If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Internal Medicine, Neuchâtel Hospital Network, Rue de la Maladière 45, Neuchâtel CH-2000, SwitzerlandDepartment of Internal Medicine, Geneva University Hospitals (HUG), Gabrielle-Perret-Gentil 4, Geneva CH-1205, SwitzerlandUniversity of Geneva, Rue Michel-Servet 1, Geneva CH-1211, Switzerland
Department of Medicine, Neuchâtel Hospital Network, Neuchâtel, SwitzerlandDepartment of Medicine, University of Lausanne, Switzerland, Division of Internal Medicine, Bern University Hospital, Bern, SwitzerlandBrigham and Women's Hospital, Harvard Medical School, Boston, MA, USA
Hospital readmission after discharge from internal medicine unit is frequent.
•
Post-discharge care are important to mitigate the readmission risk.
•
General practitioner (GP) consultation decrease hospital readmission.
•
GP consultation is associated with less emergency room visits.
•
Specialist consultation may be associated with more hospital readmission.
Abstract
Background
The period following hospital discharge is one of significant vulnerability. Little is known about the relationship between post-discharge healthcare use and the risk of readmission.
Objectives
To explore associations between medical consultations and other healthcare use parameters and the risk of 30-day unplanned hospital readmission.
Methods
Between July 2017 and March 2018, we monitored all adult internal medicine patients for 30 days after their discharge from four mid-sized hospitals. Using follow-up telephone calls, we assessed their post-discharge healthcare use: consultations with general practitioners (GPs) and specialist physicians, emergency room (ER) visits, and home visits by nurses. The binary outcome was defined as any unplanned hospital readmission within 30 days of discharge, and this was analyzed using logistic regression.
Results
Of 934 patients discharged, 111 (12%) experienced at least one unplanned hospital readmission within 30 days. Attending at least one GP consultation decreased the odds of readmission by half (adjusted OR: 0.5; 95%CI: 0.3–0.7), whereas attending at least one specialist consultation doubled those odds (aOR: 2.0; 95%CI: 1.2–3.3). GP consultations also reduced the odds of the combined risk of an ER visit or unplanned hospital readmission (aOR: 0.5; 95%CI: 0.3–0.7). ER visits were also associated with a higher readmission risk after adjusting for confounding factors (aOR: 10.0; 95%CI: 6.0–16.8).
Conclusion
GP consultations were associated with fewer ER visits and unplanned hospital readmissions.
] can be assessed directly during the hospital stay. However, physiological stresses experienced during hospitalization extend beyond discharge, resulting in a period of increased vulnerability [
General practitioners (GPs) are essential actors in healthcare transitions from hospital to ambulatory care. Misky et al. found an almost 90% reduction in 30-day readmissions for the same medical condition as the index hospitalization, after a timely follow-up consultation with a GP [
]. In a retrospective observational study, Jackson et al. found a significant reduction in readmissions among high-risk patients who consulted with a physician within seven days of discharge [
]. However, in prospective studies, the relationship between consultations with a GP and readmission rates has been inconsistent, with Field et al. finding no effects for consultations with a GP within seven days of hospital discharge [
]. Besides, a randomized trial involving intensive primary care interventions (close follow-up by a nurse and a GP) among patients suffering from diabetes, chronic obstructive pulmonary disease, or heart failure, showed even an increase in the number of 6-month readmissions compared to usual care [
]. Therefore, the relationship between post-discharge healthcare use and unplanned readmission after hospital discharge in medical patients remains unclear.
We aimed to evaluate the relationships between medical consultations after hospital discharge—dichotomized between GP and specialist consultations—and 30-day unplanned readmission risk. We also aimed to evaluate associations between other healthcare use (home visits by nurses, home support, and emergency room visits) and the 30-day unplanned readmission risk.
2. Methods
This study is part of phase 1 of the TARGET-READ study (Transition cAre intervention targeted to high-risk patiEnts To Reduce rEADmission; clinicaltrials.org NCT03496896). Patients were enrolled during a hospital stay, and their healthcare use and unplanned readmissions were monitored for 30 days.
2.1 Settings and participants
Between July 2017 and March 2018, all the adult patients admitted to general internal medicine units for 24 h or more and then discharged alive from four secondary and tertiary hospitals in Switzerland (Neuchâtel, Liestal, Bienne, and Fribourg) were consecutively included in the study. Patients previously enrolled in the study, admitted electively, living outside Switzerland, without a telephone, not speaking a national language, or unwilling or unable to give written informed consent were excluded. Each participating center's ethics committee approved the study protocol.
Everyone living in Switzerland have at least a standard health insurance that cover for all main healthcare services, including access to outpatient treatment by GP or specialists, emergency treatment and hospitalization. Patients are free to choose their own GP and may consult specialists without a referral from their GP. However, some patients may choose a cheaper insurance policy plan with the obligation to see first their GP before to be referred to a specialist.
In Switzerland, GP have no role in the care of hospitalized patients, which is assumed by physicians working solely at the hospital.
2.2 Outcomes and measurements
The primary outcome was 30-day unplanned hospital readmission. Secondary outcomes were the time to the first hospital readmission, cumulative hospital length of stay (LOS), and the number of emergency room (ER) visits.
Trained study nurses collected patients’ characteristics, demographic data, diagnoses at index hospitalization, discharge destination, and calculated their HOSPITAL score, which includes: hemoglobin level at discharge, discharge from an oncology unit or an active cancer diagnosis, sodium level at discharge, procedure during the index hospitalization, index admission type (urgent or emergent), number of hospital admissions during the previous year, and LOS ≥ 8 days [
Information on health care use (home visits by nurses, home support, number of medical consultations or ER visits) and unplanned hospital readmissions were collected using three planned, follow-up telephone calls at 2–4 days, 13–15 days, and 30 days after discharge and hospital chart screening. Death was recorded using registers of death, calls to GPs, and calls to next of kin. To limit information bias, study nurses used a standardized form to collect information from patients or, when needed, next of kin or GPs. When patients were readmitted to hospital, they and their medical professional (specialist or GP) were asked for their subjective feelings about whether the admission had been avoidable.
2.3 Statistics
The primary analysis—unplanned hospital readmissions associated with medical consultation—was made using logistic regression analysis. We repeated this separately for GPs and specialist physicians and for the combined outcome of ER visits and hospital readmissions.
The unadjusted impact of GP or specialist consultations on the time to a first hospital readmission was calculated using Kaplan–Meier survival analysis and an unweighted, two-sided, log-rank test to compare groups. The proportional hazards assumption was verified using Schoenfeld residuals and a visual inspection of the log-minus-log plots. The associations between medical consultations and cumulative LOS on readmission were tested using a linear regression model adjusted for confounding factors and in which LOS was log-transformed to correct for skewed data.
Patients with missing information on post-discharge medical information (3.5%) were excluded from the main analysis. The data was missing at random.
Associations between other healthcare use and 30-day readmission were analyzed using logistic regression analyses. All analyses were adjusted for age (continuous), HOSPITAL score (continuous), being of Swiss nationality (binary), and numbers of comorbidities (continuous).
Group characteristics were compared using the chi-squared test or Fisher's exact test, where appropriate, for categorical variables. The Mann–Whitney test was used for continuous variables, as these were not normally distributed. The significance level was set at 5%, and all analyses were performed using STATA statistical software, version 15.0 (StataCorp LP, College Station, TX, USA).
3. Results
Of 3239 patients screened, 934 were finally included in the study (Fig. 1). Within 30 days of discharge, 22 (2%) had died and 111 (12%) had experienced at least one unplanned hospital readmission. Participants’ baseline characteristics are shown in Table 1.
Table 1Characteristics of the entire patient population and their post-discharge medical consultations. Values are numbers (percentages) unless otherwise stated.
Total(N = 934)
30-day post-discharge medical consultations*
No consultations (n = 168)
≥ 1 GP consultation(n = 569)
≥ 1 specialist consultation(n = 164)
p-value
General
Age (years), median (IQR 25%–75%)
71 (58–80)
75 (62–82.5)
71 (59–80)
66 (54–75)
< 0.001
Male Female
509 (56%) 392 (44%)
96 (57%) 72 (43%)
315 (55%) 255 (45%)
98 (60%) 65 (40%)
0.53
Swiss nationality Other nationality
783 (87%) 118 (13%)
155 (92%) 13 (8%)
494 (87%) 76 (13%)
134 (82%) 29 (17%)
0.022†
Place of living Home Sheltered accommodation Nursing Home Other or unknown
882 (94) 11 (1.2) 35 (3.7) 6 (1%)
158 (94%) 2 (1%) 8 (5%) 0 (0%)
541 (95%) 7 (1%) 20 (4%) 2 (0%)
152 (93%) 2 (1%) 5 (3%) 4 (2%)
0.21†
Work Active Unemployment Receiving social or invalidity benefits Retired Other or unknown
206 (23%) 16 (2%) 55 (6%) 602 (67%) 22 (2%)
31 (18%) 2 (1%) 8 (5%) 122 (73%) 5 (3%)
123 (23%) 10 (2%) 31 (5%) 386 (68%) 13 (2%)
45 (28%) 4 (2%) 16 (10%) 94 (58%) 4 (2%)
0.15†
Health insurance None Standard Standard + Semi-private Private
* 33 participants had missing data (at random) regarding their post-discharge consultation. Therefore, the sum of the columns is 901 and not 934; † Fischer's exact test (instead of chi-squared test).
3.1 Primary care professionals and hospital readmission
During the 30-day post-discharge period, 569 (61%) and 164 (18%) patients consulted their GP or specialist, respectively, whereas 168 (18%) consulted neither (Table 2). Patients who consulted a physician within 30 days of discharge were younger, less frequently Swiss nationals, had higher rates of heart failure or active oncological disease, and had HOSPITAL scores significantly statistically different from those of patients who did not consult one (Table 1).
Table 2Association between healthcare use and hospitalization index score and risk of 30-day unplanned hospital readmission.
Unplanned 30-day hospital readmission
Yes (n = 111)
No(n = 813)
OR
Adjusted* OR
Medical professional
Patient had no medical professional
1
12
1.7 (0.2–12.9)
1.1 (0.1–8.9)
Medical consultation (any)
78
655
0.7 (0.4–1.2)
0.7 (0.4–1.1)
GP consultations
48
526
0.4 (0.3–0.7)
0.5 (0.3–0.7)
No. of GP consultations (vs. none) 1 2 or more
20 28
219 303
0.5 (0.3–0.8) 0.5 (0.3–0.8)
0.5 (0.3–0.9) 0.5 (0.3–0.8)
Specialist consultations
30
133
2.1 (1.3–3.3)
2.0 (1.2–3.3)
No. of specialist visits 1 2 or more
10 20
42 87
2.2 (1.1–4.6) 2.3 (1.3–3.9)
2.3 (1.1–4.8) 2.0 (1.1–3.5)
No medical consultation vs. GP consultation Specialist consultation
When primary care professionals were not dichotomized, univariate and adjusted analyses did not associate medical consultations with a greater risk of 30-day unplanned hospital readmission (Table 2). When these consultations were dichotomized between GPs and specialists, a GP consultation decreased the risk of readmission (OR 0.4, 95%CI: 0.3–0.7), whereas a specialist consultation increased it (OR 2.1, 95%CI: 1.3–3.3). These two associations persisted in adjusted analyses (Table 2).
Time to first hospital readmission was also associated with post-discharge GP consultations (HR: 0.47, 95%CI: 0.32–0.70; adjusted HR: 0.49, 95%CI: 0.33–0.74), and specialist consultations (HR: 2.0, 95%CI: 1.32–3.12; adjusted HR: 1.88, 95%CI: 1.21–2.93) (Fig. 2).
Fig. 2Time to 30-day hospital readmission associated with no consultation (solid line), or at least one post-discharge consultation with a GP (dash) or a specialist physician (dots). Log-rank test for survival difference p < 0.001.
GP consultations reduced the odds of an ER visit (adjusted OR: 0.62, 95%CI: 0.39–1.00) and the combined risk of an ER visit or unplanned readmission (adjusted OR: 0.51, 95%CI: 0.35–0.74). Specialist consultations, however, were associated with an increased adjusted combined risk of an ER visit or unplanned readmission (adjusted OR: 1.69, 95%CI: 1.10–2.62) (Table 3).
Table 3Secondary outcomes associated with primary care professional consultations.
Without consultation(n = 168)
With GP consultation(n = 569)
p-value
With specialist consultation(n = 164)
p-value
Unplanned hospital readmission
24 (14%)
48 (8%)
< 0.001
30 (18%)
0.002
Emergency room (ER) visits
20 (12%)
40 (7%)
0.011
19 (12%)
0.150
More than 1 visit to ER
12 (2%)
8 (1%)
-
4 (2%)
-
Hospital readmission or ER visit
34 (20%)
70 (12%)
< 0.001
39 (24%)
< 0.001
Cumulative hospital LOS of readmissions, median days (IQR 25%–75%)
8 (4.5–15)
6 (2–10)
0.11*
6 (4–9)
0.96*
ER: emergency room; GP: general practitioner; IQR: interquartile range; LOS: hospital length of stay.
There was no difference in the cumulative LOS for readmissions within 30 days between patients who had consulted their GP, those who had consulted their specialist, and those who had consulted neither (Table 3).
3.2 Other post-discharge healthcare use and readmission risks
Home visits by nurses, home support, and ER visits were associated with 30-day unplanned readmissions. However, only ER visits remained associated after adjustment for confounding factors (Table 2). Patients’ subjective feeling that their readmission to hospital had been avoidable was associated with a higher proportion of ER visits but not with their other uses of healthcare services (e.g., home visits by nurses) (Appendix Table A1).
Table A1Healthcare use by patients readmitted to hospital within 30 days of discharge and according to subjective impressions of hospitalization being avoidable according to the patient and their primary care professional.
Avoidable according to patient
P-value
Avoidable according to professional
P-value
no
yes
no
yes
Any medical consultation
34 (89%)
4 (11%)
0.357
30 (88%)
4 (12%)
0.132
Specialist consultation
24 (89%)
3 (11%)
0.742
17 (85%)
3 (15%)
0.740
Home visits by a nurse
24 (86%)
4 (14%)
0.975
21 (81%)
5 (19%)
0.760
Living with someone vs. living alone
48 (69%)
7 (58%)
0.518
41 (68%)
6 (46%)
0.200
Homecare support (any vs. none)
29 (41%)
4 (33%)
0.755
26 (43%)
3 (23%)
0.227
Emergency room (ER) visits
24 (34%)
9 (75%)
0.010
18 (30%)
6 (45%)
0.329
ER: emergency room; GP: general practitioner.
* median number of comorbidities = 1, ** median HOSPITAL score = 3.
In this multicenter prospective cohort, patients discharged from internal medicine units who consulted with their GP at least once soon afterwards were half as likely to be readmitted to hospital or to visit an ER within 30 days, compared to patients who did not consult their GP. In contrast, those who consulted a specialist physician soon after hospital discharge were twice as likely to be readmitted or to visit an ER in comparison to patients who did not visit their specialist.
To the best of our knowledge, associations between medical consultations and readmission risk have been inconsistently reported in the past. Only one randomized interventional study has shown the paradoxical increase in readmission risk following consultation [
]. The GP and specialist consultations in our study showed two opposite associations, neutralizing each other when analyzed together; thus, overall, consulting a physician was not associated with 30-day readmission risk. Variations in other studies might have come from the types of medical consultations included [
Our observed lower rate of readmissions has several possible explanations. First, patients consulting their GP are less prone to visiting ERs, which are a major stepping-stone to hospital admission. Interestingly, more readmitted patients with a ER visit within the 30 days than those without an ER visit felt that their readmission could have been avoided. Second, an early consultation with a GP could increase the potential for medication reconciliation, avoiding drug prescription errors, and identifying or treating adverse drug reactions after hospital discharge [
]. Third, the potential causative association between GP consultations and lower numbers of readmissions could, in fact, result from an inverse relationship. Readmitted patients may not have had the opportunity to consult their GP because they lacked time between the two hospitalizations. Also, readmitted patients or patients presenting at ERs could be too sick to benefit from care from their GP and may have been readmitted regardless of a consultation. To mitigate this effect, we adjusted for age, comorbidities, and HOSPITAL score (a score that has been associated with readmission and mortality risks). However, an interventional study randomly assigning patients at risk of readmission to a consultation with their GP (or not) would be required to definitively appreciate whether this observed relationship was causative or not.
The observed opposite relationship between specialist consultations and readmissions agreed with previous reports [
]. On the one hand, this inverse effect could be due to the different types of patients consulting GPs and specialists. Many end-organ failures (e.g., heart, lung, and kidney) deserving specialized care repeatedly decompensate and need frequent hospital admission [
]. Furthermore, the patients in our study consulting specialists had more oncological diseases, and oncological diseases and their treatment are known to increase admission risk [
]. On the other hand, healthcare systems oriented towards clearly separated medical specialties could lack the more general approach needed following hospital discharge (medication reconciliation, avoiding drug errors, and identifying or treating adverse drug reactions) [
]. Besides, the lack of a statistically significant difference between patients consulting specialists or not consulting a physician at all could suggest that the increased risk of readmission was mainly driven by not consulting a GP. However, this analysis was of limited statistical power.
The findings present some limitations. First, as an observational study, associations may result from unconsidered confounding factors. Second, GPs and specialists were dichotomized and mutually exclusive in this study. Indeed, patients were asked to state which type of medical professional was most involved in their post-discharge care, but both a GP and a specialist might share this responsibility. Not considering the possibility of dual care might have favored the positive results with GPs. Nevertheless, for oncological diseases, which represented a substantial proportion of our study's patients under specialized care, patients were often treated exclusively by specialists, with GPs taking over responsibility after therapy has ended. However, within our 30-day post-discharge window, few patients attended more than one consultation with a physician, thus reducing the chances of having consulted a GP and a specialist.
In conclusion, consultations with a GP were associated with fewer ER visits and unplanned readmissions, whereas consultations with a specialist physician were associated with a higher risk of these outcomes. Interventional studies are needed to explore the nature of this association in more detail.
Declaration of Competing Interest
The authors declare they have no conflict of interest
Funding
Jacques Donzé received funds from the Swiss National Science Foundation as part of the TARGET-READ study (PP00P3_170656).
References
Jencks S.F.
Williams M.V.
Coleman E.A.
Rehospitalizations among patients in the medicare fee-for-service program.
In the United States the Hospital Readmissions Reduction Program (HRRP) is a Medicare value-based purchasing program that encourages hospitals to improve communication and care coordination to better engage patients and caregivers in discharge plans and, in turn, reduce avoidable readmissions [1]. The program supports the national goal of improving health care for Americans by linking payment to the quality of hospital care. The Centers for Medicare & Medicaid Service (CMS) includes the following six condition or procedure-specific 30-day risk-standardized unplanned readmission measures in the program: Acute Myocardial Infarction, Chronic Obstructive Pulmonary Disease, Heart Failure, Pneumonia, Coronary Artery Bypass Graft Surgery, Elective Primary Hip and/or Total Knee Arthroplasty.
00014-0&id=gr1.jpg){kind=link}
00014-0&id=gr2.jpg){kind=link}