Patient activation among patients with type 2 diabetes and its association with perception of care, perceived autonomy support, sociodemographic and clinical characteristics Muzlifa Mohamed Yusoff, Syahnaz Mohd Hashim, Norfazilah Ahmad ABSTRACT Background: Patient activation is an important aspect of self-management in type 2 diabetes (T2DM) and a key component of the Chronic Care Model, whereby patients should play an active role in their care. Past studies have yet to prove the exact factors influencing patient activation. Limited studies have examined patients’ perceptions of care and whether healthcare providers are autonomy-supportive. This study primarily focused on determining factors associated with patient activation, including sociodemographic and clinical characteristics, perception of care, and perceived autonomy support. Methods: A cross-sectional study was conducted among T2DM patients at the public primary care clinic, Kepala Batas Health Clinic, Penang, the northern part of Malaysia, between 2nd December 2024 and 30th April 2025. Participants were sampled through a systematic sampling method and were given a self-administered questionnaire, comprising sociodemographic and clinical characteristics, Patient Activation Measure, Patient Assessment of Chronic Illness Care and Health Care Climate Questionnaire. Results: A total of 450 patients participated, with a response rate of 85.3%. Many were Malay (92.9%), had lower income (99.0%), attained a secondary level of education (62.0%) and were on oral hypoglycaemic agents only (OHA) (66.0%). The mean patient activation score was 59.54 (SD 14.58), and 66.7% were at a high level of activation. Multiple linear regression revealed that factors significantly associated with patient activation were male (β = 1.984, [95% CI 0.629, 3.339], p = 0.004), age (β = −0.089 [95% CI: −0.161, −0.017], p = 0.015), HbA1c (β = −6.661 [95% CI: −7.022, −6.300), p < 0.001] and on OHA only (β = −1.460 [95% CI: −2.902, −0.019), p = 0.047]. Conclusions: Patients tend to have lower activation when they are older, have higher HbA1c, and are on oral hypoglycaemic agents only. However, male patients exhibit higher activation. Interventions should focus on providing targeted, tailored support to those at risk of lower activation, to enhance their engagement in diabetes self-management and improve health outcomes. Introduction Type 2 Diabetes Mellitus (T2DM) is among the top ten major causes of morbidities and mortalities all over the world [ 1 ]. The rising number of T2DM cases globally has substantially strained the healthcare system [ 2 , 3 ]. In 2020, 445 million cases were reported and it is likely to increase to 730 million by 2050 [ 4 ]. The prevalence of diabetes in South Asia has increased from 11.29% in 2000–2004 to 22.30% in 2020–2024 [ 5 ], and in Southeast Asia, it is expected that 10.8% of adults, or one in ten, will have diabetes by 2050 [ 6 ]. Malaysia is following a similar trend, with 11.2% of adults affected in 2011 and 15.6% in 2023 [ 7 ]. Among Malaysian states, Penang has a significant number of adults with T2DM, with 106,402 individuals affected [ 7 ]. Despite extensive efforts by healthcare providers, only 34.38% of patients achieve good glycaemic control [ 8 ]. For many decades, there has been a call for a more efficient healthcare system to improve outcomes for chronic conditions, such as diabetes. Various models of care have been proposed and the most prominent is the CCM [ 9 , 10 ]. CCM identifies the vital components of a healthcare system that help to improve the outcomes of chronic illnesses [ 9 , 10 ], including Patient Activation, Goal setting, Problem-solving/Contextual Counselling, Delivery System Design/Decision Support, and Follow-up/Coordination [ 9 , 10 ]. A crucial component of CCM is patient activation, in which patients act as effective health managers, and to fulfil this role, they need knowledge, skills, and confidence [ 11 , 12 ]. The ability of patients to become health managers, i.e., ‘patient activation’ is assessed through a scale, the Patient Activation Measure (PAM) [ 11 ]. The total score of this scale can be computed and further categorised into four activation levels (levels 1-4) [ 11 ]. According to Hibbard et al. (2004), levels 1 and 2 are considered low activation, as individuals in these levels have inadequate knowledge and confidence to manage their illness. When patients reach level 3, they begin to self-manage, and at level 4, they can maintain self-management. Levels 3 and 4 are considered high activation, meaning patients are competent to become their own health managers [ 11 ]. A body of evidence showed that when patients are highly activated, their clinical outcomes, including glycaemic, lipid, and blood pressure control, improve [ 13 – 15 ]. This can enhance patients’ quality of life, reduce hospitalisation and healthcare costs [ 15 ]. Past research across various populations found that more than half of patients with T2DM are at a high level of activation, with a mean patient score of 54–58 [ 16 – 20 ]. Similar research has not yet been conducted in the local setting, as the concept of patient activation has not been widely investigated in Malaysia. Nevertheless, a local study among patients with metabolic syndrome found that the mean patient activation score was 58.9, and most were at a high activation level [ 21 ]. Previous studies have also highlighted that factors associated with patient activation include younger age [ 22 , 23 ], higher education level [ 24 , 25 ], adequate literacy skills [ 26 , 27 ], employment [ 21 , 24 , 28 ], and less financial distress [ 24 , 29 , 30 ]. Apart from these, having good knowledge [ 24 , 31 ], positive coping [ 32 , 33 ], depression [ 34 ], and good glycaemic control [ 15 , 25 ] are also significant factors. However, the outcomes of the quantitative studies showed that Nagelkerke R ², a measure of variance (16% to 27%), was not large [ 17 , 34 ], suggesting that other factors remain unexplored. It is possible that factors contributing to patient activation may not be solely based on patients’ characteristics but also on the care they receive and the attitudes of healthcare providers [ 35 ]. This view has long been expressed through the Chronic Care Model (CCM), which emphasises that patient-centred care should be established at every level of primary care, with an effective delivery system, well-organised care, and a focus on self-management and decision support [ 9 , 10 ]. The CCM was then examined through research, and the evidence is sufficient to demonstrate that high-quality care and positive attitudes among healthcare providers influence patients’ engagement and ability to manage their health [ 28 , 36 ]. In assessing patient’s perception of care, the Patient Assessment of Chronic Illness Care (PACIC) scale has been widely utilised [ 37 , 38 ], including for diabetes care [ 39 ]. Besides the quality of care, promoting patients’ autonomy has been shown to motivate them to engage in diabetes care [ 36 , 40 , 41 ]. Perceived autonomy support is vital in healthcare, as it promotes adherence to diabetes self-management [ 41 - 43 ]. This concurs with Self-Determination Theory, which posits that perceived autonomy is a fundamental element that motivates people to act [ 44 ]. The theory explains that people can thrive when they believe those in positions of authority, such as healthcare providers, allow them to make decisions about matters that are important to them, i.e., when those in authority are autonomy-supportive [ 44 ]. The concept of perceived autonomy is measured using the Health Care Climate Questionnaire (HCCQ) [ 40 ]. To date, there appears to be limited research examining the perception of care and perceived autonomy support as factors contributing to patient activation. Furthermore, efforts to assess patient activation among patients with T2DM in the local setting remain limited. Therefore, the general objective of this study was to determine the factors associated with patient activation among patients with T2DM, including sociodemographic and clinical characteristics, perceptions of care, and perceived autonomy support . Our secondary objective was to assess the proportions of patients with high and low levels of activation, and the mean scores for perception of care and perceived autonomy support. As Malaysia is currently experiencing a surge in T2DM cases, we hope this research will contribute to efforts to identify the most effective strategies to facilitate patient activation in diabetes self-management. Before developing an intervention, information is needed to identify potential factors, particularly from patients’ perspectives on their care and healthcare providers’ attitudes. This information is crucial, as high-quality, patient-centred diabetes care can only be delivered by a well-trained, efficient healthcare team. Methods Study design, setting and population This cross-sectional study was conducted at Kepala Batas Health Clinic, a public primary care clinic, in the northern state of Peninsular Malaysia, Penang, from 2 nd December 2024–30 th April 2025. The inclusion criteria for this study were (i) patients with T2DM aged 18–70 years, (ii) able to read and write in Malay or English, (iii) diagnosed with T2DM for more than a year [ 36 ]. The criteria for having T2DM for more than a year were consistent with the finding that patient activation improved at one-year follow-up, when patients had sufficient time to engage with their healthcare providers [ 36 ]. The exclusion criteria were participants who had (i) an emergency condition during the visit to the clinic, (ii) cognitive or visual impairment, (iii) psychiatric illness, and (iv) physical impairment. The sample size was calculated based on each study’s objectives. The largest sample size obtained was 357, based on the single mean formula using Z-score of 1.96 with 95% confidence interval, a margin of error of 5% [ 45 ], and a standard deviation of 0.54 for the perception of care from a previous study [ 46 ]. Considering a 20% non-response rate, 448 samples were required for this study. Data collection Approval to conduct the study was obtained from the Medical and Research Ethics Committee, Faculty of Medicine UKM, Universiti Kebangsaan Malaysia (FF-2024–330) and the Ministry of Health Malaysia (NMRR ID-24–01860-YRL (IIR). Permission to perform the study was also obtained from the Penang State Health Department, the Seberang Perai Utara District Health Office, and the Family Medicine Specialist in charge. A systematic sampling method was employed to recruit the participants. Patients with T2DM registered at the clinic’s Non-Communicable Disease Unit were identified and numbered. The principal researcher would approach every second patient and screen them for eligibility criteria. When the participants fulfilled the criteria, they were invited to participate in the study. Participants were given an information sheet, and, once they agreed to participate, they were required to sign a written consent form. Confidentiality was maintained by not including the participant’s identity. All data was entered into a password-protected computer, with the access data document restricted to the primary investigator. A self-administered questionnaire was then distributed. Study instrument and variables The study used a self-administered questionnaire in both Malay and English, consisting of five sections: A to E. Section A assessed sociodemographic characteristics ( age, gender, ethnicity, marital status, educational level, employment status, and monthly household income ). Section B measured clinical characteristics (duration of T2DM, current treatment, presence of comorbidities, number of visits to doctors per year for diabetes treatment, and recent HbA1c) . Of note, the HbA1c result was obtained from the medical records. Section C measured patients’ perceptions of care received over the past six months using the short version of the PACIC scale [ 38 ]. The brief version of this questionnaire consists of 11 items, whereby the patients were required to rate each item on an 11-point scale: none (score of 0%) to always (score of 100%) . The overall score ranged from 0 to 1100, with higher scores reflecting a better perception of care received over the past six months. The PACIC scoring method was based on the original author’s work [ 38 ]. Each participant’s average score should be obtained by adding all responses and dividing by 11, since there are 11 items in the PACIC scale. The total average score is then computed to yield the PACIC mean score [ 38 ]. Section D measured the extent to which doctors have been autonomy-supportive over the last six months using the short version of the HCCQ scale [ 47 ]. The brief version of this scale consists of six items, with options on a seven-point Likert Scale: 1 (not at all true) to 7 (very true). The overall score ranged from 6 to 42, with higher scores reflecting a greater autonomy support over the past six months. According to the original author, each participant’s average score should be calculated by summing all responses and dividing by six, since the HCCQ has six items [ 47 ]. The total average score is then computed to yield the mean HCCQ score [ 47 , 48 ]. Section E measured patient activation using the PAM scale. The scale consists of 13 items, with options on a 4-point Likert Scale: 1 (strongly disagree) to 4 (strongly agree) [ 49 ]. The overall scores were calculated using a 0-to-100-point algorithm scale [ 50 ]. Levels 1 and 2 signified a low level of activation, whereas Levels 3 and 4 signified a high level of activation. The validated PAM-13 questionnaire is available in English [ 50 ] and Malay [ 51 ]. Permission had been obtained to use both versions. The independent variables of this study are i) sociodemographic characteristics, ii) clinical characteristics, iii) mean score of perception of care measured by the PACIC scale and iv) mean score of perceived autonomy support measured by the HCCQ scale. The dependent variable is the patient activation score using the PAM scale. Translation and validation of PACIC and HCCQ scales The PACIC and HCCQ scales were originally developed in English and need to be translated into Malay. The translation and validation process begins with a panel of experts ( a family medicine specialist and a public health medicine specialist ) reviewing the original scale to assess its content for the local context. Next, two linguistic experts translated the English version into Malay, producing two translated Malay versions, which were then harmonised. The harmonised Malay version was subsequently translated into English by two other linguistic experts. The translated English version was then compared with the original one. The panel of experts reviewed each step of the forward and backward translations and agreed that the harmonised Malay version has a meaning similar to that of the original English. Permission was obtained from the original authors for the English version of the questionnaires and the translation of the Malay version. Before the actual data collection, face validity was assessed in five patients with T2DM for both the English and Malay versions of the PACIC and HCCQ scales. All five patients provided feedback that both versions were well understood. Subsequently, a pilot study involving 60 T2DM patients was performed. Thirty patients completed the Malay version, while another 30 completed the English version of both the PACIC and HCCQ scales. Construct validity for the PACIC and HCCQ was assessed using independent t-test ( t = −1.95, df = 58, p value = 0.056 and t = −1.54, df = 58, p value = 0.130), respectively. The reliability of both versions was determined using Cronbach’s alpha, a measure of internal consistency for both the PACIC and HCCQ scales. For the PACIC scale, the Cronbach’s alpha is 0.77 for Malay and 0.811 for English. For the HCCQ scale, the Cronbach’s alpha is 0.80 for Malay, and 0.87 for the English version. These results showed that both versions of the PACIC and HCCQ scales have good internal consistency. Statistical analysis All data collected were analysed using the IBM SPSS software version 27. Categorical data were described as frequency (n) and percentage (%). Numerical data were described as the mean with standard deviation (SD) or the median with interquartile range (IQR). Statistical significance was set at p < 0.05. Simple linear regression was performed first to examine the association between patient activation and each of the independent variables (age, gender, ethnicity, marital status, education level, employment status, monthly income, duration of diabetes, type of treatment, HbA1c, comorbidity status, number of visits, PACIC and HCCQ) . Multiple linear regression was then performed, including all independent variables. Backward, forward, and stepwise methods were employed to obtain the best model. The model was reasonably well. There were no interactions among the independent variables. No multicollinearity was detected and variance inflation factors are below 10. Model assumptions were fulfilled. Results Sociodemographic and clinical characteristics A total of 533 T2DM patients were invited to participate, of whom 455 agreed. Hence, the response rate was 85.3%. Five incomplete questionnaires needed to be excluded. Thus, only 450 questionnaires were included in the analysis. Fig 1 details the study flow chart. Table 1 shows that the participants’ mean age was 53.84 years (SD 9.37). The majority were Malay (92.9%), married (91.0%), and had a low monthly income (99.0%). More than half were female (57.8%), had attained up to secondary education (62.0%), and were unemployed (58.7%). Many had other comorbidities (90.0%) and 66.0% took oral hypoglycaemic agents (OHA) only. The mean duration of diabetes was 7.57 years (SD 5.53), with an average of 3.50 (SD 1.45) clinic visits for diabetes. The mean HbA1c was 8.15% (SD 1.90). 10.1371/journal.pone.0348840.t001 Table 1 Sociodemographic and clinical characteristics (N = 450). Characteristics n % Mean (SD) Age (years) 53.84 (9.37) Age (categories) 18-29 5 1.1 30-44 69 15.3 45-59 233 51.8 60-70 143 31.8 Gender Male 190 42.2 Female 260 57.8 Ethnicity Malay 418 92.9 Chinese 8 1.8 Indian 24 5.3 Marital status Single 15 3.3 Married 410 91.0 Divorced/Widow 25 5.6 Education level 1 Primary 34 7.6 Secondary 280 62.0 Tertiary 136 30.2 Employment status Unemployed 264 58.7 Employed 186 41.3 Monthly income (MYR) 2 Low 445 99.0 High 5 1.1 Duration of DM (years) 7.57 (5.53) Type of treatment Lifestyle 16 3.6 OHA 3 only 298 66.0 OHA 3 + Insulin 136 30.2 HbA1c (%) 8.15 (1.90) Comorbidity status 4 No 45 10.0 Yes 405 90.0 No. of visits 3.50 (1.45) 1 Primary: no formal education & primary school, Secondary: secondary school, Tertiary: diploma, college, matriculation, degree, masters, PhD. 2 Low: those with monthly income of