Alcohol Use Among Older Adults and Health Care Utilization

Abstract

Objective:

Examinations of the association between health care utilization and levels of alcohol use are lacking in nationally representative samples of older adults. The present study set out to fill this gap by demonstrating how health care utilization concepts are associated with alcohol use among older adults in the United States.

Method:

Cross-sectional data from 11 years of the National Health and Interview Survey were used to examine prevalence and rates of alcohol use among older adults (n = 106,511) and associations with demographic variables and health care use.

Results:

About 70% of older adults (aged 65+; mean age = 74.1, SD = 0.04) had drunk alcohol in their lifetime, and 15.8% were current moderate or heavy drinkers. Results of a multinomial logistic regression revealed that individuals with any alcohol use in their lifetime had more recent health care visits and fewer ER/ED visits, and current moderate users had fewer office visits than abstainers, controlling for sex, race, educational attainment, marital status, and concurrent tobacco use.

Conclusion:

Older adults who have any history of alcohol use are more likely than abstainers to have had recent health care visits, more office visits, and less likely to have had an emergency department visit.

Older adult substance abuse is a growing problem in the United States [1,2] and around the world [3] with significant variations between regions [4]. Alcohol abuse in particular has implications for the health care system, since physiological changes due to the aging process (i.e., decreased blood volume, among other factors) make older adults more sensitive to the risks of alcohol [5,6], which can lead to increased and exacerbated health problems, including falls, that require increases in health care use [7–10]. To design effective assessment tools and intervention plans, it is important to understand whether associations exist between older adults use of health care services and levels of alcohol use.

Alcohol Use among Older Adults

Prior to the current generation of older adults, older adults had low rates of substance use and these rates decreased as they aged [1]. The current generation of older adults have higher rates of substance use than previous generations, in part due to more tolerant societal views of alcohol and recreational drug use as they came of age [6,11]. Results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) found that alcohol was the most commonly used substance among older adults in the United States [12,13] with over half of older adults sampled in a national survey using alcohol in the previous 12 months. Other studies have found that over half of older adults have used alcohol in the past 3 months, [5] and in recent years a 19.2% increase in binge drinking was reported among this population [14,15]. Given the increased rate of alcohol use among older adults, it is expected that instances of alcohol use disorder (AUD) in older adults will drastically increase as the number of people aged 65 and over in the United States grows fivefold by 2030 [11] and, globally, reaches 2.1 billion individuals by 2050 [16].

Sensitivity to alcohol increases with age, which can lead to increased medical problems related to drinking [5]. Decreases in lean body mass, volume of body water, blood flow, and liver function are observed with age [17] and lead to an increased blood alcohol concentration and higher sensitivity to the effects of alcohol consumption. Alcohol use can also increase mortality risks in older adults, particularly males, and especially those diagnosed with certain physical and psychological comorbidities, such as gout or anxiety, for which they may be taking medications that mix poorly with alcohol [18]. Additionally, frequent alcohol use is associated with numerous health problems, including liver disease, heart disease, memory problems, cancers, and falls [19] that may require increased utilization of health care.

Health Care Utilization & Substance Use Assessment Among Older Adults

The utilization of health care services [20] increases with age as an individual’s health deteriorates and requires greater medical attention. Rates of health care utilization [20] tend to be highest among older adults compared with other age groups [21,22], and substance use affects when and how these individuals access the health care system. Data from the National Health Interview Survey (NHIS) show that the group of older adults with the most emergency department visits tends to be former drinkers or current heavy drinkers who also have chronic health conditions [5]. Yet current heavy alcohol use among older adults is associated with low primary and outpatient care, with abstainers and former drinkers having a higher probability of non-emergency health care utilization [23].

Demographic Differences

There exist important differences in rates of substance use among older adults due to demographic characteristics such as sex and race/ethnicity. Existing data indicate that males have higher rates of alcohol use [24,25] across the globe and that alcohol use disorders contribute to the global burden of disease for men [4]. Males drink twice as much alcohol on a weekly basis than females [26] and according to global estimates are four times more likely than females to have had alcohol in the past month [27]. Regarding alcohol use among older adults of racial/ethnic minority groups in the United States, African Americans have lower rates of alcohol use compared with non-Hispanic white Americans, but they have higher rates of alcohol-related problems [28,29]. Among adults aged 65 and older, Asian/Pacific Islanders are less likely to have lifetime AUD compared to whites [30].

Current Study

The current population of older adults exhibit different alcohol use patterns than previous generations requiring additional health care services. As this group ages they may be more sensitive to alcohol and might be at increased risk of alcohol use problems due to underlying health conditions and medication use, it is therefore important to examine associations of health care utilization and alcohol use in this age group. It is also important to understand how alcohol use and health care utilization rates differ based on demographics, including sex, race, education and marital status, considering the differences observed in previous studies. The current study uses data from several consecutive years of the National Health and Interview Study (2006 – 2016) [24] to examine the associations between health care utilization and levels of older adult alcohol use.

Method

Sample

Data for this study came from the years 2006 through 2016 of the National Health Interview Survey (NHIS). The National Health Interview Survey is carried out annually among the U.S. civilian, non-institutionalized population through personal interviews conducted in U.S. households to monitor the health of the country’s population. The goal of the NHIS is to monitor the health of the U.S. population by collecting and analyzing data on a variety of health topics. People in long-term care institutions and correctional facilities were excluded from the study. Out of the over 1 million participants who completed the NHIS surveys between 2006 and 2016, we retained only individuals who reported their age as 65 or older. In total, 138,676 participants who reported being aged 65 and older were included in the analytic sample.

Measures

To assess alcohol use among older adults and examine how health care utilization, in addition to demographic factors, are associated with alcohol use in this population, the following variables were retained for each of the years included.

Variables

Alcohol Use.

The NHIS survey included several questions regarding alcohol use. The categories of the original alcohol use variable¹ were recoded into the following four categories: “Lifetime abstainer”, “Former user”, “Current infrequent”, and “Current moderate”. Individuals who indicated not having had any alcohol in the past year in the original alcohol use variable (“Former infrequent”, “Former regular”, “Former unknown frequency”) were categorized as “Former user”. Individuals who in the original alcohol use variable selected “Current infrequent” or “Current light” were categorized as “Current infrequent”. All other individuals who reported some type of alcohol use in the previous year (i.e., 3–14 drinks per week for males and 3–7 drinks per week for females) were categorized in the “Current moderate” group.

Health Care Utilization.

We measured three aspects of each participant’s health care utilization by using three separate questions included in the NHIS survey.

Recency of Health Visit.

The first variable came from a survey question that asked participants to indicate the time since they had last seen or talked to a health professional. Participants selected their answers from the following categories; “Never”, “6 months or less”, “More than 6 months, but not more than 1 year ago”, “More than 1 year, but not more than 2 years ago”, “More than 2 years, but not more than 5 years ago”, and “More than 5 years ago”.

Office Visits Past Year.

Another variable asked participants to select the number of health care office visits they had in the past 12 months. Participants selected from the following categories: “None”, “1”, “2–3”, “4–5”, “6–7”, “8–9”, “10–12”, “13–15”, and “16 or more”.

ER/ED Visits Past Year.

Participants were also asked to indicate the number of ER or ED visits they had incurred in the previous year. Participants selected their answers from the following categories; “None”, “1”, “2–3”, “4–5”, “6–7”, “8–9”, “10–12”, “13–15”, and “16 or more”.

Sex.

Participants reported their sex as either “Male” or “Female”.

Age.

Participants were asked to report their age at the time of the interview. To protect the identity of participants over age 85, anyone aged 85 or older is grouped together in one category, and thus the maximum age in this sample is 85 even though there may be older individuals.

Race.

The NHIS asked participants to self-report their race and ethnicity in two separate questions. In one question participants were asked to report their race by selecting from the following categories: “Black/African American”, “Indian (American), Alaska Native”, “Asian Indian”, “Chinese”, “Filipino”, “Other Asian”, and “Multiple race, no primary race selected”. In a second question, participants were asked to report whether they were of Hispanic ancestry. A new race/ethnicity variable was created out of the three separate variables from the NHIS. The categories in the new variable are “White”, “Black/African American”, “Asian/American Indian and Alaskan Native (AIAN)”, and “Hispanic”.

Education.

Participants were asked to record the highest level of education they had completed from 22 available responses.² In keeping with past literature, we recoded the education variable to reduce it to four categories: “Less than a high school diploma”, “High School Diploma or GED”, “Some college”, and “Bachelor’s degree or higher”.

Marital Status.

The survey asked all participants about their marital status by selecting from the following categories: “Married – spouse in household”, “Married – spouse not in household”, “Married – souse in household unknown”, “Widowed”, “Divorced”, “Separated”, “Never Married”, and “Living with partner”. We recoded the marital status variable to reduce it to four categories: “Married/Cohabiting”, “Widowed”, “Divorced/Separated”, and “Never married”.

Living alone.

To assess whether participants lived alone, a variable measuring family structure from the NHIS survey was used. A new variable was created to show if a participant lived alone or not.

Tobacco Use.

To assess tobacco use, participants were first asked if they had smoked 100 cigarettes in their entire lives. If they answered “Yes” they were subsequently asked to indicate whether they smoked tobacco “Every day”, “Some days”, or “Not at all”. We created a new variable that collapsed the two variables and included individuals who had never smoked. The new variable had the following categories: “Lifetime abstainer”, “Former smoker”, “Some days”, and “Every day”.

Data Analysis

The dependent variable, alcohol use in the previous year, was an ordered categorical variable and, as a result, multinomial logistic regression was used to examine associations between covariates and each category of the dependent variable. Stata Version 14 [31] was used to analyze the data.

Results

Out of the 138,676 participants who participated in the NHIS survey years 2006–2016 and were 65 and older, 106,511 participants with complete data on all variables were included in the analysis. Table 1 shows descriptive statistics for the participants included in the analysis. Slightly more than half of the sample was female (56.9%) with an average age of 74.1 (SE = 0.04). A little over half of the participants were either lifetime abstainers of alcohol (26.8%) or former alcohol users (25.2%), with current light users making up 32.2% and current moderate users accounting for 15.8% of the sample. Most participants had not visited an ED in the past 12 months (77.3%). The remainder reported one ED visit (14.7%) or 2–3 visits (6.1%). Regarding health care office visits, 6.9% of participants had no office visits in the past 12 months, 26.1% had 2–3 visits, 20.7% had 4–5 visits, and 7.9% had 16 office visits or more.

Table 1.

Descriptive Statistics of Key Variables (N = 106,511)

Measures		Percent/M (SE)
Demographics
Female		56.9%
Age (years)		74.1 (0.04)
Race
	White	78.8%
	Black	9.3%
	Asian/AIAN	4.2%
	Hispanic	7.8%
Education
	Less than High School diploma	21.5%
	High School diploma	31.2%
	Some college	23.9%
	Bachelor’s or higher	23.5%
Marital Status
	Married/Cohabiting	56.6%
	Widowed	27.6%
	Divorced/Separated	11.7%
	Never married	4.1%
Lives Alone		31.6%
Substance Use
Alcohol Use
	Lifetime abstainer	26.8%
	Former user	25.2%
	Current infrequent/light users	32.2%
	Current moderate/heavy users	15.8%
Tobacco Use
	Lifetime abstainer	51.6%
	Former Smoker	38.9%
	Some days	1.7%
	Every day	7.8%
Health Care Utilization
Recency of Health Visit
	Never	0.6%
	6 mo or <	86.1%
	7 mo to 1 yr	8.1%
>1 yr to 2 yrs		2.6%
>2 yrs to 5 yrs		1.4%
	>5 yrs	1.2%
Office Visits Past Year
	None	6.9%
	1	10.6%
	2–3	26.1%
	4–5	20.7%
	6–7	10.4%
	8–9	5.8%
	10–12	8.9%
	13–15	3.0%
	16 or more	7.9%
ER/ED Visits Past Year
	None	77.3%
	1	14.7%
	2–3	6.1%
	4–5	1.2%
	6–7	0.4%
	8–9	0.1%
	10–12	0.1%
	13–15	0.0%
	16 or more	0.1%

Note: AIAN = American Indian and Alaska Native.

The results of the multinomial logistic regression (presented in Table 2) indicated an association between alcohol use and health care utilization when controlling for important demographic characteristics and tobacco use comorbidity.³ Compared to lifetime abstainers, all individuals who had any experience with alcohol in their lifetime (former, current light, and current moderate) reported more recent health care visits. Individuals currently using alcohol at light (RRR = 0.84, p < 0.0001) or moderate levels (RRR = 0.79, p < 0.0001) had fewer ER/ED visits than lifetime abstainers. Former users had more office visits (RRR = 1.07, p < 0.0001), while current moderate users had fewer office visits (RRR = 0.97, p < 0.0001) relative to lifetime abstainers.

Table 2.

Multinomial regression results examining associations between study variables and alcohol use (n = 106,511)^*

Reference = Lifetime Abstainer	Former User				Current Light User				Current Moderate User
	RRR (SE)		95% CI	p	RRR (SE)		95% CI	p	RRR (SE)		95% CI	p
Health Care Utilization
Recency of Health Visit (Ref = Never)
6 mo or <	2.86	(0.57)	(1.94, 4.22)	0.000	2.35	(0.36)	(1.75, 3.17)	0.000	2.92	(0.72)	(1.81, 4.73)	0.000
7 mo to 1 yr	2.60	(0.52)	(1.75, 3.86)	0.000	2.20	(0.34)	(1.62, 2.98)	0.000	2.74	(0.67)	(1.69, 4.42)	0.000
> 1 yr to 2 yrs	2.82	(0.58)	(1.88, 4.24)	0.000	2.22	(0.37)	(1.60, 3.08)	0.000	2.74	(0.72)	(1.64, 4.58)	0.000
to 2 yrs to 5 yrs	2.98	(0.68)	(1.90, 4.68)	0.000	2.20	(0.41)	(1.53, 3.17)	0.000	3.38	(0.95)	(1.95, 5.86)	0.000
>5 yrs	2.93	(0.68)	(1.86, 4.62)	0.000	1.74	(0.36)	(1.16, 2.61)	0.008	2.61	(0.76)	(1.47, 4.63)	0.001
Office Visits Past Year (Ref = None)
1	1.24	(0.11)	(1.04, 1.47)	0.015	1.62	(0.14)	(1.36, 1.92)	0.000	1.46	(0.16)	(1.19, 1.80)	0.000
2–3	1.22	(0.10)	(1.04, 1.43)	0.013	1.64	(0.13)	(1.40, 1.91)	0.000	1.46	(0.15)	(1.19, 1.79)	0.000
4–5	1.43	(0.11)	(1.22, 1.66)	0.000	1.57	(0.12)	(1.34, 1.83)	0.000	1.35	(0.14)	(1.10, 1.66)	0.005
6–7	1.57	(0.13)	(1.33, 1.85)	0.000	1.68	(0.14)	(1.42, 1.98)	0.000	1.37	(0.16)	(1.09, 1.72)	0.007
8–9	1.66	(0.15)	(1.39, 1.98)	0.000	1.75	(0.15)	(1.47, 2.07)	0.000	1.32	(0.15)	(1.05, 1.66)	0.018
10–12	1.59	(0.13)	(1.34, 1.88)	0.000	1.54	(0.13)	(1.30, 1.83)	0.000	1.17	(0.13)	(0.94, 1.47)	ns
13–15	1.60	(0.17)	(1.30, 1.97)	0.000	1.47	(0.16)	(1.19, 1.82)	0.000	1.09	(0.16)	(0.82, 1.45)	ns
16 or more	1.84	(0.16)	(1.55, 2.17)	0.000	1.44	(0.13)	(1.20, 1.73)	0.000	1.12	(0.13)	(0.89, 1.41)	ns
ER/ED Visits Past Year (Ref = None)
1	1.08	(0.04)	(1.00, 1.16)	0.044	0.94	(0.03)	(0.88, 1.01)	0.960	0.86	(0.04)	(0.78, 0.94)	0.002
2–3	1.03	(0.06)	(0.92, 1.16)	ns	0.77	(0.05)	(0.68, 0.86)	0.000	0.64	(0.05)	(0.54, 0.75)	0.000
4–5	0.82	(0.10)	(0.65, 1.04)	ns	0.51	(0.06)	(0.40, 0.66)	0.000	0.47	(0.08)	(0.34, 0.66)	0.000
6–7	0.96	(0.18)	(0.66, 1.40)	ns	0.42	(0.10)	(0.27, 0.66)	0.000	0.47	(0.20)	(0.21, 1.07)	ns
8–9	0.58	(0.17)	(0.32, 1.04)	ns	0.25	(0.10)	(0.12, 0.54)	0.000	0.09	(0.05)	(0.03, 0.27)	0.000
10–12	0.33	(0.12)	(0.16, 0.66)	0.002	0.16	(0.07)	(0.07, 0.37)	0.000	0.14	(0.08)	(0.05, 0.42)	0.000
13–15	0.98	(0.55)	(0.32, 2.96)	ns	0.31	(0.23)	(0.07, 1.34)	ns	0.21	(0.21)	(0.03, 1.57)	ns
16 or more	0.53	(0.22)	(0.23, 1.21)	ns	0.45	(0.22)	(0.18, 1.17)	ns	0.63	(0.38)	(0.20, 2.03)	ns
Tobacco Use
Tobacco Use (Ref = Lifetime abstainer)
Former Smoker	5.08	(0.15)	(4.78, 5.39)	0.000	4.00	(0.12)	(3.77, 4.24)	0.000	7.52	(0.29)	(6.98, 8.10)	0.000
Some days	3.61	(0.40)	(2.91, 4.48)	0.000	3.91	(0.39)	(3.22, 4.75)	0.000	6.19	(0.76)	(4.87, 7.87)	0.000
Every day	4.12	(0.24)	(3.57, 4.63)	0.000	3.02	(0.18)	(2.68, 3.40)	0.000	7.02	(0.43)	(6.23, 7.91)	0.000
Demographics
Sex (Ref = Male)	0.55	(0.01)	(0.52, 0.58)	0.000	0.63	(0.01)	(0.60, 0.66)	0.000	0.40	(0.01)	(0.38, 0.42)	0.000
Age	0.99	(0.00)	(0.99, 1.00)	0.003	0.96	(0.00)	(0.96, 0.97)	0.000	0.97	(0.00)	(0.96, 0.97)	0.000
Race (Ref = White)
Black	0.85	(0.04)	(0.78, 0.92)	0.000	0.51	(0.02)	(0.47, 0.56)	0.000	0.32	(0.02)	(0.28, 0.36)	0.000
Asian/AIAN	0.47	(0.03)	(0.42, 0.53)	0.000	0.31	(0.02)	(0.28, 0.35)	0.000	0.19	(0.02)	(0.16, 0.22)	0.000
Hispanic	0.67	(0.04)	(0.59 0.76)	0.000	0.68	(0.04)	(0.61, 0.76)	0.000	0.48	(0.04)	(0.41, 0.56)	0.000
Education (Ref = Less than HS)
High school diploma	1.10	(0.03)	(1.04, 1.17)	0.002	1.75	(0.06)	(1.64, 1.87)	0.000	1.82	(0.08)	(1.67, 1.99)	0.000
Some college	1.22	(0.04)	(1.14, 1.30)	0.000	2.48	(0.09)	(2.31, 2.66)	0.000	3.00	(0.14)	(2.74, 3.28)	0.000
Bachelor’s or higher	1.05	(0.04)	(0.97, 1.14)	ns	3.37	(0.14)	(3.10, 3.67)	0.000	5.76	(0.30)	(5.19, 6.39)	0.000
Marital Status (Ref = Married/Cohabiting)
Widowed	1.02	(0.05)	(0.93, 1.11)	ns	0.85	(0.04)	(0.78, 0.93)	0.001	0.65	(0.04)	(0.57, 0.73)	0.000
Divorced/Separated	1.45	(0.08)	(1.30, 1.60)	0.000	1.08	(0.06)	(0.97, 1.19)	ns	0.85	(0.06)	(0.75, 0.98)	0.020
Never married	1.13	(0.08)	(1.00, 1.30)	0.076	0.83	(0.06)	(0.72, 0.95)	0.008	0.66	(0.06)	(0.56, 0.78)	0.000
Lives Alone	0.99	(0.04)	(0.91, 1.07)	ns	1.07	(0.05)	(0.98, 1.17)	ns	1.26	(0.07)	(1.13, 1.40)	0.000

^*Note:Analyses control for the complex sampling design of the NHIS; Lifetime abstainer outcome category is the reference; Analyses control for the year of NHIS survey that the data came from. AIAN = American Indian and Alaska Native.

Women in this sample were less likely than men to use alcohol at light levels (RRR = 0.63, p < 0.0001) and at moderate levels (RRR = 0.40, p > 0.0001). In nearly all cases, individuals of racial/ethnic groups other than white were significantly less likely to be in any of the alcohol use categories (former, current light, or current moderate). For instance, African American, Asian, and Hispanic older adults were less likely to use alcohol at light levels (RRR = 0.51, p < 0.0001; RRR = 0.31, p < 0.0001; RRR = 0.68, p < 0.0001, respectively) and less likely to use alcohol at moderate levels (RRR = 0.32, p < 0.0001; RRR = 0.19, p < 0.0001; RRR = 0.48, p < 0.0001, respectively).

Regarding the co-use of alcohol and tobacco, we found that individuals who reported smoking on some days or every day had a greater likelihood of being a current light and current moderate alcohol user than lifetime abstainers. For example, individuals who reported smoking on some days had a nearly fourfold likelihood of being a current light alcohol user (RRR = 3.01, p < 0.0001) while individuals who reported smoking everyday had a seven-fold likelihood of being a current moderate alcohol user (RRR = 7.02, p < 0.0001).

Discussion

Considering the rate at which the older population is growing and the cohort-specific exposures that have led to greater levels of substance use, it is important to consider how health care utilization is associated with older adults’ use of alcohol. The results of the present study show significant associations between greater alcohol use and recency of health care visits, even after we control for important demographic differences and tobacco use comorbidities. The present study’s findings show that the majority of older adults have at least one office visit a year, with many having had more. This suggests that older adults, including moderate alcohol users, already have contact with health care professionals creating opportunities for screening and brief intervention for alcohol use and possibly for discussions about referral and other treatment options.

Despite the potential for screening and brief intervention, past research suggests that physicians are less likely to assess older adult patients for alcohol use, in part due to ageist assumptions that older adults do not have substance use problems [6,30,32]. Screening all older adults would help alleviate this bias. Alcohol use screening tools for older adults, such as the Short Michigan Alcoholism Screening Instrument – Geriatric Version (SMAST-G) do exist and are advised for use with older adults [33]. Follow up research on the frequency with which health care professionals use screening tools and discuss alcohol use with patients is needed to learn about the acceptability and effectiveness of screening tools among older adults.

Much of the previous research in this area has focused on older adult alcohol use and ED utilization [5]. The present results extend previous research by focusing on office visits in addition to ED visits and show that the pattern of health care utilization depends on alcohol use. Compared to abstainers, individuals with any history of alcohol use in this study reported fewer ED visits overall but more office visits in the past 12 months. Our current research cannot identify whether the health care visits were for the treatment of alcohol use or for alcohol-related problems and conditions or for health maintenance, but future studies may want to explore these issues further.

Our results examining sex differences are consistent with earlier research showing that older men have higher rates of alcohol consumption than older women [26]. Women have a lower likelihood of belonging to any of the alcohol use categories than men perhaps due to some lingering cohort effects regarding societal restrictions on alcohol use [34]. Our results further showed that individuals from minority racial and ethnic groups were less likely to use alcohol than whites as previous studies have found. This may be due to protective factors such as having greater access to social and community supports as has been previously shown for physical activity [35], hypertension [36] and marijuana use [37]. Previous studies further suggest that higher abstinence rates among racial/ethnic minorities might arise from experiencing greater negative social consequences from using alcohol [38]. Studies in this area are scant, so further research should be conducted to examine whether there are interactions between race and alcohol use and explore whether the associations between health care utilization and alcohol use depend on race/ethnicity and whether these relationships are influenced by social disadvantage [39].

Limitations

Our findings regarding the ways in which health care utilization and alcohol use are related among older adults have to be taken with several limitations in mind. First, the NHIS data is limited to older adults in the United Sates not in long-term care facilities, our results are limited to older adults who are healthier than those requiring long-term health care and thus do not apply to all older adults. This study also did not control for potential covariates, such as comorbid diagnoses, number of health problems, and number of years drinking and could not determine whether health care visits where due to alcohol use treatment. In addition, the NHIS data are cross-sectional, prohibiting us from determining whether there are changes over time in alcohol use due to changes in health.

Relatedly, there may be a bidirectional relationship between health care utilization and alcohol use such that alcohol use reduces engagement with the health care system or, alternatively, increased health problems due to decreased engagement from the health care system may lead individuals to self-medicate by using alcohol. The NHIS dataset is not able to distinguish the direction of causation between these two important health behaviors among older adults because it is not a longitudinal study. However, future studies may be designed with untangling this comorbidity in mind.

Notwithstanding the above limitations, our findings using nationally representative data of older adults extend current knowledge of how health care utilization is associated with alcohol use. Older adults need additional medical attention, especially if they are using alcohol in ways that may interfere with medications and with their safety. Research attention should continue to focus on this population, as with increasing alcohol use there may be a tendency to not seek medical attention and therefore exacerbate underlying health conditions.