Translation and Validity of the Chinese Version of the Hypoglycemia Confidence Scale in Adults with Type 2 Diabetes Using Both Basal and Prandial Insulins: A Psychometric Analysis with Cultural Adaptation

Abstract

Background:

Hypoglycemia confidence (HC) reflects a patient’s belief in their ability to avoid hypoglycemia-related adverse events and is crucial for optimizing glycemic management in insulin-treated type 2 diabetes mellitus. The aim of this study was to translate and validate the Chinese version of the Hypoglycemia Confidence Scale (HCS-C) in adults with type 2 diabetes (T2D) using both basal and prandial insulin administration (T2D-BP).

Methods:

A two-phase psychometric evaluation study was conducted from July to September 2024 at a tertiary hospital in eastern China. Phase 1 involved cross-cultural adaptation of the original nine-item HCS using forward-backward translation and an expert panel review. Phase 2 employed a cross-sectional design with 135 T2D-BP participants recruited from a tertiary hospital in China. The reliability, validity, and dimensionality of HCS-C were analyzed using confirmatory factor analysis (CFA), Cronbach’s α, intraclass correlation coefficients (ICC), and Spearman’s correlations. Known-group validity was assessed across educational/income levels and hypoglycemia frequency.

Results:

The Chinese version of the HCS retained nine items and demonstrated high content validity. CFA confirmed one-dimensionality (Chi-square/degree of freedom = 1.787, root mean square error of approximation = 0.077, comparative fit index = 0.982, incremental fit index = 0.982, Tucker–Lewis index = 0.973, and normed fit index = 0.961). Internal consistency was excellent (Cronbach’s α = 0.931), with strong test–retest reliability (ICC = 0.971). Convergent validity was demonstrated by the inverse correlation with fear of hypoglycemia (Fear of Hypoglycemia Scale-15: r = −0.560, P < 0.001) and fasting blood glucose (r = −0.209, P = 0.015). Known-group validity revealed significant differences in HC scores according to education (P = 0.006), income (P = 0.048), and frequency of hypoglycemia episodes (P = 0.041).

Conclusions:

The nine-item HCS-C is a culturally adapted, psychometrically robust tool for assessing HC in Chinese T2D-BP populations. Its validation addresses a critical gap in the assessment of hypoglycemia management and supports tailored interventions to enhance patient safety and glycemic outcomes.

Introduction

Optimal glycemic control remains foundational in diabetes care for reducing microvascular and macrovascular complications. ¹ However, stringent glycemic targets often increase hypoglycemia risk, creating a therapeutic dilemma between complication prevention and acute safety. ² Hypoglycemic events impose a multidimensional burden: they reduce quality of life through physical symptoms and psychological distress (e.g., fear of hypoglycemia, FoH), elevate health care costs from emergency interventions, and paradoxically hinder long-term glycemic goals by causing treatment disengagement.^3,4 This cyclical relationship necessitates a paradigm shift. Recent advancements in glucose-responsive insulin analogs, continuous glucose monitoring systems, and closed-loop insulin delivery have substantially reduced severe hypoglycemia incidence.^5,6 Consequently, contemporary diabetes management emphasizes personalized glycemic targets that balance physiological control with patient-centered outcomes, particularly safety perception and hypoglycemia confidence (HC).

HC is a new concept in hypoglycemia research that reflects a patient’s belief in their ability to avoid hypoglycemia-related adverse events. This positive cognition is distinct from FoH and is associated with better glycemic control in patients with type 2 diabetes (T2D) treated with insulin.^5,7,8 HC can empower diabetic patients to cope with hypoglycemia without resorting to avoidance behaviors that are detrimental to glycemic control. ⁹ Hypoglycemic Attitudes and Behavior Scale was used initially to assess HC (five items) as its one dimension. ⁸ However, this scale was shown not to provide a comprehensive and systematic and an in-depth assessment of HC. ⁵ This led the researchers to develop a single-dimension Hypoglycemia Confidence Scale (HCS) for measuring patients’ confidence in avoiding hypoglycemia-related complications. ⁵ HCS has been applied widely in various studies in different populations, including type 1 diabetes (T1D), T2D, and partners of adult patients with T2D.^10–12 The Turkish version of the HCS has been translated and validated for use in T1D. ⁶ Given the clinical relevance of HC and the absence of validated assessment tools in China, this study cross-culturally adapted the HCS with the objectives to (1) assess its psychometric properties (reliability and validity) and (2) examine its applicability for evaluating HC in Chinese patients with diabetes.

Methods

The study was conducted in two phases. The first phase involved translating the HCS from English to Chinese, while the second phase used a cross-sectional design to evaluate its psychometric properties.

Participants

We used convenience sampling to recruit participants from the Endocrinology Department of a tertiary hospital in Zhejiang Province, China, between July and September 2024. The sample size was determined using the Kendall estimation method, which indicated a sample size of at least 5–10 times the number of items on the scale was required to achieve sufficient statistical power. ¹³ Given that the HCS comprises nine items and has a 20% rate of invalid responses, we aimed for a target sample size of 108 patients to ensure robust psychometric analysis. The participants were included in the study if they met the following criteria: (1) a diagnosis of T2D for at least 1 year, (2) currently using both basal and prandial insulins, (3) age 18 years or older, and (4) willing to participate in the study. Participants were excluded if they had (1) a history of mental illness and/or use of antipsychotic medications, (2) inability to read or comprehend the questionnaire, or (3) presence of cognitive impairment. A total of 150 eligible patients with T2D were enrolled in the cross-sectional study. Ethical approval was obtained from the Ethics Committee of Lishui University (Approval No. 2024YR0033), with written informed consent obtained from all the participants prior to their involvement in the study.

Measurement instrument

Personal information questionnaire

We designed the personal information questionnaire based on the objectives of the study and a review of previous relevant literature. The questionnaire included variables such as age, gender, diabetes duration, educational status, income status, frequency of hypoglycemia episodes, and levels of glycated hemoglobin A1c (HbA1c) and fasting blood glucose (FBG).

Hypoglycemia Confidence Scale

Polonsky et al. developed a preliminary nine-item HCS through interviews with adult patients diagnosed with type 1 diabetes mellitus (T1DM), insulin-treated patients with type 2 diabetes mellitus, and diabetes specialists. ⁵ Subsequently, an exploratory factor analysis (EFA) was conducted on 326 adult patients with T1D, 145 adult patients with T2D-BP, and 82 adult patients with T2D only using basal insulin administration (T2D-BO). This process resulted ultimately in a unidimensional, nine-item self-reported HCS designed to assess how safe and comfortable diabetic patients felt about their ability to avoid hypoglycemia-related problems. The scale employed a 4-point Likert scoring method, with scores ranging from 1 to 4, representing “no confidence at all,” “a little confidence,” “moderate confidence,” and “very confident,” respectively. The total score ranged from 9 to 36, with higher scores indicating greater levels of HC. The EFAs of the HCS yielded a single-factor solution for each group, explaining 50.8% (T1D), 65.1% (T2D-BP), and 73.7% (T2D-BO) of the variance. The HCS was associated significantly with well-being and diabetes distress in the T1D (both P < 0.001) and T2D-BP groups (both P < 0.05). The HCS was also related significantly to self-reported HbA1c in the T2D-BP group (P < 0.05), independent of FoH. ⁵ In 2023, Gokhan et al. translated the scale into Turkish and validated it in patients with T1DM, achieving a Cronbach’s α coefficient of 0.814 and a test–retest reliability of 0.885. ⁶ These results confirmed the good psychometric properties of the HCS.

Procedure

Translation and pre-survey

Upon receiving permission, via an email from Professor Polonsky, the original author of the HCS, we proceeded with the translation of the HCS into Chinese, adhering to established guidelines for cross-cultural adaptation. ¹⁴ A panel of experts was then assembled, including four nursing specialists with expertise in diabetes care and proficiency in English, none of whom had prior exposure to the original scale, as well as two translation experts unfamiliar with the HCS. The translation process involved two nursing experts independently translating the original scale into Chinese, producing two initial versions, T1 and T2. These versions were then reviewed by a third nursing specialist, who facilitated discussions to reconcile discrepancies and synthesize a unified version, T12. Following this, two independent English translators back-translated T12 into English, resulting in versions B1 and B2. A comparative analysis of B1, B2, and the original scale was conducted by a fourth nursing specialist, culminating in the development of version B12. All the versions were subsequently reviewed by the translators, who engaged in collaborative discussions to refine and finalize the preliminary Chinese version of the HCS, designated as HCS-C1.

A panel of six experts (five females; three PhDs, two master’s degrees, and one bachelor’s degree) carried out the cultural adaptation process. Of these experts, four held senior professional titles, while two held associate senior professional titles. The expert panel comprised two clinical nurses from tertiary hospitals specializing in diabetes care, three nursing professors with expertise in medical nursing, and one endocrinologist actively engaged in research. The participants’ ages ranged from 42 to 48 years, with their professional experience between 22 and 31 years. The diverse backgrounds and extensive expertise of these experts ensured that the revised scale was both culturally appropriate and highly applicable to the nursing context in China. The expert panel evaluated each item of the Chinese version of the scale in terms of linguistic appropriateness, content relevance, and structural integrity. The evaluation was conducted using a 4-point Likert scale, with scores of 1–4 representing “not relevant,” “weakly relevant,” “moderately relevant,” and “highly relevant,” respectively. Based on the experts’ feedback, the researchers made revisions and promptly communicated these changes to the panel, ultimately resulting in the HCS-C2 version. Specific modifications included “hypoglycemia confidence” changed to “hypoglycemia coping confidence” in HCS-C2 to make it understandable in the Chinese context. The phrase in the HCS-C1 version “How confident are you that you can avoid serious hypoglycemia problems?” was revised to “How confident are you in your ability to ensure your safety during hypoglycemia?” while Item 8 “Continue to do the things you really want to do in your life, despite the risks of hypoglycemia?” was revised to “Despite the risk of hypoglycemia, continue to do what you really want to do in your life?”. Item 9 was used to assess the partners’ trust, which was inconsistent with the former item 8, leading one expert to suggest deleting the item. However, based on evidence from previous research, ⁶ our research team decided to keep the item.

In July 2024, research assistants (RAs), consisting of two diabetes specialists, approached potential participants during their hospitalization or follow-up visits to the Endocrinology Department. The RAs provided comprehensive details about the study and invited eligible individuals to participate in a pre-survey. Those who expressed willingness and met the inclusion criteria were asked to provide written informed consent. Following this, the RAs collected the participants’ clinical data, including age, gender, duration of diabetes, and levels of HbA1c and FBG. The participants were then instructed to complete self-administered questionnaires to assess their personal information that included the HCS, Fear of Hypoglycemia Scale-15 (FH-15), and Summary of Diabetes Self-Care Activities (SDSCA). Feedback from the participants on the HCS indicated that item 9 should be deleted, because they found it difficult to answer. After discussion within the research group, it was decided to retain the item for the purpose of psychometric evaluation.

From late July to September 2024, the formal survey was conducted by the RAs following the aforementioned procedures. A total of 150 participants were recruited initially, although 10 were excluded for using basal insulin alone. Of the remaining 140 participants, five submitted incomplete questionnaires, leaving 135 eligible datasets for analysis. In order to assess test–retest reliability, 30 of these participants were subsequently invited to complete the HCS again during their outpatient visits. Of these, 28 participants completed the questionnaire, while 2 were unable to attend the visit due to work commitments.

Statistical analysis

Psychometric properties in the HCS were evaluated using SPSS Version 23 and AMOS Version 26. Given the non-normal distribution of the variables, nonparametric tests were employed throughout the study. Descriptive statistics, including means and standard deviations, as well as frequencies and percentages, were calculated for demographic and clinical variables, and for the total scores of each patient-reported outcome instrument.

The internal consistency and test–retest reliability of the HCS were evaluated using Cronbach’s α and the intraclass correlation coefficient (ICC), respectively, with data collected at weeks 1 and 4. A reliability coefficient of 0.8 or higher was considered satisfactory. ¹⁵ Content validity was assessed using the item-level content validity index (I-CVI) and scale-level content validity index/average (S-CVI/Ave).

The confirmatory factor analysis (CFA) was used to assess the construct validity of the HCS. In the CFA, the indices used to evaluate the model fit included the following criteria: Chi-square/degree of freedom (CMIN/DF) < 5, root mean square error of approximation (RMSEA) < 0.08, comparative fit index (CFI), incremental fit index (IFI), Tucker–Lewis index (TLI) ≥ 0.9, and normed fit index (NFI) > 0.9. ¹⁶ To examine convergent validity, due to non-distribution of the HCS mean score, Spearman’s correlation coefficients were calculated between HCS mean scores and FH-15 scores, SDSAC scores, and HbA1c and FBG levels. A priori hypothesis posited that the HCS mean scores would correlate with the FH-15 scores, SDSAC scores, and HbA1c and FBG levels. To provide known-group validity, Kruskal–Wallis H tests were conducted to detect differences in HCS mean scores across subgroups defined by educational status, income status, and frequency of hypoglycemia episodes. The α level for all the analyses was set at P < 0.05.

Results

Characteristics of the participants

A total of 135 participants were included in the study—68 (50.4%) were female—with the study cohort having a mean age of 57.76 ± 15.05 years (range: 21–84 years), a mean duration of diabetes of 10.13 ± 7.73 years, and 43% having a primary school education or lower. Only 17 participants (12.6%) reported that their income was less than their expenditure. A total of 82 participants (60.7%) experienced hypoglycemia at least once a month, with their mean HbA1c level being 9.17 ± 2.31% and their mean FBG level 9.35 ± 3.35 mmol/L. The mean HCS score was 2.50 ± 0.78.

Content validity

Following cultural adaptation of the scale, the I-CVI for each item ranged from 0.83 to 1.00, with all the values exceeding the threshold of 0.78. The S-CVI/Ave was 0.981.

Construct validity

A total of 135 adults with T2D-BP participated in the factor analysis validation stage. Initially, we validated the original nine-item structure of the scale, but the results were unsatisfactory. Then we made several modifications to the model; the results seemed acceptable. The fit indices for the revised model were as follows: CMIN/DF = 1.787, RMSEA = 0.077, CFI = 0.982, IFI = 0.982, TLI = 0.973, and NFI = 0.961, indicating that the model’s fit indices were acceptable (Table 1). Therefore, the Chinese version of the scale comprising nine items was considered as a reasonable and accurate tool for measuring HC in adults with T2D-BP. Figure 1 presents the path diagram and CFA factor loadings, and the appendices are original HCS and HCS-C.

FIG. 1.

Confirmatory factor analysis factor pathway and loadings.

Table 1.

Fit Indices of the HCS-C for Adults with T2D-BP

Index	Good fit	Acceptable fit	HCS-C
CMIN/DF (χ2/SD)	<2	<5	1.787
CFI	>0.95	>0.90	0.982
IFI	>0.95	>0.90	0.982
NFI	>0.95	>0.90	0.961
TLI	>0.95	>0.90	0.973
RMSEA	<0.05	<0.08	0.077

CFI, comparative fit index; CMIN/DF, Chi-square/degree of freedom; HCS-C, Chinese version of the Hypoglycemia Confidence Scale; IFI, incremental fit index; NFI, normed fit index; RMSEA, root mean square error of approximation; SD, standard deviation; T2D-BP, type 2 diabetes using both basal and prandial insulins; TLI, Tucker–Lewis index.

To establish convergent validity, the correlations between the nine-item HCS-C mean score and the FH-15 score, SDSCA score, and HbA1c and FBG levels were evaluated (Table 2). The mean score of the HCS-C showed associations in the expected (and opposite) directions with the FH-15 score (r = −0.560, P < 0.001) and FBG level (r = 0.209, P = 0.015) but not with the HbA1c level (r = 0.055, P > 0.05) and SDSCA score (r = 0.067, P > 0.05).

Table 2.

The Relationship Between the Participants’ HCS Mean Score and Related Constructs

Variables	Spearman’s correlation coefficient	P value
HbA1c	0.055	0.530
FBG	−0.209	0.015
FH-15	−0.560	0.000
SDSCA	0.067	0.440

HbA1c, glycated hemoglobin A1c; FBG, fasting blood glucose; FH-15, Fear of Hypoglycemia Scale-15; SDSCA, Summary of Diabetes Self-Care Activities.

Known-group validity

The educational status (P = 0.006), income status (P = 0.048), and frequency of hypoglycemia episodes (P = 0.041) of the participants were shown to correlate significantly with the mean HCS score.

Reliability analysis

The Cronbach’s α coefficient for the HCS-C was 0.931, demonstrating excellent internal consistency. In addition, the test–retest reliability of the HCS-C was 0.971, further confirming its high reliability.

Discussion

This study translated and validated the HCS-C in adults with T2D-BP, demonstrating its reliability and validity as a nine-item psychometrically robust tool for assessing HC (see Appendix A1：Original HCS and Appendix A2：HCS-C). The results align with previous studies that validated the HCS in other languages and populations,^6,10 further supporting its cross-cultural applicability and utility in clinical and research settings.

Our analysis of the scale’s dimensionality strongly supports its unidimensional nature, with a robust fit observed in the nine-item HCS-C. This demonstrated that the selected items effectively assessed HC in Chinese adults with T2D-BP. The scale’s one-dimensionality enhanced its psychometric reliability and suggested that these nine items comprehensively captured the core construct of HC within this population.

Our expert panel identified potential construct inconsistency in Item 9 (“partner’s trust”), as its focus on interpersonal dynamics diverged from the self-referential nature of the preceding eight items that assessed personal confidence. However, our CFA result revealed that the original nine-item structure of the HCS fits well in the Chinese context. These indices suggested that the nine-item HCS-C had a good fit to the data and was a valid measure of HC in Chinese patients with T2D-BP. Our findings were supported by the original HCS validation study of Polonsky et al. ⁵ Thus, the final nine-item scale represented a clinically coherent measure of situation-specific hypoglycemia coping confidence, integrating temporal safety awareness, self-efficacy dimensions, as well as the partner’s trust.

The HCS-C had excellent internal consistency, with a Cronbach’s α coefficient of 0.945, which is consistent with the original English version (Cronbach’s α = 0.87–0.95) ⁵ and the Turkish adaptation (Cronbach’s α = 0.814–0.885). ⁶ The high internal consistency indicated that the items within the scale were closely related and measured the same underlying construct of HC. In addition, the test–retest reliability of the HCS-C was 0.972, suggesting that the scale was stable over time and capable of producing consistent results when administered to the same individuals at different time points.

The content validity of the HCS-C was also robust, with I-CVI ranging from 0.83 to 1.00 and a S-CVI/Ave of 0.981. These values exceed the recommended thresholds, indicating that the scale items were relevant and representative of the construct being measured. ¹⁵ The cultural adaptation process, which involved rigorous translation and expert review, ensured that the scale was linguistically and culturally appropriate for Chinese adults with T2D-BP.

The convergent validity of the HCS-C was supported by significant correlations with other relevant constructs. As hypothesized, the HCS-C scores correlated inversely with FoH and FBG levels. These findings were consistent with those of previous research, which showed that a higher HC was associated with lower FoH and FBG.^5,6

However, the absence of a significant correlation between the mean scores of HCS-C and HbA1c levels and self-management behaviors contradicts previous findings. A plausible explanation is that HC is more closely associated with patients’ perceptions than with objective glycemic control and behavior measures. This is in line with a previous study that reported that self-efficacy did not show a significant correlation with HbA1c in patients with T2D. ¹⁷ Nevertheless, more in-depth research is needed to thoroughly explore this correlation.

Known-group validity was also established, with significant differences in HCS-C scores observed across subgroups defined by educational status, income status, or frequency of hypoglycemia episodes. These findings suggested that the HCS-C was sensitive to differences in HC among patients with varying sociodemographic and clinical characteristics. For example, patients with higher educational levels and income may have greater access to resources and support, which could enhance their confidence in managing hypoglycemia.^18,19

Clinical implications

The HCS-C provides a valuable tool for clinicians and researchers to assess HC in Chinese adults with T2D-BP. Given the significant role of HC in diabetes management, the HCS-C can be used to identify patients who may benefit from targeted interventions to enhance their confidence and improve their ability to manage hypoglycemia. For instance, patients with low HC may require additional education, support, or psychological interventions to reduce fear and avoidance behaviors, which can negatively impact glycemic control and their quality of life.

Moreover, the HCS-C can be used to evaluate the effectiveness of interventions aimed at improving HC, such as diabetes self-management education programs or the use of advanced diabetes technologies such as continuous glucose monitoring or automated insulin delivery systems. By systematically assessing HC, health care providers can tailor interventions to meet the specific needs of their patients and improve overall diabetes outcomes.

Limitations

While the findings of this study are promising, several limitations should be acknowledged. First, the study sample was recruited from a single tertiary hospital in Zhejiang Province, which may have limited the generalizability of the results to other regions or health care settings in China. Single-center sampling lacks diversity in terms of cultural background, medical resources, and socioeconomic and self-management behavior levels. Our previous multiple-center cross-sectional study confirmed area-based differences of self-management behaviors. ²⁰ Therefore, the conclusion of this study needs to be verified through multicenter cooperation and incorporated into other diabetes groups, such as patients with T1D or T2D-BO.

Second, the cross-sectional design of the study limited our ability to draw causal inferences regarding the relationships between HC and other variables, such as FoH or self-management behavior. Longitudinal studies are therefore needed to determine how changes in HC over time impact diabetes outcomes, including glycemic control, quality of life, and health care utilization.

Conclusions

The Chinese version of the HCS is a reliable and valid tool for assessing HC in adults with T2D-BP. The scale has strong psychometric properties, including excellent internal consistency, test-retest reliability, and content validity. The nine-item structure of the HCS-C provides a culturally appropriate and clinically useful measure of HC, which can be used to inform patient care and evaluate the effectiveness of interventions aimed at improving diabetes management. Future research should focus on validating the HCS-C in more diverse populations and explore its use in longitudinal and interventional studies.