End stage liver disease (ESLD) is known to be a life limiting illness whose only cure is transplant. Patients who do not undergo transplant will die from their disease, and those awaiting transplant are also at high risk of death. Prior to death, patients often suffer from poor health-related quality of life. Data from the SUPPORT study (Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment) showed that the burden of pain for patients with ESLD was similar to patients with cancer, and communication challenges were prevalent.
Until recently, no quality indicators for high quality palliative care, and few guidelines for palliative care, have been available for patients with ESLD. Supportive care process indicators for ESLD that cover two domains (Information and Care Planning, and Symptom Management) have recently been developed by a VA-based research team including leaders in hepatology and gastroenterology, using the RAND/UCLA appropriateness method to provide tools to assess the extent to which care for patients with ESLD address the symptomatic and informational needs of patients and families. This pilot project aimed to:
1) Pilot test these quality indicators in patients at the West Los Angeles VA to evaluate sampling strategies for a national study and
2) Study the feasibility and reliability of these quality indicators among Veterans with ESLD.
We reviewed charts for Veterans with at least 1 ICD-9 code for decompensated cirrhosis in 2012 identified from the VA CDW (Corporate Data Warehouse) who had at least 2 outpatient visits or one hospitalization at the West Los Angeles VA. We selected three sampling strategies based on previous research to evaluate their sensitivity and specificity compared to the gold standard of a screening chart abstraction tool to capture Veterans with and initial diagnosis of advanced end stage liver disease (defined as an initial instance of a MELD score of 20 or greater or a CTP score of 12 or greater between 2011-2013).
We also developed a full chart abstraction tool (with guidelines and training materials) to measure the 19 palliative care quality indicators for ESLD. We evaluated the feasibility and inter-rater reliability of these quality indicators, as well as their validity using physician implicit review to streamline the indicator set for a national study.
Both the screening chart abstraction tool and the full chart abstraction tool were implemented using trained nurse abstractors and clear written guidelines that were updated iteratively as questions arose during regular team meetings.
We identified 167 patients in CDW that met one of our three sampling strategies (1-at least one decompensated cirrhosis code, 2-at least one decompensated cirrhosis code during an inpatient admission, 3-at least two ICD-9 decompensated cirrhosis codes) in 2012. Seventy one of these patients were identified to be true positives using the gold standard screening chart abstraction tool. Nine of these patients died within 30 days of diagnosis and sixty-two patients were in the final sample for full abstraction. Seventeen patients were selected from the 167 patients to calculate inter-rater agreement on final "screen in" status and we had 100% agreement. Sampling strategy 1 had the highest sensitivity at 87% and the lowest specificity at 22%. Sampling strategy 2 had the highest specificity at 59% and a sensitivity of 62%.
Regarding full abstraction of the quality indicators:
o17/19 were triggered at least twice with a denominator range of 2-62
oInter-rater reliability was good to excellent with denominator kappa scores ranging from 0.75-1 and numerator kappa scores ranging from 0.56 to 1
oKey areas identified in need of further research included:
Only 45% of patients with advanced end stage liver disease in this population had evidence of receiving palliative care and/or hospice prior to death
Only 55% of patients with advanced end stage liver disease considered for transplant and 58% of those with hepatic encephalopathy were offered advance care planning in a timely fashion.
Only 32% of patients with advanced end stage liver disease admitted to the ICU had timely goals of care discussions
Only 19% of patients with a DNR order at time of discharge from the hospital had an appropriate physician orders for life sustaining treatment form completed
39% of patients with advanced end stage liver disease were prescribed NSAIDS that should be avoided due to increased risk of renal toxicity and bleeding
This project has informed a final set of quality indicators and a sampling strategy for a planned IIR HSR&D proposal that will evaluate the quality of palliative care provided to Veterans with end stage liver disease using these newly developed process measures to identify practice gaps and guide improvements in clinical care. Findings from this pilot study identified important areas for improvement in advance care planning and palliative care at the West Los Angeles VA. A national study is needed to study if these patterns exist more broadly to inform VA strategies to improve care for Veterans with this increasingly prevalent disease.
- Walling AM, Ahluwalia SC, Wenger NS, Booth M, Roth CP, Lorenz K, Kanwal F, Dy S, Asch SM, Palliative Care Cirrhosis Quality Expert Panel. Palliative Care Quality Indicators for Patients with End-Stage Liver Disease Due to Cirrhosis. Digestive diseases and sciences. 2017 Jan 1; 62(1):84-92.
- Walling AM. Nuances of integrating palliative care and end-stage liver disease. [Editorial]. Palliative Medicine. 2018 May 1; 32(5):906-907.