Data for this analysis were derived from TRACE-CARE, an ancillary study to Transition and Risks in Acute Coronary Events–Center for Outcomes Research and Education (TRACE-CORE), a large longitudinal observational study of nearly 2,200 adults with ACS in Worcester, Massachusetts, and central Georgia. Details of TRACE-CORE (the parent study) have been previously described (Waring et al., 2012).

Briefly, TRACE-CORE study coordinators reviewed computerized hospital records daily for admission diagnoses related to ACS, lists of planned coronary interventions, and laboratory records to identify patients with elevated cardiac biomarkers. After confirming ACS diagnosis using criteria outlined by the American College of Cardiology and American Heart Association (Anderson et al., 2011; Kushner et al., 2009), patients’ eligibility for study inclusion was further examined using the following criteria: age ≥21 years and ability to communicate in English or Spanish. Patients were ineligible if they developed ACS secondary to another acute condition; had documented delirium;; were pregnant, imprisoned, or expected to move out of the area within 18 months; were admitted for palliative care only; or died during the index admission. Eligible patients were approached in-hospital by trained interviewers within 72 hr of admission and, after providing informed consent, completed a 60-minute interview that collected information on demographics, behavioral characteristics, and psychosocial factors. All study protocols were approved by institutional review boards.

From September 2011 to May 2013, consecutive TRACE-CORE participants from the Worcester sites were invited to enroll in TRACE-CARE, an ancillary study to TRACE-CORE designed to further examine psychosocial factors and caregiver support during the early post-discharge period. Details of the ancillary TRACE-CARE study have been previously described (Waring et al., 2015). Consenting participants (n = 585, 96% of invited participants) were contacted at home via telephone 1 week (range = 5-10 days) after hospital discharge. Of these 585 consenting participants, 127 could not be reached 1 week post-discharge and 32 refused participation in the follow-up interview. Four hundred twenty-six participants (72% of those eligible) completed a 30-minute interview during which they reported on functional ability, cognitive status, self-management behaviors, and support received from caregivers since discharge. Trained reviewers abstracted data on medical history, ACS type, and in-hospital procedures from participants’ medical records.

We calculated percentages of any, type-specific (i.e., instrumental or informational), and task-specific (e.g., help with housework, help with identifying symptoms) practical support for descriptive purposes. We then created a four-level outcome variable that categorized participants according to type of practical care support received: (a) instrumental only, (b) informational only, (c) both support types, or (d) neither support type. We examined bivariate associations between type of practical care support received and demographic, psychosocial, functional, and clinical predictors using analyses of variance, rank sum, and chi-square tests. Before analysis, all predictors were checked for normality and collinearity. Length of hospital stay was found to be right skewed and was log-transformed. No correlation coefficients between predictors exceeded 0.60 except for marital status and living situation (tetrachoric ρ = –.93); in cases where both variables were significantly associated with an outcome, marital status alone was included in the model.

We used multivariable-adjusted modified Poisson regression with robust error variance (Zou, 2004) to estimate the associations between demographic, psychosocial, functional, and clinical variables and types of practical support received. We chose this statistical approach because it is considered to be less biased for approximating relative risks when outcomes are common (Camey, Torman, Hirakata, Cortes, & Vigo, 2014). We fit three separate models, examining predictors of receipt of each type of support (instrumental only, informational only, and both types) compared with receipt of neither type of support (reference). Predictors were included in the multivariable-adjusted models if they were associated with the outcome at p < .10 in bivariate analyses. Due to previous evidence of gender-based differences in receipt of caregiver support according to marital status (Hammond, Mochari-Greenberger, Liao, & Mosca, 2012), we examined interaction terms for sex–marital status. We confirmed model goodness of fit with Pearson chi-square goodness-of-fit tests. We performed all analyses using SAS Version 9.3 (Cary, North Carolina) and Stata 13 (College Station, Texas).

Of the 426 participants who completed the 1 week post-discharge interview, 10 participants were excluded due to missing data on practical support and 10 participants were excluded due to missing data on covariates, leaving an analytic sample of 406 participants. Participants included in the final sample did not differ from those excluded based on sex, race, marital status, or education, but excluded participants were slightly older (63 vs. 61 years, p = .01) and were more likely to have had unstable angina (29% vs. 16%, p < .001).

The mean age of the sample was 61 (SD = 11) years, 73% were male, 96% were non-Hispanic White, 62% were married, and 21% lived alone. Approximately one third of the sample was college educated. Nearly one quarter had a history of coronary heart disease and 15% were impaired in activities of daily living at 1 week post-discharge. The demographic characteristics of our analytic cohort were similar to that of the parent study (Goldberg et al., 2015), but participants were generally younger than population-based studies of ACS (Manemann et al., 2015). This difference may be due to our criteria of including only patients who had survived their ACS hospitalization and excluding patients with dementia or delirium.

Our exploratory analyses revealed some interesting differences in characteristics among ACS patients reporting receipt of instrumental, informational, or both types of support during the early post-discharge period. We found that men were less likely than women to report receiving instrumental or both types of support, whereas married participants were more likely to report receiving instrumental or both types of support. Our finding that marriage was associated with greater receipt of practical support is not surprising, given prior knowledge that spouses are the primary source of caregiving support for adults with cardiovascular disease (Mosca et al., 2011). Our finding that men were less likely to report receiving instrumental or both types of support is less expected. It is possible that men were not actually less likely to receive these types of support than females, but rather that they were less likely to identify instrumental tasks performed by their caregivers as “support.” Research on caregiver support in heart failure (Buck, Kitko, & Hupcey, 2013) has found that caregiving roles often reflect traditional relationship roles. For example, if a male patient’s caregiver (e.g., a spouse or child) has traditionally assumed the cooking, cleaning, and other responsibilities in the household throughout his or her relationship, these activities may not necessarily be viewed as “caregiving support” by a male care recipient. It is also possible that receipt of certain types of practical support in men depends on their marital status—our mixed findings for interaction effects between sex and marital status cautiously suggest that unmarried men are less likely to receive some types of practical support compared tomarried men or married or unmarried women. Further research is needed to clarify whether male ACS patients are truly less likely to receive certain types of practical support, whether they identify receipt of caregiving support differently than female ACS patients, or if receipt of practical support in men differs according to their marital status.

We found that higher education level was associated with an increased likelihood of receipt of instrumental or informational support only. ACS patients with more formal education may more readily understand the importance of self-care disease management, as has been found in heart failure patients (Riegel et al., 2009; Rockwell & Riegel, 2001), and thus be proactive in enlisting caregivers to help meet their care goals during the early recovery period. Alternatively, higher education levels may serve as a proxy for higher socioeconomic status, which may enable access to a more resource- and support-rich social environment than patients with lower education levels. For example, caregivers of participants with higher socioeconomic status may be more likely to have their own source of transportation (for helping participants get to medical appointments) or more flexible work schedules (enabling free time to provide informational support). However, as the association between education level and likelihood of receipt of practical support was not consistent across all types of practical support (i.e., not associated with receipt of both types of support), further work is needed to clarify how patient education/socioeconomic status affects the need for or receipt of practical support after ACS.

In-hospital complications, found to be associated with higher likelihood of reporting receipt of instrumental support only, may serve as a marker of a traumatic hospital course or risk of post-hospital syndrome (Krumholz, 2013), either of which may result in greater need for instrumental support in the early post-discharge period. Impaired health literacy, found to be associated with a higher likelihood of reporting receipt of informational support only, may reflect a deficit in patients’ ability to understand how to manage their disease and the need for support in doing so. We were surprised to find that ADL impairment predicted increased likelihood of receiving both support types, but not instrumental support only—ADL impairments reflect a state of heightened dependence and such vulnerable patients may require a high level of practical support with both instrumental and informational activities during recovery.

Interestingly, age was not associated with receipt of practical support in this study, a finding consistent with another, smaller study examining caregiver support among ACS survivors (Molloy et al., 2008). This finding provides a counterfactual argument to the commonly held belief that caregiving is prevalent or important only in the context of older, debilitated adults, indicating that clinicians should be diligent in recognizing the need for and receipt of practical care support among ACS survivors across the age spectrum. An alternative explanation for this null finding is that, although over a third of the cohort was ≥65 years, the modest rates of functional and cognitive impairment suggest a younger physiologic age of these participants. Further research is needed to characterize predictors of receipt of practical support specifically among biologically (and physiologically) older cohorts with ACS.

Our findings of different predictors for receipt of instrumental and informational support (with some overlap) suggest that there are distinct ACS patient subgroups with unique practical support needs. ACS patients who receive instrumental support only during the early post-discharge period may understand their disease well but face physical challenges in taking care of themselves. Conversely, patients who receive informational support only may be physically capable of taking care of themselves but require help with increasing knowledge about their condition and how to manage it. Clinicians should assess both instrumental and informational support needs before discharging patients hospitalized with ACS and ensure that these distinct support needs can be met by informal caregivers or other resources.

Caregivers, particularly friends and family, may be a convenient, committed, and cost-effective resource to increase adherence to disease management regimens, alleviate patient burden, and improve outcomes in patients with ACS. As mentioned in recent guidelines released by the Hospital-to-Home Initiative of the American College of Cardiology (Wiggins, Rodgers, Didomenico, Cook, & Page, 2013), clinicians should aim to include caregivers in discharge education and discussions about disease management whenever possible. In addition to assessing participants’ instrumental and informational support needs, the clinical care team should also assess caregivers’ capacity to meet these support needs and tailor disease management regimens to fit the patients’ (and caregivers’) capacity for burden.

This study is strengthened by use of data from a prospective cohort that captured detailed information on demographic, clinical, and psychosocial factors during hospitalization as well as receipt of caregiver support during the critical early post-discharge period. We collected information on the types of practical support received by ACS patients, which has not been previously reported, and linked type of support received to patient characteristics. However, our sample, although substantially larger than a previous study examining practical support after ACS (Molloy et al., 2008), may not have been sufficiently large to detect significant associations for the rarest support group (informational support only, n = 25) or infrequent characteristics (e.g., cognitive impairment). Receipt of coronary artery bypass grafting was perfectly predictive of receipt of practical care support and thus could not be included in the regression models. Furthermore, we did not have information about the frequency or intensity of practical support received. However, 95% of participants reported that their support needs were met, so the amount of support received was perceived as adequate by the vast majority of participants. Finally, our sample was composed of mainly non-Hispanic White participants; this may limit the generalizability of our findings.