Novel, nurse-led early postdischarge clinic is associated with fewer readmissions and lower mortality following hospitalisation with decompensated cirrhosis

Abstract

Objective

Patients hospitalised with decompensated cirrhosis have high rates of early unplanned readmission. Many readmissions are avoidable with secondary preventative strategies, but patients are often readmitted prior to outpatient review. To address this, we established a novel, nurse-led early postdischarge (EPD) clinic delivering goal-directed care for cirrhosis complications and evaluated the impact.

Methods

Retrospective cohort study comparing outcomes in 78 patients seen in the EPD clinic with 91 phenotypically matched controls receiving standard, consultant hepatologist care. Follow-up for 12 months from index admission with endpoints including survival, time to readmission, number of readmissions and healthcare burden.

Results

Median time to readmission was 51 days in controls and 98 days in the intervention group (p<0.01). The intervention cohort had significantly fewer readmissions at 30 days (12% vs 30%, p<0.01) and 90 days (27% vs 49%, p<0.01) but not significantly at 12 months (58% vs 68%, p=0.16) with an overall reduction in bed day usage of 29%. Mortality for the control group was 4% at 30 days with no deaths in the intervention group. There were significantly fewer deaths in the intervention group at 90 days (5% vs 15%, p<0.05) and 12 months (22% vs 41%, p<0.01).

Conclusions

Following an index hospitalisation with decompensated cirrhosis, goal-directed postdischarge care can be effectively delivered by specialist nurses, prior to outpatient review by hepatologists. This model was associated with significantly fewer readmissions, lower bed day usage and a reduced mortality. Our data suggest such models of care deserve wider implementation and further evaluation.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Patients admitted to hospital with decompensated cirrhosis have high rates of unplanned readmission within 30 days of discharge. The majority of these admissions are due to complications such as ascites or hepatic encephalopathy, which are potentially avoidable if medical treatment is optimised. Unfortunately, many patients are readmitted before their planned outpatient review with a consultant hepatologist.

WHAT THIS STUDY ADDS

• We demonstrated that care provided by specialist nurses working in a dedicated early postdischarge clinic was associated with significantly fewer readmissions and a reduced mortality rate.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• A nurse-led early postdischarge clinic is an effective way of delivering goal-directed care for the complications of cirrhosis and can achieve improved outcomes. We suggest such a model of care should be considered for evaluation in other disease settings.

Background

Recent decades have seen a rise in the incidence of chronic liver disease in the UK with a consequent increase in liver-related hospital admissions and deaths.¹ Importantly, patients admitted with decompensated cirrhosis have high rates of early, unplanned readmission with associated morbidity, mortality and increased healthcare burden, with over a third readmitted within 1 month.²

The most frequent causes of readmission are recurrent ascites or hepatic encephalopathy (HE)^{3 4} and many are potentially avoidable by optimising secondary preventative therapies.^{5 6} Unplanned readmissions may indicate wider deficiencies in coordinating care for people with cirrhosis, effectively a ‘canary in the coalmine’ for the quality of hepatology services.⁷

At our hospital, there are often long waiting times for consultant outpatient appointments. Consequently, patients admitted with decompensated cirrhosis may be discharged with plans for early outpatient review but are readmitted to hospital before their consultant appointment.

To break this cycle and deliver more effective secondary prevention, we established an early postdischarge (EPD) clinic provided by specialist hepatology nurses delivering goal-directed care within a maximum of 2 weeks from discharge. The EPD clinic offers holistic care, with a range of interventions (box 1), for patients discharged with decompensated cirrhosis including treatment of ascites, HE, portal hypertension and alcohol misuse.

Box 1. Interventions offered by the nurse-led decompensated cirrhosis early postdischarge clinic:

• Patients seen within 1–2 weeks of discharge following admission with complication of cirrhosis.

• Two senior nurse prescribers drawn from wider team of five hepatology nurse specialists.

• Goal-directed care of ascites, hepatic encephalopathy, portal hypertension and alcohol misuse.

• Review of medications, including dose titrations and compliance.

• Blood test monitoring.

• Ensure appropriate secondary prevention for varices, hepatic encephalopathy and spontaneous bacterial peritonitis.

• Nurse-delivered, daycase paracentesis for ascites where required.

• Active management of alcohol or illicit substance misuse.

• Frailty and nutritional assessments with dietician support.

• Link with community palliative care services where required.

• Weekly multidisciplinary meeting for case discussions with consultant hepatologists.

We hypothesised that the nurse-led EPD clinic would improve care and reduce readmissions. The primary aim of our study was to evaluate the impact of the clinic on readmission and mortality rates at 30 days, 90 days and 12 months. Additional endpoints included the number of readmissions, inpatient bed days and emergency department (ED) attendances.

Methods

In this retrospective cohort study, patients with cirrhosis and an index hospital admission with decompensation (ascites, HE, jaundice, spontaneous bacterial peritonitis and/or variceal haemorrhage) were identified using a departmental database and hospital coding. Diagnosis of decompensated cirrhosis was confirmed by clinical record review.

Cases were classified into an intervention and control cohort based on their discharge follow-up. The intervention cohort was seen in the nurse-led EPD clinic. In cases with multiple admissions, the earliest episode was selected as the index admission. Controls were patients receiving standard consultant-led outpatient follow-up prior to the introduction of the EPD clinic in 2017. The groups were phenotypically matched by age, gender, disease aetiology and severity. Follow-up period from index admission was 12 months for both cohorts.

The study design is summarised in figure 1.

Figure 1.

Overview of study design.

Patients in the nurse-led clinic were seen face-to-face at each visit by one of two senior hepatology nurse specialists (NHS Agenda for Change Band 7). The consultations took place in a dedicated clinical area with access to a procedural suite for paracentesis if required. The nurses’ input included nutritional advice and prescription of supplements and a specialist dietician could also be called if required by the nurses.

The nurse prescribers take a ‘goal-directed’ approach such as titration of lactulose versus stool frequency or adjustment of diuretics against ascites. Management steps for complications of cirrhosis are based on published guidelines, including the British Society of Gastroenterology guidelines on ascites,⁸ local protocols and a formal Local Safety Standard for Invasive Procedures for people undergoing paracentesis.

Data collected included demographics, liver disease aetiology, number of comorbidities, index admission details (including length of stay, reason for admission, medications at discharge, presence of infection or acute kidney injury (AKI)), Model for End-stage Liver Disease-sodium (MELD-Na), platelet count and serum albumin at discharge.

We analysed the time taken to first unplanned readmission and time to death within the 12-month follow-up. Readmission and mortality rates were analysed at 30 days, 90 days and 12 months following index admission. The impact of healthcare burden was based on ED attendances, total number of inpatient bed days and number of times seen in the nurse-led clinic within 12 months of the index hospitalisation.

Statistical analysis was performed using Wilcoxon Rank-sum test for continuous variables and χ² or Fisher’s exact test for categorical data. A two-tailed p value <0.05 was considered statistically significant.

The study was registered and approved by Portsmouth Hospitals University NHS Trust clinical audit department, audit number 4888.

Results

There were 78 patients in the intervention group and 91 patients with standard consultant-led follow-up in the control group. Baseline characteristics at index admission are summarised in table 1. Demographics were similar for both groups, with a median age of 60 years old in the control group and 58 years old in the intervention group, with most patients being men (65% in control group, 55% in intervention group). The most frequent cause of cirrhosis in both groups (>85%) was alcohol. All other characteristics (number of comorbidities, MELD-Na score, presence of hepatocellular carcinoma, AKI, infection, number of medications, serum albumin and platelet count) at index admission were similar between both groups with no statistical differences.

Table 1.

Comparison of results for control (standard follow-up) and intervention group (nurse-led early post-discharge clinic)

	Control group (n=91)	Intervention group (n=78)	P value
Baseline characteristics
Age, median (years; range)	60 (31–85)	58 (34–87)	0.238
Male gender, n (%)	59 (65%)	43 (55%)	0.198
Aetiology of liver disease, n (%):
Alcohol	78 (86%)	66 (85%)	0.216
Non-alcoholic steatohepatitis	14 (15%)	6 (8%)
Hepatitis C	3 (3%)	3 (4%)
Other	1 (1%)	4 (5%)
Unknown	1 (1%)	3 (4%)
Number of comorbidities, median (IQR)	2 (1–3)	1 (1–2)	0.126
MELD-Na score at discharge, median (IQR)	19 (13–22)	20 (17–23)	0.072
Hepatocellular carcinoma, n (%)	8 (9%)	3 (4%)	0.194
Acute kidney injury during index admission, n (%)	28 (30%)	21 (27%)	0.679
Infection during index admission, n (%)	23 (25%)	26 (33%)	0.25
Number of medications at discharge, median (IQR)	8 (6–11)	7 (5–9)	0.194
Serum albumin (g/L) at discharge, median (IQR)	27 (23–30)	27 (23–30)	0.984
Platelet count (×109/L) at discharge, median (IQR)	128 (90–171)	134 (87–208)	0.465
Readmission
Time to first readmission, median (IQR)	51 days (13–93)	98 days (33–168)	0.003*
First readmission within:
30 days	27 (30%)	9 (12%)	0.004*
90 days	45 (49%)	21 (27%)	0.003*
12 months	62 (68%)	45 (58%)	0.16
Reasons for readmission, n (%):			0.207
Ascites	30 (48%)	29 (62%)
Hepatic encephalopathy	19 (31%)	10 (22%)
Portal hypertensive bleed	12 (19%)	10 (22%)
Spontaneous bacterial peritonitis	1 (2%)	2 (4%)
Mortality, n (%)
30 days	4 (4%)	0 (0%)	0.125
90 days	14 (15%)	4 (5%)	0.031*
12 months	37 (41%)	17 (22%)	0.009*

*Statistically significant result.

MELD-Na, Model for End-stage Liver Disease-sodium.

The intervention group had a significantly longer time to first readmission (98 days) versus controls (51 days). There was a significant reduction in readmission rates in the intervention group at 30 days (12% vs 30%) and 90 days (27% vs 49%) but not at 12 months (58% vs 68%). Readmission rates are displayed in table 1 and figure 2. There was no significant difference in the cause of readmission although ascites, or complications relating to ascites, was the most common reason for readmission in both groups.

Figure 2.

Comparison of time taken to first readmission within 12 months of index admission.

Survival (see table 1 and figure 3) was improved in the intervention group compared with controls. At 30 days following discharge, 4% of patients had died in the control group with no deaths in the intervention group. Mortality was lower at 90 days (5% vs 15%) and 12 months (22% vs 41%) in the intervention group, with both achieving statistical significance.

Figure 3.

Comparison of probability of survival in patients receiving EPD nurse-led follow up in the intervention group and standard follow-up in control group. EPD, early postdischarge.

Table 2 summarises the differences in healthcare usage between the groups. Unplanned emergency healthcare usage was lower in the intervention group compared with the control group across all measures but only reached statistical significance for the reduction in ED attendance. Patients were seen a median of 8 times in the nurse-led EPD clinic within 12 months of admission.

Table 2.

Comparison of hospital episodes

	Control group	Intervention group	P value
Index admission length of stay, median (IQR)	10 days (5–19)	8 days (5–16)	0.12
Number of readmissions within 12 months, median (IQR)	1 (0–3)	1 (0–3)	0.25
Number of bed days within 12 months, median (IQR)	7 days (0–29)	5 days (0–21)	0.11
Number of emergency department attendances within 12 months, median (IQR)	1 (0–3)	0 (0–2)	0.02*
Number of visits to nurse-led clinic within 12 months, median (IQR)	NA	8 (4-12)	NA

Bold indicates P value of <0.05.

*=P value of <0.05.

We did not find a significant difference in attendance rates between clinics. The nurse-led EPD clinic had a ‘did not attend’ (DNA) rate of 6.7%, which was broadly similar to DNA rates in consultant delivered clinics (which ranged from 5.3% to 8.3%).

Discussion

Many cirrhosis complications can be reduced by secondary preventative approaches.⁹ Despite this, patients admitted to hospital with decompensated cirrhosis have high rates of early, unplanned readmission, which may reflect failures to coordinate or deliver effective care.

Our data demonstrate that, following an index hospital admission, goal-directed care can be effectively delivered by specialist nurses with significant reductions in unplanned readmissions and improved survival compared with a purely consultant-led service.

Published data show the presence of HE, ascites or ongoing alcohol misuse are major predictors of unplanned readmission within 30 days of discharge.⁴ However, optimising management of HE, for example, through use of rifaximin in appropriate patients, leads to fewer readmissions and ED attendances¹⁰ and may correlate with improved mortality compared with historical cohorts.¹¹

Patients with ascites have markedly reduced quality of life¹² and their condition is often suboptimally managed, leading to recurrent admissions.¹³ Similarly, a recent study of over 3500 patients with a first variceal haemorrhage, showed only a third received a non-selective beta blocker and repeat endoscopy within 120 days of admission.¹⁴

One way of reducing readmissions is to ensure patients are discharged on effective secondary prevention, either by using prompts to physicians completing discharge summaries⁵ or through use of a specific ‘discharge care bundle’.¹⁵

Unfortunately, delays in subsequent outpatient clinic visits due to long waiting times may mean patient care is suboptimal or poorly coordinated, leading to early readmission. Two recent studies have shown the benefit of a hepatologist-led EPD follow-up model for patients admitted with decompensated cirrhosis, demonstrating reductions in readmission and mortality rates with fewer ED attendances.^{16 17}

However, given the sometimes long waiting times for consultant appointments, the purpose of our study was to show whether a nurse-led service could achieve similar results.

Our data show a clear positive impact of the nurse-led EPD clinic on readmission rates and mortality compared with standard consultant-led outpatient follow-up. The time to first readmission was 47 days longer for the intervention group compared with the controls. This led to a significant reduction in early (<30 days) and medium-term (<90 days) readmission rates. However, readmission rates were not statistically different at 12 months, possibly relating to disease progression in this high-risk patient group. The EPD clinic model was also associated with improved mortality rates with no deaths seen within 30 days of admission and significantly lower mortality at 90 days and 12 months.

We also demonstrated lower unplanned healthcare utilisation in the intervention cohort. The number of readmissions and ED attendance within the 12-month study period were all lower in the intervention group, although only the number of ED visits achieved statistical significance. These markers represent a reduced healthcare burden and potentially reduced costs. The median number of times a patient was seen in the EPD clinic was eight times over the following 12-month study period. Although there is a nursing resource cost of staffing the service, the interventions offered are relatively inexpensive compared with the cost of emergency admissions and we would anticipate a cost saving, although we did not conduct a formal health economic analysis.

However, the benefits of a physician-led decompensated cirrhosis early follow-up clinic on outcomes and healthcare costs were demonstrated by Morales et al with their day hospital model (HEPACONTROL).¹⁷ This led to improved early (<30 days) readmission rates, lower 60-day mortality and reduced net healthcare costs. There were several differences in the baseline characteristics of patients in the Morales study compared with our study. The most notable were a lower average MELD-Na score (15 compared with 19–20 across both groups in our study) and aetiology of liver disease, with higher rates of hepatitis C (40% vs 4%) and lower alcohol-related liver disease (40% vs 85%) compared with our patient cohort. Another key difference was the set up and provision of their service. Patients visited the physician-led HEPACONTROL clinic within 7 days of discharge and were seen an average of 4.3 times during the 10-month follow-up period. The median time from discharge to first review in our nurse-led EPD clinic was 11 days (IQR 8–19 days) with patients attending the service an average of 8 times over the following 12 months. Despite our higher baseline MELD-Na scores, we saw similar improvements in 30 and 90-day readmission rates and, more importantly, improved mortality up to 12 months after index admission. A key advantage of our nurse-led model is the reduced staff cost compared with physician-delivered care and, therefore, the ability to offer more intensive follow-up. This may be one of the factors that led to the more sustained improvement in mortality seen in our study.

With over 85% of patients having alcohol excess as an aetiology of their cirrhosis, we evaluated the impact of ongoing alcohol consumption as a possible confounder of our study. However, ongoing postdischarge alcohol misuse was documented in 30% of patients in the intervention group compared with 29% in the control group, with no statistical significance.

We examined other factors, which may have impacted on the differences seen in our study. For example, it could be argued a benefit of the nurse-led care was achieving an early outpatient review and it might be possible to achieve similar results by delivering earlier appointments in the consultant clinic. The median time from discharge to first consultant-led outpatient appointment was 68 days in the control group and 58 days for the intervention group, which was not statistically different. An argument can be made that improved care of patients could be made by improving consultant clinic waiting times. However, the 2–3-month lag in outpatient follow-up seen in our study is not uncommon in hospitals around the UK and there is a need for more hepatologists.¹ This is not a problem that can be easily solved and nurse-led services may offer a faster solution, enabling patients to be seen within 2 weeks.

The nurses working in the EPD clinic have undertaken a physical examination course, are qualified prescribers and are trained in performing large-volume paracentesis. However, it may also be feasible for such care to be delivered by other suitably trained healthcare professionals such as physician associates.

Our study has some potential limitations. First, the retrospective nature of this controlled study requires accurate coding and classification of admissions, although we used a departmental database of admissions and there were no changes in coding practice during the study period. Second, it may be that regular interaction with healthcare services, such as the specialist nurses, could have increased a patient’s interest in their own health and well-being. However, this behaviour change is one of the aims of the service and by observing the similar baseline characteristics between our intervention and control cohorts and setting objective study variables, we have attempted to reduce these limitations.

Third, we did not record patients’ levels of social support or socioeconomic deprivation, which may be contributory factors to unplanned hospital admissions, particularly as cirrhosis disproportionately affects the deprived and marginalised.¹

In conclusion, we have demonstrated that following an index hospitalisation with decompensated cirrhosis, a novel nurse-led, EPD clinic is associated with improved patient outcomes, including fewer readmissions, fewer ED attendances and improved mortality rates. We have shown that goal-directed care can be effectively delivered by specialist nurses prior to outpatient review by consultant hepatologists. Our data suggest such models of care deserve wider implementation and further evaluation in other settings.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study involves human participants and was approved by Audit number 4888, approved and registered with the Portsmouth Hospitals University NHS Trust clinical audit department. As a registered clinical audit, this study did not require ethical approval. This was an approved clinical audit and does not include any data on identifiable individuals. As an approved clinical audit, individual patient consent was not required.