Setting of Alcohol Use Disorder Diagnosis among Women Enrolled in Medicare

Abstract

Objective:

This study assessed setting of initial diagnosis of alcohol use disorder among Medicare-enrolled women.

Methods:

Using 2016–2022 Medicare fee-for-service claims, this study identified the first alcohol use disorder diagnosis for women 18–64 years versus 65+ and examined the setting of diagnosis. For diagnoses during a hospitalization, the study determined how often alcohol use disorder was the principal diagnosis.

Results:

Initial diagnoses for two-thirds of women 18–64 and women 65+ occurred in a hospital or emergency department. Only 10–14% of diagnoses were by primary care physicians and 2–5% by behavioral health specialists in community-based practices. Alcohol use disorder was the principal diagnosis for ≤10% of hospitalizations where a patient was initially diagnosed, suggesting it was often identified during acute treatment for other conditions.

Conclusions:

This suggests the need for continued efforts to increase screening outside of acute care settings for earlier diagnosis and engagement in community-based care.

Rates of binge drinking, alcohol use disorder, and alcohol-related medical conditions have increased among women in recent years.^1–4 While alcohol-related mortality rates remain higher among men than women, the longstanding gap has narrowed due to more rapid increases among women.^5,6 Rates of binge drinking and alcohol use disorder have been rising in women over 50 for some time,⁷ but recent trends have diverged somewhat by age,³ with particularly notable increases in heavy drinking among women in their 30s and 40s,⁸ and a rise in serious alcohol-related medical complications among women aged 40–64.⁴

Timely access to evidence-based treatment and recovery support can mitigate the health consequences of heavy alcohol use, yet only 6.8% of women with alcohol use disorder report past-year treatment use.^1,9–11 Given increasing rates of alcohol-related harms among women, ^2–4 improving early identification and treatment in this population is a public health priority.¹² However, little is known about where alcohol use disorder among women is initially identified by clinicians in the health care system. This paper uses claims data to identify the setting of initial diagnosis among women insured by Medicare, which covers more than 31 million women over 65 and approximately 4 million women under age 65.¹³

Methods

Using Medicare fee-for-service claims, the authors identified the first (i.e., “index”) alcohol use disorder diagnosis (ICD-10 F10.XX) in any field for beneficiaries 18–64 years and those 65+ during 2016–2022. For cohort inclusion, beneficiaries must have been continuously enrolled in Medicare Parts A and B with no alcohol use disorder diagnoses for the twelve months before the index diagnosis.

The key variable was the setting of index diagnosis: inpatient hospital, emergency department (ED), or outpatient care provided by a primary care physician, behavioral health specialist (psychiatrist, addiction medicine physician, psychologist, or social worker), advanced practice nurse or physician assistant (APRN/PA), gastroenterologist, obstetrician/gynecologist, cardiologist, or other type of clinician. We defined outpatient settings using provider specialty codes. We also examined patient race and ethnicity, dual eligibility for Medicaid, and rurality of residence; whether the index diagnosis encounter was via telemedicine; whether it reflected moderate or severe disease (vs. mild); and clinical comorbidities in the past 12 months (mental health disorder, other substance use disorder, and acute or chronic alcohol-related medical conditions). Finally, we categorized the principal discharge diagnosis for patients whose index diagnosis occurred during a hospitalization as alcohol, other substance use, or mental health disorder; medical conditions that are 100% attributable to alcohol use (e.g., alcohol-associated liver disease); medical conditions linked to alcohol use but not 100% attributable to it (e.g., gastroesophageal hemorrhage); other medical conditions that alcohol may worsen or place patients at greater risk for (e.g., migraines); or other conditions (see Supplement).

We examined the under- and over-65 groups separately given the difference in their primary reason for Medicare eligibility (i.e., age vs. disability). Among the subset of beneficiaries whose initial diagnosis occurred during an inpatient hospitalization, we also examined the distribution of principal discharge diagnosis on the index claim.

Results

We identified 116,226 women 18–64 and 264,857 women 65+ with an index diagnosis. A sizeable proportion of women in both age groups (77.8% for 18–64 and 49.7% for 65+) had a mental health comorbidity, while 36.0% of women 18–64 and 7.2% of women 65+ had a comorbid substance use disorder (Table 1).

Table 1:

Characteristics of Women Enrolled in Medicare Aged 18–64 and 65+ with Index Alcohol Use Disorder Diagnosis, 2016–2022

	Age 18–64 (n=116,226)		Age 65+ (n=264,857)
	N	%	N	%
Age
18–34	10,911	9.4%	--	--
35–44	19,164	16.5%	--	--
45–54	31,930	27.5%	--	--
55–64	54,221	46.7%	--	--
65–74	--	--	158,893	59.9%
75–84	--	--	80,572	30.4%
85+	--	--	25,392	9.6%
Race & Ethnicity
Black	20,426	17.6%	16,695	6.3%
Hispanic	3,695	3.2%	1,721	0.6%
Asian	672	0.6%	1,228	0.5%
North American Native	2,618	2.3%	2,454	0.9%
White	86,674	74.6%	238,175	89.9%
Other	604	0.5%	1,421	0.5%
Unknown	1,537	1.3%	3,163	1.2%
Dually eligible for Medicaid	53,603	46.1%	25,502	9.6%
Rurality
Rural	4,835	4.2%	10,731	4.1%
Micropolitan	23,641	20.3%	43,512	16.4%
Metropolitan	87,750	75.5%	210,614	79.5%
Index diagnosis (Dx) of moderate/severe alcohol use disorder	52,492	45.2%	123,166	46.5%
Mental health Dx	90,470	77.8%	131,569	49.7%
Acute alcohol-related medical condition Dx	2,244	1.9%	3,242	1.2%
Chronic alcohol-related medical condition Dx	13,205	11.4%	25,723	9.7%
Substance use disorder Dx other than alcohol use disorder	41,791	36.0%	19,064	7.2%
Setting of index alcohol use disorder Dx
Inpatient hospital	47,638	41.0%	116,890	44.1%
Emergency department	32,465	27.9%	60,890	23.0%
Primary care physician	10,771	9.3%	36,249	13.7%
Behavioral health specialist	6,200	5.3%	4,673	1.8%
APRN/PA	3,995	3.4%	8,907	3.4%
Gastroenterologist	1,667	1.5%	4,423	1.7%
OB/GYN	408	0.4%	656	0.2%
Cardiologist	410	0.4%	2,363	0.9%
Other	12,672	10.9%	29,806	11.3%
Index alcohol use disorder visit via telemedicine	4,656	4.0%	8,414	3.2%

Note: All co-morbidity variables reflect diagnoses in the 12-month period before the index alcohol use disorder diagnosis. APRN refers to advanced practice registered nurse. PA refers to physician assistant. OB/GYN is obstetrician/gynecologist. Behavioral health specialists include psychiatrists, addiction medicine physicians, psychologists, and social workers. The other setting category included unknown physical specialty, clinical laboratory, and other types of specialists, including general surgery, ambulance services, orthopedic surgery, neurology, among others, The proportion of index visits via telemedicine covers the entire study period, although telemedicine was only commonly used after the onset of the COVID-19 pandemic (2020–2022).

Approximately two-thirds of women in both age groups received their index diagnosis in an acute care setting (e.g., for 18–64, 41.0% inpatient and 27.9% ED; for 65+, 44.1% inpatient and 23.0% ED) (Table 1). Among diagnoses occurring during a hospitalization, alcohol use disorder was the principal discharge diagnosis for only a small proportion (8.5% for women 18–64 and 5.5% for women 65+) (Supplement). Among women 18–64 whose index diagnosis occurred during a hospitalization, the principal discharge diagnosis was mental health-related for 27.7%, a medical condition 100% attributable to alcohol use for 4.1%, a medical condition linked to but <100% attributable to alcohol for 9.5%, a condition for which alcohol use carries greater risks for 10.6%, and another principal diagnosis for 36.3%. In contrast, among women 65+, the principal discharge diagnosis for the hospitalization was mental health for only 4.2%, a medical condition 100% attributable to alcohol for 3.0%, a medical condition linked to but <100% attributable to alcohol for 17.5%, a condition for which alcohol use carries greater risks for 15.1%, and another principal diagnosis for 54.3%.

Approximately 9.3% of index diagnoses for women 18–64 and 13.7% for women 65+ were made in outpatient settings by primary care physicians. Only 5.3% of diagnoses for women 18–64 and 1.8% of those 65+ were made by behavioral health specialists, and less than 1% were made by obstetricians/gynecologists. A small proportion of women’s index diagnoses over the study period were made via telemedicine (4.0% among 18–64-year-olds and 3.2% among 65+).

Discussion

Approximately two-thirds of new diagnoses of alcohol use disorder among women in Medicare occur in inpatient hospitals and EDs. The fact that alcohol use disorder was not the principal discharge diagnosis for most of these hospitalizations suggests that it was identified during the treatment of another condition, which could have implications for follow-up care (because the focus of instructions for discharge and follow-up care would likely be the other condition). Particularly for women 18–64 years, the principal discharge diagnosis was often for conditions that are commonly comorbid with or related to alcohol, including alcohol-related medical conditions, other substance use disorders, and mental health disorders. Only 10–14% of index diagnoses were made by PCPs and 2–5% by behavioral health specialists. This points to the need to improve recognition and diagnosis in outpatient settings. Earlier identification and management of alcohol use disorder in outpatient settings could help to avert acute care utilization for the serious conditions for which many beneficiaries in our sample were hospitalized.

Findings should be considered in the context of study limitations. Our study did not include Medicare Advantage beneficiaries. Although about half of beneficiaries 65 and older are now enrolled in these private plans, this is less common among beneficiaries 18–64, which would have impeded our cross-age group comparison. It is possible that individuals in our study were diagnosed with alcohol use disorder prior to our 12-month washout period. We cannot observe diagnoses within encounters not reimbursed by Medicare during the study period or for which a diagnosis is not recorded on the claim. The COVID-19 pandemic occurred during our study period and may have affected engagement with primary preventative care during this period. With these analyses we are unable to disentangle causal factors for these patterns or fully explore disparities across groups, necessitating additional research.

Conclusions

As alcohol-related mortality rates increase among women, most receive an initial diagnosis only in acute care settings, and often in the context of treatment for alcohol-related or commonly comorbid conditions. Ongoing efforts have focused on increasing screening for alcohol use disorder in primary care and, to a lesser extent, specialty care settings commonly used by these patients (e.g., hepatology¹⁴). Our findings suggest the need for enhanced efforts to encourage early identification and treatment outside of acute care settings to diagnose women earlier and engage them in community-based care. Earlier diagnosis and treatment could stem the rapid rise in alcohol-related morbidity and mortality among women in the U.S.

Highlights:

• Given increasing rates of alcohol use disorder and alcohol-related medical conditions coupled with low rates of alcohol treatment among women, improving early identification and treatment engagement in this population is a public health priority.

• Two-thirds of Medicare-enrolled women newly-diagnosed with alcohol use disorder were diagnosed in an inpatient hospital or emergency department; 10–14% occurred during an outpatient visit with a primary care physician and 2–5% during a visit with a behavioral health specialist.

• Enhanced efforts to encourage early identification and treatment outside of acute care settings to diagnose women earlier and engage them in community-based care are needed.