|Home | About | Journals | Submit | Contact Us | Français|
Most individuals with alcohol use disorders receive no treatment for their disorder. Past research suggests that a major reason for this is that individuals with alcohol use disorders do not perceive a need for treatment. The current research had two objectives: (i) to provide updated estimates of the percentage of individuals with alcohol use disorders who perceived a need for treatment, and, among those, the percentage who received any alcohol use disorder treatment; and (ii) to investigate the determinants of perceived need for and utilization of alcohol use disorder treatment.
Secondary data analysis of two national surveys, the National Epidemiologic Study on Alcohol and Related Conditions (NESARC, n=3,305 individuals with alcohol use disorders) and the National Survey of Drug Use and Health (NSDUH, n=7,009 individuals with alcohol use disorders).
In both surveys fewer than 1 in 9 individuals with an alcohol use disorder perceived a need for treatment. In predicting perceived need, the explanatory power of diagnostic variables was much greater than that of demographic variables. Among those with perceived need, 2 in 3 reported receiving treatment in the past year.
Our results suggest that failure to perceive need continues to be the major reason individuals with alcohol use disorders do not receive treatment. On the other hand, among those with perceived need, the majority receive treatment. It is likely that high levels of unmet need for alcohol use disorder services will persist as long as perceived need is low. Efforts are needed to increase levels of perceived need among those with alcohol use disorders.
Alcohol use disorders (abuse and dependence) are common, occurring in 4 to 9% of the U.S. population in a given year (1-4) and cause substantial morbidity (5) accounting for about five percent of all disability (6). The negative social and health consequences associated with alcohol use disorders are protean (7) and include increased suicidal behaviors, (8,9) high rates of criminal justice involvement and violence, (10) and substantial medical/physical consequences (7,11). The medical consequences of alcohol use disorders, such as cirrhosis and premature death, are particularly high among Hispanics, Native Americans, and African Americans compared to whites (12,13).
Despite widespread public skepticism regarding alcohol use disorder services, there are effective, evidence-based psychosocial and pharmacological treatments for alcohol use disorders. These include brief primary care interventions and interventions based on motivational interviewing (14-19). As with other areas in medicine, actual treatment may not be high quality or guideline concordant (20,21). These alcohol use disorder treatments are as effective as those for other chronic conditions, including heart disease, asthma, and diabetes (22), although many individuals experience remission of their alcohol use disorder without formal treatment. It has been suggested in an Institute of Medicine monograph that improving the quality of treatment for alcohol use disorders and other behavioral health disorders would decrease the mortality, morbidity, and societal costs of these disorders (20).
However, a large majority of individuals with alcohol use disorders receive no treatment for their disorder (2). A major reason for this may be that individuals with alcohol use disorders do not perceive a need for treatment; (23-25) in the 1992 National Longitudinal Alcohol Epidemiologic Survey (NLAES) only 12.7% of persons with alcohol use disorders perceived a need for treatment (23). This is not surprising, given that a hallmark of alcohol use disorders is denial.
Thus, perceiving a need for alcohol use disorder treatment is likely often a rate-limiting step in receiving treatment. Further, perceived need is potentially modifiable through educational efforts, which could occur anywhere from the clinician-patient interaction to population level-media efforts, as has occurred with direct-to-consumer advertising for antidepressants.
While there has been considerable research in the past 15 years to improve detection tools for alcohol use and develop brief primary care interventions for alcohol use disorders (14-18), we do not know whether rates of perceived need for alcohol use disorder treatment are increasing over time. In addition, the determinants of perceiving a need for treatment are not well understood; if we better understood the determinants we could possibly use this information to design educational efforts to enhance awareness of alcohol use disorders and increase perceived need for alcohol use disorder services in at-risk populations. Using data from the National Comorbidity Survey (NCS), Mojtabai found that men, married individuals, the uninsured, younger individuals (ages 15 to 24) and those with less severe psychiatric illness were less likely to perceive need for substance use disorder and/or mental health services (26). In our analysis of Health Care for Communities (HCC) data, men, the elderly, the less educated, and those who were married were less likely to perceive a need for any substance use disorder/mental health services (27). However, since the determinants of need for any substance use disorder or mental health disorder service was studied as a single outcome in both of these studies, and since perceived need for mental health treatment is greater than the perceived need for substance use disorders treatment, the extent to which these results can be extrapolated to perceived need for substance use disorders or alcohol use disorder treatment is unknown.
In this paper we had two objectives: (i) to provide updated estimates of the percentage of individuals with alcohol use disorders who perceived a need for treatment, and, among those, the percentage who actually utilized any alcohol use disorder treatment; and (ii) to investigate the determinants of perceived need for and utilization of alcohol use disorder services. While past studies have looked at substance use disorders/mental health disorders as determinants of need, we were particularly interested in conducting a more detailed analysis of the effects of individual alcohol use disorder symptoms, as we were interested in whether certain symptoms made individuals with alcohol use disorders more especially likely to perceive a need for treatment. Further, we were interested in comparing the overall explanatory power of sociodemographic factors versus clinical factors in explaining perceived need.
We used two large national surveys, the National Epidemiologic Study on Alcohol and Related Conditions (NESARC) and the National Survey of Drug Use and Health (NSDUH), to investigate these issues. Several characteristics of these studies make them ideal for our purposes. First, investigating the determinants of perceived need for alcohol use disorder services requires large samples, as perceived need for alcohol use disorder services is an infrequent event in the general population: alcohol use disorders occur in about 5 to 10% of the population, and in past studies, among individuals with alcohol use disorder, perhaps 10 to 15% perceive a need for treatment. Second, they are nationally-representative. Third, they have detailed measures of clinical need, along with sociodemographic measures.
The National Epidemiologic Survey on Alcohol and Related Conditions was conducted by NIAAA in 2001−2002 to provide data for the adult U.S. population on alcohol and drug use, abuse and dependence, and associated psychiatric and physical comorbidities (2). Potential respondents were selected by multi-stage probability sampling from the Census 2000/2001 Supplementary Survey and the Census 2000 Group Quarters Inventory. NESARC had a sample size of 43,093 individuals in private residences and certain group quarters housing with a response rate of 81%. NESARC oversampled Hispanics, non-Hispanic blacks, and younger adults (age 18 to 24). Face-to face interviews were conducted by trained lay interviewers from the Census Bureau.
The National Survey on Drug Use and Health is an annual survey sponsored by the Substance Abuse and Mental Health Services Administration (SAMHSA) (28-30) to provide national data on the incidence and prevalence of illicit drug, alcohol, and tobacco use. Each year roughly 80,000 individuals are selected by multistage probability sampling to be representative of the U.S. civilian, non-institutionalized population aged 12 and older. Interviews are face-to-face. Data from the 2004 and 2005 surveys were used for this study.
The response rate for the 2004 survey was 75%, with a total sample of 67,760 individuals. In 2004, a split sample design was implemented, where adult respondents were divided into two samples. Adults in Sample A were administered the Adult Mental Health Module, but not the Adult Depression Module. Adults in Sample B were administered the Adult Depression Module but only six core questions from the Adult Mental Health module. For this study, sample B was used. The overall response rate for the 2005 sample was 74.4 percent, with a total sample of 68,308.
As we were interested in investigating the determinants of perceived need for alcohol treatment services in adults, we limited our samples to those adults (age 18 and older) who met criteria for alcohol abuse or dependence in the past year (n=3,305 for NESARC, and n=7,009 for NSDUH). The samples are described in Tables 1 and and2.2. The study was approved by the University of Arkansas for Medical Sciences Institutional Review Board.
In NESARC, a respondent was classified as having 12-month perceived need for alcohol treatment if he or she either (i) reported thinking he/she should have received treatment but did not go or (ii) reported receiving treatment in the past 12 months. This is similar to the definition Mojtabai used in his definition of perceived need for alcohol, drug, and mental health problems (26). Regarding (i) the respondent was asked “Was there ever a time you thought you should see a doctor, counselor, or other health professional or seek any other help for your drinking, but you didn't?” The respondent needed to additionally to affirm that this perceived need happened in the last 12 months.
Regarding (ii), the respondent was asked about seeking treatment from the following 13 sources: Alcoholics Anonymous/Narcotics or Cocaine Anonymous; family services or other social service agency; alcohol or drug detoxification ward or clinic; inpatient ward of a psychiatric or general hospital or community mental health program; outpatient clinic, including outreach programs and day or partial patient programs; alcohol or drug rehabilitation program; emergency room for any reason related to drinking; halfway house, including therapeutic communities; crisis center for any reason related to drinking; Employee Assistance Program; clergyman, priest or rabbi for any reason related to drinking; private physician, psychiatrist, psychologist, social worker or other professional; or any other agency or professional.
The NSDUH measure of perceived need for alcohol treatment was derived in a similar fashion: (i) the respondent reported a need for treatment or counseling, or additional treatment or counseling for their alcohol use, in the past 12 months, or (ii) the respondent reported receiving treatment or counseling in the past 12 months for their alcohol use.
Past 12-month treatment was defined in (ii) above.
NESARC utilizes the NIAAA Alcohol Use Disorder and Associated Disabilities Interview Schedule-DSM-IV Version (AUDADIS-IV), a structured interview designed to be administered by lay interviewers. Studies have demonstrated generally good to excellent reliability and validity (2,31-33). The alcohol diagnostic section from NESARC contains 35 questions used to assess the 7 alcohol dependence criteria and the 4 abuse criteria. Thus there are multiple questions for each dependence and abuse criterion.
For each set of questions corresponding to a particular dependence or abuse criterion the number of affirmative responses was summed. For example, in NESARC respondents were asked four questions for the tolerance to alcohol criterion, and thus could endorse 0, 1, 2, 3, or 4 alcohol tolerance symptoms. These different levels were then coded with binary indicator variables. To create a more parsimonious model (decrease the number of degrees of freedom) used by these abuse and dependence symptoms, when odds ratio estimates were similar for two values, they were collapsed into fewer categories. For example, the symptom count for tolerance to alcohol was entered into the final model as 0, 1 or 2, or 3 or more symptoms.
In NSUDH, each alcohol and abuse criterion was assessed with a single question, which was coded with a binary indicator in our analyses. The use of a single question (rather than multiple questions) for each criterion could be seen as a limitation, although many sophisticated, validated instruments, used for research and clinical care for behavioral disorders utilize only one question to assess each criterion.
The NESARC AUDADIS-IV instrument contains measures of 12 month and lifetime: major depression, dysthymia, mania and hypomania, generalized anxiety, panic disorder (with and without agoraphobia), agoraphobia without panic, social phobia, specific phobias, PTSD, retrospective childhood ADHD, and self-report of schizophrenia/psychosis. For our analyses, we utilized measures of the number of mental health disorders in the past 12-months (0, 1, 2, 3+). Personality disorders included paranoid, schizoid, anti-social, histrionic, avoidant, dependent, and obsessive compulsive. We coded the presence of a Cluster A personality disorder (paranoid, schizoid), the presence of a Cluster B personality disorder (anti-social, histrionic) and Cluster C personality disorder (avoidant, dependent, and obsessive-compulsive) with binary indicators.
The K6 is a measure of psychological distress that was developed for use in the National Health Interview Survey and subsequently included in the NSDUH (34,35). The K6 includes six questions that measure on a zero to four scale how frequently respondents experience symptoms of psychological distress (nervousness, hopelessness, restlessness, depressed, feeling worthless, feeling that everything is an effort) during the month in the past year when they were feeling their worst emotionally. Respondents with scores of 13 and higher based on a simple count of the endorsed items are considered to have serious psychological distress (28,29). NSDUH also contains a depression module, which allows for the construction of a variable measuring major depression in the past 12 months.
Sociodemographic variables included gender, age, race, income, education, marital status, and insurance status.
Our analytical plan was identical in the NESARC and NSDUH samples. First, we used a logistic regression to assess the effects of our predictors on perceived need.
The analytical sample was individuals with an alcohol abuse or dependence diagnosis. Then among those with perceived need, we regressed any alcohol treatment on the same set of predictors. The sample characteristics are shown in Tables 1 and and2.2. To measure strength of association, we used the generalized coefficient of determination (R2) described by Cox and Snell (36).
In NSDUH, 10.4% of individuals with an alcohol use disorder had any perceived need for treatment in the past year (Table 1). Table 3 shows the multivariate predictors of any perceived need. In a multiple logistic regression, the explanatory power of the diagnostic variables (partial pseudo R2=.13) was much greater than the explanatory power of the demographic variables (partial pseudo R2=.01). Five of the 7 dependence criteria and 3 of the 4 abuse criteria significantly predicted perceived need, and the results were generally highly significant (e.g., p<.001). The effect of each symptom was moderate to strong (e.g. OR's greater than 1.30). The strongest predictor of perceived need was recurrent alcohol-related legal problems (OR=4.82, 95% CI=3.82−6.09, p<.001). Symptoms not related to perceived need were: tolerance; taken in larger amounts or over a longer period of time than intended; and recurrent use in situations where it is physically hazardous. Individuals with serious psychological distress were more likely to perceive need (OR=1.76, 95% CI=1.42−2.18, p<.001).
Among the sociodemographic variables race was a significant predictor of perceived need (chi-square=16.1, df=3, p=.001), with whites having the highest rates of perceived need. Low-income individuals (OR=1.52, CI=1.21-1.91, p<.004) and unmarried individuals (OR=1.52, CI=1.21-1.91, p<.004) were more likely to have perceived need. Age was highly significant, with younger individuals much less likely to perceive need, and middle-aged individuals the most likely to perceive need (chi-square=42.9, df=2, p<.001). Gender, education, and insurance status were not significant.
Among those with perceived need in NSDUH, 70% (weighted) reported receiving any treatment in the past 12 months. Again, in a multiple logistic regression clinical factors had the greatest explanatory power (partial pseudo R2=.13), compared to sociodemographics (partial pseudo R2=.03). In the logistic regression, among the diagnostic symptoms, 2 of the 7 dependence and 1 of the abuse criteria significantly predicted receiving treatment. Low income individuals (chi-square=13.1, df=3, p=.004) were less likely to receive treatment, and individuals who were not married were more likely than those who were to receive treatment (OR=1.62, CI=1.01-2.60, p=.04). Gender, education, race, marital status, and age were not significant.
Similar to NSDUH, in NESARC a small percentage of individuals with an alcohol use disorder in the past year perceived a need for alcohol treatment, 11.0% (Table 2). Also, as in NSDUH, in a logistic regression investigating perceived need, diagnostic factors had the greatest explanatory power (partial pseudo R2=.20), compared to demographic factors (partial pseudo R2=.01). In the logistic regression most of the abuse/dependence symptom variables were significant (Table 4). Tolerance, drinking in larger quantities or longer periods than anticipated, reduced social activities from alcohol use, failure to fulfill role obligations, and recurrent use in which it was physically hazardous were not significant predictors of perceived need. On the other hand, among the sociodemographic factors, age was the only significant predictor, with younger individuals significantly less likely to perceive a need for treatment (Chi-square=19.7, df=1, p<.001). Race, which was significant in our NSDUH analyses, was not significant in our NESARC analyses. Gender, income, education, insurance, marital status, urbanicity, and self-rated health were all also non-significant.
In NESARC, among those with perceived need, 64% reported receiving some alcohol use disorder treatment in the past year. Again, diagnostic factors had the greatest explanatory power (partial pseudo R2=.14), versus sociodemographic factors (partial pseudo R2=.09). Several sociodemographic groups were significantly less likely to receive treatment, including the uninsured, Hispanics, and those who were married.
We used two different nationally representative surveys to investigate perceived need for alcohol use disorder services. Given methodological differences inherent across all surveys, the similarity in the results between the two studies is striking. In both surveys we found that fewer than 1 in 9 individuals with an alcohol use disorder perceived a need for treatment. Further, our definition of perceived need, based on whether the individual reported thinking he/she needed treatment in the past year, or actually received treatment, was liberal. It likely included individuals in treatment not because they felt they needed treatment, but rather, in treatment at the behest of family, friends, or the legal system. In both surveys we found that the explanatory power of diagnostic variables was substantially larger than the explanatory power of sociodemographic factors.
The issue of improving treatment for alcohol use disorders and other behavior health conditions has received considerable attention, including a recent report from the Institute of Medicine (20). It has been suggested that improving the quality of treatment for alcohol use disorders (and other substance and mental health disorders) could decrease the mortality, morbidity, and the societal cost of these disorders. However, because the large majority of individuals with alcohol use disorders receive no treatment at all, improving the quality of care for alcohol use disorders while not increasing the proportion of individuals who receive services might have only a modest impact on morbidity, mortality, and societal costs on the population level.
Our results suggest that failure to perceive need continues to be the major reason individuals with alcohol use disorders do not receive treatment, as only a small proportion of individuals with alcohol use disorders perceived need. On the other hand, among those with perceived need, the majority receive treatment. Further, our results offer little reason for optimism concerning perceived need, as the percentage with perceived need in the National Longitudinal Alcohol Epidemiologic Survey (NLAES) , conducted in the early 1990's, was slightly higher, indicating no progress on this important front.
Factors related to access to alcohol treatment, such as insurance parity for the treatment of substance use disorders, have received considerable attention in the past decade. One concern is that lack of parity contributes significantly to unmet need. However, in our analyses, insurance status did not predict perceived need, but did predict receiving treatment only among the relatively few who actually perceive a need for a treatment. This suggests that attempts to make alcohol use disorder services more accessible through efforts such as parity legislation will at best increase modestly the number of individuals receiving alcohol use disorder services, although this should not be seen as a reason not to implement parity. On the other hand, our results suggest that even modest success in increasing the percentage of individuals with perceived need would dramatically increase the number of individuals in treatment. For example, currently more than 8 out of 9 individuals do not perceive a need for a treatment. Thus, if efforts to increase perceived need were successful in 1 out of 8 individuals, the overall number of people receiving alcohol use disorder services could almost double, an increase that would likely overwhelm the capacity of the alcohol use disorder treatment system. Because of this, if efforts to increase perceived need were successful, treatment would have to be better targeted to those with the most severe disorders, or the capacity of the alcohol use disorder treatment system would need to be expanded.
Increasing the proportion of individuals with alcohol use disorders who perceive a need for treatment could possibly be done in several ways. First, in a situation where individuals do not perceive a need for treatment, primary care screening, as recommended by U.S. Preventive Services Task Force (37), is essential in detecting disorders. For policy purposes, it is important to know what proportion of individuals are screened; if screening rates are already high, then there may be little room for improvement. Unfortunately, estimates of alcohol use disorder screening rates in primary care settings vary widely. Clinician surveys suggest that screening occurs frequently (38,39); patient surveys offer less reason for optimism (40-42). However, studies which indicate that about half of primary care patients with alcohol abuse/dependence are known by the clinician to have an alcohol problem (43-45) suggest that patient surveys might be more accurate than physician surveys. Similarly, it is important to know whether screening rates are increasing, decreasing, or static, but we know of no study that assesses this important question.
Clinical studies suggest that approximately half of patients with a mental health or substance use disorder recognized by their clinician as having a disorder receive treatment (43-51), although the proportion receiving guideline concordant treatment is likely lower. Thus screening does not necessarily lead to quality care. Primary care screening could be followed by brief primary care alcohol interventions, or specialty referrals for those with relatively severe disorders. The development of screening instruments and brief interventions has been the subject of intense research efforts in the past 15 years, although most efforts have focused on efficacy and effectiveness studies. We believe that the next logical step is efforts to widely implement these evidence-based screening and brief interventions in community settings.
Along with efforts to increase screening, we also need more effective public health efforts to increase perceived need among individuals with alcohol use disorders and their families on a population basis. While we know of no experimental data on this issue, such efforts appear to have been effective for depression. For example, use of antidepressants increased dramatically during the 1990s (52-55) and it is believed that direct-to-consumer advertising both increased recognition of the symptoms of depression and diminished stigma associated with the disease, resulting in higher treatment rates (53).
However, increasing perceived need among individuals with alcohol use disorders is likely to be more difficult than increasing perceived need for depression treatment. First, while there are FDA approved medications for the treatment of alcohol use disorders (e.g., acamprosate, naltrexone, and antabuse), there has been no significant direct to consumer adverting for these products, or even detailing to physicians. Second, the stigma associated with alcohol use disorders may be greater than depression-related stigma. Third, alcohol use disorders are addictive disorders, while depression is not.
It would be useful to know which sociodemographic groups are most likely to not perceive a need for treatment, to better target these groups with appropriate interventions. We found that in both NSDUH and NESARC the explanatory power of the sociodemographic variables in predicting perceived need was relatively small, especially compared to the explanatory power of the diagnostic variables. Age was the only significant sociodemographic predictor of perceived need in both NSDUH and NESARC, although all sociodemographic groups had high levels of not perceiving need for treatment. This suggests that efforts to increase rates of perceived need should be targeted broadly to all sociodemographic groups, although the problem of perceived need is particularly acute in younger individuals.
NESARC and NSDUH are both ongoing studies, and thus are valuable for assessing the evolution of perceived need. Repeated cross-sectional waves, as employed in NSDUH, are the preferred methodology for assessing whether a belief or characteristic, such as perceived need, is increasing over time in a population (56). On the other hand, the panel design of the NESARC survey will allow us to investigate how perceived need for treatment changes in individuals over time. However, neither survey is able to give us detailed insight into the important issue of why individuals do not perceive need for treatment, an issue that we believe is best addressed initially with qualitative interviews.
It is likely that high levels of unmet need for alcohol use disorder services will continue to persist as long as perceived need is low. Efforts are needed to both increase levels of perceived need among those with alcohol use disorders, and to improve the quality of care they receive.
NIAAA R01 AA016299
Disclosures: None for any author.