Table 2.
Synthesis of Study Populations, Interventions, and Results
| AUTHOR (S) (YEAR) | SETTING | PATIENT INCLUSION CRITERIA | INTERVENTION DESCRIPTION | OUTCOMES INVESTIGATED | REPORTED RESULTS |
|---|---|---|---|---|---|
| Adams and Nielson (2012) | Single ED Psychiatric ED census 6,000 |
Psychiatric patients returning to the ED within 30 days of inpatient discharge from single included hospital | An alert notified ED psychiatric providers if patients had been discharged from inpatient psychiatric setting within 30 days and provided decisional support through accessing care plans and outpatient services | 30-Day psychiatric ED revisit rates | Preintervention 30-day revisit rates: 6.51% (monthly range 1.83–9.53%) Postintervention 30-day revisit rates: 4.3% (monthly range 3.53–5.56%) NB: Statistical testing not reported |
| Chan et al. (2009) | Single ED ED census 39,000 |
1. Reside in 15 zip-code regions served by three clinics 2. State they do not have a primary care physician |
Electronic referral system with embedded scheduling capabilities in the EHR for follow-up at clinics (IMPACT-ED) | 1. Frequency of 2-week follow-up at clinics 2. Frequency of 2-week ED return visits 3. Frequency of referrals during 6-month intervention period to one of three local community clinics from the ED |
2-Week absolute change in frequency of follow-up visits with an outpatient provider among patients receiving referrals: +23.8% (95% CI = 19.1–28.6), p < 0.05 Preintervention percentage of patients with 2-week ED return visits: 14.8% (95% CI = 11.3–18.3) Postintervention percentage of patients with 2-week ED return visits: post 11.7% (95% CI = 8.2–15.2), p > 0.05 6-Month change in frequency of referrals: no change |
| Hardin et al. (2017) | Single ED ED census 80,000 |
1. Referred to have a Complex Care Map by any hospital, emergency, or community health professional who believed patients could benefit from the program based on their current healthcare utilization 2. 3+ ED visits within 12 months before intervention 3. 18 Years or older |
A pop-up alert fired the first time any hospital provider (including ED providers) opened the medical record of a patient with a Complex Care Map; this alert is linked to the patient's care plan | 1. Frequency of 12-month healthcare utilization (ED visits; urgent care visits; observation/inpatient admissions LOS; computed tomography scans ordered) 2. Change in 12-month status of social and healthcare access (safe housing; insurance; primary care provider) |
12-Month change in total number healthcare visits: −5.581; percentage change: −37.4, p < 0.001 12-Month change in mean number ED visits: −4.419; percentage change: −43.1, p < 0.001 12-Month change in mean number IP visits: −0.575; percentage change: −44.4, p < 0.001 12-Month change in mean number OP visits −0.582; percentage change: −17.3, p < 0.001 12-Month change in mean number of CT scans: −0.918; Percentage change: −62, p < 0.001 12-Month change in mean LOS −0.918; Percentage change: −40.5, p < 0.001 12-Month gross healthcare cost charges ($) Total difference in means = −17.764, p < 0.001 ED difference in means = −6,290, p < 0.001 12-Month stable housing difference in means 11.5; percentage change: 14, p < 0.001 12-Month identifiable PCP difference in means: 12.1; percentage change: 14.9%, p < 0.001 12-Month insured difference in means: 13.0, percentage change: 15.6%, p < 0.001 |
| Mercer et al. (2015) | Multiple EDs 924-bed academic tertiary care center, with 36,000 inpatient discharges annually |
1. 3+ ED visits or admissions within 6 months 2. Medical, social, or behavioral complexity (defined as multiple medical comorbidities, concomitant psychiatric illness, substance abuse, homelessness, etc.) determined by a multidisciplinary team of hospital medicine, emergency medicine, psychiatry, ambulatory care, social work, nursing, risk management, and performance services (system analysts) |
Multidisciplinary teams created individual care plans for patients in the Complex Care Plan program (those with 3+ ED visits or admissions within 6 months and medical, social, or behavioral complexity). These care plans were embedded in the EHR and became visible when a patient registered in the ED. At the same time they registered, a secure page was sent to the inpatient team; the Complex Care Plan chairperson received an automated e-mail notification of the patient's ED visit at the point of patient ED registration | 1. 6- and 12-month ED visits 2. 6- and 12-month inpatient LOS 3. 6- and 12-month inpatient admissions 4. 6- and 12-month 30-day inpatient readmissions 5. 6- and 12-month direct costs (ED and inpatient) |
ED visits 6-Month change in total ED visits per enrolled patient: +42.9% (SD = 148.4%), p = 0.836; change in mean ED visits per enrolled patient: 0.7 (SD = 11.92), p = 0.836 12-Month change in total ED visits per enrolled patient: +48.4% (SD = 145.1%), p = 0.941; change in mean number of ED visits per enrolled patient: 10.2 (SD = 43.19), p = 0.941 Inpatient LOS 6-Month change in total inpatient LOS per enrolled patient: −50.8% (SD = 51.4%), p = 0.506; change in mean inpatient LOS per enrolled patient: −0.3 (SD = 4.3) 12-Month change in total inpatient LOS per enrolled patient: −37.8% (SD = 78.8%), p = 0.910; change in mean LOS per enrolled patient: −0.3 (SD = 2.27), p = 0.910 Inpatient admissions 6-Month change in total number of inpatient admissions per enrolled patient: −56.0% (SD = 41.6%), p < 0.001; change in mean number of inpatient admissions per enrolled patient: −3.9 (3.76), p < 0.001 12-Month change in total number of inpatient admissions per enrolled patient: −50.5% (SD = 43.9%), p = 0.003; change in mean number of inpatient admissions per enrolled patient: −6.1 (SD = 6.02) 30-Day inpatient readmissions 6-Month change in total number 30-day readmissions per enrolled patient: −66.0% (SD = 32.4%), p < 0.001; change in mean number of 30-day readmissions per enrolled patient: −3.7 (3.79) 12-Month change in total number of 30-day readmissions per enrolled patient: −51.5% (SD = 32.0%), p = 0.002; change in mean number 30-day readmissions per enrolled patient: −5.1 (5.71), p = 0.002 ED direct costs ($) 6-Month change in total ED costs per enrolled patient: +12.5% (SD = 147.5%), p = 0.143; change in mean ED costs per enrolled patient: −852.4 (2780.01), p = 0.143 12-Month change in total ED costs per enrolled patient: +48.0% (SD = 161.8%), p = 0.850; change in mean ED costs per enrolled patient: 1319.7 (SD = 10348.89), p = 0.850 Inpatient direct costs ($) 6-Month change in total inpatient costs per enrolled patient: −47.7% (SD = 52.3%), p < 0.001; mean inpatient costs per enrolled patient: −14264.90 (SD = 19301.75), p < 0.001 12-Month change in total inpatient costs per enrolled patient: −35.8% (SD = 76.1%), p = 0.052; change in mean inpatient costs per enrolled patient: −19923.2 (SD = 31891.69), p = 0.052 |
| Murphy and Neven (2014) | Single ED ED census 80,000 |
1. Enrollment in ED care coordination program (Consistent Care Program) based on referrals by ED physicians and/or Medicaid managed care plans 2. Enrollment in program between January 1, 2008 and December 31, 2010 3. Complete data for year before enrollment in program 4. At least three ED visits in year before index date 5. 18 years or older |
EHR push notification to the ED provider, stating that patient was a member of the care coordination program and that the patient's care plan existed in the hospital system | 1. 12-Month number of ED visits 2. 12-Month ED visit direct treatment cost 3. 12-Month ED visit direct treatment cost per visit 4. 12-Month hospital net income |
Extreme users (12 or more ED visits in prior year) 12-Month change in median no. of ED visits: −15 (95% CI −17 to −13), p < 0.001 12-Month change in direct treatment costs: −$6,091; (95% CI −$8,998 to −$4,298), p < 0.001 12-Month change in direct treatment cost per visit: −133 (95% CI −$211 to −$50), p < 0.001 12-Month change in net income: $1,925 (95% CI $1,093 to $3,159), p < 0.001 Frequent users (3–11 ED visits in prior year) 12-Month change in median no. of ED visits = −5 (95% CI −5 to −2), p < 0.001 12-Month change in direct treatment costs: −$1,285 (95% CI −$2,364 to −$492), p < 0.001 12-Month change in direct treatment cost per visit: −$88 (95% CI −$150 to −$33), p < 0.001 12-Month change in net income $431 (95% CI $112 to $878), p < 0.001 |
| Neven et al. (2016) | Multiple EDs Three EDs in separate health systems with combined ED census of 112,000 |
Selection from aggregated patients from all three hospitals: 1. 5+ ED visits at study hospitals over previous 12 months 2. ≥50% of ED visits for pain complaints or drug-seeking behavior 3. 18+ years old |
Multidisciplinary teams created care coordination plans for patients randomized to the intervention. Within 3 min of patient presenting to study ED, automatic fax of the patient-specific 8-point care plan (including recommendations against opioid prescriptions) was sent to the ED provider. In addition, a case manager met with patient when they presented in ED or, if unable to meet in ED, called to follow-up the next day | 1. 12-month ED visits (mean number and incidence) 2. Prevented fraction of opioid use (morphine milligram equivalents and opioid prescriptions with a refill, measured at month 10 of intervention) 3. 12-month total treatment costs (third-party) 4. 12-month ROI of intervention |
12-Month ED visits: Mean no.: 5.59 (SD 4.65) in intervention vs. 8.49 (SD 7.02) in control, p = 0.0003 OR 0.673, 95% CI = 0.538–0.841 p < 0.001 12-Month opioid prescriptions from the ED: Mean no.: 0.28 (SD = 0.74) in intervention vs. 1.44 (SD 2.05) in control, p < 0.001 OR = 0.21, 95% CI = 0.122–0.353, p = 0.001 Prevented fraction of opioid use at month 10: Morphine milligram equivalents: 43.7% (95% CI = 41.4–45.9), p < 0.001 Opioid prescriptions with a refill: 58.4% (95% CI = −27.8 to 86.5), p = 0.136 Estimated 12-month costs: Estimated 12-month third-party cost difference: $5,785 in intervention vs. $8,985 for control, (−$3,200 cumulative cost difference, SE = 1,345), p = 0.002 Estimated 12-month ROI of intervention: $3.39 per $1 spent (SE = 1.85), p = 0.07 |
| Rathlev et al. (2016) | Multiple EDs Two EDs within the same health system with censuses of 16,000 and 27,000 |
Monthly review by the High Frequency User Task Force of all potentially eligible patients, defined as:1. 1. 4+ ED visits to Baystate Health System in previous 12 months 2. A history of opioid use disorder (defined by E-codes for opioid overdose, a referral to the High Frequency User Task Force by a treating provider, or supportive evidence from the local EHR or state Prescription Drug Monitoring Program) per review of the prescription Drug Monitoring Program review of the State Prescription Drug Monitoring Program 3. 18+ years old |
A push electronic alert appeared the first time any hospital provider (including ED, inpatient, and nursing staff) accessed the chart of a patient who had a care plan developed by the High Frequency User Task Force and who was randomized to intervention | 1. Opioids prescribed to patients upon discharge and administered to ED and inpatients 2. 12-Month total medical charges within hospital system 3. 12-Month number of ED visits 4. 12-Month number of ED visits with advanced radiologic imaging 5. 12-Month number of inpatient admissions |
Morphine milligram equivalents administered to ED/inpatients: ratio treatment vs. control: 0.29 (95% CI = 0.07–1.12), p = 0.07 Morphine milligram equivalents prescribed to discharged patients: ratio treatment vs. control = 0.11 (95% CI = 0.01–0.92), p = 0.04 12-Month total medical charges: ratio treatment vs. control: 0.92 (95% CI = 0.31–2.7), p = 0.88 12-Month no. of ED visits: treatment mean change = −10.7 (95% CI = −17.5 to −4.0), control mean change −12.8 (95% CI = −19.8 to −5.8), p = 0.68 12-Month no. of ED visits with CT/MRI: treatment mean change = −5.7 (95% CI = −10.0 to −1.4), control mean change −5.8 (95% CI = −9.1 to −2.5), p = 0.98 12-Month no. of inpatient admissions: treatment mean change = −2.6 (95% CI = −5.0 to −0.2), control mean change −1.3 (95% CI = −2.8 to 0.2), p = 0.46 |
| Ringwalt et al. (2015) | Multiple EDs Thirteen of the EDs in a single health system (unknown total number of EDs) with 12 million patient interactions each year |
1. 10+ ED visits between October 2010 and September 2011 2. At least two of those visits with a discharge code related to nonspecific subjective pain 3. 18–89 years old |
EHR alert to ED providers of patient making multiple visits for treatment of chronic noncancer pain, notifying them that a multidisciplinary team suggests the patient receive care from a community-based provider. The alert also prompted providers to provide a list of community resources for the patient and not prescribe opioids. Letters were also sent to patients and community-based providers stating the patient would no longer receive opioids from the ED. | 1. 12-Month number of ED visits 2. Opioid administration (in the ED, and prescribed at discharge) |
12-Month no. of ED visits: 11.9 (SD 13.8) in intervention vs. 16.6 (SD 14.8) in control, unadjusted, p < 0.0001 12-Month mean no. of ED visits: 0.690 (95% CI = 0.57–0.84), p < 0.0002 Opioids administered in ED: intervention 16% of visits (95% CI = 0.15–0.18), control 26% of visits (95% CI = 0.25 to −0.28), p-value not reported No. of prescribed opioids, adjusted for baseline opioids: intervention vs. control rate ratio 0.57 (95% CI = 0.46–0.70), p < 0.0001 |
| Stokes-Buzzelli et al. (2010) | Single ED census 95,000 |
1. Query of 100 highest number of ED visits within ED 2. 18+ Years old 3. Without sickle cell anemia (managed in a separate program) |
Multidisciplinary “Community Resources for Emergency Department Overuse [CREDO]” volunteer team summarized (twice per month) all eligible ED patients' medical histories, and created/reviewed individualized care plans for them that were embedded in the EHR. These patients were flagged in the EHR so that all other providers had access to the previously determined care plan | 1. 6-Month number of laboratory studies ordered 2. 6-Month total number of visits 3. 6-Month total ED contact time 4. 6-Month average LOS 5. 6-Month total ED charges |
6-Month no. of laboratory studies ordered: preintervention mean: 1,847 (SD = 1,826); postintervention mean = 1,328 (SD = 1,191), (95% CI = −1,252 to −26) p = 0.04 6-Month average no. of ED visits: preintervention mean = 67.4 (SD = 47.4); postintervention mean = 50.5 (SD = 49.0), (95% CI = −33 to −0.3), p = 0.046 6-Month ED LOS (minutes): preintervention mean = 388 (SD = 186); postintervention mean = 342 (SD = 180), (95% CI = −98 to 6) p = 0.08 6-Month total emergency department contact time (hours): preintervention mean = 443.7 (SD = 381.7); postintervention mean = 270.6 (SD = 245.8), (95% CI = −17,072 to −3,701); decrease of 39%, p = 0.003 (Represents a mean of 7.21 days less in the ED) 6-Month total ED charges: preintervention mean = $64,721 (SD = $52,448); postintervention mean = $49,208 (SD = $49,239), (95% = CI −$30,943 to $83), p = 0.049 |
| Stowe (2011) | Single ED Census not specified |
Any cancer patient presenting to a single ED in a healthcare trust | RAPA system to provide real-time notifications to nominated health workers when known cancer patients presented to the ED | 1. No. of alerts received in 16 weeks 2. LOS for nonelective admissions for 16 weeks 3. Hospital system savings (British pound and bed-days) 4. User experience |
16-Week total no. of alerts received: 155 16-Week estimated no. of bed-days saved: 650 (Results reported for lung cancer patients only) Projected annual savings: 143,000 pounds Qualitative data: positive provider experience, system promoted better communication between outpatient providers and ED providers |
CI, confidence interval; CT, computed tomography; ED, emergency department; EHR, electronic health record; IP, hospital inpatient; LOS, length of stay; MRI, magnetic resonance imaging; OP, hospital outpatient; RAPA, recurring admissions patient alert; ROI, return on investment; SD, standard deviation; SE, standard error; TAU, treatment as usual.