Participants, settings and design
The present study combines data collected as part of two randomized controlled trials (RCTs) investigating the effect of smartphone-based monitoring in patients with BD (the MONARCA I trial and the MONARCA II trial) 29,30. In addition, a group of HC collecting smartphone-based data on cognitive function was included for comparison 31. Thus, data was collected as part of studies published in study protocols á priori.
Patients with bipolar disorder
The MONARCA I trial: The patients were recruited from The Copenhagen Clinic for Affective Disorders, Copenhagen, Denmark during a period from September 2011 to March 2013. The clinic is a specialized outpatient clinic with a catchment area consisting of the Capital Region in Denmark corresponding to 1.4 million people. Patients with a newly diagnosis of BD or with treatment-resistant BD were referred to the clinic. The staff consists of specialists in psychiatry, psychologists, nurses, and a social worker, all with specific experience and knowledge regarding BD. Treatment at the clinic comprises a two years program including combined evidence-based psychopharmacological treatment and supporting therapy, including group psychoeducation 32. The trial had a six-month follow-up period. Inclusion criteria: BD diagnosis according to ICD-10 using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) interview 33, age between 18 to 60 years, a Hamilton Depression Rating Scale 17-item (HDRS-17) score ≤ 17 34 and Young Mania Rating Scale (YMRS) score ≤ 17 35 at the time of inclusion. Exclusion criteria: Pregnancy, a lack of Danish language skills, inability to learn the technicalities for using a smartphone, unwilling to use the trial smartphone as the primary cell phone, and severely physical illness or schizophrenia, schizotypal or delusional disorders according to the SCAN interview.
The MONARCA II trial: All patients with a diagnosis of BD who had previously been treated at the Copenhagen Clinic for Affective Disorder, Copenhagen, Denmark (as described above) in the period from 2004 to January 2016 and who at the time of recruitment were being treated at community psychiatric centres, private psychiatrists and general practitioners were invited to participate in the trial. Patients were included in the study for a nine-month follow-up period if they had a BD diagnosis according to ICD-10 using the Schedules for Clinical Assessments in Neuropsychiatry (SCAN) 33 and previously were treated at the Copenhagen Clinic for Affective Disorder. Patients with schizophrenia, schizotypal or delusional disorders, previous use of the MONARCA system, pregnancy and lack of Danish language skills were excluded. Patients with other comorbid psychiatric disorders and substance use were eligible for the trial.
As part of the MONARCA I trial and the MONARCA II trial, patients were randomized to either using a smartphone-based monitoring system (the Monsenso system) for daily self-monitoring (the intervention group) or to treatment as usual (the control group). Patients included in the intervention group from both trials collected daily smartphone-based self-monitoring data on cognitive function and were included in the analyses in the present report. Inclusion and exclusion criteria were investigated and assessed by two clinical researchers (MFJ and ASJ).
Healthy control individuals
In the present study, a group of HC was included in the analyses to investigate differences in the association between cognitive functioning and stress, quality of life and functioning, respectively with those in the patient population. The HC were part of a larger cohort study 31 and recruited consecutively from the Blood Bank at Rigshospitalet, Copenhagen University Hospital, Denmark, by approaching blood donors in the waiting room on random occasions from September 2015 to August 2016. The inclusion criteria were: age over 18 years, no history of psychiatric illness and no first-generation family history of psychiatric illness. The exclusion criteria were: lack of Danish language skills and pregnancy. The healthy individuals participated as part of a larger cohort study 31.
Daily smartphone-based monitoring
On a daily basis during the follow-up, participants in the trials used a smartphone with the Monsenso app installed and were instructed to use the system for evaluation 29. The app allowed for daily evaluation of cognitive function (evaluated from not present, present to some degree or present (scale: 0, + 1, +2)). Further details regarding the Monsenso system are described elsewhere 29.
Clinical measurements
In the MONARCA I trial outcome measurements were conducted monthly for the entire trial period of six months. In the MONARCA II trial outcome measurements were conducted at baseline, after 4 weeks, 3 months, 6 months and 9 months. The HC used the smartphone-based system on a daily basis for four months and outcome measurements were conducted at baseline and after four months.
All clinical assessments were conducted by researchers (MFJ and HÞ), who were blinded to all smartphone-based data. Thus, data on severity of depressive and manic symptoms as well as functioning were collected rater-blinded.
Clinical rater-blinded assessments: The severity of depressive and manic symptoms were clinically assessed using the HDRS 34 and the YMRS 35. Functioning were clinically rated using the Psychosocial Assessment Short Test (FAST), which is an 24-item interviewer-administrated interview concerning autonomy, occupational functioning, cognitive functioning, financial issues, interpersonal relationships and leisure time 36. Each item is scored from 0 to 3. Higher scores indicate functional impairment and scores above 11 indicate impaired functioning.
Patient-reported questionnaires: The following questionnaires were filled in by the patients at all visits with the researcher: Perceived stress according to Cohen’s Perceived stress scale (PSS), which is a 14-item questionnaire measuring the degree to which situations in one’s life are appraised as stressful. Each item is scored from 0 to 4. Higher scores indicate higher perceived stress 37; Quality of life according to WHO Quality of Life-BREF (WHOQoL), which is a 26-item questionnaire concerning physical health, psychological health, social relationships, environment. Each item is scored from 1 to 5. Higher score indicates better quality of life 38.
Assessments of cognitive function: As part of the MONARCA I trial objectively-evaluated cognitive function with the Screen for Cognitive Impairment in Psychiatry (SCIP) 39,40 and patient-evaluated cognitive function with the Massachusetts General Hospital Cognitive and Physical Functioning questionnaire (CPFQ) 41 were collected at baseline, after 3 months and six months. The SCIP consist of five short performance-based tests that assesses verbal learning, delayed verbal memory, working memory, verbal fluency and processing speed with a high validity and reliability in patients with BD 39,42. The SCIP exists in three parallel versions to minimize learning effects with repeated testing, and in the present study the administration order of the SCIP versions was defined using a randomization list to randomly one of the three SCIP versions to each patient during follow-up. The CPFQ is a 7-item questionnaire which measures patients’ experience of cognitive and physical symptoms. The CPFQ was included to increase the internal validity of the findings by investigating a well-known subjective measure of cognition against both the smartphone-based cognitive function as well as the performance-based cognition measure, SCIP.
Statistical methods
The hypotheses and statistical analyses for the present study were defined á priori.
Since the PSS questionnaire reflects perceived stress during the two previous weeks, the WHOQoL questionnaire reflects quality of life for the previous two weeks, and the FAST rating scale reflects psychosocial functioning during the previous week, mean measures of cognitive function measured using smartphones for the days the scales were reflecting were used in the present report. Calculated total scores for the PSS and the WHOQoL were used in the present study. Since five subitems on the FAST scale reflect cognitive function, in addition to total scores of the FAST, analyses excluding measures on cognitive function (item 10, 11, 12, 13 and 14) on the FAST scale as well as analyses only including measures of cognitive function (item 10, 11, 12, 13 and 14) were conducted in the present study.
In relation to aim 1: To investigate the association between both objectively-measured cognitive function and PSS, WHOQoL and FAST, respectively as well as the association between patient-evaluated cognitive function and PSS, WHOQoL and FAST, respectively, a two-level linear mixed effect model was employed (Tables 2 and 3). In all models, we first considered an unadjusted model (model 1). Secondly, we considered a model adjusted for age and gender as possible covariates. Thirdly, we considered a model adjusted for age, gender, and severity of depressive and manic symptoms according to HDRS and YMRS, respectively as possible covariates. In sensitivity analysis, we restricted all statistical analyses in relation to aim 2 (Tables 2 and 3) to include patients with a HDRS score ≤ 14 and a YMRS score ≤ 14, and in addition for the results presented in Table 4 to include patients with a HDRS score ≤ 7 and a YMRS score ≤ 7.
Table 2
Associations between objective clinically evaluated cognitive function and patient evaluated perceived stress, quality of life and functioning, respectively in patients with bipolar disorder, N = 67
Model 1a Model 2a |
| B | 95% CI | p | B | 95% CI | p |
SCIPb | | | | | | |
PSS | -0.23 | -0.45; -0.0084 | 0.042 | -0.22 | -0.43; -0.013 | 0.038 |
WHOQoL | 0.15 | -0.0043; 0.30 | 0.057 | 0.14 | -0.010; 0.28 | 0.067 |
FAST | -0.13 | -0.28; 0.013 | 0.075 | -0.13 | -0.28; 0.011 | 0.072 |
HDRS | -0.28 | -0.55; -0.0072 | 0.044 | -0.27 | -0.54; -0.0024 | 0.048 |
YMRS | -0.11 | -0.47; 0.25 | 0.55 | -0.15 | -0.50; 0.20 | 0.40 |
a Model 1: Unadjusted. Model 2: Adjusted for age and gender. SCIP: The Screen for Cognitive Impairment in Psychiatry; PSS: Perceived stress measured using Cohen’s Perceived Stress Scale; WHOQoL: Quality of life measures using WHO Quality of Life BREF; FAST: Psychosocial functioning measured using Functional Assessment Short Test; HDRS: Hamilton Depression Rating Scale; YMRS: Young Mania Rating Scale |
Table 3
Associations between patient-evaluated cognitive function measured using smartphone and patient-evaluated perceived stress, quality of life and functioning, respectively in patients with bipolar disorder, N = 117
Model 1a Model 2a Model 3a |
| B | 95% CI | p | B | 95% CI | p | B | 95% CI | p |
Cognitive function measures using smartphonesb | | | | | | | | | |
PSS | 0.033 | 0.024; 0.042 | < 0.0001 | 0.033 | 0.024; 0.042 | < 0.0001 | 0.027 | 0.017; 0.036 | < 0.0001 |
WHOQoL | -0.023 | -0.031; -0.016 | < 0.0001 | -0.023 | -0.031; -0.016 | < 0.0001 | -0.024 | -0.032; -0.017 | < 0.0001 |
FAST | 0.021 | 0.014; 0.027 | < 0.0001 | 0.022 | 0.015; 0.028 | < 0.0001 | 0.014 | 0.0062; 0.021 | < 0.0001 |
FAST excluding cognitive items (item 10–14) | 0.023 | 0.015; 0.031 | < 0.0001 | 0.023 | 0.015; 0.032 | < 0.0001 | 0.017 | 0.0087; 0.025 | < 0.0001 |
FAST including cognitive items (item 10–14) | 0.066 | 0.044; 0.088 | < 0.0001 | 0.069 | 0.047; 0.091 | < 0.0001 | 0.041 | 0.018; 0.064 | 0.001 |
a Model 1: Unadjusted. Model 2: Adjusted for age and gender. Model 3: Adjusted for age, gender, Hamilton Depression Rating Scale 17-items score and Young Mania Rating Scale score. b Scored on a scale from 0, 1, 2. PSS: Perceived stress measured using Cohen’s Perceived Stress Scale; WHOQoL: Quality of life measures using WHO Quality of Life BREF; FAST: Psychosocial functioning measured using Functional Assessment Short Test |
Table 4
Associations between patients-evaluated cognitive function and objective clinically evaluated cognitive function in patients with bipolar disorder
Model 1a Model 2a Model 3a |
| B | 95% CI | p | B | 95% CI | p | B | 95% CI | p |
Cognitive function measured using smartphonesb, n = 33 | | | | | | | | | |
SCIP, overall scorec | -0.00075 | -0.0096; 0.0081 | 0.87 | 0.0017 | -0.0078; 0.011 | 0.73 | 0.00090 | 0.0017; 0.016 | 0.015 |
Verbal learning and memory | 0.0041 | -0.022; 0.030 | 0.76 | 0.0080 | -0.018; 0.034 | 0.55 | 0.025 | 0.0050; 0.045 | 0.013 |
Delayed memory | -0.0063 | -0.039; 0.026 | 0.71 | 0.0038 | -0.031; 0.039 | 0.83 | 0.033 | 0.0055; 0.061 | 0.018 |
Working memory | 0.0032 | -0.016; 0.023 | 0.75 | 0.0054 | -0.014; 0.025 | 0.60 | 0.0031 | -0.14; 0.021 | 0.72 |
Word mobilization | -0.037 | -0.051; 0.044 | 0.88 | 0.0031 | -0.045; 0.052 | 0.90 | 0.031 | -0.0067; 0.069 | 0.11 |
Processing speed | -0.032 | -0.074; 0.011 | 0.14 | -0.026 | -0.076; 0.023 | 0.30 | -0.0068 | -0.051; 0.037 | 0.76 |
CPFQd | 0.023 | 0.014; 0.032 | < 0.0001 | 0.023 | 0.014; 0.032 | < 0.0001 | 0.019 | 0.0087; 0.030 | < 0.0001 |
CPFQd n = 67 | | | | | | | | | |
SCIP, overall scorec | -0.091 | -0.20; 0.022 | 0.11 | -0.077 | -0.20; 0.042 | 0.21 | -0.00030 | -0.092; 0.092 | 0.99 |
Verbal learning and memory | -0.49 | -0.82; -0.15 | 0.005 | -0.46 | -0.82; -0.11 | 0.010 | -0.018 | -0.30; 0.27 | 0.90 |
Delayed memory | 0.014 | -0.30; 0.39 | 0.94 | 0.069 | -0.33; 0.46 | 0.73 | 0.13 | -0.17; 0.43 | 0.40 |
Working memory | 0.016 | -0.21; 0.24 | 0.89 | 0.023 | -0.21; 0.26 | 0.85 | 0.0065 | -0.20; 0.18 | 0.95 |
Word mobilization | -0.30 | -0.31; -0.09 | 0.024 | -0.67 | -1.31; -0.26 | 0.041 | -0.087 | -0.59; 0.42 | 0.74 |
Processing speed | -0.50 | -1.06; 0.051 | 0.075 | -0.43 | -1.06; 0.20 | 0.18 | -0.19 | -0.66; 0.29 | 0.44 |
a Model 1: Unadjusted. Model 2: Adjusted for age and gender. Model 3: Adjusted for age, gender, Hamilton Depression Rating Scale 17-item score and Young Mania Rating Scale score b Scored on a scale from 0, 1 ,2 c SCIP: The Screen for Cognitive Impairment in Psychiatry d CPFQ: The Massachusetts General Hospital Cognitive and Physical Functioning Questionnaire |
In relation to aim 2: To investigate the association between patient-evaluated cognitive function and objectively-measured cognitive function a two-level linear mixed effect model, which accommodates both variation of the variables of interest within patients (intra-individual variation) and between individuals (inter-individual variation) was employed. The models included a fixed effect of time (measurement visit number) and a patients-specific random effect allowing for individual intercept and a slope for each participant (Table 4). In addition, the association between these measures according to the sub-domains on the SCIP reflecting verbal learning and memory, delayed memory, working memory, verbal fluency and processing speed were employed. Model 1 reflects unadjusted analyses, model 2 includes age and gender as covariates, and model 3 includes age, gender, HDRS and YMRS score as covariates.
In relation to aim 3: To investigate differences in the association between cognitive function measures using smartphones and perceived stress, quality of life and functioning in patients with BD and HC, interactions between the groups (BD or HC) and outcome measures were investigated for each of the considered models and reported accordingly.
As few prior studies have investigated associations between daily patient-reported cognitive function and perceived stress, quality of life and functioning, respectively, in patients with BD, we were not able to make statistical power analyses prior to the study since potential effects were unknown. However, to account for multiple testing in the statistical models the Bonferroni correction method was used. Thus, p-values below 0.0023 in the individual analyses were considered statistically significant. Model assumptions were checked visually by means of residuals and QQ plots for each of the statistical analyses. Data were entered using Excel and Epidata®, STATA (StataCorp LP, College Station, TX, USA) version 13 was used for statistical analyses.
Ethical considerations
The trials were approved by the Regional Ethics Committee in the Capital Region of Denmark (H-2-2011-056, H-2-2014-059 and H-7-2014-007) and the Danish Data protection agency (2013-41-1710). The law on handling of personal data was respected. Prior to commencement the trials were registered at ClinicalTrials.gov (NCT01446406 and NCT02221336). Electronic data collected from the smartphones were stored at a secure server at Concern IT, Capital Region, Denmark (I-suite number RHP-292 2011-03). The trial complied with the Helsinki Declaration of 1975, as revised in 2008.