Data collection for this work was completed while Dr. Rao was a faculty member of Indiana University School of Medicine and prior to her employment with the Centers for Disease Control and Prevention. The views expressed in this paper are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.
Predicting reports of unmet expectations among rheumatology patients†
Article first published online: 5 APR 2004
Copyright © 2004 by the American College of Rheumatology
Arthritis Care & Research
Volume 51, Issue 2, pages 215–221, 15 April 2004
How to Cite
Rao, J. K., Weinberger, M., Anderson, L. A. and Kroenke, K. (2004), Predicting reports of unmet expectations among rheumatology patients. Arthritis & Rheumatism, 51: 215–221. doi: 10.1002/art.20246
- Issue published online: 5 APR 2004
- Article first published online: 5 APR 2004
- Manuscript Accepted: 26 JUN 2003
- Manuscript Received: 22 FEB 2003
- Picker-Commonwealth Faculty Scholars Program
- Department of Veterans Affairs Research Career Development Program
- Department of Veterans Affairs Research Career Development Program
- Unmet expectations;
- Longitudinal study;
- Health status
Approximately 25% of patients report unmet expectations after their doctor visits. In a longitudinal study of rheumatology patients, we examined whether changes in health status could predict unmet expectations.
Arthritis patients (n = 177) responded to 2 surveys (baseline and 6-month followup). Both surveys contained questions on health status (functional status, pain, helplessness, psychological status) and visit duration. The followup survey contained questions on postvisit unmet expectations. Factors associated with unmet expectations were determined.
Fifty-eight patients (33%) reported unmet expectations, most often for information (47%) and new medications (31%). Unmet expectations were more common among patients with greater baseline helplessness (odds ratio [OR] 1.9, 95% confidence interval [95% CI] 1.0–3.6) and short doctor visits at followup (OR 5.6, 95% CI 2.4–13.1). Unmet expectations were less common among those experiencing a decline in pain (OR 0.3, 95% CI 0.1–0.9).
Attention to the patient's level of pain and helplessness and duration of the visit may limit reports of unmet expectations.
Changes to the structure and financing of health care and the current emphasis on quality (1) have prompted health systems and physicians to adopt patient-centered approaches to the delivery of care. Patient-centered care is intended to enhance medical decision making and improve health care experiences (2). One important aspect of patient-centered care is recognizing that patients often have specific expectations when visiting their physicians (3, 4). Patients may want the physician to address specific illness-related concerns or to order diagnostic tests, specialty referrals, or medication changes (3, 5–8). The term “expectations” is frequently used to describe desires or beliefs about what will happen, regardless of whether patients explicitly verbalize these perceptions to the physician during the visit (9–12).
Most patients come to the encounter with expectations and ∼15–25% leave with unmet expectations (3, 13). Those who report unmet expectations are less likely to be satisfied with their care (8, 13–18). Patients who are not satisfied with their care are less likely to adhere to medical treatment or experience symptom improvement (19–21); they also are more likely to switch physicians and sue for malpractice (22–27).
This study builds upon previous research about unmet expectations in 2 ways. First, prior investigations have generally focused on primary care patients (3, 6, 7, 13, 16, 18, 28–30), but we focus on patients with chronic rheumatologic conditions. These patients often develop relationships with rheumatologists for the ongoing management of chronic pain and other symptoms (31, 32). As cognitively oriented subspecialists, rheumatologists are experiencing pressures to deliver high-quality, efficient care to maintain their patient panels. Second, we use longitudinal data to examine the relationship between changes in patients' health status and unmet expectations. With few exceptions, previous investigations have been performed within the context of single clinic visits and none, to our knowledge, has examined whether health status changes are associated with unmet expectations. Because chronic rheumatologic conditions are characterized by waxing and waning symptoms, arthritis can provide an important framework for examining patients' expectations in relation to improvement or deterioration in health status.
This study examined whether demographic characteristics, clinic-related factors, or changes in self-reported health status between visits are associated with reports of unmet expectations among patients who receive continuous care from rheumatologists. We describe the types and percentages of unmet expectations reported by rheumatology patients, correlates of unmet expectations, and factors that are associated with unmet expectations.
Study sample and procedures.
This article presents the results of secondary analyses of data from a longitudinal study of patients with arthritis. The original study was designed to examine behaviors and outcomes related to patients' use of complementary therapies. Our recruitment strategy and study methods are described in detail elsewhere (33).
In brief, during a 2-week period in 1997, the patients were enrolled from 6 outpatient sites: 3 university and 3 private rheumatology practices. The university sites included a municipal hospital that provides care for medically and socioeconomically vulnerable patients, a Veterans Affairs medical center, and a fee-for-service specialty practice. In the original study, we were interested in obtaining information from adult patients who had established relationships with rheumatologists. Thus, patients were excluded from the study if they were new to the practice, were younger than 18 years, were demented, or resided in a nursing home. The Institutional Review Board of Indiana University-Purdue University at Indianapolis approved the data collection.
At study enrollment, medical records were reviewed to obtain demographic information and all rheumatologic diagnoses for each patient who consented to participate in the study. Patients underwent an initial telephone survey within 2 weeks following their rheumatology appointment, and subsequent surveys were administered at 6 and 12 months. The 12-month survey contained questions on unmet expectations. Because we wished to examine the potential relationship between changes in self-reported health status and unmet expectations, we limited our sample to participants who completed the 6- and 12-month surveys, which, for the purposes of this article, we shall refer to as the baseline and followup surveys, respectively.
The baseline and followup surveys included questions from 4 sources: the Modified Stanford Health Assessment Questionnaire (MHAQ; 8 items, each item scored from 1 [without any difficulty] to 4 [unable to do]) (34), a 10-point numerical pain scale (35), a psychological scale (4 items on sleep, anxiety, stress, and depression, each item scored from 1 [without any difficulty] to 4 [unable to do]) (36), and the Rheumatology Attitudes Index (a learned helplessness scale with 5 items, each item scored from 1 [strongly disagree] to 5 [strongly agree]) (37). The MHAQ has been shown to be sensitive to change (38, 39). Changes in the Arthritis Helplessness Index, a learned helplessness scale upon which the Rheumatology Attitudes Index is based, correlates with changes in functional status (40). Patients reported their perceptions regarding the amount of time spent with the rheumatologist at their last visit.
The followup survey also contained items on postvisit unmet expectations. Respondents were asked, “Sometimes patients have expectations of what the doctor should do at their visit. Remembering your last visit to the rheumatologist, do any of the following statements apply to you?” The respondents were then asked to chose from a list of 6 statements pertaining to unfulfilled desires, as follows: “I wished the doctor would: 1) change my medication, 2) refer me to a surgeon or other specialist, 3) give me information on my disease or treatment, 4) order laboratory tests, 5) spend more time examining me, 6) other (please specify)” (6, 13). The patients were asked to report all statements that applied to their last visit with the rheumatologist.
Our main study outcome was whether patients reported any unmet expectation, defined as a yes response to at least 1 of the 6 unmet expectation items on the survey. The independent predictors were grouped into 3 categories: demographic, clinic-related, and health-status variables. The demographic variables included age, race, sex, education level, and rheumatologic diagnosis. At study enrollment, rheumatologic diagnosis was defined according to a 4-level hierarchical variable (e.g., rheumatoid arthritis, fibromyalgia, osteoarthritis, or other rheumatologic conditions) based on the rheumatologist's diagnosis as documented in the chart (33). In this scheme, patients with rheumatoid arthritis, even if they had another diagnosis (i.e., fibromyalgia or osteoarthritis), were considered to have rheumatoid arthritis. Patients with a diagnosis of fibromyalgia, but not rheumatoid arthritis, were classified as having fibromyalgia. Finally, those with osteoarthritis, but not fibromyalgia or rheumatoid arthritis, were classified as having osteoarthritis. The clinic-related variables included site of care and perceived time spent with the doctor at the last visit. Site of care was defined as university or nonuniversity. Perceived time spent with the doctor was classified as a short visit (e.g., ≤10 minutes) or a longer visit (>10 minutes). In other studies, a visit lasting <10 minutes has been associated with lower patient satisfaction with care (41, 42). The baseline health status variables included pain level, functional status, psychological status, and helplessness. Finally, we calculated change scores for these health status measures between the baseline and followup surveys.
All analyses were performed using PC-SAS version 8 for Windows (SAS Institute, Cary, NC). Patients who responded to both the baseline and followup surveys were included in our analyses. We used chi-square or t-tests to compare the background characteristics of responders to both surveys with those of the participants who were lost to followup between surveys.
The analysis was conducted in 2 phases. First, we compared demographic, clinic-related, and health-status variables between patients who reported unmet expectations and those without unmet expectations. In these analyses, we used chi-square tests for categorical variables and t-tests for continuous variables. For the health status variables, we examined the relationships between unmet expectations and both the baseline values and the change scores for these variables. Because of the multiple comparisons, we considered as significant only those factors that showed an association with unmet expectations of P < 0.01.
Second, we were interested in determining whether a change in self-reported health status between baseline and followup could predict reports of unmet expectations. We fit multivariate logistic regression models that included unmet expectations as the dependent variable; the candidate independent predictor was the change score for the health status variable that was significantly associated with unmet expectations in the bivariate analyses. The model controlled for this health status variable at baseline and also adjusted for baseline functional status. Short doctor visit at followup, race, and education level were forced into the model as covariates because these variables have been associated with expectations (8, 12, 16) or satisfaction with care (42) in prior studies. Other candidate covariates included the predictors that were significantly associated with unmet expectations in the bivariate analyses. We calculated the odds ratio (OR) and 95% confidence interval (95% CI) for each independent predictor in the model.
Of the 203 respondents at baseline, 177 (87%) responded to the followup survey. Survey respondents who were lost to followup were similar to those who responded to both surveys with respect to age, sex, race, education, rheumatologic diagnosis, disease duration, site of care, pain, and functional status. The respondents' average age was 56.5 years, 74% were women, and 89% were white. Half (51%) were seen in a university clinic. The average duration of disease was 11.1 years; 39% had rheumatoid arthritis, 19% had fibromyalgia, 15% had osteoarthritis, and 27% had other rheumatologic conditions. Changes were found in functional status (P = 0.006) from baseline to followup, but not in psychological status (P = 0.02) or learned helplessness (P = 0.15).
Overall, 33% of respondents (58 of 177) reported at least 1 unmet expectation after their clinic visits; 119 reported no unmet expectations. Of the 58 respondents with unmet expectations, 57% had 1 unmet expectation, 31% had 2 unmet expectations, and 12% had 3 or more unmet expectations. Most common were unfulfilled expectations for information (n = 27, 47%), medication changes (n = 18, 31%), and physical examination (n = 17, 29%). Eleven (19%) and 13 (22%) of the respondents reported unfulfilled desires for referrals and laboratory testing, respectively, and 9 (16%) reported other unmet expectations not included on the structured list.
In bivariate analyses, age, sex, site of care, and rheumatologic diagnosis were not associated with unmet expectations (Table 1), but unmet expectations tended to be more common among nonwhite respondents (17% versus 8%; P = 0.05). Unmet expectations were significantly more likely among those reporting poorer functional and psychological status at baseline, higher levels of learned helplessness and pain at baseline, and short doctor visits at followup. Changes in functional status, psychological status, or helplessness scores from baseline to followup were not associated with reports of unmet expectations. However, there was a significant association between changes in pain score over time and unmet expectations (P = 0.002). Respondents reporting unmet expectations had an increase in pain score (0.98) whereas those who did not report unmet expectations experienced a slight decrease in pain score (−0.08) over time.
|Variable||Reported unmet expectations (n = 58)||Did not report unmet expectations (n = 119)||P*|
|Age, mean ± SD||55.9 ± 14.5||56.9 ± 13.4||0.7|
|College degree, %||19||31||0.09|
|Other rheumatologic disease||26||28|
|Seen at university clinic, %||48||52||0.6|
|Short visit with doctor at followup, %||49||24||0.001|
|Baseline scores, mean ± SD|
|Functional status (range 1–4)||1.8 ± 0.5||1.5 ± 0.5||0.003|
|Psychological status (range 1–4)||2.0 ± 0.6||1.8 ± 0.6||0.006|
|Helplessness (range 1–5)||3.5 ± 0.8||2.9 ± 0.9||<0.001|
|Pain (range 1–10)||5.5 ± 2.3||4.4 ± 2.4||0.008|
|Change scores,† mean ± SD|
|Functional status||0.08 ± 0.4||0.08 ± 0.4||0.9|
|Psychological status||0.09 ± 0.5||0.08 ± 0.5||0.9|
|Helplessness||0.09 ± 0.8||0.08 ± 0.7||0.9|
|Pain||0.98 ± 2.2||−0.08 ± 1.9||0.002|
Using multivariate models, we examined the relationships between changes in pain scores over time and reports of unmet expectations. In our initial model, we included the pain change score as a single continuous independent variable and adjusted for baseline pain, functional status, helplessness, race, educational level, and short doctor visits at followup. Baseline psychological status also showed a significant association with unmet expectations in the bivariate analyses. Because of concerns regarding multicollinearity, we did not include this variable in the model because of its correlation with helplessness (Pearson's correlation coefficient = 0.62, P < 0.0001) and functional status (Pearson's correlation coefficient = 0.57, P < 0.0001). In this initial model, a 1-point increase (or worsening) in pain score between baseline and followup corresponded to a 50% increase in the odds (OR 1.5, 95% CI 1.2–1.9) of unmet expectations (data not shown).
Our initial model did not allow us to compare the reports of unmet expectations based on whether the patient experienced an improvement or worsening of pain during the study. Therefore, we performed another multivariate analysis to examine the relationship between the change in pain measured as a categorical variable (i.e., improved pain [change score <0], worse pain [change score >0], or no change [change score equal to 0]) and unmet expectations. Compared with those who reported no change in their pain, respondents who experienced a decline in pain were significantly less likely to report unmet expectations (OR 0.3, 95% CI 0.1–0.9; Table 2). In addition, respondents who perceived that they had short doctor visits were significantly more likely to report unmet expectations (OR 5.6, 95% CI 2.4–13.1). Finally, higher baseline levels of learned helplessness were associated with unmet expectations, but this relationship had borderline significance (OR 1.9, 95% CI 1.02–3.6). When we substituted baseline psychological status for helplessness in the multivariate model, the model estimates did not change.
|Variable||Adjusted odds ratio*||95% confidence interval|
|Change in pain (baseline to followup)†|
|No change (referent)||1.0||-|
|Baseline functional status||0.9||0.4–2.5|
|Short doctor visits at followup||5.6||2.4–13.1|
In this longitudinal study, one-third of arthritis patients reported they had unmet expectations after visiting their rheumatologists. Consistent with other data (3, 43), the patients most often reported unfulfilled expectations for information. In bivariate comparisons, unmet expectations were more common among patients with poorer functional status and higher levels of pain and learned helplessness at baseline and among those who perceived their clinic visits as short in duration. Although a change in pain level between visits was a strong predictor of unmet expectations, we did not find an association with changes in functional status or in learned helplessness. In multivariate analyses, patients who experienced a decline in pain level over time were significantly less likely to report unmet expectations than those whose pain level did not change. Although patients experiencing an increase in their pain were more likely to report unmet expectations, the relationship was not significant.
The finding that changes in pain had a stronger association with reports of unmet expectations than changes in functional status probably relates to our population of patients with chronic arthritis. Because of the chronicity of their symptoms (mean duration 11.1 years), these patients may have had prior experiences with the health care system, health care providers, or treatments that influenced their expectations for specific physician actions (44). Many arthritis patients believe that despite good medical care, gradual declines in physical functioning are likely to occur as a result of their condition (44–47). Perhaps these patients considered changes in pain as an immediate issue that warranted action, either from themselves or from their physicians. For example, arthritis patients who have severe pain are more likely than those without severe pain to use complementary therapies to treat their symptoms (33, 47). Our results suggest that paying close attention to changes in patients' level of pain over time may be one strategy to reduce unmet expectations.
The association between learned helplessness and unmet expectations is notable in light of previous research. Learned helplessness theory describes a psychological state in which an individual believes that his or her efforts to control the outcome will be ineffective (37). Personal helplessness may be manifested as dysfunctional patient behaviors (e.g., noncompliance, passivity, and decreased problem solving) (40, 48–50), cognitive deficits (e.g., poor sense of self efficacy, perceived control) (51), or emotional difficulties (e.g., depression, anxiety, hostility) (51). For arthritis patients in particular, a high degree of helplessness is associated with higher levels of pain and functional impairment, greater use of passive pain-coping strategies, and nonadherence with treatment (40, 48, 49). Given the potential for poor health outcomes among patients with learned helplessness, physicians should be aware that cognitive–behavioral interventions may restore the patient's self efficacy and sense of control over their situation (50).
Patients' perception of visit length was another important predictor of unmet expectations. Time during the visit has become an important commodity for physicians and patients alike. Many physicians believe that the amount of time available to see patients has declined because of the current emphasis on visit efficiency (52–54). Patients, however, may have certain expectations regarding visit duration (42). For instance, patients who are new to the practice, those with health-related concerns, and those with lower self-rated health are likely to desire longer visits (42). Visit duration may have an influence on the interactions between physicians and patients. In the 1980s, Roland et al (41) and Morrell et al (55) randomly assigned primary care patients to clinic visits of 5-, 7.5-, or 10-minute duration. Audiotapes of the visits revealed that physicians were more likely to spend time identifying patients' problems and explaining proposed management strategies in the 10-minute visits than in those booked at shorter intervals (41, 55). In several studies, short visits have been associated with incomplete assessments of health risks and health problems, inappropriate prescribing practices, and lower rates of preventive health care delivery (54, 56–59). Patients who have shorter visits report lower scores on measures of satisfaction and the doctor-patient relationship (42, 55, 59).
Although visit duration may be important, the nature of the communication between patients and physicians (25, 60) may help decrease reports of unmet expectations. In a recent study of primary care visits, patients were allowed an average of 23 seconds to express their concerns before being interrupted by physicians; these interruptions occurred most often after the patients had stated their first concern (61). Furthermore, physicians did not inquire about the patients' expectations in 24.6% of visits (61). Failure to solicit the patients' expectations increases the possibility that they will raise their concerns late in the visit (62) or not at all, perhaps resulting in unmet expectations afterwards. Physicians may find communication techniques, such as facilitating remarks or open-ended statements, helpful in obtaining information on the patients' expectations. Orienting patients to the process of the visit may also help them understand when to disclose their expectations and how to develop priorities for the visit (25). Finally, use of methods to identify the patient's expectations immediately before the visit and presenting this information to the physician may reduce unmet expectations without causing a substantial increase in visit length (13).
Our study is subject to several limitations. First, all of our patients had established relationships with rheumatologists, and our results may not apply to patients who are seen by primary care providers for their arthritis. The findings may also differ among patients with a new diagnosis of arthritis, as they may report different expectations than our patients with chronic arthritis. Second, we measured unmet expectations at 12 months only and therefore were unable to assess whether the same patients were likely to have unmet expectations at other time points. It would have been ideal to measure patients' expectations after each clinic visit rather than at the prespecified intervals of our longitudinal survey. However, this study contributes to the literature on unmet expectations by describing the characteristics of patients who report unmet expectations after their visits to rheumatologists. Third, our findings are based solely on patients' postvisit expectations because we did not inquire about their expectations before the visit. Although a postvisit survey should be adequate to ascertain patients' perceptions regarding expectation fulfillment (8), we did not have feedback from physicians or access to objective evidence of what transpired during the visit. Thus, we were unable to examine whether the physicians actually addressed the patients' expectations or whether the visit length corresponded with the patients' perceptions. Finally, because of the small sample size, we were unable to stratify patients according to their level of pain or other clinical characteristics. Furthermore, it is possible that our study lacked sufficient power to detect associations between changes in functional status, psychological status, and learned helplessness and unmet expectations.
Future research should examine the relationship between learned helplessness and other psychosocial factors and patients' tendencies to verbalize their expectations during the visit. Studies should also examine arthritis patients' expectations as they relate to disease duration and the health care setting in which the patient is seen (primary care versus rheumatology). Finally, longitudinal studies of unmet expectations should be performed to determine whether the same patients are likely to report unmet expectations over a series of visits or whether expectation fulfillment has an impact on arthritis-related outcomes.
In conclusion, at least one-third of arthritis patients report unmet expectations after visiting their rheumatologists. Attention to the patients' sense of helplessness and pain level, as well as the amount of time spent with the patient during the clinic visit, may prevent reports of unmet expectations.
The authors wish to thank the 6 rheumatologists in Indianapolis who kindly allowed us to recruit their patients for this study. We also thank Kimberly Mihaliak, BSc, and Maureen Reindl, BS, for their assistance with patient recruitment and data collection.
- 1Committee on Quality of Health Care in America, Institute of Medicine. Crossing the quality chasm. Washington (DC): National Academy Press; 2001.
- 2Through the patient's eyes: understanding and promoting patient-centered care. San Francisco: Jossey-Bass; 1993., , , .
- 36A “psychological” 4-item questionnaire in the format of the Modified Health Assessment Questionnaire (MHAQ) to assess difficulties with sleep, anxiety, depression, and stress in the standard HAQ format. Arthritis Rheum 1997; 40 Suppl 9: S111., .
- 39Leflunomide Rheumatoid Arthritis Investigators Group. Clinical improvement as reflected in measures of function and health-related quality of life following treatment with leflunomide compared with methotrexate in patients with rheumatoid arthritis: sensitivity and relative efficiency to detect a treatment effect in a 12-month, placebo-controlled trial. Arthritis Rheum 2000; 43: 506–14., , , , , , et al,
- 45“Arthritis”: variation in beliefs about joint disease. Med Anthropol 1983; 7: 29–46..