Type 2 diabetes mellitus (DM) is a prevalent and increasingly common disease worldwide.1 It is estimated to affect 10% of the Hong Kong (HK) population (approximately 700 000 people).2 Achieving satisfactory DM control during the early disease course can reduce DM-induced microvascular and macrovascular complications (ie the ‘legacy effect’).3 4 These benefits were maintained in patients in a tight DM-control group even though their glycosylated haemoglobin (HbA1c) level became similar to those in the control group after the end of the United Kingdom Prospective Diabetes Study.4 It was proposed that a ‘reverse legacy effect’ also persists: “intensive glycaemic intervention started late in the natural course of diabetes seems disappointingly ineffective in limiting cardiovascular events”.5 6 Very tight control may even result in mortality.7 8 Therefore, achieving tight HbA1c control early via lifestyle changes and the use of medications including insulin is important.

Because of the progressive nature of DM, most patients eventually require insulin.9 Despite robust evidence of the benefits of early strict HbA1c control, patients often delay insulin initiation and titration. In a UK study, 50% of patients with DM delayed insulin initiation despite suboptimal control for 5 years, regardless of the presence of complications.10 Their reluctance to initiate insulin use10 11 12 and its subsequent titration13 is known as ‘psychological insulin resistance’ (PIR). The prevalence of PIR has been estimated to be higher in Singapore (70.6%)11 than in western countries (approximately 20%-40%).12 A HK survey of 97 participants found a similarly high prevalence of PIR (72.1%).14 Previous studies conducted in western countries have identified several factors that can lead to PIR.11 12 13 These reasons might differ in Asian countries, however.15 16 Recently, a local primary care research group developed a scale, Chinese Attitudes to Starting Insulin questionnaire, to identify barriers to insulin initiation in insulin-naïve patients with DM.16 These investigators found that Asian patients might be more affected by the availability of social support and that cultural differences might also play a role. For example, Chinese patients are more likely to combine western medical treatments with traditional Chinese medicine17 and might believe that hypoglycaemic agents cause renal toxicity.18

Doctors, particularly primary care physicians, can be insensitive to patients’ psychological needs; physicians often fail to recognise psychological needs19 and might incorrectly identify the reasons for a patient’s PIR.20 21 Identifying one’s psychological needs might be hindered in HK due to short consultation times (lasting an average of 5-7 minutes per consultation). A limited number of longer sessions may be offered to DM patients with difficult glycaemic control, but the time limit would be 10 to 14 minutes. Therefore, a quick tool to help identify PIR and its underlying causes might help general practice physicians optimise care for their patients with DM.12 The Insulin Treatment Appraisal Scale (ITAS) was developed for this purpose.22 The Chinese version of the ITAS (C-ITAS) was validated in Taiwan,23 and has been used in Taiwan15 to investigate the underlying causes of PIR. Validating C-ITAS scores might enable direct comparisons of data between local and international studies. The C-ITAS might also be used to help local primary care clinicians identify PIR and offer appropriate counselling. The ITAS is sensitive to changes in PIR throughout the course of DM.24

This study is the first to be conducted in HK to examine the prevalence of PIR and the validity and reliability of the C-ITAS in our local population.

This research has been approved by the Research Ethics Committee at Kowloon West Cluster, Hospital Authority.

Participants were recruited from a government-funded primary care general out-patient clinic in HK from July to September 2013. Written consent was obtained when the participants were approached by the research assistant. The investigator’s contact information was given to each participant if they had concerns after the administration of the questionnaire. Patients who fulfilled the following criteria were recruited: (1) diagnosed with type 2 DM as defined by the World Health Organization25 for ≥6 months; (2) aged 30 years or above; (3) of Chinese ethnicity; (4) able to communicate effectively in Cantonese or Mandarin; and (5) had the mental capacity to provide informed written consent. The exclusion criteria were severe sensory deficits, severe mental illness (eg dementia, psychosis, or mental retardation), or any other health condition that compromised the ability to comprehend and complete the questionnaire. The required sample size was calculated from the estimated prevalence rate of PIR in the primary care setting. To achieve a 95% confidence interval with a margin of error of 5% and an estimated 70% prevalence of PIR among patients with DM in public primary care,11 14 the required sample size was estimated to be 312 patients. To compensate for the predicted 20% refusal rate, at least 390 patients were recruited.

A list of DM patients who would attend the clinic the next day was obtained daily. From that list, 40 patients were randomly selected by computer (25 in the morning and 15 in the afternoon). A reminder was set in the clinical computer system such that clinic staff were alerted once the patient attended his or her appointment. The procedure was repeated until the number of patients recruited exceeded 390, which was checked at clinic closing time.

Patients were encouraged to complete the questionnaire unaided because the C-ITAS is self-administered. Because the majority of patients who attend public primary care clinics are of lower socio-economic status and education level, those who had difficulty completing the questionnaire were assisted by research assistants who were trained by the principal investigator.

Each patient was asked whether he or she was willing to have insulin started or titrated upon his or her case doctor’s suggestion. The response options included “strongly unwilling”, “unwilling”, “might consider it”, “willing”, and “very willing”. Demographic data were collected, and clinical data (eg the presence of DM complications, insulin use, and control of DM and lipid levels) were retrieved from a computer database.

The ITAS is a 20-item instrument that contains 16 negative and four positive statements that appraise insulin treatment. Each statement is ranked using a 5-point Likert-type scale from 1 to 5. Positive scores are reversed to allow for summation. The total possible score ranges from 0 to 80. A higher score signifies a more negative appraisal of insulin. The ITAS was developed for clinical use to measure PIR.22 No cut-off score is used to diagnose PIR. Of those who completed the clinical interview, 26 were selected for phone interview 2 to 4 weeks later to examine test-retest reliability. Because of the lack of a written language difference between Taiwan and HK, the validated C-ITAS was used with the permission of the Taiwan research group.

The C-ITAS was examined for its internal reliability using Cronbach’s alpha, the test-retest reliability was assessed using Pearson’s correlation of test scores and retest scores, and construct validity was assessed using an exploratory factor analysis (EFA) [using Oblimin rotation as this was used in the original development study of ITAS22]. Patients who answered “strongly unwilling” or “unwilling” to the question “Would you agree to start or titrate insulin treatment if advised by your case doctor?” were classified as having PIR. Descriptive statistics were used to describe the prevalence of PIR. Each C-ITAS item was dichotomised as “unwilling” (scores of 1 and 2) or “neutral/willing” (scores of 3 to 5); this dichotomy was created to assess the difference between patients with and without PIR. The responses of the patients with or without PIR were compared using a Chi squared test.

A total of 399 patients with DM were randomly selected from the clinical database and approached by the research team (Fig 1). Of them, 42 patients were excluded due to the following circumstances: two patients were incorrectly diagnosed with DM; 27 had severely impaired hearing not compensated for with the use of hearing aids; three spoke languages other than Cantonese or Mandarin; eight had severe psychiatric illness such as dementia, psychosis, or mental retardation; one left at the beginning of the interview when called into a consultation room; and one was excluded for checking all boxes of the questionnaire.

In addition to the insulin-naïve patients with DM who were recruited as outlined above, all of the current insulin users who were not interviewed during the above period (47 patients) were invited to participate in this study and were interviewed over the phone; of whom three were excluded for the following reasons: one could not speak Cantonese or Mandarin, one was out of HK during the interview period, and one questionnaire was invalid due to a missing subject case number.

The overall response rate was 89.8% (n=360): 89.5% (n=314) for the insulin-naïve patients and 92.0% (n=46) for the insulin users. Other demographic data are shown in Table 1.

A total of 12.8% (n=46/360) of participants were insulin users. Patients with HbA1c ≥7% (≥53 mmol/mol; 21.6%) were more likely to be on insulin than those with HbA1c <7% (<53 mmol/mol; 2.9%; P<0.001). The HbA1c level was not significantly associated with the presence of DM complications in the current study. Of all participants, 96.3% received DM complication screening within 2 years, which was a nurse-led clinical service to screen for DM complications and provide counselling.

Non-respondents were significantly older (mean age=72.32 vs 67.17 years, t test: P<0.001), less likely to agree to titrate insulin (for current insulin users), and less educated (91.7% educated up to primary level vs 68.9%; Chi squared test; P=0.004). The differences with regard to the other demographics, including DM complication rate, insulin use status, marriage, work, family income, and gender were not significant.

The prevalence of PIR was 44.9% (141/314; 95% confidence interval [CI], 39.4% to 50.4%) in insulin-naïve patients; in contrast, the PIR rate was 6.8% (3/44; 95% CI, -0.64% to 14.24%) in current insulin users.

The internal consistency of the C-ITAS questionnaire was high, with Cronbach’s alpha of 0.78. The original ITAS was designed to have 16 negative and four positive statements. Cronbach’s alpha was calculated separately for the negative and positive statements, yielding values of 0.812 and 0.738, respectively. Within the negative statement scale, removing two negatively stated questions individually, including Q1, “Insulin signifies failure with pre-insulin therapy”, and Q18, “Taking insulin causes family/friends to be more concerned” improved the overall Cronbach’s alpha to 0.819 and 0.825, respectively.

Of the 20 individual questions within the C-ITAS, answers to 17 questions were significantly different in the expected direction between patients with PIR and those without. Importantly, Q18, “Taking insulin causes family/friends to be more concerned” was originally designed to detect a negative view towards insulin use; however, more insulin-accepting patients agreed with the statement (Table 2).

The total C-ITAS scores, as described above, were higher among participants who refused insulin initiation (42.42 vs 35.78; t test, P<0.001). The test-retest reliability for each question ranged from 0.294 to 0.725, and 13 questions were significant (P<0.05). The test-retest reliability of the overall scores as defined above was 0.571 (P=0.002).

The EFA identified five factors with an eigenvalue of >1. Nonetheless, the scree plot correctly identified two factors within the questionnaire. When two factors were extracted using an Oblimin rotation, a few negative statements including Q18 were significantly associated with the other positive statements (Table 3). The three-, four-, and five-factor solutions were calculated as suggested by the eigenvalue, which did not provide better representation of the latent structure of ITAS.

In the EFA, the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.834 and Bartlett’s test of sphericity was significant (P<0.001), and signified adequate sample size for the test.

Because the participants were old and not well educated, difficulties in answering the C-ITAS were expected. This assumption was further supported by the fact that the non-respondents were less educated and were older than the respondents. Nevertheless a high proportion of participants (89.8%) were able to complete the entire questionnaire. Additional research might be necessary to assess the response rate if the questionnaire is self-administered because the staffing at our public out-patient clinics was limited. The use of ITAS might be limited if it cannot be self-administered because it was developed as a self-administered tool.

It is surprising that the prevalence of PIR was not as high as reported by previous studies.11 14 More than 50% of patients were willing to consider or accept insulin if suggested by their primary doctor. This finding might be because of differences in the patient cohorts or the improvements made to the PIR over the years due to patient education. Only 53 patients with DM out of the thousands of patients followed up in our clinic were started on insulin. Alternative reasons might explain the low rates of insulin use (eg physician beliefs and competencies regarding the use of insulin), and might merit additional research.

The C-ITAS was reliable because it yielded high Cronbach’s alpha scores (0.738-0.812) and correctly provided a higher score for patients who resisted insulin use. It identified many different attitudes towards insulin use; in the current study, answers to 17 out of 20 of the C-ITAS items significantly differed between patients who resisted insulin and those who did not, whereas a previous study showed that only four questions were able to make this distinction.12 This may be because individual patients had multiple concerns and many different attitudes towards insulin use.

Although the test-retest reliability value of all ITAS items was positive, the values were low, ranging from 0.294 to 0.725 for individual C-ITAS questions. In the present study, the C-ITAS was completed either via a personal interview with a research assistant or by self-administration. Retests were administered via telephone interviews by either the research assistant or the principal investigator. Therefore, the low test-retest reliability scores might be because of the different means of administration or due to the different interviewers. Conversely, this difference might reflect the actual low test-retest reliability of the current C-ITAS that requires additional validation.

Question 18, “Taking insulin causes family/friends to be more concerned”, merits additional discussion. Originally designed as a negative statement, it is expected that patient resistance to insulin would positively predict the score. The reverse was true, however, in the current study (Table 2). When the statement was reviewed by six family physicians and one psychiatrist, the word “concerned” (關心) was translated into a word in Chinese that can also mean “caring” (使用胰島素使家人和朋友對我更關心). It is likely that patients understood the question as, “Taking insulin causes my family and friends to be more caring toward me”. Because Q18 was meant to be a negative statement, it is more appropriate to translate its meaning to “worry”. This supposition is supported by both the Cronbach’s alpha analysis, in which exclusion of Q18 improved the value of Cronbach’s alpha, and the factor analysis, where Q18 was regarded as a factor with the other positive statements. The factor analysis did not show a two-factor structure within the ITAS, as in the previous study.22 As the factor analysis table notes (Table 3), when set as a two-factor construct, no trend can be drawn for these two groups. The factor analyses of the first study on the development of the ITAS22 and the validation study in Taiwan23 both showed a two-factor construct, with the two factors being positive statements and negative statements. This finding might reflect the previously noted translation problem; alternatively, our local community might have had a different set of causes for PIR. This finding suggests that a dialectic or cultural difference remains between HK and Taiwan, despite a shared written language.26 27 Additional validation of the C-ITAS in our local population is likely necessary.

The strengths of our study include its large sample size, the use of random sampling, and the high response rate. The use of an internationally validated questionnaire might aid comparison with results from other countries. The C-ITAS, however, might require additional validation as noted above.

The statement proposing the use of insulin to patients was hypothetical. For example, estimated PIR rates might be lower when patients perceive their disease as having deteriorated so that additional intervention is necessary.

This study was conducted in a major government-funded clinic in Hong Kong, and the demographics of the participants were more similar to those of other government clinics than to the general population (Fig 2). The extent to which the results can be generalised to other countries and to other social classes (eg wealthy patients attending private primary clinics) is not known.

A majority of the patients in the current study were insulin-naïve. Despite including all available insulin users in the clinic, the number of insulin users was small, and limits the potential applicability of this study’s results to secondary or tertiary care where many patients may be on insulin.

The study also did not distinguish between questionnaires that were completed with the help of research staff and those that were self-administered. The influence of different administration methods on the outcome has not been previously described. For example, when participants did not understand a statement, the trained research assistant may use her own words to elaborate and explain it to the participant and thus may alter the statement’s original sentence structure or intended meaning.

Another weakness was that data on macrovascular complications were not collected. Microvascular complications were well documented during the DM complication screening and were easily traceable. The tracing of macrovascular complications, however, was difficult because diagnostic coding needed to be entered or the complication needed to be mentioned in the latest case record by the respective doctors, and missed coding for macrovascular complications was not uncommon.

The prevalence of PIR was 44.9% in our population, which is less than that previously estimated. Tools such as the C-ITAS can improve physician’s understanding of patient views on insulin and might help physicians to appropriately counsel their patients. The C-ITAS may provide clues to patients’ knowledge about insulin use, eg the risk of hypoglycaemia or the side-effects of obesity. Despite good psychometric properties such as high internal consistency, there is a translation issue in at least one of the 20 statements. Health care professionals who wish to use the C-ITAS clinically should be aware of the instrument’s limitations.

The author expresses gratitude to Prof Sandra Chan for her teaching and guidance regarding this research; to Prof Samuel Wong for his kind and timely advice; and to Drs YK Yiu and SN Fu and the Department of Family Medicine, Kowloon West Cluster, HK for their research support. The author would like to thank Ms Man-ping Chang and her team for the development of the C-ITAS and for allowing the use of the C-ITAS in the current study.

The author has disclosed no conflicts of interest.

1. International Diabetes Federation. IDF diabetes atlas update 2012. Available from: http://www.idf.org/sites/default/files/EN_6E_Atlas_Full_0.pdf. Accessed 1 Apr 2013.

2. Hong Kong Department of Health. Hong Kong reference framework for diabetes care for adults in primary care settings. Available from: http://www.pco.gov.hk/english/resource/professionals_diabetes_pdf.html. Accessed 1 Apr 2013.

3. Genuth S, Eastman R, Kahn R, et al. Implications of the United Kingdom prospective diabetes study. Diabetes Care 2003;26 Suppl 1:S28-32. Crossref

4. Davis TM, Colagiuri S, United Kingdom Prospective Diabetes Study. The continuing legacy of the United Kingdom Prospective Diabetes Study. Med J Aust 2004;180:104-5.

5. ADVANCE Collaborative Group, Patel A, MacMahon S, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med 2008;358:2560-72. Crossref

6. Riddle MC, Yuen KC. Reevaluating goals of insulin therapy: perspectives from large clinical trials. Endocrinol Metab Clin North Am 2012;41:41-56. Crossref

7. Action to Control Cardiovascular Risk in Diabetes Study Group, Gerstein HC, Miller ME, et al. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med 2008;358:2545-59. Crossref

8. ACCORD Study Group, Gerstein HC, Miller ME, et al. Long-term effects of intensive glucose lowering on cardiovascular outcomes. N Engl J Med 2011;364:818-28. Crossref

9. Turner RC, Cull CA, Frighi V, Holman RR. Glycemic control with diet, sulfonylurea, metformin, or insulin in patients with type 2 diabetes mellitus: progressive requirement for multiple therapies (UKPDS 49). UK Prospective Diabetes Study (UKPDS) Group. JAMA 1999;281:2005-12. Crossref

10. Rubino A, McQuay LJ, Gough SC, Kvasz M, Tennis P. Delayed initiation of subcutaneous insulin therapy after failure of oral glucose-lowering agents in patients with type 2 diabetes: a population-based analysis in the UK. Diabet Med 2007;24:1412-8. Crossref

11. Wong S, Lee J, Ko Y, Chong MF, Lam CK, Tang WE. Perceptions of insulin therapy amongst Asian patients with diabetes in Singapore. Diabet Med 2011;28:206-11. Crossref

12. Woudenberg YJ, Lucas C, Latour C, Scholte op Reimer WJ. Acceptance of insulin therapy: a long shot? Psychological insulin resistance in primary care. Diabet Med 2012;29:796-802. Crossref

13. Jenkins N, Hallowell N, Farmer AJ, Holman RR, Lawton J. Participants’ experiences of intensifying insulin therapy during the Treating to Target in Type 2 diabetes (4-T) trial: qualitative interview study. Diabet Med 2011;28:543-8. Crossref

14. Yiu MP, Cheung KL, Chan KW, et al. A questionnaire study to analyze the reasons of insulin refusal of DM patients on maximum dose of oral hypoglycemic agents (OHA) among 3 GOPCs in Kowloon West Cluster. 2010. Available from: http://www.ha.org.hk/haconvention/hac2010/proceedings/pdf/Poster/spp-p5-38.pdf. Accessed 1 Apr 2013.

15. Chen CC, Chang MP, Hsieh MH, Huang CY, Liao LN, Li TC. Evaluation of perception of insulin therapy among Chinese patients with type 2 diabetes mellitus. Diabetes Metab 2011;37:389-94. Crossref

16. Fu SN, Chin WY, Wong CK, et al. Development and validation of the Chinese Attitudes to Starting Insulin questionnaire (Ch-ASIQ) for primary care patients with type 2 diabetes. PLoS One 2013;8:e78933. Crossref

17. Ma GX. Between two worlds: the use of traditional and western health services by Chinese immigrants. J Community Health 1999;24:421-37. Crossref

18. Lai WA, Lew-Ting CY, Chie WC. How diabetic patients think about and manage their illness in Taiwan. Diabet Med 2005;22:286-92. Crossref

19. Martin A, Rief W, Klaiberg A, Braehler E. Validity of the Brief Patient Health Questionnaire Mood Scale (PHQ-9) in the general population. Gen Hosp Psychiatry 2006;28:71-7. Crossref

20. Peyrot M, Rubin RR, Khunti K. Addressing barriers to initiation of insulin in patients with type 2 diabetes. Prim Care Diabetes 2010;4 Suppl 1:S11-8. Crossref

21. Brod M, Kongsø JH, Lessard S, Christensen TL. Psychological insulin resistance: patient beliefs and implications for diabetes management. Qual Life Res 2009;18:23-32. Crossref

22. Snoek FJ, Skovlund SE, Pouwer F. Development and validation of the insulin treatment appraisal scale (ITAS) in patients with type 2 diabetes. Health Qual Life Outcomes 2007;5:69. Crossref

23. Chang MP, Huang CY, Li TC, Liao LN, Chen CC. Validation of the Chinese version of the insulin treatment appraisal scale. J Diabetes Investig 2010;1(Suppl 1):88.

24. Hermanns N, Mahr M, Kulzer B, Skovlund SE, Haak T. Barriers towards insulin therapy in type 2 diabetic patients: results of an observational longitudinal study. Health Qual Life Outcomes 2010;8:113. Crossref

25. World Health Organization. About diabetes 2013. Available from: http://www.who.int/diabetes/action_online/basics/en/index1.html. Accessed 1 Feb 2013.

26. Chow KM, Chan CW, Choi KC, et al. Psychometric properties of the Chinese version of Sexual Function After Gynecologic Illness Scale (SFAGIS). Support Care Cancer 2013;21:3079-84. Crossref

27. Tang DY, Liu AC, Leung MH, Siu BW. Antisocial Personality Disorder Subscale (Chinese version) of the Structured Clinical Interview for the DSM-IV Axis II disorders: validation study in Cantonese-speaking Hong Kong Chinese. East Asian Arch Psychiatry 2013;23:37-44.