Interrelationships Among Inflammation, Physical Symptoms, Mental Symptoms and Quality of Life in Subjects With Chronic Inflammatory Physical Diseases.
Interrelationships Among Inflammation, Physical Symptoms, Mental Symptoms and Quality of Life in Subjects With Chronic Inflammatory Physical Diseases.
The prevalence of common mental disorders is high in patients with chronic inflammatory physical diseases(e.g., autoimmune or infectious diseases). The traditional explanatory causation model in which physical symptoms and related disability drive mental health problems is now called into question, and evidence has accumulated supporting more complex interactions whereby psychiatric disorders can both result from and contribute to the progression of physical diseases. In the present project, the investigators will focus on comorbidity of depression and anxiety symptoms or syndromes with chronic inflammatory skin diseases (psoriasis, hidradenitis suppurativa and atopic dermatitis) or chronic infectious diseases (chronic HBV and HIV infection).
The study is aimed to clarify the mechanisms underlying the high frequency of those comorbidities. It will overcome the main limitations of previous investigations and use innovative statistical tools to model complex interrelationships and causal links among the assessed variables.
The identification of key variables driving the causal chain of determinants of poor global health and quality of life may impact treatment outcome and models of care.
STATE OF ART:
Healthcare professionals often underestimate the effects of chronic inflammatory processes on patients' quality of life (QoL). The prevalence of common mental disorders, such as depression and anxiety, is high in patients with physical diseases entailing chronic inflammatory processes (e.g., autoimmune or infectious diseases), adding to the psychosocial burden of these disorders. In fact, mental symptoms may harm response to treatment and may have a negative impact on the physical diseases, which, in their turn, may worsen mental symptoms, and contribute, together with them, to undermine persons' QoL.
Associations of depression and anxiety with dermatological and infectious diseases (entailing chronic inflammatory processes) have been the subject of an extensive literature. Depression is reported in up to 42% and 60% of people with hidradenitis suppurativa (HS) and psoriasis (PS), respectively. Patients with atopic dermatitis (AD) and HS seem to have a high suicidal risk. High comorbidity with depression and anxiety is also reported in chronic infectious diseases such as chronic hepatitis C or B (HCV, HBV), and human immunodeficiency virus (HIV) infection. Prevalence of anxiety symptoms is as high as 80% in people living with HIV, with 20% meeting criteria for generalized anxiety disorder. HBV and HCV are associated with clinically significant depression in up to 30% of cases.
Research, so far, has not succeeded at fully clarifying mechanisms underlying the high frequency of comorbidity between common mental disorders and chronic inflammatory diseases.
The presence of mental disorders adds to the psychosocial burden of chronic inflammatory skin diseases or chronic infectious diseases, may harm response to treatment and may have a negative impact on the physical diseases themselves, which, in their turn, may worsen mental symptoms, and contribute, together with them, to undermine persons' QoL.
The project's ambitions are manifold and are listed below.
In addition, the investigators will use network analyses to model the complex interrelationships and causal links among the assessed variables. So far, no study modelled the comorbidity between mental and physical symptoms in chronic inflammatory diseases by using this approach. The network analysis does not rely on an a priori model of cause-effect relationships among variables, as current alternative statistical models (e.g., structural equation models) do; it is largely used in physical and communication sciences to study how complex interactions among sets of variables maintain complex systems. It has been used to improve knowledge on the complex phenomena of psychiatric comorbidity, and identify high value targets for interventions. More recently, network modeling algorithms have been used to study causal relationships among sets of variables and identify those variables that in the causal chain hold precedence and represent "activators,", "mediators,", and/or "products" within a syndromic condition. The identification of key variables in the network, and in particular of those that represent causal antecedents in the chain of variables leading to poor QoL, is crucial to design treatment programs that target those variables, increasing the likelihood to improve patients' QoL. In addition, as the analysis enables the possibility to identify the influence of unmeasured ("latent") variables not included in the model, the project's findings may guide further research in the field.
OBJECTIVES:
The project aims to clarify the mechanisms underlying the high frequency of comorbidity of physical diseases entailing chronic inflammatory processes (e.g., autoimmune or infectious diseases) with common mental disorders.
The traditional explanatory causation model in which physical symptoms and related disability drive mental health problems is called into question by the relevant literature, and evidence has accumulated supporting more complex interactions whereby psychiatric disorders can both result from and contribute to the progression of physical diseases. In this frame, recognizing the co-existence of physical and mental disease symptoms pertains not only to the need to treat both conditions, but also to the need to avoid a vicious circle in which mental symptoms may harm response to treatment and may have a negative impact on the physical ones that, in their turn, may worsen mental symptoms, and contribute, together with them, to undermine person's disability and quality of life.
To these aims, the present project will pursue the following objectives.
To investigate the frequency of depressive and anxiety symptoms and disorders in 500 clinically stable, outpatients affected by one of the following dermatological diseases: psoriasis, hidradenitis suppurativa or atopic dermatitis, who are being treated for the disorder; as well as in 500 clinically stable outpatients with documented HIV or HBV infection, on treatment with antiviral therapy.
To demonstrate that in both samples the presence of clinically significant depressive or anxiety symptoms is associated with worse global health and QoL.
To identify mechanisms responsible for the vicious circle in which mental symptoms can be promoted by proinflammatory cytokines and immune deficiency; in their turn can stimulate the production of proinflammatory cytokines and down-regulate the cellular immune response, and can contribute to prolonged infection and delayed healing, as well as to functional decline.
To test the hypotheses listed below by the network analysis, using a machine-learning algorithm:
To identify key determinants of disability and poor quality of life, and improve treatment planning by addressing possible unmet treatment targets.
To uncover the possible influence of unmeasured variables not included in the model and to be investigated in future studies.
STATISTICAL ANALYSIS
Data will be checked for missing values, inconsistencies, and outliers. Continuous variables will be summarized by the following descriptive statistics: number of observations, mean, standard deviation, median, minimum, and maximum; while categorical ones by counts of patients and percentages.
The normality assumption will be tested with Shapiro-Wilk test. Pairwise correlations between continuous study measures will be calculated using Pearson correlation coefficients or Spearman rank correlation coefficients if needed.
The statistical testing will be conducted at the two-sided a = 0.05 and 95% confidence interval (CI) will be computed, unless otherwise specified.
For continuous variables, differences among groups will be tested by means of t-test or one-way ANOVA and Mann-Whitney or Kruskal-Wallis tests when appropriate. For categorical variables, differences among groups will be tested by means of chi-square or Fisher's exact test when needed.
The study hypotheses will be tested in the overall sample and separately in the two subgroups of patients with dermatological diseases or infectious diseases using network analysis, an exploratory technique that is increasingly used to identify causal relationships within and between manifestations of psychiatric disorders.
A causal network is represented graphically as a set of nodes (variables) connected through edges (directed arrows), that may be interpreted as causal relations. Specifically, the complex relationships among variables will be investigated by means of partial ancestral graphs (PAGs), which use an enriched set of edge types to convey both edge orientation and the possible influence of unmeasured variables not represented in the model. Analyses will be conducted using the machine-learning Greedy Fast Causal Inference (GFCI) algorithm, which uses a 2-step process to make causal inferences that can be represented in graphical format as PAGs. In the first step, all possible non-recursive causal relationships among the variables will be searched automatically to establish which variable pairs are potentially causally related using a method called Fast Greedy Equivalence Search. In the second step, the preliminary assessment will be refined by performing a series of conditional independence tests to iteratively rule out causal models that imply conditional independence statements not found to be true of the data. At the end of this process, it will be possible to interpret the links between pairs of variables as direct causal relationships, indirect causal relationships, possible causal relationship confounded by unmeasured variables, or correlations, in which the direction of causality cannot be defined. Thus, the GFCI algorithm will help understand which of the study hypotheses is/are more plausible and compatible with the data.
All statistical analyses and data processing will be performed using R version 3.3.3 (R Foundation for Statistical Computing).
Sample size considerations:
Among constraint-based causal search algorithms, GFCI showed better results in terms of precision and good performance in the simulations with a sample size of 200. Our sample size of 1000 patients (500 patients for each subsample) is adequate to detect causal relationships among the study variables, with a ratio patients/variables of about 30/1 in each subgroup.
IMPLEMENTATION: PERFORMANCE OF THE INTERMEDIATE AND FINAL OBJECTIVES
During the first 3 months of the project, the following procedures will be completed: enrollment of five research assistants (one biologist or one psychiatrist with experience in procedures for assessment of plasma biomarkers of inflammation, two dermatologists, one infectious disease specialist and one statistician for data quality control); acquisition of equipment and consumables for the study (7.5 Mhz probe for color-doppler instrument, bone quantitative ultrasound device, ELISA kits for dosage of blood markers; computerized neurocognitive test battery CANTAB; 5 portable computers, 1 personal computer, 3 iPAD, 3 STATA/SE version 15; stationery); implementation and testing of the study database by a computer engineer under the supervision of the coordinating unit; training of the researchers to the use of the study instruments and to the correct study procedures, carried out by each research unit for the respective assessments.
At the beginning of month 4, the recruitment of the subjects will start and will be carried out during the following 18 months.
Confirmation of patients' eligibility will be done by associate professors and researchers of Psychiatry, Dermatology and Infectious diseases research units, who will review the inclusion/exclusion criteria relevant to their discipline, such as the diagnosis of skin diseases or infectious diseases listed in the inclusion criteria, as well as lifetime psychiatric diagnoses. Associate professors and researchers from the Dermatology and Infectious diseases research unit will also obtain informed consent from eligible patients.
Trainees in Psychiatry from the Psychiatry research unit will collect data relevant to socio-demographic characteristics, psychopathology, disability, quality of life, cognition, coping skills and stigma.
Assessment of dermatological indices will be performed by the dermatologists enrolled as research assistants in the project, as well as by the researcher and trainee of the Dermatology research unit.
Assessment of markers of subclinical inflammatory damage will be done by the associate professor and research assistant of the Infectious Diseases research unit. Measurement of plasma levels of inflammatory markers will be done by a research assistant enrolled in the project (a biologist or a psychiatrist with experience in procedures for assessment of plasma biomarkers) under the supervision of a researcher of the Psychiatry research unit.
Trainees will fill in data collected for their respective research unit in the electronic database.
To control for missing data or insufficient sample size, every three months, recruitment strategies and subject enrollment will be reviewed by an ad hoc appointed internal audit committee. Incomplete data for any recruited subject will be allowed in the measure of 5%. Subjects missing all data in one of the main assessment areas (biomarkers, mental symptoms, physical and primary disease assessments) will be replaced.
Intermediate objectives of the study are the followings: 1) recruitment of at least 400 patients (200 from the Dermatological unit and 200 from the Infection disease unit) by the end of month 12; 2) Study registration and publication of the study protocol and its dissemination in national and international congresses by the end of month 9; 3) publication of two systematic reviews on the comorbidity of common mental disorders with chronic inflammatory skin and infectious diseases by month 12.
Final objectives of the project are: 1) the recruitment of at least 700 subjects with complete data by month 21; 2) consultation with a statistician with specific expertise in network analysis and completion of final statistical analyses by the end of month 22; 3) publication of the results in high-impact international journals and their dissemination in national and international congresses.
Inclusion criteria for subjects with dermatological diseases;
Inclusion criteria for subjects with HIV or HBV:
Inclusion criteria for both groups:
Exclusion Criteria:
- a history of intellectual disability, bipolar disorder, psychosis or schizophrenia, or current high suicide risk.