Self-management program for patients with chronic kidney disease (SMP-CKD) in Southern China: protocol for an ambispective cohort study

Chronic kidney disease (CKD) is a serious global public health problem. It is reported a prevalence rate of 9.1% in global and 10.8% in Mainland China adult population [1, 2]. CKD can cause numerous complications, high morbidity, increased mortality and heavy economic burden especially progress to the end-stage kidney disease (ESKD) [3, 4]. Blood pressure control with angiotensin-converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARB) is considered the main pharmic way to prevent CKD progression [5]. However, the traditional treatment method involves many challenges, such as insufficient knowledge acquired by patients, low participation in disease treatment decision-making, and low compliance in the treatment [6]. Hence, finding a new strategy for slow CKD progression is urgent.

Self-management, defined as the active management by individuals of their treatment, symptoms, lifestyle and the physical and psychological consequences inherent in living with a chronic condition, is an established treatment for managing CKD in recent years [7‐12]. It has contributed to a paradigm shift from a paternalistic model to a more equitable and collaborative model between nephrologists and patients [13, 14]. Dietary adjustment is an important component of the self-management intervention due to disordered nutritional status and commonly appeared protein-energy wasting condition in the CKD population [15]. A scoping review revealed that diet/nutrition intervention accounts for over 60% of all list self-management topics [16]. A nutrition intervention, together with patient’s education, lifestyle change, risk factors control, psychological support, pharmacist medication review, etc. is recommended as preferred non-pharmacological alternative strategies for managing CKD patients [5].

Studies showed this emerging strategy with the function to relieve renal-related symptoms, enhance patient-centeredness and decrease patients’ concerns with depression, anxiety, self-perceived burdens [5, 15, 17‐20]. However, the evidence of self-management for renal outcomes is contradictory. A randomized controlled trial (RCT) in Taiwan showed self-management may slow the reduction of estimated glomerular filtration rate (eGFR) in late-stage CKD patients within 12 months follow up, it suggested that self-management may become an effective strategy for managing CKD patients, but systematic review and meta-analysis in RCT have shown that although self-management interventions can lower 24-h urinary protein excretion, improve self-care activities and systolic blood pressure than usual care, it did not provide additional benefits for eGFR and other renal outcomes in non-dialysis patients [17‐20]. While considering the median follow-up time of these studies is 13.44 months, in such a short follow-up period, it may be difficult to distinguish the real effects of self-management [20]. Furthermore, self-management is not easy to evaluate, especially in the long-term study. Adherence to the intervention plays a crucial role in the effectiveness of nonpharmacologic treatment strategies [21]. But self-management has low adherence rates, and long-term persistence is difficult [22]. During the study period, most patients’ clinical status and their participation in self-management changed dynamically, which can lead to inaccurate measures of interventions and get biased research conclusions. Therefore, its effect on thwarting the progression of CKD and other outcomes still needs further research.

What’s more, self-management interventions can be seen as complex interventions, it means not only number of and interactions between components within the interventions, but also the complexity of how the intervention works [23]. However, many studies give only a conceptual or general definition of self-management interventions, or no definition at all [18‐20, 24]. When the variety among self-management interventions is not taken into account and no clear operational definition is posited, this might lead to incorrect conclusions [25, 26]. Hence, a theory-based model is preferred, it can causally link behavioral determinants, through behavior, to physiological and biochemical variables, and health outcomes [27]. This theoretical basis approach ensures interventions have a greater chance of being effective, and the reasons for these effects can be deduced [28]. However, existing research showed only 20% of studies were designed based on a theoretical framework. More knowledge about how self-management will work for CKD is still need to be studied [16].

Considering the problems and gaps above, Self-Management Program for Patients with Chronic Kidney Disease Cohort (SMP-CKD cohort), a theory-based intervention cohort study alongside a process evaluation aiming to improve the lives of patients with CKD by slowing progression and improving overall Quality of Life (QoL) will be performed. In this protocol, we present the rationale and design of this ongoing hospital-based cohort study. The objective of this cohort include i) to evaluate the long-term effectiveness of the theory-based self-management program in the Chinese CKD population; ii) to assess its longitudinal relationship to surrogate markers and renal outcomes, and; iii) to explore the working mechanism of this complex intervention.

We selected the occurrence of composite endpoints as the primary study outcomes to calculate the sample size. Previous studies have reported that the composite endpoints rate range from 18.1–126/1,000 person-years [46, 47]. Based on these studies and our previous data, we estimated that the incidence rate of a composite endpoint in our cohort would be 60/1,000 person-years. There are about 30 variables that may affect prognosis, and there should be at least 10 endpoints per variable. Thus, we anticipate that recruiting 1,200 participants with 300 events over a 5-year follow-up will ensure adequate statistical power.

Categorical demographic and other baseline characteristics variables will be summarized by frequencies and percentages. Continuous variables will be summarized by both mean and standard deviation for data with normal distributions or median and inter-quartile range for non-normally distributed data. Either a Chi-square or Fisher’s exact test will be used to compare categorical variables between groups, while Mann–Whitney U-test or Student’s t-test will be used for continuous variables, if necessary.

The primary objective of this cohort is to evaluate the effectiveness of the self-management program, we plan to perform three separate analyses for the primary outcome. First, the incidence rate of composite endpoint events will be analyzed by the Chi-squared test. Second, time to occurrence of composite clinical endpoints will be analyzed using log rank test. The hazard ratio (HR) with 95% CI will be estimated by using the Cox proportional hazards model or competitive risk model. Third, growth mixture models and group-based trajectory models (GBTM) will be used to analyze the trajectories of exposure and outcome variables [48]. Covariates for the models will be selected based on prior knowledge and published papers. Logistic regression or Poisson regression will be used to analyze categorical outcomes. Liner mixed-effect model will be applied to the longitudinal data, such as the absolute change, and slope of eGFR and other secondary outcomes.

Even with some retention strategies, considerable missing data is expected over 5-years follow-up. The proportion of missing data will be summarized in each group and at each visit point. If there is less than 20% of data missing for the covariate data, we will perform a complete case analysis. If there is more than 20% of missing data, we will perform Little’s test and use multiple imputations under the assumption of missing is at random. Subgroup analyses will be conducted across baseline eGFR levels, demographic status and other factors, if possible. To test the credibility and robustness of the findings, several sensitivity analyses will be performed, including E-value and other advanced models [49]. Effect modification will be investigated using stratified analyses and formal tests of interaction. Mediation will be analyzed by adding the hypothesized mediators, one by one, to the confounder-adjusted models to study the extent to which they explain any association found.

All analysis will be performed in PASW 18.0 (SPSS Inc. IBM Corporation, Armonk, New York, USA) and STATA SE/15.0 (Stata Corp. College Station, Texas, USA).

To the best of our knowledge, this is the first theory-based self-management cohort conducted in Mainland China for CKD patients. This paper presents the protocol for the design and process evaluation of this self-management cohort. We believe there are several elements in the design and implementation of this study that allow advantages for quality data to inform the management of CKD in China.

Previous studies have regarded self-management as a plausible way to slow CKD progression [20]. However, due to the short follow-up time, their results have certain limitations. This is because self-management adherence is high in these studies, but patients experience ‘active’ self-management initially whereas self-management behavior is neglected as time progresses [50]. Meanwhile, applying self-management to CKD management needs flexible appointments and re-planning options between nephrologists and patients which is more suitable to achieve by cohort research. Therefore, it is worthwhile to study the effect of self-management through a long-term follow-up in a real-world medical environment.

However, it’s not easy to maintain patients’ compliance in the long-term follow-up study. Results from other studies showed theory-based intervention can improve medication and self-management adherence in other chronic conditions [51, 52]. Guidelines also recommend the explicit use of behavior change for addressing lifestyle risk factors when designing and reporting interventions for patients with CKD [53, 54]. In this cohort, behavior change is based on the SCT framework [55]. The application of this theory has shown positive effects on health behaviors in many chronic disease settings such as diabetes, heart disease and neurological disorders [56, 57]. According to this theory, this cohort study will determine the specific items to personalize educational content and behavior change strategies based on psychosocial determinants of self-efficacy, self-regulation, skill mastery, and outcome expectations. Health education and self-management programs focus not only on changing the patient’s awareness of the disease, but also on improving their personalized self-efficacy attainment strategies addressing motivation to change motivational enhancement strategies, behavior modification, and emotion management and stress management issues [58, 59].

Since self-management is a complex intervention, how to deliver a reasonable self-management intervention is still being explored [60]. We use process evaluation to provide novel information and understanding about how self-management works in clinical practice. Process evaluation will evaluate the components of execution, the influence mechanism and context to identify and understand the core components of self-management [61]. This not only can increase understanding of how the process is delivered, but also help us derive more reliable conclusions about the effectiveness of interventions in terms of whether they can be accepted by patients and healthcare professionals [62].

However, this study has some limitations. Firstly, potential selection bias or confounder effect may exist because it is a hospital-based design study, the results may not sufficiently generalize to other populations. But we try to select 3 branches hospitals, while the source populations of these three hospitals have different demographic and socioeconomic characteristics, it can increase the confidence in the generalizability of the study. And more so than other cohorts, its hospital-based design can ascertain that CKD patients’ diagnoses and the interpretations of the renal outcomes are accurate which means the risk of misclassification bias is low [63]. Secondly, the long study period is a double-edged sword. Although a considerable amount of follow-up data can be obtained during the study which allows us to have the possibilities to dynamically evaluate patients’ condition and study the relationship among longitudinal trajectories of exposure, surrogate markers, and renal outcomes [48]. High rates of patient's lost-to-follow-up are inevitable, how to minimize the drop-outs and deal with the missing data is a big challenge. One solution is the patients enrolled in this cohort are registered in the chronic disease management clinic in GPHCM, they will return to the outpatient clinic for regular checkups because we are the designated hospital for patients' medical insurance. The other solution is some retention strategies have been adopted, such as reducing barriers to participation, setting reminder strategies and gift/freebies incentives.

In a word, although self-management has been recommended in CKD guidelines, it is still in its infancy in China. Better knowledge about the topic may highlight the need for CKD patients and healthcare professionals to facilitate lifestyle change, manage symptoms, and offer rational support. The results of this project are expected to provide evidence of high methodological quality on the effectiveness of self-management intervention for CKD patients in order to improve the lives of patients with CKD by slowing progression, improving psychological well-being and overall quality of life.