Identifying Quality Gaps in Preventive Care for Outpatients With Cirrhosis Within a Large, Academic Health Care System

Cirrhosis is a common condition associated with significant morbidity and mortality.^{(1, 2)} It is the twelfth leading cause of death in the United States⁽³⁾ and carries a burden of disease that is projected to increase over the next decade due to a rising incidence of alcohol-associated liver disease and nonalcoholic steatohepatitis (NASH).^{(4, 5)} Cirrhosis is also associated with significantly decreased quality of life and high health care use.^{(6, 7)} Compared to patients with congestive heart failure, patients with cirrhosis experience higher rates of hospitalizations, longer hospital stays, and more readmissions, resulting in substantial and increasing morbidity and economic burden.^(8-10)

Clinical guidelines for managing and preventing complications of cirrhosis are well-established, and several process measures including hepatocellular carcinoma (HCC) surveillance and antibiotic prophylaxis for ascites have been linked to improved survival.^(11-13) Unfortunately, adherence to many of these practices remains poor.^(14-18) Pooled data from national cohorts of patients with cirrhosis have shown that less than 20% receive appropriate HCC surveillance.^{(19, 20)} In a retrospective study in the Veterans Affairs (VA) health care system, only 24% of patients with cirrhosis received an upper endoscopy within a year of diagnosis.⁽²¹⁾

Most prior quality improvement efforts in cirrhosis have focused more narrowly on improving only a few process measures in an effort to reduce health care use and improve patient-reported outcomes.^(22-24) However, to guide stakeholders into promoting and prioritizing quality improvement efforts within a health system for patients with cirrhosis, we must first more broadly evaluate and identify quality gaps. These data can then lay the groundwork for future studies, identifying barriers and facilitators to achieve higher rates of important process measures with the goal of developing interventions to improve cirrhosis quality of care (QOC).

In this study, we aimed to evaluate the receipt of evidence-based preventive quality indicators (QIs) provided to outpatients with cirrhosis in our health system, a high-volume, academic tertiary health care system, and to determine factors associated with higher QOC.

Methods

Results

Discussion

In this retrospective cohort study measuring the QOC provided to an ethnically diverse cohort of outpatients at a high-volume tertiary center, we found both low uptake as well as wide variation in the receipt of several QIs. In our sample, only 2% of patients received appropriate secondary prophylaxis for variceal bleeding, yet 83% of patients with alcohol-associated cirrhosis received counseling to abstain from alcohol consumption. Despite 93% of patients in our sample receiving care by a gastroenterologist or hepatologist, there was suboptimal adherence for several QIs, including SBP prophylaxis, primary and secondary prevention of variceal bleeding, and surveillance for HCC. Finally, we found few patient-level factors associated with higher QOC.

Low receipt of QIs in our study mirrored the results from investigations that have measured individual QIs in the literature. Results from population-based cohorts and meta-analyses of single center studies have demonstrated low rates of HCC surveillance in patients with cirrhosis.^{(30, 31)} One study found that patients were up to date with surveillance 23% of the time⁽³⁰⁾; another showed a 52% overall adherence rate to appropriate surveillance guidelines.⁽³¹⁾ In our study, only 24% of patients with cirrhosis received appropriate surveillance. Data from the VA have also reported major gaps in adherence to recommendations for ascites-related care, as well as varices screening and prophylaxis. In a cohort of 774 veterans with cirrhosis and ascites, only 30% of patients received recommended antibiotics for secondary prophylaxis after having a prior documented episode of SBP.⁽³²⁾ This is in line with our data, in which only 21% of patients with a documented episode of SBP received long-term outpatient antibiotics for secondary prophylaxis. Similarly, initial screening rates for varices in patients with a new diagnosis of compensated cirrhosis remained low at 32% in our cohort, paralleling a VA cohort in which 24% of patients received an upper endoscopy within a year of their cirrhosis diagnosis.⁽²¹⁾ QI adherence for secondary prevention of variceal bleeding was even lower; notably, among patients who had prior variceal bleeding, only 2% had variceal band ligation performed every 4 weeks until obliteration, a beta-blocker initiated within 4 weeks of bleeding, or a combination of the two.

Although more data is needed regarding the patient-level and provider-level factors associated with lower QOC, one possible explanation for the consistently low adherence to cirrhosis QIs is that general cirrhosis quality measures are not tied to Medicare reimbursement or to most hospitals’ compensation or incentive programs, unlike those for heart failure or hepatitis C virus screening and treatment. Our data add to a growing body of evidence showing that major gaps in QOC exist across a variety of different practice settings. To bridge these gaps, we recommend that the liver community not only explore the root causes of these variabilities in QI adherence in order to design meaningful interventions, but also use the electronic health record to track changes in quality measures.

Next, we were interested in identifying patient-level and provider-level factors that may be associated with higher value care. We found few variables that were associated with high QOC, which may be due to underpowering, as many individual quality measures had a small number of eligible patients. Interestingly, we found that receiving more visits with specialists within our health system was not associated with higher quality, which diverges from other studies showing greater adherence rates for HCC surveillance, ascites care, and varices surveillance with access to specialty care.^{(19, 21, 32)} One possible explanation for this difference is that our study may not have been powered to show an association, as only 8% of patients (n = 35) were not seen by a specialist. We did find that non-Hispanic Asian race was associated with higher rates of HCC surveillance as well as higher value care overall, compared with non-Hispanic white patients. The explanation for this association is not entirely clear, although one possible reason is that a greater proportion of these patients had chronic hepatitis B (31% vs. 2% in non-Hispanic whites in our cohort), which has been associated with higher rates of surveillance in prior studies.⁽³³⁾ Future studies are needed to examine the potential interaction between race and other factors, such as cirrhosis etiology, in impacting QOC. Decompensated cirrhosis was also associated with higher HCC surveillance rates, with a 2.1 greater odds of receiving appropriate and timely abdominal imaging even after controlling for the number of specialty visits and referral for liver transplant evaluation. This is in contrast to a prior study showing that the presence of comorbid conditions was associated with lower likelihood of receiving routine HCC surveillance.⁽¹⁹⁾ One potential explanation for our finding is that sicker, decompensated patients may receive more intensive outpatient or inpatient care due to the inherent complexity of their illness.

The information from this study will be presented to key stakeholders (including clinicians, staff, quality officers, and patients) within our health system, to determine the process measures to prioritize in future quality improvement efforts. After achieving consensus regarding the QI measures to target initially, potentially with modified Delphi methods,⁽³⁴⁾ our next goal will be to identify specific barriers and facilitators to each process of care before developing and testing interventions. We believe that similar gap analyses performed by health systems will be helpful to create interventions and can positively and sustainably affect health outcomes in patients with cirrhosis at a population level.⁽³⁵⁾ Recently, more updated guidelines on cirrhosis quality measures were established by the Practice Metrics Committee of the American Association for the Study of Liver Diseases.⁽³⁶⁾ The main differences between the new guidelines and prior established guidelines⁽²⁸⁾ are the use of more updated recommendations for cirrhosis management as well as the inclusion of outcome measures, including both clinical and patient-reported, as new quality metrics. Although these new guidelines were not relevant to our cohort, which received care before these new quality measures were published, it is important that future quality improvement studies use these updated recommendations. Furthermore, longitudinal studies are needed to evaluate the impact of patient nonadherence on QOC.

We acknowledge a few limitations to our study. First, we could not account for care that was provided outside of our health system. However, we mitigated this problem by only including patients who were receiving their primary care at UCLA and by including all outside information that was available within the electronic health record. We also acknowledge that we may have missed patients who carried a diagnosis of cirrhosis by using the ICD coding system; however, the ICD-9 and ICD-10 systems have been previously well-validated as accurately identifying cirrhosis and cirrhosis-related complications and have formed the basis for many clinical, epidemiologic, and health services research studies. Next, lower QOC scores may reflect patient preference, which should be factored in how these results are interpreted. We hope that future studies exploring barriers and facilitators to specific practices will help elucidate this further. We also could not account for the potential lack of documentation, which may have led to underestimation of QI rates. For example, providers may have provided verbal counseling for driving to patients with hepatic encephalopathy, but this may not have been documented in the electronic medical record. However, only a few process measures are susceptible to underdocumentation, as many others are easily captured by objective laboratory or imaging data. Additionally, we chose 75% or more QIs achieved to define higher QOC, although we acknowledge that this is not a standard, defined endpoint. However, we also explored other cutoffs, including receipt of 50% or more QIs and of 75% or more of class I QIs and found similar patterns in multivariable regression analyses. Next, time of follow-up was variable within our cohort, although most of the QIs we evaluated were not time-dependent measures. Finally, our study took place at a single, tertiary care center, where most patients were seen by a specialist; thus, our findings may be less generalizable to other centers or the larger populations of U.S. patients with cirrhosis, many of whom are managed by primary care providers.

There were several strengths of our study. First, our cohort consists of an ethnically diverse population of patients with cirrhosis that is representative of people living with chronic liver disease across the United States. Second, we used previously validated definitions of cirrhosis in our inclusion criteria and additionally performed manual chart review to confirm a cirrhosis diagnosis. Our data also had the granularity and detail necessary to examine a range of patient-level variables that might affect adherence to QIs, allowing us to identify variations in care by race/ethnicity and illness trajectory. Finally, our study examined a wide range of preventive QIs using a population of non-VA patients receiving care at a large academic health system with excellent access to specialty care services. Ultimately, this study provides data on adherence rates to cirrhosis quality measures in a practice setting, and lays the groundwork for future studies examining more detailed patient, provider, and system-level factors that might inform interventions and impact QOC in this complex patient population.

In conclusion, we examined the quality of outpatient preventive care in a large, ethnically diverse cohort of outpatients with cirrhosis at a high-volume tertiary liver transplant center. Receipt of evidence-based QIs was variable and overall low. These findings suggest that larger system redesign efforts are needed to improve the QOC, beyond just improving access to specialty providers. We believe that our approach to broadly examine gaps to high-quality care is crucial for organizing an approach on how to lead quality improvement efforts that can be replicated by any health system. Future studies are needed to identify modifiable factors to develop appropriate, targeted interventions that standardize care delivery, reduce practice variability, and improve the QOC among patients with cirrhosis.