Psychiatric Comorbidity in Back Pain Disorders

Aim of the thesis

The aim of this PhD thesis is to use population-based registry data to:

1. Estimate the occurrence of psychiatric comorbidity in a Danish nationwide population of patients with back pain disorders (BPD) diagnosed in 2010-2014; 2) Investigate whether presence of psychiatric comorbidity affects utilization of pharmacological, surgical and other types of treatment for patients with back BPD; 3) Investigate whether presence of psychiatric comorbidity affects utilization and dosages of pharmacological treatment for patients with BPD.

Clinical contributions of this PhD: This PhD study will contribute with new knowledge on the association between different types of back pain disorders (BPD) and psychiatric comorbidity. As BPD affects a large part of the population, this is a very important area of research from a public health and clinical perspective. Access to nationwide registries facilitates a comprehensive and population-based dataset, ensuring the ideal circumstances for investigating the association between BPD and potential psychiatric comorbidity. Furthermore, the results can be used to increase clinical awareness of vulnerable patients in the intersection between rheumatology and psychiatry and subsequently improve prevention and clinical management for a complex and resource-demanding patient population.

Materials and methods

Setting and study population: This PhD project includes a population-based nationwide cohort based on data from Danish health registries. Danish health registries, including The Danish National Patient Registry (DNPR), a subdivision of the DNPR, the National Patient Registry - Psychiatry (NPD-Psych), The Danish National Prescription Database and The Danish National Health Service Register, contains complete data on hospital contacts, prescription drug use and services provided by health contractors. By using the Danish Civil Registry and the unique personal identification number assigned to all Danish citizens at birth, data across registries can be linked on an individual level.

The study population will be defined as all adult patients (18+) identified in the DNPR, covering all inpatient and outpatient services in Denmark, in the time period 1th of January 2010 to 31th of December 2014 with a diagnosis of BPD according to the ICD-10 Classification of Disease (DM*).

Psychiatric comorbidity will be defined as the presence of any of the following diagnosis, according to ICD-10 classification, obtained from the National Patient Registry - Psychiatry (NPD-Psych), covering all inpatient and outpatient psychiatric services in Denmark in the time period 1th January 2007 to 31th December 2017: Organic, including symptomatic, mental disorders (F00-DF09), Mental and behavioral disorders due to psychoactive substance use (F10-F19), Schizophrenia, schizotypal and delusional disorders (F20-F29), Mood [affective] disorders (F30-F39), Neurotic, stress-related and somatoform disorders (F40-F48), Behavioral syndromes associated with physiological disturbances and physical factors (F50-F59), Disorders of adult personality and behavior (F60-F69), Mental retardation (F70-F79), Disorders of psychological development (F80-F89), Behavioral and emotional disorders with onset usually occurring in childhood and adolescence (F90-F98) and, Unspecified mental disorder (F99). To qualify as a psychiatric comorbidity to back pain disorder, we define that the psychiatric diagnosis must be given no earlier than three years prior to back pain disorder diagnosis and no later than three years after.

Study 1) Occurrence of Psychiatric comorbidity in back pain disorders

Aim: 1) To estimate occurrence of psychiatric comorbidity in patients with BPD compared to the background population and 2) occurrence of psychiatric comorbidity in patients with unspecific BPD compared to patients with specific BPD in a Danish nationwide cohort from 2010-2014.

Hypothesis: 1) There is a higher occurrence of psychiatric comorbidity in BPD patients compared to the background population and 2) patients with unspecific BPD have a higher occurrence of psychiatric comorbidity compared to patients with specific BPD.

Design: The study will be a nationwide register-based matched cohort study.

Procedure: Selected BPD patients with the following diagnosis will be identified via the DNPR: Spondylopathies DM45-49, Other dorsopathies DM 50-54 and Segmental and somatic dysfunction DM99. Each patient will be matched 1:5 on age and sex with patients without back disorders by random at Statistics Denmark. The selection of diagnosis is to be able to match the population with non-BPD patients in the Danish registries.

Information on somatic comorbidity for calculation of Charlson Comorbidity index (CCI) at time of BPD diagnosis, will be retrieved from the DNPR. Data on psychiatric comorbidity will be identified using NPD-Psych. Information on marital status, vital status, and immigration status will be retrieved from the Civil Registration System (CPR). Patients that are not alive or have emigrated within three years after BPD diagnosis are considered lost to follow-up and excluded from the cohort. Descriptive data on the socioeconomic status, education level and equivalent available income will be retrieved from Statistics Denmark. All data comprising each variable will be retrieved from 1th of January 2007 to 31th of December 2017 (three years before and after study period).

Statistics: We will use Fisher's exact test and t-test to compare baseline characteristics with 95% confidence intervals (CI) and p-values. Crude risks of psychiatric comorbidity will be calculated and logistic regression model will be used, to investigate differences between the distribution of risk in BPD patients and the background population. A logistic regression model will also be used to investigate differences between the distribution of psychiatric comorbidity in the unspecific and specific BPD group compared to the background population. The logistic regressions will be reported using 95% CI and adjusted for the following confounders: age, sex, equivalent available income, educational level, CCI, marital status.

Study 2) The effect of psychiatric comorbidity on type of treatment for BPD

Aim: 1) To investigate whether treatment of BPD in patients with psychiatric comorbidity differ from treatment of patients with BPD without psychiatric comorbidity; 2)To investigate whether treatment of BPD differs in the following subgroups: a) specific BPD and no psychiatric comorbidity; b) specific BPD with psychiatric comorbidity; c) unspecific BPD with no psychiatric comorbidity; d) unspecific BPD with psychiatric comorbidity. Investigation will be performed in a nationwide Danish cohort in 2010-2014.

Hypothesis: 1) Psychiatric comorbidity affects type of treatment in BPD patients and 2) psychiatric comorbidity affects type of treatment in subgroups of BPD.

Design and study population: The study will be a nationwide register-based cohort study. The study population consists of patients with BPD and in the sub-analysis the study population consists of patients with BPD and psychiatric comorbidity.

Procedure: BPD patients with any BPD diagnosis (DM*) will be identified via the DNPR. Surgical treatment will be defined as the presence of any procedure code for back surgery retrieved from DNPR (NOMESCO: KNA-W). Pharmacological treatment of BPD will be defined as any treatment with the following drugs: NSAID (ATC: N01A*), acetominophen (ATC: N02BE01), opioids (N02A*), antiepileptica (ATC: N03A) and tricyclic antidepressants (ATC: N06AA)13 identified using the Register of Medicinal Products Statistics. Data on other types of treatment (defined as any treatment not surgical or pharmacological) such as physiotherapist, psychologist, chiropractor or occupational therapist (covered via tax-funded health insurance) will be obtained via The Danish National Health Insurance Service Register (SSR). All BPD treatment-related information will be retrieved up to three years after the date of BPD diagnosis.

Information on somatic comorbidity for calculation of Charlson Comorbidity index(CCI) at time of BPD diagnosis, will be retrieved from the DNPR. Data on psychiatric comorbidity will be identified using NPD-Psych. Information on psychiatric comorbidity will be used to define whether each patient had a psychiatric diagnosis prior to BPD diagnosis, after BPD diagnosis or no psychiatric diagnosis. If a psychiatric diagnosis occurs both prior to and after BPD diagnosis, the patient will be defines as having a psychiatric comorbidity prior to BPD diagnosis.

Information on marital status, vital status, and immigration status will be retrieved from the Civil Registration System (CPR). Patients that are not alive or have emigrated within three years after BPD diagnosis are considered lost to follow-up and excluded from the cohort. Descriptive data on the socioeconomic status, education level and equivalent available income will be retrieved from Statistics Denmark. All data comprising each variable will be retrieved from 1th of January 2007 to 31th of December 2017 (three years before and after study period).

Statistics: We will use Fisher's exact test and t-test to compare patients' baseline characteristics with 95% confidence intervals (CI) and p-values. For each treatment type we will use logistic regression model to investigate differences in outcome distribution between patients with and without psychiatric comorbidity.The logistic regressions will be performed with 95% CI adjusted for the following confounders: age, sex, CCI, level of education, equivalent available income, type of BPD (IBD, DBD or UBP) and marital status.

Study 3) Effect of psychiatric comorbidity in pharmacological treatment of BPD.

Aim: 1) To compare dosages of pharmacological treatment in BPD patients with and without the presence of psychiatric comorbidity in a nationwide Danish cohort study in 2010-2014. 2) To compare high vs. low dosages of pharmacological treatment for BPD in patients with BPD and psychiatric comorbidity.

Hypothesis: 1) Patients with BPD and psychiatric comorbidity are prescribed a higher dosage of pain medication for BPD compared with BPD patients without psychiatric comorbidity. 2) BPD patients with a psychiatric comorbidity with a high (defined as above the 50th percentile) vs. low (defined as below the 50th percentile) dosage of pain medication are also prescribed higher dosages of psychiatric medication.

Design and study population: The study will be a nationwide register-based cohort study. The study population consists of patients with BPD that received pharmacological treatment for BPD. In the secondary analysis the study population consists of patients with BPD and a psychiatric comorbidity, who received pharmacological treatment for BPD.

Procedure: BPD patients with any BPD diagnosis (DM*) will be identified via the DNPR. Pharmacological treatment of BPD will be defined as any treatment with the following drugs: NSAID (ATC: N01A*), Acetominophen (ATC: N02BE01), Opioids (N02A*). Pharmacological treatment for psychiatric comorbidity will be defined as any treatment with the following drugs: Antipsychotics (ATC:N05A), Antidepressants (ATC: N06A), Selective Serotonin Reuptake Inhibitors (SSRI) (ATC: N06AB), and anxiolytics (ATC: N5B*) Pharmacological treatment that can be used in the treatment of both BPD and psychiatric disorders will be defined as any treatment with the following drugs: Antiepileptica (ATC: N03A) and Tricyclic antidepressants (ATC: N06AA). All data on pharmacological treatment will be identified using the Register of Medicinal Products Statistics. From time of BPD diagnosis and the following three years, pharmacological treatment for both somatic illness and psychiatric comorbidity in Daily Defined Dosage (DDD) per year will be calculated.

Information on somatic comorbidity for calculation of Charlson Comorbidity index (CCI) at time of BPD diagnosis, will be retrieved from the DNPR. Data on psychiatric comorbidity will be identified using NPD-Psych. Information on marital status, vital status, and immigration status will be retrieved from the Civil Registration System (CPR). Patients that are not alive or have emigrated within three years after BPD diagnosis are considered lost to follow-up and excluded from the cohort. Descriptive data on the socioeconomic status, education level and equivalent available income will be retrieved from Statistics Denmark. All data comprising each variable will be retrieved from 1th of January 2007 to 31th of December 2017 (three years before and after study period).

Statistics: We will use Fisher's exact test and t-test to compare patients' baseline characteristics with 95% confidence intervals (CI) and p-values. Time dependent Cox regression will be performed separately for patients with specific and unspecific BPD to examine dosage of pharmacological treatment over time comparing patients with and without psychiatric comorbidity. The analysis will be performed with 95% CI and adjusted for the following confounders: age, sex, CCI, level of education, equivalent available income, marital status and the presence of psychopharmacological treatment.

To examine dosage of psychopharmacological treatment over time comparing patients with psychiatric comorbidity and low vs high dose of pharmacological treatment of BPD, time dependent Cox regression will be performed reporting 95% CI and adjusting for the following confounders: age, sex, CCI, level of education, equivalent available income and marital status.