Subjects and Study Design
A cross-sectional survey using self-administered questionnaires and a retrospective health record review was conducted at two polyclinics (public primary care centres) in the northeast region of Singapore. Patients were recruited on a case-encounter basis at the patient monitoring stations within the polyclinic during their routine medical reviews.
Patients aged 40 to 79 years with T2DM on follow-up for more than 1 year, had a latest HbA1c of ≥8% and who could understand and comply with written and/or verbal instructions were eligible to participate in the study. Patients were excluded if they were on treatment for psychological disorders, mentally incapacitated or pregnant.
The questionnaire was printed in the English language, and was tested and revised in a pilot study. Interviewers were trained to use a common script to field questions, so as to ensure standardisation. For participants who spoke only Mandarin or Malay, the questionnaire was translated to the participants by interviewers who were native Mandarin and Malay speakers respectively.
Clinic staff referred patients fulfilling the eligibility criteria to a study team member, who explained the study and obtained their consent in a quiet, closed room. Following consent, the subjects filled out the questionnaire and returned it back to the study team member to check for completeness. If clarifications or translation of the questionnaire were required, the study team member would read from the standardised script based on the study protocol.
Sample Size Calculation
The sample size was calculated to be 335, assuming a similar DRD prevalence of 32% as in a local tertiary clinic [19], to achieve a confidence level of 95%. This number was multiplied by 10% to allow for drop-outs and missing data, and rounded up to give a final count of 370.
Study Questionnaire
Data on demographics was collected directly from subjects using the questionnaire. The information included gender, ethnicity, marital status, education level, employment and living arrangements. For financial status, the Community Health Assist Scheme (CHAS), a local tiered health financing support programme, was used to identify lower- to middle-income households, while receipt of Medifund assistance (another health financing support scheme targeted at people from low socioeconomic status) was used to identify needy subjects who have difficulties with their medical bills despite government subsidies. Self-reported information on smoking history, exercise frequency, hypoglycaemic frequency and history of previous consults with a dietician and/or counsellor/psychologist was obtained from the participants.
Questions on other psychological and social factors such as diabetes-related family arguments and financial concerns were included to identify psychosocial issues. These questions were adapted from the second Diabetes Attitudes, Wishes and Needs (DAWN-2) study [9]. The latter was a global study to assess diabetes care and self-management of PWDs, family members and healthcare professionals, and to identify determinants of effective treatment and self-management [20]. Responses were recorded on a 5-point Likert scale, 5 representing “strongly agree” and 1 representing “strongly disagree”.
Scales and Tools
(i) Problem Area In Diabetes (PAID) Scale
DRD was measured using the PAID scale (Annex 1). This instrument is a 20-item questionnaire, where items are rated on a 5-point Likert scale, with 0 representing “no problem” and 4 representing “a serious problem”. The scores are summed up and multiplied by 1.25 to give a total score from 0 to 100. A PAID score of ≥40 suggests distress at a level warranting clinical attention [1]. The psychometric properties of the English version [19] as well as the Chinese translation [21] have been studied and found to be valid and reliable for use in Singapore.
In this report, the terms “DRD” and “elevated distress” will be used interchangeably to refer to a PAID score of ≥40.
(ii) Patient Health Questionnaire (PHQ-9) and Generalised Anxiety Disorder (GAD-7)
The PHQ-9 is a 9-item self-report tool used to screen for depression. Subjects are asked to rate the frequency of symptoms over the past 2 weeks on a 4-point Likert scale, with 0 representing “not at all” and 3 representing “nearly every day”. The possible score range is from 0 to 27. Scores of 5, 10, 15 and 20 represents cut-off points for mild, moderate, moderately severe and severe depression respectively [22]. The PHQ-9 has been shown to be valid and reliable (Cronbach’s α = 0.87) for use in Singapore [23].
Similarly, the GAD-7 is a 7-item self-report tool designed to screen for anxiety, where subjects rate the frequency of symptoms over the past 2 weeks on the same 4-point scale. The total score ranges from 0 to 21. Scores of 5, 10 and 15 represents cut-off points for mild, moderate and severe anxiety respectively. The scale is valid and reliable to measuring anxiety in the general population based on overseas studies [24].
A cut-off score of 10 and above was used to define cases of clinical significance for both the PHQ-9 and GAD-7.
(iii) 5-Level EuroQol 5 Dimensions Scale (EQ-5D-5L)
The EQ-5D-5L is a simple self-report survey developed by the EuroQoL Group to measure health-related quality of life [25]. It consists of two parts. The first measures health status in five domains (mobility, self-care, usual activities, pain/discomfort and anxiety/depression) across five response levels. The results may be presented as a descriptive profile or a single index value. Value sets for the latter are available for each country [26].
The second part records the respondent’s self-rated health on a visual analogue scale (EQ-VAS), ranging from 0, which represents the “worst imaginable health state”, to 100, which represents the “best imaginable health state”.
The original instrument utilises a three-level response, with the English, Chinese and Malay versions previously validated for use in Singapore [27, 28, 29]. The newer EQ-5D-5L, which saw the introduction of two new response levels to improve sensitivity and reduce ceiling effects, has been found to be more discriminative compared to the three-level version in PWDs in Singapore [30].
Electronic Medical Records Review
Clinical data were extracted from the electronic medical records. The data included birth year, latest body mass index (BMI) and HbA1c reading, co-morbidities and diabetes-related complications, number of doctors and nurses visits, duration of T2DM and the number of long term medications.
Statistical Analyses
Data analysis was performed using SPSS Version 27.0. Statistical significance was defined as p<0.05. Chi-square test or Fisher’s exact test was used to analyse categorical variables. For continuous variables that are normally distributed, independent t-test and one-way ANOVA were used for two groups and three or more groups respectively. Continuous variables that are non-parametric were analysed with the Mann-Whitney U and Kruskal-Wallis tests for two groups and three or more groups respectively. Potential factors with p-values less than 0.2 were included in the multiple logistic regression to obtain the adjusted odds ratio (AOR).