To the Editor:
Adults discharged from the hospital with skilled home health care (HHC) are at high risk for preventable adverse events, including medication errors and hospital readmissions.1–3 Health Information Exchanges (HIEs) are increasingly being used in multiple care settings across the United States.4 HIEs are designed to allow clinicians in different care settings to access a patient’s medical information electronically, and have potential to improve information exchange between hospitals and other care settings.5–7 The association between HIE participation during hospital to home health agency transitions and hospital readmissions has not previously been examined. Therefore, we completed a secondary data analysis using the statewide Colorado all payer claims database (COAPCD) to examine if hospital and/or home health agency participation in a regional HIE is associated with reductions in 30-day readmissions.
The study sample included Medicare and Medicare Advantage (MA) beneficiaries hospitalized in Colorado between January 1, 2014, and August 31, 2018. The primary outcome was 30-day hospital readmission and the primary independent variables were HIE participation by the hospital and/or home health agency for each hospital to HHC transition. In multivariable generalized linear regression, additional patient, hospital, and home health agency characteristics that might influence the association between readmissions and HIE participation were included in the model.
After exclusions (see Supplementary Figure 1), this analysis included 46,903 individuals with 54,016 hospital to home health transitions. Patients were 75 years old (median), 57.8% were female, 82.8% had Medicare, and 17.2% had MA insurance (see Supplementary Table 1). The overall 30-day readmission rate was 12.8%. Table 1 shows patient, hospital, and home health agency characteristics by 30-day readmissions. Hospital HIE participation [odds ratio (OR) 0.84, 95% confidence interval (CI) 0.79–0.89] but not home health agency HIE participation (OR 1.02, 95% CI 0.97–1.07) was associated with lower odds of 30-day readmissions in unadjusted models. After adjusting for multiple covariates, neither hospital nor home health agency HIE participation was associated with lower 30-day readmissions (adjusted hospital OR 0.94, 95% CI 0.87–1.02; adjusted home health agency OR 0.99, 95% CI 0.93–1.05).
Table 1.
Patient, Hospital. and Home Health Agency Characteristics by 30-day Readmission (Row Percentages)
| Variable | Total n = 54,016 | 30-Day Readmission |
P Value | |
|---|---|---|---|---|
| No n = 47,116 | Yes n = 6900 | |||
| Patient characteristics | ||||
| Age, median (IQR) | 75.0 (68.0–83.0) | 75.0 (68.0–83.0) | 76.0 (69.0–84.0) | <.001 |
| Length of stay, median (IQR) | 3.0 (2.0–5.0) | 3.0 (2.0–5.0) | 4.0 (2.0–6.0) | <.001 |
| Sex | <.001 | |||
| Female | 57.8 (30,981) | 87.7 (27,177) | 12.3 (3804) | |
| Male | 42.2 (22,604) | 86.6 (19,570) | 13.4 (3034) | |
| Payer | .828 | |||
| Medicare | 82.8 (44,750) | 87.2 (39,040) | 12.8 (5710) | |
| Medicare Advantage | 17.2 (9266) | 87.2 (8076) | 12.8 (1190) | |
| Charlson Comorbidity Index score | <.001 | |||
| 0 | 30.1 (16,265) | 92.9 (15,107) | 7.1 (1158) | |
| 1 | 25.1 (13,579) | 88.8 (12,054) | 11.2 (1525) | |
| 2 | 17.1 (9217) | 85.3 (7860) | 14.7 (1357) | |
| 3+ | 27.7 (14,955) | 80.9 (12,095) | 19.1 (2860) | |
| Primary diagnoses for hospitalization | ||||
| Acute myocardial infarction | .002 | |||
| Yes | 1.5 (793) | 83.6 (663) | 16.4 (130) | |
| No | 98.5 (53,223) | 87.3 (46,453) | 12.7 (6770) | |
| Chronic obstructive pulmonary disease | <.001 | |||
| Yes | 1.8 (994) | 83.5 (830) | 16.5 (164) | |
| No | 98.2 (53,022) | 87.3 (46,286) | 12.7 (6736) | |
| Pneumonia | .004 | |||
| Yes | 2.8 (1498) | 84.8 (1270) | 15.2 (228) | |
| No | 97.2 (52,518) | 87.3 (45,846) | 12.7 (6672) | |
| Stroke | .004 | |||
| Yes | 2.1 (1108) | 90.1 (998) | 9.9 (110) | |
| No | 97.9 (52,908) | 87.2 (46,118) | 12.8 (6790) | |
| Heart failure | <.001 | |||
| Yes | 2.6 (1381) | 80.1 (1106) | 19.9 (275) | |
| No | 97.4 (52,635) | 87.4 (46,010) | 12.6 (6625) | |
| Total hip or knee arthroplasty | <.001 | |||
| Yes | 21.9 (11,807) | 96.0 (11,336) | 4.0 (471) | |
| No | 78.1 (42,209) | 84.8 (35,780) | 15.2 (6429) | |
| Hospital characteristics | ||||
| Hospital type | <.001 | |||
| Acute care hospitals | 98.4 (53,141) | 87.2 (46,316) | 12.8 (6825) | |
| Critical access hospitals | 1.6 (875) | 91.4 (800) | 8.6 (75) | |
| Hospital compare quality star rating | <.001 | |||
| 2 | 3.2 (1701) | 83.4 (1418) | 16.6 (283) | |
| 3 | 39.7 (21,129) | 86.4 (18,246) | 13.6 (2883) | |
| 4 | 46.9 (24,979) | 87.6 (21,874) | 12.4 (3105) | |
| 5 | 10.2 (5432) | 89.1 (4840) | 10.9 (592) | |
| Discharges from hospital (total in cohort) | <.001 | |||
| 0–139 | 1.6 (862) | 89.9 (775) | 10.1 (87) | |
| 140–616 | 8.4 (4542) | 89.3 (4057) | 10.7 (485) | |
| 617–1622 | 28.4 (15,343) | 86.2 (13,231) | 13.8 (2112) | |
| >1622 | 61.6 (33,269) | 87.3 (29,053) | 12.7 (4216) | |
| Ownership - Hospital | <.001 | |||
| Government | 15.7 (8474) | 85.8 (7269) | 14.2 (1205) | |
| Nonprofit | 68.5 (37,016) | 87.2 (32,273) | 12.8 (4743) | |
| Proprietary | 15.8 (8526) | 88.8 (7574) | 11.2 (952) | |
| Hospital - HIE use at discharge | <.001 | |||
| Yes | 80.1 (43,292) | 87.6 (37,935) | 12.4 (5357) | |
| No | 19.9 (10,724) | 85.6 (9181) | 14.4 (1543) | |
| Home health agency characteristics | ||||
| Home health agency location | <.001 | |||
| Rural | 11.4 (6183) | 88.9 (5494) | 11.1 (689) | |
| Urban | 88.6 (47,833) | 87.0 (41,622) | 13.0 (6211) | |
| Home health compare quality star rating | .002 | |||
| ≤ 2.5 | 7.3 (3938) | 86.0 (3388) | 14.0 (550) | |
| 3 | 13.4 (7193) | 88.1 (6334) | 11.9 (859) | |
| 3.5 | 19.1 (10,222) | 87.8 (8971) | 12.2 (1251) | |
| 4 | 28.6 (15,325) | 87.2 (13,367) | 12.8 (1958) | |
| 4.5 | 23.7 (12,693) | 86.5 (10,981) | 13.5 (1712) | |
| 5 | 7.9 (4213) | 87.6 (3691) | 12.4 (522) | |
| Number of referrals to home health agencies (total in cohort) | <.001 | |||
| 0–92 | 2.7 (1444) | 84.3 (1217) | 15.7 (227) | |
| 93–212 | 7.8 (4236) | 86.4 (3658) | 13.6 (578) | |
| 213–471 | 17.5 (9439) | 86.7 (8185) | 13.3 (1254) | |
| >471 | 72.0 (38,897) | 87.6 (34,056) | 12.4 (4841) | |
| Ownership – Home Health Agencies | <.001 | |||
| Government | 2.3 (1253) | 87.5 (1097) | 12.5 (156) | |
| Nonprofit | 40.3 (21,580) | 86.6 (18,678) | 13.4 (2902) | |
| Proprietary | 57.4 (30,751) | 87.7 (26,957) | 12.3 (3794) | |
| Home health agency – all services offered | .431 | |||
| No | 4.1 (2214) | 87.8 (1943) | 12.2 (271) | |
| Yes | 95.9 (51,370) | 87.2 (44,789) | 12.8 (6581) | |
| Home health agency - HIE use at discharge | .450 | |||
| No | 43.9 (23,697) | 87.3 (20,699) | 12.7 (2998) | |
| Yes | 56.1 (30,319) | 87.1 (26,417) | 12.9 (3902) | |
Cells are % (n) by row, or median [interquartile range (IQR)]; P values from Pearson’s χ2 test or the Wilcoxon Rank Sum test.
These findings should be interpreted in the broader context of care transition interventions, in which multiple systematic reviews have identified that transitions of care interventions are more likely to reduce readmissions when they include multiple components.8–10 For example, in a systematic review with a meta-analysis of 42 studies, care transition interventions that were composed of 5 or more unique components had a lower readmission risk compared with control groups (OR 0.63, 95% CI 0.53–0.76). In this review, examples of unique components included telephone follow-up, home visits, timely communication with the primary care provider, and timely postdischarge follow-up. As a result, it might be reasonable to conclude that although HIE participation may not alone reduce 30-day readmissions, HIE participation could support other components of a care transitions program to ultimately reduce readmissions.
This analysis has multiple limitations. First, the observational nature of this study does not fully account for the self-selection of hospitals and home health agency participation in the HIE. Although we used multivariable regression to include variables that could influence the association between HIE participation and readmissions, given the complexity of information exchange and care transitions, it is likely that unmeasured confounders influenced our findings, such as actual use of the HIE for individual patients. In addition, although the COAPCD has multiple strengths, including the ability to follow individuals longitudinally across settings and payers, it lacks comprehensive race and ethnicity data, as well as home health assessment and functional data that could have enhanced this analysis and the interpretation of findings. Finally, this analysis did not capture use of information exchange mechanisms outside of HIE, such as direct access to hospital electronic health records for home health agencies, which may represent an important and unmeasured variable given the increasing use of these as supplements or alternatives to HIEs for information exchange.
In sum, although HIEs are increasingly available to promote better coordination across care settings, HIE participation for hospitals and/or home health agencies was not significantly associated with reductions in 30-day readmissions when included in a full regression model. Future work to understand how best to implement and integrate the HIE into HHC workflow could provide important insights to optimize HIE use. In addition, because multiple components are frequently included in successful care transitions initiatives, future work could aim to understand how HIE use by hospitals and/or home health could be combined with other components of transitional care (eg, HHC frontloading, primary care follow-up) to improve outcomes beyond readmissions (eg, functional outcomes, patient quality of life) for patients as they transition from the hospital to HHC.
Supplementary Material
Acknowledgments
The authors acknowledge the Colorado Regional Health Information Organization for providing key data about hospital and home health agency HIE participation.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the University of Colorado School of Medicine, the Agency for Healthcare Research and Quality, the Colorado Department of Health Care Policy and Financing, the Center for Improving Value in Health Care or the Colorado Regional Health Information Organization.
This work was supported by the Data Science to Patient Value Center at the University of Colorado School of Medicine, United States. Christine Jones is supported by grant number K08HS024569 from the Agency for Healthcare Research and Quality, United States. The Colorado Department of Health Care Policy and Financing, United States provided financial support of the acquisition of the Colorado All Payers Claims Database from the Center for Improving Value in Health Care. The institutions supporting this work had no role in the design or completion of this analysis.
Contributor Information
Christine D. Jones, Division of Hospital Medicine, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, CO, USA; Denver/Seattle Center of Innovation for Veteran-Centered and Value, Driven Care, VHA Eastern Colorado Healthcare System, Aurora, CO, USA.
Jacob Thomas, Adult and Child Consortium for Health Outcomes Research and Delivery Science (ACCORDS), University of Colorado Anschutz Medical Campus, Aurora, CO, USA.
Kate Ytell, Data Science to Patient Value Program, ACCORDS, University of Colorado Anschutz Medical Campus Aurora, CO, USA.
Marisa L. Roczen, Division of Health Care Policy and Research, Department of Medicine, University of Colorado School of Medicine, Aurora, CO, USA.
Cari R. Levy, Denver/Seattle Center of Innovation for Veteran-Centered and Value, Driven Care, VHA Eastern Colorado Healthcare System, Aurora, CO, USA; Division of Health Care Policy and Research, Department of Medicine, University of Colorado School of Medicine, Aurora, CO, USA.
Sarah R. Jordan, Division of Geriatric Medicine, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, CO, USA.
Hillary D. Lum, Division of Geriatric Medicine, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, CO, USA; VA Eastern Colorado Geriatrics Research Education and Clinical Center, Rocky Mountain Regional VA Medical Center, Aurora, CO, USA.
Mark Gritz, Adult and Child Consortium for Health Outcomes Research and Delivery Science (ACCORDS), University of Colorado Anschutz Medical Campus, Aurora, CO, USA; Data Science to Patient Value Program, ACCORDS, University of Colorado Anschutz Medical Campus, Aurora, CO, USA; Division of Health Care Policy and Research, Department of Medicine, University of Colorado School of Medicine, Aurora, CO, USA.
References
- 1.Brody AA, Gibson B, Tresner-Kirsch D, et al. High prevalence of medication discrepancies between home health referrals and Centers for Medicare and Medicaid Services home health certification and plan of care and their potential to affect safety of vulnerable elderly adults. J Am Geriatr Soc 2016;64: e166–e170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sterling MR, Kern LM, Safford MM, et al. Home health care use and post-discharge outcomes after heart failure hospitalizations. JACC Heart Fail 2020; 8:1038–1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jones CD, Falvey J, Hess E, et al. Predicting hospital readmissions from home healthcare in Medicare beneficiaries. J Am Geriatr Soc 2019;67: 2505–2510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rahurkar S, Vest JR, Finnell JT, Dixon BE. Trends in user-initiated health information exchange in the inpatient, outpatient, and emergency settings. J Am Med Inform Assoc 2021;28:622–627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jones CD, Jones J, Bowles KH, et al. Quality of hospital communication and patient preparation for home health care: Results from a statewide survey of home health care nurses and staff. J Am Med Dir Assoc 2019;20: 487–491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cross DA, Adler-Milstein J. Investing in post-acute care transitions: Electronic information exchange between hospitals and long-term care facilities. J Am Med Dir Assoc 2017;18:30–34. [DOI] [PubMed] [Google Scholar]
- 7.Cross DA, McCullough JS, Adler-Milstein J. Drivers of health information exchange use during postacute care transitions. Am J Manag Care 2019;25: e7–e13. [PubMed] [Google Scholar]
- 8.Leppin AL, Gionfriddo MR, Kessler M, et al. Preventing 30-day hospital readmissions: A systematic review and meta-analysis of randomized trials. JAMA Intern Med 2014;174:1095–1107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fønss Rasmussen L, Grode LB, Lange J, et al. Impact of transitional care interventions on hospital readmissions in older medical patients: A systematic review. BMJ Open 2021;11:e040057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Morkisch N, Upegui-Arango LD, Cardona MI, et al. Components of the transitional care model (TCM) to reduce readmission in geriatric patients: A systematic review. BMC Geriatr 2020;20:345. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.