Social Science & Medicine 49 (1999) 267278

Goal-setting in clinical medicine

Elizabeth H. Bradley a,*, Sidney T. Bogardus Jr. b, Mary E. Tinetti b, Sharon K. Inouye b
a

Department of Epidemiology and Public Health, Yale School of Medicine, New Haven, CT 06520-8034, USA b Department of Internal Medicine, Yale University School of Medicine, New Haven, CT 06520-8034, USA

Abstract The process of setting goals for medical care in the context of chronic disease has received little attention in the medical literature, despite the importance of goal-setting in the achievement of desired outcomes. Using qualitative research methods, this paper develops a theory of goal-setting in the care of patients with dementia. The theory posits several propositions. First, goals are generated from embedded values but are distinct from values. Goals vary based on specic circumstances and alternatives whereas values are person-specic and relatively stable in the face of changing circumstances. Second, goals are hierarchical in nature, with complex mappings between general and specic goals. Third, there are a number of factors that modify the goal-setting process, by aecting the generation of goals from values or the translation of general goals to specic goals. Modifying factors related to individuals include their degree of risk-taking, perceived self-ecacy, and acceptance of the disease. Disease factors that modify the goal-setting process include the urgency and irreversibility of the medical condition. Pertinent characteristics of the patientfamilyclinician interaction include the level of participation, control, and trust among patients, family members, and clinicians. The research suggests that the goal-setting process in clinical medicine is complex, and the potential for disagreements regarding goals substantial. The nature of the goal-setting process suggests that explicit discussion of goals for care may be necessary to promote eective patientfamily clinician communication and adequate care planning. # 1999 Elsevier Science Ltd. All rights reserved.
Keywords: Goals; Doctorpatient relationship; Communication; Dementia

Introduction Goals, dened in previous research as the `internal representations of desired states' (Austin and Vancouver, 1996), are pervasive motivating forces in human behavior. Goals are the object or aim of an action (Locke et al., 1981), and may be considered

* Corresponding author. Tel.: +1-203-785-2937; fax: +1203-785-6287. E-mail address: elizabeth.bradley@yale.edu (E.H. Bradley)

synonymous with other terms such as aim, objective, object, intent, purpose, end, or desired outcome (Locke and Latham, 1990). Although the topic of goal-setting has received relatively little attention in the medical literature, evidence from the social sciences suggests that the articulation of goals may be fundamental to planning appropriate medical care and achieving desired clinical outcomes. Goal-setting may enhance both the process and outcome of clinical care in several ways. First, goals may help clinicians link recommended plans of care to desired outcomes. Outcomes-based plans of care can

0277-9536/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 9 5 3 6 ( 9 9 ) 0 0 1 0 7 - 0

268

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

help ensure the relevance of both diagnostic and therapeutic interventions. Second, explicit linking of plans of care with desired outcomes may clarify for patients and family members the purpose of various recommendations. If patients and family members understand the purpose of various recommendations, they may be more likely to accept and adhere to recommended plans for care. Third, explicit discussion and articulation of goals may help resolve conicts among clinicians, family members, and patients about desired outcomes. Conicts about goals, both among clinicians and between clinicians and patients or families, may be especially prevalent in cases of chronic illness when multiple parties are often involved in a patient's care, and the goals of care are not always obvious since `cure' is not feasible. In such cases, discussions of goals of care may expedite identication of sources of disagreement and negotiation of shared goals, resulting in improved patient and family satisfaction with care. By promoting outcomes-based plans of care, enhancing acceptance of clinical recommendations, and facilitating the resolution of conicting goals, a clear statement of goals may increase the chances of achieving desired outcomes, a result valued by clinicians, patients, and families. Most previous research on goal-setting has been in psychology and organizational behavior. These studies suggest that explicit goal-setting is fundamental to goal achievement (Locke and Latham, 1990; Hadavi and Krizek, 1993). Goal-setting has been found to have a positive eect on several important outcomes, including individual behavior (Burke et al., 1978), worker productivity (Locke and Latham, 1990; Hadavi and Krizek, 1993), and job satisfaction (Pearson, 1987). Further, this body of research suggests that goal-setting is most eective when the goals are specic (Early et al., 1987; Locke et al., 1989), appropriately challenging (Locke et al., 1989; Locke and Latham, 1990), and developed in a participative manner (Dachler and Wilpert, 1978; Tjosvold, 1988; Tjosvold et al., 1992). In contrast to the large body of research on goals in psychology and organizational behavior, evidence concerning the nature and inuence of goals in clinical medicine is more limited. Evidence from the treatment adherence literature supports the importance of communication in enhancing adherence (Carter et al., 1982; Stewart, 1984; Squier, 1990), although the role of goals per se has not been assessed. Additional literature on goals in clinical medicine has focused on a technique for monitoring patients' progress toward pre-established goals. This technique, known as `goalattainment scaling' (Kirusek et al., 1994) has been shown to be useful in the care of patients with brain injuries (Joyce et al., 1994) and with infants with motor delays (Palisano et al., 1992; Palisano, 1993).

This literature suggests that the goal-attainment scaling technique focuses multidisciplinary teams on specic problems and helps to evaluate their progress in resolving these dened problems (Malec et al., 1991; Joyce et al., 1994). Applied to geriatric care, the goal-attainment scaling technique has been found to facilitate selection of pertinent patient-specic goals and ecient evaluation of clinically important changes (Rockwood et al., 1993; Rockwood et al., 1996; Hartman et al., 1997). Several recent articles in both the nursing (Ryden and Feldt, 1992; Ugarriza and Gray, 1993) and the medical literature (Mold et al., 1991; Sachs, 1997; Von Kor et al., 1997; Bogardus et al., 1998) have discussed the importance of caregiving goals. However, although both goalattainment scaling and discussions of goals have shown promise in specic settings, less attention has been focused on the goal-setting process, itself. This paper focuses on the care of patients with dementia as a case study of the larger topic of goalsetting in clinical medicine because dementia represents a prevalent, chronic disease for which establishing goals is particularly complex. Dementia is a disease for which `medical cure', in the usual sense, is typically not a feasible goal. Thus, the predominant objective in the care of persons with dementia is adaptation to the condition. Because the meaning of adaptation to disability is multifaceted and subjective, articulating goals of care may be more complex than in cases where curing disease is an option. In addition, dementia care often involves both formal and informal care givers, each with potentially dierent perspectives and opinions about appropriate goals and plans for care. Patients' cognitive capacity to participate in discussions of care planning is often unclear and dicult to assess (Lo, 1990; Kapp and Mossman, 1996), and family members and clinicians often must develop plans for care without full participation of the patient due to the patient's inability to reason about and express their goals. The inevitable complexity in medical decision-making for this population presents an opportunity to witness and describe a range of factors in the goal-setting process and a variety of patient and family roles in this process. The objective of this paper is to develop a theory of goal-setting applicable to clinical medicine. The theory was derived empirically; data from open-ended interviews with patients, family members, and clinicians were categorized to create the framework for the goal-setting theory. A better understanding of goal-setting in clinical medicine may help to enhance the process of clinical care by revealing potential causes of conicts in goals and identifying opportunities for focused communication and negotiation of shared goals.

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

269

Methodology Study design The study involved a qualitative analysis of openended interview data from consecutive patients seen in the outpatient geriatric assessment center of Yale/New Haven Hospital, a major teaching hospital. The study proceeded with the following steps: (1) open-ended interviews with patients, their primary family caregiver, and clinicians; (2) transcription of audiotape of each interview; (3) categorization of goals and factors in goal-setting using line-by-line review of transcripts by the investigative team; (4) resolution of discrepancies in categorizations among investigators using a process of negotiation and consensus-building; (5) development of an integrated conceptual framework incorporating the categorizations, i.e., development of the goal-setting theory. Setting The study was performed at the geriatric assessment center aliated with Yale/New Haven Hospital, a center that conducts approximately 50 new patient assessments each month and receives referrals from a wide geographical area. About 30% of new patients are referred by their primary care physicians; 20% are referred by a community-based agency (such as a home care agency or an adult day care center) or other institutions, and 50% come at the direct request of their families. Geriatric assessment represents a comprehensive approach to caring for elderly patients and their families. An assessment involves examination of a number of clinical factors that may limit the cognitive, emotional, physical, or social functioning of patients. Interdisciplinary teams are essential to geriatric assessment, a process that often involves medical and cognitive examination of the patient as well as interviews with family members or other informal caregivers. In this study, geriatric assessments typically lasted for one and a half to two hours, and often resulted in followup visits to the center for further evaluation. Because geriatric assessment focuses on multiple aspects of the patient's health and future needs for care, it represented an appropriate setting in which to observe the goal-setting process taking place. Participants Consecutive new patients presenting to the center for an assessment were enrolled in the study and interviewed. The primary care giver, case manager, and physician for each patient was also enrolled in the study and interviewed. Thus, for every patient, four

interviews (with patient, primary care giver, case manager, and physician) were attempted. Interviews with patients and primary caregivers were conducted on the day of their initial visit to the center, at the conclusion of the clinical encounter. Case managers and physicians were interviewed within 48 h of the completed assessment. The study protocol was approved by the Human Investigation Committee of Yale School of Medicine; written consent was obtained from the primary care giver, and verbal assent from the patient, prior to the interviews. Interviews were conducted until no new concepts were elicited, i.e. until the point of `theoretical saturation' (Glader and Strauss, 1967). Ten sets of participants were enrolled; two patients could not be interviewed due to the severity of their cognitive impairment; one family caregiver refused interview, and one physician was not available for interview. Thus, a total of 36 interviews were completed. Data collection Patient data, including sociodemographic information, functional status, and cognitive status, were abstracted from clinic medical records. Functional status was measured with instrumental and basic activities of daily living IADL, BADL). Each of seven IADL were scored on three levels: needs no assistance (scored 0), needs some assistance (scored 1), needs much assistance (scored 2). The scores were summed for the total IADL score, with higher scores reecting higher functioning. Activities included telephone use, shopping, meal preparation, housework, transportation, ability to take medications, and management of nances. Similarly, each of seven basic activities of daily living were scored on the same 02 scale and summed to create a BADL score, with higher scores reecting higher functioning. Basic activities included eating, dressing, grooming, walking, transferring, bathing, and continence. The Folstein Mini-Mental State score (Folstein et al., 1975) was used as a measure of cognitive impairment. Data about goals and goal-setting were collected using open-ended interviews, with two or three investigators present at each interview and during subsequent analysis of interview data. To enhance the collection and interpretation of data, the research team included four investigators with varying backgrounds, including health services research, medicine, and public health. Interviews generally required about 1030 min to complete and were audiotaped and transcribed for precision and accuracy. Interviews included broad, open-ended questions to elicit goals for health care and how goals were developed, i.e. the factors involved in selecting goals. Because little is known about the language used by patients, care givers, and clinicians to describe goals, a more structured interview might have constrained re-

270

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278 Table 1 Patient characteristics Characteristic Age < 75 7680 >80 Gender Male Female Mental status Widowed Married Never married Primary care giver Spouse Child Other Cognitive function MMSEa <18 MMSE 1823 MMSE >24 IADL scoreb Mean (range) BADL scorec Mean (range)
a b

sponses, limiting the accuracy and depth of the inquiry. The interview guide began with a broad question, ``what are your goals for the patient's (or your) care?'' If participants appeared to have diculty with the term, `goal', alternative language was used, including ``desired outcomes'' or ``what you would like to achieve''. Standardized probes were used to prompt discussion of both general and specic goals, their optimism in achieving the goals, and how they arrived at the goals. In all cases, participants were asked to expand their responses with examples and detailed stories that illustrated their statements. To enhance the direct quotations from the audiotape, data concerning body language, tone, and emotional responses were noted by the interviewers.

Number 1 6 3 3 7 4 5 1 2 6 2 5 3 2 6.1 (0, 14) 11.6 (6, 14)

Data analysis Data were analyzed using the `constant comparative method' of qualitative data analysis (Glaser and Strauss, 1967). The investigators jointly reviewed each transcript line by line, coding comments and quotations that appeared to embody distinct concepts. Consistent with qualitative data analysis techniques, interview data were analyzed soon after each interview was completed and prior to completing all interviews. During the coding process, new interview data were compared with previous quotations or comments in the same category. Similar concepts were grouped into a list of distinct categories as negotiated by the investigators. Categories were expanded or combined as transcripts were analyzed. New probes were added to successive interviews when data analysis suggested the existence of new categories to ensure that all concepts and their dimensions were fully explored. Transcripts were reviewed several times over the course of the study to ensure all relevant data were accounted for and appropriately assembled under dened categories. This data analysis process required a total of approximately one and a half hours of research team meetings per transcript. At the conclusion of the interviews, when no new concepts were being elicited in the interviews, categories and their properties were described in full by the investigators through a series of group meetings in which diering viewpoints were explicitly discussed and considered. Dierences in interpretation were discussed and negotiated until consensus was reached. The objective of this process was to attain valid interpretations concerning the integration of quotations, the meaning of conceptual categories, and the interrelationship of various concepts. This synthesis of concepts formed the basis of the theory of goal-setting developed in this paper.

Mini mental state examination. Instrumental activities of daily living score. c Basic activities of daily living score.

Findings Participant characteristics Thirty-six interviews were completed. Of nine family interviews, three included two family members, and all nine included the primary family caregiver. The ten case manager interviews were performed among three case managers, and the nine physician interviews were performed among six geriatricians, capturing the full spectrum of clinicians practicing at the study site. As displayed in Table 1, the average age of the patients was 72 years, with a range from 69 to 84. Seventy percent of the patients were female, and patients represented a broad range of cognitive and functional ability, and family structure. A theory of goal-setting in clinical medicine The proposed theory of goal-setting in the care of chronically ill patients encompasses four fundamental components: values (as distinct from goals), a hierarchy of goals (general versus specic goals), the relationships among dierent goals, and modifying factors that inuence the translation of general goals

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

271

Fig. 1. Theory of goal setting in clinical medicine.

into selected specic goals. These components interrelate as displayed in Fig. 1, and will be further described below.

She likes to be her own person ... She is very independent. She likes doing things on her own. (#3F) The distinction between values and goals is subtle; however, recognition of this distinction may be important for directing care decisions that depend on specic clinical circumstances. For instance, in the case cited, although the mother generally desired to be `on her own', it was unclear what her goal might be in any specic circumstance in which alternatives and trade os with other desires may be necessary. Discussions of values without discussion of goals, themselves, may not be sucient for planning appropriate care. The hierarchy of goals: higher and lower order goals Previous literature on the nature of goals describes a hierarchy of goals (Locke and Latham, 1990; Austin and Vancouver, 1996). The hierarchy of goals includes both general (or `higher order') and specic (or `lower

Values and goals Several individuals described not only goals but also what we term, `values'. In this study, values refer to fundamental beliefs about global states of being and are individual-specic. Values remain stable despite changing circumstances. They are general in nature and reect an individual's self-concept. Although goals can correspond closely to values (Austin and Vancouver, 1996), values are specic to the individuals whereas goals may depend on an individual's context or circumstance. To demonstrate the description of values when discussing her goals, one care giver said of her mother,

272

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

order') goals. General goals are situation-specic and pertain to one's overall desired state given contextual factors and feasible alternative states. Specic goals are linked to immediate tasks at hand. Specic goals are aimed at components of the general goal, and thus help move an individual along the path toward the general goal. Almost always, there are many dierent possible paths towards a general goal, and hence there may be many dierent possible specic goals. As an example of a general goal, the spouse of a patient who was experiencing a precipitous decline in cognition expressed this goal, I'd like to see him more or less stabilized, closer to the way he used to be ... to make him normal. (#2F) In contrast, a specic goal was expressed by this physician: To establish whether she has an isolated memory impairment or dementia; to further monitor her to conrm our initial impression that she does not have depression, and to make sure that she doesn't lose any more weight. (#4D) Based on the construct of hierarchical goals, the interview data illustrated that the range of general goals can be characterized by a limited set of goal categories. These categories, described below, include safety, independence and day-to-day functioning, social and family relationships, personal health, economic stability, and autonomy or dignity. Safety refers to protection from harm of any kind including physical, psychological, or nancial harm. Independence is the ability to function and/or live without assistance or with minimal assistance. Social and family relationships refer to having positive relationships with family members and friends, and participation in leisure or work activities. Personal health comprises the broadest conception of health, i.e. the absence of physical, emotional, mental, and/or spiritual disabilities. Economic stability refers to a number of nancial concerns such as future nancial planning and maintaining adequate economic resources. Autonomy refers to the ability to make one's own decisions about care, treatment options, and choices in life, and dignity, a related concept, means the maintenance of one's respect by self and others. As an expression of the goal `safety', a daughter-inlaw stated about her parents-in-law, I would like to see them safely taken care of. I want someone to be sure they are safe, and I want them to let [their son] take care of them. They are

not eating right, not cooking. She couldn't follow a recipe if she had to. (#1F) Articulating the goal of `independence and day-today functioning', one patient responded, I drive my own car; I get groceries; I do everything myself. When I couldn't do things on my own, then I'm an old lady. I want to be myself. I don't want to bother them; they have their children and are working. I want to stay as I am. (#3P) Expressing a goal related to `social and family relationships', one patient said, I'd like to get well enough to travel again ... to do things like visit cousins and play cards with my dear friends ... I miss them. (#8P) The goal that we term `personal health' encompassed physical, cognitive, emotional, and behavioral aspects of health, as well as expressions of general well-being. As a daughter expressed her goal for her mother's care, The short term goal is we want her to stay healthy ... make sure she gets the proper care she needs. I just want her to stay healthy; that's all. The longer term goal is that she stay healthy. (#3F) `Economic stability' was articulated as a goal by some patients and caregivers. One daughter considering nursing home placement for her father stated, I am of course concerned about the monetary situation ... because the pocket book has been eaten up recently from overnight care, day care, and the rest ... (#6F) Finally, `autonomy' and `dignity' were expressed as goals in several cases. For example, caregivers described as their goals, That she have her own space, her own belongings ... she is so attached to her own belongings. She has that little bit of freedom to move around her building. A more rigid situation might be detrimental to her right now. (#5F)

Basically it [the goal] is just to create variety and a sense of dignity in life style ... (#7F) In addition to these general goals, respondents articulated many specic goals, which spanned a large range of items and which varied in terms of their level of diculty or challenge, the time frame for attaining

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

273

the goal, and their level of specicity. A taxonomy that characterizes goals along these dimensions has been previously described based on these data (Bogardus et al., 1998). Examples of specic goals included ascertaining a diagnosis, increasing supervision for the patient, promoting acceptance of the disease by family members, reducing aggressive behavior of the patient, or enhancing family members' knowledge about future care needs. The participants in the study articulated more than 40 dierent lower order goals, and the list of possible lower order goals is extensive. Integration among goals Fundamental to goal-setting is the integration of goals. This refers to both the relationships among specic or among general goals, as well as the integration of specic goals with general goals. The relationships among goals can be complex, and one individual may have multiple goals. At times, these goals may be well-ordered in their importance. As an example of multiple and well-ordered goals, the son of a patient described his goals for the care of his mother, Her well-being is probably number one. And to make sure she has someone there. Flexibility is another thing that's pretty important too. And her safety and comfortability [sic] are important to us too. (#8F) In this cases and many others, the goals are not conicting. In fact, it is likely that all goals can be met simultaneously. However, in other cases, multiplicity of goals led to conict among goals expressed by one person. In the following instance, given the patient's cognitive impairment, the caregiver's goal of independence may threaten the caregiver's goal of patient safety. A daughter-in-law speaking of her husband and his parents, who were patients at the assessment center, said: They [the parents] had an accident in the previous car. I'm just afraid they are going to get lost, but my husband hasn't wanted to limit them, especially their mobility. (#1F) In addition to the relationships and potential conicts among general goals, the translation of general goals into specic goals demonstrated substantial complexity and variability among the participants. As an example, the general goal of maintaining a patient's safety may map to any number of specic goals including increasing supervision with home care sta, having the patient live with relatives, or modifying sources of

potential harm in the patient's home. In this example, the general goal of enhancing safety maps to multiple specic goals. Conversely, a single specic goal may map to multiple general goals. For instance, the specic goals of gaining knowledge about available adult day care services may map to both enhancing the patient's social connectedness and promoting his/her safety. Several themes that emerged from the interviews suggested that this mapping, or translation, of general goals to specic goals may be governed by a set of factors that we have termed modifying factors in the goal-setting process. Modifying factors in the goal-setting process The data suggested that several factors may inuence the setting of goals by patients, family members, and clinicians. These factors can be characterized in three broad categories: characteristics of individuals, characteristics of disease or illness, and characteristics of patientcaregiverphysician interactions. The categories encompass a range of factors that may inuence the translation of general goals into specic goals. Characteristics of individuals Characteristics of individuals that inuenced goalsetting by participants in the study included risk-taking, personal self-ecacy, and the degree of acceptance of the patient's current state. Respondents varied in the degree to which they expressed risk-taking, and willingness to address an existing problem in the face of potential costs of intervention(s), either medical, behavioral, or social. The willingness to take risks was fundamental to the choice (or non-choice) of specic goals. As expressed by one patient, Truthfully, we'd be just happy to continue as we are. We're very simple people ... you just take it as it is handed to you. (#1P) In contrast, a more risk-taking approach was voiced by a daughter and related by a case manager, We actually told the patient we thought she had early Alzheimer's disease, and I heard from the daughter the next day that [her] mom had not slept all night because she was so distressed by that fact [the diagnosis], and they decided that they would try Aricept just because she wanted to try something. (#3CM) The quotations, which are representative of a number of such expressions, illustrate the variation in willingness to take action in order to potentially improve either general health or specic aspects of functional or cognitive status.

274

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

A second characteristic of the individuals that emerged as important in goal-setting was personal selfecacy. Self-ecacy refers to an individual's perceptions about his or her ability to perform a task (Locke and Latham, 1990). A patient's low self-ecacy was expressed by one daughter, I expect there will be problems. At [her] age, you expect [she's] not going to be realistically able to do everything. You try to get everything organized as best you can. She won't make a move on her own ... They'd sit right where they are until the bottom fell out. (#4F) In contrast, a physician noted high self-ecacy in achieving her goals for the patient's care. [The goal is] not to have the patient drive and [for her] to be accepting of somebody coming into the home to help. The driving goal will happen because we are going to make it happen by calling the Department of Motor Vehicles. (#2D) The data suggested that perceptions of self-ecacy may inuence not only the likelihood of setting explicit goals but also the nature of goals themselves. Low self-ecacy resulted in either the setting of no or less challenging goals. The eect of self-ecacy was evidenced among all types of respondents patients, family members, and clinicians. In addition to risk-taking and self-ecacy, individuals' varying levels of acceptance of or resistance to the clinical circumstance seemed to inuence the process of setting specic goals. Previous research on individuals' willingness to adhere to treatment regimens or change an approach to care has suggested that individuals move along a trajectory of resistance, acceptance, and adaptation to diseases and conditions (Prochaska et al., 1992). Although the cross-sectional data from this study could not conrm the degree to which individuals move through these stages, variation in the level of acceptance and adaptation among respondents was apparent. In addition, some respondents suggested that their understanding and acceptance had changed over time. One spouse referring to her severely impaired husband said, Up to now, I didn't want to face it. But the doctors said it had been coming on for a while, but there was no way we saw it. We ignored it I guess, or it seemed like the symptoms would come and go. Some people told me, but I didn't really see it the inevitable would be a convalescent home, but I don't know about being ready to do that. (#9F) Other individuals describe more acceptance of their condition,

I think your mind doesn't retain everything that you've learned. You think, ``I could sign and cosign anything when I was in college, but that's gone by the board.'' I think it is because you just don't have access to it anymore, but I like my life. (#8P) In some cases, patients, families, and clinicians varied substantially in their degree of acceptance of the patient's conditions. For instance, a patient related, [My daughter] thought I was getting feeble. I think I am ne. I do everything, I do my washing. I shoveled snow this morning. My kids think I am an old woman, but I don't. (#3P) The daughter of this same patient said, I had a concern about my mom forgetting things. I wanted to know if this was normal progression. She doesn't think there is anything to worry about. We tried to do planning for the future. (#3F) The case manager expressed her role as helping move the daughter to acceptance of Alzheimer's disease in the mother. She related, Our role is to help her daughter think about changes that were going to be happening probably over the next year ... so just help her daughter think about what was probably Alzheimer's disease and what's going to happen to her. They need to start planning. (#3CM) The physician summed up the visit, In terms of the patient, her primary goal was to get out of here. She didn't recognize much of a problem at all. The discussion of diagnosis and potential implications came as a shock to her. (#3D)

Characteristics of the disease or illness Characteristics of the disease that may modify the process of setting goals include the reversibility of the disease and the urgency of the problem(s) presented. For irreversible conditions, specic goal choices involving cure are eliminated, and the general goal involving personal health might be substantially threatened. Further, the permanence that irreversibility implies suggests that adjustments and adaptations are necessary for the long term, not only for a limited time. This permanence may limit the feasible goals and bring a dierent set of concerns to the goal-setting process, fundamentally aecting the translation of general to specic goals. One patient described her goals,

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

275

What can you do? You take one day at a time. There's not much you can do at our age ... I'm alive and kicking, and that's it. I'd like to continue like this. (#4P) Similarly, the urgency of the presenting problems may aect the process of goal-setting. While a sense of urgency promoted the articulation of specic goals, it also delayed formulation of longer term goals. As stated by a case manger very concerned about a patient's safety, We're not chasing a diagnosis now. He was diagnosed in 1993 with Alzheimer's. Diagnosis is not where it is at; emergency measures is where it is at. These are emergency measures now F F F What I did not deal with was the long range planning. (#2CM) The apparent level of urgency in this expression contrasted to many instances in which the situation was perceived as less dire. Perceptions of urgency were not always uniform across the patient, caregiver, case manager, and physician.

Control, as distinct from participation, in decisionmaking was a factor that also aected the goal-setting process. Central to the care of patients with dementia, the degree to which patients' control in decision making was limited highlighted a dicult tension in goalsetting. This tension is suggested by one physician who described goal-setting for a patient with moderate cognitive impairment, There is sort of an interesting decision about how autonomous the patient can be. It's sort of a funny justication of trying to maintain as much independent decision making for her as we can. We are trying to preserve [her] independence in getting to the end point that we unilaterally have decided she needs to go. (#2D) Issues of control were also apparent within families, in which informal care were often faced with limiting the control of the patient out of necessity. One case manager describes, The son came last week with his father because his father had become paranoid and agitated. The son was frightened by that. This was the rst time the son had wrested control and did something. If the son is not cooperative, then we will have to call Protective Services to help do what he cannot do. These two [the father and mother] are at risk. (#1CM) The quotation demonstrated a common complexity in goal-setting which is exacerbated in the care of dementia. Understanding who is in control and agreeing on who should be in control of decision making is frequently unclear or contested. Yet, the distribution of control remains central to both the choice and acceptance of the goals for care. The nal factor that seemed to inuence the goalsetting process was trust. Many have written on the inuence of trust in the medical encounter (Anderson and Dedrick, 1990; Cassel, 1991; McWilliam et al., 1994; Scott et al., 1995; McWilliam et al., 1997), demonstrating in a variety of medical situations that trust can enhance compliance, satisfaction, and outcomes of care. This study suggests that one of the mechanisms by which trust may inuence care is through the goal-setting process. Trust seemed to help patients and families accept recommendations from clinicians. Conversely, distrust eliminated whole sets of possible goals. For instance, in reference to potential nursing home placement of her father, one daughter expressed her distrust of nursing homes, You worry about them taking him to the bathroom or feeding him or just caring. If they should refuse

Characteristics of the interaction In addition to the modifying factors related to individuals or diseases, characteristics of the interaction among patients, families, and clinicians may inuence the goal-setting process. The characteristics of the interactions identied in this study were the degree of participation in the decision making process, perceived control in decision making, and the level of trust among the patient, family member, and clinician. The data revealed that patients often were not full participants in decisions regarding plans for care. In many cases, this was due to the cognitive incapacity and lack of insight of patients to participate meaningfully in decisions. Patients were often brought to the geriatric assessment under pressure from family members or referring physicians. One patient describes, I came because my daughter told me to come. She thought I was getting feeble (patient laughed, believing this was not a legitimate concern). (#3P) Alternatively, a physician described a more participative approach to identifying goals and planning care, one that involved patients and families more fully. The rst thing was to try to nd out from the family what were their questions. Then I talk to the patient directly to kind of get a feel from her of why she was here, what she wanted from the evaluation. I don't know that I ever say, `goals' but [rather] ``what is it that you want out of this visit?'' Then we try to work together. (#8D)

276

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

to take their medication the rst time, they [the sta] just don't have to give it to him ... they [nursing homes] are so scary. (#6F) The lack of trust of nursing home care resulted in the daughter's delay in considering this as a goal, despite recommendations of clinical sta. In a similar example, a daughter-in-law described bringing her father-in-law to the clinic, Getting him to go to any doctor at all is very hard. I have to lie to him. It is a matter of pushing him here. He thinks doctors are out to take his money. (#2F) The quotations demonstrated the challenge of setting goals without adequate trust. Given the myriad conicts in specic goals, general goals, and the integration of the two, trust may be a fundamental factor to facilitating communication in a way that shared goals can be negotiated eectively.

Discussion This study forms the basis of a theory of goal-setting for clinical medicine in the care of patients with dementia and other chronic illnesses. The propositions of the theory are summarized as follows. First, individuals come to the clinical encounter with a set of values. These values, which are specic to individuals not situations, are relatively stable, embedded in one's accumulated knowledge and experiences. From these embedded values are generated goals, or desired outcomes, which are specic to contextual factors. Second, there is a hierarchy of goals from general goals to more specic goals. General goals are closely related to values and typically seek to maximize particular values, under existing circumstances. Specic goals, given existing circumstances, govern specic actions. For instance, in the case of clinical medicine, specic goals may be used to generate plans of care and recommended treatments or interventions. Third, the relationships among general and specic goals is highly individualized and complex, necessitating discussions not only of specic goals but also how these goals relate to other goals. The translation of general goals into specic goals is the cornerstone of the goalsetting process. This translation is inuenced by a set of modifying factors. The factors can be broadly dened as characteristics of individuals involved in the clinical encounter, characteristics of the disease itself, and characteristics of interactions among patients, families, and clinicians. The distinction between general and specic goals is consistent with goal-setting literature in psychology

(Locke and Latham, 1990; Austin and Vancouver, 1996). While the concept of a hierarchy of goals is well accepted (Austin and Vancouver, 1996), the translation of general goals into specic goals that govern more immediate plans and behaviors is less well described. The translation is especially complex in clinical medicine in which identication of feasible and meaningful specic goals may require substantial experience and knowledge concerning diagnosis and treatment of disease. Patients and families, who possess general goals, must depend on clinicians' expertise for identifying specic recommendations. At the same time, clinicians, who are less knowledgeable about individual patients' and families' values and general goals, must form appropriate and feasible plans of care consistent with patients' and primary care givers' values and general goals. As a result, achieving consistent hierarchies of goals (and, thus, reasonable plans for care) may require substantially more communication and negotiation in clinical medicine than in tasks in which values and subjective judgments are not as fundamental. The research suggests a number of factors that may be inuential in the goal-setting process, or the selection of specic goals based on general goals and values. While previous research has identied dimensions of goals, themselves (Locke and Latham, 1990; Bogardus et al., 1998), little is known about the factors which modify the translation of values and general goals into specic goals. Because they form the basis for plans of care and choice of interventions, specic goals are at the heart of the clinical encounter. This research suggests that particular characteristics of the individual, the disease, and the interactions of patients, families, and clinicians are central components in the goal-setting process. Several implications are apparent from the proposed theory of goal-setting in clinical medicine. First, if goals are not articulated, it will be dicult to determine an appropriate plan of care. Second, conicting goals or varying weighting of goals may result in substantial disagreement concerning the plan for care. Third, even if general goals are shared and similarly weighted, specic goals among patients, families, and clinicians may conict due to the inuence of several modifying factors in the goal-setting process. Fourth, goal conict and potential disagreements concerning the plan of care may be minimized by modifying or understanding the factors in the goal-setting process. For instance, understanding the movement of individuals from non-recognition of the illness to acceptance and then to adaptation to the illness may focus individuals on specic goals of care that are most appropriate and plans for care that are most likely to be implemented. Finally, the theory suggests that, because the setting of specic goals necessitates the understand-

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278

277

ing of values and general goals, explicit participation by individuals with decisional capacity and building of trust among patients, families, and clinicians can enhance the goal-setting process and result in more appropriate, shared goals. There are several limitations of the study that warrant recognition. The study, using a purposeful sample, is preliminary in nature, designed to generate a new theory, rather than conrm aspects of existing theory. Additional studies that are quantitative in nature and include larger samples of patients, primary care givers, and clinicians would be useful to augment the proposed theory. For instance, availability of resources (of clinicians, families, and patients) may inuence the type of goals selected. Although our research in the outpatient setting did not reveal the availability of resources as a factor in goal-setting, resource availability may be more inuential in other medical arenas. In addition, although data collection and analyses were performed by multiple investigators from varying backgrounds, the study is based on qualitative data, and thus may have been subject to the biases of the investigators. Finally, although the study provides a framework for studying goal-setting in clinical medicine, the data are based on patients with dementia. Replication of these data among patients with other chronic diseases and in other clinical settings would augment the proposed goal-setting theory. Previous literature in medicine has shown that enhanced communication results in a number of positive outcomes, including greater patient and family satisfaction (Putnam et al., 1985; Buller and Buller, 1987; Bertakis et al., 1991; Ong et al., 1995), enhanced patient empowerment (McWilliam et al., 1994), greater adherence to treatment plans (Stewart, 1984; Ong et al., 1995), and improved health outcomes such as enhanced functioning, emotional health, and symptom relief (Stareld et al., 1981; Kaplan et al., 1989; Stewart, 1995). The consistency of the ndings is impressive; however, the causal mechanism for the result remains unknown. One mechanism may involve the process of goal-setting, which, when embraced in a participative and explicit way, may help patients, families, and clinicians forge shared and meaningful goals for care and, in turn, enhance desired outcomes of the health care process. Acknowledgements The authors are grateful to Terri Fried, Nora Groce, Sarah Horwitz, and Mark Schlesinger for their comments on previous versions of this paper. They would also like to thank Michael Jeong and Grace Shih for their research assistance, and Carolyn Cummings for her careful manuscript preparation. This research was

supported by a grant from the Claude D. Pepper Older Americans Independence Center of the Yale School of Medicine. References
Anderson, L.A., Dedrick, R.F., 1990. Development of the trust in physician scale: a measure to assess interpersonal trust in patientphysician relationships. Psychological Reports 67, 10911100. Austin, J., Vancouver, J., 1996. Goal constructs in psychology: structure, process, and content. Psychological Bulletin 120, 338375. Bertakis, K.D., Roter, D., Putnam, S.M., 1991. The relationship of physician medical interview style to patient satisfaction. The Journal of Family Practice 32 (2), 175181. Bogardus, S.T., Bradley, E.H., Tinetti, M., 1998. A taxonomy for goal-setting in the care of persons with dementia. Journal of General Internal Medicine 13, 675680. Buller, M.K., Buller, D.B., 1987. Physicians' communication style and patient satisfaction. Journal of Health and Social Behavior 28, 375388. Burke, R., Weitzel, W., Weir, T., 1978. Characteristics of eective employee performance review and development interviews: replication and extension. Personnel Psychology 31, 903919. Carter, W.B., Inui, T.S., Kukull, W.A., Haigh, V.H., 1982. Outcome-based doctorpatient interaction analysis. Medical Care 20 (6), 550566. Cassel, E., 1991. The Nature of Suering and the Goals of Medicine. Oxford University Press, New York. Dachler, H., Wilpert, B., 1978. Conceptual dimensions and boundaries of participation in organizations: a critical evaluation. Administrative Science Quarterly 23, 135. Early, P., Wojnaroski, P., Prest, W., 1987. Task planning and energy expended: exploration of how goals inuence performance. Journal of Applied Psychology 72, 107114. Folstein, M., Folstein, S., McHugh, P., 1975. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatry Research 12, 189198. Glader, B., Strauss, A., 1967. The Discovery of Grounded Theory. Aldine de Gruyter, New York. Hadavi, A., Krizek, R., 1993. Short-term goal-setting for construction. Journal of Construction Engineering and Management 119, 622630. Hartman, D., et al., 1997. Use of goal attainment scaling in a dementia special care unit. American Journal of Alzheimer's Disease 12 (3), 111116. Joyce, B., Rockwood, K., Mate-Kole, C., 1994. Use of goal attainment scaling in brain injury in a rehabilitation hospital. American Journal of Physical Medicine & Rehabilitation 73, 1014. Kaplan, S.H., Greeneld, S., Ware, J.E., 1989. Assessing the eects of physicianpatient interactions on the outcomes of chronic disease. Medical Care 27 (3), S110S127. Kapp, M.B., Mossman, D., 1996. Measuring decisional capacity: cautions on the construction of a `Capacimeter'. Psychology, Public Policy and Law 2 (1), 7395.

278

E.H. Bradley et al. / Social Science & Medicine 49 (1999) 267278 Putnam, S., Stiles, W., Jacob, M., James, S., 1985. Patient exposition and physician explanation in initial medical interviews and outcomes of clinic visits. Medical Care 23, 7483. Rockwood, K., Stolee, P., Fox, R., 1993. Use of goal attainment scaling in measuring clinically important change in the frail elderly. Journal of Clinical Epidemiology 46, 11131118. Rockwood, K., Stolee, P., Howard, K., Mallery, L., 1996. Use of goal attainment scaling to measure treatment eects in an anti-dementia drug trial. Neuroepidemiology 15, 330338. Ryden, M.B., Feldt, F.S., 1992. Goal-directed care: caring for aggressive nursing home residents with dementia. Journal of Gerontological Nursing 18 (1), 3541. Sachs, G.A., 1997. Dementia and the goals of care. Journal of American Geriatrics Society 46, 782783. Scott, R.A., Aiken, L.H., Mechanic, D., Moravcsik, J., 1995. Organizational aspects of caring. Milbank Quarterly 73 (l), 7795. Squier, R.W., 1990. A model empathic understanding and adherence to treatment regimens in practitionerpatient relationships. Social Science & Medicine 30, 325. Stareld, B., Wray, C., Kess, K., Gross, R., Birk, P.S., D'Lugo, B.C., 1981. The inuence of patientpractitioner agreement on outcome of care. American Journal of Public Health 71 (2), 127131. Stewart, M., 1984. What is a successful doctorpatient interview? A study of interactions and outcomes. Social Science & Medicine 19, 167175. Stewart, M.A., 1995. Eective physicianpatient communication and health outcomes: a review. Canadian Medical Association Journal 152 (9), 14231433. Tjosvold, D., 1988. Cooperative and competitive dynamics within and between organizational units. Human Relations 41, 425436. Tjosvold, D., Andrews, I., Struthers, J., 1992. Leadership inuence: goal interdependence and power. The Journal of Social Psychology 132, 3950. Ugarriza, D.N., Gray, T., 1993. Alzheimer's disease: nursing interventions for clients and caretakers. Journal of Psychosocial Nursing 31 (10), 710. Von Kor, M., Gruman, J., Schaefer, J., et al., 1997. Collaborative management of chronic illness. Annals of Internal Medicine 127, 10971102.

Kirusek, T.J., Smith, A., Cardillo, J.E., 1994. Goal Attainment Scaling: Applications, Theory, and Measurement. Lawrence Erlbaum Associates, Hillsdale, NJ. Lo, B., 1990. Assessing decision making capacity. Journal of Law, Medicine & Health Care 18 (3), 193200. Locke, E.A., Latham, G.P., 1990. The Theory of Goal-Setting and Task Performance. Prentice Hall, Englewood Clis, NJ. Locke, E.A., Shaw, K., Saari, L., Latham, G.P., 1981. Goalsetting and task performance: 19691980. Psychological Bulletin 90, 125152. Locke, E., Chah, D., Harrison, S., Lustgarten, N., 1989. Separating the eects of goal specicity from goal level. Organizational Behavior and Human Decision Processes 43, 270287. Malec, J.F., Smigielski, J.S., DePompolo, R.W., 1991. Goal attainment scaling and outcome measurement in postacute brain injury rehabilitation. Archives of Physical Medicine & Rehabilitation 72 (2), 138143. McWilliam, C.L., Brown, J.B., Carmichael, J.L., Lehman, J.M., 1994. A new perspective on threatened autonomy in elderly persons: the disempowering process. Social Science and Medicine 38 (2), 327338. McWilliam, C.L., Stewart, M.A., Brown, J.B., et al., 1997. Creating empowering meaning: an interactive process of promoting health with chronically ill older Canadians. Health Promotion International 12 (2), 111123. Mold, J.W., Blake, G.H., Becker, L.A., 1991. Goal-oriented medical care. Family Medicine 23, 4651. Ong, L.M.L., DeHaes, J.C.L.M., Hoos, A.M., Lammes, F.B., 1995. Doctorpatient communication: a review of the literature. Social Science & Medicine 40 (7), 903918. Palisano, R.J., 1993. Validity of goal attainment scaling in infants with motor delays. Physical Therapy 73, 651660. Palisano, R.J., Haley, S.M., Brown, D.A., 1992. Goal attainment scaling as a measure of change in infants with motor delays. Physical Therapy 72, 432437. Pearson, C., 1987. Participative goal setting as a strategy for improving performance and job satisfaction: a longitudinal evaluation with railway track maintenance gangs. Human Relations 40, 473488. Prochaska, J.O., DiClemente, C.C., Norcross, J.C., 1992. In search of how people change: applications to addictive behaviors. American Psychologist 47 (9), 11021114.