Shared Decision Making Among Older Adults With Advanced CKD

Abstract Rationale & Objective: Older adults with advanced chronic kidney disease (CKD) face difficult decisions about dialysis initiation. Although shared decision-making (SDM) can help align patient preferences and values with treatment options, the extent to which older CKD patients experience SDM remains unknown. 

Study Design: A cross-sectional analysis of patient surveys examining decisional readiness, treatment options education, care partner support, and SDM. 

Setting & Participants: Adults ages 70 years and older with non-dialysis advanced CKD from Boston, Chicago, San Diego, and Portland (Maine). 

Predictors: Decisional readiness factors, treatment options education, and care partner support. 

Outcomes: Primary: SDM measured by the SDM-Q-9 instrument, with higher scores reflecting greater SDM. Exploratory: Factors associated with SDM. 

Analytical Approach: We used multivariable linear regression models to examine the associations between SDM and predictors, controlling for demographic and health factors. 

Results: Among 350 participants, the mean age was 78 ± 6 years, 58% were male, 13% identified as Black, and 48% had diabetes. The mean SDM-Q-9 score was 52 ± 28. SDM item agreement ranged from 41% of participants agreeing that "My doctor and I selected a treatment option together" to 73% agreeing that "

CLICK HERE TO GET HERBAL FORMULATION OF CISTANCHE FOR KIDNEY

My doctor told me that there are different options for treating my medical condition." In a multivariable analysis adjusted for demographics, lower eGFR, and diabetes, being "well informed" and "very well informed" about kidney treatment options, having higher decisional certainty, and attendance at a kidney treatment options class were independently associated with higher SDM-Q-9 scores. 

Limitations: The cross-sectional study design limits the ability to make temporal associations between SDM and the predictors. 

Conclusion: Many older CKD patients do not experience SDM when making dialysis decisions, emphasizing the need for greater access to and delivery of education for individuals with advanced CKD. 

Keywords: shared-decision making, older adults, chronic kidney disease, dialysis, education; 

Plain Language Summary: Older adults with advanced kidney disease face difficult treatment decisions. Dialysis offers uncertain survival benefits but has significant quality-of-life implications. 

Shared decision-making (SDM) may help patients choose treatment options that best align with their goals and values. We performed a cross-sectional analysis among older adults with advanced kidney disease to examine SDM in nephrology clinics, using the 9-item Shared Decision-Making Questionnaire (SDM-Q-9). SDM was suboptimal, with mean SDM-Q-9 score 52 (possible score 0-100). Being "well informed" and "very well informed" about kidney treatment options, having higher decisional certainty, and attending a kidney treatment options class were associated with higher SDM. Our research highlights the need to improve SDM for older adults facing dialysis decisions. 

Introduction 

Older adults are the fastest-growing demographic receiving dialysis in the United States.1 Importantly, dialysis and conservative management of advanced chronic kidney disease (CKD) appear associated with similar survival among older adults, particularly those 80 and older, while quality of life may vary between the two treatment approaches.2-5 The burdens of dialysis can be substantial, especially among the elderly, who commonly have significant comorbidities and poor functional status.6, 7 Older adults with CKD may also have different values than their younger counterparts and are more likely to prioritize the quality of life, decreasing caregiver burden, and preserving autonomy over survival.8, 9 Accordingly, more so than with many healthcare decisions where there is a clear survival benefit, dialysis decisions for older adults are preference-sensitive. 

Shared decision-making (SDM) may help optimize decisional outcomes.10 SDM is a collaborative discussion-based model that engages patients, clinicians, and often care partners in the decisional domains of agenda setting, information sharing, deliberation, and decision-making.11-13 With SDM, both the physician and patient are involved in the decision-making process. The physician shares pertinent medical information, while the patient and care partners share their values and concerns. Patients and their clinicians make a decision together, one that balances medical risks and benefits with patient preferences.11 SDM addresses decisional needs and enhances decisional outcomes with complex medical choices, in which there is not one clearly superior option.14, 15 SDM increases patient engagement with the decision-making process and helps patients make medical choices that better align with their values.16 

Nephrology professional organizations recommend using SDM to discuss with patients available treatment options for advanced CKD.17-19 However, many nephrologists struggle to implement SDM, especially with older adults facing dialysis decisions.20, 21 Previous qualitative studies reveal that many older adults do not perceive dialysis initiation as a choice, feel disengaged from the decision-making process, and experience regret after starting dialysis.22-24 Lack of understanding about treatment options may also contribute to decision dissatisfaction, as studies have found that poor medical knowledge is associated with decreased feelings of self-efficacy.25, 26

Gaps remain in quantifying the extent of SDM experienced by older adults who are nearing the point at which dialysis decisions are needed. Few quantitative studies have evaluated SDM in discussions about advanced CKD treatment options that take place in routine clinical practice. Studies that address this topic have mostly included younger patients, did not use validated measures, were limited geographically, or were retrospective, evaluating treatment decision-making in individuals who have already initiated dialysis, thus excluding those who chose conservative management.27-29 Increased knowledge about SDM in nephrology clinics will help identify areas for improvement, implement interventions to increase SDM, and track SDM utilization over time.

To improve our understanding of SDM among older adults with advanced CKD, we examined SDM in nephrology clinics from four geographically diverse sites across the United States, using baseline data from the Decision Aid for Renal Therapy (DART) Trial. 

Methods 

Study design and sample

In 2018-2019, we recruited 400 patients and randomized 363 of these patients from nephrology clinics in Greater Boston, Portland (Maine), San Diego, and Chicago to participate in the DART Trial (ClinicalTrials.gov NCT03522740), a randomized controlled trial investigating the effectiveness of a web-based decision aid compared to routine in-person education in reducing decisional conflict.30, 31 Eligible patients had non-dialysis advanced CKD and were age 70 or older, English speaking, and established in a nephrology clinic. To determine eligibility by eGFR, one of the last two eGFR values had to be <30 ml/min/1.73m2 with the other <35 ml/min/1.73m2. Scheduled kidney transplantation or dialysis initiation were exclusion criteria. All participants provided written informed consent, and the Tufts Health Sciences Institutional Review Board approved and served as a single IRB for this study

Our study is a cross-sectional, observational study with data derived from the baseline survey of the DART Trial. Research coordinators administered the survey in person after participants provided informed consent, but before randomization into the DART Trial; thus participants had not yet received the intervention. Of the 363 patients who participated in the DART study, 350 participants answered questions from the SDM-Q-9 survey and were included into this study. 

Outcome 

The primary outcome was patient-perceived SDM, measured using the validated, widely used 9-item Shared Decision-Making Questionnaire (SDM-Q-9), which consists of 9 statements about the decision-making process (Figure 1).32 The core components of SDM addressed in the SDMQ-9 include agenda setting (including identifying a decision to be made and clarifying the degree to which patients want to be involved in the decision-making process), information sharing (including medical information from the nephrologist and discussion of patient values), deliberation, and decision-making, with the input from both the clinician and patient.11, 12, 33 Item responses were assessed using a 6-point Likert scale from 0 (strongly disagree) to 5 (strongly agree). The overall score is standardized to a 0 to 100 point scale by calculating the average score among the 9-item responses and multiplying this number by 20. Higher scores indicate greater SDM. We evaluated the SDM-Q-9 using the COnsensus-based Standards for the selection of health status Measurement Instruments (COSMIN) criteria.34 Psychometric testing demonstrated high reliability, validity, and acceptance in English and German versions.32, 35, 36 As an exploratory analysis, we investigated the factors associated with SDM, examining demographics, health-related factors, decisional readiness characteristics, and education and support. 

Covariates 

The Ottawa Decision Support Framework (ODSF) was used to guide the analysis. The ODSF suggests that increased decisional needs, if not adequately addressed, affect patients' decision-making and perceived SDM. 14 To assess decisional needs, we examined decisional readiness characteristics, defined as patients' perceptions about their treatment options knowledge, their certainty about their decision, and the quality of their medical care. We also examined kidney treatment options class attendance, care partner support, and demographic and health factors as decisional needs that may affect SDM (Figure 2).37 

To examine decisional readiness characteristics, we used 3 items adapted from the DECISIONS survey.38 One item asked how informed participants felt about their kidney disease treatment options, scored on a 4-point Likert scale from "not at all informed" to "very well informed." The second question asked if participants had decided on treatment if their kidneys failed, using "yes" or "no" answer choices. The third question asked how certain participants were about their choice on a 10-point scale, with 1 = "not at all sure" and 10 = "completely sure/certain." A single item assessed affective forecasting, querying about the ability to imagine life with hemodialysis, using a 4 point Likert scale from "very easy time picturing what to expect," to "absolutely no idea what to expect."39 The Canadian Health Care Evaluation Project (CANHELP) Lite Questionnaire was used to assess satisfaction with medical care, with scores ranging from 0 to 100 and higher scores reflecting greater satisfaction with care.40

To assess other decisional needs, participants were asked if they had attended a kidney treatment options education class and had a carepartner. Health factors included estimated glomerular filtration rate (eGFR) reported in the local electronic health record (all sites used the 2009 4- variable CKD-EPI equation), urine albumin to creatinine ratio (UACR), and comorbid conditions. Self-reported health was assessed using a single-item "Health Slider" from the EQ- 5D questionnaire, in which participants rated their overall health on a 0-100 scale, with higher numbers representing better health.41 Demographic characteristics included self-reported age, sex, self-identified race, marital status, and educational attainment. 

Statistical analysis

Bivariate associations between the variables listed above and SDM quartiles were examined using Chi-squared tests for categorical variables and ANOVA for continuous variables. To assess the cross-sectional associations between decisional factors and SDM, we used a multivariable linear regression model that included demographic and health characteristics, decisional readiness, options class attendance, and involvement of an identified care partner. Two exploratory models used backward selection to examine which modifiable factors contributed most to SDM. In each model, age, sex, and race were fixed, while clinical characteristics (diabetes, UACR, eGFR, and health slider) were included in the selection process along with either decisional readiness factors or options for class attendance and care partner presence. We verified that there was no violation of normality or constant variance assumption in the outcome models. 

Missing data for covariates are shown in Table S1. To minimize the loss of power when fitting multivariable models and assuming data were missing at random, we used multiple imputations with chained equations to create 20 multiple complete datasets. The imputation model included all covariates in Table 1. Linear regression models were fitted for each imputed dataset and averaged using Rubin's rule.42 All analyses were performed using the SAS Enterprise Guide (Version 7.14, Cary, NC).