Taylor et al. (Reference Taylor, Jones and Dening2014) raise an important issue concerning the detection of alcohol problems in older adults. The authors identify a number of age-related factors playing a role in the detection of alcohol problems, such as stigma, the concept of alcohol use disorder diagnoses and their standardized assessment, and drinking levels. They list a comprehensive review of tools for assessing alcohol problems among older adults, including laboratory tests, questionnaires, and interviews.

To this list, we would like to add results from a large representative study in six European countries on alcohol problems and alcohol use disorders in primary care (Rehm et al., Reference Rehm2015a). In this study, we used the Composite International Diagnostic Interview (CIDI) to assess alcohol problems – considered to be the gold standard for detecting alcohol use disorders (for instance, in the NICE guidelines).

In short, we found that the older the client, the more problems there had been for the CIDI to detect alcohol dependence, taking the diagnosis of general practitioners (GPs) as gold standard (Rehm et al., Reference Rehm2015b). Figure 1 of the cited publication illustrates how both assessments of alcohol dependence were related to age. Of the oldest age group studied, i.e. the patients in primary care between 55 and 64 years of age, the CIDI could only detect 38.6% of people with alcohol use disorders as diagnosed by the GP (95% CI: 32.3–44.9%).

This finding corroborates the conclusion of Taylor et al. (Reference Taylor, Jones and Dening2014) that standardized instruments have problems in identifying people with alcohol use disorders in older adults. In addition to some diagnostic criteria not being applicable to older adults, older adults may no longer recognize or remember their own adaptation processes leading to tolerance – a core diagnostic criterion of alcohol dependence. For a 20 old, the transition from no alcohol to regular alcohol consumption with the inevitable increase of tolerance is easy to remember – as opposed to older adults. In addition, older adults may not recognize loss of control as such and judge their overall social situation as acceptable because of long-term habituation. However, GPs’ identification of alcohol use disorders was found to be strongly related to the criterion on continuation of drinking despite social consequences (Web Appendix 4 from Rehm et al., Reference Rehm2015a), which may explain the gap between CIDI and GPs in detecting alcohol use disorders among older adults.

Changing standardized questionnaires/interviews by including somatic symptoms may improve performance in detecting alcohol problems in older adults (Web Appendix 4 from Rehm et al., Reference Rehm2015a). We also agree with the authors to include questions on drinking levels in order to enhance sensitivity and specificity (Saha et al., Reference Saha, Stinson and Grant2007). However, from a biological point of view, there are two pharmacokinetic processes that make older people more susceptible to the impact of alcohol: First, as the liver shows reduced degradative activity, more alcohol may reach the blood stream for any given intake; and second, due to on average reduced body water the same amount of alcohol is associated with higher blood alcohol concentration (Ferreira and Weems, Reference Ferreira and Weems2008). Thus, different norms for detecting problems should be developed.

As such instruments have been, are, and will nevertheless be used in epidemiological research, the resulting incidence and prevalence figures are considerable underestimates. Improved standardized assessments and a combination with clinical judgments from treating GPs may constitute a strategy to obtain more reliable estimates.