Burden and Cost of Gastrointestinal, Liver, and Pancreatic Diseases in the United States: Update 2021

Abstract

Background:

Gastrointestinal diseases account for considerable health care use and expenditures. We estimated the annual burden, costs, and research funding associated with gastrointestinal, liver, and pancreatic diseases in the United States.

Methods:

We generated estimates using data from National Ambulatory Medical Care Survey, National Hospital Ambulatory Medical Care Survey, Nationwide Emergency Department Sample, National Inpatient Sample, Kids’ Inpatient Database, Nationwide Readmissions Database, Surveillance, Epidemiology, and End Results program, National Vital Statistics System, Centers for Disease Control Wide-ranging Online Data for Epidemiologic Research, MarketScan Commercial Claims and Encounters data, MarketScan Medicare Supplemental data, United Network for Organ Sharing registry, Medical Expenditure Panel Survey, and National Institutes of Health.

Results:

Gastrointestinal health care expenditures totaled $119.6 billion dollars in 2018. Annually, there were more than 36.8 million ambulatory visits for gastrointestinal symptoms and 43.4 million ambulatory visits with a primary gastrointestinal diagnosis. Hospitalizations for a principal gastrointestinal diagnosis accounted for more than 3.8 million admissions with 403,699 readmissions. A total of 22.2 million gastrointestinal endoscopies were performed and 284,844 new gastrointestinal cancers diagnosed. Gastrointestinal diseases and cancers caused 255,407 deaths. The National Institutes of Health supported $3.1 billion dollars (7.5% of the NIH budget) for gastrointestinal research in 2020.

Conclusion:

Gastrointestinal diseases are responsible for millions of health care encounters and hundreds of thousands of deaths which annually costs billions of dollars in the US. To reduce the high burden of gastrointestinal diseases, focused clinical and public health efforts, supported by additional research funding, are warranted.

INTRODUCTION

Gastrointestinal (GI) diseases are common, cause considerable suffering, and can be fatal. GI diseases account for substantial health care use and expenditures.^1–3 Reports detailing the burden of GI diseases are necessary for clinical research, decision-making, and priority-setting. Our aim was to describe health care use, expenditures, and research funding across GI, liver, and pancreatic diseases in the United States (US). To that end, we used multiple data sources to produce GI-specific summary statistics on office-based and emergency department (ED) visits, adult and pediatric hospitalizations, readmissions, and mortality. We report statistics and temporal trends in GI endoscopy, cancers, and organ transplants. We estimated GI-specific health care expenditures and summarized National Institutes of Health (NIH) funding for GI research.

METHODS

Symptoms and Diagnoses across Ambulatory Settings

We used the 2016 National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS) to tabulate the leading GI symptoms and diagnoses in the US for outpatient office-based and emergency department visits. NAMCS and NHAMCS are national surveys conducted annually by the US Centers for Disease Control and Prevention (CDC) (http://www.cdc.gov/nchs/ahcd.htm). The NAMCS surveys non-federal employed office-based physicians or non-physician clinicians primarily engaged in direct patient care. The NHAMCS collects data on ED visits exclusive of Federal, military, and Veterans Administration hospitals. We downloaded public use data files from the CDC website to perform our analyses. Patient-reported symptoms are available in both NAMCS and NHAMCS. We used the most important complaint (variable RFV1) for our analyses. We combined related symptoms and totaled data from office visits and ED visits to present the top 10 most common symptoms. We categorized physician and non-physician diagnoses into relevant disease categories based on clinical expertise using International Classification of Diseases, 10^th Revision, Clinical Modification (ICD-10-CM) (online supplement). We used the primary diagnosis code only. After combining the related diagnoses, we created a rank order list. NAMCS and NHAMCS are based on probability samples. Therefore, sampling weights were applied to all analyses to generate national estimates. When there were fewer than 30 observations for a specific condition, the estimates are unreliable and should be interpreted with caution. Both analyses include children and adults.

Emergency Department Visits

The most common GI diagnoses in the ED were compiled using discharge data from the Nationwide Emergency Department Sample (NEDS), part of the Agency for Healthcare Research and Quality (AHRQ) Healthcare Cost and Utilization Project (HCUP). The NEDS is the largest, publicly available, all-payer ED database in the US. The 2018 NEDS contains information from 990 hospitals participating in HCUP across 36 States and the District of Columbia, approximating a 20-percent stratified sample of US hospital owned EDs. The NEDS sampling frame includes discharge information on patients admitted to the hospital from EDs, patients treated and discharged from EDs, and patients transferred to another hospital from an ED.

We queried the database for the rank order of the principal visit diagnosis (i.e., ICD-10 codes) from the ED for children and adults in all hospitals. We identified all GI diagnoses and symptoms, which were subsequently rank ordered after combining related codes (online supplement) and present the top 20 GI diagnoses and symptoms. Diagnosis categories and associated codes were determined using previously published GI coding categories mapped to ICD-10 codes.^{2, 4–7} Weighted national estimates for ED visits in 2018 were generated, including estimates of the total number of visits and rate of visits per 100,000 persons for each individual ICD-10 code (or group of codes). We included the number of secondary diagnoses for each GI category. Secondary diagnoses are additional diagnoses that appear on the discharge record with the first-listed or principal diagnosis.

Additional information on the most common GI diagnoses among patients seen in the ED according to Clinical Classifications Software Refined (CCSR) categories is provided in the online supplement, including mean age, numbers admitted to the hospital, and deaths. CCSR is a tool developed by AHRQ for clustering diagnoses into a manageable number of clinically meaningful policy-relevant categories. A complete list of GI-related CCSR diagnoses is also included in the online supplement and was used to estimate the total number of ED visits in the US with a principal diagnosis code for a GI disease or symptom.

Hospitalizations

The most common GI diagnoses from hospital admissions were compiled using discharge data from the National Inpatient Sample (NIS), part of HCUP. The NIS approximates a 20-percent stratified sample of all discharges from US community hospitals, excluding rehabilitation and long-term acute care hospitals. The NIS contains information on all hospital stays, regardless of expected payer, including persons covered by Medicare, Medicaid, private insurance, and the uninsured. The sampling frame for the 2018 NIS comprises over 97 percent of the US across 47 states and the District of Columbia.

We queried the database for the rank order of the principal discharge diagnosis (i.e., ICD-10) for all patients in all hospitals. For inpatient stays, the principal diagnosis is that condition thought to be chiefly responsible for the patient’s admission to the hospital. We identified the top 15 GI diagnoses and symptoms, which were subsequently rank ordered after combining related diagnosis codes. Diagnosis categories and associated codes were determined using previously published GI coding categories mapped to ICD-10 codes.^{2, 4–7} Specific groupings of ICD-10 codes for each category are detailed in the online supplement. Weighted national estimates for visits in 2018 were generated. We also included the number of secondary diagnoses for each GI category. Secondary diagnoses are additional diagnoses that appear on the discharge record with the first-listed or principal diagnosis.

A complete list of GI-related CCSR diagnoses is included in the online supplement and was used to estimate the total number of admissions with a principal diagnosis code for GI diseases or symptoms. The list includes median length of stay, median charges and costs, and number of inpatient deaths associated with each CCSR category. Information on the most common GI procedure categories related to GI hospital admissions are also provided in the online supplement.

Pediatric Hospitalizations

We performed a cross-sectional analysis of the 2016 data contained within the Kids’ Inpatient Database (KID), a nationally representative sample of pediatric (age <21 years at admission) hospitalizations from HCUP (http://hcup.ahrq.gov/hcupnet.jsp). KID is the largest, publicly available, all-payer pediatric inpatient database and is designed to be representative of pediatric hospital care across the US. In 2016, the KID data set included pediatric hospitalization data from 44 states including the District of Columbia and represented nearly 4,200 participating community hospitals (defined as short-term, nonfederal, general and specialty) excluding rehabilitation hospitals.

We queried the 2016 KID database and analyzed the principal discharge diagnosis (i.e., ICD-10) for all hospitalizations from participating hospitals. We identified GI diagnoses, which were subsequently rank ordered in frequency after combining related diagnosis codes (online supplement). Weighted national estimates for diagnosis frequency estimates were generated using 2016 KID weighting coefficients provided by HCUP/AHRQ. Median length of stay in days and interquartile range (IQR; 25%–75%) were analyzed in addition to median hospital charges (dollars) per hospitalization and IQR. Hospital charges represent hospital billing for covered hospital services but do not include professional fees or non-covered charges.

Readmissions

Using the 2018 Nationwide Readmissions Database (NRD), a resource of HCUP State Inpatient Databases, we compiled 30-day all cause readmission rates after an index stay for a GI indication. The NRD is a publicly available all-payer inpatient database from the US that is designed to be nationally-representative of readmission rates for all payers and the uninsured. The NRD includes data from 17,686,511 admissions across 28 states and utilizes individual linkage identifiers within a state to allow for the assessment of readmissions. This linkage allows identification of readmissions to hospitals where the index admission did not occur. We identified all patients with an ICD-10 code for the most common inpatient GI discharge diagnoses in the NRD. Any patient with an ICD-10 diagnosis code as a primary diagnosis on the index admission was eligible for inclusion in this study. Diagnosis categories and associated codes (online supplement) were determined using previously published GI coding categories.^{2, 4–7}

We included patients age 18 years and older. Patients were required to have a full 30 days between the date of discharge from their index admission and December 31, 2018. Patients with an index hospitalization that ended in death or transfer to another acute care faculty were excluded. The primary outcome was the first all-cause readmission within 30 days of the index hospitalization. We calculated weighted national estimates for index stays and first readmissions. We estimated the total charges associated with both index admissions and the first readmission.

Cancer Incidence and Mortality

We estimated incidence and mortality rates of GI cancers in adults (age ≥20 years) during 1992–2018 using data from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program of cancer registries and the National Center for Health Statistics. SEER routinely collects data on patient demographics, primary tumor site, tumor morphology, and stage for all cancers diagnosed in defined geographic regions. The SEER 13 registries cover approximately 14% of the US population and include Alaska, Atlanta, Connecticut, Detroit, rural Georgia, Hawaii, Iowa, Los Angeles, New Mexico, San Francisco-Oakland, San Jose-Monterey, Seattle-Puget Sound, and Utah. Mortality data are collected and compiled by the National Center for Health Statistics from death certificates filed in all 50 US states and the District of Columbia.

We estimated age-adjusted (to the 2000 U.S. standard population) incidence and mortality rates of esophageal, colorectal, gastric, liver, pancreas, gallbladder, and small intestine cancer as rates per 100,000 persons using SEER*Stat version 8.3.9, overall and by year and race/ethnicity.⁸ Race/ethnicity included non-Hispanic White, non-Hispanic Black, Hispanic (any race), non-Hispanic Asian/Pacific Islander, and non-Hispanic Alaska Native/American Indian.

Non-Cancer Mortality

We determined the most common causes of non-cancer related GI deaths in the US from the Centers for Disease Control Wide-ranging Online Data for Epidemiologic Research (CDC WONDER) (http://wonder.cdc.gov). CDC WONDER is a publicly-available, online database developed by the CDC. Causes of death are derived from physician-completed death certificates and are classified in accordance with ICD-10. The underlying cause of death is defined as the disease that initiated the train of events leading to death. A disease listed as a contributing cause of death is classified on the death certificate as the underlying cause or any of 20 additional diseases leading to death. For this analysis, we used the 2019 public use data files from CDC Wonder for underlying cause of death and multiple cause of death. The 15 most common non-malignant GI causes of death were identified using ICD-10 codes (supplemental tables) and were ranked by underlying cause of death. Diagnoses were combined to create clinically meaningful categories. The crude rate per 100,000 deaths was calculated by dividing the number of deaths listed as an underlying cause by the total U.S. population of children and adults in the US in 2019 (328,239,523 from the U.S. Census Bureau) and multiplying by 100,000. We also calculated numbers of deaths and mortality rates stratified by ethnicity and race.

Endoscopy Use and Trends

Using IBM MarketScan Commercial Claims and Encounters database and IBM MarketScan Medicare Supplemental database, we examined patterns of endoscopy use in adults (age ≥18 years) during 2002–2019. We examined temporal trends in upper endoscopy, colonoscopy, flexible sigmoidoscopy, Endoscopic Retrograde Cholangiopancreatography (ERCP) and upper and lower endoscopic ultrasound (EUS). MarketScan is an employer-based claims database that includes 77 contributing employers and 12 contributing health plans, with 126 unique carriers and 8 Medicaid states representing approximately 165 million covered lives. Medicare Supplemental database includes beneficiaries ≥65 years with an employer sponsored Medicare supplemental plan. We summed the total number of months persons aged 18–64 years (MarketScan) and ≥65 years (Medicare Supplemental) were enrolled in their insurance plan in each calendar year as standardized denominators of “enrollee-time.” We then depicted time trends by calculating a rate of the procedure per 1,000 enrollee-years in each calendar year, assuming constant rates within each calendar year. We examined rates by age group (18–29, 30–39, 40–49, 50–64, 65–74 and ≥75 years). We estimated the number of procedures performed in 2019 by standardizing the number of procedures in each database to 2018 US Census Bureau data (within age categories).

Organ Transplant

We used information available through the United Network for Organ Sharing (UNOS) registry to tabulate data related to liver, intestine, and pancreas organ transplants. UNOS is a non-profit scientific and educational organization that administers the Organ Procurement and Transplantation Network (OPTN) in the US. An online database system, UNet^SM, was developed to collect, store, analyze, and publish all OPTN data. Data are collected through an online application from transplant professionals in hospitals, histocompatibility laboratories, and organ procurement organizations across the country. The data we included are publicly available as de-identified data at: optn.transplant.hrsa.gov/data/. We included patients with a history of liver, intestine, or pancreas transplant during 1988–2020. Data extracted included number of patients on the wait list, type of transplant, patient demographics at the time of transplant, recipient’s primary disease, and donor characteristics. We report frequencies and proportions for these data.

Expenditures

Using data from the 2018 Medical Expenditure Panel Survey (MEPS) (https://meps.ahrq.gov/), we estimated total expenditures (not charges) for GI diseases and symptoms. The MEPS is a set of large-scale surveys of families and individuals, their medical providers (including doctors, hospitals, pharmacies), and employers across the US. MEPS collects data on the use of specific health services, how frequently they are used, and the cost of these services. These surveys are designed to collect data from a nationally representative sample of households in the US. In the 2018 MEPS Household Component, 29,415 persons were surveyed. This survey represents the civilian noninstitutionalized population. Expenditures are only available for CCSR categories. All GI-related categories available in MEPS were pulled for this analysis. The full year consolidated data file for 2018 was used to estimate expenditures. All estimates were weighted by the MEPS person-level weight (PERWT18F) to produce national estimates of expenditures.

National Institutes of Health Categorical Spending

We gathered estimates of annual GI-specific and all cancer funding from the NIH between 2011–2022 (https://report.nih.gov/funding/categorical-spending#/). Estimates were selected from NIH determined research areas. Actual expenditures are reported when available, otherwise the values were estimated. Individual research projects could be included in multiple categories. We also report the total NIH budget between 2011–2022 from the NIH Office of Budget (https://officeofbudget.od.nih.gov/spending_hist.html). We calculated the percentage of the NIH budget spent on digestive and liver diseases and the percentage of NIH cancer funding spent on GI cancer research.

RESULTS

Symptoms and Diagnoses across Ambulatory Settings

The leading GI symptoms prompting a visit are shown in Table 1. Abdominal pain was responsible for more than 19.0 million visits, followed by vomiting (5.4 million visits), nausea (3.9 million visits), diarrhea (2.6 million visits), and GI bleeding (1.5 million visits). Constipation, anorectal symptoms, heartburn, decreased appetite, and dysphagia accounted for an additional 4.0 million visits. Abdominal pain is also the most frequent diagnosis (vs. symptom) with 15.7 million annual visits (Table 2). There were more than 4.7 million visits with GERD and reflux esophagitis diagnoses, and hemorrhoid diagnoses accounted for nearly 1.9 million visits.

Table 1.

Leading Gastrointestinal Symptoms Prompting an Ambulatory Visit, 2016

		Estimated number of annual visits
Rank	Symptoms	Office visits	Emergency department visits	Total
1	Abdominal pain	6,893,881	12,524,768	19,418,649
2	Vomitinga	2,596,369	2,768,558	5,364,928
3	Nauseaa	1,636,346	2,304,775	3,941,121
4	Diarrheaa	1,915,475	667,584	2,583,060
5	Gastrointestinal Bleedinga	659,135	874,612	1,533,747
6	Constipationa	864,103	378,331	1,242,434
7	Anorectal symptomsa,b	786,668	94,993	881,661
8	Heartburn and indigestiona,b	680,031	43,432	723,463
9	Decreased appetitea,b	511,705	59,662	571,367
10	Dysphagiaa,b	412,870	154,853	567,723
Total				36,828,153

Source: The 2016 National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS), Emergency Department only (https://www.cdc.gov/nchs/ahcd/index.htm))

^aDenotes categories reported from the NAMCS with <30 observations that should be interpreted with caution.

^bDenotes categories reported from the NHAMCS (Emergency Department only) with <30 observations that should be interpreted with caution.

Table 2.

Leading Physician Diagnoses for Gastrointestinal Disorders in the United States, 2016

	Estimated number of annual visits
Diagnosis	Office visitsc	Emergency department visitsc	Total
Abdominal pain	7,979,815	7,693,205	15,673,020
Nausea and vomiting	2,569,377	2,469,273	5,038,650
GERD and reflux esophagitis	4,345,680	317,964	4,663,644
Constipation	2,226,455	854,399	3,080,854
Abdominal wall and inguinal hernia	2,468,048	289,434	2,757,482
Diarrheaa	1,362,383	626,030	1,988,413
Hemorrhoidsa, b	1,803,032	86,514	1,889,546
Gastritis and dyspepsiaa	1,128,641	490,448	1,619,089
Cholelithiasisa	863,398	326,985	1,190,383
Dysphagiaa,b	1,049,240	89,325	1,138,565
Diverticular disease of the colona	783,928	288,183	1,072,111
Malignant neoplasm of the colon or rectuma,b	1,046,452	13,806	1,060,258
Chronic liver disease and cirrhosisa,b	330,201	129,724	459,925
Hepatitis C infectiona,b	383,650		383,650
Rectal bleedinga,b	235,322	139,448	374,770
Pancreatitis - acute and chronica,b	70,201	194,468	264,669
Crohn's diseasea,b	230,770	18,874	249,644
Irritable bowel syndromea,b	245,673	1,531	247,204
Appendicitisa,b	119,691	54,302	173,993
Benign neoplasm of colon and rectuma, b	43,720	3,145	46,865
Total			43,381,926

Source: The 2016 National Ambulatory Medical Care Survey (NAMCS) and the 2016 National Hospital Ambulatory Medical Care Survey (NHAMCS), Emergency Department only (https://www.cdc.gov/nchs/ahcd/index.htm))

^aDenotes categories reported from the NAMCS with <30 observations that should be interpreted with caution.

^bDenotes categories reported from the NHAMCS (Emergency Department only) with <30 observations that should be interpreted with caution.

^cGray shading denotes categories that decreased in visits by 50% or more compared to 2014 data.

Emergency Department Visits

In 2018, there were 17.9 million ED visits with a principal diagnosis code for a GI diagnosis or symptom (comprehensive list in the online supplement). The most common GI diagnoses are detailed in Table 3. Abdominal pain was the most common principal diagnosis with 5.8 million visits. Nausea/vomiting (2.2 million visits), constipation (1.1 million visits), and GI bleeding (941,658 visits) were also high frequency primary diagnoses. A foreign body in the GI tract accounted for 201,613 visits. Abdominal pain (12.0 million visits), gastroesophageal reflux disease (9.3 million visits), and nausea/vomiting (8.5 million visits) were common secondary diagnoses that appear on the discharge record with the principal diagnosis. Additional details on the most common GI diagnoses and symptoms among patients seen in the ED are available in the online supplement.

Table 3.

Most Common Gastrointestinal, Liver and Pancreatic Principal Diagnoses from US Emergency Department Visits, 2018

Rank	Diagnosis/symptom	Annual # visits (principal diagnosis)	Rate of visits per 100,000 persons	Annual # visits (all listed diagnoses)
1	Abdominal pain	5,755,363	1762	12,017,501
2	Nausea and vomiting	2,160,754	661	8,487,667
3	Noninfectious gastroenteritis/colitis	1,261,948	386	1,835,848
4	Constipation	1,084,724	332	3,448,680
5	Gastrointestinal bleeding	941,658	288	1,707,704
	Upper GI bleeding	431,141	132	1,009,733
	Lower GI bleeding	271,575	83	459,028
	GI bleeding not otherwise specified	238,942	73	238,943
6	Cholelithiasis and cholecystitis	726,636	222	1,453,137
7	Gastritis/duodenitis/ulcers (nonbleeding)	645,327	198	1,653,652
8	GI infectiona	618,243	189	1,003,089
9	Diarrheab	595,133	182	3,162,456
10	Diverticulitis	462,954	142	629,139
11	Pancreatitis	419,344	128	890,382
	Acute pancreatitis	381,741	117	626,769
	Chronic pancreatitis	37,603	12	263,613
12	Liver disease and viral hepatitis	400,567	123	4,471,214
	Alcoholic liver disease	130,154	40	613,948
	Hepatic encephalopathy	54,589	17	196,699
	Ascites or SBP	53,873	16	431,043
	Viral hepatitis	25,594	8	1,005,812
	Acute liver failure	25,089	8	131,755
13	Appendicitis	379,197	116	417,028
14	Gastroesophageal reflux	354,698	109	9,279,382
15	Intestinal obstruction and ileus	350,045	107	703,280
	Obstruction	300,267	92	431,722
	Ileus	49,778	15	271,558
16	GI disorders during pregnancy	308,898	95	440,810
17	Foreign body in GI tract	201,613	62	236,204
18	Hemorrhoidsc	193,353	59	645,928
19	Inflammatory bowel disease	148,923	46	548,217
	Crohn’s disease	94,711	29	381,856
	Ulcerative colitis	54,212	17	166,361
20	Abdominal abscess/peritonitisd	97,780	30	251,409

Total ED visits in 2018 with a principal diagnosis code for a GI diagnosis or symptom (see online supplement for complete list) = 17,943,067

Source: Healthcare Cost and Utilization Project Nationwide Emergency Department Sample (NEDS) (https://hcupnet.ahrq.gov)

^aIncludes C difficile infection

^bDoes not include diarrhea coded as gastroenteritis, Clostridioides difficile infection, GI infection

^cExcludes hemorrhoids of pregnancy (included with GI disorders during pregnancy)

^dExcludes liver abscess, spontaneous bacterial peritonitis

Hospitalizations

In 2018, there were 3.9 million adult hospital admissions with a principal diagnosis code for a GI diagnosis or symptom (comprehensive list in the online supplement). The estimated aggregate charges (“the national bill”) and costs for GI hospitalizations in 2018 were more than $200 billion and $47 billion respectively. The most common GI diagnoses associated with hospital admissions are detailed in Table 4. GI bleeding (530,855 visits), cholelithiasis and cholecystitis (316,020 visits), pancreatitis (299,150 visits), and liver disease (280,645 visits) were the most common principal GI discharge diagnoses. Gastroesophageal reflux disease (6.0 million visits), liver disease (3.7 million visits), functional/motility disorders (4.0 million visits), and GI bleeding (1.3 million visits) were common secondary diagnoses that appear on the discharge record with the principal diagnosis. Length of stay, median charges and costs, and deaths by category are detailed in the online supplement. The most common GI procedures related to a hospital admission were esophagogastroduodenoscopy, cholecystectomy, colectomy, paracentesis, and appendectomy (online supplement).

Table 4.

Most Common Gastrointestinal, Liver and Pancreatic Principal Diagnoses from US Hospitals, 2018

Rank	Diagnosis/symptom	Annual # admissions (principal diagnosis)	Rate of visits per 100,000 persons	Annual # admissions (all listed diagnoses)
1	GI bleeding	530,855	166.6	1,275,025
	Upper GI bleeding	311,015	97.6	786,470
	Lower GI bleeding	113,020	35.5	211,350
	GI bleeding NOS	110,030	34.5	277,205
2	Cholelithiasis and cholecystitis	316,020	99.2	741,060
3	Pancreatitis	299,150	93.9	685,880
	Acute pancreatitis	288,220	90.4	492,555
	Chronic pancreatitis	10,930	3.4	193,325
4	Liver disease	280,645	88.1	3,682,015
	Alcoholic liver disease	104,920	32.9	548,940
	Hepatic encephalopathy	48,765	15.3	189,945
	Acute liver failure	26,480	8.3	169,700
	Viral hepatitis	15,600	4.9	722,920
	Ascites or spontaneous bacterial peritonitis	12,025	3.8	395,410
5	Intestinal obstruction	275,355	82.1	417,890
6	Diverticulitis	207,150	65.0	323,760
7	Obesity	193,840	60.8	5,172,209
8	Appendicitis	158,265	49.7	190,770
9	Functional/motility disordersa	137,495	43.1	3,970,930
10	Colorectal cancer	125,805	39.5	259,235
11	Noninfectious gastroenteritis/colitis	123,235	38.7	508,760
12	GI infectionb	105,095	33.0	248,385
13	Inflammatory bowel disease	99,165	31.1	346,710
	Crohn’s disease	60,880	19.1	216,695
	Ulcerative Colitis	38,285	12.0	130,015
14	Clostridioides difficile infection	88,815	27.9	307,075
15	Abdominal pain	58,990	18.5	450,790

Total admissions in 2018 with a principal diagnosis code for a GI diagnosis or symptom (see online supplement for complete list) = 3,873,354

Source: Healthcare Cost and Utilization Project National Inpatient Sample (http://hcup.ahrq.gov/hcupnet.jsp)

^aIncludes esophageal (e.g. achalasia), gastric (e.g. dyspepsia), and intestinal (e.g. irritable bowel syndrome) functional/motility syndromes, as well as constipation and diarrhea

^bIncludes Salmonella, Shigella, E.coli, and other viral, bacterial, and parasitic GI infections. Does not include Clostridioides difficile infection (reported separately).

Pediatric Hospitalizations

In 2016, there were an estimated 202,647 pediatric hospital admissions with a principal diagnosis code for a GI disease or symptom. The most common GI diagnoses are detailed in Table 5. Appendicitis was the most frequent discharge diagnosis (58,017 visits) and was associated with a median charge of nearly $35,000. Intestinal infection (25,550 visits) was the second leading discharge diagnosis among pediatric hospitalizations. Of those infectious hospitalizations, viral gastroenteritis was the most frequent with nearly 14,000 hospitalizations. Congenital malformations were the 6^th leading discharge diagnosis, but the second highest median total cost per hospitalization driven predominately by small bowel malformations.

Table 5.

Most Common Pediatric Gastrointestinal, Liver and Pancreatic Principal Diagnoses in US Hospitals, 2016

	Principal Diagnosis	Annual # admissions (principal diagnosis)	Median length of stay (IQR)(Days)		Median total charges per hospitalization (IQR)(US Dollars)
1	Appendicitis	58,017	1.7	(0.7 – 3.7)	34,710	(24,043 – 51,060)
2	Intestinal infection	25,550	1.7	(0.8 – 3.0)	14,179	(8,537 – 24,709)
	Viral	13,932	1.4	(0.6 – 2.4)	12,381	(7,669 – 20,838)
	Bacterial	9,322	2.4	(1.4 – 4.0)	17,709	(10,481 – 31,672)
	Protozoan	347	2.4	(1.3 – 4.3)	17,012	(8,958 – 31,846)
	Unspecified	1,949	1.5	(0.7 – 2.5)	14,313	(8,871 – 23,513)
3	Inflammatory bowel disease	24,238	2.3	(1.0 – 4.6)	22,090	(11,066 – 45,074)
	Crohn’s disease	7,604	3.5	(2.0 – 6.2)	35,570	(20,239 – 62,460)
	Ulcerative colitis	5,243	3.9	(2.2 – 6.7)	37,277	(20,131 – 69,160)
	Other noninfective gastroenteritisa	11,391	1.3	(0.5 – 2.4)	12,904	(7,494 – 22,539)
4	Pancreaticobiliary diseases	22,813	2.3	(1.2 – 3.9)	31,484	(18,265 – 52,208)
	Cholelithiasis/cholecystitis	10,948	1.9	(1.0 – 3.2)	37,607	(24,557 – 56.575)
	Acute pancreatitis	9,134	2.7	(1.4 – 4.4)	22,817	(12,969 – 42,226)
	Otherb	2,731	2.9	(1.4 – 5.8)	33,325	(19,216 – 59,794)
5	Upper alimentary tract diseases	17,930	2.0	(1.0 – 4.0)	21,896	(11,380 – 40,211)
	GERD	8,151	1.8	(0.8 – 3.8)	16,701	(8,776 – 35,224)
	Esophagitis / gastritis / duodenitis	4,456	1.9	(0.8 – 3.4)	21,496	(11,834 – 35,230)
	Other diseases of the upper tract	3,498	2.5	(1.1 – 5.3)	29,136	(16,518 – 52,902)
	Ulceration of the upper tract	1,480	2.6	(1.4 – 4.4)	31,439	(20,228 – 50,436)
	Functional dyspepsia	345	3.6	(1.9 – 6.9)	30,709	(17,167 – 60,881)
6	Congenital malformations	15,922	2.5	(1.3 – 6.4)	33,670	(20,951 – 75,388)
	Upper alimentary tract	8,513	1.6	(1.0 – 2.7)	24,423	(17,831 – 36,389)
	Otherc	4,193	4.6	(2.4 – 8.9)	53,767	(31,286 – 101,317)
	Large intestine malformation	1,646	3.8	(2.0 – 8.4)	53,211	(31,516 – 97,567)
	Biliary malformation	947	6.6	(3.3 – 12.2)	84,066	(39,537 – 167,355)
	Small intestine malformation	623	23	(13.6 – 44.1)	217,392	(124,347 – 413,120)
7	Paralytic ileus	14,060	2.2	(1.0 – 4.6)	20,259	(10,509 – 44,482)
8	Functional intestinal disorder	11,169	1.5	(1.0 – 2.7)	12,922	(7,356 – 23,700)
9	Abdomen and/or pelvic pain	8,482	1.1	(0.4 – 2.3)	16,190	(9,508 – 26,623)
10	Nausea and vomiting	4,466	1.5	(0.6 – 2.9)	14,261	(8,234 – 24,973)
Total		202,647

Source: Healthcare Cost and Utilization Project Kid Inpatient Database

^aIncludes allergic and dietetic gastroenteritis and colitis, food protein-induced enteropathy, indeterminate colitis, eosinophilic gastritis, gastroenteritis or colitis, other specified noninfective gastroenteritis and colitis, and noninfective gastroenteritis and colitis, unspecified

^bIncludes other diseases of the gallbladder, biliary tract, and pancreas

^cIncludes other congenital malformations of digestive system

Readmissions

In 2018, there were an estimated 403,699 readmissions within 30 days of an index admission for a GI condition. The most common readmissions are detailed in Table 6. GI bleeding, liver disease, and pancreatitis had the highest number of all-cause readmissions within 30 days. Liver disease (31.4%), chronic pancreatitis (27.0%), and Clostridioides difficile infection (22.8) had the highest rates of all-cause readmissions. Median charges for readmissions were higher than median charges for an index admission for most disease categories.

Table 6.

All-cause 30-day Readmissions Ranked by Most Frequently Readmitted Gastrointestinal, Liver or Pancreatic Conditions in US

Rank	Principal diagnosisb for index hospital stay	Number of index stays	Number of 30-day all cause readmissions	% Readmitted	Median Charge per Index stay (US Dollars)	Median Charge per Readmission (US Dollars)
1	Gastrointestinal bleedinga	486,447	84,533	17.4	32,450	36,546
	Upper	286,676	50,519	17.6	33,340	36,590
	Lower	103,361	15,619	15.1	30,933	36,252
2	Liver disease	200,633	63,072	31.4	34,792	36,120
3	Pancreatitis	269,247	42,727	15.9	26,471	28,944
	Acute	259,284	40,036	15.4	26,434	29,024
	Chronic	9,963	2,691	27.0	27,816	27,416
4	Intestinal obstruction	279,871	40,144	14.3	27,680	35,788
5	Cholelithiasis and cholecystitis	286,711	27,098	9.5	46,734	39,041
6	Diverticulitis	204,805	24,874	12.1	30,552	38,740
7	Functional/motility disordersb	101,699	19,876	19.5	27,625	35,785
8	Clostridioides difficile infection	80,842	18,438	22.8	27,734	33,686
9	Colorectal cancer	113,519	17,955	15.8	68,727	39,639
10	Inflammatory bowel disease	88,113	16,285	18.5	31,770	36,206
	Ulcerative colitis	32,667	5,686	17.4	34,276	38,684
	Crohn’s disease	55,446	10,599	19.1	30,312	34,633
11	Non-infectious gastroenteritis/colitis	112,408	16,224	14.4	24,894	34,556
12	Abdominal pain	54,246	9,101	16.8	24,523	33,220
13	Gastrointestinal infectionc	80,525	9,097	11.3	22,224	32,740
14	Obesity	175,528	6,830	3.9	43,259	31,061
15	Appendicitis	108,454	6,676	6.2	40,826	32,105
16	Gastritis	37,985	6,287	16.5	27,493	29,912
17	Gastroesophageal reflux disease	29,217	3,574	12.2	30,213	36,638
Total			403,699

Source: Healthcare Cost and Utilization Project Nationwide Readmissions Database (https://hcupnet.ahrq.gov)

^aThe code “Gastrointestinal hemorrhage NOS” (578.9) was included in the overall GI bleeding category but was not included in subcategories of upper and lower GI bleeding, because it is nonspecific.

^bIncludes esophageal (e.g. achalasia), gastric (e.g. dyspepsia), and intestinal (e.g. irritable bowel syndrome) functional/motility syndromes, as well as constipation and diarrhea.

^cIncludes Salmonella, Shigella, E.coli, and other viral, bacterial, and parasitic GI infections. Does not include Clostridioides difficile infection.

Cancer Incidence and Mortality

In 2018, there were 284,844 new diagnoses and 155,090 deaths from GI cancers (Table 7). Between 1992–2018, the incidence and mortality of esophageal cancer (Figures 1A and B) decreased among non-Hispanic Blacks, Asians, and Hispanics. Over the same time, esophageal cancer incidence and mortality increased in non-Hispanic Whites. The incidence and mortality of colorectal cancer (Figures 2A and B) decreased among all race/ethnicities but remains higher among non-Hispanic Blacks compared to other racial/ethnic groups. The incidence and mortality of gastric cancer (Figures 3A and B) decreased among all race/ethnicities, and incidence rates are now highest among Hispanics. Liver cancer incidence (Figure 4A) began decreasing among all race/ethnicities between 2007–2015 but mortality increased (Figure 4B) between 1992–2018 for all race/ethnicities except non-Hispanic Asians. The incidence of pancreatic cancer (Figure 5A) increased among non-Hispanic Whites, Asians, and Hispanics between 1992–2018 and remained highest among non-Hispanic Blacks. Pancreatic cancer mortality has changed little over this time (Figure 5B). Given the small population size, cancer rates for Alaska Native/American Indian were not stable and are included in the online supplement. We observed increasing incidence and mortality for almost all GI cancers in this population between 1992 – 2018. Because gallbladder and small intestine cancers are rare, temporal trends of these cancers are included in the online supplement only.

Table 7.

New Diagnoses and Deaths from Gastrointestinal, Pancreatic and Liver Cancers in the United States, 2018

Cancer Site	Number of new diagnoses	Number of deaths
Colon and Rectum	141,074	52,163
Pancreas	52,546	44,914
Liver and Intrahepatic Bile Ducts	34,638	27,685
Stomach	24,101	11,043
Esophagus	18,364	15,419
Small Intestine	9,810	1,686
Gallbladder	4,311	2,180
Total	284,844	155,090

Sources: National Program of Cancer Registries and National Center for Health Statistics, 2018

Figure 1A.

Age-adjusted incidence of esophageal cancer by race/ethnicity, National Cancer Institute’s Surveillance, Epidemiology, and End Results program of cancer registries, 1992 – 2018

Figure 2A.

Age-adjusted incidence of colon and rectal cancer by race/ethnicity, National Cancer Institute’s Surveillance, Epidemiology, and End Results program of cancer registries, 1992 – 2018

Figure 3A.

Age-adjusted incidence of stomach cancer by race/ethnicity, National Cancer Institute’s Surveillance, Epidemiology, and End Results program of cancer registries, 1992 – 2018

Figure 4A.

Age-adjusted incidence of liver and intrahepatic bile duct cancer by race/ethnicity, National Cancer Institute’s Surveillance, Epidemiology, and End Results program of cancer registries, 1992 – 2018

Figure 4B.

Age-adjusted mortality of liver and intrahepatic bile duct cancer, National Center of Health Statistics, 1992 – 2018

Figure 5A.

Age-adjusted incidence of pancreas cancer by race/ethnicity, National Cancer Institute’s Surveillance, Epidemiology, and End Results program of cancer registries, 1992 – 2018

Figure 5B.

Age-adjusted mortality of pancreas cancer, National Center of Health Statistics, 1992 – 2018

Non-Cancer Mortality

Among the top 15 causes of death from non-malignant GI diseases in 2019, there were a total of 100,317 total deaths (Table 8). The most common causes of non-malignant GI-related mortality were alcohol-associated liver disease, underlying 24,110 deaths, and serving as a contributing factor for more than twice that number, hepatic fibrosis/cirrhosis of any etiology (20,184 deaths), gastrointestinal bleeding (9,548 deaths), vascular disorders of the intestine (7,757 deaths), and paralytic ileus and intestinal obstruction (6,943 thousand deaths). More than half (56%) of deaths from all non-malignant GI diseases were attributable to liver disease. When stratified by race/ethnicity, a similar distribution of causes of death was observed (online supplement).

Table 8.

Causes of Death from Non-Malignant Gastrointestinal, Pancreatic and Liver Diseases in the United States, 2019

		Women			Men			Total
Rank	Disease	Deaths underlyingc	Deaths contributingd	Crude mortality rate (per 100,000)a	Deaths underlyingc	Deaths contributingd	Crude mortality rate (per 100,000)a	Deaths underlyingc	Deaths contributingd	Crude mortality rate (per 100,000)a
1	Alcohol-associated liver disease	7,460	9,294	4.5	16,650	22,339	10.3	24,110	31,633	7.3
2	Hepatic fibrosis/cirrhosis (all-cause)b	8,763	18,473	5.3	11,421	26,381	7.1	20,184	44,854	6.1
3	Gastrointestinal bleeding, unspecified	4,642	16,184	2.8	4,906	19,662	3.0	9,548	35,846	2.9
4	Vascular disorders of the intestine	4,781	9,082	2.9	2,976	6,395	1.8	7,757	15,477	2.4
5	Paralytic ileus and intestinal obstruction	3,934	10,225	2.4	3,009	8,314	1.9	6,943	18,539	2.1
6	Hepatic failure (acute and chronic) b	2,095	12,173	1.3	2,516	15,746	1.6	4,611	27,919	1.4
7	Clostridioides difficile	2,644	4,569	1.6	1,889	3,557	1.2	4,533	8,126	1.4
8	Ulcers (gastric/duodenal/peptic)	1,746	3,328	1.0	1,798	3,708	1.1	3,544	7,036	1.1
9	Chronic hepatitis C	1,252	4,008	0.8	2,245	10,223	1.4	3,497	14,231	1.1
10	Fatty change of liver-not elsewhere specified	2,158	3,667	1.3	1,315	2,987	0.8	3,473	6,654	1.1
11	Diverticular disease	1,931	3,042	1.2	983	1,738	0.6	2,914	4,780	0.9
12	Acute pancreatitis	1,118	2,311	0.7	1,693	3,367	1.0	2,811	5,678	0.9
13	Perforation of intestine (non-traumatic)	1,676	4,369	1.0	1,021	3,131	0.6	2,697	7,500	0.8
14	Cholecystitis	1,290	1,950	0.8	1,305	2,098	0.8	2,595	4,048	0.8
15	Cholangitis	536	1,144	0.3	564	1,249	0.3	1,100	2,393	0.3
Totals								100,317	234,714

Source: Centers for Disease Control Wide-ranging Online Data for Epidemiologic Research (http://wonder.cdc.gov). See supplementary table for specific groupings of ICD-10 CM codes for each category

^aCrude rate per 100,000 deaths was calculated by dividing the number of deaths listed as an underlying cause by the total U.S. population in the United States in 2019 then multiplying by 100,000

^bDoes not include deaths from Liver disease, unspecified (K76.9), which accounted for 3,711 deaths in 201

^cUnderlying cause of death is defined as the disease that initiated the train of events leading to death.

^dA contributing cause of death is classified on the death certificate as the underlying cause or any of 20 additional diseases leading to death

Endoscopy Use and Trends

In 2019, an estimated 13,837,748 colonoscopies, 7,459,419 upper endoscopies, 379,883 flexible sigmoidoscopies, 290,655 upper EUSs, 177,508 ERCPs and 17,428 lower EUSs were performed in adults (Table 9). Colonoscopy use in adults aged 50–74 years increased every year between 2011 and 2019, while rates in adults age ≥75 years decreased and then plateaued (Figure 6A). Since 2002, rates of upper endoscopies increased slightly in adults age 18–74 years and decreased in the oldest age group (Figure 6B). Rates of flexible sigmoidoscopies decreased across all age groups between 2002 and 2010 and then plateaued (Figure 6C). The rates of ERCPs have remained stable across all age groups between 2002 and 2019 (Figure 6D), while upper EUS use increased each year (Figure 6E). The rates of lower EUSs have declined since 2012 (Figure 6F).

Table 9.

Estimated Annual Number of Endoscopic Procedures in the United States, 2019

Procedure	Numbers
Colonoscopy	13,837,748
Upper endoscopy	7,459,419
Flexible sigmoidoscopy	379,883
Upper endoscopic ultrasound	290,655
Endoscopic retrograde cholangiopancreatography	177,508
Lower endoscopic ultrasound	17,428
Total	22,162,641

Source: MarketScan Commercial Claims and Encounters and Medicare Supplemental database

Figure 6A.

Colonoscopies performed per 1,000 enrollee-years, by age group, MarketScan Commercial Claims and Encounters and Medicare Supplemental database, 2002–2019

Figure 6B.

Upper endoscopies performed per 1,000 enrollee-years, by age group, MarketScan Commercial Claims and Encounters and Medicare Supplemental database, 2002–2019

Figure 6C.

Flexible sigmoidoscopies performed per 1,000 enrollee-years, by age group, MarketScan Commercial Claims and Encounters and Medicare Supplemental database, 2002–2019

Figure 6D.

Endoscopic retrograde cholangiopancreatographies performed per 1,000 enrollee-years, by age group, MarketScan Commercial Claims and Encounters and Medicare Supplemental database, 2002–2019

Figure 6E.

Upper endoscopic ultrasound performed per 1,000 enrollee-years, by age group, MarketScan Commercial Claims and Encounters and Medicare Supplemental database, 2002–2019

Figure 6F.

Lower endoscopic ultrasound performed per 1,000 enrollee-years, by age group, MarketScan Commercial Claims and Encounters and Medicare Supplemental database, 2002–2019

Organ Transplant

In 2020, there were 12,610 persons added to the liver transplant wait list, as well as 138 intestine, 446 pancreas, and 1275 kidney/pancreas candidates added to their respective wait lists. The number of liver transplants performed has steadily increased over the past three decades with a total of 8,906 liver transplants performed in 2020 (Figure 7A). Alcohol-associated cirrhosis is the leading indication for liver transplantation, accounting for 25.9% of liver transplants in 2020, followed by cirrhosis from nonalcoholic steatohepatitis (14.7%) and hepatocellular carcinoma (14.1%). In contrast to liver transplants, there has been a decrease in the number of intestine and pancreas single organ transplants throughout the past decade, while the number of combined kidney/pancreas transplants has remained stable (Figure 7B–D). A total of 91 intestine and 962 pancreas (135 pancreas only, 827 simultaneous kidney/pancreas) transplants were performed in 2020. Short gut syndrome accounted for 45.1% of all intestine transplants and 13.2% were related to intestine graft failure and need for retransplant. A primary diagnosis of diabetes accounted for 82.7% of pancreas transplants. The characteristics of the transplant recipients and donors are described in Table 10.

Figure 7A.

Numbers of persons added to the waiting list and liver transplants performed, United Network for Organ Sharing registry, 1988–2020

Figure 7B.

Numbers of persons added to the waiting list and intestine transplants performed, United Network for Organ Sharing registry, 1988–2020

Table 10.

Recipient and donor characteristics of liver, pancreas, and intestine transplants in the United States, 2020

	Liver	Pancreas +/− Kidney	Intestine
Transplants performed	8,906	962	91
Age, n (%)
< 1 year	139 (1.6)	5 (0.5)	5 (5.5)
1–5 years	166 (1.9)	8 (0.8)	16 (17.6)
6–10 years	64 (0.7)	4 (0.4)	7 (7.7)
11–17 years	133 (1.5)	6 (0.6)	6 (6.6)
18–34 years	606 (6.8)	226 (23.5)	23 (25.3)
35–49 years	1,751 (19.7)	477 (49.6)	18 (19.8)
50–64 years	4,054 (45.5)	228 (23.7)	13 (14.3)
≥ 65 years	1,993 (22.4)	8 (0.9)	3 (3.3)
Sex, n (%)
Male	5,562 (62.5)	559 (58.1)	52 (57.1)
Female	3,344 (37.5)	403 (41.9)	39 (42.9)
Race/Ethnicity, n (%)
White	6,118 (68.7)	495 (51.5)	53 (58.2)
Black	705 (7.9)	269 (28.0)	25 (27.5)
Hispanic	1,539 (17.3)	151 (15.7)	9 (9.9)
Asian	381 (4.3)	35 (3.6)	3 (3.3)
American Indian / Alaska Native	71 (0.8)	4 (0.4)	0 (0.0)
Pacific Islander	19 (0.2)	2 (0.2)	0 (0.0)
Multiracial	73 (0.8)	6 (0.6)	1 (1.1)
Hepatitis C virus positive donors, n (%)	805 (9.0)	39 (4.1)	3 (3.3)
Hepatitis B virus core positive antibody donors, n (%)	413 (4.6)	10 (1.0)	0 (0.0)
High risk donors, n (%)	2,454 (27.6)	241 (25.1)	14 (15.4)
All donation after circulatory death + non-donation after circulatory death, n (%)
Brain death donor	7,587 (85.2)	943 (98.0)	91 (100.0)
Donation after circulatory death donors	831 (9.3)	28 (2.9)	0 (0.0)
Living donors	491 (5.5)	0 (0.0)	0 (0.0)

Source: United Network for Organ Sharing registry

Expenditures

Health care expenditures for GI conditions totaled $119.6 billion annually (Table 11). Among the 23 condition categories available, the five most expensive categories were biliary tract disease ($16.9 billion), esophageal disorders ($12.1 billion), abdominal pain ($9.5 billion), abdominal hernias ($9.0 billion) and diverticular disease ($9.0 billion). Prescription medications accounted for 53% of expenditures for esophageal disorders and 71% of expenditures for inflammatory bowel diseases.

Table 11.

Total Expenses Gastrointestinal, Pancreatic and Liver Diseases in the United States, 2018

Condition Categorya	Total Expenditures (in millions)	Distribution of Total Expenditures by Type of Service
		Office-Based Provider Visits %	Outpatient Department Visits %	Hospital Inpatient Stays %	Emergency Room Visits %	Prescribed Medicines %	Home Health %
Biliary tract diseases	$16,912.6	21%	13%	61%	4%	1%	1%
Esophageal disorders	$12,119.0	21%	10%	13%	3%	53%	2%
Abdominal pain and other digestive/abdomen signs and symptoms	$9,511.1	14%	9%	38%	20%	20%	7%
Abdominal hernias	$9,018.8	27%	32%	34%	6%	2%	6%
Diverticulosis and diverticulitis	$8,984.8	4%	3%	82%	10%	1%	1%
Other gastrointestinal disorders	$8,706.4	11%	13%	31%	10%	36%	5%
Regional enteritis and ulcerative colitis	$8,520.7	9%	1%	18%	2%	71%	1%
Other liver diseases	$5,756.6	4%	8%	50%	2%	37%	5%
Intestinal obstruction and ileus	$5,325.1	2%	1%	94%	4%	0%	1%
Pancreatic disorders (excluding diabetes)	$4,974.7	26%	2%	44%	3%	26%	1%
Appendicitis/appendiceal conditions	$4,769.4	2%	4%	80%	14%	0%	1%
Anal and rectal conditions	$3,823.5	3%	0%	94%	2%	0%	1%
Other disorders of stomach and duodenum	$3,231.5	20%	8%	9%	9%	54%	11%
Hepatitis	$3,050.9	6%	6%	*	*	88%	*
Noninfectious gastroenteritis	$2,905.8	20%	5%	39%	9%	28%	0%
Gastritis and duodenitis	$2,811.9	9%	6%	22%	46%	18%	*
Intestinal infections	$2,425.8	9%	5%	53%	29%	4%	*
Nausea and vomiting	$2,155.5	13%	3%	39%	15%	29%	1%
Colorectal cancers	$1,349.3	51%	26%	18%	0%	6%	36%
Hemorrhoids	$1,324.6	26%	45%	21%	3%	6%	*
Dysphagia	$840.8	30%	38%	28%	3%	1%	*
Gastroduodenal ulcers	$819.8	8%	9%	72%	2%	10%	*
Gastrointestinal bleeding	$301.7	19%	15%	8%	39%	19%	21%
Total	$119,640.30

Source: Medical Expenditure Panel Survey data, Household Component 2018 survey (https://meps.ahrq.gov/)

^aThe condition categories were defined using the Clinical Classifications Software Refined (CCRS) which is a tool developed by AHRQ for clustering diagnoses into a manageable number of clinically meaningful policy-relevant categories

^*No survey participants reported services from which expenditures can be extrapolated

National Institutes of Health Categorical Spending

The NIH supported $2.3 billion dollars in digestive diseases research and $845 million dollars in liver disease research in 2020 (Table 12). The total NIH budget increased from 30.6 billion in 2011 to 41.5 billion in 2020. Funding for GI research has kept pace with increases in the NIH budget. In 2020, $319 million dollars financed colorectal cancer research, $130 million for liver cancer research, $230 million for pancreatic cancer research, $24 million for stomach cancer research, and $37 for esophageal cancer research (Table 13). Of the $7.0 billion in funded cancer research at NIH in 2020, 10.5% supported GI cancer research.

Table 12.

National Institutes of Health Estimates of Funding for Select Gastrointestinal Disease Categories and Total Budget, 2011–2022

Gastrointestinal Disease Category (Dollars in millions and rounded)	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021 estimated	2022 estimated
Digestive diseases	$1,698	$1,719	$1,575	$1,607	$1,684	$1,745	$1,881	$2,242	$2,173	$2,273	$2,316	$2,390
Inflammatory Bowel Disease	$113	$121	$114	$125	$128	$126	$134	$144	$163	$177	$180	$186
Liver diseases	$623	$632	$594	$605	$616	$635	$691	$802	$851	$845	$860	$892
Chronic Liver Disease & Cirrhosis	$303	$288	$282	$293	$295	$293	$285	$324	$351	$368	$373	$388
Hepatitis	$208	$210	$195	$251	$262	$267	$306	$349	$378	$362	$371	$385
Hepatitis A	$4	$2	$2	$3	$4	$3	$4	$5	$3	$5	$5	$5
Hepatitis B	$58	$51	$48	$48	$42	$47	$42	$55	$67	$70	$72	$74
Hepatitis C	$114	$112	$101	$111	$96	$107	$114	$129	$150	$120	$123	$128
National Institutes of Health total budget	$30,630	$30,802	$29,129	$30,019	$30,293	$32,258	$34,147	$36,642	$39,420	$41,525	-	-
% National Institutes of Health budget spent on digestive and liver diseases	7.6%	7.6%	7.4%	7.4%	7.6%	7.4%	7.5%	8.3%	7.7%	7.5%	-	-

Sources: National Institutes of Health Research Portfolio Online Reporting Tools Categorical Spending and National Institutes of Health Office of Budget

Table 13.

National Institutes of Health Estimates of Funding for Gastrointestinal Cancers and All Cancer, 2011–2022

Gastrointestinal Cancer (Dollars in millions and rounded)	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021 estimated	2022 estimated
Colorectal cancer	$313	$302	$281	$271	$309	$274	$270	$314	$294	$319	$325	$335
Liver cancer	$74	$73	$71	$74	$85	$83	$90	$113	$127	$130	$133	$137
Pancreatic cancer	$112	$127	$125	$123	$174	$168	$199	$215	$219	$230	$233	$239
Stomach cancer	+	+	+	+	+	+	+	+	+	$24	$25	$25
Esophageal cancer	+	+	+	+	+	+	+	+	+	$37	$38	$39
NIH spending on all cancer research	$5,448	$5,621	$5,274	$5,392	$5,389	$5,589	$5,980	$6,335	$6,520	$7,035	$7,176	$7,435
% NIH cancer funding spent on GI cancers	9.2%	8.9%	9.0%	8.7%	10.5%	9.4%	9.3%	10.1%	9.8%	10.5%	10.5%	10.4%

Source: National Institutes of Health Research Portfolio Online Reporting Tools Categorical Spending

DISCUSSION

GI, liver, and pancreatic diseases account for a substantial burden of health care in the US. Expenditures total at least $119.6 billion dollars annually using data from the MEPS. This conservative estimate did not include most GI cancers and likely underestimated the costs associated with some GI conditions. For example, the MEPS estimate associated with GI bleeding was $300 million. In comparison, the aggregate cost of GI bleeding was more realistically $3.7 billion dollars as estimated using inpatient data from NIS. The NIH supported $3.1 billion dollars for GI specific research in 2020 with GI research continuing to capture roughly ~7% of the NIH budget.

Using discharge data from the NIS, we determined the most common primary GI diagnoses leading to hospitalizations.^{9, 10} We also reported the total number of secondary diagnoses for each GI category because GI conditions may develop after admission to the hospital (e.g. GI bleeding) or may be an underlying condition for hospitalized patients (e.g. liver disease). We found substantial numbers of GI conditions and symptoms listed in secondary positions on the discharge record. For example, liver disease accounted for 280,645 discharges with a primary diagnosis, however there were 13-fold as many discharges (3.6 million in 2018) with liver disease as a secondary diagnosis.¹¹ Including all diagnoses captures a burden of GI disease not previously reported.

We used data from the KID database to specifically estimate the most common primary GI diagnoses among children. Compared with a similar report using data from 1997, the total number of admissions for pediatric intestinal infections has decreased by more than 50% over the last 20 years.¹² This decline may be attributable to the introduction of the rotavirus vaccine in 2006, which is now a routine immunization for all US infants.¹³ Admissions for abdominal pain have also declined, while acute pancreatitis increased. This change is potentially due to an increased awareness of pediatric pancreatitis and recategorization from abdominal pain to pancreatitis.^{14, 15} Admissions for inflammatory bowel disease have increased, which is likely due to increased disease incidence.^{16, 17} Similar to 1997, appendicitis remains the most common primary GI diagnosis among children in the United States.¹⁸

Readmissions represent a significant burden for patients admitted with a primary diagnosis of a GI condition. In our evaluation of readmissions using the NRD, patients with a primary admission for GI bleeding or liver disease demonstrated the highest total number of 30-day all cause readmissions. The burden of readmissions among patients with liver disease is noteworthy and markedly higher compared to our last report.³ We suspect that the high burden of readmissions in the liver disease population is accurate given that a similarly high proportion of 30-day readmissions has been noted in other recent evaluations of patients with a primary diagnosis of a liver disease (particularly those with complications of cirrhosis).^19–23 In comparison to readmission data from 2015 NRD, readmissions for a primary diagnosis of obesity and appendicitis have decreased. However, it should be noted that there were significant changes in coding from ICD-9 to ICD-10 within the 2015 data. Additionally, there is a high likelihood that patients admitted with a primary diagnosis of obesity or appendicitis may have undergone a surgical procedures during the primary admission potentially leading to anomalies in readmission counts in prior evaluations using a different ICD coding system.

GI cancers account for a large number of diagnoses and deaths annually, with persistent disparities in incidence and mortality rates by race/ethnicity. Despite an uptake in colorectal cancer screening and surveillance, colorectal cancer remains the leading cause of mortality among GI cancers. Pancreas cancer mortality rates in 2018 are approaching rates for colorectal cancer mortality. If that trend continues, we may soon see pancreas cancer become the leading cause of GI-cancer related death. Liver cancer mortality continues to increase over time. We found that of the $7.0 billion in cancer research funded at NIH in 2020, 10.5% supported GI cancer research.

We determined the most common causes of non-cancer related GI deaths in the US using data from the CDC. A substantial proportion of deaths from all non-malignant GI diseases were attributable to liver disease. Mortality from alcohol-associated liver disease and non-alcoholic fatty liver disease continue to increase over time.²⁴ Hepatitis C related mortality is declining (rate per 100,000: 2.3 in 2012, 1.8 in 2016, 1.1 in 2019), likely as a result of the availability of effective hepatitis C therapy.^25–27 Mortality from Clostridioides difficile infection has declined since 2011, likely as a result of a true decline in Clostridioides difficile infection and potentially from more effective therapy.²⁸

Using data from commercial claims, we estimated the annual number of gastrointestinal endoscopies in the US and described patterns of use between 2002 and 2019. Colonoscopy use in adults ages 50–74 years has increased every year since 2011. This trend is likely driven by an uptake in colorectal cancer screening and surveillance.²⁹ Rates of ERCP have remained stable, which is expected given that indications for this procedure have not changed in the last 20 years.³⁰ Compared to ERCP, upper EUS is relatively new with growing and evolving utility beyond staging of malignancy. Rates of upper EUS have increased over time as EUS becomes a standard approach to tissue acquisition in pancreatic cancer, and a tissue diagnose is required in these patients for required for delivery of neoadjuvant therapy.^{31, 32} Additionally, therapeutic endoscopy is increasingly being performed for drainage of pancreatic pseudocysts and walled-off necrosis, biliary drainage, and creation of enteric anastomoses among other procedures. Racial, ethnic, and regional disparities in access to most GI endoscopy procedures exist, which suggests an unmet need for GI procedures across the US.^33–35

We gathered information available through the UNOS registry to examine national characteristics and temporal trends related to liver, intestine, and pancreas organ transplants. Liver transplants account for a vast majority of all gastrointestinal-related solid organ transplants. Over the past three decades, we have seen a steady increase in the number of liver transplants performed each year. This growth is, in part, due to an increase in living donor transplants, use of organs from donations after circulatory death and from donors with positive HCV status. Liver transplants from living donors began in 1989 and has grown from 0.1% to 5.5% in 2020. Additionally, livers from donations after circulatory death accounted for 1.1% of all liver transplants in 1993 and has increased to 9.3% of transplants in 2020, and liver donations from HCV positive donors began in 1994 and has increased from 1.2% to 9.0%. Remarkably, the number of liver transplants in 2020 (8,906 transplants) did not decline during the COVID-19 pandemic compared to 2019 (8,896 transplants).³⁶

This work has limitations. We relied on data collected by the CDC on leading gastrointestinal symptoms prompting an ambulatory visit and leading physician diagnoses in the ambulatory setting. The number of sampled visits in the 2016 dataset were substantially reduced and as a consequence some diagnoses from the prior report (e.g. ulcerative colitis) had too few visits in 2016 to generate estimates. Compared with our last report, the numbers of ambulatory visits reported in Tables 1 and 2 have decreased. The decreased number of visits may relate to changes in sampling or may be due to a downward trend in outpatient visits in the primary care setting.^{3, 37} Since our last report, diagnosis coding has transitioned from ICD-9 to ICD-10. We mapped the codes from the last report. In addition to mapping, we consulted available lists of ICD-10 codes for specific gastrointestinal conditions to ensure that important diseases were not overlooked. Due to differences between the ICD-9 and ICD-10 codes, caution should be used when comparing our results in this report to prior estimates. We may have overestimated the number of GI procedures performed annually because we do not have data from the uninsured and those covered by Medicaid. Our estimates appear to be conservative compared with others.^{38, 39} Using data from the Medical Expenditure Panel Survey, total expenditures for GI diseases fell from $135.9 billion in 2015 to 119.6 in 2018. A decrease in hepatitis expenditures ($23.2 billion to $3.1 billion) accounted for most of this difference. There is evidence that drug expenditures for hepatitis C medications peaked in 2015 in the United States and have decreased over time.⁴⁰ It is also important to note that between 2015 and 2018, diagnosis coding transitioned from ICD-9 to ICD-10 and the Medical Expenditure Panel Survey switched from Clinical Classifications Software codes to Clinical Classification Software Refined codes. Given these changes, the Agency of Health Research and Quality recommend extreme caution when comparing data expenditures before and after this transition. The limitations of each data source are described in the online supplement.

It essential that we understand the basic epidemiology of digestive and liver diseases in the United States and identify changes in this epidemiology over time. Carefully examining the data in this report can help generate areas for future investigation, prioritize research funding, identify areas of unmet need or disparities and provide an important overview of the impact of digestive and liver conditions. By periodically updating the data we can map trends and identify gaps. During the COVID-19 pandemic, elective endoscopy stopped, and patients delayed or lost access to health care. In the future, we expect to find a greater burden of mortality from GI diseases and cancers because of the impact of COVID-19 on healthcare. Finally, we hope that others will use this report as motivation to take a deeper dive into individual diseases. There is much to learn from carefully studying existing data sources.

Figure 1B.

Age-adjusted mortality of esophageal cancer, National Center of Health Statistics, 1992 – 2018

Figure 2B.

Age-adjusted mortality of colon and rectal cancer, National Center of Health Statistics, 1992 –2018

Figure 3B.

Age-adjusted mortality of stomach cancer, National Center of Health Statistics, 1992 – 2018

Figure 7C.

Numbers of persons added to the waiting list and pancreas transplants performed, United Network for Organ Sharing registry, 1988–2020

Figure 7D.

Numbers persons added to the waiting list and of kidney-pancreas transplants performed, United Network for Organ Sharing registry, 1988–2020

Abbreviations:

AHRQ

Agency for Healthcare Research and Quality

CCSR

Clinical Classifications Software Refined

CDC

US Centers for Disease Control and Prevention

CDC WONDER

Centers for Disease Control Wide-ranging Online Data for Epidemiologic Research

ED

Emergency Department

GI

Gastrointestinal

ERCP

endoscopic retrograde cholangiopancreatography

EUS

endoscopic ultrasound

HCUP

Healthcare Cost and Utilization Project

ICD-10-CM

International Classification of Diseases, 10^th Revision, Clinical Modification

IQR

interquartile range

KID

Kids’ Inpatient Database

MEPS

Medical Expenditure Panel Survey

NAMCS

National Ambulatory Medical Care Survey

NEDS

Nationwide Emergency Department Sample

NIH

National Institutes of Health

NIS

National Inpatient Sample

OPTN

Organ Procurement and Transplantation Network

NRD

Nationwide Readmissions Database

SEER

Surveillance, Epidemiology, and End Results

UNOS

United Network for Organ Sharing

US

United States