community Care Division, Population Health Campus, National Healthcare Group, #03-08 Nexus @ One-north, 3 Fusionopolis Link, Singapore, 138543, Singapore

    Ramakrishnan Karthigayan, Hui Ting Foo, Marcus Tang, Rochelle Chua & Chin Fung Ong

Authors
  1. Lixia Ge
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Joseph Molina
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Ramakrishnan Karthigayan
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Hui Ting Foo
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Marcus Tang
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Rochelle Chua
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Chin Fung Ong
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

L.G. conceptualised the manuscript, analysed and interpreted the data, and was a major contributor in writing the manuscript and preparing the tables and figures. C.F.O. and J.M. conceptualised, reviewed, and edited the manuscript. L.G., R.K., H.T.F., M.T., R.C., and C.F.O. were involved in assent taking and data collection at schools. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Lixia Ge.

Ethics declarations

Ethics approval and consent to participate

This study received approval from the institutional ethics committee of the National Healthcare Group Domain Specific Review Board, Singapore (Reference Number: 2021/01112) and was conducted in strict adherence to the principles outlined in the Declaration of Helsinki. Additionally, the study obtained approval and support from the Ministry of Education and the respective schools involved. Prior to engaging individual students, parents were contacted to provide detailed information about the study. Consent from parents was secured before approaching the students. The individual students, whose parents had consented, were fully informed about the study’s purpose, procedures, the voluntary nature of participation, and the confidentiality of the data collected. All the participants provided their written consent prior to the commencement of data collection.

Competing interests

The authors declare no competing interests.

Consent for publication

Not applicable.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below are the links to the electronic supplementary materials.

Supplementary Material 1

Supplementary Material 2

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ge, L., Molina, J., Karthigayan, R. et al. Development and validation of the Health Activation Scale for Children (HAS-C): an important intermediate outcome measure for health promotion initiatives. BMC Health Serv Res 24, 1120 (2024). https://doi.org/10.1186/s12913-024-11526-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12913-024-11526-7

Keywords

BMC Health Services Research

ISSN: 1472-6963

Contact us