Analysing Patient Trajectories of Individuals with Alcohol Use Disorders (PRAGMA)
Study Protocol of a Data-Linkage Study
Abstract
Abstract:Aim: Alcohol use is causing a considerable health burden for individuals and society in Germany. To reduce the burden from alcohol use, ensuring optimal treatment for those who are in need, is key. With this data-linkage study, we aim to provide a comprehensive description of healthcare service use among individuals with alcohol use disorders (AUD) in Hamburg, the second-largest German city. Methods: The study population is defined as adults living in Hamburg, currently insured by one of two statutory health insurance funds and with at least one alcohol-specific ICD-10 code between 2016 and 2021. Additionally, we will obtain data from pension funds and the Hamburg basic data monitoring system of outpatient addiction aid. By using unique identifiers, individual register data from these three sources will be linked. Hypotheses and qualitative analyses are presented in the form of research questions to analyse administrative prevalence rates, patient trajectories and predictors of treatment success as well as to estimate the impact of prototypical care pathways and the COVID-19 pandemic on utilization of alcohol-specific healthcare services. Discussion: The study ‘Patient Routes of People with Alcohol Use Disorders in Germany’ (PRAGMA) will be the first to provide an in-depth understanding of treatment provision for people with AUD in Germany. Following up a heterogeneous sample of people with AUD for six years will provide a unique opportunity to compare current with recommended care pathways as well to identify options for care improvements.
Zusammenfassung:Zielsetzung: Alkoholkonsum stellt eine erhebliche gesundheitliche Belastung für den Einzelnen und die Gesellschaft in Deutschland dar. Um diese Belastung zu reduzieren, ist die Sicherstellung einer optimalen Behandlung für diejenigen, die sie benötigen, von zentraler Bedeutung. Ziel dieser Data-Linkage-Studie ist eine umfassende Beschreibung der Inanspruchnahme von Gesundheitsleistungen durch Personen mit Alkoholkonsumstörungen in Hamburg. Methoden: Die Studienpopulation ist definiert als in Hamburg lebende Erwachsene, die aktuell bei einer von zwei gesetzlichen Krankenkassen versichert sind und in den Jahren 2016 bis 2021 mindestens eine alkoholspezifische ICD-10-Diagnose aufwiesen. Zusätzlich werden Daten der Rentenversicherungsträger und des Hamburger Basisdatenmonitorings der ambulanten Suchthilfe genutzt. Mittels eineindeutiger Identifikatoren werden die individuellen Registerdaten aus den drei Quellen miteinander verknüpft. In Form von Forschungsfragen werden Hypothesen und qualitative Analysen vorgestellt, um administrative Prävalenzraten, Patient_innenwege und Prädiktoren für einen Behandlungserfolg zu untersuchen sowie die Auswirkungen prototypischer Versorgungspfade und der COVID-19-Pandemie auf die Inanspruchnahme von alkoholspezifischen Gesundheitsleistungen abzuschätzen. Diskussion: Mit der Studie „Patient_innenwege von Menschen mit einer Alkoholabhängigkeit in Deutschland“ (PRAGMA) wird erstmals ein umfassendes Verständnis der Behandlungsangebote für Menschen mit Alkoholkonsumstörungen in Deutschland bereitgestellt. Die Nachverfolgung einer heterogenen Stichprobe von Menschen mit Alkoholkonsumstörungen über sechs Jahre hinweg bietet die einzigartige Möglichkeit, die tatsächlichen mit den empfohlenen Behandlungspfaden zu vergleichen und Möglichkeiten zur Verbesserung der Versorgung zu identifizieren.
Introduction
In Germany, levels of alcohol use are above the global and European average (Manthey et al., 2019). The morbidity and mortality burden from alcohol use is high, with about 14 000 deaths and 145 000 hospitalisations caused directly by alcohol use in the year 2020 (Manthey et al., 2023). Considering the contribution of alcohol to all injuries and diseases (e. g., cancer and cardiovascular problems), about 12 % and 6 % of all deaths were attributable to alcohol use in 2021, for men and women respectively (Kraus et al., 2023). Accordingly, the economic burden of alcohol use on German society is substantial (Effertz et al., 2017).
To alleviate the societal burden stemming from alcohol use, the provision of effective interventions for people with risky drinking patterns and those with an alcohol use disorder (AUD) diagnosis is key (Rehm et al., 2013). To reduce heavy alcohol use and to improve the social and health status of this population, several entry points are summarised for Germany below.
The German Health Care System
People with risky drinking patterns or manifest AUD can receive addiction-specific interventions in four different outpatient and inpatient care settings: primary health care, other health care professionals in outpatient settings (e. g., psychotherapists), hospitals (including emergency rooms), outpatient addiction care services, inpatient rehabilitation. First, in outpatient primary health care settings, only a few receive appropriate care (e. g., brief interventions, pharmacotherapy) (Manthey et al., 2021; Frischknecht et al., 2021) but this setting is important for early identification of heavy drinking patterns. Second, low-threshold counselling services for substance users and their relatives, in addition to group therapy and outpatient rehabilitation are offered in outpatient addiction care services. Third, short- or long-term (1 vs 3 weeks) withdrawal therapies are conducted in (specialized) hospitals (e. g. psychiatric wards; inpatient). Fourth, inpatient rehabilitation treatment is provided after withdrawal therapy to stabilize patients and increase the likelihood of abstinence.
Prevalence and Treatment of Risky Drinking Patterns and Alcohol Use Disorders
In Germany, an estimated 22 % of the adult working-age population drinks above defined drinking levels (men: 24 g pure alcohol per day; women: 12 g pure alcohol per day; Rauschert et al., 2022). According to the German S3 guideline on ‘Screening, Diagnosis, and Treatment of Alcohol-related Disorders’, brief interventions are recommended for this population in primary health care (Kiefer et al., 2022), however, only a small minority of alcohol users (with risky drinking patterns) are estimated to have had a conversation about alcohol with their primary health care physician (Kastaun et al., 2022; Manthey, Lindemann, Verthein et al., 2020). In other words, most people who progress from risky drinking patterns to AUD do so without their primary health care physician knowing – despite regular contact.
The prevalence of AUD in the German population aged 18–64 is estimated at 6 % (Atzendorf et al., 2019) and about one-third of individuals with AUD were identified by general practitioners (Kraus et al., 2015; Kuitunen-Paul et al., 2017). For the year 2009, it was estimated that only about 16 % received treatment in hospital or outpatient addiction care and 2 % underwent rehabilitation treatment (Kraus et al., 2015). Data from the year 2016 confirmed low treatment rates, which were below average among 21- to 39-year-olds (Manthey, Lindemann, Verthein et al., 2020). With the onset of the COVID-19 pandemic a decrease in alcohol-specific hospitalisations in Germany has been found, for both males and females (Manthey et al., 2023).
Who Receives Treatment for Alcohol Use Disorders?
Findings from various studies suggest that predictors for seeking treatment for AUD problems are lower education, later AUD onset (Blanco et al., 2015), higher drinking levels, a higher degree of comorbidity (Rehm et al., 2015), and previous treatment for other substance use disorders or mental health problems (Blanco et al., 2015). The role of comorbidity and treatment seeking is complex and intertwined with age. In Germany, the average age of individuals with AUD in inpatient treatment settings is considerably higher (47 years) than for people with opioid (40 years), stimulant (32 years) or cannabis use disorders (30 years; similar distribution for outpatient settings, see Schwarzkopf et al., 2021). On average, the onset of AUD occurs at the age of 30, about 15 years after the onset of drinking (15 years; Martens & Neumann-Runde, 2021). Thus, people seeking treatment for their drinking have usually engaged more than one or two decades in heavy alcohol use. As heavy alcohol use over a prolonged period commonly damages the liver and/or the cardiovascular system, it is not surprising to observe a range of somatic health problems from drinking in the treatment-seeking population (Rehm et al., 2015).
There are inconsistent findings with regard to sex as a predictor for treatment seeking (Blanco et al., 2015; Bourdon et al., 2020). One study found no age differences between males and females with AUD entering outpatient treatment, but a shorter history of heavy drinking among females (Bravo et al., 2013).
What Are the Predictors for Treatment Success?
Several factors are key to the success of AUD treatment, that can be defined as abstinence or reduction of alcohol use but also by health indicators or survival. First, temporal variations in drinking patterns, including abstinence and heavy drinking periods, are not only inherent to the course of AUDs (Cranford et al., 2014) but are also predictive of treatment outcomes (Maisto et al., 2021; Wallach et al., 2022). Those who achieve sustained periods of abstinence or experience fewer relapses during or shortly after therapy are more likely to lower their alcohol use (Maisto et al., 2021) and experience better psychosocial outcomes (see Supplement of Maisto et al., 2021). Yet, other variables that do not reflect alcohol use, such as comorbidities and psychosocial functioning (how physical and mental symptoms affect daily life) are also crucial for long-term health outcomes (Witkiewitz et al., 2020).
Relapses after AUD treatment are common. In Germany, about 25 to 50 % of patients with AUD experience relapse within 12 months after discharge from an inpatient rehabilitation facility, with the majority of relapses occurring within 30 days after discharge (Bachmeier et al., 2021). Several international studies have reported sex and socioeconomic differences regarding treatment success. While females appear to experience greater health risks from drinking the same amount of alcohol (see e. g., for liver diseases Llamosas-Falcón et al., 2022), single studies report that – when compared to males – females show improved drinking outcomes after treatment (Bravo et al., 2013; Running-Bear et al., 2022) and benefit from women-specific treatment programs (Berge et al., 2022). However, a systematic review on sex differences and treatment outcomes found the evidence to be rather mixed (McCrady et al., 2020). People with lower SES, defined as unstable housing or not seeking employment, were at increased risk for relapse in a US sample (Running-Bear et al., 2022). Another US study found that lower education and being below the poverty line were predictive for lower functioning three years after completing AUD treatment (Swan et al., 2021). The same study also found environmental variables reflective of structural disadvantages, such as income levels among employees or coverage of health insurance in the community to predict treatment outcomes (Swan et al., 2021). Similarly, rural residents with AUD in Ontario, Canada were less likely to receive appropriate care and more likely to die after an alcohol-related hospitalization (Friesen et al., 2022).
Within PRAGMA treatment outcome will be measured by regular termination of treatment episodes, documented provider data on treatment outcomes as well as by the rate of re-hospitalisations after regular and premature termination of treatment episodes.
Research Motivation
Alcohol use is causing a considerable individual and societal health burden in Germany. There is a large potential in reducing this burden since only a fraction of the population in need of treatment is receiving formal help. The fractioned healthcare system in Germany has been a barrier to a detailed, comprehensive understanding of the treatment received by people with risky drinking patterns or manifest AUD. In a recent pilot study, Möckl and colleagues (2023) used routinely collected register data from different key stakeholders (health insurance, pension funds, outpatient addiction care) to overcome this barrier and to describe individual pathways of patients with alcohol dependence in Bremen in the years 2016/17. Although more than half of the individuals with alcohol dependence were documented in the health system, the utilization rates of addiction-specific treatments were low.
The PRAGMA (‘Patient Routes of People with Alcohol Use Disorders in Germany’) research project will extend the observation period to up to 10 years. By including data of the Hamburg monitoring system of outpatient addiction aid (BADO), the data set for some of the patients will be substantially enlarged. In addition to other socio-demographic information, the BADO data also include past life events, data on physical, psychological, and family problems as well as on resources, professional development, and financial situation. The combination of the extended observation period together with the enlarged dataset will enable a comprehensive and systematic analysis of patient trajectories in the care system.
Study Aim
With this data-linkage study, we aim to provide a comprehensive description of healthcare service use among individuals with risky drinking patterns and those with AUD, in addition to identifying entry points for improving health and social outcomes in this population.
Methods
Study Population and Data Sources
The study population is defined as adults (18 or older) living in the city of Hamburg (North Germany), currently insured by one of two statutory health insurance (SHI) funds and with at least one alcohol-specific ICD-10 code within January 1st 2016 and December 31st 2021 (see Electronic Supplementary Material [ESM] 1). For this study population, we will obtain routine register data for services reimbursed by the SHI, which includes outpatient (including primary health care but also other medical specialities and psychotherapies), inpatient (emergency departments and hospital admissions with overnight stay), outpatient surgeries (hospital admissions without overnight stay), as well as prescription data (only outpatient). Importantly, this will cover addiction-specific but also any other healthcare service that is documented by the data owners (e. g., dentistry, psychotherapy for anxiety disorders).
As not all addiction services are covered by the SHI, we will retrieve further data from other agencies. First, information on outpatient or inpatient rehabilitation services will be obtained from the German pension funds (PF). Second, information on services provided by outpatient addiction care services of the Hamburg basic data monitoring system of outpatient addiction aid (BADO) in Hamburg. The type of health care services covered by the different data owners and the temporary availability of the data is summarized in Table 1.
Data Linkage Procedures
To determine the healthcare utilization trajectory of the same individual across different settings, the register data from the different sources will be linked using individual identifiers. As the health registry data from the different sources are not meant to be combined, the following steps need to be undertaken to link the data while preserving anonymity.
First, the data will be pseudonymized by each data holder by creating an individual identifier based on information about the persons’ names (third character of the first name; length of the first name; third character of surname; length of surname), their sex (man/woman/other), and year of birth (for more information on the identifier, see below). The health registry data with the person identifier will then be encrypted using a 12-digit ‘Keyed Hash-Message Authentication Code’ (H-MAC Algorithm, ISO-Norm 9797). At all times, the encryption code will remain with the data owners, i. e., with SHI, PF and BADO, disallowing decrypting the identifier and thus minimizing the chance to identify natural persons through information about their name, sex, or birthday.
Second, the data owners will transfer the health registry data together with the encrypted personal identifier to an independent third party not involved in data analyses. Here, the various data sets from SHI, PF, and BADO will be checked for plausibility and variables with low frequencies will be recoded into broader categories to minimize the risk of re-identification of natural persons.
Third, the checked and cleaned health registry data from each data owner will be sent to the researchers at the University Clinic Hamburg-Eppendorf (principal investigators). These data sets will contain the encrypted individual identifiers that will then be used to identify the same persons across the different data sources. For example, this allows us to determine whether an inpatient withdrawal treatment (reimbursed by SHI) was followed by residential treatment (reimbursed by PF).
Mitigating Data Linkage Errors
Previous projects have successfully used the same approach to link data sets across different sources based on an identifier that is constructed on predetermined information (names, sex, birthday) (e. g. Möckl et al., 2023). This deterministic approach is different from probabilistic approaches in which a range of information is used to determine the probability that two given records belong together. Despite being a deterministic approach, errors may still occur. There are two types of errors (Harron, 2022). The first error occurs if two records match although belonging to different individuals (=false match), which can happen if two individuals of the same sex share similarities in their names and were born in the same year. The second error occurs if two records of the same individual do not match (=missed match), which can happen if, for example, the last name changed after marriage and that change was registered in one data source (e. g., SHI) but not in another. To identify false matches, we will use five different approaches for developing the human-identifier variable for each person. For missed matches, there will be no method to ascertain concerned cases.
Quantitative Analyses
In this research project, the quantitative analyses are structured by six research questions. For each research question, we have specified accompanying hypotheses and provided information on the definition of the analytical sample and the planned statistical methods. A more granular description of the statistical analyses can be accessed in the publicly available study plan (https://osf.io/gc29n).
Research Question 1: What Is the Administrative Prevalence of Alcohol-Related Diagnoses in Hamburg, Germany?
Hypothesis
The two-year administrative prevalence of alcohol-related diagnoses in Hamburg, Germany, is expected to be 3 % (Möckl et al., 2023), with more than half of cases being diagnosed with alcohol dependence.
Definition of the Analytical Sample
For the main analysis, the study population will be limited to individuals insured by one of the two SHIs. For calculating the prevalence, reference populations will be obtained by the SHIs. In sensitivity analyses, we will also include diagnoses in settings covered by PF or BADO data (in the same target population).
Statistical Methods
The annual administrative prevalence will be the percentage of adults (aged 18 or older) insured with one of the two SHIs who have been diagnosed with any alcohol-specific condition in settings covered by the SHI in the period between 2016 and 2021: outpatient, inpatient, and rehabilitation services. Additional analyses will stratify four groups of alcohol-specific conditions based on ICD-10 diagnoses: a) alcohol dependence (F10.2), b) AUD (F10.1-F10.9), c) harms indicative of acute alcohol use (F10.0; R78.0; T51.0; T51.9), and d) harms indicative of chronic heavy alcohol use (E24.4; G31.2; G62.1; G72.1; I42.6; K29.2; K70; K85.2; K86.0; O35.4). All prevalence estimates will be provided for the target population as well as stratified by gender-, age-, and income quintiles (or equivalent definition of socioeconomic position that is available in the data).
Research Question 2: Which Short- and Long-Term Trajectories of Addiction-Specific Healthcare Utilization Can Be Observed Among Individuals Newly Diagnosed with Alcohol Use Disorders?
We will perform exploratory analyses to describe the short- (6 months) and long-term (2 years) trajectories of people with newly diagnosed AUD.
Definition of the Analytical Sample
The study population will be limited to SHI-insured individuals that have been diagnosed with AUD in any setting. For both short- and long-term trajectories, there need to be at least 12 months of complete data available before the index date (look-back window), i. e., before the first diagnosis registered in the data.
Statistical Methods
We will first distinguish between the following five addiction-specific interventions: counselling (delivered in outpatient addiction care services), psychotherapy (delivered in inpatient or outpatient settings), pharmacotherapies as prescribed by medical professionals in outpatient settings, (qualified) withdrawal treatment delivered in hospitals, rehabilitation treatment (delivered in inpatient or outpatient settings; see also Table 2).
Statistical Analysis
The data will contain individual information on the utilization of various types of addiction-specific services over a given time period. Two steps will be considered: first, a cross-sectional reduction of the data using latent class analyses, allowing to identify groups of services that are typically used by the same persons. The class membership probabilities will be associated with individual-level covariates, such as sociodemographics or comorbidities. Second, to consider the longitudinal nature of the data, we will consider methods such as repeated measures latent class analyses, latent transition analyses or latent growth mixture modelling. Which of these methods will eventually be applied, primarily depends on the structure of the data. Specifically, the decision will depend on the groupings of the interventions and the granularity of the time resolution, which cannot be assessed a priori.
Research Question 3: How Do Patients with Different Addiction-Specific Health Care Utilization Trajectories Differ in Terms of Morbidity Indicators?
Hypothesis
Among persons with an incident F10.2 diagnosis in outpatient settings, lower rates of healthcare utilization (inpatient or outpatient) can be observed if they received subsequent withdrawal and/or rehabilitation treatment – as recommended in the guidelines.
Definition of the Analytical Sample
All SHI-insured persons with an F10.2 diagnosis in outpatient settings, who did not have this diagnosis in outpatient settings in the 12 months prior and who also did not enter withdrawal treatment in that preceding period, will be included. At least 12 months of data before and after the diagnosis is required.
Statistical Methods
Those with an incident F10.2 diagnosis will be grouped according to entering withdrawal treatment (and rehabilitation). Dependent variables will be calculated on the entire follow-up period and will comprise 1) the number of all-cause hospital days and 2) the number of all-cause outpatient contacts. To account for varying follow-up periods, the days/contacts will be divided by the number of follow-up days and the time spent in the hospital for withdrawal/rehabilitation will be subtracted from the follow-up period. Differences in the two dependent variables will be examined using generalized linear models with appropriate link functions, controlling for sociodemographic variables and comorbidity in the 12 months prior to the incident diagnosis. Additional analyses using propensity score matching will be considered.
Research Question 4: What Are the Effects of Prototypical Addiction-Specific Healthcare Utilization Trajectories on Avoidable Hospitalisations?
In a modelling study, we will estimate how many hospital days could be avoided if more people with an incident F10.2 diagnosis entered the guideline-recommended withdrawal treatment.
Definition of the Analytical Sample
This modelling study will be based on the sample and the findings from research question 3.
Statistical Methods
Different scenarios of treatment uptake will be considered, e. g., 10 %, 25 % and 50 % of people with optimal utilization of addiction-specific interventions. The number of avoidable hospital days will be estimated as the difference between the number of hospital days among people with no or insufficient utilization of addiction-specific interventions compared to the number of hospital days among people with optimal utilization of addiction-specific interventions. Simulation of the impact of different scenarios will be performed similar to previous studies (Manthey, Lindemann, Kraus et al., 2020; Manthey et al., 2021).
Research Question 5: Which Factors Predict the Utilization of Alcohol-Specific Services and/or Premature Completion of Treatment Programs?
In an explorative study, we will identify sociodemographic, health, and psychosocial factors linked to treatment uptake and early dropout.
Definition of the Analytical Sample
The analyses will be conducted for two samples: 1) SHI-insured people; 2) SHI-insured people with valid data from BADO.
Statistical Methods
The dependent variable will be discontinued addiction-specific intervention, which can be defined as withdrawal treatment (SHI data), rehabilitation (PF data), and psychotherapy. For withdrawal and rehabilitation, discontinuation is explicitly coded in the data. For psychotherapy, we will compare patients in the lowest quantile (e. g., 20 %) in terms of intervention length against the remaining sample. With logistic regression analyses, we will model the odds of discontinuation depending on sociodemographic variables, comorbidity, and previous experiences of addiction-specific interventions (as defined in research question 2). Repeating these analyses for the sample of SHI-insured people with valid BADO data, we will be able to also include a range of other possible determinants, such as drinking patterns, familial SUD history, or criminal justice problems.
Research Question 6: How Did the COVID-19 Pandemic Impact on the Utilization of Alcohol-Specific Services?
Hypothesis
At the beginning of 2020, the utilization of five types of alcohol-specific services has declined in general and across all types of treatment. With the COVID-19 pandemic progressing, declines in withdrawal therapies were more pronounced than outpatient interventions (psychotherapy, medication, addiction counselling), which were easier to maintain with physical distancing measures.
Definition of the Analytical Sample
We will aggregate the number of people who have entered addiction-specific intervention programs for their alcohol problems (see research question 2). Depending on the data availability and structure, we will aggregate the data on a weekly, monthly, or quarterly basis.
Statistical Methods
Time series a) for all types of interventions and b) stratified by type of intervention will be created. Following guidelines for time series analyses in addiction research (Beard et al., 2019), we will identify possible changes in the outcome variables for various periods of the COVID-19 pandemic (e. g., the first peak in March/April 2020, the second peak in winter 2020/2021).
Ethics
The deterministic data linkage approach together with the study background, the research questions and data protection procedures has been detailed in a data protection concept which has been submitted and approved by the Federal Office for Social Security, the Department of Social Affairs, Health, Youth, Family and seniors of the State of Schleswig-Holstein and the Ministry of Labour, Health and Social Affairs of the State of North Rhine-Westphalia (Reference number: GZ: 117 – 1010904#00001#0050). A separate ethics review will not be undertaken.
Public Involvement
Using a nominal group technique, the results from this quantitative data analysis will be explored with patient representatives, service providers and representatives of health insurances as well as pension funds. The aim is to identify opportunities towards to an improved patient and result-oriented care for people with AUD in Germany. For this, we will compare priorities between different groups (e. g., between patients and service providers or service providers and health insurances) and to identify divergent views regarding possible solutions.
Dissemination
The study results will be disseminated through publications in scientific journals and presentations on local, national, and international conferences. Open access publication will be used to ensure a high visibility and widespread dissemination of the findings to different stakeholders.
Funding
This study protocol is based on the project ‚Patient Routes of People with Alcohol Use Disorders in Germany‘ (PRAGMA) which was funded by the Innovation Committee of the Federal Joint Committee (Gainsayer Bundesausschuss, GBA) under the funding code 01VSF21029. The funder was not involved in the study design, collection, analysis, interpretation of data or preparation of this publication.
Discussion
This study protocol of the PRAGMA study describes a comprehensive and systematic data analysis on the utilization of health care services and patient trajectories among individuals with AUD over a period of up to 10 years. The quantitative analyses provide the opportunity to generate comprehensive and largely representative insights into the care pathways of SHI insured patients with AUD in Germany. To date, only few studies have analysed the utilization of alcohol specific care of patients with AUD based on register data (see Kraus et al., 2015; van der Linde et al., 2014), but only one of them used data linkage to combine register date from different stakeholders (Möckl et al., 2023).
A strength of the study is the high expected sample size and the observation time of up to 10 years, that for the first time will allow empirical results on typical short- and long-term trajectories of AUD patients and the impact of those trajectories on morbidity. To approximate the expected sample size, we assume an administrative prevalence of 2.9 % for ICD-10 F10 as determined in Bremen in 2016/2017 (Möckl et al., 2023). As the two SHI cover about 400 000 insured persons, we expect at least about 11 600 persons to be included in the present analyses. Furthermore, the analyses will identify patient groups with and without use of health care services over a defined observation period as well as on patient groups with frequent use or with premature termination of treatment and/or care. With the inclusion of the BADO data, we will be able to identify factors for the (non) utilization or (non) completion of alcohol specific in- and outpatient services, at least for a part of the study population. The study results will be based on routine data from Hamburg but can be extrapolated to other urban regions in Germany as the alcohol-related healthcare system is not assumed to significantly differ between the federal states. Several limitations need to be considered. First, register data from SHI or PF are limited to health care interventions that have been billed by services providers. Hence, the data was not collected primarily for research purposes and the information contained in the data may be subject to biases by the service providers. Certain information that would be of interest for this study, e. g., delivery of alcohol screening or brief interventions, are not recorded in the data and thus cannot be analysed. Second, we will employ measures to mitigate any errors that may be incurred by linking data sets from various sources (e. g., two records of the same person are not successfully linked), but we do not have the means to verify our measures or to estimate the scope of errors in the data. Third, we follow a population-based approach and will be able to include a large number of people affected by AUD, however, we may not be able to cover the most disadvantaged persons, such as homeless, incarcerated, and undocumented populations. Fourth, our trajectory analyses will only be a snapshot of people using addiction-specific services at some point after AUD onset. We seek to control for previous treatments using a 12-months look-back window, but we cannot rule out the utilization of any services during previous periods or even adjust for that in the analyses.
With the identification of healthcare trajectories of patients with AUD the project will extend our understanding of patients’ health care utilization and contribute to improve alcohol-specific health care provision. The unique approach to link data from individuals across several data sources may serve as scaffold for future research projects aimed at understanding the provision of healthcare for people affected by addiction. The PRAGMA project started in 2022, the main outcomes will be available by the end of 2024.
Electronic Supplementary Material
The electronic supplementary material (ESM) is available with the online version of the article at https://doi.org/10.1024/0939-5911/a000877
References
(2019). The use of alcohol, tobacco, illegal drugs and medicines. Deutsches Aerzteblatt International, 116, 577–584. 10.3238/arztebl.2019.0577
(2021). Effektivität der stationären Suchtrehabilitation. FVS-Katamnese des Entlassjahrgangs 2018 von Fachkliniken für Alkohol- und Medikamentenabhängige [
Effectiveness of inpatient addiction rehabilitation. FVS catamnesis of the 2018 discharge cohort from specialist clinics for alcohol and drug addicts ]. https://www.suchthilfestatistik.de/fileadmin/user_upload_dshs/05_publikationen/katamnesen/FVS_fachkliniken_alkohol_medikamente/FVS_Katamnese_-_Alkohol_Medikamente_SuchtAktuell__2021-1.pdf(2019). Understanding and using time series analyses in addiction research. Addiction, 114(10), 1866–1884. 10.1111/add.14643
(2022). Gender specific early treatment for women with alcohol addiction (EWA): Impact on work related outcomes. A 25-year registry follow-up of a randomized controlled trial (RCT). Drug and Alcohol Dependence, 239, Article 109600. 10.1016/j.drugalcdep.2022.109600
(2015). Probability and predictors of treatment-seeking for substance use disorders in the U. S. Drug and Alcohol Dependence, 149, 136–144. 10.1016/j.drugalcdep.2015.01.031
(2020). Characterization of service use for alcohol problems across generations and sex in adults with alcohol use disorder. Alcoholism, Clinical and Experimental Research, 44(3), 746–757. 10.1111/acer.14290
(2013). Gender differences in the long-term outcome of alcohol dependence treatments: an analysis of twenty-year prospective follow up. Drug and Alcohol Review, 32(4), 381–388. 10.1111/dar.12023
(2014). Trajectories of alcohol use over time among adults with alcohol dependence. Addictive Behaviors, 39(5), 1006–1011. 10.1016/j.addbeh.2014.02.009
(2017). The costs of hazardous alcohol consumption in Germany. European Journal of Health Economics, 18(6), 703–713. 10.1007/s10198-016-0822-1
(2022). Rural-urban disparities in post-discharge outcomes following alcohol-related hospitalizations in Ontario, Canada: A retrospective cohort study. Drug and Alcohol Dependence, 238, Article 109568. 10.1016/j.drugalcdep.2022.109568
(2021). Diagnostik und Behandlung alkoholbezogener Störungen – Versorgerbefragung zur Erarbeitung von Strategien der Leitlinienimplementierung [
Diagnosis and treatment of alcohol-related disorders – A survey of health care providers to develop strategies for the implementation of guidelines ]. Suchttherapie, 22, 141–151. 10.1055/a-1265-4687(2022). Data linkage in medical research. BMJ Medicine, 1(1), Article e000087. 10.1136/bmjmed-2021-000087
(2022). Prevalence and characteristics of hazardous and harmful drinkers receiving general practitioners’ brief advice on and support with alcohol consumption in Germany: Results of a population survey. BMJ Open, 12(9), Article e064268. 10.1136/bmjopen-2022-064268
(2022). German guidelines on screening, diagnosis, and treatment of alcohol use disorders: Update 2021. European Addiction Research, 28(4), 309–322. 10.1159/000522335
(2023). Trends in alcohol-attributable morbidity and mortality in Germany from 2000 to 2021: A modelling study (Manuscript submitted for publication).
(2015). Inanspruchnahme gesundheitlicher Versorgung durch Alkoholabhängige [
Utilisation of health care by alcohol addicts ]. Suchttherapie, 16(1), 18–26. 10.1055/s-0034-1376999(2017). Alkoholabhängigkeit in der primärärztlichen Versorgung: Welche Patienten werden erkannt? [
Alcohol dependence in primary medical care: Which patients are recognised? ]. Suchttherapie, 18(02), 82–89. 10.1055/s-0042-113143(2022). Sex-specific association between alcohol consumption and liver cirrhosis: An updated systematic review and meta-analysis [Systematic Review]. Frontiers in Gastroenterology, 1, Article 1005729. 10.3389/fgstr.2022.1005729
(2021). Within-AUD outpatient treatment heavy drinking transitions and associations with long-term outcomes. Drug and Alcohol Dependence, 228, Article 108968. 10.1016/j.drugalcdep.2021.108968
(2023). Changes in the alcohol-specific disease burden during the COVID-19 pandemic in Germany: Interrupted time series analyses. European Journal of Public Health, 33(4), 645–652. 10.1093/eurpub/ckad103
(2020). The potential effects of an extended alcohol withdrawal treatment programme on morbidity and mortality among inpatients in the German city of Bremen: A simulation study. Substance Abuse Treatment, Prevention, and Policy, 15(1), Article 1. 10.1186/s13011-019-0249-7
(2020). Versorgung von Personen mit riskantem Alkoholkonsum und schwerer Alkoholkonsumstörung in Bremen: bedarfsgerecht und leitlinienkonform? [
Care for people with risky alcohol consumption and severe alcohol use disorders in Bremen: Needs-based and in line with guidelines? ]. Bundesgesundheitsblatt – Gesundheitsforschung – Gesundheitsschutz, 63(1), 122–130. 10.1007/s00103-019-03072-z(2019). Global alcohol exposure between 1990 and 2017 and forecasts until 2030: A global modelling study. The Lancet, 393(10190), 2493–2502. 10.1016/S0140-6736(18)32744-2
(2021). Can alcohol consumption in Germany be reduced by alcohol screening, brief intervention and referral to treatment in primary health care? Results of a simulation study. PloS One, 16(8), Article e0255843. 10.1371/journal.pone.0255843
(2021). Suchthilfe in Hamburg. Statusbericht 2020 der Hamburger Basisdatendokumentation in der ambulanten Suchthilfe und der Eingliederungshilfe [
Addiction care system in Hamburg. Status report 2020 of the Hamburg basic data documentation in outpatient addiction support and integration assistance ]. BADO e. V.(2020). Treatment interventions for women with alcohol use disorder. Alcohol Research, 40(2), 1–18. 10.35946/arcr.v40.2.08
(2023). Estimating the prevalence of alcohol-related disorders and treatment utilization in Bremen 2016/2017 through routine data linkage. Frontiers in Psychiatry, 14, Article 1002526. 10.3389/fpsyt.2023.1002526
(2022). The use of psychoactive substances in Germany: Findings from the epidemiological survey of substance abuse 2021. Deutsches Aerzteblatt International, 119(31–32), 527–534. 10.3238/arztebl.m2022.0244
(2015). Who receives treatment for alcohol use disorders in the European Union? A cross-sectional representative study in primary and specialized health care. European Psychiatry, 30(8), 885–893. 10.1016/j.eurpsy.2015.07.012
(2013). Modeling the impact of alcohol dependence on mortality burden and the effect of available treatment interventions in the European Union. European Neuropsychopharmacology, 23(2), 89–97. 10.1016/j.euroneuro.2012.08.001
(2022). Factors associated with readmission to alcohol and opioid detoxification in the Alaska Interior. American Journal on Addictions, 31(5), 406–414. 10.1111/ajad.13288
(2021). Suchthilfe in Deutschland 2020. München: Institut für Therapieforschung.
(2021). Individual and community social determinants of health and recovery from alcohol use disorder three years following treatment. Journal of Psychoactive Drugs, 53(5),394–403. 10.1080/02791072.2021.1986243
(2014). Sekundärdatenanalyse der Prävalenz der Alkoholabhängigkeit (F10.2) in Deutschland [
Secondary data analysis of the prevalence of alcohol dependence (F10.2) in Germany ]. Deutsche Medizinische Wochenschrift, 139(45), 2285–2289. 10.1055/s-0034-1387353(2022). Predictors of abstinence, no heavy drinking days, and a 2-level reduction in World Health Organization drinking levels during treatment for alcohol use disorder in the COMBINE study. Alcoholism: Clinical and Experimental Research, 46(7), 1331–1339. 10.1111/acer.14877
(2020). Can alcohol use disorder recovery include some heavy drinking? A replication and extension up to 9 years following treatment. Alcoholism, Clinical and Experimental Research, 44(9), 1862–1874. 10.1111/acer.14413
