The association between patient activation and self‐care practices: A cross‐sectional study of an Australian population with comorbid diabetes and chronic kidney disease

Abstract Objective This study aimed to examine the association between performance of self‐care activities and patient or disease factors as well as patient activation levels in patients with diabetes and chronic kidney disease (CKD) in Australia. Methods A cross‐sectional study was conducted among adults with diabetes and CKD (eGFR <60 mL/min/1.73m2) who were recruited from renal and diabetes clinics of four tertiary hospitals in Australia. Demographic and clinical data were collected, as well as responses to the Patient Activation Measure (PAM) and the Summary of Diabetes Self‐Care Activities (SDSCA) scale. Regression analyses were performed to determine the relationship between activation and performance of self‐care activities. Results A total of 317 patients (70% men) with a mean age of 66.9 (SD=11.0) years participated. The mean (SD) PAM and composite SDSCA scores were 57.6 (15.5) % (range 0‐100) and 37.3 (11.2) (range 0‐70), respectively. Younger age, being male, advanced stages of CKD and shorter duration of diabetes were associated with lower scores in one or more self‐care components. Patient activation was positively associated with the composite SDSCA score, and in particular the domains of general diet and blood sugar checking (P<.05), but not specific diet, exercising and foot checking. Conclusion In people with diabetes and CKD, a high level of patient activation was positively associated with a higher overall level of self‐care. Our results identify subgroups of people who may benefit from tailored interventions to further improve their health outcomes. Further prospective studies are warranted to confirm present findings.


| INTRODUCTION
Patient activation specifies the level of patients' involvement with their health care and refers to the extent to which they have the knowledge, motivation, belief, confidence and skills to manage chronic disease, access health care and to partner with health-care providers for disease management. [1][2][3] Patient activation is an important concept in chronic disease management driven by a person-centred approach and chronic care models. 1,4 Higher levels of patient activation are associated with better patient outcomes compared to lower levels of activation, in chronic diseases. 1,3,[5][6][7] Individuals with low activation are more likely to be hospitalized, 8,9 have a longer length of stay in hospital, 10 have greater health-care costs, 11 are less likely to participate in self-management activities such as blood pressure monitoring 12 and have worst care experiences 13 compared to those with higher activation levels.
Patient self-management is a patient's ability to participate in the management of symptoms, treatment and the physical, psychological and lifestyle consequences associated with chronic disease. 14 There is growing evidence to suggest an association between patient activation levels and performance of self-care activities for single chronic diseases including human immunodeficiency virus, 15 congestive heart failure, 16 schizophrenia 17 and diabetes. 18,19 Patient activation predicts a variety of behaviours such as engaging in exercises, healthy diet and other disease-specific self-care and consumeristic behaviours. 6,12 However, studies are inconsistent in demonstrating an association between patient activation and self-management for patients with diabetes and other long-term diseases including chronic obstructive pulmonary disease (COPD), depression and musculoskeletal pain. 5,12,18,20 The PAM has previously been used as a screening tool for tailoring self-management interventions or as a quality indicator for with the composite SDSCA score, and in particular the domains of general diet and blood sugar checking (P<.05), but not specific diet, exercising and foot checking.

Conclusion:
In people with diabetes and CKD, a high level of patient activation was positively associated with a higher overall level of self-care. Our results identify subgroups of people who may benefit from tailored interventions to further improve their health outcomes. Further prospective studies are warranted to confirm present findings.

K E Y W O R D S
chronic kidney disease, diabetes, patient activation, self-care, self-management F I G U R E 1 Patient inclusion flow diagram delivery of care. 21 In the UK, one health service has redesigned the diabetes review process according to the individual's level of activation. 22 Additionally, tailored coaching following activation assessment has resulted in improved clinical indicators and decreased health-care utilization in patients with asthma, coronary artery disease, congestive heart failure, COPD and diabetes. 23 Similarly, tailored care according to activation levels has been used to empower patients to ask questions during clinical reviews. 24 There is a knowledge gap regarding the relationship between patient activation and self-management in instances of comorbidity and multimorbidity such as diabetes and chronic kidney disease (CKD).
This gap is important given that multimorbidity is increasing globally 25,26 and CKD commonly coexists with diabetes 27 and is complex to manage. Moreover, greater understanding of how patient activation may influence performance of self-care activities will be important in the design of interventions to increase self-management.
The purpose of this study was to examine the association between performance of self-care activities and patient or disease factors as well as patient activation levels in patients with diabetes and CKD.

| Study design and participants
The design and recruitment of participants for this study have been described in great detail previously. 28

| Patient activation
The American version of the PAM-13 30 was used to evaluate the patients' level of involvement in their health care. The PAM scale examines participants' beliefs, knowledge and confidence in performing several self-management activities and then yields a score based on patients' answers to the 13 questions. 34 There are four alternative responses to each of the 13 items namely, "disagree strongly, disagree, agree and agree strongly" and fifth response option "not applicable" (N/A) was available for all items.
The authors used a standardized spreadsheet provided by Insignia Health ® to calculate the PAM score. 35 We excluded participants who responded to less than 7 items or if all questions were answered with "disagree strongly" or "agree strongly." The mean PAM score was then calculated on all items leaving out the ones thought to be nonapplicable by the participants. The raw mean score was converted into a standardized activation score ranging from 0 to 100 creating the PAM scores which were classified into the four levels of activation:

| Outcomes
Self-management was evaluated by the SDSCA questionnaire, 31 a self-report measure of how often participants perform diabetes self-care activities. The SDSCA questionnaire has been utilized in several studies and settings and is deemed to be reliable, valid and sensitive [36][37][38] in evaluating diabetes self-management in adults. This T A B L E 2 (Continued) study used a version of the SDSCA questionnaire that comprised of items assessing five domains of diabetes self-management which are "general diet (2 items), specific diet (2 items), exercise (2 items), blood glucose testing (2 items) and foot care (2 items)". 31 The medication self-management component was excluded based on previous reports of its "ceiling effects and lack of variability among participants". 31 The smoking self-management component was also excluded because smoking behaviour was relevant to smokers only.

| Statistical analysis
Results are presented as mean and standard deviation (SD) and median and interquartile range (IQR) for normal and non-normally

| Patient characteristics
A total of 3028 patients were screened and 305 were included in the analyses after exclusion of nine patients who had their eGFR misclassified (>60 mL/min/m 2 ) and three patients who had incomplete PAM data ( Figure 1). There were no differences in age, gender and stage of kidney disease between responders and non-responders (Table S1)

| Association between self-care activities and patient or disease factors
Patient factors associated with self-care activities are shown in Table 2. On multivariable analysis, younger age was associated with lower scores in the general diet domain (all P value <.05). Male patients had lower scores in the blood sugar checking domain where they scored 1.6 points less than female patients. A shorter duration of diabetes was associated with lower composite scores, and with lower scores in the blood sugar checking and foot checking domains (all P<.05). (Figure 2). Patients with stage 5 kidney disease scored 1 point less than patients with stage 3a disease in the exercising domains. No association was found between socio-economic status and the composite score or any specific self-care domain.

| Association between self-care activities and patient activation
With decreasing patient activation level, the mean scores for the com-  Table 2).
In univariable and multivariable analyses, the level of patient activation was positively associated with the composite self-care score and the domains of general diet and blood sugar checking (all P<.05) but not the domains of specific diet, exercising and foot checking (Table 2). When patient activation was included in the models as a continuous variable, the results remained similar (data not shown).

| DISCUSSION
In our study, among patients with comorbid diabetes and CKD, we have demonstrated an association between patient activation and diabetes self-care activities. A higher patient activation level was associated with a higher overall self-care score. However, this association was not observed for all specific self-care domains; only for general diet and blood sugar checking. Additionally, different patient and disease characteristics were associated with diabetes self-care: younger age and male gender were associated with less home blood glucose monitoring, more severe CKD was associated with less foot checking and exercising, and a shorter duration of diabetes was associated with lower overall self-care score as well as less blood sugar checking and foot checking.
In patients with comorbid diabetes and CKD, higher patient activation levels were associated with higher composite self-care scores.
Previous studies have only examined this association for single chronic diseases, such as diabetes. 18 In patients with diabetes, the relationship is inconsistent, with some studies showing a positive association between patient activation and self-care activities 5,39 and others showing no association. 18 In patients with CKD, this association has not been explored. Our study adds to the literature by showing that in the setting of multimorbidity, the association is positive and independent of certain potential confounding patient or disease factors such as age, gender and disease duration.
Interestingly, the association between patient activation levels and diabetes self-care was not observed for all specific self-care domains. While there was a positive association between general diet and blood sugar checking, there was no association between patient activation levels and specific diet, exercising and foot checking domains. This suggests that an activated patient may not necessarily or automatically participate in all self-care activities-they not only need to have knowledge, motivation and skills to self-manage, but they also need to have the physical and financial ability to self-manage across all domains of diabetes self-management. A possible reason for the lack of association between PAM and exercise and foot checking is that both these activities require a certain degree of physical fitness and ability, which is compromised in patients with diabetes and CKD due to comorbidity. 40,41 Similarly, a lack of association between PAM and a specific diet could be that the specific diabetes diet may be financially prohibitive. 42,43 These results highlight the importance of addressing all self-care domains to improve self-management for patients with comorbid diabetes and CKD across all spectrums of activation.
We found an association between younger ages and lower selfcare scores in the domain of general diet independent of patient activation. The explanation is likely to be multifactorial, but we hypothesize that younger patients may be less motivated to self-manage compared to older patients, as risk perception is altered in younger populations, especially in males 44,45 and they have competitive priorities that take precedence such as socializing and work commitments. 46 Lack of knowledge may also contribute but less so than other factors given that younger patients are reported to have greater diabetes knowledge than older patients. 47 Additionally, we found that a shorter duration of diabetes was associated with lower self-care scores. Previous studies among patients with diabetes have not been consistent with some reporting an association between lower self-care scores with a shorter duration of diabetes, 48,49 while others reported an association between lower self-care scores and longer duration of diabetes. 50,51 In patients with comorbid diabetes and CKD, we found a shorter duration of diabetes to be associated with lower self-care scores. This suggests that patients with a shorter duration of diabetes may not be exposed to sufficient diabetes education or have not yet mastered self-management skills, and should be targeted by interventions to improve self-management such as tailored Diabetes Self-Management Education and support 52 .
Alternatively, participants with a longer duration of diabetes are likely to be older and may have some physical limitations such that they receive more attention and social support to improve their ability to self-manage. 53 More advanced CKD was associated with lower scores in the selfcare domains of exercising and foot checking. Exercising and foot checking require a certain level of mobility and physical fitness such The other strengths include the inclusion of several demographic and clinical variables as potential predictors for diabetes self-management behaviour, and the use of valid and reliable tools to measure patient activation 30 and diabetes self-management. 31 Additionally, the study population was drawn from multiple hospitals across Australia, increasing generalizability of our findings. Potential limitations are due to the cross-sectional nature of the study design, which did not allow us to track patient activation patterns over time. Assessment of patient activation over time permits an early identification of patients in whom a change in activation levels may flag a change in health status. Moreover, longitudinal PAM data can be used to develop risk prediction models that predict adverse patient outcomes. 56 Another apparent limitation was the modest response rate of 38.5%, which is, however, comparable to other studies in people with diabetes. 18, 57 We did not collect data on some factors such as depression and health literacy, which have been found to be associated with patient activation in different population groups. 58,59 In addition, our sample of participants who attend hospital may be a biased group from the aspect of utilizers of the service.
Our findings have important implications for practice and future research. First, targeted multifactorial risk reduction interventions focusing on subgroups of patients identified in this study, who are likely to perform poorly in self-care activities, may improve health outcomes. There is evidence that such interventions could be delivered optimally through collaborative care, 60 a key feature of combined diabetes kidney specialist clinics, which often have a multidisciplinary team. 61 Second, we have shown that highly activated patients are more likely to participate in self-care activities than those with low activation levels. Additionally, assessment of patient activation in this patient group, which is already suffering a double burden of chronic disease, 62,63 ensures that resources are directed to those who need them most, thereby improving on resource utilization and reduction in health inequalities. Our study, being of an exploratory nature, opens up opportunities for future research, which should include welldesigned and disease-specific longitudinal studies to validate and extend our findings.
In patients with comorbid diabetes and CKD, although a high level of patient activation in self-care is associated with a high level of patient self-management in general, this is not the case across all individual domains of diabetes self-care. Patient age, gender, duration of diabetes and stage of CKD may also influence patient selfmanagement in comorbid diabetes and CKD.

ACKNOWLEDGEMENT
We acknowledge S Chaviaras, D Giannopoulos, R McGrath and S Coggan for help in study conduct.