This review showed high acceptance rates for alcohol screening and intervention as well as adherence to intervention among emergency and surgical patients with AUDs. Four in five ED patients accepted alcohol screening compared to two in three surgical patients, whereas the intervention acceptance rate was to two out of three in ED patients compared to almost 100% among surgical patients. Though, as only a minority of the studies was conducted among surgical patients, no conclusions can be made regarding possible differences in emergency and surgical patients' acceptance of alcohol screening and interventions. Adherence to the alcohol intervention programs was above 60% for up to twelve months in both ED and surgical patients. Overall, the numbers needed to screen to identify one eligible AUD patient and to get one eligible patient to accept intervention varied from a few up to 70 patients.
Acceptance of alcohol screening among AUD patients in intervention trials was not reduced compared to studies performing screening exclusively [49–52], despite the fact that the consequences of being identified with an AUD in an intervention trial can be more comprehensive regarding study participation and possible changes in alcohol consumption.
The NNS are important when planning future interventions. Here the NNS to identify one AUD patient is not different from that described in primary care . However, the NNS for acceptance of alcohol intervention is up to four times higher in primary care. This indicates a larger potential for conducting alcohol interventions in hospital settings.
In addition, the acceptance and adherence rates are comparable to those in smoking cessation intervention studies conducted in similar patient groups [54–57].
We were not able to show that any of the study characteristics facilitated or hindered acceptance or adherence, but the rates were probably influenced by multiple other factors for example patients sex, age and lifestyle .
Bias and limitations
The weaknesses of this review are closely related to the weaknesses of the individual studies. In general, the included studies were heterogeneous. The intervention and primary outcomes differed from study to study. Screening methods used for detecting AUDs varied notably between the studies from blood tests to interviews and different questionnaires. The included questionnaires mainly focus on alcohol abuse and dependency and have mostly been developed in non-surgical settings . These questionnaires may therefore not be useful for detecting a current hazardous alcohol intake in surgical patients, which is the most clinical relevant outcome . Also, in emergency departments there does not appear to be a gold standard tool for screening for AUDs .
Moreover, the studies used different questionnaires and in the same questionnaires different cut-of points were sometimes applied. As a result a patient identified with AUD in one study would not necessarily be identified or included in another study. This could affect the NNS. In addition, differences in other inclusion and exclusion criteria between the studies could also influence the NNS.
Many studies had not reported the total patient population and number of patients accepting screening and alcohol intervention. Also, numbers of patients available for follow-up were not given in a few studies. Missing data from several studies may have affected the acceptance and adherence rates as well as the value of the subgroup analyses, where data in many cases was limited. Furthermore, as almost all studies described randomized trials and brief interventions comparisons among the few studies having other characteristics may not be valid.
No comparable systematic or narrative reviews were found. Other studies on patient opinions and experiences found that patients in EDs were positive towards alcohol screening including blood tests [61, 62]. Also, in interviews with AUD patients following intervention they described the ED as an appropriate setting . A recent study among acute surgical patients with AUDs sustained that alcohol intervention is relevant in relation to surgery .
Though generalization of the results should be considered carefully, the ED patients and patients undergoing surgery do not seem to form a major barrier for introducing alcohol screening and intervention programs. The high acceptance and adherence rate may also reflect staff effort and compliance to research protocol. Compared to project staff, the clinical staff may experience other barriers for alcohol screening and intervention such as missing knowledge and training as well as lack of time, appropriateness of setting and implementation into daily routines [65–67]. These barriers may, however, be overcome by prioritizing the area through professional teaching and training as well as using lessons learned regarding appropriate setting and resources.
The clinical routines differ from study settings in several ways. In general, projects and studies benefit from a very professional approach from trained and experienced staff, who are also highly dedicated to screen patients, intervene and follow-up. Study settings have also been tailored to meet the requirements for completing the project parts. In this way, several of the barriers above are overcome.
A background WHO-paper concluded that there are no technical barriers for handling alcohol intervention as well as other hospital-based health promoting activities in the DRG-system . Furthermore, a simple model for documentation of patient need for alcohol intervention and the related health promotion activity was shown to be useful, applicable and understandable when evaluated internationally in clinical settings . These international tools are an integrated part of WHO standards and they all fit into the hospitals' quality management .