Health-related quality of life among older patients following an emergency department visit and emergency department-to-home coaching intervention: A randomized controlled trial

Abstract

Introduction

Many older patients experience repeated emergency department visits and hospitalizations and inadequate links to primary care. This fragmented care can result in anxiety, uncertainty, and poor health outcomes. This study compares the impact of an emergency department-to-home coaching intervention to usual, post-emergency department care on patient-reported health-related quality of life measures: information support, anxiety, and physical function.

Methods

This was a randomized controlled trial. Seven hundred and forty-nine chronically ill older adults presenting to emergency departments were randomized into Intervention (emergency department-to-home coaching) or Usual Care groups. Participants completed baseline, in-person and follow-up, telephone surveys. Within- and between-group differences in health-related quality of life were assessed using unweighted linear regression and propensity-weighted difference-in-difference analyses. Three Patient Reported Outcomes Measurement System measures were assessed: social health (informational support), mental health (anxiety/emotional distress), and physical health (physical functioning).

Results

Usual Care participants experienced statistically significant declines in informational support in unweighted (−3.13) and weighted (−2.84) analyses not observed in the Intervention group (−0.91 and −1.45, respectively). Self-reported anxiety was lowest and physical function highest at the time of emergency department visit. Patient-reported anxiety increased and physical function declined statistically significantly in Intervention participants.

Conclusion

Among older emergency department patients, health-related quality of life was highest at the emergency department visit and declined following the encounter. The emergency department visit per se appears to provide needed information, and in the short term, reassurance about patients’ medical conditions. The coaching intervention blunts the fall in informational support observed after usual, post-emergency department care but may heighten patients’ anxiety and awareness of chronic health conditions.

Keywords

Coordination health services research quality of life emergency services—hospital patient-reported outcomes

Many countries including the United States see increased demand for emergency department (ED) services.^1–3 Across a variety of health systems, a significant portion of ED utilizers are older individuals with complex care needs who are likely to have had prior hospital care.^3,4 There is some evidence that these older ED patients are more likely to have a connection with a primary care provider.^3,4 However, in Australia, ED patients with lower clinical urgency may actually have limited access to primary care.²

Patients in the United States often get caught in a cycle of crisis care characterized by episodic ED visits, hospital admissions, and readmissions with little connection to continuous primary or specialty care.⁵ To reduce unnecessary ED visits and hospitalizations, current U.S. health policy, as well as new payment models and incentives, emphasize the need to improve care transitions—the activities that ensure the coordination and continuity of care across health care providers and settings.⁶

Care transition interventions reduce the likelihood of readmission for patients transitioning from acute inpatient stays back to the community, in part by strengthening patients’ connections to primary care and specialists.⁷ Systematic reviews of interventions in various countries aimed at reducing ED utilization provide mixed assessments of their effectiveness.^3,8–13 There is evidence that intensive case-management interventions linked to community-based programs are associated with reductions in ED use.^9,10,12,13 However, most reviews conclude that there is limited high-quality evidence to support the development and implementation of optimal ED to home transition strategies.^3,8–13 This is unfortunate for two reasons. First, the ED is the primary portal of entry for nearly 60% of hospital admissions and 80% of readmissions among Medicare beneficiaries and some of these hospitalizations might be avoided if timely access to provider follow-up or linkages to community-based social support could be established during or immediately after the ED visit.⁵ Medicare is the U.S. health insurance program for the elderly and some individuals with disabilities. Second, as noted by the U.S. based National Quality Forum, an ED visit often represents a critical inflection or crisis point in a patient’s health trajectory. This makes transitions into and out of the ED of key importance.¹⁴ Inadequate care transitions after an emergency visit can lead to poor health-related quality of life (HRQoL) for patients,¹⁵ need for further hospital-based care,¹⁶ and suboptimal health outcomes.¹⁷ This is especially true for older adults with chronic illnesses who are frequent ED visitors.³

We recruited patients and patient caregivers as research partners to inform the development of an ED-initiated care transition intervention. Patient stakeholders emphasized that in order to interrupt the cycle of crisis care, an ED-to-home care transition intervention must address their HRQoL. Stakeholders indicated that the reasons they seek emergency care and the outcomes most important to them following ED discharge include: (1) the need for informational support about their symptoms, health conditions, and navigating the health care system, (2) reduced anxiety and emotional stress associated with health concerns, and (3) the return to baseline physical function.

Consistent with our stakeholders’ input, published evidence confirms that seniors’ HRQoL declines following discharge from acute care settings, including insufficient informational support, heightened anxiety, and loss of physical function.^15,16 Further, it is increasingly recognized that uncertainty and fear are key drivers of health care utilization including ED visits.¹⁸ This and other evidence suggests that HRQoL measures are not only important indicators of patient well-being in their own right but likely represent a key mechanism through which subsequent hospital use occurs.¹⁹

This study is part of a larger patient-centered outcomes research project designed to assess whether an ED-initiated coaching intervention can interrupt the cycle of repeat ED visits and hospital admissions, enhance linkages with primary care, and improve patient-reported, HRQoL among seniors with chronic conditions.²⁰ The purpose of the current analysis was to assess the impact of the randomized, ED-to-home transitional care intervention on patient-reported quality of life relative to usual post-ED care. In this study, we hypothesized that one-on-one coaching assistance provided by the ED-to-home care transition intervention would improve quality of life in older adults with chronic illnesses who presented for emergency care. The trial is registered at Clinicaltrials.gov (NCT02079987).

Methods

This study is a randomized controlled trial that compares outcomes in an intervention and Usual Care group. In this analysis, we used a difference and difference approach to compare baseline and follow-up HRQoL scores in the two groups.

Study population and setting

The study population was older Medicare fee-for-service (FFS) beneficiaries with chronic medical conditions who were scheduled for ED discharge, hospital observation prior to ED discharge, or patients who were frequent ED visitors (≥3 annual visits/year). Patients were eligible for study enrollment if they resided within the geographical area defined by specific zip/postal codes (to enable coach home visits), had a working telephone, and had one or more chronic medical conditions reported in the electronic health record (EHR). Patients were excluded if they had a current diagnosis of psychosis, were undergoing active cancer treatment, had dementia without a live-in caregiver, or were on hospice care or living in a skilled nursing facility.

The study was conducted in two communities with EDs associated with an academic health science center. Site 1 ED (89,000 visits/year) is a tertiary referral center serving an urban community of one million. Site 2 ED (90,000 visits/year) is a tertiary referral center with a 13-county catchment area serving a community of 250,000 and college campus. The study was approved in advance by the University of Florida Institutional Review Board (UF IRB Number 20140000).

Our sample size determination was based on anticipated reduction in hospital-based care (our primary outcome). Based on pilot data, a power analysis was conducted to determine the sample size needed to detect a 30% reduction in preventable hospital-based care (ED visits and hospital admissions) in the intervention compared to the usual-care group. Three-hundred participants in both the intervention and usual-care groups (600 total) were needed to detect a 32% reduction in composite repeat ED visits and hospital admissions using Medicare claims data with a power of 80% and alpha of 0.05 at each site.

Intervention and usual post-ED care

Participants were randomly assigned to ED-to-home care transition intervention (Intervention) or usual care (Usual Care) at each study site prior to study initiation. Participants were randomized using simple random sampling in three blocks. In the first two blocks, first one for Site 1 and the second one for Site 2, 330 participants were randomly selected from the first 660 participants to be assigned to the intervention arm and the rest to the usual care condition. Additional participants were deemed necessary for Site 2. Therefore, in a third block, an additional 25 participants were randomly selected from the next 50 participants to be assigned to the intervention arm and the rest to the Usual Care group. The sample command in the statistical programming language R with seed set to 4444 was used to initiate the first two randomizations and with seed set to 8888 for the third randomization. The randomization sequence was conducted by the project biostatistician (BB). Recruitment began 1 May, 2014 and was completed on 30 November, 2015.

Trained research associates asked patients who met inclusion criteria to participate after which informed consent was obtained. Study personnel were blinded to patient assignment until after completion of a baseline, in-person survey.

The four-week intervention was modeled on the Coleman Care Transitions Intervention (CTI), an evidence-based program designed to empower patients, improve health care quality, and reduce health care costs. The program reduced readmissions and lowered costs in hospitalized patients.⁷

Trained Area Agency on Aging (AAA) health care coaches conducted the home visit and up to three coaching phone calls to assist ED patients with the four-component intervention: (1) scheduling and attending a follow-up doctor’s office visit, (2) recognizing and responding to signs and symptoms of worsening disease, (3) reviewing and addressing concerns about current medications, and (4) facilitating patient–provider communication with a personal-health record. AAAs are designed to provide social support and other services to individuals 60 years and older in every local community in the United States. Based on our stakeholders’ advice and input from the AAA coaches, lack of transportation and difficulties obtaining adequate nutrition were identified as challenges facing seniors with chronic conditions. Therefore, the intervention also included 10 home-delivered meals to participants who needed nutritional support (n = 108) and financial assistance for transportation to post-ED doctor visits for those who qualified (n = 36). This was paid for by the research study.

Participants assigned to the Intervention received written and verbal discharge instructions from the treating ED physician and nurse and information about scheduling a home visit with the AAA health care coach. Usual Care participants received written and verbal discharge instructions from the treating ED physician and nurse and advice to follow up with their provider. Health care coaches recorded intervention activities (home visit, coaching phone calls, meals, transportation assistance).

The home visit, which lasted approximately 60 min, took place as soon as possible based on patient preference, preferably within two days of ED discharge. Each of the three coaching phone calls lasted approximately 15 min and focused on the four-program components. Caregivers were included in patient interaction during home visits.

Measures

Baseline, in-person survey

The baseline, in-person survey included validated measures of global self-rated health and patient-reported measures of social health (informational support), mental health (anxiety/emotional distress), and physical health (physical functioning) obtained from the Patient Reported Outcomes Measurement Information System (PROMIS™).²¹ PROMIS measures use a T-score metric in which 50 is the mean of a relevant reference population and 10 is the standard deviation (SD) of that population. Thus, a score of 40 is one SD below the reference population mean and a score of 60 is one SD above the mean.

Measures of sociodemographic, socioeconomic, and health-status characteristics were obtained using modified items from the 2011 Medicare Current Beneficiaries Survey (MCBS).²² Patients self-reported their chronic medical conditions. Health literacy was assessed during the baseline survey with the 66-item Rapid Estimate of Adult Literacy in Medicine (REALM).²³ Research associates read survey questions aloud and recorded participant responses on a secure, web-based database (Research Electronic Data Capture, REDCap). The baseline, in-person survey was conducted at the time of the index ED visit and before patient discharge.

Follow-up telephone survey

All participants were contacted within 31–60 days of ED discharge to participate in a telephone follow-up survey administered using best practices (e.g., 10 call attempts, rotating call attempts, refusal conversion).²⁴ Patients were again asked to report informational support, anxiety/emotional distress, and physical functioning using PROMIS™ measures.

Data analysis

We first compared sociodemographic, health status, and baseline HRQoL measures between Intervention and Usual Care groups for all participants with Chi-square and analysis of variance (ANOVA) for categorical and continuous measures, respectively. We accounted for differences in baseline patient characteristics that occurred by chance and differential loss to follow-up between groups that could influence study findings with propensity score adjustment.²⁵ Propensity score weights were estimated by first conducting separate multivariable logistic regression models to predict Intervention Group assignment and loss to follow-up. Second, from these models, predicted probabilities were estimated and multiplied. The inverse of the product of probabilities was used in weighted models. Propensity models were constructed to achieve covariate balance between the Intervention and Usual Care groups (Table 1).

Table 1.

Unweighted and weighted characteristics of participants by intervention and Usual Care group (N = 749).

	Unweighted				Weighted
	Overall (N = 749)	Intervention (N = 384)	Usual care (N = 365)		Overall (N = 749)	Intervention (N = 384)	Usual care (N = 365)
	%	%	%	P value	%	%	%	P value
Mean age (SD)	72.0 (8.2)	71.9 (8.0)	72.2 (8.5)	0.63	72.3 (9.0)	72.3 (8.8)	72.3 (9.2)	0.96
Gender				0.99				0.99
Female	64	64	64		61	61	61
Male	36	36	36		39	39	39
Non-white				0.33				0.73
Yes	51	53	49		54	55	54
No	49	47	51		46	45	46
Medicaid	43	47	39	0.02	45	45	45	0.92
Private	26	26	25	0.91	22	22	22	0.76
Other insurance	25	23	27	0.19	24	23	25	0.69
Education
Less than higher secondary	33	35	32	0.68	36	36	35	0.88
Higher secondary	29	28	29		30	30	31
Some college or more	38	37	39		34	34	34
Health literacy				0.25				0.26
Adequate	43	43	42		40	41	38
Inadequate	41	43	40		43	43	42
Unknown	16	14	18		17	15	20
Marital status				0.58				0.78
Single/never married	11	12	10		13	14	13
Separated/divorced	29	29	28		27	27	28
Married	33	30	35		30	28	31
Widowed	28	28	27		30	31	28
Emergency Severity Index^a								0.53
High acuity	42	43	42	0.78	43	45	42
Less urgent	58	57	58		57	55	58
Comorbidity count				0.13				0.35
0–2	17	15	20		18	16	20
3–4	37	39	34		37	38	35
5 or more	46	46	46		45	46	45
Chronic conditions
Heart	55	56	55	0.67	55	56	55	0.87
Hypertension	82	84	79	0.11	83	84	81	0.27
Diabetes	39	38	39	0.79	40	40	40	0.98
Depression	32	31	34	0.48	32	32	32	0.82
COPD	35	34	37	0.41	35	33	36	0.51
Mean quality of life at baseline (SD)
Physical	37.5 (8.9)	37.4 (8.8)	37.6 (9.0)	0.68	37.1 (9.5)	37.1 (9.8)	37.1 (9.2)	0.9
Information	54.1 (10.7)	53.5 (10.3)	54.8 (11.1)	0.09	54.2 (11.5)	54.2 (10.3)	54.3 (12.6)	0.92
Anxiety	56.6 (10.0)	56.3 (9.9)	56.9 (10.1)	0.39	56.4 (10.4)	56.1 (10.2)	56.8 (10.6)	0.34

Bolded values indicate significant differences between a given characteristic and intervention/usual care group (p < 0.05).

^aCategorized as high acuity (ESI = 1, 2) or less urgent (ESI = 3, 4, 5). COPD: chronic obstructive pulmonary disease; ESI: emergency severity index. Agency for Healthcare Research and Quality Emergency Severity Index (ESI): A Triage Tool for Emergency Department. Content last reviewed February 2013. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/professionals/systems/hospital/esi/index.html (accessed 14 September 2017).

We assessed between-group differences in HRQoL measures between the baseline and follow-up time points in two ways. First, we assessed within-group differences in the change in these scores with t-tests. Second, we conducted an unweighted and propensity weighted difference-in-differences (DID) analysis to assess between-group changes in HRQoL measures over time. The DID approach ensures background trends in outcomes unrelated to the program are not responsible for Intervention effects by comparing outcomes in the Intervention group to a group that is experiencing the same background trends but is not exposed to the program.²⁶ Unadjusted and adjusted findings are presented as mean differences to enhance the interpretability of changes in the PROMIS measures used in the study.

Heterogeneity of Intervention effect was assessed by testing a three-way interaction between the Intervention/Usual Care groups, time (baseline vs. follow-up), and a priori-defined subgroups (race, education level, comorbidity count, and self-reported health condition) to determine if the impact of the Intervention differed within subgroups.²⁷

Analyses followed an intention-to-treat approach where all participants were retained according to their original study assignment regardless of whether the participant completed the study protocol. This approach allows for an assessment of the effectiveness of the intervention, as opposed to the efficacy, by allowing differences in dropout rates and treatment protocol fidelity to be factored into the study results.²⁸ Intention-to-treat provides a better assessment of the effectiveness of treatment in a real-world setting compared to a treatment-received approach that only retains individuals who complete the full course of treatment.

Statistical significance was determined at the p < 0.05 level. However, of primary interest were changes in HRQoL scores that have implications for patient care. Therefore, we also estimated minimally important differences in these outcome measures.²⁹ Based on existing literature, we established a lower PROMIS score change threshold of three points.³⁰ All analyses were performed using Stata v.13.

Results

Of the 505 and 498 patients randomized to the Intervention and Usual Care groups respectively, 76% of the Intervention (n = 384) and 73% Usual Care participants (n = 365) completed the follow-up telephone survey (see Figure 1). Approximately 60% of patients assigned to the Intervention completed the home visit. Of patients who completed a home visit, 91% participated in at least one follow-up phone call with the health coach. Thirty-six patients received transportation and 108 received nutritional services.

Figure 1.

CONSORT flow diagram.

The average age of participants was 72 years (SD 8.2 years). The majority had a high school degree or less and three or more chronic conditions (Table 1). There was a statistically significant between-group difference in the proportion of patients dually eligible for Medicare and Medicaid at the time of enrollment (47% Intervention, 43% Usual Care), but there were no other statistically significant baseline between-group differences. Medicaid is the health insurance program for low-income individuals. It can serve as a form of supplemental insurance for low-income Medicare beneficiaries. Women and Medicaid beneficiaries were more likely to complete the follow-up telephone survey, while older participants and those with lower education, limited health literacy, more chronic conditions, and worse baseline physical function were more likely to be lost to follow-up. Once data were weighted by the inverse probability of group assignment and nonresponse, there were no statistically significant between-group differences (p > 0.05; Table 1).

Informational support

For all participants, informational support was highest at the time of the ED visit and fell during the follow-up period (Table 2). Patients assigned to Usual Care had a statistically significant decline in informational support score over time in both unweighted (−3.13 points) and weighted (−2.84 points) analyses not observed in patients assigned to Intervention (−0.91 and −1.45 points, respectively). The weighted average between-group difference was not statistically significant by DID analysis (p > 0.05). Taken together, the findings suggest the Intervention blunted the fall in self-reported informational support observed in Usual Care participants. Interaction findings suggest the effect may be stronger for patients with five or more comorbidities though the three-way interaction was not statistically significant (Appendix 1 in Supplemental Digital Content).

Table 2.

Unweighted and propensity-weighted average PROMIS scores for the intervention and usual care group during the baseline and follow-up periods, intention to treat (N = 749).

	Usual care			Intervention			Average difference in difference	P value
	Baseline	Follow-up	Difference	Baseline	Follow-up	Difference	Average difference in difference	P value
Unweighted
Informational support	54.83	51.70	−3.13*	53.48	52.57	−0.91	−2.22	0.05
Anxiety	56.95	56.90	−0.05	56.32	57.67	1.35*	−1.39	0.19
Physical functioning	37.64	36.48	−1.16	37.37	35.77	−1.60*	0.44	0.61
Propensity weighted
Informational support	54.26	51.42	−2.84*	54.17	52.72	−1.45	−1.39	0.25
Anxiety	56.81	57.02	0.21	56.08	57.61	1.53*	−1.32	0.23
Physical functioning	37.06	36.22	−0.83	37.14	35.67	−1.47*	0.64	0.49

*Statistically significant within-group change in PROMIS scores from baseline to follow-up (p < 0.05).

Bolded values indicate statistically significant difference in PROMIS score change between the intervention and Usual Care groups (p < 0.05).

Anxiety

Overall, anxiety increased statistically significantly from baseline to follow-up surveys in the Intervention Group in both unweighted (56.32–57.67) and weighted (56.08–57.61) analyses (p < 0.05). The between-group DID was not statistically significant (Table 2) and anxiety did not differ by patient subgroup (Appendix 2 in Supplemental Digital Content).

Physical function

Self-reported physical function declined in all participants from the index ED visit to follow-up, but declined statistically significantly in the Intervention Group in both unweighted and weighted analysis (Table 2). There were no statistically significant between-group differences in either the unweighted or weighted analyses. Patients with more comorbidities, chronic obstructive pulmonary diseases, and depression reported the lowest levels of physical function at baseline and at follow-up (Appendix 3 in Supplemental Digital Content).

Discussion

To our knowledge, this is the first study to assess the impact of an ED-to-home care transition intervention on three HRQoL patient-reported outcomes among chronically ill seniors: informational support, anxiety, and physical function. Other studies have focused on subsequent health care utilization such as return ED visits, inpatient stays, and primary care visits.¹¹ In addition, our study is one of a few that directly recruited patients at the time of their ED visit.¹¹ Understanding the impact of an ED visit per se and the transition intervention on the reported HRQoL of older, chronically ill ED patients is a critical prerequisite to developing, targeting, and tailoring strategies aimed at interrupting the cycle of hospital-based crisis care.

We hypothesized that the one-on-one coaching intervention would improve self-reported HRQoL by directly addressing older patients’ information needs and concerns regarding their symptoms and navigating the health care system. Contrary to expectations, our analysis showed that among older adults who experienced a health crisis and sought emergency care, HRQoL was highest for all participants at the time of the ED visit and declined in the days after the emergency encounter. The Usual Care group saw statistically significant declines in information support over the study period, while the Intervention group did not. Based on existing reports, the observed differences in informational support experienced by the Usual Care but not the Intervention Group were of the magnitude that reflects a clinically meaningful change in PROMIS measures. That is, the demonstrated changes in informational support in this study are consistent with changes previously reported to reflect a minimally important difference.³⁰ This is an active area of research inquiry and no change threshold exists for all populations.²⁹ Despite some uncertainty surrounding minimally important differences, our study population, with physical health on average 1.4 SDs below the national norm, reflects a vulnerable population in which additional declines in informational support may be particularly problematic. We also observed an increase in self-reported anxiety and worse physical function between baseline and follow-up among Intervention but not Usual Care participants.

Our results on informational support are consistent with an earlier pilot study where we addressed the impact of a similar Intervention on patient engagement. In that study, engagement fell post-ED visit for both Intervention and Usual Care groups, but the fall in engagement was statistically significantly less for those assigned to the coaching intervention.³¹ While a decline in HRQoL scores in both Intervention and Usual Care participants may seem counterintuitive, prior studies suggest an important component of an ED visit is the provision of advice and reassurance around a specific episode of care.^32,33 We and others have also reported that over 90% of ED patients, regardless of visit acuity, consider the situation that prompted the ED visit a true emergency and are too worried about their condition to seek care in settings other than the ED.^32,33 These results suggest the ED visit provides needed information, and in the short term, provides patients with immediate reassurance about their medical condition. This may be what leads to perceived higher levels of patient engagement and HRQoL at the index ED visit. In the days following the ED encounter, patients return home and are often on their own with limited access to advice and reassurance. Over time, the immediate benefits of the ED visit may decline causing HRQoL to fall, including perceptions of less informational support, more anxiety, and worse physical function.

According to the National Quality Forum, improving management of care transitions into and out of the ED has the potential to improve person-centered care, value, and cost efficiency.¹⁴ However, little is known about how to improve this critical transition. Inadequate information sharing between providers could cause patient anxiety and uncertainty, unnecessary or duplicative testing, return ED visits and hospital readmissions.³⁴ However, recent evidence and feedback from our patient stakeholders indicate that it is patients’ need for information and concern about the cause and course of symptoms that are the key drivers of ED use. In reality, most individuals regardless of literacy, education, or socioeconomic status, need some informational support to help them manage their health, effectively access health services and understand process and act on the information available to them. The fact that the ED-to-home intervention arrested a statistically significant decline in the unweighted informational support relative to usual post-ED care is an encouraging first step in improving the ED-to-home transition, particularly in vulnerable and hard-to-reach populations. Our findings also suggest that the ED-to-home intervention may be particularly well-suited for specific ED patients, such as those with more comorbidities.

Increased informational support may have a downside. As we noted, self-reports of anxiety increased and physical function declined in the Intervention Group. Providing information through a coaching intervention may increase anxiety and worsen perceived physical function in the short term as patients are confronted with news about serious health conditions or the actions needed to manage those conditions including following up with their provider. Indeed, published evidence suggests that nurse follow-up calls may actually sensitize ED patients to their unmet health needs.³⁵ Our patient stakeholders agreed with this possibility but also suggested anxiety could motivate patients to take action to manage their health. Further research on ED-initiated care transitions and coaching programs must determine what information to provide, how to sustain informational support over the longer term, how to eliminate or use anxiety to motivate patient engagement, and how to strengthen patient linkages to primary and specialty care after an emergency visit.

This study has limitations. First, though a 75% response rate is high, given the vulnerability of the ED population, approximately 25% of participants did not respond to the follow-up telephone survey. However, the availability of a comprehensive set of patient characteristics at baseline allowed for a robust propensity score weighting and intention-to-treat approach to minimize the impact of missing responses. Second, uptake of coaching in the Intervention Group was 60%. The vulnerability of some of the seniors who go to the ED may also explain the modest Intervention uptake. For example, our patient stakeholders suggested patients from impoverished neighborhoods may be reluctant to allow a coaching home visit for fear of being “judged” or letting someone they do not know into their homes. It is important that health services research continue to focus on the understanding of how to deliver health services to vulnerable and hard-to-reach patients, especially those who present to the ED, in order to achieve better health care outcomes for those who need the proposed services most. Last, the current study assessed outcomes at two time points, which precluded the ability to assess nonlinear changes in HRQoL over time.

Despite these limitations, this study suggests the ED visit support patients’ HRQoL during their health crisis likely by providing informational support, relief from anxiety/emotional distress, and a focus on their physical function. Therefore, despite policy and payment incentives that attempt to steer patients away from emergency care, our results demonstrating that self-reported HRQoL is highest during the ED visit relative to the period after a visit, suggest vulnerable seniors with chronic health conditions will continue to use the ED for health crises unless effective ED-initiated transition programs can be developed. Results from our pilot and now this larger randomized controlled trial indicate ED-initiated care transition interventions blunt the decline in engagement and informational support seen in the days following ED discharge.³¹ The ED may be an ideal setting to deploy care transition interventions that leverage the positive impact of the emergency encounter and help even the most vulnerable and hard-to-reach patients make more informed decisions about their health and health care.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Pines

Hilton

Weber

et al . International perspectives on emergency department crowding. Acad Emerg Med 2011; 18: 1358–1370.

Lowthian

Smith

Stoelwinder

et al . Why older patients of lower clinical urgency choose to attend the emergency department. Intern Med J 2013; 43: 59–65.

Aminzadeh

and Dalziel

WB.

Older adults in the emergency department: a systematic review of patterns of use, adverse outcomes, and effectiveness of interventions.

Ann Emerg Med 2002; 39: 238–247.

Soril

Leggett

Lorenzetti

et al . Characteristics of frequent users of the emergency department in the general adult population: a systematic review of international healthcare systems. Health Policy 2016; 120: 452–461.

Morganti

Bauhoff

Blanchard

et al . The evolving role of emergency departments in the United States. Santa Monica, CA: RAND Corporation, 2013.

Burton

Improving care transitions. Health Affairs Health Policy Brief, http://www.healthaffairs.org/healthpolicybriefs/brief.php?brief_id=76 (2012, accessed 1 June 2017).

Coleman

Parry

Chalmers

et al . The care transitions intervention: results of a randomized controlled trial. Arch Intern Med 2006; 166: 1822–1828.

Raven

Kushel

et al . The effectiveness of emergency department visit reduction programs: a systematic review. Ann Emerg Med 2016; 68: 467–483 e415.

Fan

Lukin

Zhao

et al . Interventions targeting the elderly population to reduce emergency department utilisation: a literature review. Emerg Med J 2015; 32: 738–743.

10.

Karam

Radden

Berall

et al . Efficacy of emergency department-based interventions designed to reduce repeat visits and other adverse outcomes for older patients after discharge: a systematic review. Geriatr Gerontol Int 2015; 15: 1107–1117.

11.

Lowthian

McGinnes

Brand

et al . Discharging older patients from the emergency department effectively: a systematic review and meta-analysis. Age Ageing 2015; 44: 761–770.

12.

Kumar

and Klein

Effectiveness of case management strategies in reducing emergency department visits in frequent user patient populations: a systematic review.

J Emerg Med 2013; 44: 717–729.

13.

Althaus

Paroz

Hugli

et al . Effectiveness of interventions targeting frequent users of emergency departments: a systematic review. Ann Emerg Med 2011; 58: 41–52 e42.

14.

National Quality Forum. Emergency Department Quality of Transitions of Care, http://www.qualityforum.org/ProjectDescription.aspx?projectID=83442 (accessed 7 November 2016).

15.

Chin

Jin

Karrison

et al . Older patients’ health-related quality of life around an episode of emergency illness. Ann Emerg Med 1999; 34: 595–603.

16.

Mistiaen

Duijnhouwer

Wijkel

et al . The problems of elderly people at home one week after discharge from an acute care setting. J Adv Nurs 1997; 25: 1233–1240.

17.

Lee

Stukel

Austin

et al . Improved outcomes with early collaborative care of ambulatory heart failure patients discharged from the emergency department. Circulation 2010; 122: 1806–1814.

18.

Rising

Hudgins

Reigle

et al . “I’m just a patient”: fear and uncertainty as drivers of emergency department use in patients with chronic disease. Ann Emerg Med 2016; 68: 536–543.

19.

Rising

Padrez

O’Brien

et al . Return visits to the emergency department: the patient perspective. Ann Emerg Med 2015; 65: 377–386 e373.

20.

Lutz

Hall

Vanhille

et al . A framework illustrating care-seeking among older adults in a hospital emergency department. Gerontologist 2017. DOI: 10.1093/geront/gnx102 [epub ahead of print].

21.

Cella

Riley

Stone

et al . The patient-reported outcomes measurement information system (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005-2008. J Clin Epidemiol 2010; 63: 1179–1194.

22.

U.S. Centers for Medicare and Medicaid Services. Medicare Current Beneficiary Survey: Details for 2011 Questionnaires, https://www.cms.gov/Research-Statistics-Data-and-Systems/Research/MCBS/Questionnaires-Items/2011-Questionnaires.html?DLPage=1&DLEntries=10&DLSort=0&DLSortDir=descending (accessed 3 May 2017).

23.

Davis

Crouch

Long

et al . Rapid assessment of literacy levels of adult primary care patients. Fam Med 1991; 23: 433–435.

24.

Fowler

FJ.

Survey research methods. 5th ed. Thousand Oaks, CA: Sage Publications, 2013.

25.

Rosenbaum

and Rubin

DB.

The central role of the propensity score in observational studies for causal effects.

Biometrika 1983; 70: 41–55.

26.

Dimick

and Ryan

AM.

Methods for evaluating changes in health care policy: the difference-in-differences approach.

JAMA 2014; 312: 2401–2402.

27.

Kent

Rothwell

Ioannidis

et al . Assessing and reporting heterogeneity in treatment effects in clinical trials: a proposal. Trials 2010; 11: 85.

28.

Gupta

SK.

Intention-to-treat concept: a review.

Perspect Clin Res 2011; 2: 109–112.

29.

Wyrwich

Bullinger

Aaronson

et al . Estimating clinically significant differences in quality of life outcomes. Qual Life Res 2005; 14: 285–295.

30.

Yost

Eton

Garcia

et al . Minimally important differences were estimated for six Patient-Reported Outcomes Measurement Information System-Cancer scales in advanced-stage cancer patients. J Clin Epidemiol 2011; 64: 507–516.

31.

Schumacher

Lutz

Hall

et al . Feasibility of an ED-to-home intervention to engage patients: a mixed methods investigation. West J Emerg Med 2017; 18: 743–751.

32.

Schumacher

Hall

Davis

et al . Potentially preventable use of emergency services: the role of low health literacy. Med Care 2013; 51: 654–658.

33.

Ragin

Hwang

Cydulka

et al . Reasons for using the emergency department: results of the EMPATH Study. Acad Emerg Med 2005; 12: 1158–1166.

34.

Jencks

Williams

and Coleman

EA.

Rehospitalizations among patients in the medicare fee-for-service program.

N Engl J Med 2009; 360: 1418–1428.

35.

Wong

Chow

Chang

et al . Effects of nurse follow-up on emergency room revisits: a randomized controlled trial. Soc Sci Med 2004; 59: 2207–2218.