The national alcohol helpline in Sweden: an evaluation of its first year

Heavy or risky alcohol use is a threat to public health as it is causally linked to a wide range of harms and diseases [1]. Screening and performing a brief intervention is an evidence-based and cost-effective strategy for the early management of alcohol-use disorder [2–7]. Hazardous use of alcohol can be regarded as a level of consumption that entails a risk of harmful use of alcohol or alcohol dependence in the future.

Telephone helplines addressing the general public are well-established enterprises, e.g. for smoking cessation. The worldwide success of telephone helplines in relation to smoking cessation [8, 9] suggests that a comparable approach to the risky use of alcohol may be a viable option. Many studies have found telephone-based support to be a successful and cost effective method of treatment, such as in the aftercare of substance abusers [10–27]. However, judging from the scientific studies describing telephone-based alcohol counseling, the Swedish Alcohol Helpline seems to be among the first to offer such a service as a primary-care option for risky users of alcohol.

Although people with alcohol dependence, the most severe state of alcohol-related disorder, may require specialized treatment [28, 29], brief clinically based counseling episodes have been found to be an effective way of reducing consumption and risk among people with hazardous alcohol consumption [30–32].

About 14 percent of the Swedish population have been estimated to have hazardous alcohol consumption [33]. Hazardous use of alcohol is usually identified by screening instruments, which measure alcohol consumption and its consequences for the individual’s physical and mental health [34]. The more severe states, harmful use of alcohol and alcohol dependence, are indicated by several behavioral, cognitive, and physiological symptoms. They are generally detected in clinical interviews, but brief and straightforward proxies for clinical interviewing are also provided in questionnaires [35]. WHO makes different recommendations according to the severity of the disorder [35, 36]. For hazardous users simple advice suffices, but for harmful use WHO recommends not only simple advice but also brief counseling and continued monitoring. Alcohol dependence should prompt a referral to specialists for diagnostic evaluation and treatment.

Typically, the most severe state, i.e., dependence, is characterized by a strong desire to consume alcohol, impaired control of its use, persistent drinking despite harmful consequences, greater priority given to drinking than to other activities and obligations, increased tolerance of alcohol, and physical withdrawal reactions when alcohol use is discontinued. Not all hazardous drinkers become dependent over time. But once dependency on alcohol develops, it may become more difficult to stop or decrease consumption. Nonetheless, even if people with alcohol dependence are at a greater risk of incurring high levels of harm, the bulk of harmful consequences is to be found among people with hazardous use of alcohol, simply because of their greater number – the so-called prevention paradox [37].

Any telephone helpline is reliant on people’s own identification of their behavior as a problem. Whether hazardous alcohol users spontaneously seek help to change their behavior from a service like the telephone helpline is not clear. Telephone support offers anonymity and is easy accessible; even when ill and in bed, you can always make a call. On the other hand, people may prefer face-to-face contact [38], and the telephone can be a problematic option for some, e.g., for persons with a speech or hearing impairment. There may also be other disadvantages, such as the demands placed on existing infrastructure, and a lack of guidelines. There are methodological concerns in relation to previous research on helplines, counselor fatigue, etc.

Alternative techniques for improving early detection, such as screening or incorporating relevant questions into clinical interviews in health care have been recommended [39]. But there have also been obstacles to the implementation of these methods, such as lack of knowledge and skills, lack of time, and financial disincentives [2]. Moreover, not all people visit primary care. The threshold for actively seeking help from a telephone helpline, where you can choose to be anonymous, is likely to be lower than in primary care.

The Swedish Alcohol Helpline is a nationwide service operated by the Stockholm County Health Service in collaboration with Karolinska Institutet. Due to the success of quitlines for smoking cessation in high-income countries, including Sweden [8, 9], it was decided to develop and test a comparable approach to the risky use of alcohol. Together with several other organizations, Stockholm County Council started to operate a smoking quitline in 1998 [40, 41]. The “know how” and the techniques (computer systems, logistics, and many other factors) from the tobacco quitline were applied to the development of the Alcohol Helpline.

The Swedish Alcohol Helpline, a nation-wide service, has been in operation since January 2007. Its principal objective has been to encourage people contemplating change in their alcohol drinking habits, but it also offers help in sustaining change and preventing relapse among persons who have already decreased their alcohol consumption.

The primary aim of the present study is to describe callers to the alcohol helpline and, in particular, to assess the characteristics of callers requesting counseling. This involves considering distributions by gender, age, and alcohol use (i.e., non-hazardous use, hazardous use, harmful use, and dependence). A secondary aim, in a pilot study, is to follow up a sample of callers who enrolled for more intensive (proactive) support to assess their goals, perceived successes, and possible changes in alcohol use over time.

Of all the 480 callers to the Help Line during its first calendar year, 47 percent (n = 226) were people seeking support for their own alcohol-related problems (clients). Other callers were relatives of persons with alcohol problems, 31 percent (n = 147). There were also calls on other matters (19%, n = 91), e.g., health care personnel wanting information, prank calls, and wrong numbers.

The proportions of hazardous users, harmful users, and persons dependent on alcohol were similar in the sub-cohort of 40 clients followed up in the pilot study (Sample 2) to those in the baseline assessment, which included all the 120 clients completing the AUDIT (Sample 1); the proportions of hazardous users were 25/120 (20%) and 10/40 (25%), respectively (not shown in the table).

In the follow-up sample, the total AUDIT-score was reduced from 21 to 14 points (see Table 2). The AUDIT scores suggest that men and women did not have significantly different [t(38) = -1.1, p = 0.26] drinking profiles at baseline. There were no significant differences between the age groups, with the exception of the oldest age group, which had a lower mean score than that of the others [15–39 years: t(37) = -0.4, p = 0.69; 40–49 years: t(37) = -1.4, p = 0.18; 50–64 years: t(37) = -0.4, p = 0.68; 65+ years: t(37) = 2.7, p = 0.01].

Both men and women decreased in total AUDIT score significantly from baseline to follow-up (Table 2). Similarly, separate analyses of the age groups indicated that decreases occurred over all ages. The separate analyses of the AUDIT scores for the three different categories at baseline showed that persons with harmful use or dependence decreased their scores significantly, but no significant decrease was detected for hazardous users (Table 2). The analyses also showed that a decrease in AUDIT score among those who received other professional help and those only receiving support from the Helpline were both significant.

Next, we examined changes as transitions between the different categories of alcohol use. At follow-up, 18 (45%) of the 40 clients showed a positive change from a higher category of use to a lower; 18 (45%) showed no change; and 4 (10%) showed negative change from a lower category of use to a higher. Of the 40 clients, 5 (12%) had changed to the non-hazardous group.

Not shown are the estimated intercepts (in ORs) for the change model. In the last model the odds for all steps of change differed significantly from each other. In the model without independent variables, all steps except the step from negative change to no change were significant (p < 0.05).

Assessment of the primary personal goals revealed that 72.5% (29/40) of the follow-up clients had primarily wanted to “gain control over their drinking”, and approximately 22.5% (9/40) stated that they wanted to “quit drinking completely” as a primary goal. Two clients (5%) were not able or willing to formulate primary goals for contacting the Helpline (not shown in the tables).

Perceived success at follow-up of achieving initial goals indicated partial success. The average score was 6.2 (SD = 2.8), and there was no significant difference (t(39) = 0.8, p = 0.45) between the hazardous users (M = 5.6, SD = 3.1) and the others (M = 6.4, SD = 2.6) in this respect (not shown in the tables).

Although not conclusive, the results from the present study represent a first step in describing professionally run telephone helplines as a primary care option for people with alcohol-related problems. Of the 480 callers to the Alcohol Helpline during its first year, about half were clients, whereas one-third wanted to discuss the alcohol behaviors of relatives. As a consequence, as time has passed, the explicit aim of the Helpline has been extended to include other help, e.g., help to relatives and other people close to persons with alcohol-behavior issues. In comparison with other treatment options in Sweden, the telephone line reaches a high proportion of women [52]. Of the 120 clients with baseline AUDIT assessment, a substantial proportion (n = 25, 20%) were categorized as hazardous alcohol users on the AUDIT scale, indicating that the service may have the potential to reach people at early stages of alcohol abuse.

In the general Swedish population, 14 percent of men and 13 percent of women have been estimated to have hazardous alcohol use, while four percent of the men and one percent of the women may be classified as alcohol dependent, i.e., an AUDIT score of 15+ for men, and 13+ for women [33]. The Helpline’s clients have an average AUDIT score of 21 points, which equals the score of patients at a psychiatric emergency care unit [48]. In a US sample of alcohol-dependent individuals entering outpatient care, the average AUDIT score was 26 points, and the distribution was 4, 11 and 85 percent for hazardous use, harmful use, and dependence, respectively [51]. The corresponding distribution in the present study was 21, 27, and 52 percent, respectively. In comparison with web-based support [53, 54], or screening at primary care receptions [55], a clientele with more severe alcohol use seems to be reached by the Helpline.

The decrease in AUDIT score over time noted in the high-consumer group fits the idea that “treatment effects” are usually stronger than “prevention effects”. Thus, those with the initially highest problems have the highest scope for improvement, whereas those with the lowest problems have the lowest scope for change. It also fits a regression-to-the-mean tendency, meaning here that heavy consumers are more likely to decrease their drinking than average consumers [56]. The design of the current study does not enable us properly to evaluate these different explanations of change in AUDIT score.

The population of callers may also have changed over time, since information about the Helpline has become more widespread. Such changes in the study population limit the generalizability of our results. Our findings indicate that people with hazardous alcohol use are a difficult target group to reach. People with less severe drinking problems seem to be less likely to seek treatment or support.

A majority of the clients chose reactive services. Consequently, the pilot-study sample (Sample 2) comprises a select group of clients who accepted being registered as proactive. Although there were no differences in AUDIT scores between the proactive and reactive clients, or between the whole sample and the follow-up at baseline, it remains unclear how selection might have biased our results. The clients choosing proactive support may, for example, have been highly motivated people who were about to change anyway, i.e., that change would come with non-regarding treatment Also, the observed change in the follow-up group was not unusual size [30]. Studies of brief interventions in primary care of people with high alcohol use have consistently reported decreases in both treated and untreated (control) individuals [57]. Further, studies also suggest an effect of the screening procedure [58].

The counseling protocol in the present study appears to have been helpful and, although the present pilot study is not conclusive, the potential and novelty of the approach deserve further investigation. One encouraging finding from the pilot study is that many clients experience that they have achieved their personal goals for contacting the Helpline. Nevertheless, it remains unclear whether the Helpline’s treatment is effective.

Apparently, the counseling offered by the Alcohol Helpline is often conceived as a supplementary support when people are trying to change their alcohol habits. According to the follow-up, more than half of the clients had received additional professional help to modify their alcohol habits during the follow-up period. This may partly be due to the helpline support protocol which encourages severe alcohol users also to seek other forms of help. Unfortunately, we can only speculate on this issue, since it was not systematically assessed in the follow-up interview.

In an intervention concerning university students, the AUDIT score in the high-risk group was reduced from 11.5 to 7.5 after 6 months [49]. Our follow-up indicates a general change of about the same relative size, but from higher initial AUDIT scores.

In the present pilot study, no significant behavior change was noted for hazardous users during the study period. A longer follow-up period may be needed to capture change in this group of hazardous alcohol users. Also, it is possible that the helpline clients with hazardous use may differ from the corresponding target group in the general population. The AUDIT instrument discriminates alcohol-related behavior and consequences only during the previous year. It would have been desirable also to measure alcohol problems earlier in life, since clients of this kind may have been utilizing the service as an aid to preventing relapse into riskier alcohol habits. Thus, whether the Helpline prevents hazardous drinkers from progressing into more severe drinking is unclear.

Using medication was associated with lower odds for making positive change. Possibly, the use of medicines can be seen as a marker of more serious or long-lasting alcohol problems. On the other hand, such an interpretation goes against other results, namely that positive change is primarily found among persons with harmful use or dependence. The small sample size and a lack of randomization inhibit the interpretation of our results.

Most randomized control studies of telephone counseling have indicated positive treatment results in health care circumstances [10–16, 20, 21, 23, 24], and also elsewise [17–19]. In one study, primary care patients received up to six episodes of telephone counseling after systematic screening [13]. The results showed that the intervention group exhibited greater decline in risky drinking days compared with the controls who simply received an information pamphlet. In another study, all patients in 81 German general practices were screened for alcohol use disorder [14]. Eligible participants received computerized interventions plus a maximum of four brief telephone counseling sessions. A small to a medium effect size of the treatment was found for the intervention group compared with controls. But no difference was found between a group of patients who received stepped care, i.e., computerized intervention plus up to three 40-min telephone based counseling episodes (depending on the success of the previous intervention),and a group who received care equaling the maximum amount of stepped care.

Although non-randomized descriptive studies, such as the present one, are useful when priorities for further investigations are identified or when randomized trials are unethical [59–63], a future study of treatment effects in a telephone helpline setting needs to include a randomized control group. A control group would capture the naturally occurring change in this particular self-selected group of people.