The objectives of this study are: Firstly, to investigate the association between psychiatric disorders and functional somatic disorder (FSD). Secondly, to investigate whether psychiatric disorders are risk factors for newly developed (incident) FSD after a 5-year follow-up period.

Functional somatic disorder (FSD) is a common condition characterized by a persistent pattern of impairing physical symptoms. Our knowledge about what causes the development and perpetuating of these disorders is still sparse, however, it is well accepted that they have a multifactorial aetiology, comprising both biological, psychological, and social factors.

The relationship between FSD and other conditions have been investigated in different manners: A study in patients with severe FSD have shown an increased number of other physical diagnoses compared to healthy controls. A general population-based study has shown FSD to be associated with other self-reported physical diseases, depression, and anxiety. Further, psychopathology has shown to predict irritable bowel syndrome and fibromyalgia, while irritable bowel syndrome has shown to predict common mental disorders (5). However, for some diagnoses, e.g. whiplash associated disorders, the relationship with psychiatric disorders is more inconsistent.

The previous research may carry some limitations: Clinical studies into highly selected patient samples may carry risk of selection bias, and most population-based studies have used self-report for establishing FSD diagnoses and psychiatric diagnoses which may bring risk of misclassification. Hence, more studies with a sound methodology into these aspects are needed.

The proposed longitudinal study includes data from two investigations of the same cohort from the general Danish population. Both validated symptom questionnaires and diagnostic interviews will be used for the establishment of FSD diagnoses and psychiatric discharge diagnoses and prescription medication will be obtained from the comprehensive Danish Central Registries.

Objective

The objective of this study is twofold:

1. To investigate the association between psychiatric disorders and FSD
2. To investigate whether psychiatric disorders are risk factors for development of incident FSD in a 5-year follow-up period

Hypotheses:

1. Psychiatric disorders (across all included diagnoses) are positively associated with FSD
2. Anxiety disorder, depression, bipolar disorder, personality disorder, and stress-related disorder as individual diagnoses are positively associated with FSD
3. Psychiatric disorders (across all included diagnoses) are associated with newly developed (incident) FSD (FSD negative at baseline and FSD positive at follow-up)
4. Anxiety disorder, depression, bipolar disorder, personality disorder, and stress-related disorder as individual diagnoses are associated with newly developed (incident) FSD (FSD negative at baseline and FSD positive at follow-up)

Data from the DanFunD (Danish Study of Functional Disorders) baseline and 5-year follow-up investigations will be included. The baseline cohort is a random sample selected through the National Civil Registration system among people living in 10 municipalities in the western part of greater Copenhagen, Denmark, aged 18 to 76 years. The baseline cohort constitutes data from self-reported validated symptom questionnaires and diagnostic interviews. The follow-up cohort consists of participants all born in Denmark, between 24 and 84 years of age. The follow-up cohort also constitutes data from self-reported validated symptom questionnaires and diagnostic interviews.

Psychiatric diagnoses will be obtained from the Danish National Patient Registry according to the International Classification of Diseases, 10th Revision (ICD-10). In order to obtain cases with more 'mild' psychiatric disorders that do not require hospital admission, prescriptions for relevant pharmacological treatment of these disorders will be obtained from the Danish National Prescription Registry.

Primary outcome (dependent variables):

Participants with FSD will be defined as follows:

• Participants with FSD operationalised by the Bodily Distress Syndrome single- and multi-organ type will be defined with both questionnaires and diagnostic interviews
• Participants with a functional somatic syndrome, i.e. irritable bowel, chronic widespread pain, and chronic fatigue will be defined with questionnaires

Primary explanatory/independent variables:

Psychiatric disorders will be identified by means of

1. inpatient and outpatient hospital diagnoses (both primary and secondary diagnoses) coded in the Danish National Patient Registry according to ICD-10, and
2. prescriptions for relevant pharmacological treatment of these disorders from the Danish National Prescription Registry.

The time frame will be 10 years before the DanFunD baseline investigation and the 5-year period between the baseline and follow-up investigation.

All analyses will be performed using STATA version 17.0. Descriptive tables will be performed with number and percentages of individuals with psychiatric diagnoses (overall psychiatric diagnosis and specific categories: anxiety disorder, depression, bipolar disorder, personality disorder, and stress-related disorder) and use of prescription medication across FSD diagnoses and controls.

Hypothesis 1:

Relevant regression models with FSD obtained at baseline as primary outcome (baseline FSD positive = 1, baseline FSD negative = 0) and a dichotomous variable constituting overall psychiatric diagnoses and/or prescription medicine (at least one: yes = 1, no = 0) received within a period of 10 years before baseline as primary explanatory variable will be performed.

Hypothesis 2:

Relevant regression models with FSD obtained at baseline as primary outcome (baseline FSD positive = 1, baseline FSD negative = 0) and a dichotomous variable constituting each specific diagnosis and prescription medication for the specific diagnoses (anxiety disorder, depression, bipolar disorder, personality disorder, and stress-related disorder) (yes = 1, no = 0) received within a period of 10 years before baseline as primary explanatory variable will be performed.

As some of the prescription medication may be prescribed for several of the included diagnoses, a sub analysis will be performed for hypothesis 2, where only the specific diagnoses will constitute the primary outcome variable, i.e. without prescription medication.

For the analyses of hypotheses 1 and 2, prevalence odds ratios (POR) with 95% confidence intervals (CI) will be used as measure of association. A POR > 1 supports the hypotheses. Depending on number of cases, the analyses will be adjusted for as many of the following covariates (prioritized order) as possible without over fitting the data: 1. Sex, 2. age, 3. social status, and 4. severe physical disease (measured with the Charlson comorbidity index).

Hypothesis 3:

Relevant regression models with newly developed FSD at follow-up as primary outcome (follow-up FSD positive = 1, baseline FSD negative = 0) and a dichotomous variable constituting overall psychiatric diagnoses and/or prescription medicine (at least one: yes = 1, no = 0) received within a period of 10 years before baseline as primary explanatory variable will be performed.

Hypothesis 4:

Relevant regression models with newly developed FSD at follow-up as primary outcome (follow-up FSD positive = 1, baseline FSD negative = 0) and a dichotomous variable constituting each specific diagnosis and prescription medication for the specific diagnoses (anxiety disorder, depression, bipolar disorder, personality disorder, and stress-related disorder) (yes = 1, no = 0) received within a period of 10 years before baseline as primary explanatory variable will be performed.

As some of the prescription medication may be prescribed for several of the included diagnoses, a sub analysis will be performed for hypothesis 4, where only the specific diagnoses will constitute the primary outcome variable, i.e. without prescription medication.

For the analyses of hypotheses 3 and 4, odds ratios (OR) with 95% CI will be used as measure of association. An OR > 1 supports the hypotheses.

Depending on number of cases, the analyses will be adjusted for as many of the following covariates (prioritized order) as possible without over fitting of the data: 1. Sex, 2. age, 3. social status, and 4. severe physical disease (measured with the Charlson comorbidity index).

Inclusion Criteria:

• None

Exclusion Criteria:

• Not born in Denmark
• Not being a Danish citizen
• Pregnancy

A total of 25,368 participants were randomly drawn from the adult general Danish population by means of the the Danish Civil Registration system.

The DanFunD baseline cohort comprises a total of 7,493 (29.5% of invited participants) men and women aged 18-76 years, born in Denmark, and living in the Western part of greater Copenhagen.

The follow-up cohort (gathered in the years 2018-2020) consists of participants all born in Denmark, between 24 and 84 years of age. The follow-up cohort constitutes data from self-reported questionnaires (n=4,288) and diagnostic interviews data (n=1,094).

Restrictions apply to the availability of data according to Danish law, which means that data cannot be made publicly available. A request for collaboration or access to data can be sent to ckff@regionh.dk. An application for data should consist of a short synopsis of the study planned. Please see our website (www. danfund.org) for more information.