Structured Abstract
Background:
Health outcome measures for chronic diseases have focused on survival, condition-specific indicators, and general health-related quality of life. These measures do not assess success or failure in meeting patient-centered goals. Goal-attainment scaling (GAS) focuses on a patient's individual health goals and how well these goals are being met.
Objectives:
Our research aimed to (1) develop a standard set of goals for patients with dementia and their caregivers; (2) develop a methodology to revise goals when the clinical or social situation changes; and (3) integrate goal setting and attainment into clinical care.
Methods:
We recruited patients with dementia and their caregivers from the University of California Los Angeles Alzheimer's and Dementia Care program. The study had 3 phases. In phase 1, we conducted focus groups with 7 patients with early dementia and 36 caregivers to produce an inventory of goals for dementia care. In phase 2, we pilot tested this inventory in a sample of 6 patients with dementia and 26 caregivers to determine feasibility and refine the goals. In phase 3, we studied 101 patient/family caregiver pairs over a period of 6 to 12 months to assess the importance and difficulty of achieving different goals and estimate how often goals were attained. At 6 months, we asked participants whether their goals remained the same, whether they wanted to rescale the responses of their goals, or whether they wanted to create new goals. We also elicited caregiver perceptions of how valuable this process was and whether it captured something different from usual medical care.
Patient and caregiver outcomes:
The final inventory resulting from the 3 phases identified 53 goals for dementia care that fit within 5 domains (medical care, physical quality of life, social and emotional quality of life, goals for accessing services and support, and caregiver support goals). The vast majority of goals (87%) were nonmedical. Caregiver support goals were most commonly selected (34%). Goals related to patient physical (30%) and social/emotional (16%) aspects of life were the next most commonly selected. With the help of nurse practitioner dementia care managers (DCMs), caregivers and patients were able to operationalize these goals for clinical use. The majority (60% at 6 months and 72% at 12 months) of participants felt it was not difficult to rate and rank the importance of goals. Patient- and caregiver-identified goals were integrated into the clinical care of 101 patients; 96% felt the goals they set were meaningful; and 79% felt the process captured something different from usual medical care. At 6 and 12 months, 75% and 74% of participants, respectively, had attained the goals they had set.
Conclusions:
The study demonstrated patients with dementia and their caregivers can establish patient-centered goals that can be operationalized, worked toward, and measured. Further research is needed to integrate this approach into routine clinical care and adapt it to other conditions.
Limitations:
This research was limited to a single site and a single condition, dementia. Because relatively few participants changed their goals, inferences about new goals is limited.
Background
For most chronic diseases, there are no cures. Hence, the focus turns to prolonging survival and ameliorating symptoms with the intent of maintaining or improving quality of life. Outcomes measurement for chronic diseases has traditionally focused on survival, condition-specific indicators (eg, glycated hemoglobin levels and hypertension readings), and disease and symptom control. In addition, general health-related quality of life measures that assess functioning and well-being are widely used.1 These outcomes are “universal” goals that most people with a long-life expectancy and who are free of multiple comorbidities would want. However, these universal goals may not be possible or desired by all patients. That is, for some patients and particularly those with advanced illness, these outcome measures may not reflect individual preferences or goals and are not patient centered.
The inability to measure outcomes that reflect clinical care that meets patients' goals is problematic for several reasons.2 First, it means important outcomes for those with advanced illness (defined as “occurring when 1 or more conditions become serious enough that general health and functioning decline, and treatments begin to lose their impact”)3 are ignored. Advanced illness is a process that continues to the end of life and is associated with high costs. By defining appropriate patient-centered outcome measures that respect people's preferences instead of relying on universally desired outcomes, it may be possible to provide more desired care (eg, care that avoids intensive care units and ventilators) at lower costs.
Second, if only universally desired outcomes are considered, the quality of care provided to a person with advanced illness may be rated as low. In other words, a positive outcome from an individual-patient perspective would either not be captured or would contribute negatively to measures of the quality of a clinician's care and population health. For example, extended physical therapy might enable a hip fracture patient transition from needing an assistive device to ambulating independently.4 However, the patient may perceive the amount of work required to make this transition as not worth the effort for the amount of gain. Thus, a health plan might be negatively rated on a measure of functional status following hip fracture although the patient's preference was accommodated. For these reasons, the field of outcomes measurement needs to develop a set of quality metrics that capture individual goal elicitation and attainment as part of goal-directed, patient-centered decision-making.
One approach to providing patient-centered, individualized care is to focus on a patient's individual health goals within or across a variety of dimensions (eg, symptoms, physical, role and social functioning) and determine how well these goals are being met. This approach has been called goal-attainment scaling (GAS). Researchers have been using GAS for decades,5 to measure the effect of treatment of conditions such as dementia6 and in comprehensive geriatric assessment.7 It has also been used in clinical rehabilitation programs, though with skepticism about whether it is truly an outcome or “stand-alone” measure.8
Patients may have goals that are not well captured by established, validated patient-reported outcome measures (PROMs). For example, the 12-Item Short Form Survey (SF-12) will not capture a patient's goal of improving transfers using a walker. GAS allows for greater personalization of a goal. For example, a goal of pain control could be specified using GAS as not requiring opioids for pain control or improvement on a pain scale. GAS allows for flexibility in how a patient operationalizes a goal in a general domain (eg, function, symptom control, social roles, quality of life).
PROMs also have the disadvantage of being geared toward better health (eg, improving symptoms and well-being) and do not fit well for patients who are unlikely to improve or who are expected to decline. Also, the qualitative meaning of a PROM score and the change in score that is clinically meaningful may not be clear to clinicians or patients. GAS has the advantage of capturing a health-related goal that is more personalized to an individual patient. Because scaling a goal requires specifying what would be better or worse than the expected goal attainment outcome, it may also better facilitate care planning toward meeting a goal as compared with PROMs.
We designed this study to apply and evaluate the utility of GAS as an outcome measure for patient-centered outcomes research. The research used dementia, including Alzheimer disease, as a prototypic disease for development of patient-centered outcomes. In the United States, an estimated 5.5 million people are affected by Alzheimer disease.9 Moreover, the burden of dementia is even higher, as Alzheimer disease accounts for only 60% to 80% of cases of dementia.
Dementia is a progressive, incurable disorder. At best, some of the medications slow the progression and may help manage severe behavioral disturbances. Dementia is a disorder that exemplifies the conundrum in which existing outcome measures have little meaning, thus demonstrating the need for the development of new measures that can capture success or failure in meeting patient-centered goals when universally desired goals may not be in the range of possibilities. Moreover, dementia is usually a long-term disease, with median survival of 6 to 8 years. During this time, the needs, goals, and potential to achieve initial goals change. Early on, the goal may be to preserve functional status, including the ability to work (eg, Glen Campbell performed >100 concerts when his disease was still in the early stages). However, as the disease progresses, goals shift to maintaining ability to live at home; maintaining mobility; controlling behavioral and psychological complications; reducing caregiver burden; and eventually ensuring a dignified, peaceful death. As a result, dementia is a disorder in which all decisions should be considered within the context of the stage of the disease and the individual patient's goals, as well as the preferences and the tradeoffs required to achieve them.
The goal of this study was to create a goal-oriented approach to making health care decisions and assessing outcomes for dementia care that frames the discussion in terms of individually desired health states. This approach simplifies decision-making for patients with multiple conditions by focusing on outcomes that span conditions and aligning treatments toward common goals. Thus, patients can be in control when treatment options require tradeoffs (eg, better symptom control at the expense of potentially shorter life span). If successful, this approach could be adopted for other chronic incurable diseases. The study's aims were the following:
Further develop GAS for dementia by partnering with patients and their families to develop a standardized set of goals.
Validate this approach against other clinical outcome measures being collected (eg, caregiver burnout and depression, neuropsychological symptoms) as well as caregiver ratings of care.
Develop an approach for assessing revised goals when the clinical or social situation changes, thus permitting a more precise measurement of evolving goals.
By imposing structure and rigor upon GAS, we aimed to create a measure that could be used in comparative effectiveness research and quality improvement efforts.
Participation of Patients and other Stakeholders in the Design and Conduct of Research and Dissemination of Findings
Our engagement plan included convening a Study Oversight Committee (SOC) with broad representation of researchers and stakeholders that included 2 patients with early dementia, 4 caregivers who cared for someone with dementia, representatives from 2 community-based organizations that provide dementia services, 3 Spanish-speaking members, 4 physicians, and 2 nurse practitioner dementia care managers (DCMs). The SOC had a complementary function with the research team and shared responsibilities for the major decisions of the study. SOC functions included (1) providing advice on selection of samples of participants for all phases of the study, (2) review and approval of all study materials before use in each study phase, (3) review of results of each study phase, and (4) ensuring the study outcomes remained patient centered. Overall, 10 SOC meetings were held on a quarterly basis throughout the project period to allow SOC members to closely review study documents and procedures, providing comments and suggestions for revisions.
The broad-based SOC was invaluable. The preparation work for the SOC helped us determine how to communicate complex ideas to members of various backgrounds and medical sophistication and was critical in maintaining the focus of the study and ensuring that the outcomes would be meaningful. The comments of the SOC during the meetings were provocative and challenged our a priori assumptions (eg, during the May 2016 meeting, SOC members prompted a discussion about assessing goal attainment at follow-up that helped us clarify the 3-prompt approach we used to assess goal attainment). Finally, input in between meetings (eg, feedback on instruments and design) was exceptionally valuable. As a result, we plan to establish similar committees for all future research that involves patients. The SOC was also an exceptionally valuable way to engage stakeholders.
The fact that study participants had already established meaningful relationships with their DCMs through the University of California Los Angeles (UCLA) Alzheimer's and Dementia Care (ADC) Program was very helpful in engaging patients and their caregivers in the study. Not only were they more likely to participate, the DCMs were also able to provide insight into approaches that might help keep patients and caregivers engaged. For example, they knew when life events or health changes might preclude ongoing participation or require a delay. In addition, they were able to provide insight into how best to contact participants (eg, email vs telephone). The most notable impact of engaging patients and caregivers in the project was identification of patient-centered goals, particularly through the phase 1 focus groups and ongoing in the context of identifying individual goals as the study progressed. This input was not only critical to completing the study; it created exceptionally strong face validity for the goals. In addition, the input of our stakeholders provided key insights into the value of this approach.
The most useful impact of engaging patients and stakeholders is the potential for widespread uptake (adopt and adaptation) by health systems. A major facilitator of this was inclusion of the DCMs, who provided guidance along the way and participated in 2 focus groups, the first about the process of GAS and the second about its clinical value. Another important, though unanticipated, stakeholder was the National Committee for Quality Assurance, which learned of the study and incorporated the GAS approach into a national learning collaborative with 7 practice sites. The GAS approach was piloted with 127 patients showing almost two-thirds of patients meeting their goal (62%) at 6 months. If successful, that project could lead to widespread dissemination, including potential use of GAS as a quality measure.
There were several impacts of engagement on the investigators. First, we were able to obtain insights from different types of people, including patients, front-line community-based organization workers, and health care systems. The study participants benefited from engagement through the continual input in revising instruments and approaches. The health care system (ie, clinical care) was affected by patient engagement, including conversations between individual patients and providers about goals of care as well as incorporating these into the electronic health record notes.
Methods Overview
GAS is 1 method that could be used in clinical care for dementia to measure personalized goal attainment. GAS encourages the specification of a personal goal by describing the expected level of goal achievement within a given follow-up period as well as achievements that would be better, much better, worse, or much worse than expected (5-point scale) in that time period. GAS can accommodate diverse preferences within the context of achieving an overall goal (eg, safety may be specified as avoiding falls or preventing wandering) and allows goals to be modified or discarded if the clinical or social situation changes. Thus, in GAS an individualized outcome measure is created in which goals are set and followed over the course of a trial, observational study, or clinical care. The goals are personalized (ie, set according to the patient's needs), and what is standardized is the extent of their attainment. For patients with dementia who have cognitive impairment, there is generally no alternative to caregivers making decisions and planning. Caregivers have been used extensively as proxies to report traditional outcomes in dementia (eg, behavioral symptoms, depression).
Researchers have used GAS in clinical rehabilitation, in mental health, and in comprehensive geriatric assessment. Although GAS has good psychometric properties, it has rarely been used as a primary outcome measure in clinical medicine. Current limitations to the use of GAS in the care of patients with dementia include limited use of GAS in dementia cohorts, limited data on its feasibility in clinical care, and differences in administration and scoring (eg, use of weights, differences in baseline scoring). The use of caregivers more often than individuals with dementia in this study was a pragmatic approach and reflects real-world situations, in which individuals with dementia are often too cognitively impaired to participate in the process.
Although data from this study show that GAS is feasible and adds value to usual medical care for dementia, practical barriers to the implementation of GAS remain. Provider and health system barriers include visit time constraints, a culture of disease-specific outcomes and cure, and a workforce unprepared to use GAS. There are also patient and caregiver barriers to implementation, including clinically unrealistic goals (eg, cure dementia, resume driving), and situations in which the goals of the patient and caregiver do not align. Ongoing study of the feasibility and acceptability of GAS to guide and organize clinical care around achieving personalized goals is needed.
The development of GAS-directed care has several applications. Personalized goals can be used in clinical settings as prompts to help patients and caregivers obtain what is important to them. In tandem, health care providers can organize care toward meeting these goals and away from undesired outcomes. How well personalized goals are attained might also be used to measure the quality of care provided by a health system or how well different models of dementia care perform. As the specification and measurement of individual goals are developed, the attainment of personal goals can emerge as an increasingly valuable outcome of dementia care and can be applied to the care of patients with other advanced illnesses and with multiple morbidities. The overall study design aimed to (1) develop a standard set of goals for patients with dementia and their caregivers, (2) develop a methodology to revise goals when the clinical or social situation changes, and (3) integrate goal-setting and attainment into clinical care.
The study had 3 phases. In phase 1, we conducted focus groups with patients with early dementia and caregivers to produce an inventory of goals for dementia care. In phase 2, we pilot tested this inventory in a sample of patients and caregivers to determine feasibility and refine the goals. In phase 3, we studied patient/family caregiver pairs over a period of 6 to 12 months to assess the importance and difficulty of achieving different goals and estimate how often goals were attained. All phases of the study included patients with all types of dementia.
Phase 1. Qualitative Study to Develop a Standard Set of Goals for Patients with Dementia and their Caregivers
Methods for Phase 1
Study Design
We conducted a qualitative study from April to June 2014 to explore the goals of people with dementia, both from the perspective of people living with early-stage disease and from the perspective of caregivers of people with all stages of dementia. We conducted 1 focus group with 6 people with early-stage dementia and 4 focus groups with a total of 36 caregivers of people with dementia (one of the groups consisted of 14 Spanish-speaking caregivers) and conducted a semistructured interview with 1 Spanish-speaking female with dementia. We used the qualitative data from these focus groups to create an inventory of dementia care goals (37 goals identified in focus groups and 4 added by the SOC).
Forming the Study Cohort
Focus group participants were recruited from 5 community-based organizations that provide dementia-related services in partnership with the UCLA ADC program, a comprehensive dementia care management program. Participants were recruited using flyers posted in clinical areas and at community-based organizations. Mailers were also sent to patients and caregivers enrolled in the ADC program. Patients and caregivers who agreed to be contacted were invited to participate in the focus groups. We used purposive sampling to select participants with early-stage dementia and to ensure inclusion of Spanish-speaking participants. For the Spanish-speaking focus group, the project coordinator made brief announcements in Spanish at a community-based organization that hosted Spanish-speaking support groups for caregivers. The Spanish-speaking focus group had more participants because it was conducted at a community-based organization at the time of its caregiver support group. Fourteen people wanted to participate, and we allowed them all to participate. The other 4 focus groups were conducted on the UCLA campus in a conference room in the Division of Geriatrics office suite. To create focus groups with a balanced representation of various perspectives, we asked potential study participants their gender, patient's disease stage (mild, moderate, and severe) and the caregiver's relationship to the person with dementia (spouse, child, other family member, or friend). We did not use these variables for inclusion/exclusion, but rather to ensure a balanced perspective in the focus groups. Participants with early-stage dementia were required to have a physician diagnosis of dementia and be able to give informed consent. Caregiver participants could be unpaid family members or friends or paid caregivers and could care for a person with dementia at any stage of disease.
Study Setting
The Spanish-speaking focus group was conducted at a community-based organization (Leeza's Care Connection) at the time of its caregiver support group. The other 4 focus groups were conducted on the UCLA campus in a conference room in the Division of Geriatrics office suite. Focus groups ranged from 5 to 14 participants. The 3 English-speaking caregiver focus groups included 8 participants, 5 participants, and 9 participants, respectively. The Spanish-speaking caregiver focus group included 14 participants, and the focus group of persons with early-stage dementia consisted of 6 participants. Dr Xavier Cagigas conducted 1 semistructured interview with 1 Spanish-speaking female with early-stage dementia,. We were unable to recruit additional Spanish-speaking persons with early-stage dementia to form a focus group.
The English focus groups were co-led by 2 study investigators, Dr Cagigas and Dr Lee Jennings. Dr Cagigas led the Spanish-speaking focus group alone as Dr Jennings does not speak Spanish. Dr Cagigas is a psychologist with expertise in care of patients with dementia and support of their caregivers and leads several Alzheimer's disease support groups. His research and clinical interests include the dementia care experience of racial and ethnic minority groups, particularly Latinos. He is bilingual in English and Spanish. Dr Jennings is a geriatrician and health services researcher, formerly an assistant professor at UCLA and now at the University of Oklahoma, and her work focuses on improving care delivery and health outcomes of older adults with dementia. Both Drs Cagigas and Jennings have prior qualitative research experience.
Interventions
Not available.
Follow-Up
Phase 1 did not have a follow-up visit. Participation consisted of 1 focus group.
Study Outcomes
We identified 41 discrete goals (36 by the focus groups and 5 by the SOC) and grouped them into 5 domains (medical care, physical quality of life, social and emotional quality of life, goals for accessing services and support, and caregiver support goals) to create a goal inventory for dementia care. We further refined the goal inventory in preparation for phase 2 and produced a patient version with 29 goals and a caregiver version with 8 additional caregiver-related goals (total of 37), which we used in phase 2 (see Appendix B).
Data Collection and Sources
We conducted 5 focus groups with 43 participants. English caregiver group 1 had 8 participants; English caregiver group 2 had 5 participants; and English caregiver group 3 had 9 participants. The Spanish caregiver group had 14 participants and the English patient group had 6 participants; as described earlier, Dr Cagigas interviewed 1 Spanish-speaking patient.
We developed a focus group guide based on prior work eliciting patient-defined goals and treatment preferences in older adults with dementia, other chronic illnesses, and near the end of life and based on input from Study Oversight Committee members, including geriatricians, nurse practitioner DCMs, representatives from community-based organizations that provide dementia-related services, and individuals with dementia and their caregivers. In the focus group guide, all participants (ie, both patients and caregivers) were first asked open-ended questions about their goals for their dementia care: “What are the important goals for your dementia care?” and “Why are these goals important?” Caregivers were asked to answer the questions considering the person with dementia's values and preferences. Participants were then prompted to consider goals relating to specific domains, including medical care, social functioning, emotional well-being, physical functioning, safety, and end-of-life care, and were also asked if certain situations would not be desired, such as a prolonged hospitalization or living in a nursing home. Participants were prompted to consider how their goals might change over time as the disease progressed and how the goals of a caregiver might be different from the goals of the person with dementia. Participants also completed a brief demographic questionnaire that included age; gender; and, for caregiver participants, their relationship to the person with dementia and whether they live with the person with dementia for whom they provide care.
Focus groups ranged from 6 to 14 participants, lasted approximately 90 to 120 minutes, and were digitally recorded and transcribed verbatim by a professional transcriptionist. The Spanish-speaking focus group had more participants because it was conducted at a community-based organization at the time of its caregiver support group. Fourteen people wanted to participate, and we allowed them all to participate. This was important for the engagement of our community partner. The focus group was longer (about 2 hours) to accommodate more people.
Analytical and Statistical Approaches
Verbatim focus group transcripts were independently coded line-by-line by 2 study investigators (L.J. and A.P. [English transcripts]; M.C. and K.R. [Spanish transcripts]) using both deductive (a priori) and inductive (emerging from the data) coding approaches10-12 to develop a coding scheme. Coding and analyses were iterative, allowing modification of the focus group guide to more deeply explore emerging goal domains. Noting the range of themes conveyed by participants in communicating their goals for dementia care, we formulated 5 domains in which to categorize participant goals. The full study team discussed emerging domains, discrete goals, and exemplary texts and settled any differences in coding by group consensus. We also shared emerging goal domains with the SOC for relevance and comprehensiveness of the domains and goals within them. Committee members agreed with all participant-elicited goal domains and no new domains emerged; however, committee members added 4 new discrete goals to the emerging list.
Conduct of the Study
The final study protocol is included as Appendix A. We conducted phase 1 from April to June 2014 to explore the goals of people with dementia, both from the perspective of people living with early-stage disease and from the perspective of caregivers of people with all stages of dementia. The IRB at UCLA approved this study.
Results for Phase 1
Flow sheets are not available for phase 1. The participant characteristics for phase 1 is Table 1.
Table 1
Phase 1 Participant Characteristics, N = 43 (7 Patients, 36 Caregivers).
Outcomes and Statistical Analysis
In phase 1, the mean age of caregiver participants was 63 years (range, 42-85 years); 72% were female; 67% were spouses or children of the person with dementia; 92% were unpaid caregivers; and 39% were Spanish speaking. Five of 7 participants with early-stage disease were men; mean age was 81 years (range, 73-92 years).
Patients with dementia, caregivers, and SOC members identified 41 discrete goals for dementia care within 5 domains: medical care, physical quality of life, social and emotional quality of life, accessing services and supports, and caregiver support (see Appendix B for a table of goal domains, discrete goals, and exemplary quotes). Although the importance of different discrete goals varied some among groups of participants, the range of goals was similar among groups, and all goal domains were described in every focus group.
Participants also articulated the need to reevaluate goals over time as the needs of the person with dementia changed with progression of the disease and considered goals in the context of the severity of the disease:
The goal, to me, was trying to keep her happy, but then that became secondary after a while because her physical condition became more of an issue as the dementia progressed. (daughter)
It's very hard to set in stone goals because they change … so you have to change with them and adapt, add or subtract from your goals or your plan depending on where they [person with dementia] are. (nephew)
Domains
Medical Care
Participants identified medical goals pertaining to receiving high-quality dementia care (eg, having doctors who work with them, avoiding medications with adverse effects); intensity of treatment (eg, avoiding burdensome medical care, staying out of the hospital); and end-of-life care (eg, dying peacefully and living as long as possible). When prompted about health states or situations that may not be desired, caregivers gave examples of intensive, burdensome medical care their loved ones had received and said they would avoid such care in the future. Caregivers also identified goal setting as a way to better communicate with their family member's physician, particularly about prioritizing medical interventions and therapies:
We need someone that's going to help us to tell the doctor, ‘Let's set some goals. What is going to help him? Let's just deal with that.’ (wife)
In contrast to the English-language focus groups, avoiding burdensome medical care did not emerge as a prominent goal for dementia care in the Spanish-language caregiver focus group; however, the Spanish-speaking participant with early-stage dementia mentioned limiting burdensome care near the end of her life as a goal.
Participants with early-stage dementia said an important goal was to not be a burden on family members, especially near the end of life. Both participants with dementia and caregivers identified dying peacefully as a goal.
The Spanish-language caregiver group identified the importance of receiving medical care for dementia in Spanish to facilitate better communication among the patient, caregiver, and health care provider. They noted negative experiences they had encountered in seeking health care for their family members with dementia secondary to linguistic and cultural barriers and identified obtaining dementia care in their preferred language and greater access to high-quality dementia care as prominent goals:
There are very few doctors who speak Spanish so you have to wait a long time to get an appointment, and meanwhile things are getting worse. (son)
When my wife started with her dementia … we were told there is nothing else you can do … but there has to be a way … to figure out how to manage it. (husband)
Some Spanish-speaking caregivers also equated eating with good overall physical and mental health and identified maintaining adequate nutrition for the person with dementia as a goal.
Quality of Life—Physical, Social, and Emotional
Participants with early-stage dementia emphasized the importance of maintaining a good quality of life as their disease progressed:
My 1 goal is to live my life in the remaining years until I start to go downhill as well as I can. For my wife and I to enjoy our lives … that's my goal—to do it up good. (man with early-stage dementia)
Who wouldn't want to continue to live forever? But the longer you live the more problems will ensue … it's not just a question of living forever, but it's the quality of the life. (man with early-stage dementia)
Participants with dementia identified maintaining physical functioning (eg, performing self-care and household activities or continuing to be physically active) and engaging in meaningful activity (eg, work, hobbies, or social interaction with family and friends) as goals central to maintaining a good quality of life.
Some participants with dementia and caregivers said it was very important for the person with dementia to continue to live at home as the disease progressed. However, participants also stated the goal to identify other high-quality long-term care options should the need arise. Caregivers were concerned about the affordability of long-term care and felt confused and overwhelmed when seeking information about the options available to their family members.
Ensuring the safety of the person with dementia was an important goal among caregivers. Caregivers identified areas where their goal to ensure safety was in conflict with the goal of their family member with dementia to maintain functional independence. One caregiver articulated the challenge of choosing between her goal of minimizing home hazards for her mother and her mother's desire to continue to live alone in her own home:
This balance between the autonomy of the patient and the decision-making of caregivers is a great challenge. One, because of course we want to respect their autonomy, but at the same time, clearly, their judgment is severely impaired … you try to not intervene, but at a certain point, you have to do that in order to keep them safe.
Similarly, another caregiver expressed concerns about his mother being at risk for fraud or financial abuse and her desire to continue to manage her own financial affairs.
Accessing Services and Supports
Assistance with legal affairs and financial concerns were 2 goals voiced by caregivers. Adult children caregivers conveyed concern their parents did not have adequate financial resources for care and were confused about how to seek advice to help address these concerns. Caregivers also had questions about how to establish a durable power of attorney for health care or for financial affairs and where to seek legal assistance for this. Caregivers expressed confusion about how to access community resources for dementia and difficulty understanding what benefits may be available to their family member with dementia.
Some Spanish-speaking participants indicated a lack of understanding about dementia in general within the Latino community and stated improving community awareness and education about dementia as a goal. They also reported a need to give back to the community by educating others about dementia:
In our culture, when an older adult begins to forget things, it is attributed to old age … disregarded as not important … and the problem is neglected … this situation happens a lot. (son)
One main goal would be to accept that there is a problem … and the way we can accept dementia is by educating ourselves and knowing what we're dealing with. (cousin)
I looked after my mom for 12 years … I decided the day she died that it was my time to come back to this great place (community organization) and do something for it. (daughter)
Caregiver Support
Caregivers also identified managing stress related to caregiving responsibilities, finding respite care, and receiving social support (eg, finding a support group) as goals. Caregivers spoke about difficulty managing behavioral symptoms related to dementia and wanted to improve their caregiving skills and feel more confident as caregivers. Minimizing family conflict regarding caregiving was also a goal, especially for caregivers who shared caregiving responsibilities with siblings or other family members. Finally, caregivers reported they had delayed seeking health care for their own medical problems due to the stress of caregiving and identified maintaining their own health as a goal.
Phase 2. Pilot Study to Determine Feasibility and Develop a Methodology to Revise Goals when the Clinical or Social Situation Changes
Methods for Phase 2
Study Design
We evaluated the goal inventory developed in phase 1 with a pilot sample of 6 patients and 26 caregivers to determine the feasibility of using the inventory to identify and select goals in clinical practice. We tested the GAS approach to identifying goals and assessing their progress with 11 participants in preparation for phase 3.
Forming the Study Cohort
Participants were recruited from the UCLA ADC program population. The ADC is a comprehensive dementia care management program. In the ADC cohort, 88% of individuals had Alzheimer's disease, a mixed etiology of dementia including Alzheimer disease, or an unspecified type of dementia. Other dementia types included vascular dementia (4%), Lewy body dementia (4%), frontotemporal dementia (2%), and Parkinson disease with dementia (2%). At the time of enrollment in the UCLA ADC program, patients and caregivers were asked if they would be interested in being contacted about participation in research studies. Patients and caregivers who agreed to be contacted were invited to participate with mailers if they met the study eligibility criteria. Recruitment was augmented by efforts of the DCMs to inform patients and caregivers about this project. The study staff reviewed the ADC clinic appointment calendar to identify patients scheduled for a standard of care ADC visit. Study staff then reviewed the scheduled patient's record to prescreen for the following criteria:
Patient:
Mini-Mental State Examination (MMSE) score of 23 to 30 in the initial ADC visit note
If MMSE <23, we identified an eligible caregiver
Caregiver:
Is listed as a caregiver for the patient in the ADC initial visit note
Is attending the scheduled ADC visit with the patient
We aimed for a variety of participants representing different dementia types and stages of dementia, caregiver relationship to the patient, and gender, as well as English and Spanish speakers.
Study Setting
The study took advantage of the UCLA ADC, which received a Centers for Medicare & Medicaid Services (CMS) Innovations Award. This unique program provides comprehensive, coordinated, patient-centered care for patients with Alzheimer disease and other dementias. Study visits occurred in the UCLA ADC clinic and by phone.
Interventions
Not available.
Follow-Up
Phase 2 did not have a follow-up visit. Participation consisted of 1 visit.
Study Outcomes
We chose to focus on GAS (ie, whether participants could set goals for their care and whether these could be measured, and the degree of attainment).
Data Collection and Sources
Participants completed the goal inventory at home Before the meeting with the DCM for the visit. The inventory asked for ratings of the importance of goals using 4 response options: not important at all, somewhat important, very important, or extremely important. To ensure that important goals for an individual were captured, the survey then asked, “Were there any other goals that were not listed that you consider important in the next 6 months? Please list them below and indicate their importance to you.” Study participants were next asked, “Taking into account the patient's and the caregiver's preferences, what are your 5 most important goals from the list above for the next 6 months?” During the baseline visit, DCMs reviewed the top 5 goals with participants and proceeded with a discrete choice experiment (DCE) with the top 3 goals. We used DCEs to determine whether participants were able to prioritize among important goals and conducted it primarily with caregivers. Nevertheless, we conducted 6 DCEs with patients; only 1 of the patients was unable to complete the DCE without help, providing evidence that individuals with early-stage disease could consider important goals and prioritize among them. The DCMs also piloted the GAS approach with 11 participants using their top goal. A research assistant observed and took notes while the DCM reviewed the goal ranking exercise, conducted the DCE, and tested the GAS approach. After the DCM completed this, a research assistant conducted a structured cognitive interview to determine whether patients and caregivers understood the system of weighing the importance of different goals, both as individual goals and relative to other goals. The cognitive interview gathered information about the clarity and comprehension of the goal ranking exercise, understanding how the goal ranking process can be most useful.
Analytical and Statistical Approaches
We evaluated the extent of agreement in the importance ratings by different study participants using a standard analysis of variance (ANOVA) model that estimates between-variance vs within-variance. This approach partitions the variance in importance ratings between the 45 goals vs the within rating variance for 124 participants (caregivers or patients from phases 2 and 3 who rated their individual goals, inventories of which included 45 different goals). Multiple respondents chose the same goals. Research assistant notes from the phase 2 cognitive interviews were reviewed by study investigators and were supplemented by audio recordings of the interviews if clarification was needed. We used these qualitative data to inform the development of phase 3, including study procedures and the preparation of study materials. However, these data were not formally qualitatively coded. The original plan was to use discrete choice scenarios to weight goals, but respondents had great difficulty with the task. Discrete choice experiments have been used successfully in marketing when there are clearly discrete options, but the extension to preferences in more complicated tradeoffs is sometimes problematic. For example, when we used the discrete choice method to elicit preferences for health states for the Patient Reported Outcomes Measurement Information System health states, the average score assigned to the US general population was about 0.20 on the 0 (dead) to 1 (optimal health) continuum. Weighting the responses by perceived importance is more valid than a “validated” method when respondents are confused by the preference elicitation task (eg, conjoint analysis).
Conduct of the Study
The final study protocol is included as Appendix A. We conducted phase 2 from December 2014 to May 2015 to field test the inventory of goals and determine feasibility as well as pilot test the GAS instrument into a mix of ADC program follow-up evaluations. The IRB at UCLA approved this study.
Results for Phase 2
Participant flow charts are not available for phase 2. The participant characteristics for phase 2 are listed in Table 2.
Table 2
Phase 2 and Phase 3 Participant Characteristics, N = 133.
Outcomes and Statistical Analysis
In phase 2, 6 patients and 26 caregivers participating in a dementia care management program completed a survey rating the importance of 38 patient-centered goals for dementia care and ranked the 3 goals they considered most important to achieve in the next 6 months. A 2-way random-effects ANOVA model indicated substantial agreement between study participants in ratings of the importance of the different goals (reliability = 0.979, intraclass correlation = 0.272).
Participants were able to add goals to the survey and were asked to consider both the patient's and caregiver's preferences when rating and ranking goals, which are included with phase 3 goals below.
Phase 3. Longitudinal Pilot Study of Patient-Caregiver Dyads Testing the Feasibility of Integrating Goal Setting and Attainment into Clinical Care
Methods for Phase 3
Study Design
We tested the feasibility of goal ranking and rating and using GAS in a longitudinal study of 101 patient-caregiver dyads to evaluate how well GAS captures patient-oriented care compared with traditional measures such as cognition, functional status, and mortality. Because this was an exploratory proof-of-concept study, the sample size for phase 3 was not powered for hypothesis testing. Rather, we designed it to have a sample size that was large enough to provide preliminary estimates and be conducted within the time and budgetary parameters of the award. We used a single group pre-post comparison, during which a DCM worked with patients and caregivers to identify and take specific steps to help them achieve their goals.
Forming the Study Cohort
Participants were recruited from the UCLA ADC program population using the same methods as in phase 2 to form the study cohort. The eligibility criteria were the following:
Patient:
MMSE score of 23 to 30 in the initial ADC visit note
If MMSE <23, we identified an eligible caregiver
Caregiver:
Is listed as a caregiver for the patient in the ADC initial visit note
Is attending the scheduled ADC visit with the patient
We aimed for variety of participants representing different dementia types and stages of dementia, caregiver relationship to the patient, and gender, as well as English and Spanish speakers.
Study Setting
Study visits occurred in the UCLA ADC clinic and by phone.
Interventions
Not available.
Follow-Up
The follow-up schedule included a baseline visit, a 6-month follow-up visit, and a final 12-month visit. Efforts to prevent missing data included offering visits by phone and working with DCMs to communicate with participants and finding appropriate times within the follow-up window to conduct study visits. We also offered respite care if needed to facilitate participation of caregivers.
Study Outcomes
Similar to phase 2, we chose to focus on GAS as the primary outcome measure. We also obtained study outcomes (see Table 3) that could be used to explore how GAS compared with traditional outcome measures.
Table 3
Health Outcomes Assessed in CMS Innovations Challenge Award.
Data Collection and Sources
Prior to first study visit, participants were asked to complete the dementia goals exercise at home and asked to identify their top 5 dementia care goals to use during the first study visit. At the first study visit, DCMs reviewed the top 5 goals and the completed GAS with top goal identified by participants. The specific attainment levels were created by each patient/caregiver together with the DCM as they went through the process of discussing goal setting. The DCM would then work with patients and their caregivers to set a SMART (specific, measurable, attainable, relevant, and time-specific) goal to work toward over the next 6 months.18 During the discussion, they identified possible outcomes for meeting or not meeting the goal using a 5-point scale (much less than expected, less than expected, expected, better than expected, and much better than expected; see Table 4). We monitored goal attainment with 6- and 12-month follow-up surveys. New goals proposed by participants during the goal ranking process in phase 3 were qualitatively coded by 4 study investigators, grouped into domains, and added to the inventory. At 6- and 12-month phone calls, the research assistant assessed the participant's perception of the quality of their dementia care, reviewed the goal identified as most important at the time of the first study visit, discussed progress toward meeting the top goal, assessed the participant's perception of how their dementia care was helping them meet their goal, and reviewed the GAS process with the participant. At 6 months, participants were asked whether their goals remained the same, whether they wanted to rescale the responses of their goals (ie, identifying different situational states to define each outcome category in the 5-category attainment scale), or whether they wanted to pick a new goal to work on for the next 6 months. At the end of the study, 2 focus groups were held with DCMs to understand the process of goal setting and assessment of GAS and also to better understand its clinical value.
Table 4
Example Baseline Goals and GAS.
Analytical and Statistical Approaches
There is no gold standard regarding how to ask people about whether they have attained their goals. We were concerned that simply asking participants whether they had attained the goal would not be a sufficient stimulus for them to provide this information. Hence, we used 2 additional prompts and compared 3 different methods of determining whether patients and caregivers had attained their highest-priority goal. Participants were asked to rate how well they had attained their goal:
Without cueing. Participants were asked to rate how well they had attained the goal they had set without any prompts other than the general goal description.
With cueing. Participants were shown the scaled goal they had created at the initial visit and asked to rate goal attainment.
With itemizing. Participants were shown the scaled goal and asked to confirm the things they were able to do from each of the 5 response categories, and then asked to rate their goal attainment.
The purpose of this approach was to understand how an individual's rating of goal attainment might change based on the method used to review the goal and assess goal attainment. Tables 5, 6, and 7 show the number of patient-caregiver dyads who changed their ratings of goal attainment across the 3 different assessment approaches. We concluded that the method used to assess goal attainment (how researchers ask patients about attainment) does influence the patient's attainment rating.
Table 5
Phase 3—How Goal Attainment Ratings Changed During Process of Assessing Attainment (6 Months).
Table 6
Phase 3—How Goal Attainment Ratings Changed During Process of Assessing Attainment (12 Months).
Table 7
Phase 3—Goal Attainment at 6 and 12 Months by Attainment Question.
Goals were also rated by importance and difficulty to allow for weighting and standardizing goal attainment across goals of variable importance and difficulty (eg, more difficult goals are harder to achieve). Weighting was planned from the beginning, but we did not know exactly how we would end up weighting the goals until after we tried various options. We did not conduct a formal analysis of the qualitative difference between GAS and other commonly measured health outcomes. The wording of the goals and the personalized categories of level of attainment are very different from standard patient-reported outcome measures, which typically assess personal responses to standardized questions. In addition, 79% of participants at 6 months and 80% at 12 months answered affirmatively when asked, “Did the goal-setting process get at something that is different from usual medical care?”
The importance of achieving the goal was rated by patient-caregiver dyads, and the difficulty of achieving the goal was rated by DCMs. To score goals, we first created a standardized unweighted goal-attainment score (Tgas) for the participant's goal by assigning value levels of attainment:
Much worse than expected (−2): Tgas = 30
Worse than expected (−1): Tgas = 40
Expected (0): Tgas = 50
Better than expected (+1): Tgas = 60
Much better than expected (+2): Tgas = 70
To obtain a weighted GAS, goals rated as a little important (2) received a weight of 75%, goals rated as moderately important (3) received a weight of 90%, and goals weighted very important (4) received a weight of 100%. We did not weight based on difficulty because 5 DCMs rated difficulty, and we could not conduct interrater reliability of ratings.
We linked demographic and clinical data from the UCLA ADC database to the primary data collected for this study, which enabled us to examine whether baseline characteristics (eg, sex, race, cognitive status) predicted goal attainment and how GAS correlated with other outcomes (eg, behavioral problems, caregiver strain) of the program. We used descriptive statistics to generate the percentage of goals attained and to summarize patient and caregiver demographic data and survey data collected as part of UCLA ADC clinical care (ie, MMSE, MCSI, NPI-Q symptom, and FAQ scores). For analyses, we dichotomized goal attainment (yes/no), defined as a GAS score of 0, +1, or +2 (yes) vs a GAS score of −1 or −2 (no). We performed basic bivariate statistics (eg, 2-sample t tests, Fisher exact test) to compare groups. We used a Bonferroni correction to adjust for multiple comparisons. After adjustment for 18 tests, ɑ = .00278 (.05/18).
All subgroup analyses examining associations with the goal attainment outcome were considered exploratory. We had 3 prespecified subgroup analyses. We hypothesized that more severe caregiver strain or more severe dementia-related symptoms would negatively affect goal attainment. However, given the small sample size, this analysis was still considered exploratory. We also hypothesized that the goals chosen by dyads would differ by severity of disease with persons with early-stage dementia choosing goals that were related to maintaining function or meaningful activity and persons with moderate/late stage disease more often choosing goals related to meeting caregiver needs or end of life care. Thus, we stratified goals by 3 categories of MMSE score as a proxy for disease stage.
We also developed a methodology to reexamine and revise goals when the clinical or social situation changes, thus permitting a more precise measurement of evolving goals. These changes might reflect an intercurrent medical illness or a loss of a caregiver. Typically, the DCM would initiate the reconsideration. We decided to treat goals that had changed as new goals and use simple averages of the 2 goals (baseline and new goal at 6 months). At 6 months, some patients/caregivers had attained their initial goals or decided to switch goals. For these dyads, we included both the 6- and 12-month attainment of goals in the assessment of their goal attainment and scored both as a simple average. For example, if a participant attained his or her 6-month goal but not his or her 12-month goal, his or her score would be 0.5.
We quantified responses from structured interview questions about the goal-setting process completed with dyads at 6 and 12 months as yes/no or into categories. Study investigators reviewed any additional qualitative comments provided by participants during the interviews, supplemented by audio recordings if clarification was needed. However, these comments were not formally qualitatively coded.
Phase 3 focus groups conducted with 5 DCMs were transcribed and independently coded by 2 study investigators (L.J. and D.R.) using both deductive (a priori) and inductive (emerging from the data) coding approaches to develop a coding scheme. After completing analyses from the first focus group (conducted at the study midpoint), the guide for the second focus group (conducted at the conclusion of the study) was created to allow exploration of emerging themes. Study investigators discussed emerging themes, and any differences in coding were settled by group consensus.
There was very little missing data for study outcomes in phase 3. The 33 follow-up interviews we missed were due to 2 reasons. First, recruitment was more difficult than initially planned, and we had to extend our recruitment period into our follow-up period. Second, the project deliverables were due before 29 of these participants were due for the 12-month follow-up visit. Because of these timeline changes, we did our best to gather at least the 6-month follow-up for as many participants as possible (86% having a 6-month follow-up survey). The help and insight from the DCMs ensured these high retention rates.
Conduct of the Study
The final study protocol is included as Appendix A. We conducted phase 3 from June 2015 to December 2016 to determine the stability of goals and how well the GAS approach captures perceived patient-oriented care and to explore the best psychometric approaches to a scoring system. A protocol change was the modification of timing of the primary outcome (attainment of goals) to be collected at 6 months rather than 12 months. This allowed the largest sample of outcomes as some participants enrolled later in the study. The institutional review board at UCLA approved this study.
Results for Phase 3
Figure 1 depicts the phase 3 flow of participants. The participant characteristics for phase 3 are listed in Table 2.
Figure 1
Phase 3—Flow of Participants.
Outcomes and Statistical Analysis
The most important goals were identified by 101 patient and caregiver dyads (Table 8). Examples of baseline goals and how they were operationalized are presented in Table 4. Caregiver support goals were most commonly selected (by 34%). Goals related to physical aspects of the patient's quality of life and social and emotional aspects of quality of life were next most common (by 30% and 16%, respectively). Medical goals and goals for accessing services and support release were less commonly selected (by 14% and 7%, respectively). There was no clear pattern to goal selection by cognitive status as assessed by the MMSE score. In Table 9, we present ratings of goal importance (as rated by patients and caregivers) and difficulty (as rated by DCMs). Almost all (94%) goals selected were rated as extremely or very important and most were rated as a little or moderately difficult.
Table 8
Top Ranked Goals for Dementia Care.
Table 9
Phase 3—Ratings of Goal Importance and Difficulty.
Tables 5, 6 and 7 present pairwise comparisons of goal attainment as assessed without cueing and after itemizing. Responses on the diagonal indicate assessed goal attainment was the same by both methods. Responses to the right of the diagonal in Tables 6 and 7 indicate participants rated their goal attainment higher without cueing, and responses to the left of the diagonal indicate participants rated their goal attainment lower without cueing. At 6 months, 52% (n = 44) did not change their rating of attainment after reviewing accomplishments (with itemizing), 24% (n = 20) rated their attainment lower after reviewing accomplishments, and 24% (n = 20) rated their attainment higher after reviewing accomplishments (with itemizing). At 12 months, 55% (n = 29) did not change their rating of attainment after reviewing accomplishments, 21% (n = 11) rated their attainment lower after reviewing accomplishments, and 25% (n = 13) rated their attainment higher after reviewing accomplishments.
Based on this analysis, we concluded the method used to assess goal attainment (how researchers ask about attainment) does influence the person's attainment rating. We chose “with itemizing” for our final analyses, as this approach asks the individual to review all 5 parts of the scaling and to comment on whether each part was achieved. We believed that this was the most thorough and detailed recall of goal attainment.
We completed 6-month follow-up assessments of 86 participants and 12-month follow-up assessments of 54 participants (Figure 2), which included ascertaining revised goals and determining if their original goal was met.
Figure 2
Phase 3—Goal Attainment at 6 and 12 Months.
Subgroup Analyses
At 6 months, goal attainment did not differ by race (White vs other), DCM, MMSE/dementia stage, goal domain, caregiver relationship, caregiver gender, whether caregiver lives with patient, patient age, MCSI score, NPI-Q severity or distress scores, FAQ score, instrumental activities of daily living score, goal difficulty rating by DCM, goal importance rating by dyad, or dyad rating of quality of dementia care at 6 months. Baseline median ADL (activities of daily living) scores were higher (more independent) among those who attained goals compared with those who did not attain goals: 4.0 (interquartile range [IQR], 2-6) out of 6 independent ADLs at baseline for those who attained goals vs 2.5 (IQR, 1-4) independent ADLs at baseline for those who did not attain goals (P = .03, rank sum test, n = 48). At 6 months, 90% of dyads in which the person with dementia was male (n = 26/29) had attained goals compared with 65% of dyads in which the person with dementia was female (n = 36/55; Fisher exact P = .02). Goal attainment differed by dyad rating of whether the patient's dementia care helped them attain their goal: 92% of dyads who attained their goal felt dementia care helped, while 68% of dyads who did not attain their goal felt dementia care helped (P = .01). However, none of these subgroup analyses were significant after adjustment for multiple comparisons.
Weighted mean GAS scores for the cohort were similar to unweighted scores (Table 10). Only 11 participants changed goals at 6 months.
Table 10
Phase 3—Mean (SD) GAS Scores.
Table 11 indicates caregivers' and patients' perceptions, obtained through structured interviews, of the goal-setting process and approach. Although at 6 months, 40% indicated it was difficult to rate and rank the importance of goals, by 12 months only 28% of respondents rated this as difficult. Goal setting as an approach to planning for future care was rated as helpful by an extremely high percentage of participants (89% and 88% at 6 and 12 months, respectively) and most felt the process achieved something different from usual medical care (79% and 80% at 6 and 12 months, respectively). Virtually all (96% at 6 and 12 months) regarded the goals they set as meaningful. About half (55% at 6 months and 5% at 12 months) would have preferred to work on more than 1 goal. Participants felt that the DCM was the right person with whom to discuss goals, and most felt a time frame of 6 months to obtain their goals was about right. Most had discussed their goals with their health care providers and family members.
Table 11
Dyad Perceptions of the Goal-Setting Process.
Item-level response rates at 6 months range from 93% to 100% (n = 80-86 for each item). Item-level response rates at 12 months range from 91% to 100% (n = 49-54 for each item).
In the focus groups, DCMs emphasized the importance of receiving training in GAS, including observation and role play, to become more comfortable with the process of making a personalized goal actionable. They also offered several open-ended questions as example “conversation starters” to help elicit values and care preferences from patients and caregivers (eg, “I want to hear your goals rather than tell you what to do,” “I'm thinking along the lines of … would you agree or disagree?”) They felt the goal-setting process improved their understanding of what was most important to their patients and identified goals they might have otherwise missed. They identified as challenging situations when patients or caregivers chose goals they felt were clinically unrealistic (eg, improve cognitive processes) or very difficult to achieve (eg, live at home without additional caregivers). They also reported the GAS process helped set expectations about disease progression and care needs for patients and caregivers. Some patients and caregivers had difficulty completing the extremes of the GAS scale (much worse or much better than expected), and they suggested a 3-point scale as an alternative for some patients.18 DCMs thought the GAS process empowered caregivers who did not feel confident in their decision-making and provided positive reinforcement to caregivers and patients who were able to accomplish part of their goal. GAS also provided a framework for helping people create a plan to accomplish their goal as well as follow through with this plan. They reported the GAS process added about 20 minutes to their clinic visit, on average (see Appendix C).
Discussion
This study is proof of concept that GAS can be used in the clinical care of patients with dementia to elicit, specify, measure attainment of, and revise personalized goals for care. We used GAS to facilitate the process of making a personalized goal actionable or SMART,18 and as a measurement tool, allowing the standardized assessment of goal attainment across the cohort at 6 and 12 months. Persons with dementia and their caregivers (dyads) were able to identify and prioritize goals that were important to them, and most goals were not related to medical care; rather, they focused on broader aspects of quality of life as well as caregiver support. Dyads with guidance from their nurse practitioner DCMs were able to operationalize personal goals, and most of the dyads had attained their goals at follow-up. Nearly all participants (96%) felt the process of goal setting was meaningful and more than three-fourths felt it achieved something more meaningful than usual medical care.
Decisional Context
Decision-making is made in clinical care by the clinician and patient, if capable, or appropriate surrogate, if not. We did not design this project to facilitate identification of the decision maker but rather to facilitate the decision-making process. In this context, our findings have important implications for both dementia care and the care of patients with multiple chronic conditions for whom traditional outcome measures may not be achievable.19,20 By shifting the focus of care toward goals that may be achievable and are important to patients, the success or failure of health care interventions can be better evaluated. However, to do so requires the specification of goals and how meeting them would be operationalized. For example, what states (eg, living at home) or processes (eg, hiring a caregiver) would meet an individual goal? Making patient goals explicit using GAS also allows health care providers to recognize when goals are inconsistent with clinical reality (eg, curing dementia, reversing cognitive decline) and provides an opportunity for patient and caregiver education.
Personalized goals can be used in clinical settings as prompts to help patients and caregivers obtain what is important to them. In tandem, health care providers can organize care toward meeting these goals and away from undesired outcomes. The degree to which personalized goals are attained might also be used to measure the quality of care provided by a health system or how well different models of dementia care perform.21 As the specification and measurement of individual goals is further developed using GAS, the attainment of personal goals can emerge as an increasingly valuable outcome of dementia care and might be applicable to the care of patients with other advanced illnesses and with multiple morbidities.
Study Results in Context
We found that people with early-stage dementia and caregivers were able to clearly articulate goals for dementia care, and these goals were perceived by patients and caregivers as achieving something different from usual care. Specifically, most goals addressed aspects of overall health, not just medical care. For example, participants noted the importance of meaningful activity and social interaction for emotional well-being, maintaining independence, addressing needed services for dementia outside of the traditional health care system, and providing caregiver support. Although some of these goals were similar to existing measures (eg, functional status, social functioning), others were not. Moreover, they were personalized and could be acted on. Many of these goals were also relevant to persons with multiple medical conditions who do not have dementia.22,23
Many of the goals identified (eg, low caregiver strain, management of behavioral symptoms, ensuring safety, avoidance of pain and depression, living at home, as much functional independence as possible, eventually dying with dignity) were generally applicable for dementia patients and their caregivers. However, what it meant to achieve these “universal” outcomes was often very individualized. For example, 1 caregiver defined safety as avoiding falls while another defined safety as preventing wandering. GAS can accommodate such diverse preferences within the context of achieving an overall goal (eg, patient safety). Moreover, there were tradeoffs even among these commonly desired health outcomes (eg, a caregiver accepting a more intensive caregiving role to avoid long-term nursing home placement of a parent). Some goals were more idiosyncratic and patient specific (eg, being able to attend a family reunion) and would likely have been missed as an important outcome for the patient in routine care because the goals fall outside of traditional medical care domains. The ability of GAS to capture change in attainment across a wide range of personalized goals is also an advantage over other PROMs.
The goals identified in this study and compiled in the final inventory confirm and extend prior qualitative work examining health- and care-related values among people with dementia and their caregivers.24-28 The similarity of goal domains articulated across studies reinforces the importance of these issues to patients and families in the care of people with dementia and underscores the need to measure patient-centered outcomes within these domains.
In this study, we used GAS to specify personalized goals and measure goal attainment in the clinical care of dementia. Researchers have used GAS in clinical rehabilitation,29,30 for mental health,31 and in comprehensive geriatric assessment.32 Current limitations to the use of GAS in the care of patients with dementia include limited use of GAS in patients with cognitive impairment,33,34 limited data on its feasibility in typical care settings,35 and differences in administration and scoring (eg, use of weights, differences in baseline scoring).36,37
This study begins to address some of these limitations by showing it is feasible to use GAS in the clinical care of patients with dementia in an outpatient setting. We also used a goal-setting process that involved the patient-caregiver dyad, as achieving the patient's goal is often dependent on caregiver support. The process of goal setting as a dyad also facilitated discussion between patients and their caregivers about the patient's values and care preferences, which could be used to guide the caregiver in future decision-making.
We also found that about 10% of participants changed their goals over 6 months and developed an approach to review and revise goals at follow-up. We used simple averages to calculate a summative attainment score if goals changed over time. This is particularly important in dementia as goals are likely to evolve as the disease progresses. Finally, we revised the GAS weighting system to be more clinically meaningful based on the patient's rating of the importance of the goal.
Implementation of Study Results
Although this study was successful at showing proof of concept, practical barriers to implementation of GAS remain. Provider and health system barriers include visit time constraints, a culture of disease-specific outcomes and cure, and a workforce unprepared to use GAS. There are also patient and caregiver barriers to implementation, including unrealistic goals (eg, reverse cognitive impairment, drive a car again), and situations in which the goals of the patient and caregiver do not align (eg, autonomy vs safety). We have included a set of implementation tips as Appendix C. Ongoing study of the implementation (ie, acceptability, feasibility) of GAS in clinical care is needed to further understand these barriers and to further develop GAS as a measurement tool that is useful in clinical care of patients with dementia.
Generalizability
This study piloted GAS in a single dementia care management program in a community-dwelling, insured population in an urban academic health system. Thus, these findings may not be applicable to some other dementia care settings. Additionally, GAS was performed with highly trained nurse practitioner DCMs, and these findings may not translate to other types of care providers or to laypersons in health coaching roles. However, 1 strength of the study that increases the generalizability of the findings is the diversity of the UCLA ADC patient population, which includes persons with all stages of dementia, middle-aged adults with early-onset disease, racial and ethnic minorities, those dually insured by Medicare and Medicaid, and both child and spouse family caregivers.
Subpopulation Considerations
All subgroup analyses were considered exploratory. We hypothesized patients with early-stage disease would have goals focused on functional independence while caregivers of patients with more advanced dementia would choose goals focused more on caregiver support. However, there was no clear pattern to goal selection by cognitive status, as defined by MMSE score. We also hypothesized more severe caregiver strain or more severe dementia-related symptoms might negatively affect goal attainment. However, we found no difference in goal attainment based on dementia severity (ie, MMSE), neuropsychiatric symptom score, caregiver relationship, or caregiver strain score at baseline. However, those who attained goals had greater functional independence at baseline, as measured by ADL score; however, this was not significant after correction for multiple comparison. There was also no difference in goal attainment by whether the goal focused primarily on caregiver support or on a patient need.
Goal attainment also differed by gender, with a greater percentage of men with dementia attaining goals. This was an unexpected finding and may deserve further study; however, this finding was also not significant after correction for multiple comparisons. Further multivariate modeling in larger cohort studies may help elucidate confounding factors related to gender and goal attainment. We also compared goal attainment with other patient-reported outcome measures (ie, caregiver strain, neuropsychiatric symptom severity) collected 3 months or more after setting the goal. However, the sample size was too small to make conclusions and is an area that also deserves further study.
We conducted 1 focus group that was exclusively Spanish speaking, and while there were slight differences in goals identified by the Spanish-speaking compared with the English-speaking participants, goal domains were similar. 38 It is likely that other meaningful personal goals will need to be identified and accommodated in goal-oriented care for patients with dementia.
Study Limitations
When interpreting these findings, the limitations of the study should be considered. The inventory of dementia care goals is based on goals elicited from 176 participants across all 3 study phases; however, the sample included only a small number of participants with early-stage disease (n = 29 [16%]). More early stage participants may have identified additional goals. As there were few early-stage participants, the goal-attainment data largely reflect caregiver input as most participants' cognitive impairment was too advanced to participate in rating goal attainment at follow-up. This use of 2 study populations (patient and caregiver) is a limitation but also occurs in other research on populations in which participants are not able to speak for themselves (eg, young children, developmentally disabled persons). Also, while this study did compare goal attainment to selected patient-reported outcome measures (ie, caregiver strain, neuropsychiatric symptom severity), the small sample size and missing 12-month and clinical data limited the robustness and generalizability of the findings. Because of methodologic limitations, we did not include degree of difficulty in the weighting of GAS. Finally, we must acknowledge the difficulty of including persons with dementia in the oversight of the study. It was hard to identify persons with dementia who were interested in participating and could follow the aims and methodology of the study.
Future Research
Future work should study the implementation of goal setting and measurement using GAS in other care settings (ie, primary care, nursing homes, home-based care) and with other provider types (ie, social workers, nurse case managers, physicians, community health coaches). Future studies could also adapt GAS for use in the care of other chronic conditions (eg, heart failure, chronic pain, cancer). An implementation science framework using mixed methods could be used to examine the acceptability, adoption, feasibility, and barriers to use of GAS in these different settings and populations. As most health care providers are unfamiliar with GAS, there is also need for further development of GAS training materials, including training in motivational interviewing, eliciting values, specifying goals using the GAS framework, and measuring attainment.
To further develop GAS as a rigorous outcome measure in dementia research,32 future work should improve the process (eg, adding difficulty as a weighting factor) and further validate GAS by comparing the method with other PROMs, where available, to examine the construct validity of goal attainment as an outcome. For example, the attainment of a personalized goal to reduce caregiver stress might correspond to a decrease in an index of caregiver burden. While this study did compare goal attainment to selected PROMs (eg, MCSI and Neuropsychiatric Inventory-Severity scale), the small sample size limited the robustness and generalizability of the findings.
Finally, to address the health care and research culture largely focused on disease-specific outcomes and cure, we need further engaged research with patients, caregivers, providers, payers, and researchers that incorporate measures, like goal attainment, that span disease conditions and address meaningful, relevant outcomes for persons with dementia and for other older adults with advanced, life-limiting illnesses.
Conclusions
This methods study aimed to further develop GAS as an outcome measure for dementia that can be used in studies and clinical practice. The findings of the study suggest goal setting using GAS can be incorporated into clinical care for persons with dementia and may have potential value as an outcome measure in dementia clinical care and research.
Using patient and caregiver input, we created a relatively small (53 goals) inventory of goals for dementia care, most of which were nonmedical and focused on patient quality of life as well as caregiver health and well-being. Patients and caregivers were able to select a most important goal and then, with the help of DCMs, operationalize what it would mean to attain these goals and plan the care necessary to attain them. While most dyads met their goal at 6 months, about a fourth did not, demonstrating GAS can capture failure as well as success in meeting outcomes. Almost all felt the goals set were meaningful (96%) and most (79%) felt the goal-setting process captured something different from usual medical care.
Although we conducted the study at 1 site within the context of a structured care program focusing on a single, specific disorder, our research demonstrates that this approach is practicable and might have applications to goal setting and measurement of goal attainment as outcomes for other chronic conditions and advanced illnesses.
References
- 1.
Working Group on Health Outcomes for Older Persons With Multiple Chronic Conditions. Universal health outcome measures for older persons with multiple chronic conditions. J Am Geriatr Soc. 2012;60(12):2333-2341. doi:10.1111/j.1532-5415.2012.04240.x [PMC free article: PMC3521090] [PubMed: 23194184] [CrossRef]
- 2.
ReubenDB, TinettiME. Goal-oriented patient care—an alternative health outcomes paradigm. N Engl J Med. 2012;366(9):777-779. doi:10.1056/NEJMp1113631 [PubMed: 22375966] [CrossRef]
- 3.
What is advanced illness?Coalition to Transform Advanced Care. Accessed October 11, 2012. http://advancedcarecoalition.org/
- 4.
BinderEF, BrownM, SinacoreDR, Steger-MayK, YarasheskiKE, SchechtmanKB. Effects of extended outpatient rehabilitation after hip fracture: a randomized controlled trial. JAMA. 2004;292(7):837-846. [PubMed: 15315998]
- 5.
KiresukTJ, ShermanRE. Goal attainment scaling: a general method for evaluating community mental health programs. Community Ment Health J. 1968;4(6):443-453. [PubMed: 24185570]
- 6.
RockwoodK, FayS, SongX, MacKnightC, GormanM; Video-Imaging Synthesis of Treating Alzheimer's Disease (VISTA) Investigators. Attainment of treatment goals by people with Alzheimer's disease receiving galantamine: a randomized controlled trial. CMAJ. 2006;174(8):1099-1105. [PMC free article: PMC1421447] [PubMed: 16554498]
- 7.
RockwoodK, StadnykK, CarverD, et al.A clinimetric evaluation of specialized geriatric care for rural dwelling, frail older people. J Am Geriatr Soc. 2000;48(9):1080-1085. [PubMed: 10983907]
- 8.
Turner-StokesL. Goal attainment scaling and its relationship with standardized outcome measures: a commentary. J Rehabil Med. 2011;43(1):70-72. [PubMed: 21174054]
- 9.
Alzheimer's Association. California Alzheimer's statistics. 2011 Alzheimer's Disease Facts and Figures Report. Accessed February 1, 2013. https://www.alz.org/national/documents/Facts_Figures_2011.pdf
- 10.
BernardHR, RyanGW. Analyzing Qualitative Data: Systematic Approaches. Sage Publications; 2010.
- 11.
PopeC, ZieblandS, MaysN. Qualitative research in health care. Analyzing qualitative data. BMJ. 2000;320:114-116. [PMC free article: PMC1117368] [PubMed: 10625273]
- 12.
MaysN, PopeC. Qualitative research in health care. Assessing quality in qualitative research. BMJ. 2000;320:50-52. [PMC free article: PMC1117321] [PubMed: 10617534]
- 13.
KauferDI, CummingsJL, KetchelP, et al.Validation of the NPI-Q, a brief clinical form of the Neuropsychiatric Inventory. J Neuropsychiatry Clin Neurosci. 2000;12:233-239. [PubMed: 11001602]
- 14.
PfefferRI, KurosakiTT, HarrahCHJr, ChanceJM, FilosS. Measurement of functional activities in older adults in the community. J Gerontol. 1982;37(3):323-329. [PubMed: 7069156]
- 15.
AlexopoulosGS, AbramsRC, YoungRC, ShamoianCA. Cornell Scale for Depression in Dementia. Biol Psychiatry. 1988;23(3):271-284. [PubMed: 3337862]
- 16.
FolsteinMF, FolsteinSE, McHughPR. “Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189-198. [PubMed: 1202204]
- 17.
ThorntonM, TravisSS. Analysis of the reliability of the modified caregiver strain index. J Gerontol B Psychol Sci Soc Sci. 2003;58(2):S127-S132. [PubMed: 12646602]
- 18.
MeyerPJ. What would you do if you knew you couldn't fail? Creating S.M.A.R.T. Goals. In: MeyerPJ. Attitude Is Everything: If You Want to Succeed Above and Beyond. Paul J. Meyer Resources; 2003.
- 19.
Krasny-PaciniA, PaulyF, HiebelJ, GodonS, Isner-HorobetiME, ChevignardM. Feasibility of a shorter Goal Attainment Scaling method for a pediatric spasticity clinic - the 3-milestones GAS. Ann Phys Rehabil Med. 2017;60(4):249-257. doi:10.1016/j.rehab.2017.01.005 [PubMed: 28365157] [CrossRef]
- 20.
FriedTR, McGrawS, AgostiniJV, et al.Views of older persons with multiple morbidities on competing outcomes and clinical decision-making. J Am Geriatr Soc. 2008;56:1839-1844. [PMC free article: PMC2596278] [PubMed: 18771453]
- 21.
TinettiME, NaikAD, DodsonJA. Moving from disease-centered to patient goals-directed care for patients with multiple chronic conditions: patient value-based care. JAMA Cardiol. 2016;1:9-10. [PMC free article: PMC6995667] [PubMed: 27437646]
- 22.
LynnJ, McKethanA, JhaAK. Value-based payments require valuing what matters to patients. JAMA. 2015;314:1445-1446. [PubMed: 26378889]
- 23.
BaylissEA, BondsDE, BoydCM, et al.Understanding the context of health for persons with multiple chronic conditions: moving from what is the matter to what matters. Ann Fam Med. 2014;12:260-269. [PMC free article: PMC4018375] [PubMed: 24821898]
- 24.
NaikAD, MartinLA, MoyeJ, et al.Health values and treatment goals of older, multimorbid adults facing life-threatening illness. J Am Geriatr Soc. 2016;64:625-631. [PMC free article: PMC5001155] [PubMed: 27000335]
- 25.
BogardusSTJr, BradleyEH, TinettiME. A taxonomy for goal setting in the care of persons with dementia. J Gen Intern Med. 1998;13:675-680. [PMC free article: PMC1500896] [PubMed: 9798814]
- 26.
BradleyEH, BogardusSTJr, van DoornC, et al.Goals in geriatric assessment: are we measuring the right outcomes?Gerontologist. 2000;40:191-196. [PubMed: 10820921]
- 27.
DucharmeJK, GeldmacherDS. Family quality of life in dementia: a qualitative approach to family-identified care priorities. Qual Life Res. 2011;20:1331-1335. [PubMed: 21279738]
- 28.
PhinneyA, ChaudhuryH, O'ConnorDL. Doing as much as I can do: the meaning of activity for people with dementia. Aging Ment Health. 2007;11:384-93. [PubMed: 17612802]
- 29.
KuluskiK, GillA, NaganathanG, et al.A qualitative descriptive study on the alignment of care goals between older persons with multi-morbidities, their family physicians and informal caregivers. BMC Fam Pract. 2013;14:133. [PMC free article: PMC3844424] [PubMed: 24010523]
- 30.
Turner-StokesL. Goal attainment scaling (GAS) in rehabilitation: a practical guide. Clin Rehabil. 2009;23:362-370. [PubMed: 19179355]
- 31.
OttenbacherKJ, CusickA. Goal attainment scaling as a method of clinical service evaluation. Am J Occup Ther. 1990;44:519-525. [PubMed: 2353720]
- 32.
Krasny-PaciniA, HiebelJ, PaulyF, GodonS, ChevignardM. Goal attainment scaling in rehabilitation: a literature-based update. Ann Phys Rehabil Med. 2013;56:212-230. [PubMed: 23562111]
- 33.
Bovend'EerdtTJ, BotellRE, WadeDT. Writing SMART rehabilitation goals and achieving goal attainment scaling: a practical guide. Clin Rehabil. 2009;23:352-361. [PubMed: 19237435]
- 34.
ClineDW, RouzerDL, BransfordD. Goal-attainment scaling as a method for evaluating mental health programs. Am J Psychiatry. 1973;130:105-108. [PubMed: 4682740]
- 35.
RockwoodK, HowlettS, StadnykK, CarverD, PowellC, StoleeP. Responsiveness of goal attainment scaling in a randomized controlled trial of comprehensive geriatric assessment. J Clin Epidemiol. 2003;56:736-743. [PubMed: 12954465]
- 36.
RockwoodK, FayS, SongX, MacKnightC, GormanM. Attainment of treatment goals by people with Alzheimer's disease receiving galantamine: a randomized controlled trial. CMAJ. 2006;174:1099-1105. [PMC free article: PMC1421447] [PubMed: 16554498]
- 37.
RockwoodK, HowlettSE, HoffmanD, SchindlerR, MitnitskiA. Clinical meaningfulness of Alzheimer's Disease Assessment Scale-Cognitive subscale change in relation to goal attainment in patients on cholinesterase inhibitors. Alzheimers Dement. 2017;13(10):1098-1106. [PubMed: 28341540]
- 38.
TotoPE, SkidmoreER, TerhorstL, RosenJ, WeinerDK. Goal Attainment Scaling (GAS) in geriatric primary care: a feasibility study. Arch Gerontol Geriatr. 2015;60(1):16-21. [PubMed: 25465505]
Acknowledgment
Research reported in this report was [partially] funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (#ME-1303-5845) Further information available at: https://www.pcori.org/research-results/2013/testing-way-help-patients-dementia-and-their-caregivers-set-goals-care
Appendices
Appendix A.
Patient and Caregiver Goal Study Phase 1, 2 and 3 Protocol (PDF, 305K)
Appendix B.
Patient- and Caregiver-Identified Goal Domains and Discrete Goals for Dementia Care (PDF, 125K)
Appendix C.
UCLA DCM Focus Group Summary (PDF, 112K)
Original Title: Developing Patient-centered Outcomes for Dementia: Goal Setting and Attainment
PCORI ID: ME-1303-5845
Suggested citation:
Reuben DB, Wenger N, Jennings L, Hays R. (2019). Testing a Way to Help Patients with Dementia and Their Caregivers Set Goals for Care. Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/10.2019.ME.13035845
Disclaimer
The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.
