Validity and reliability of the Mandarin version of the Treatment Burden Questionnaire among stroke patients in Mainland China Free

Abstract

Objectives

To examine the validity and reliability of the Mandarin version of the Treatment Burden Questionnaire (TBQ) among stroke patients.

Background

Stroke patients need long-term management of symptoms and life situation, and treatment burden has recently emerged as a new concept that can influence the health outcomes during the rehabilitation process.

Methods

The convenience sampling method was used to recruit 187 cases of stroke patients in a tertiary grade hospital in Tianjin for a formal investigation. Item analysis, reliability and validity tests were carried out. The reliability test included internal consistency and test–retest reliability. And as well as content, structure and convergent validity were performed for the validity test.

Results

Of the 187 completed questionnaires, only 180 (96.3%) were suitable for analysis. According to the experts’ evaluation, the I-CVI of each item was from 0.833 to 1.000, and the S-CVI was 0.967. The exploratory factor analysis yielded three-factor components with a cumulative variation of 53.054%. Convergent validity was demonstrated using measures of Morisky’s Medication Adherence Scale 8 (r = –0.450, P < 0.01). All correlations between items and global scores ranged from 0.403 to 0.638. Internal consistency reliability and test–retest reliability were found to be acceptable, as indicated by a Cronbach’s α of 0.824 and an intraclass correlation coefficient of 0.846, respectively.

Conclusions

The Mandarin TBQ had acceptable validity and reliability. The use of TBQ in the assessment of treatment burden of stroke survivor may benefit health resources allocation and provide tailor therapeutic interventions to construct minimally disruptive care.

Introduction

Stroke is a common chronic disease-causing significant disability and mortality worldwide and has been listed as a key neuropsychiatric disorder (1). Nowadays, the annual incidence rate of stroke is 8.7 per 1000, resulting in costs in excess of more than $100 billion in China (2). Approximately, 75% of stroke survivals had a disability to varying degrees, affecting patient’s daily life and activities (3). These stroke survivals may experience a variety of physical and psychosocial problems, requiring long-term rehabilitation and supports from families and communities (4–6). In long-term rehabilitation, stroke survivals also need to perform ongoing extensive self-management procedures to take care of themselves with a large expenditure of time and efforts, such as medication administration, self-monitoring, visits to the physician, laboratory tests and lifestyle restrictions (7–9). All efforts have the potential to compound a patients’ sense of treatment burden, defined as the ‘workload’ resulting from their treatments and disease-related self-care (10).

Treatment burden, an emerging concept, is increasingly being recognized as a crucial outcome for disease management and rehabilitation (11). Researches revealed that excessive treatment burden is related, independently of illnesses, to worse health outcomes including reduced quality-of-life, non-adherence and wasted resources (12,13). While health care workers tend to escalate treatment to meet the challenge of poor outcomes, requiring more ‘workload’ from patients, such as taking more medications, more frequent self-monitoring and greater lifestyle restrictions (14). This vicious cycle further aggravates the level of treatment burden. One cross-sectional study of 1 424 378 participants in the UK showed that multimorbidity and polypharmacy were strikingly more common in those with a diagnosis of stroke compared with those without, which means stroke survivals may have a heavier treatment burden than other chronic diseases (15). There is clearly a need to ascertain the level of treatment burden experienced by people with stroke. Therefore, it is imperative that researchers and clinics have access to effective assessment tool of treatment burden concerning individuals with stroke.

To date, a few instruments relevant to treatment burden have been reported (16–20), including the Treatment Burden Questionnaire (TBQ) (16). Compared with other measurements, the TBQ measures a broader range of treatment burden and the reliability and validity have been well demonstrated in different populations (16–18). The TBQ, including 13 items, was derived from a literature review and qualitative semi-structured interviews with patients in France (16). After adding two items about financial treatment burden and relationships between patients and medical workers, a more comprehensive 15-item English version was subsequently developed and exhibited similar validity to the original TBQ in a broader range of countries (17). The psychometric properties of TBQ have also been well reported across patients with multiple sclerosis in Spain (18). Recently, 15-item TBQ has also been rigorously translated into Cantonese and confirmed to have acceptable reliability and validity in the Cantonese-speaking population with chronic disease in Hong Kong (21). Unfortunately, the existing Cantonese version of the TBQ cannot be implemented directly in Mainland China due to the obvious difference in a cultural, economic, linguistic and political environment. Besides, Mainland China is the largest or broadest Chinese-speaking region; despite the shared language between Mainland China and Hong Kong, it is necessary to carry out separate validity and reliability of the Mandarin TBQ in Mainland China (22,23).

Therefore, the aims of our study were to translate and adapt the 15-item TBQ cross-culturally into Mandarin Chinese and examine its psychometric properties in Mainland China.

Methods

Study design and participants

A cross-sectional design was adopted. STROBE checklist for cross-sectional studies was followed to enhance methodological rigor (Supplementary material). The recruitment of a convenience sample of 187 patients was conducted at the neurology department of Tianjin Huanhu Hospital in Tianjin, China from July to September 2019. The patients had to meet the following criteria: (i) between 18 and 80 years old; (ii) being clinically diagnosed with ischemic or haemorrhagic stroke; (iii) over three months post-stroke with stable disease conditions; (iv) being able to understand and write Chinese; (v) volunteering to participate and sign written informed consent. Exclusion criteria included: (i) had been previously diagnosed with psychotic and mental disorders by any available diagnostic criteria or other validated scales and tests; (ii) had other severe diseases (such as cancer and heart failure). Ethical approval was obtained from the Ethics Committee of Tianjin Huanhu Hospital, and informed consent was obtained from all participants before data collection.

Translation of the TBQ into Chinese

After the permission to translate and use the questionnaire from the initial author and copyright company, the TBQ was translated into Chinese according to a classic ‘forward–backward’ translation method (24): an attending physician of the department of neurology with a high level of English and a neurological nurse with overseas study experience independently translated the original English scale into Chinese. The two experts then discussed their outputs to reach a consensus on the initial draft of the Mandarin version of TBQ. Then, the Mandarin TBQ was translated back into English by two different bilingual native Chinese speakers (one medical English teacher and the other who studied and worked as a doctor in an English-speaking country for over 10 years). There were 20 stroke patients were recruited for pilot language and reading tests of the Mandarin version of the TBQ. Finally, the formal Mandarin version of the TBQ was produced.

Measures

General demographic and clinical information

The information included questions on age, gender, residency, economical income, marital status, educational level, employment status, stroke type, comorbidity, types of the drugs taken orally, times of hospitalization, duration of disease, medical insurance and other clinical information.

Treatment Burden Questionnaire (TBQ)

The English version of TBQ is a 15-item questionnaire with one dimension that assesses the patient’s workload to maintain their health and its impact on the quality of life. Each item has 11-point response options from 0 to 10, which translates to low treatment burden to high treatment burden, respectively. The total scores, ranging from 0 to 150, are then calculated, with a higher score indicating higher treatment burden. The TBQ has been validated in several countries for multi-setting use. It has been shown to be reliable and valid, with good internal consistency and a high test-retest reliability (16–18,21).

Morisky’s Medication Adherence Scale 8 (MMAS-8)

The MMAS-8 is an 8-item scale that was used to assess medication adherence. Response choices are ‘yes’ or ‘no’ for items 1 through 7. Each ‘no’ response is scored 1 and each ‘yes’ response is scored 0 except for item 5, in which each ‘yes’ response is rated as 1 and each ‘no’ response is rated as 0. The last question is answered using a 5-point Likert scale. Responses of ‘never’, ‘once in a while’, ‘sometimes’, ‘usually’ and ‘all the time’ were scored 1, 0.75, 0.50, 0.25 and 0, respectively, whereas for the item were scored ‘1’ for ‘never’ and ‘0’for other responses. The total scores ranged from 0 to 8. It has been translated into different languages internationally and used for measuring medication adherence in multiple chronic conditions, including stroke worldwide (25,26). The reported reliability and validity among Chinese stroke patients were acceptable, with Cronbach’s α of 0.74 (27).

Statistical analysis

All data were analysed by SPSS 21.0. Descriptive statistics were conducted to summarize the basic demographic and clinical characteristics. The level of statistical significance was 0.05.

Item analysis was evaluated using item-total scale correlation with values >0.4 demonstrating adequate correlation.

Content validity was judged by a six-member expert panel (including two neurologists, two head nurses in the neurology department, a nursing management specialist and a professor of nursing). All experts were experienced in neurology or nursing and they had at least 10 years of working experience in senior clinical roles. The Likert 4-point scale was used, where ‘1–4’ respectively indicated ‘no relevance’ to ‘very much relevance’. Content validity index (CVI) was used to calculated content validity with experts’ score responses. Both scale-level CVI (S-CVI) and item-level CVI (I-CVI) were calculated for content validity. The I-CVI value ≥0.78 is considered acceptable, and S-CVI >0.8 means good content validity, and an S-CVI is hoped to achieve 0.90 (28,29). They were also consulted for revisions and whether any important content needs to add to meet the Chinese cultural environment.

We chose the exploratory factor analysis (EFA) to explore the underlying latent structure of the 15 items. To assess data factorability, the Kaiser–Meyer–Olkin (KMO) index of sampling adequacy and Bartlett’s test of sphericity coefficient were carried. An eigenvalue of over 1 was retained for factor extraction. Factor loadings >0.40 were considered to be significant (30,31).

Convergent validity tests how well two methods that theoretically measure the same construct are related to each other (32). Based on the previous studies, medication adherence is closely associated with the treatment burden during post-stroke recovery (17,18). Thus, to further explore the convergent validity of the Chinese TBQ, Pearson’s correlation coefficients (r) between the TBQ and MMAS-8 were calculated, with values of ＞0.40 indicating adequate correlation. We expected a negative correlation between treatment burden and medication compliance.

Cronbach’s α was calculated to analyse the internal consistency with the minimally acceptable alpha values of 0.7 of the TBQ. Test–retest reliability was assessed using the intraclass correlation coefficient (ICC) in a sub-sample of 20 randomly selected patients at the two-time points, where values＞0.70 are considered adequate (33,34).

Results

A total of 187 questionnaires were distributed. A final sample size of 180 patients with stroke was used for the analysis. The questionnaire was easily understood and required an average of 12.5 minutes to complete. Table 1 shows the details.

Item analysis

All correlations between items and total scores ranged from 0.403 to 0.638, indicating that each item is correlated with the overall Mandarin version of the TBQ.

Validity

The I-CVI values of the TBQ items ranged from 0.833 to 1.000. The value of the S-CVI was 0.967 (Table 2).

The KMO value was 0.777, and Bartlett’s test of sphericity coefficient was 851.635 (P < 0.001), which means TBQ is suitable for performing EFA. The principal component analysis with the varimax rotation yielded three-factor components with a cumulative variation of 53.054% according to the standard. Finally, based on the factor analysis combined with our research team’s professional experience, six items clustered together to form the first factor labelled ‘self-management’. Five items loaded onto the second factor labelled ‘medical consultation’, and the third factor is named as ‘life-style’ including four items (Table 3).

Convergent validity analysis revealed a moderate negative Pearson correlation between the MMAS-8 and the Mandarin TBQ total score (r = –0.450, P < 0.01), implying that convergent validity was satisfied.

Reliability

The Cronbach’s α for the total TBQ was 0.824 and the values of Cronbach’s α of the subscales ranged from 0.724 to 0.774, indicating adequate internal consistency and scale robustness. Test–retest reliability was supported by an ICC of 0.840 for the total and the ICC of the subscales ranged from 0.837 to 0.955.

Discussion

Stroke, as a chronic illness, is now in epidemic proportions globally, but the burden of treatment on patients remains poorly understood. Effective clinical intervention should assess treatment burden and adopt strategies to minimize the ‘work’ of being a patient, otherwise, it can lead to treatment non-adherence, side effects, worsening or recurrence of symptoms, poor quality of life, and ineffective use of definite health resources (35,36). A reliable assessment tool to determine the extent of the treatment burden and tailored education intervention projects for stroke patients are extremely urgent under the current situation.

This is the first study to have applied the TBQ for a psychometric study to the treatment burden of stroke patients. The Mandarin version of the TBQ showed good construct and convergent validity with a three correlated-factor structure. In this study, the Mandarin TBQ was developed based on several forward and backward translations with cross-cultural validation according to international studies.

Obviously, the item-total correlations for the 15 items (Pearson’s coefficients ranging from 0.403 to 0.638) were all higher than the minimum recommended level of 0.4 (37), indicating that each item assessed the same concept or construct.

In our study, the Mandarin TBQ showed good content validity: the I-CVI values of all items ranged from 0.83 to 1, and the value of the S-CVI was 0.967. As a criterion, the I-CVI ≥ 0.78 indicated that the item has an acceptable content validity (38,39).

In terms of construct validity, EFA results showed that the Mandarin TBQ can be characterized by three factors, explaining 53.054% of the total variance. To be specific, the first domain, labelled ‘Self-management’, included six items (items 1–6) and reflected the burden from the frequency, taste, and side effects of medication, as well as the burden from the maintenance of effective self-management. The second factor, ‘Medical consultation’, included items 7–11 and reflected the burden from the health care reservation, communication with medical staffs, and even huge expenditures of health care. The third factor, ‘lifestyle’, is composed of 4 items (items 12, 13, 14 and 15) and mainly assesses the burden from health maintenance, such as proper diet and regular exercise in daily life. Although the three-factor structure of Mandarin TBQ is inconsistent with the original English version, researchers from different countries and cultures have also suggested a similar structure among patients with chronic diseases (18,21). However, we noticed that there were a few differences in the attribution of items. The possible reason was the differences in the country’s economic level, cultural background, and the current policies on health care for patients with chronic diseases.

Convergent validity is the extent to which a new instrument is correlated with other comparators based on prior hypotheses. In this study, the Pearson correlation coefficients of MMAS-8 and the Mandarin TBQ was used to evaluate the convergent validity. As expected, an inverse association was observed between the TBQ global score and medication adherence (r = –0.450, P < 0.01), indicating that Mandarin TBQ has a satisfactory convergent validity. This is in line with the findings of the treatment burden among patients with multiple sclerosis by Ysrraelit et al. (18).

High internal consistency of Mandarin TBQ was also observed (with Cronbach’s α of 0.824), which is slightly lower than the Cantonese version (0.842) and the original English version (0.90) (16,21). Test–retest reliability refers to the temporal stability of an instrument over time. As with other researches, the test–retest reliability of both overall (with ICC of 0.840) and three factors (ranged from 0.837 to 0.955) after a 14 days interval was satisfactory.

In the current study, a high completion rate of 96.3% can be observed, which indicated that the Mandarin version has good cultural acceptance among stroke patients. Some limitations also should be addressed. First, due to limited resources, the data for our study were obtained from only one hospital in Tianjin. Although this hospital treats stroke patients from around the nation, the generalizability of the findings may not be generalized to individuals living in other geographic regions and populations. Future researchers need larger and more diverse populations to test the finding. Second, criterion validity was not evaluated in our study because no gold standard measure of treatment burden currently exists worldwide. Third, our study was a cross-sectional study, and longitudinal research is needed to validate the sensitivity and responsiveness of the Mandarin version of the TBQ. Fourth, compared to other chronic diseases, stroke has severe physical and psychological disability, more dimensions or specific items may be needed to explore treatment burden from particular sequelae of stroke. Finally, the use of drug profile, drug burden, post disease outcomes can also be changed by experiences of stroke, furthermore comprehensive clinical information is needed to explore treatment burden among patients with stroke.

Conclusion

The Mandarin version of TBQ is valid and reliable. Health care providers can use it in clinical practice and research to identify stroke patients who are overwhelmed by the complexity of their treatment, which may benefit patient-centred doctor-patient interactions in Mainland China.

Acknowledgements

We thank Professor Viet-Thi Tran, who provided the treatment burden questionnaire for our research. We also express our thanks to all bilingual and medical experts for assistance in the process of translation.

Declarations

Funding: This research received no specific grant from any funding agency.

Ethical approval: This study was approved by the Institutional Review Board of Tianjin Huanhu Hospital (number 2020–27).

Conflict of interest: All authors declare that they have no competing interests.

References

Author notes