Abstract
Aim
To assess knowledge, attitudes, and practices (KAP) among family caregivers of patients with acute exacerbations of chronic obstructive pulmonary disease (AECOPD).
Methods
This cross-sectional survey recruited caregivers through Yizheng People’s Hospital and associated community settings in Jiangsu Province, China, from July to September 2024. A self-developed questionnaire was administered online, on paper, or with staff assistance. Associations were examined using Spearman correlations, multivariable logistic regression, and structural equation modeling (SEM). Psychometric properties were additionally evaluated using a six-expert content-relevance assessment and a three-factor measurement model.
Results
Among 500 caregivers, 278 (55.6%) were female. Mean knowledge, attitude, and practice scores were 14.07 ± 5.15, 33.08 ± 3.71, and 36.14 ± 4.10, respectively. Using a standardized-score threshold >70%, 231 (46.2%) met the operational knowledge criterion, 401 (80.2%) met the positive-attitude criterion, and 338 (67.6%) met the proactive-practice criterion. Higher knowledge (adjusted odds ratio [OR] = 1.093, 95% confidence interval [CI]: 1.043–1.144) and attitude scores (OR = 1.255, 95% CI: 1.164–1.353) were associated with proactive practice; higher age was inversely associated (OR = 0.974, 95% CI: 0.950–0.998). SEM showed knowledge–attitude (β = 0.488, P = 0.001) and attitude–practice associations (β = 0.714, P = 0.001). The indirect knowledge–practice association through attitude was significant (β = 0.349, P = 0.001), whereas the direct path was not (β = 0.179, P = 0.064). Model fit and convergent-validity evidence were mixed.
Conclusion
The surveyed caregivers reported limited knowledge but generally positive attitudes and proactive practices. The indirect statistical relationship through attitude does not establish causal mediation. Prospective studies are needed to validate the questionnaire further and evaluate caregiver-support interventions.
Keywords: chronic obstructive pulmonary disease, acute exacerbation, knowledge, attitudes, practices, family caregivers, structural equation modeling
Plain Language Summary
An acute worsening of chronic obstructive pulmonary disease can make breathing much more difficult. Family caregivers often help patients follow treatment, attend appointments, avoid smoke and other triggers, and manage everyday care. We surveyed 500 caregivers to understand how familiar they were with information about these episodes, how they viewed caregiving, and what care they reported providing. Many caregivers reported positive attitudes and active involvement, although their familiarity with several topics, particularly breathing exercises and rehabilitation, was more limited. Caregivers with higher knowledge and attitude scores were more likely to report proactive care. Statistical analysis suggested that the relationship between knowledge and practice was partly represented through attitude; a separate direct relationship was not clearly established after attitude was considered. However, all information was collected at one time, and the questionnaire still needs further validation. These findings cannot show that education causes better caregiving or improves patients’ health. They may help identify topics for future caregiver-support programs, whose benefits would need to be tested.
Introduction
Chronic obstructive pulmonary disease (COPD) is a major cause of respiratory disability and mortality worldwide.1,2 The China Pulmonary Health study estimated a COPD prevalence of 13.7% among adults aged 40 years or older and approximately 100 million affected adults nationwide.3 Acute exacerbations of COPD (AECOPD) involve worsening respiratory symptoms and can require additional treatment or hospitalization. Hospitalized exacerbations are associated with substantial clinical burden and a risk of in-hospital death.2,4
Care demands extend beyond the hospital. During an exacerbation and subsequent recovery, family caregivers may support medication use, symptom monitoring, daily activities, communication with healthcare professionals, and decisions about seeking care. Qualitative studies describe both the practical tasks and the emotional responsibilities involved in caregiving for people with COPD.5,6 Understanding caregivers’ perspectives and reported activities is therefore relevant to the design of support that addresses their actual needs.
Non-pharmacological management and pulmonary rehabilitation are important components of COPD care. The American Thoracic Society recommends pulmonary rehabilitation for adults with stable COPD and after hospitalization for an exacerbation.7 Rehabilitation combines individualized exercise training with education and self-management support. Education may address symptom recognition, breathing strategies, physical activity, treatment use, and avoidance of respiratory irritants.8 Caregivers may help patients apply these strategies during recovery and home management. This context is particularly relevant when questionnaires assess familiarity with breathing exercises and bedside rehabilitation activities; it does not imply that every exercise or device is appropriate for every patient during an acute episode.
Knowledge, attitude, and practice (KAP) surveys characterize what people know about a health issue, how they view it, and what behaviors they report undertaking. They can identify educational needs and explore relationships among these domains, but they do not by themselves establish an inevitable sequence from knowledge to attitude and behavior.9,10 In this study, KAP provided an organizing framework for examining concurrent statistical associations rather than a basis for causal claims.
Previous studies have identified disease- and treatment-related knowledge gaps in COPD patient–caregiver dyads and among primary family caregivers.11,12 However, evidence simultaneously examining knowledge, attitudes, and reported practices specifically in the AECOPD caregiver context remains limited. This study assessed these domains, examined factors associated with proactive practice, and evaluated a prespecified statistical pathway linking knowledge, attitude, and practice in a single-center and community-associated sample.
Methods
Study Design and Participants
This cross-sectional study was conducted through Yizheng People’s Hospital and selected associated community settings in Jiangsu Province, China. Questionnaire records included in the analysis were submitted between July 25 and September 19, 2024. The study targeted adult family caregivers who had provided care for at least one month to a patient diagnosed with AECOPD and were currently responsible for the patient’s daily care. Eligibility was restricted to family caregivers; the broad “Other” response category in the questionnaire did not expand eligibility beyond family members. Individuals unable to communicate adequately or not currently involved in daily care were excluded. Participants were recruited using convenience sampling through the respiratory outpatient clinic, a chronic disease management group, wards, and selected community settings.
The study was approved by the Ethics Committee of Yizheng People’s Hospital (Ethical Approval Number: YPH Medical Scientific Research 2023–23), and informed consent was obtained from all study participants. The study was conducted in accordance with the Declaration of Helsinki.
Questionnaire Development and Scoring
The questionnaire was independently developed using literature on family caregiving during COPD exacerbations, disease-specific education, and rehabilitation.5,13,14 It comprised demographic and patient-related characteristics, 11 knowledge items, eight attitude items, and nine practice items. In a pilot survey of 50 participants, the overall Cronbach’s α was 0.926; domain-specific values were 0.935 for knowledge, 0.911 for attitude, and 0.835 for practice. These pilot results assessed internal consistency rather than establishing comprehensive validity. The administered item wording and original response categories are reproduced in English in Supplementary Material 1.
Knowledge responses were scored as 2 for “Very familiar”, 1 for “Heard of it”, and 0 for “Not sure”, yielding a theoretical total of 0–22. This domain assessed self-reported familiarity with AECOPD information rather than objectively verified factual knowledge. Attitude responses were scored as 5 for “Strongly agree”, 4 for “Agree”, 3 for “Neutral”, 2 for “Disagree”, and 1 for “Strongly disagree”, yielding a total of 8–40.
Practice responses were scored as 5 for “Completely consistent”, 4 for “Consistent”, 3 for “Uncertain”, 2 for “Inconsistent”, and 1 for “Very inconsistent”, yielding a total of 9–45. These English labels translate the original Chinese response anchors. The practice domain assessed how well statements described respondents’ reported care; it did not use an “Always” to “Never” frequency scale. Higher scores indicated greater familiarity, more favorable attitudes, or more proactive self-reported practice. No items were reverse-scored.
To compare domains with different theoretical minima and maxima, standardized percentage scores were calculated as [(observed score − theoretical minimum)/(theoretical maximum − theoretical minimum)] × 100%. A standardized score >70% was used as an operational classification criterion. The corresponding raw-score thresholds were ≥16 for the operational knowledge criterion, ≥31 for the positive-attitude criterion, and ≥35 for the proactive-practice criterion. This threshold was not a clinically validated cutoff for AECOPD caregivers; the categorized results were interpreted accordingly.
Supplementary Content-Relevance Assessment
After completion of the 2024 survey, six experts independently evaluated the relevance of all 28 KAP items in a supplementary post hoc assessment. Their expertise comprised pulmonary and critical care medicine (a chief physician), respiratory nursing (a senior professional title), pulmonary rehabilitation or respiratory therapy, chronic disease management or community nursing, epidemiology and medical statistics, and nursing research or questionnaire methodology. This assessment was not part of the original pre-survey development process and did not change the items administered in 2024.
Relevance was rated from 1 (not relevant) to 4 (highly relevant); scores of 2 indicated major revision was needed and scores of 3 indicated minor revision was needed. Ratings of 3 or 4 were treated as relevant. The item-level content validity index (I-CVI) was the proportion of experts rating an item as 3 or 4. The scale-level averaging index (S-CVI/Ave) was the mean I-CVI, and universal agreement (S-CVI/UA) was the proportion of items rated relevant by all six experts. Modified kappa coefficients were calculated to adjust for chance agreement. I-CVI ≥0.78 and S-CVI/Ave ≥0.90 were used as reference criteria for item relevance and average scale-level agreement.15,16 The assessment addressed relevance; it did not independently establish comprehensiveness or caregiver comprehension of all items.
Questionnaire Administration and Data Quality
The questionnaire was created in Wenjuanxing and distributed using a QR code through the respiratory outpatient clinic and chronic disease management group. Paper questionnaires and staff-assisted face-to-face completion were also used to improve accessibility. Trained respiratory medicine staff provided instructions and assistance to respondents who had difficulty completing the survey. Participant-level administration mode was not retained as a separate analysis variable, so mode-specific comparisons were not possible.
Data quality checks excluded records without consent for research use, records from people not currently responsible for daily care, completion times below 90 seconds, malformed household-income responses, and uninterpretable age entries. The income exclusions concerned invalid text in the response field rather than high or low income. No data were missing for the variables included in the final analytic dataset, and no imputation was performed.
Sample Size
The sample size for a cross-sectional proportion estimate was calculated using the following formula:
 |
Here, n denotes the required sample size, α = 0.05, Z1₋α/2 = 1.96, p = 0.50, and the allowable error δ = 0.05. The calculation gave n = 384.16, or approximately 384 participants. Allowing for approximately 20% nonresponse or exclusion, the planned target was approximately 480 records. The final sample of 500 exceeded the calculated requirement. This calculation was for the survey sample, not a separate SEM-specific power calculation.
Statistical Analysis
Statistical analyses were performed using SPSS 27.0 and AMOS 26.0 (IBM, Armonk, NY, USA). Continuous variables are summarized as means and standard deviations (SDs), and categorical variables as frequencies and percentages. Mann–Whitney U or Kruskal–Wallis tests were used for comparisons of KAP scores between demographic groups, and Spearman correlations described associations among domain totals. These group comparisons were exploratory; P values were not adjusted for multiple comparisons.
For analysis, master’s and doctoral education were combined, as were divorced and widowed status, retired and unemployed status, and the two highest income categories. The original 7–10-day and >10-day exacerbation-duration categories were combined. The exact original options are retained in Supplementary Material 1; analysis categories are distinguished from those options in Tables S1 and S3.
Proactive practice was the binary outcome: a practice score ≥35 was coded 1 and a score <35 was coded 0. Univariable logistic regressions examined each measured characteristic. Variables associated with proactive practice at P <0.05 in the univariable analyses were entered into the multivariable model. The multivariable model therefore included knowledge score, attitude score, caregiver age, residence, and education. Education was included as a categorical variable with secondary school or below as the reference. Odds ratios (ORs), 95% confidence intervals (CIs), and two-sided P values are reported; age and KAP-score ORs are per one-year or one-point increase. Statistical significance was defined as P <0.05.
The final analytic sample was used to assess Cronbach’s α and explicitly report the prespecified three-factor confirmatory measurement model. Each item loaded on its corresponding knowledge, attitude, or practice factor. Standardized loadings, composite reliability (CR), and average variance extracted (AVE) were examined, with CR ≥0.70 and AVE ≥0.50 used as conventional reference values.17 Heterotrait–monotrait (HTMT) ratios were calculated from inter-item correlations, with values below 0.85 taken as supportive, rather than conclusive, evidence of construct separation.18 All 28 original items were retained. No post hoc item deletion or correlated-error modification was used to improve fit.
The SEM used maximum likelihood estimation with numerically coded item responses and three prespecified structural paths: knowledge to attitude, attitude to practice, and knowledge to practice. It estimated direct, indirect, and total standardized associations; the indirect association was the product of the knowledge–attitude and attitude–practice coefficients. Bias-corrected 95% CIs and two-tailed P values were estimated using 5000 nonparametric bootstrap resamples. These analyses describe concurrent statistical pathways, not temporal or causal mediation.
Model fit was evaluated using the chi-square/degree-of-freedom ratio (CMIN/DF), root mean square error of approximation (RMSEA), incremental fit index (IFI), Tucker–Lewis index (TLI), and comparative fit index (CFI). CMIN/DF <3 was treated as preferable, while <5 was a lenient reference; RMSEA <0.08 and incremental indices ≥0.90 were used as conventional reference values. Fit indices were interpreted jointly rather than as rigid pass–fail rules.19
Results
Participant Flow and Characteristics
A total of 595 questionnaire records were submitted. One was excluded because the respondent did not consent to research use, and 67 because the respondents were not currently responsible for daily care. Of the remaining 527 records, 18 were excluded for completion times <90 seconds, seven for malformed household-income responses, and two for uninterpretable age entries. The final sample comprised 500 caregivers, representing 84.0% of submitted records. This percentage is an analytic-retention proportion, not a population response rate; the number of eligible people approached but not participating was not recorded.
Among the 500 caregivers, 278 (55.6%) were female, mean age was 48.62 ± 9.17 years, 346 (69.2%) lived in urban areas, 226 (45.2%) had secondary-school education or below, and 405 (81.0%) lived with the patient. The “Child” relationship category was selected by 185 (37.0%). Among patients, 277 (55.4%) were male, mean age was 71.66 ± 11.96 years, 150 (30.0%) had COPD for 1–3 years, 274 (54.8%) were reported to smoke, and 268 (53.6%) had a previous exacerbation. Detailed characteristics and group-specific scores are presented in Table S1.
KAP Scores and Item Responses
Mean knowledge, attitude, and practice scores were 14.07 ± 5.15, 33.08 ± 3.71, and 36.14 ± 4.10, respectively. With the operational standardized-score threshold >70%, 231 caregivers (46.2%) met the operational knowledge criterion, 401 (80.2%) met the positive-attitude criterion, and 338 (67.6%) met the proactive-practice criterion. The remaining 162 (32.4%) were classified as non-proactive for logistic regression. Knowledge scores differed by education, relationship category, cohabitation, marital status, occupation, and several patient-related characteristics. Attitude and practice scores also varied across selected groups (Table S1). These comparisons do not establish independent associations.
The highest proportions of “Not sure” knowledge responses concerned breathing exercises and respiratory resistance training (K11, 15.4%), exercise or bedside rehabilitation activities (K10, 14.6%), and additional treatment beyond regular COPD management (K5, 13.8%). The largest proportions of negative attitude responses (“Disagree” or “Strongly disagree”) concerned timely preventive or control measures (A2, 13.8%), trust in the treatment plan (A6, 4.0%), and regular check-ups and rehabilitation (A5, 3.8%). Negative practice responses (“Inconsistent” or “Very inconsistent”) were reported for follow-up attendance (P4, 5.2%), dietary management (P5, 4.6%), smoking avoidance (P6, 4.4%), psychological support (P8, 4.4%), and keeping emergency medication and oxygen equipment at home (P9, 4.4%). Complete distributions, including zero frequencies, are shown in Table S2.
Correlations and Factors Associated with Proactive Practice
Spearman correlations were positive between knowledge and attitude (r = 0.348), knowledge and practice (r = 0.410), and attitude and practice (r = 0.499); all P <0.001 (Table 1).
Table 1.
Correlations Among Knowledge, Attitude, and Practice Scores
| Domain | Knowledge | Attitude | Practice |
|---|---|---|---|
| Knowledge | 1.000 | — | — |
| Attitude | 0.348 (<0.001) | 1.000 | — |
| Practice | 0.410 (<0.001) | 0.499 (<0.001) | 1.000 |
Notes: Values are Spearman correlation coefficients (two-sided P values), based on 500 caregivers. The em dash (—) denotes the unreported upper triangle of the symmetric correlation matrix. Bold type is used only for column headings.
In the multivariable logistic model, higher knowledge (OR = 1.093, 95% CI: 1.043–1.144, P <0.001) and attitude scores (OR = 1.255, 95% CI: 1.164–1.353, P <0.001) were associated with higher odds of proactive practice. Higher caregiver age was associated with lower odds (OR = 0.974, 95% CI: 0.950–0.998, P = 0.035). Residence was not independently associated after adjustment (urban versus rural: OR = 1.356, 95% CI: 0.868–2.119, P = 0.181). Compared with secondary school or below, neither associate/bachelor’s education (OR = 1.229, 95% CI: 0.773–1.954, P = 0.384) nor master’s education or above (OR = 0.556, 95% CI: 0.283–1.091, P = 0.088) was statistically significant. Table S3 provides all univariable and multivariable results.
Content Relevance and Psychometric Properties
In the six-expert post hoc assessment, 20 items were rated relevant by all six experts (I-CVI = 1.000), and eight by five experts (I-CVI = 0.833): K5, K10, K11, A2, A8, P2, P5, and P9. Overall S-CVI/Ave was 0.952, with domain values of 0.955, 0.958, and 0.944 for knowledge, attitude, and practice. Overall S-CVI/UA was 0.714; modified kappa ranged from 0.816 to 1.000 (Table S4). These results support item relevance and high average expert agreement, but not universal agreement on every item.
In the final sample, Cronbach’s α was 0.921 for knowledge, 0.747 for attitude, and 0.755 for practice; the overall 28-item α was 0.898. Standardized loadings ranged from 0.652–0.758, 0.451–0.584, and 0.450–0.605, respectively. CR values were 0.921, 0.756, and 0.757, and AVE values were 0.514, 0.281, and 0.258 (Table S5). Internal consistency and CR were acceptable, whereas AVE supported convergent validity more clearly for knowledge and indicated limited convergence for attitude and practice. The overall α does not demonstrate that the questionnaire is unidimensional. HTMT ratios were 0.492 for knowledge–attitude, 0.549 for knowledge–practice, and 0.803 for attitude–practice (Table S6), providing supplementary evidence of separation without negating the convergence limitations.
SEM Analysis
SEM showed statistical paths from knowledge to attitude (β = 0.488, bias-corrected 95% CI: 0.392–0.590, P = 0.001) and from attitude to practice (β = 0.714, 95% CI: 0.589–0.842, P = 0.001). The indirect knowledge–practice association through attitude was significant (β = 0.349, 95% CI: 0.275–0.478, P = 0.001). In contrast, the direct knowledge–practice path was not statistically significant (β = 0.179, 95% CI: −0.011–0.309, P = 0.064). The total knowledge–practice association was β = 0.527 (95% CI: 0.397–0.641, P = 0.001) (Table 2 and Figure 1). Minor differences between sums of displayed coefficients arise from rounding.
Table 2.
Standardized Direct, Indirect, and Total Associations in the Structural Equation Model
| Model path | Association Type | Standardized β | Bias-Corrected 95% CI | Two-Tailed P |
|---|---|---|---|---|
| Knowledge → Attitude | Direct | 0.488 | 0.392–0.590 | 0.001 |
| Attitude → Practice | Direct | 0.714 | 0.589–0.842 | 0.001 |
| Knowledge → Practice | Direct | 0.179 | −0.011–0.309 | 0.064 |
| Knowledge → Practice through Attitude | Indirect | 0.349 | 0.275–0.478 | 0.001 |
| Knowledge → Practice | Total | 0.527 | 0.397–0.641 | 0.001 |
Notes: Standardized effects, bias-corrected 95% confidence intervals, and two-tailed P values were estimated using 5000 bootstrap resamples (N = 500). The indirect effect represents the statistical pathway through attitude. CI, confidence interval. The arrow (→) indicates the direction of the prespecified structural path, not demonstrated causality. Estimates are rounded independently to three decimals. Bold type is used only for column headings.
Figure 1.
Standardized Structural Equation Model of the Relationships among Knowledge, Attitude, and Practice. Values adjacent to the arrows represent standardized path coefficients. Solid arrows indicate statistically significant structural paths, whereas the dashed arrow indicates the nonsignificant direct path from knowledge to practice. The indirect statistical association between knowledge and practice through attitude was significant (β = 0.349, bias-corrected 95% CI: 0.275–0.478, P = 0.001). Because the study was cross-sectional, the directional arrows represent prespecified statistical pathways and should not be interpreted as evidence of causality.
Model fit was mixed: CMIN/DF = 3.163 was within the lenient <5 reference but exceeded 3, and RMSEA = 0.066 indicated acceptable approximate fit. IFI = 0.852, TLI = 0.838, and CFI = 0.851 were below 0.90, indicating suboptimal incremental fit (Table 3). The structural findings therefore require cautious interpretation rather than a claim of uniformly satisfactory fit.
Table 3.
Fit Indices of the Three-Factor Structural Equation Model
| Fit index | Conventional Reference | Observed Value | Interpretation |
|---|---|---|---|
| CMIN/DF | <3 preferred; <5 is a lenient reference | 3.163 | Within the lenient reference; above 3 |
| RMSEA | <0.08 | 0.066 | Acceptable approximate fit |
| IFI | ≥0.90 | 0.852 | Below reference |
| TLI | ≥0.90 | 0.838 | Below reference |
| CFI | ≥0.90 | 0.851 | Below reference |
Notes: Indices were interpreted collectively, not as absolute diagnostic cutoffs. Model fit was mixed, with suboptimal incremental fit. Bold type is used only for column headings.19
Abbreviations: CMIN/DF, chi-square divided by degrees of freedom; RMSEA, root mean square error of approximation; IFI, incremental fit index; TLI, Tucker–Lewis index; CFI, comparative fit index.
Discussion
In this single-center and community-associated sample, caregivers reported comparatively limited knowledge alongside generally positive attitudes and proactive practices. Using the operational standardized-score criterion, fewer than half met the knowledge threshold, while most met the attitude and practice thresholds. Knowledge, attitude, and caregiver age were associated with proactive practice in the adjusted model. In the latent-variable analysis, the indirect knowledge–practice relationship through attitude was significant, but the direct path was not. These results must be interpreted alongside the questionnaire’s mixed psychometric evidence and the cross-sectional design.
The coexistence of limited knowledge and favorable reported practice is not unique to this sample. A study of caregivers of people with hypertension in rural Malawi also reported poor knowledge alongside favorable attitudes and good practices.20 In COPD patient–caregiver dyads, disease knowledge has been found to be limited, particularly for treatment-related topics.11 Direct comparisons of score levels are not appropriate because the questionnaires and classifications differ, but these studies support discussing knowledge and reported behavior as distinct rather than interchangeable domains.
Qualitative COPD research shows that caregivers contribute to practical routines, medication support, monitoring, healthcare communication, and emotional support.5,6 Such activities may reflect accumulated routines, professional guidance, and perceived family responsibility as well as formal knowledge. These are possible explanations, not mechanisms measured in this study. Moreover, the knowledge domain measured self-reported familiarity, and the practice domain measured agreement that statements described care. Socially desirable responding could inflate reported practice. Future studies should compare questionnaire responses with documented follow-up attendance, medication-management records, smoking-exposure measures, or direct observation of selected care activities.
The significant bivariate knowledge–practice correlation and the nonsignificant direct SEM path are not contradictory. The Spearman correlation describes an unadjusted association between observed domain totals, whereas the latent-variable model separates direct and indirect components while accounting for its specified measurement structure. The two estimates are therefore not numerically interchangeable. Within the SEM, the total association (β = 0.527) comprised a direct component (β = 0.179, P = 0.064) and an indirect component through attitude (β = 0.349, P = 0.001). Failure to reject the direct path does not establish that the true direct association is exactly zero or prove “complete mediation”. All domains were measured concurrently, so temporal ordering, reverse relationships, and alternative explanations remain unresolved. Cross-sectional mediation estimates cannot establish a longitudinal causal process.21
Higher caregiver age was associated with lower odds of meeting the proactive-practice criterion. The mechanism was not assessed. Education was included in the adjusted model, but neither education contrast was statistically significant. Digital literacy, health literacy, adaptability, caregiving burden and intensity, and access to online information were not directly measured. We therefore cannot attribute the age association to greater digital engagement or educational attainment among younger caregivers. Future work should measure these characteristics rather than infer them from age.
Group comparisons showed higher attitude and practice scores among urban caregivers, although residence was not independently associated with proactive practice in the adjusted model. COPD-related urban–rural disparities have been documented in other settings, including a national US study, but those findings do not establish the explanation for differences in this Chinese sample.22 Similarly, differences by education, occupation, cohabitation, or patient history identify descriptive variation rather than evidence that changing these characteristics would improve care. Several patient-related variables were reported by caregivers, and interpretation should acknowledge this source of information.
Uncertainty was greatest for rehabilitation-related knowledge items, including breathing exercises and bedside activities. This is relevant because pulmonary rehabilitation is recommended for stable COPD and after hospitalization for an exacerbation, with education and self-management support forming part of rehabilitation.7,8 The questionnaire combines several techniques within some items and does not identify exactly which component a respondent understood. In addition, an endorsement of a general rehabilitation statement is not proof of competence, clinical appropriateness, or safe technique. Exercise intensity, airway-clearance techniques, respiratory devices, and oxygen use require individualized professional assessment; the original items should not be read as universal treatment instructions during an acute episode.
The item-level findings identify possible targets for future caregiver-support research, including exacerbation recognition, rehabilitation-related information, follow-up attendance, dietary care, and avoidance of tobacco smoke or other triggers. Candidate delivery approaches could include brief face-to-face instruction, written materials, demonstrations, or accessible digital resources. No educational or behavioral intervention was tested here. Feasibility, acceptability, safety, and effectiveness should therefore be assessed prospectively before these approaches are recommended on the basis of this study.
These findings may be relevant to multidisciplinary caregiver support involving pulmonary physicians, respiratory nurses, pulmonary rehabilitation professionals, and community or chronic disease management teams, particularly when identifying education and follow-up needs.
The psychometric analyses also showed important limitations. Expert relevance ratings and acceptable internal consistency do not compensate for low AVE or prove a sound factor structure. Attitude and practice had several modest loadings and AVE below 0.50; incremental fit was also below 0.90. All original items were retained to preserve the administered instrument and its content coverage, supported by the relevance assessment, rather than deleting items solely to improve fit. This does not imply that every item is optimal. Prospective work should reassess item wording, comprehensiveness, caregiver understanding, and the dimensionality of these domains, followed by external validation.
This study also has several strengths. The final sample exceeded the prespecified survey requirement; a complete English translation of the administered questionnaire and item-level response distributions are provided; and observed-score regression analyses are complemented by a latent-variable model with transparent reporting of measurement properties.
Limitations
First, the cross-sectional design cannot establish temporal ordering or causality, and residual confounding is possible. Digital and health literacy, caregiving duration and intensity, burden, and information access were not measured in sufficient detail to explain the observed age association. No prospective patient outcomes or objectively observed caregiving behaviors were collected.
Second, convenience recruitment through a hospital, chronic disease management network, and associated community settings may preferentially include caregivers already engaged with healthcare services, while QR-code distribution may favor smartphone access or digital familiarity. These mechanisms could bias knowledge and proactive-practice estimates upward, but the magnitude cannot be quantified because information on eligible nonparticipants was unavailable. Mixed self-administered and staff-assisted approaches may also introduce mode-related differences; participant-level mode was not separately recorded, so such effects could not be formally assessed. These factors may limit generalizability to caregivers less connected to hospital services, with limited digital access, or in other regions and health systems. The broad “Other” relationship category was not further subclassified.
Third, the supplementary expert assessment was conducted after the survey and addressed item relevance rather than all aspects of content validity. Attitude and practice showed limited convergent validity, and incremental model fit was suboptimal. The 70% classification threshold was operational rather than clinically validated for AECOPD caregivers. Further prospective questionnaire refinement and external validation in independent samples are warranted.
Conclusions
The surveyed caregivers reported limited familiarity with AECOPD information but generally positive attitudes and proactive practices. Higher knowledge and attitude scores and lower caregiver age were associated with meeting the operational proactive-practice criterion. In the SEM, the indirect knowledge–practice association through attitude was significant, whereas the direct path was not. These concurrent relationships, together with mixed questionnaire-validity evidence, should inform hypotheses rather than causal claims. Longitudinal and intervention studies are needed to determine whether caregiver-support strategies improve care or patient outcomes.
Acknowledgments
The authors thank the participating caregivers and the six experts who contributed to the supplementary item-relevance assessment.
Funding Statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Abbreviations
AECOPD, acute exacerbation of chronic obstructive pulmonary disease; AVE, average variance extracted; CFI, comparative fit index; CI, confidence interval; CMIN/DF, chi-square divided by degrees of freedom; COPD, chronic obstructive pulmonary disease; CR, composite reliability; HTMT, heterotrait–monotrait ratio; I-CVI, item-level content validity index; IFI, incremental fit index; KAP, knowledge, attitudes, and practices; OR, odds ratio; RMSEA, root mean square error of approximation; S-CVI/Ave, scale-level content validity index based on averaging; S-CVI/UA, scale-level content validity index based on universal agreement; SD, standard deviation; SEM, structural equation modeling; TLI, Tucker–Lewis index.
Data Sharing Statement
The de-identified data supporting the findings of this study are available from the corresponding author upon reasonable request, subject to applicable ethical and institutional requirements.
Ethics Approval and Consent to Participate
This study was approved by the Ethics Committee of Yizheng People’s Hospital (Ethical Approval Number: YPH Medical Scientific Research 2023-23), and informed consent was obtained from all study participants. The study was conducted in accordance with the Declaration of Helsinki.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Disclosure
The authors declare that they have no competing interests.
References
- 1.Soriano JB, Abajobir AA, Abate KH.; GBD 2015 Chronic Respiratory Disease Collaborators. Global, regional, and national deaths, prevalence, disability-adjusted life years, and years lived with disability for chronic obstructive pulmonary disease and asthma, 1990–2015: a systematic analysis for the global burden of disease study 2015. Lancet Respir Med. 2017;5(9):691–11. doi: 10.1016/S2213-2600(17)30293-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Celli BR, Wedzicha JA. Update on clinical aspects of chronic obstructive pulmonary disease. N Engl J Med. 2019;381(13):1257–1266. doi: 10.1056/NEJMra1900500 [DOI] [PubMed] [Google Scholar]
- 3.Wang C, Xu J, Yang L, et al. Prevalence and risk factors of chronic obstructive pulmonary disease in China (the China Pulmonary Health [CPH] study): a national cross-sectional study. Lancet. 2018;391(10131):1706–1717. doi: 10.1016/S0140-6736(18)30841-9 [DOI] [PubMed] [Google Scholar]
- 4.Lima FV, Yen TY, Patel JK. Trends in in-hospital outcomes among adults hospitalized with exacerbation of chronic obstructive pulmonary disease. Copd. 2015;12(6):636–642. doi: 10.3109/15412555.2015.1020151 [DOI] [PubMed] [Google Scholar]
- 5.Aasbø G, Rugkåsa J, Solbrække KN, Werner A. Negotiating the care-giving role: family members’ experience during critical exacerbation of COPD in Norway. Health Soc Care Community. 2017;25(2):612–620. doi: 10.1111/hsc.12350 [DOI] [PubMed] [Google Scholar]
- 6.Pendoni R, Albanesi B, Clari M, Pecorari G, Matarese M. Contributing to self-care of a person with chronic obstructive pulmonary disease: a qualitative study of the experiences of family caregivers. J Adv Nurs. 2024;80(5):1927–1942. doi: 10.1111/jan.15939 [DOI] [PubMed] [Google Scholar]
- 7.Rochester CL, Alison JA, Carlin B, et al. Pulmonary rehabilitation for adults with chronic respiratory disease: an official American thoracic society clinical practice guideline. Am J Respir Crit Care Med. 2023;208(4):e7–e26. doi: 10.1164/rccm.202306-1066ST [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Blackstock FC, Lareau SC, Nici L, et al. Chronic obstructive pulmonary disease education in pulmonary rehabilitation. An official American thoracic society/thoracic society of Australia and New Zealand/Canadian thoracic society/British thoracic society workshop report. Ann Am Thorac Soc. 2018;15(7):769–784. doi: 10.1513/AnnalsATS.201804-253WS [DOI] [PubMed] [Google Scholar]
- 9.World Health Organization. Advocacy, Communication and Social Mobilization for TB Control: A Guide to Developing Knowledge, Attitude and Practice Surveys. Geneva: World Health Organization; 2008. ISBN:9789241596176. [Google Scholar]
- 10.Andrade C, Menon V, Ameen S, Praharaj SK. Designing and conducting knowledge, attitude, and practice surveys in psychiatry: practical guidance. Indian J Psychol Med. 2020;42(5):478–481. doi: 10.1177/0253717620946111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Matarese M, Lyons KS, Piredda M, De Marinis MG. Disease-related knowledge in people with chronic obstructive pulmonary disease and their informal caregivers: a multilevel modelling analysis. J Clin Nurs. 2023;32(13–14):3543–3556. doi: 10.1111/jocn.16433 [DOI] [PubMed] [Google Scholar]
- 12.Wang N, Zhu Y, Zhang Y. Demographic and educational factors associated with COPD knowledge among primary family caregivers: a cross-sectional study at a tertiary hospital. Medicine. 2025;104(36):e44228. doi: 10.1097/MD.0000000000044228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Janaudis-Ferreira T, Carr SJ, Harrison SL, et al. Can patients with COPD assimilate disease-specific information during an acute exacerbation?: results of a pilot randomized controlled trial. Chest. 2018;154(3):588–596. doi: 10.1016/j.chest.2018.05.028 [DOI] [PubMed] [Google Scholar]
- 14.Vincent EE, Chaplin EJ, Williams JEA, et al. Experiences of patients undergoing pulmonary rehabilitation during an exacerbation of chronic respiratory disease. Chron Respir Dis. 2017;14(3):298–308. doi: 10.1177/1479972317695812 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Polit DF, Beck CT, Owen SV. Is the CVI an acceptable indicator of content validity? Appraisal and recommendations. Res Nurs Health. 2007;30(4):459–467. doi: 10.1002/nur.20199 [DOI] [PubMed] [Google Scholar]
- 16.Polit DF, Beck CT. The content validity index: are you sure you know what’s being reported? Critique and recommendations. Res Nurs Health. 2006;29(5):489–497. doi: 10.1002/nur.20147 [DOI] [PubMed] [Google Scholar]
- 17.Fornell C, Larcker DF. Evaluating structural equation models with unobservable variables and measurement error. J Mark Res. 1981;18(1):39–50. doi: 10.1177/002224378101800104 [DOI] [Google Scholar]
- 18.Henseler J, Ringle CM, Sarstedt M. A new criterion for assessing discriminant validity in variance-based structural equation modeling. J Acad Marking Sci. 2015;43(1):115–135. doi: 10.1007/s11747-014-0403-8 [DOI] [Google Scholar]
- 19.Hooper D, Coughlan J, Mullen MR. Structural equation modelling: guidelines for determining model fit. Electron J Bus Res Methods. 2008;6(1):53–60. doi: 10.21427/D7CF7R [DOI] [Google Scholar]
- 20.Maluwa C, Kapira S, Chuljerm H, Parklak W, Kulprachakarn K. Determinants of hypertension-related knowledge, attitude, and practices (KAP) among caregivers in Neno, rural Malawi: a cross-sectional study. Heliyon. 2025;11(1):e41546. doi: 10.1016/j.heliyon.2024.e41546 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Maxwell SE, Cole DA. Bias in cross-sectional analyses of longitudinal mediation. Psychol Methods. 2007;12(1):23–44. doi: 10.1037/1082-989X.12.1.23 [DOI] [PubMed] [Google Scholar]
- 22.Gaffney AW, Hawks L, White AC, et al. Health care disparities across the urban-rural divide: a national study of individuals with COPD. J Rural Health. 2022;38(1):207–216. doi: 10.1111/jrh.12525 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The de-identified data supporting the findings of this study are available from the corresponding author upon reasonable request, subject to applicable ethical and institutional requirements.