Big Data Model Predicts Rise in Alcohol-Related Liver Disease

A big data model projected that without any changes in alcohol consumption or interventions to address high-risk drinking, deaths from alcohol-related liver disease will rise significantly in the US, according to a study published in the Lancet Public Health.

Alcohol-related liver disease is currently the most common reason for liver transplant in the US. In the past several years, high-risk drinking of alcohol has increased in nearly all sociodemographic groups, especially in women.

Researchers from Massachusetts General Hospital, Harvard Medical School, and Georgia Tech set out to project trends in alcohol-related cirrhosis and deaths in the US up to 2040 and assess the impact of potential changes in alcohol consumption on those trends.

The team developed a model of drinking patterns and alcohol-related liver disease in people born in 1900 to 2012 in the US. The model leveraged data from multiple sources, including the National Epidemiological Survey on Alcohol and Related Conditions, the National Institute of Alcohol Abuse and Alcoholism, the US Death Index, and numerous published studies.

Researchers validated the model by closely reproducing the trends in deaths due to alcohol-related liver disease that were observed from 2005 to 2018.

Future trends in alcohol-related liver disease were modeled under three potential scenarios based on the level of interventions to address high-risk drinking.

The results showed that without any changes in trends in alcohol consumption, more than one million people could die from alcohol-related liver disease by 2040, of whom 35 percent are projected to be younger than 55 years. Reducing high-risk drinking rates to 2001 levels could prevent 35,000 deaths during the same period.

In contrast, decreasing the high-risk drinking rate by 3.5 percent per year could prevent 299,000 deaths compared with the status quo scenario – a 30 percent decrease.

Additionally, the study showed that in comparison with the status quo, decreasing the high-risk drinking rates by 3.5 percent per year could prevent alcohol-related liver cancer and decompensated cirrhosis by 30 percent from 2019 to 2040.

“Our study underscores the need to bring alcohol-related disease to the forefront of policy discussions and identify effective policies to reduce high-risk drinking in the US,” said senior author Jagpreet Chhatwal, a senior scientist at the MGH Institute for Technology Assessment and an assistant professor at Harvard Medical School.

Researchers also argued that increasing rates of alcohol consumption should be regarded as critical a public health crisis as the opioid epidemic.

“With substantial investment of government resources to address the opioid epidemic, deaths due to opioid overdoses could decrease substantially in the near future. We believe similar investment of funding for alcohol-related liver disease is needed to curb increasing trends in associated deaths,” researchers said.

To stem rising rates of high-risk drinking, the team noted that lawmakers could implement social interventions similar to policies enacted in the 1960s to reduce tobacco consumption.

“The US has previously dealt with impending health crises due to substance abuse. Rigorous scientific study on the effects of smoking and policy experimentation has led to substantial improvements in US healthcare results related to smoking cessation,” researchers stated.

Researchers pointed out that early interventions will be necessary to significantly curb deaths from alcohol-related liver disease.

“Many, if not most, people with alcohol-related liver disease are diagnosed at a late stage, frequently with advanced or decompensated cirrhosis, when the only effective strategies to improve survival are sobriety and liver transplantation, with many patients who achieve sobriety still needing transplantation,” researchers said.

“To reverse these trends clinical interventions and novel pharmacological treatments are urgently needed for alcohol use disorder in patients who have advanced liver disease.”

With this model, researchers expect that patient populations will recognize their risk before it’s too late.

“Our model highlights the long-term risk, especially for younger generations whose drinking continues to outpace older generations,” said Jovan Julien, MS, a PhD student at Georgia Tech.

The study also demonstrates researchers’ ability to examine big data for trends and patterns that could improve population health.

“Given the practical infeasibility of clinical trials to study long-term effects of increased alcohol consumption, data-driven modeling studies such as ours are valuable for understanding the implications of high-risk drinking and making informed policy decisions,” said Turgay Ayer, PhD, MSc, an associate professor at Georgia Tech.