Comparison of data sources on alcohol problems: An exploratory exercise using surveys vs. national healthcare registers

Abstract

Aims: Both survey and healthcare register data struggle as data sources to capture the phenomenon of alcohol problems. We study a large group of people for whom survey data and two types of register data are available, and examine the overlaps of similar or related measures in the different data sources to learn about potential weaknesses in each. We also examine how register-based data on the prevalence of alcohol problems change depending on which register data are used. Design: We use data from the Regional Health and Wellbeing Study (ATH) of the adult Finnish population collected in 2013 and 2014 ( n = 69,441), individually linked with data on two national healthcare registers (Care Register for Health Care; Register of Primary Health Care visits) for the survey year and previous year. Results: The prevalence of substance-abuse-related healthcare was almost two-fold if data on outpatient primary care visits were included in addition to hospitalisations. Forty-six per cent of the survey respondents self-reporting substance-abuse-related healthcare service use were identified in the registers, and 22% of all respondents with such service use according to registers reported this in the survey. Records of substance-abuse-related healthcare service use, controlled for self-reported alcohol use and self-reported substance-abuse-related service use, were found more often for men, the middle-aged, people with basic education only, and the non-employed. Conclusions: The results are suggestive of underreporting in both data sources. There is an evident need to develop recording practices in the healthcare registers regarding substance use disorders.

Attention deficit hyperactivity and oppositional defiant disorder symptoms in adolescence and risk of substance use disorders—A general population‐based birth cohort study

“…Despite substantial sample size, it was not possible to stratify our analyses by sex due to power issues. Furthermore, 3.1% of the whole study population was diagnosed with SUD which is likely to be an underestimate 47 . Possible reasons are underdiagnosis and the fact that part of the substance use services are administered in social services which does not contribute data to the registers used in this study.…”

Using the alcohol use disorders identification test to predict hospital admission for alcohol‐related conditions in the Danish general population: a record‐linkage study

“…Finnish healthcare registers also include information on procedures and interventions delivered as part of primary healthcare (e.g. substanceabuse-related guidance and outpatient detoxification) [61]. This type of register data on primary care contacts is not available in Denmark;…”

“…Register‐based AUDIT validation studies could be extended using non‐hospital medical registers. For instance, in Finland, International Classification of Primary Care codes are used to classify presenting problems and diagnoses related to healthcare visits to general practice [ 61 ], including short‐term and long‐term alcohol misuse [ 61 ]. Finnish healthcare registers also include information on procedures and interventions delivered as part of primary healthcare (e.g.…”

“…Finnish healthcare registers also include information on procedures and interventions delivered as part of primary healthcare (e.g. substance‐abuse‐related guidance and outpatient detoxification) [ 61 ]. This type of register data on primary care contacts is not available in Denmark; however, future research in other Nordic countries could involve assessing the predictive validity of the AUDIT in terms of general practice diagnosis related to drinking and comparing such results to the prediction of hospital‐based ARDD.…”

Dealing with qualitative work

“…In a study by Mäkelä and colleagues (2021) , the phenomenon of alcohol problems is encircled by different types of register data which points out serious underreporting of the total burden of problems. The prevalence of substance-abuse-related healthcare was almost twofold if data on outpatient primary care visits were included in addition to hospitalisations.…”