Hidden in Plain Sight: Medical Visit Companions as a Resource for Vulnerable Older Adults

Background  Sicker patients are less satisfied with the quality of health care they receive than their healthier counterparts.

Methods  A sample of 12 018 community-dwelling Medicare beneficiaries 65 years or older from the 2004 Medicare Current Beneficiary Survey was studied. Multivariate regression was used to describe whether beneficiaries' self-reported satisfaction with their usual-care physician was related to the presence or functions assumed by visit companions.

Results  Overall, 38.6% of beneficiaries reported being typically accompanied to routine medical visits. Accompanied beneficiaries were older, less educated, and in worse health than their unaccompanied counterparts. More than 60% of companions facilitated visit communication by recording physician instructions (44.1%), providing information regarding patients' medical conditions or needs (41.6%), asking questions (41.1%), or explaining physicians' instructions (29.7%). After controlling for sociodemographic and health differences, accompanied beneficiaries were more highly satisfied with their physician's technical skills (odds ratio [OR], 1.15; 95% confidence interval [CI], 1.02-1.30), information giving (OR, 1.19; 95% CI, 1.05-1.35), and interpersonal skills (OR, 1.18; 95% CI, 1.03-1.35) than unaccompanied beneficiaries. Accompanied beneficiaries whose visit companions were more actively engaged in communication rated physician information giving (OR, 1.42; 95% CI, 1.14-1.77) and interpersonal skills (OR, 1.29; 95% CI, 1.05-1.59) more favorably. This relationship was strongest among beneficiaries with the worst self-rated health.

Conclusions  Visit companions are commonly present in older adults' routine medical encounters, actively engaged in care processes, and influential to patients' satisfaction with physician care. More systematic recognition and integration of visit companions in health care processes may benefit quality of care for a particularly vulnerable patient population.

Concerns regarding deficits in care quality and rising health care costs have grown in recent years, fueled in part by anticipated, demographically driven increases in health services use. In response, numerous quality improvement activities have been undertaken, including advances in information technology, the creation of novel chronic care delivery models, and performance-based physician payment schemes. Yet few of these endeavors have successfully addressed deficiencies in quality of care¹ or dissatisfaction with care^2,3 among those patients who are the sickest. Moreover, these initiatives seldom address the experience of patients and their families during face-to-face medical encounters, even though these experiences are thought to represent “true north” in quality care.⁴

There has been a growing appreciation of families' relevance to patient care^5-8; however, specific knowledge is limited regarding what attributes of their involvement are valued or helpful to patients or efficacious in terms of quality-of-care outcomes. Likewise, information regarding the frequency, characteristics, and impact of family involvement during medical visits has thus far been limited to small studies^9-11 with restricted samples of patients. This study contributes to knowledge on the topic by presenting population-based estimates of the number and characteristics of Medicare beneficiaries who report being typically accompanied to routine medical visits. Attributes of family members and friends who accompany patients (for simplicity, referred to as visit companions) and the functions that they perform are also described. Finally, we examine whether accompaniment status of beneficiaries, or functions assumed by visit companions, have a bearing on beneficiaries' satisfaction with their usual physician.

The Medicare Current Beneficiary Survey (MCBS) is a rotating panel survey of a nationally representative sample of Medicare beneficiaries. The sample for this study consists of 12 018 community-dwelling Medicare beneficiaries 65 years or older who were continuously enrolled throughout the 2004 calendar year. In-person interviews were conducted with the beneficiary or with his or her proxy if the beneficiary was unable to respond. With sample weights, study participants are representative of approximately 30 million community-dwelling Medicare beneficiaries.

Accompaniment status was ascertained by the participants' responses as to whether they are “usually” accompanied to their usual source of care. Visit companion identity refers to the relationship of the accompanying individual to the study participant. Respondents endorsed up to 9 functions indicating the reasons for which participants were accompanied, including assistance with logistics (transportation, physical assistance, appointment scheduling), engagement in care processes and visit communication (taking notes, asking questions, providing information to the physician about the patient's medical condition, explaining physicians' instructions to the patient, and acting as an English language translator), and the provision of company and moral support.

Beneficiaries' age, sex, education, self-rated health, as well as 14 self-reported chronic conditions, hearing, and vision impairment were reported. Disability was defined as having received help with performing—or not performing owing to a health or physical problem—any of 5 instrumental activities of daily living (IADLs) or 5 activities of daily living (ADLs), categorized as “not disabled,” “IADL disabled only,” or “ADL disabled.” Total annual Medicare expenditures, summarized from administrative claims data, are presented for the subset of participants who were continuously enrolled in fee-for-service Medicare throughout the calendar year.

Ratings of beneficiaries' satisfaction with their usual physician were constructed from a series of 12 questions. Responses were elicited on a 4-point Likert scale and were recoded as necessary so higher values corresponded to more favorable ratings of beneficiaries' usual physician.

Factor analysis was used to reduce the number of items and create a coherent structure for measuring beneficiaries' satisfaction with their usual physician and functions performed by visit companions. Responses regarding satisfaction with participants' usual physician were grouped into 3 subscales: “technical quality,” “information giving,” and “interpersonal skills” (range, 0-12; Cronbach α = 0.85, 0.92, and 0.84, respectively), which is consistent with other reports regarding the psychometric properties of these measures.¹² To construct more meaningful measures and to delineate those beneficiaries who rated their usual physician most favorably, we created dichotomous categorical variables for each subscale at the top third of satisfaction ratings.

Aside from translating English language, the 4 functions describing companion engagement in visit communication (taking notes, asking questions, providing information to the physician about the patient's medical condition, explaining physicians' instructions to the patient) loaded on a single factor and demonstrated good internal consistency (Cronbach α = 0.77). These items were combined to create a summary scale ranging from 0 to 4, indicating the number of aggregate communication functions assumed by accompanied beneficiaries' visit companions.

We employed χ² and t tests to examine the strength of differences in beneficiaries' sociodemographic and health characteristics by accompaniment status. Visit companions' relationship to beneficiaries was similarly examined, as well as tasks they were reported to fulfill, stratified by the beneficiaries' self-rated health. Beneficiaries' satisfaction with their usual physician was examined, stratified by accompaniment status, and adjusted for sociodemographic and health characteristics of the 2 groups. Last, multivariate regression models were constructed to examine communication functions performed by visit companions in relation to measures of accompanied beneficiaries' satisfaction with their usual physician, adjusting for sociodemographic and health characteristics. All analyses were conducted with SAS statistical software (version 9.1; SAS Inc, Cary, North Carolina) using sampling weights in conjunction with SAS survey procedures to account for the complex multistage survey design of the MCBS.

Analyses excluded 531 of the 12 018 participants (4.4%) who refused or did not respond to the question regarding accompaniment to routine medical encounters. Proxy respondents (n = 956) are included in all analyses except those related to beneficiaries' satisfaction with their usual physician. Proxy respondents were more likely to report for participants who were accompanied to routine medical encounters and who were older, male, less educated, more disabled, and had received a diagnosis of stroke, dementia, hip fracture, Parkinson disease, psychiatric disorder, or had vision or hearing impairment. Sensitivity analyses that separately included all participants regardless of item nonresponse on physician satisfaction measures or reliance on proxy respondents did not meaningfully change findings.

More than one-third (38.6%) of beneficiaries reported being usually accompanied on medical visits by a companion (Table 1). Accompanied beneficiaries were on average more than 3 years older (77.7 vs 74.1 years) and approximately twice as likely to have less than a high school education and to report fair or poor self-rated health than those who were unaccompanied. Accompanied beneficiaries carried a greater burden of aggregate chronic disease; each of the 16 chronic conditions listed in Table 1 was more highly prevalent among accompanied beneficiaries. Disability was also substantially greater among accompanied beneficiaries, both in terms of IADL (20.0% vs 5.1%) and ADL (16.5% vs 1.4%) impairment. Accompanied beneficiaries incurred mean annual Medicare expenditures that were more than twice as high as their unaccompanied counterparts ($8326 vs $4097; P < .001 for all comparisons).

Visit companions were spouses (53.3%), adult children (31.9%), other relatives (6.8%), roommates, friends, or neighbors (5.2%), or nonrelatives (2.8%); less than 1% of beneficiaries reported their visit companion to be a nurse, nurse aide, or legal or financial officer (data not shown). The relationship of visit companion to beneficiary varied by health; with worse self-rated health, spouses were significantly less likely, and adult children or other nonrelatives more likely, to serve as visit companion (P < .001).

As evident in Table 2, visit companions were commonly engaged in visit communication. They were reported to record physician comments and instructions (44.1%), communicate information about beneficiaries' medical condition to their physician (41.5%), ask questions (41.0%), explain physicians' instructions (29.7%), and translate English language (3.3%). Most visit companions (63.8%) engaged in at least 1 of these activities. Visit companions who performed at least 1 of the activities were fairly evenly distributed across number of functions performed, with 22.4% engaged in a single task, 26.6% in 2 tasks, and 24.4% in 3 tasks, and 26.7% performing all 4 communication functions (data not shown).

Visit companions were also reported to be present for company and moral support (28.4%) and to assist with logistics such as transportation (52.3%), scheduling appointments (16.6%), and the provision of physical assistance (8.4%). Given worse self-rated health, visit companions were incrementally more likely to fulfill each of the specified functions, with the exception of providing company and moral support. This was reported most often among beneficiaries who rated their health to be “excellent.”

The results of simple and multivariate logistic regression models examining the relationship between beneficiaries' accompaniment status and satisfaction with their usual physician are displayed in Table 3. Accompaniment status was unrelated to the likelihood of reporting high satisfaction in unadjusted models. However, after adjusting for sociodemographic and health differences in the 2 groups, accompanied beneficiaries were more highly satisfied with their usual physician's technical skills (odds ratio [OR], 1.15; 95% confidence interval [CI], 1.02-1.30), information giving (OR, 1.19; 95% CI, 1.05-1.35), and interpersonal skills (OR, 1.18; 95% CI, 1.03-1.35).

Multivariate regression models were constructed to examine the relative effects of patients' sociodemographic and health characteristics and the intensity of companion engagement in visit communication on accompanied beneficiaries' satisfaction (Table 4). Beneficiaries whose visit companion performed more than 2 of 4 communication functions were significantly more highly satisfied with their physicians' information giving (OR, 1.42; 95% CI, 1.14-1.77) and interpersonal skills (OR, 1.29; 95% CI, 1.05-1.59) than those whose companions were less active in visit communication. A similar, albeit weaker, effect was found with regard to the relationship between actively engaged visit companions and beneficiaries' satisfaction with physicians' technical skills (OR, 1.22; 95% CI, 0.98-1.52). Having less than a high school education (compared with greater education) and “fair” or “poor” self-rated health (compared with “excellent,” “very good,” or “good” health) were also notable correlates of accompanied beneficiary satisfaction, the former being somewhat stronger than the latter across all 3 satisfaction dimensions.

A graphic representation of this relationship is depicted in the Figure, stratified by beneficiaries' self-rated health. Accompanied beneficiaries in “fair” or “poor” health whose companion was not reported to have engaged in any communication functions rated all 3 aspects of satisfaction more negatively than their counterparts in “excellent”/“very good” or “good” health. However, the inverse relationship between health status and satisfaction diminishes incrementally with greater companion engagement in communication processes. This relationship is less apparent for beneficiaries' ratings of physicians' technical skills than for other satisfaction dimensions.

This study presents evidence regarding a potentially valuable but largely unrecognized quality of care resource. Findings establish that visit companions, most often spouses and adult children, are commonly present in older adults' routine medical encounters, actively engaged in the exchange of health information between patients and their physicians, and influential in patients' perceptions of physician interpersonal rapport and information giving. Moreover, visit companions tend to accompany patients who are especially vulnerable; in this study, accompanied patients were older, less educated, and in worse health than their unaccompanied counterparts. We believe these results, taken together, merit the attention of policy makers, physicians, and researchers interested in improving quality of care for the sickest patients, who tend to be least satisfied with the quality of their health care.

Our most notable finding was that accompanied beneficiaries' experience of care, as reflected by their satisfaction with physicians' interpersonal and information-giving skills, was at least in part a function of the intensity of companions' involvement in visit communication. The strength of this relationship was most notable among individuals with the worst self-rated health. Specifically, for the sickest patients, the greater number of communication functions they reported being performed by their visit companion, the more highly they rated the information giving and interpersonal skills of their usual-care physician. That the inverse relationship between beneficiaries' health status and satisfaction with their usual physician was attenuated with greater companion involvement in visit communication suggests that the presence and actions of visit companions may bridge communication barriers in the care of especially vulnerable patients.

The dynamics of physician-patient communication have been found to suffer with greater complexity of illness. Physicians have been found to less comprehensively solicit information from sick patients,^13,14 to be more task-oriented in their interactions,^3,15,16 and to be less apt to engage in social dialogue.^15,17,18 Physicians report liking sicker patients less than patients who are healthy; they rate corresponding visits as less satisfying.^17,18 In turn, individuals in poor health like their physicians less and are less satisfied with their care.^2,3,18,19 This inverse relationship between patients' health and satisfaction with care is of concern in that satisfaction is a quality-of-care indicator and is linked to valued outcomes for both patients and physicians, ranging from patient adherence and appointment keeping, to physician satisfaction, burnout, and malpractice litigation.^18,20,21

Because information in this study was limited to usual behaviors for a defined set of tasks, we can only speculate as to the mechanism by which visit companions influenced participants' perceptions of care. However, observational studies are consistent with our results and provide further insight to the processes at play. Two parallel studies^9,22 of audiotape-recorded medical encounters found visit companions to commonly facilitate physician and patient understanding by clarifying or expanding on patient history, asking questions of the physician, and prompting patient involvement. Patients who were older and in worse health held greater expectations for companion involvement,²² and their companions were in turn more verbally active.^9,22 In 1 study,⁹ patients whose visit companions prompted their engagement in medical dialogue were more than 4 times as likely to be actively engaged in decision making. Collectively, these studies suggest that the presence and roles assumed by visit companions are salient to quality of patient-physician communication and delivery of patient-centered care.

To our knowledge, this is the first study to draw from a nationally representative survey to examine patient accompaniment to medical encounters, and results are consistent with the published literature in documenting that accompanied patients are not only disproportionately older^11,22,23 and in worse health,^11,22,24 but that they also tend to be less educated than their unaccompanied counterparts.^11,22,25 Health literacy deficits have been linked to a host of unfavorable health and health services outcomes, including suboptimal physician-patient communication, as well as poor recall and understanding of treatment recommendations.²⁶ The extent to which the presence of a visit companion has a bearing on the amelioration of health literacy deficits merits further attention and research.

As with any survey, information collected is only as good as the questions asked and the interpretation, memory, veracity, and accuracy of the respondents. In this regard, we were restricted to questions fielded by the MCBS. Some respondents may have interpreted “usually accompanied to your medical visit” as including companions who remained in the waiting room. Because of this, we could not accurately distinguish whether visit companions identified as providing only logistical assistance, such as transportation, were actually present during the encounter. Therefore, we restricted our analyses of companion functions to those specific to medical encounter communication processes. In a similar vein, questions regarding disruptive or unhelpful companion behaviors were not included in the survey, and these may have acted to impede rather than facilitate communication for some patients. However, 2 observational studies found unhelpful companion behaviors to occur less frequently and to be of less consequence than facilitative behaviors, which suggests that the absence of these questions may not have dramatically changed our conclusions.^9,22

Although we believe that it is the communication functions assumed by visit companions, rather than their mere presence in the visit, that are most influential to patients' satisfaction, we cannot be sure of the causal mechanism. It is possible that the relationship is due to unmeasured factors, such as physicians' spending more time with actively accompanied patients or physicians’ simply being more pleasant when a companion is perceived to be an engaged advocate.

We recognize that patient ratings of satisfaction are subjective in nature and may diverge from more objective indications of clinical quality.²⁷ That said, patients' satisfaction has been established as conveying meaningful information regarding the experience of receiving care, such as unmet expectations, quality of information giving and communication, and patient-centeredness.^21,28,29 The quality of patient-physician communication, in turn, has been shown to have clinically important implications for a range of valued outcomes, such as patients' knowledge of their illness, treatment regimen adherence, and physiologic health outcomes.^20,21,30

Findings from this study are limited to routine medical encounters among community-dwelling older adults. For example, prenatal visits, oncology care, and treatment for acutely ill, hospitalized, cognitively impaired, or dying patients all may involve family members or friends to varying degrees. Elucidating the frequency, roles, and implications of visit companions to the experiences of these populations is beyond the scope of this study, although the positive observed effects may be operable in broader health contexts.

For physicians, these data highlight the potential value of visit companions to assist them in meeting the informational and interpersonal needs of their most vulnerable older patients. In this regard, it is especially notable that the magnitude of positive effect on patient satisfaction related to having an engaged companion is comparable to or greater than the negative impact of either low education or poor health status. Those who fear that the presence of a family member may intrude on the intimacy of the physician-patient relationship, jeopardize confidentiality, or constitute an assault on patient autonomy^8,10 may be comforted and surprised by the generally positive effect of visit companions on the care experience of their elderly patients.

Given that patient-physician communication has been shown to be amenable to improvement, educational interventions that prepare physicians for interactions with accompanied patients could enhance their ability to capitalize on the presence of visit companions. Similarly, interventions that help companions clarify their functions in the visit and align these with patient expectations may enhance the effect of their efforts.²² In light of the general success of patient activation interventions,^20,31,32 reliance on visit companions to facilitate patient activation merits consideration.

To the extent that visit companions' involvement with patients extends beyond the medical encounter, these findings are relevant to initiatives aimed at improving patient safety and treatment adherence. To this point, a large cohort study³³ of Medicare enrollees found some 20% of serious adverse drug events in the ambulatory setting to be potentially avoidable and related to miscommunication and misunderstanding. It is also reasonable to surmise that families and friends who are engaged in medical visit dialogue may be more knowledgeable and better prepared to support health-related activities, such as medication compliance, in the community. In this regard, a meta-analysis³⁴ reported individuals who received practical assistance from others to be 3.6 times more adherent to their treatment regimens.

Several successful innovations in geriatric care have explicitly targeted high-risk older adults and their families^35-37 and suggest potential synergies emanating from integrated partnerships between trained health professionals and older adults' informal support networks. In light of widespread fragmentation and quality deficiencies in health care, much may be gained from more explicitly recognizing and better integrating family visit companions as contributing members of older adults' health care team. Results presented in this article suggest that patients' visit companions, hidden, but in plain sight, are a valuable quality of care resource whose efforts, if further optimized, could enhance the experience of care for millions of vulnerable Americans.

Correspondence: Jennifer L. Wolff, PhD, Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, 624 N Broadway, Room 692, Baltimore, MD 21205 (jwolff@jhsph.edu).

Accepted for Publication: January 13, 2008.

Author Contributions:Study concept and design: Wolff and Roter. Acquisition of data: Wolff. Analysis and interpretation of data: Wolff and Roter. Drafting of the manuscript: Wolff and Roter. Critical revision of the manuscript for important intellectual content: Wolff and Roter. Statistical analysis: Wolff and Roter. Administrative, technical, and material support: Wolff. Study supervision: Roter.

Financial Disclosure: None reported.