Determinants of patient activation and its association with cardiovascular disease risk in chronic kidney disease: A cross‐sectional study

Abstract Background Patient activation describes the knowledge, skills and confidence in managing one's own health. Promoting patient activation is being prioritized to reduce costs and adverse outcomes such as cardiovascular disease (CVD). The increasing prevalence of chronic kidney disease (CKD) presents a need to understand the characteristics that influence patient activation and the effect on health outcomes. Design Cross‐sectional study. Setting and participants Patients with non‐dialysis CKD recruited from 14 sites (general nephrology and primary care) in England, UK. Outcome measures Patient activation was measured using the PAM‐13. Demographic and health‐related variables, self‐reported symptom burden, health‐related quality of life (HRQOL), socioeconomic status (SES), were assessed as determinants of patient activation. Major CVD risk factors included hypertension, dyslipidaemia, obesity and hyperkalaemia. Results 743 patients were included (eGFR: 32.3 (SD17.1) mL/min/1.73 m2, age 67.8 (SD13.9) years, 68% male). The mean PAM score was 55.1 (SD14.4)/100. Most patients (60%) had low activation. Those with low activation were older (P<.001), had lower eGFR (P = .004), greater number of comorbidities (P = .026) and lower haemoglobin (P = .025). Patients with low activation had a 17% greater number of CVD risk factors (P < .001). Risk factors in those with low activation were being older (P < .001) and having diabetes (P < .001). Conclusion This study showed that only a minority of CKD patients are activated for self‐management. Our findings help better understand the level of activation in these patients, particularly older individuals with multimorbidity, and further the knowledge regarding the characteristics that influence activation. Patient or Public Contribution Patients were involved in the design of main study.


| INTRODUC TI ON
In a growing multi-morbid population, managing the demand for health services is a challenge faced by physicians and policymakers.
In the UK, 70% of National Health Service (NHS) expenditure is spent on patients with long-term conditions (LTCs) such as chronic kidney disease (CKD). 1 With <1% of time spent in contact with health-care professionals, many patients are expected to self-manage their condition 1 particularly those with less advanced disease. The importance of self-management is recognized in the NHS Universal Personalised Care and Long-Term plans, 2 and self-management is integral to any model of care for LTCs. 3,4 'Patient activation' describes the knowledge, skills and confidence a person has in managing their own health. 5 The 'Patient Activation Measure' (PAM) is the most widely used instrument for measuring patient activation 6 and was piloted in the NHS through the UK Renal Registry (UKRR). Increased patient activation is associated with improved health outcomes in many LTCs including premature mortality and hospitalization. 7,8 Patient activation is closely related to the engagement of preventive health behaviours with empirical studies indicating activated patients are more likely to attend screenings, check-ups and immunizations, as well as engage in healthy behaviours such as eating a balanced diet. 9,10 Few studies have administered the PAM to patients with kidney disease, and as such, information regarding the determinants and outcomes in this population is sparse. 11 Nonetheless, promoting patient activation in kidney disease care is increasingly being prioritized and has recently emerged as central to legislative policy in the United States 11 and UK. 12 Studies in CKD have largely taken place in Europe, 13 the USA, [14][15][16][17] Australia 18-21 and Asia 22 and often report associations between patient activation and clinical characteristics. In general, lower activation levels are associated with older age, receiving in-center haemodialysis, poorer perceptions of health-related quality of life (HRQOL), higher decisional conflict with respect to modality choice and lower medication adherence (summarized in Nair and Cavanuagh 11 ). In the UK, Hamilton et al 23 found higher patient activation was associated with better medication use, a younger age of renal replacement therapy (RRT) in 590 dialysis and kidney transplant recipients. Data collected as part of the UKRR 12 found higher patient activation levels in younger individuals and those with a kidney transplant, although no associations between patient activation and clinical biomarkers were found. Nonetheless, these findings are limited as data were almost exclusively collected in those on RRT with eGFR data missing for all of the non-dialysis group.
The increasing prevalence of CKD presents an urgent need to understand the role of patient activation. 15 Further evidence is also needed to show higher levels of patient activation are associated with clinically meaningful outcomes. 11 Cardiovascular disease (CVD) remains a leading cause of premature mortality and morbidity in CKD, 24 and reducing CVD remains a mainstay of conventional CKD management. Patient activation may be a fundamental component that can mediate the presence of CVD risk factors.
The aim of this study was to (a) explore the prevalence of patient activation across a range of adult CKD participants, (b) identify factors associated with patient activation and (c) explore the association with CVD risk factors.

| Study design and setting
This analysis consists of PAM data taken from the multi-center ob- Participants completed a self-administered paper survey made up of different questionnaires designed to assess physical function and activity, symptoms and diet. Participants were recruited from general nephrology clinics and from GP practices between July 2018 and February 2020 across 14 sites (all sites recruited from general nephrology clinics, whilst one site (University Hospitals of Leicester NHS Trust) also recruited from 8 local GP practices) in England, UK (supplementary material 1)). Participants were included if they had been (a) diagnosed with a kidney condition (CKD 1-5 not requiring dialysis), (b) were aged ≥ 18 years and (c) were able to provide informed consent. Participants with or receiving RRT (dialysis, transplant) were excluded. The study was granted national research ethical approval (18/EM/0117). All patients provided informed written consent, and the study was conducted in accordance with the Declaration of Helsinki.

| Patient activation measure-13 (PAM-13)
The PAM-13 is a validated tool of 13 questions which assesses a patient's knowledge, skills and confidence in managing their own health. The PAM-13 has demonstrated good internal consistency as well as adequate reliability and validity. 5,6 Answers are weighted and combined to provide a score on a scale from 0 to 100. A score is generated where participants have answered ≥10 questions. The PAM allows respondents to be categorized into one of four levels with lower levels indicating low activation and higher levels indicating high activation: Level 1 (<47.0), disengagement and disbelief about one's own role in self-management; Level 2 (47.1-55.1), increasing awareness, confidence and knowledge in self-management tasks; Level 3 (55.2-67), readiness and taking action; and Level 4 (>67.1), sustainment.

| Demographic and clinical variables
Demographic (age, sex, ethnicity) and health-related variables (comorbidities, smoking status, body mass index (BMI)) were selfreported. Upon receipt of the survey, participant's most recent clinical data, including recent renal function (estimated glomerular filtration rate, eGFR), cause of disease, haemoglobin, potassium, albumin, C-reactive protein (CRP), lipid profile, and blood pressure, were extracted from the medical records by the research team at each site.

| Symptom burden
Symptom burden was self-reported using the Kidney Symptom Questionnaire, 25 a validated questionnaire surveying 13 commonly reported symptoms. Patients rated the frequency and importance of each symptom on a 5-point Likert scale (from 0 (never/not intrusive) to 4 (every day/extremely intrusive)). Total symptom burden was determined by combining the frequency (/52) and importance (/52) score to give a total score /104.

| Quality of life
Self-reported HRQOL was assessed using the SF-12 questionnaire, a validated assessment of quality of life that provides a physical (PCS) and mental (MCS) component score. The final score of the SF-12 ranges from 0-100, where a higher score indicated a better HRQOL.
A score below or above 50 indicates a, respectively, worse or better quality of life than a pre-defined general population reference group.
The SF-12 has good agreement with other measures of HRQOL including the KDQOL-36. 26

| Cardiorespiratory fitness
Cardiorespiratory fitness was estimated from the Duke Activity Status Index (DASI) questionnaire, a brief 12-item questionnaire assessing the capability to complete activities of daily living (ADLs). Each activity is weighted with a metabolic equivalent of tasks value which is summed to produce a score (0 to 58.2).
This was transformed into VO 2peak using a previously published equation. 27

| Dietary intake assessment
Self-reported dietary intake was assessed using the European Prospective Investigation of Cancer in Norfolk Food Frequency Questionnaire (FFQ). It has been validated against diet diaries and biological markers and widely used in the UK general population. 28 The FFQ measured participant's food intake during the previous year. For each food item, participants were asked to indicate their usual consumption from nine categories ranging from never or <1/ month to ≥6 times per day. The FFQ was used to extract fruit and vegetable intake (g/day) and alcohol intake (g/day).

| Socioeconomic status (SES)
Socioeconomic status (SES) was measured using the Index of Multiple Deprivation (IMD). It is comprised of seven distinct domains of deprivation which, when combined and appropriately weighted, form the IMD score. Income (22.5%) and employment (22.5%) make up the two largest components of an area's IMD score. An individual's IMD was calculated from their postcode. An IMD ranges from 1 (most deprived area) to 32,844 (least deprived area). IMD deciles are calculated by ranking all 32,844 neighbourhoods in England from most to least deprived, before dividing them into 10 groups; decile 1 is the most deprived, decile 10 is the least deprived.

| Statistical analysis
Descriptive and frequency data are presented as mean (standard deviation (SD)) or number (percentage). Statistical testing was conducted using IBM SPSS V26, and a P-value of <.050 was considered significant. eGFR was calculated using the EPI formula. 29 Activation levels were dichotomized into low activation (Levels 1 and 2) and high activation (Levels 3 and 4) as reported previously. 19 Differences between those with low and high activation were explored using general univariate models or chi-square testing. A multivariable linear model was used to assess the association between PAM-13 score and the following variables: age, eGFR, sex, BMI, ethnicity, IMD, number of comorbidities, symptom burden and number of medications. These variables were pragmatically chosen a priori based on hypothesized associations but also to maintain the largest sample size. Binominal logistic regression, adjusted for eGFR, was used to explore the association between different CVD risk factors and having low activation. Missing data were handled using pairwise deletion to maximize all data available per analysis. Due to missing data, CRP and dyslipidaemia were excluded to maintain a sample of 340. Data are shown as odds ratio with 95% confidence intervals. Missing data for each variable can be found in Table 1. Table 1 were assessed using chi-square testing or univariate regression with disease stage as the independent variable; significant ß was identified as P < .050, as appropriate.

| Summary of participant characteristics
743 patients were included, the majority (92%) in Stage 3 and 4-5, with a mean eGFR of 32.3 ( SD 17.1) mL/min/1.73 m 2 . The mean age for the cohort was 67.8 ( SD 13.9) years, and 68% were male. Most patients were White British (94%). The population were largely overweight with a BMI of 29.4 ( SD 7.4) kg/m 2 and were comorbid with a mean additional number of comorbidities of 3.3 ( SD 1.7) (shown in Table 1).

| Patient activation status across disease stages
The distribution of PAM-13 scores across disease stage is shown in     Table 3 shows differences in characteristics between those with low and high PAM-13 levels. Low activated patients were older (P < .001), had a lower eGFR (P = .004), had a greater number of comorbidities (P = .026) and had lower haemoglobin levels (P = .025).

| Determinants of patient activation status
There was no difference in other variables including IMD score. Low activation was associated with a 22% reduction in cardiorespiratory fitness compared to those with higher activation (P < .001). Those with low activation had a 17% lower PCS (P < .001) and 6% lower

| Association between patient activation and CVD risk factors
Low activated patients had a 17% greater number of CVD risk factors than those with high activation (P < .001) with a greater number of patients demonstrating ≥5 CVD risk factors (P = .007). Significant risk factors found in those with low activation were being older (P < .001) and having diabetes (P < .001) ( Table 5). Results of the logistic regression model showed that those older than 71 years were 3 times likely to have low activation (OR = 3.295, P < .001) and those with diabetes were 1.7 times more likely to have low activation (OR = 1.735, P = .049) (Figure 3). Cardiovascular disease remains a leading cause of mortality and morbidity in CKD. 24 We found patients with low activation had a greater number of CVD risk factors. A key CVD risk factor is age.

| D ISCUSS I ON
F I G U R E 1 Frequency distribution curves for each disease stage. PAM, Patient Activation Measure. Data shown as relative frequency as a percentage (%) The evidence between age and patient activation is inconsistent, although it has been hypothesized that younger patients may have poorer coping strategies which lead to low activation. 33   is associated with fatigue, and it may be that these patients find self-management tasks, like exercise, difficult. However, the UKRR data showed no association between patient activation and calcium, phosphorus, or haemoglobin, and in our data, the difference in haemoglobin between those with high/low activation was 5 g/L and should be interpreted with caution. Socioeconomic status is often considered to be an important factor in health-care engagement.
Studies have shown patient activation is only moderately correlated with SES 41 and that education and income account for only <5 to 6% of the variation in patient activation. 9 In contrast to what we expected, in a multivariate model, patients with low activation had a higher IMD, indicative of greater SES. However, differences were small and no difference in IMD was observed between the low and high activation groups.
We found patients with low activation had lower cardiorespiratory fitness and HRQOL. This supports a plethora of research in older adults and LTCs. 4,[7][8][9] In CKD, data from the UKRR 12 found those who  Contrary to what one may expect, we found no difference in fruit and vegetable intake between those with high and low activation.
Despite healthy eating being mentioned in two questions in the PAM-13, there is limited research investigating the role of patient activation on dietary intake. One previous study found that intervention-derived increases in activation failed to change participants self-reported adherence to a low-fat diet. 42 Our findings may be explained by the choice of self-reported FFQ, although it may be the PAM-13 is insensitive to detect such differences given that healthy eating is combined with other lifestyle behaviours (e.g. exercising) in each question.
We found the variables included explained 27% of the PAM-13 score. The remaining variance may be explained by other factors influencing activation for self-management, for example self-efficacy, knowledge or the support from health-care professionals. Previous research has shown greater activation is associated with greater knowledge of condition. 43 The PAM-13 includes items focussing on self-efficacy, and whilst self-efficacy measurement was not included in this study, in Social Cognitive Theory self-efficacy is an important factor in self-management skills and behavioural change. 44

| Strengths and limitations
Our study has several important strengths including the comprehensive range of biological and non-biological variables analysed in a large population. The study was conducted across multiple sites from both secondary and primary care increasing the generalizability of our results. We used validated instruments for measuring HRQOL and patient activation. The limitations include the crosssectional design which does not allow assessment of temporal effects or the potential for reverse causality. Longitudinal studies are needed to better understand the effects over time of factors influencing patient activation. The use of self-reported questionnaires may introduce misclassification due to socially desirable responses.
Whilst the PAM-13 is widely used in LTCs, it is limited by its selfassessment of a patient's perceived ability to manage their own care, rather than the direct measurement of self-management behaviour itself. Furthermore, patient activation in the setting of kidney disease may require knowledge and skills that are CKD-specific; whether the PAM-13 is an appropriate measure in kidney disease is unknown. 11

| Clinical recommendations
Interest in the PAM has been growing in nephrology, and further knowledge of characteristics associated with activation for self- interventions (e.g. a low-intensity web-based self-management intervention may demonstrate good outcomes with highly activated patients but may be ineffective in those with lower activation where more intensive intervention is needed). 4 In kidney disease, how to maintain or remediate decline in patient activation is unknown. 11 It is important to note that whilst in England, UK, the licence cost associated with the PAM-13 is funded by NHS England and Improvement as part of a national agreement, this may not be the same in other health-care organizations across the world.

| CON CLUS ION
This study showed only a minority of CKD patients are activated metric has the potential to target individuals at greatest risk, 1 additional evidence is needed to better understand the role of patient activation on patients living with kidney disease.

ACK N OWLED G M ENTS
This report is independent research supported by the National

Institute for Health Research Leicester Biomedical Research
Centre. The views expressed are those of the author(s) and not necessarily those of the Stoneygate Trust, NHS, National Institute for Health Research Leicester BRC or the Department of Health.
We are grateful to all the participants who took part in the studies and other researchers in the different sites whom assisted with data collection.

CO N FLI C T O F I NTE R E S T
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article. The authors declare that they have no competing interests. All authors have full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors gave final approval.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.