A Longitudinal Evaluation of Patients' Perceptions of Parkinson's Disease

Abstract

Purpose: Parkinson's disease (PD) is a chronic progressive neurological disorder that frequently results in nearly total disability. This study examined changes over 3 years in patients' experiences living with PD, and explored how participants' health perceptions and predicted mortality at baseline related to their actual death by Year 3. Design and Methods: Data were obtained from a local sample of 109 participants by in-home interviews. Results: The collected results (1) suggested a disconnection between the participants' responses to open-ended questions about the disorder, compared with their responses to questions structured by the investigators; (2) demonstrated changes in responses to structured and open-ended questions; and (3) demonstrated that respondents who declined to predict whether they would be living in 10 years were three times more likely to die by Year 3 than those who answered the question. Implications: Results demonstrate the importance of identifying the most important issues for the individual with PD and suggest that these issues may change over time. Results also raise issues surrounding how patients' perceptions influence the course of their disease.

Decision Editor: Laurence G. Branch, PhD

Parkinson's disease (PD) is a chronic progressive neurological disorder with potentially devastating consequences for the patient. The patient's symptoms are likely to increase over time, with progressive loss of independence and increasing financial and social burden (Schenkman, Zhu, Cutson, and Whetten-Goldstein 2001). Overall, patients in later stages of the disease report a lower quality of life than do those early in the disease (Koplas et al. 1999). Despite an evolving understanding of the consequences of PD, patients' perceptions of the illness are only beginning to be appreciated.

Recent advances over the past two decades have substantially clarified the broader implications of chronic disorders for the patient. For example, disablement models, developed from the perspectives of medical sociology (Nagi 1965) and medicine (World Health Organization 1980) provide frameworks for analyzing interrelationships of pathology, impairments, their functional consequences (e.g., functional limitations, disability), and overall quality of life. These models serve as a stimulus for development of new measures (e.g., de Boer, Sijker, Speelman, and de Haes 1996; Peto, Jenkinson, Fitzpatrick, and Greenhall 1995), as well as other investigations (e.g., Schenkman et al. 2001; Whetten-Goldstein, Sloan, Kulus, Cutson, and Schenkman 1997) that better reflect the broad impact of health conditions. Despite these strides, remarkably little is known about the experience of living with chronic health conditions in general, and PD in particular, as expressed by the patient. A few investigators have examined patients' perceived difficulty with function in relation to the amount of assistance that they need to perform tasks (Abudi, Bar-Tal, and Ziv 1997; van Bennekom, Jelles, Lankorst, and Kuik 1996). Data from these investigations demonstrate differences between patients' functional status and their perception of problems with function. A few investigators have examined patients' perceptions by asking open-ended questions regarding the impact of specific disorders (e.g., stroke; Mumma 2000). Such investigations illustrate the importance of understanding the impact of disorders from the perception of the patient.

Two aspects of patients' perceptions are of particular concern for those living with PD. The first is the patient's experience of living with the disease. The second is the broader issue of the patient's perceived health and predicted mortality in relation to actual death. Exploration of these issues can provide valuable information to caregivers who must determine the most effective strategies of care for their patients. Such investigations are critical to researchers because they clarify further the processes and impact of disablement.

In the few available studies of individuals' perception of PD, investigators typically have developed survey instruments that structure the possible responses (de Boer et al. 1996; Peto et al. 1995). In one study (Abudi et al. 1997), the investigators also asked participants to rate both perceived severity and the presence of specific symptoms. Results demonstrated agreement between experts and patients for some symptoms; but, in some instances, the symptoms were less disturbing to the patients than the experts would have predicted. To date, no investigators have reported on patients' perception of PD as evidenced by response to open-ended questions.

Furthermore, information obtained regarding the impact of PD is usually cross-sectional in nature. Longitudinal data are particularly pertinent to PD because of the chronic, progressive nature of the disease; such data may elucidate important differences in patients' perception across time, lead to a deeper understanding of the causes and effects of relationships among perception and disease, and provide insight that may be useful in directing intervention strategies as the disease progresses.

A second, more global aspect of patients' perceptions is the relationship among the individual's self-reported health, predicted mortality, and actual mortality. The individual's perception of health status, as is increasingly apparent, is a powerful predictor of mortality. In examining this relationship, Idler and Benyamini 1997 recently reviewed over two dozen studies that demonstrated that the single question, "Would you rate your health as very good, good, fair, or poor?," is as powerful a predictor of mortality as are more specific questions regarding health status, health utilization, and medical history.

Most of the studies examining relationships between self-rated health and mortality have been carried out with community-dwelling samples that did not contain many chronically ill persons (Idler and Benyamini 1997). Studies are needed to examine this relationship in individuals with chronic progressive disorders such as PD. For people with PD, in which the ultimate outcome is likely to be severe and unremitting disability, there may be a very different relationship between perceived health and mortality than that observed in community-dwelling samples without specific disorders.

An appreciation of the patient's perception of disease and functional limitations broadens the perspective of health care providers regarding the impact of PD, and provides a basis for future conceptual work related to the impact of the disablement process on individuals with chronic health conditions. Therefore, the aim of this study was to examine perceptions of individuals with PD from the following three perspectives: (1) relationships between individuals' responses to standardized and open-ended questions regarding impairments and functional ability; (2) changes in perceptions of impairments, functional limitations, and social/emotional aspects of the disease over a 3-year period; and (3) the impact of participants' self-rated health and their predictions of their future life expectancy with regard to actual mortality by Year 3.

Methods

Data for this investigation were part of a larger database that was designed to analyze the financial impact of PD (Schenkman et al. 2001; Whetten-Goldstein et al. 1997). Data were collected during in-home interviews utilizing surveys of health status, disease symptoms, functional status, and costs of the disease (Whetten-Goldstein et al. 1997). The survey instrument was approved by the Institutional Research Board of the Duke University Medical Center; all participants gave informed consent. Interviews were conducted by trained research assistants. The survey lasted approximately 90 minutes.

Sample

Participants were recruited from several sources, including participants from previous studies of PD at the Duke University Center on Aging, members of local PD support groups, and individuals who were patients of one of the neurologists at Duke University Medical Center. The sample consisted of PD participants living within 30 miles of Duke University in Central North Carolina. There were 108 participants entered into the original study in 1995. An additional 8 entered in 1996, and 44 entered in 1997. These 160 patients were followed for 3 years, until 1999. In this article, Year 1 denotes the first year of patient participation, regardless of calendar year. By Year 3, 10 participants declined to continue in the study and 41 participants were lost to follow-up (e.g., moved, phone disconnected). The analysis sample in this study consisted of the remaining 109 participants, of whom 91 were living and 18 had died by Year 3.

Instruments

The survey questions, described previously (Whetten-Goldstein et al. 1997), were based on the National Health Interview Survey (National Center for Health Statistics 1994) and the Assets and Health Dynamics (AHEAD) of the oldest old survey (Myers 1997). Structured questions included the following: (1) demographics (including wealth and income); (2) health status (including comorbid illnesses, self-rated health, estimated probability that they would be alive in 10 years; (3) PD specific impairments (e.g., tremor, rigidity); (4) general impairments (e.g., fatigue, pain); (5) basic activities of daily living and instrumental activities of daily living; and (6) cognitive/emotional status. In these surveys, depression was assessed by a single question, "In the past year, have you had two weeks or more during which you felt sad, blue, or depressed, or lost pleasure in things that you usually cared about or enjoyed." Immediate recall was assessed by reading participants a list of 10 words and asking for recall. Responses to these structured questions were differentiated in the following subsets: cardinal signs of PD, secondary symptoms, and functional limitations.

Participants were asked to predict the likelihood that they would be alive in 10 years (Myers 1997). This question is part of the AHEAD survey and provides a means of identifying whether individuals expect to live a long or short time. Data are provided on a percentage basis. For purposes of analyses, data were collapsed into two categories: probability <50% and probability ≥50%.

After completing the structured survey questions, participants were asked to respond to the following: "We are interested in knowing more about PD than can be captured through list of signs and symptoms. Please tell me what you consider to be the worst aspect of PD."

Participants' responses to this request were recorded by the interviewer (written form) using short sentences or phrases. To analyze the participant's responses to this open-ended question, the following strategy was used. First, a research assistant coded the answers (e.g., driving, fatigue, tremor, stiffness). He identified the maximum number of distinct responses (i.e., unless respondents used the same word, the information was considered a distinct response). Three investigators (M.S., T.C., C.Z.) then abstracted themes (e.g., if one participant used the word ‘tired’ but another used the word fatigue, data were collapsed into a single category called fatigue). Then reponses were grouped into the following consensus-generated categories: "impairments," including 11 items (e.g., rigidity/stiffness, tremor, memory/dementia, pain, fatigue); "functional limitations," including 11 items (e.g., loss of independence, change in lifestyle, mobility/walking, falls, speech, coordination, sleeping, driving, dressing, and eating/swallowing); and "social/emotional" problems, including four items (fear of future/disease progression, emotional problems, stigma/embarrassment, and economic burden worry).

The percentage of participants was first determined who identified at least one item among the possible 26 items. This variable was called "any problems." Then, for each category of responses, the percentage of participants identifying at least one item within that category was determined, called "any impairments," "any functional limitations," and "any social/emotional problems." We also determined, for each category of responses, the number of items identified by each respondent (e.g., 1, 2, 3, …). Finally, the percentage of participants identifying each of the 26 items also was determined. Only those items were further analyzed if at least 5% of the participants identified them as important.

Interviewers

All interviewers were trained at the Center for Health Policy Law and Management where interviews are routinely conducted in the field. As part of the training, one of the investigators (K.W.G.) observed three interviews. These observation sessions were used to ensure consistency in how the questions were read and to ensure that the interviewer did not use verbal or body language that might bias the participant's response. With regard to open-ended questions, interviewers were instructed to write the entire responses and allow participants to say as much as they wanted.

Data Analysis

Descriptive statistics at baseline were determined for those participants who survived and those who died by Year 3. Comparisons between these two groups were made using a chi-square test for dichotomous variables and Student's t test for continuous variables. Probit regressions (StataCorp 1999) were used to examine the predictors of death by Year 3 of the study. For this analysis, we first determined the bivariate associations between potential independent variable and death by Year 3. Death by Year 3 correlated significantly with the following variables: age, years of school, refusing to answer the question on probability of living in 10 years, and years since symptoms started. These variables were included in the probit regression.

Results

The Sample

The sample was comprised of 78 men, age 59–88 (mean = 73), and 30 women, age 50–86 (mean = 70). Demographic and health characteristics of the sample are shown in Table 1 . Those who survived until Year 3 on average were significantly younger (70.9 vs. 77.6, p < .0001), less likely to be widowed (12.5% vs. 35.3%, p = .02), had better immediate recall (4.5 vs. 2.9, p = .0085), and were more educated (15.4 vs. 13.7, p = .03). Additionally, differences approached significance for duration of the disease and number of comorbid conditions.

Baseline responses of the sample to the structured and open-ended questions are shown in Table 2 . In response to structured questions about the disease (Panel 1), a majority of the participants acknowledged rigidity and bradykinesia, and more than 40% acknowledged tremor. Between 34% and 70% of the respondents also acknowledged secondary symptoms of PD (e.g., excessive saliva/dry mouth, fatigue, pain) and functional limitations (speech, dressing, eating/swallowing). In multivariate analysis, differences between the two groups were not significant. Chi-squared test for the variable bladder/bowel disturbances was significantly different between those who survived and those who died by Year 3 (p = .05).

With regard to responses to the open-ended questions (Panel 2), more than three-fourths of the participants identified some symptoms (impairment, functional limitation, social/emotional) as "among the worst aspects of PD" (referred to as "any problems" in the table). No difference was observed between those who survived and those who died by Year 3. When we examined impairments, functional limitations, and social/emotional concerns separately, it appeared that functional limitations were of the most concern to the participants.

With regard to impairments, only one third or fewer of the respondents identified at least one impairment (any impairment) as a "worst aspect of PD." Indeed, for most of the responses, fewer than 10% of the respondents identified any individual impairments as "worst aspects of PD." The exceptions were tremor (14.3% of the survivors) and excess saliva/dry mouth (11.8% of those who died by Year 3). Excess saliva/dry mouth was significantly more problematic for those who died (p = .001).

In contrast, participants (48.4% of those who survived and 70.6% of those who died) more frequently identified functional limitations among the "worst aspects of PD." Mobility/walking, change in lifestyle/curtailing activities, and speech were most frequently cited. With regard to social and emotional problems, fear of future disease progression was identified by 13.2% of those who survived. Differences between those who survived 3 years and those who did not were not significant for functional limitations or social/emotional concerns.

At baseline, participants were asked to predict the likelihood that they would be alive in 10 years (Panel 3). A substantial percentage (20.9%) of those who lived to Year 3 declined to answer this question and significantly more (47.1%) of those who died declined to answer this question (p = .02). Fully half of those who survived to Year 3 rated their probability of living at >50%, whereas nearly one third of those who had died by Year 3 also rated their probability of living 10 years at >50%. Differences were not statistically significant. The mean scores among those who answered this question were 64.1 and 57.4 for those who survived and those who died by Year 3 respectively.

Changes in Perception Over Time

For those participants who survived to Year 3, we next examined how the participants' responses changed over time (Table 3 ). Responses to the structured questions (Panel 1) were significantly different for two of the cardinal signs—rigidity/stiffness and bradykinesia/slowness of movement (p ≤ .0001); two of the secondary symptoms (excessive saliva/dry mouth and fatigue/weakness, p ≤ .03); and all of the functional limitations (p ≤ .003). For each variable tested, fewer of the respondents acknowledged the problem at Year 3 than at Year 1.

With regard to the open-ended questions (Panel 2), a similar pattern emerged. Overall, fewer participants identified "any problems" by Year 3 (84.6% vs. 45.1%, p < .0001). When we examined specific symptoms, only tremor (p = .021), mobility/walking (p = .0001), and fear of disease progression (p = .0362) were significantly different at Year 3.

To understand these data further, we also examined the total number of problems reported by participants within each category (data not tabulated). Considering only those participants who identified specific areas as burdensome, most respondents identified only one or two concerns for each category. The mean (SD) number of problems identified at baseline (any category) was 1.20 (.78) for those who survived and 1.06 (.75) for those who died by Year 3. The mean (SD) number of problems identified in Year 3 by the survivors was .79 (.71). Thus, participants in general appeared to focus on a problem in only one of the three possible categories (impairment, functional limitation, or social/emotional).

When considering problems identified within the three categories, the respondents were more likely to identify a functional limitation as problematic. At baseline, the mean number of functional limitations identified was 0.57 (0.65) for those who survived and 0.82 (0.64) for those who died. At Year 3, the mean number of functional limitations identified was 0.41 (0.63). In contrast, the mean number of impairments ranged from .18 to .41 and social/emotional concerns ranged from .10 to .22 for any group and any time point.

Self-prophesy of Mortality

Finally, we examined the relationship between self-prophesy of mortality in 10 years and actual death by Year 3 (Table 4 ). For these analyses, we used a probit regression analysis. The independent variables included in the model (those with significant bivariate correlations to actual death) included age, education, duration of PD symptoms, and a dummy variable indicating the respondent declined to answer the question on probability of living in 10 years. By far, the most significant predictor of actual death by Year 3 was respondents' refusal to answer the question on self-prophesy of mortality in 10 years.

Respondents who declined to answer the question whether they believed they would be living in 10 years were three times more likely to die by Year 3 than those who answered the question. Respondents who were older were more likely to have died by Year 3 than younger respondents. The odds of dying by Year 3 were 16% higher as age increased by 1 year (p = .007). A 1-year increase duration since symptoms started increased the odds of dying by Year 3 by 14% (p = .05). Although only marginally significant (p = .08), years of schooling decreased the probability of dying by Year 3 by 20%.

We examined the bivariate correlation of self-reported fair or poor health status and death by Year 3, but it was not statistically significant (r = .13, p = .20) and therefore was excluded from the analysis. Furthermore, the probability of living in 10 years estimated by the respondents was not significantly correlated to respondents' actual death in Year 3 (r = −.15, p = .11) and was excluded from the probit analysis.

Discussion

Results of this investigation provide several unique insights into the patient who lives with PD. These insights relate to the following: (1) Participants' perception of the disease, (2) changes in perception over time, and (3) the relationship between the participants' predictions of mortality at 10 years and their likelihood of surviving for 3 years.

Participants' Perceptions of PD

Participants acknowledged both impairments and functional limitations in response to structured questions. However, they were much more likely to identify functional limitations as particularly burdensome while rarely identifying impairments. This finding is consistent with the shifting focus of disablement to aspects of disease and injury beyond impairments and functional limitations.

With regard to impairments, when responding to structured questions, a majority (64.7–81.3%) of the participants acknowledged rigidity and bradykinesia, and more than 40% reported tremor. This finding is consistent with the impairments that would be expected for individuals with PD. With the exception of pain, there were no differences between those who survived and those who had died by Year 3, most likely because of lack of sensitivity of the measures used.

In contrast, when responding to the open-ended question, participants' answers contrasted markedly; very few of the respondents noted that those same impairments were among the "worst aspects" of the disease. Indeed, with the exceptions of tremor (Year 3), fewer than 10% of the participants identified any impairment in response to the open-ended question (Table 1 ). Only excess saliva/dry mouth differed significantly (p = .001) between Years 1 and 3. In part, this discrepancy between responses to structured questions and open-ended questions might be explained by the fact that participants appeared to focus on a single issue (e.g., impairment, functional limitation, social/emotional) as most burdensome, so that the overall percentage reporting any particular impairment was low. In addition, however, the overall percentage of participants identifying any impairment as burdensome (34.1% of those who survived and 18.7% of those who died) is lower than the percentage of participants (44.0–76.9%) acknowledging specific impairments in response to structured interview questions (Table 3 ). This finding is consistent with observations regarding patients with PD (Abudi et al. 1997) and with stroke (van Bennekom et al. 1996), supporting the observation that patients may be more concerned with different issues than are "experts."

Possibly, participants adjust to impairments to some extent, learning to cope with them so that the impairments have relatively less impact than health care professionals might predict. In this context, Abudi and colleagues 1997 observed that health care professionals' judgments of participants' difficulties may be affected by their own values concerning health. Thus, the health care professional may be more focused on specific impairments than perceived by the patient or may expect a higher level of disability than is experienced by the patient. These results point out the importance of health care professionals carefully observing their own prejudices and preconceived notions regarding the patient's experience living with a chronic progressive disorder such as PD.

With regard to functional limitations, in response to the structured questions, most participants reported difficulty with functional activities that typically are problematic for those with PD. In response to the open-ended question, a larger percentage of participants identified functional limitations as burdensome, emphasizing that functional ability is a more pressing concern to the patient than are the specific impairments around which health care teams tend to focus.

These findings have implications for members of the health care team. Health care professionals need to be aware of those impairments, functional limitations, and social/emotional concerns that are most pressing to the individual patient. Although the professional may be successful in controlling or improving specific underlying impairments (e.g., tremor, rigidity), he or she also needs to know whether the impairments that improve are of concern to the patient. Indeed, patients often identified difficulties with speech and falling as burdensome, yet these functional limitations are not easily treated pharmacologically. Hence, good pharmacological control of impairments could be perceived as insufficient from the point of view of the patient.

Predictions of the probability of surviving 10 years were surprising. Nearly one third of the respondents who had died by Year 3 had—at baseline—estimated the probability of living 10 years at greater than 50%. This suggests that a substantial number of individuals either were unaware of or unable to accept their actual condition. In interpreting these data, it is important to note that the sample size for those who died was only 17 individuals. This finding should be examined in a larger data set for confirmation.

Also surprising were the number of participants who declined to answer this question. It is noteworthy that this subgroup of participants had a higher percentage of women than did those who answered the question. One possible explanation, noted anecdotally by the interviewers, was that male spouses, when present during the interview, appeared to encourage the wife to refuse to answer this question, perhaps in an attempt to "protect" their wives with PD. Additionally, these participants tended to become teary-eyed and appeared distressed by the question. Together, these anecdotal observations suggest that the participants who declined to predict mortality may have difficulty facing that reality. Future investigations should explore reasons for declining to consider end of life. Furthermore, future studies should determine whether it is beneficial or detrimental to encourage particular patients to acknowledge the likelihood of death and then develop coping mechanisms for the physical, social, and emotional impact of PD.

Changes Across Time

Our second purpose was to examine changes in participants' perceptions over time. Across all of the variables examined, fewer participants acknowledged problems at Year 3, suggesting that the participants adjusted to PD so that they no longer were as aware of specific impairments, functional limitations, and social/emotional burden of the disease. This finding held for responses to both the open-ended and structured questions, and is consistent with Starkstein's cross-sectional data indicating that depression is highest in early and late stages of PD and is lower in middle stages of the disease (Starkstein et al. 1998). Taken together, results of the present investigation and of Starkstein and colleagues suggest that people with PD may change or modify their roles and expectancy so that the impact of the disease is less burdensome during the middle years of the disease. These findings are consistent with the observation that one's perception or response to disability is relative and may shift (referred to as recalibration) based on experience and expectations (Daltroy, Larson, Eaton, Phillips, and Liang 1999).

Perception in Relation to Mortality by Year 3

In recent years, attention has turned toward patients' attitudes about disease and the impact of those attitudes on outcome. For example, individuals with breast cancer who have a fighting spirit or denial/avoidance are more likely to remain alive than those who have a fatalistic approach or helplessness/hopelessness (Greer, Morris, and Pettingale 1979; Greer, Morris, Pettingale, and Haybittle 1990). Similar findings have been reported for individuals who have sustained a stroke (Lewis, Dennis, O'Rourke, and Sharpe 2001).

One of the most powerful predictors of mortality is the individual's perception of health status (Idler and Benyamini 1997). In the present investigation, we examined both self-rated health and the participants' estimated probability of survival to their actual survival at 3 years. In contrast with other investigators (Idler and Benyamini 1997), self-rated health did not correlate significantly with death by Year 3 of the study. Several possible reasons may account for this discrepancy. First, the participants in this sample all had PD—a chronic disorder that is likely to end in significant disability before death. Thus, the sample differed substantially from many of the previous studies of community-dwelling adults without specific disorders (Idler and Benyamini 1997). Of the investigations that have examined people with specific disorders, most have examined patients with conditions that are not progressive [e.g., hip fracture, stroke (Dasbach, Lein, Lein, and Moss 1994; Wolinsky and Fitzgerald 1994)]. These studies did not examine mortality as an outcome, but rather focused on recovery from the event. The only study of patients with a progressive condition focused on diabetes (Menec, Chipperfield, and Perry 1999), a chronic condition in which disability typically is not as totally encompassing. In that study, higher mortality rates were associated with poor perceived health. Possibly the difference between those with diabetes and those with PD relates to the specific difference between patients' perceptions of these two conditions. Further examination of the relationship between self-rated health and mortality is necessary—as suggested by Idler and Benyamini 1997—to determine the conditions under which this relationship holds with patients who have specific underlying conditions.

In contrast to self-rated health, response to the question regarding probability of living 10 years was predictive of death by Year 3. Those who declined to estimate a probability of living 10 years were nearly three times as likely to die by Year 3 as were those who answered the question. (This was true whether or not those who answered predicted a low or high probability of living 10 years and indeed estimates of probability of living, whether high or low, did not correlate with actual death by Year 3.) This finding is consistent with our impression that those who declined to predict probability of living 10 years were least able to cope with their disease. What remains unclear in this sample (and should be explored further) is whether declining to contemplate dying drives death or whether impending death causes individuals to be unwilling to contemplate their mortality.

A few investigators have examined interactions between self-efficacy beliefs and declines in functional ability (Seeman, Unger, McAvay, and Mendes de Leon 1999). A future, comprehensive investigation of relationships among self-efficacy, self-prophesy of mortality, perception of health, and functional decline and death will help to elucidate further the role of perception in decline and death of those with PD. From a clinical perspective, the patient's perceptions of his or her mortality should be considered in the total scheme of disease management to best meet the needs of the patient.

Limitations

This work had several limitations. First, this was a preliminary study with a local sample of individuals who volunteered to participate. Some of the participants had taken part in previous studies at the Center for Aging, Duke University. Some were members of local PD support groups. Others volunteered in response to advertisements in local newspapers or were contacted through their neurologist. Thus, the sample was one of convenience. Furthermore, a disproportionate percentage of the sample (77% of those who survived and 64.7% of those who died) was men. This percentage is consistent with many other studies related to PD in which more men participate than women. Most epidemiological studies have shown that PD is more common among men than women, although some authors report equal prevalence (Ben-Shlomo 1996; Tanner and Goldman 1996).

Second, the sample was relatively small. Hence, we were limited with respect to number and extent of analyses. With a larger sample, additional relationships may emerge. In addition, because the sample was small and death rate was high, it was not possible to follow an adequate number of participants to carry out analyses beyond 3 years. Future studies should follow individuals longer (5 years or more) and should examine additional issues (e.g., relationship of ethnicity, depression, cognition, and gender to perception).

Third, because this was a preliminary study with in-home interviews carried out by research assistants, impairments of PD (e.g., rigidity, tremor, bradykinesia) were identified only by patient self-report rather than by direct neurologic examination. Thus, we were unable to quantify disease state [e.g., Hoehn and Yahr stage of disease (Hoehn and Yahr 1967), unified PD rating scales (Fahn, Elton, & Members of the UPDRS Development Committee, 1987)]. To understand fully the patient's perceptions of this disease, it will be important to examine the impact of stage of disease on participants' responses to the question, "What are the worst aspects of PD for you?" Additionally, the questions regarding cognition and depression (from the AHEAD survey and the National Health Interview Survey) are limited in scope. More detailed information regarding depression and cognition should be obtained in future studies using instruments specifically designed to examine these issues (e.g., the Center for Epidemiologic Studies-Depression Scale and the Short Portable Mental Status Questionnaire [SPMSQ]; Pfeiffer 1975; Radloff 1977).

Finally, this study was a part of a larger investigation into the economic impact of PD (Schenkman et al. 2001; Whetten-Goldstein et al. 1997). Participants answered the open-ended question at the end of a 90-minute interview. Previous interview questions could have influenced their responses. For example, many questions in the larger study probed the cost of PD. Thus, participants may already have felt that they answered the question of financial implications of the disease and therefore may have neglected to bring this issue up in response to the open-ended question. Alternatively, the cost of PD may not be one of the biggest concerns for these respondents. Also, participants may have been fatigued by the end of the questionnaire, limiting their responses. In particular, fatigue may have influenced the respondents who were in the poorest health and closest to end of life. Further investigation is needed to resolve these issues.

Despite the limitations associated with this preliminary investigation, these longitudinal data provide unique insights into the patient with PD. Findings from this investigation emphasize the possible discrepancy between clinicians' and patients' perception of the burden of PD and the patients' changing perceptions as the disease progresses. In addition, results of this investigation raise several issues surrounding how patients' perceptions influence the course of their disease. Future investigations—including a larger sample, neurological examination, and carried out for five or more years—will provide additional insight into patients' responses to living with this chronic progressive condition, answer some of the questions that could not be resolved in the present investigation, and assist the clinicians to work most effectively with patients who have PD. Results of this and related investigations can provide insight into further development of theoretical constructs regarding the role of patient's perception in his or her experience of chronic illness.

This work was supported by funds from the Duke University Medical Center Small Grants Program (MS), the National Institutes of Health, National Institute on Aging, Claude D. Pepper Older Americans Independence Center [Grant 5 P60 11268 (to M.S., T.C., and K.W.G.)], and the National Institutes of Health [Grant 5 T32 AG00029 (to C.W.Z.)]. We thank the interviewers who dedicated themselves over five summers to this project: Jen Liu, Jennifer Zartarian, Dan Dillon, Brenda Erhart, Kattran Rhodes, Christopher Keir, Bernadette Hall, Lauren Hollingsworth, and Jennifer Flythe. We thank Dr. Burton Scott for referral of participants to the study.

References