Key words affective disorders - bipolar disorder - longitudinal tracking - digital mood charting
Introduction
Affective disorders, including major depressive disorder (MDD) and bipolar disorders
(BD), are common, recurrent, and complex mental disorders with enormous consequences
for patients, families, and society [1 ].
Depressive disorders are a major burden on a society, including medical costs, drug
costs, workplace absenteeism, and reduced productivity [2 ]
[3 ].
The lifetime course of depressive illness affects individual suffering and
functioning as well as the socio-economic burden [4 ]
[5 ]
[6 ]
[7 ].
The median age of onset of MDD and BD is between 20 and 30 years [8 ], followed by a large variation in disease
trajectories. Although some people only experience a single episode of depression
[9 ], about 70–90% of
patients with affective disorders have at least one recurrence [10 ]
[11 ].
For both MDD and BD, the risk of recurrence increases with the number of prior
episodes [12 ].
Importance of longitudinal study
The frequent recurrences and serious consequences of affective disorders
emphasize both the need to understand individual experiences and the importance
of a collaborative approach to treatment [13 ]. A longitudinal study of the course of affective disorders is an
essential methodology for understanding disease trajectories, treatment effects,
symptom changes, and long-term outcomes. From a historical perspective, Emil
Kraepelin used Zählkarten (diagnostic cards) to describe patient
demographics and psychopathology and to study the longitudinal course of
patients with psychotic disorders [14 ]
[15 ]. The data on these
cards eventually led to Kraepelin's dichotomous classification: the
distinction between manic-depressive insanity and dementia praecox
(schizophrenia).
Today, the value of a longitudinal study of the core symptoms of affective
disorders is well established. Longitudinal tracking is essential for detecting
prodromal phases of recurrences, recognizing patterns of longitudinal disease
course, and evaluating treatments. Teaching patients to recognize early signs of
recurrence and disease patterns is an important, effective, and empowering
approach to improving disease management [16 ].
Tools for the longitudinal study
One commonly recommended intervention for the longitudinal study of affective
disorders is daily self-charting of mood, sleep, medication, life events, and
other illness-related variables. Daily patient mood charting complements
clinician monitoring and can easily be combined with other psychosocial
interventions that benefit both the patient and the clinician. Daily recording
may capture observations that would be missed if data were only collected at
clinic visits, and help differentiate between episodic and non-episodic
disorders and detect subsyndromal symptomatology. Several paper-based
instruments for daily mood recording have been developed, including the Life
Chart Methodology [17 ] and STEP-BP Mood
Chart [18 ]. ChronoSheet is a paper-based
form for self-reporting developed by Peter Whybrow during the 1980s to follow
the clinical course of patients with rapid-cycling variants [19 ]. Paper-based self-reporting instruments
have been successfully used in clinical practice and in longitudinal research
studies to characterize the long-term course of affective disorders. However,
there are problems associated with paper-based tracking, including poor
compliance, data entry errors, high costs for data entry, and limited feedback
for the patient and clinician [20 ].
Automated longitudinal tracking
Until the early 2000s, longitudinal tracking and daily self-assessment of
patients with AD was primarily paper-based. With the widespread acceptance of
home computers in the early 2000s, automated patient tools were developed for
mood charting, such as ChronoRecord [21 ],
based upon the ChronoSheet [19 ]. Automated
charting tools eliminated many of the problems of paper-based tools, including
the expense of data entry for research studies. Patients installed ChronoRecord
on a home computer and recorded mood, sleep, medications, and life events daily,
and body weight weekly. ChronoRecord provided monthly charts that displayed the
daily patient data that could be printed on demand for both patients and
physicians ([Fig. 1 ]). ChronoRecord was
validated by patients with bipolar disorder, demonstrating that patients would
accept a technology solution and reliably report daily mood [21 ]
[22 ]. The immediate feedback improved motivation of the subject to
complete a long-term study for many patients and enabled the timely use of
clinical information captured during a longitudinal study.
Fig. 1 ChronoRecord: 180 days of electronic charting of a
41-year-old male with bipolar I disorder.
This paper summarizes the daily mood, sleep, and medication data collected with
ChronoRecord, and highlights some key research findings. Lessons learned from
implementing a computerized tool for patient self-reporting are also
discussed.
Methods
Over the last 20 years, ChronoRecord software was used to collect data from 609
patients with affective disorder. In total, 692 patients were registered to use
ChronoRecord. Eighty patients did not return any data and 3 patients were missing a
diagnosis and excluded from all analyses. All study participants were outpatients
with a diagnosis of affective disorder based on DSM IV or DSM 5 criteria. The
diagnosis of affective disorder was made by a psychiatrist in a clinical interview,
and all patients received treatment as usual. All participants were volunteers
recruited from a university mood clinic or private practice at multiple sites [21 ] and were informed about the study prior to
providing written informed consent. The study was approved by each local
institutional review board. ChronoRecord software was translated into several
languages, including German and Spanish, which allowed the standardization of
international data for analysis. All ChronoRecord data stored on the computer of the
patient or physician, or the transferred data, are encrypted using a unique
password. To maximize patient security and privacy, no patient data is stored on a
web server.
Demographic data was obtained by a clinician at the time of enrollment. All data on
medications taken, mood, and sleep were entered daily by the patient using
ChronoRecord software in their native language. Patents received about a half hour
of training on the use of ChronoRecord software, in person or by telephone, before
entering data. During the training session, a medication list was created for each
patient that included all drugs prescribed for bipolar disorder and any other
prescribed or over-the-counter (OTC) drugs that the patient felt impacted their
mood. The prescribed psychiatric drugs were selected from a list in the software,
displayed by both brand and generic names. The patient could add a drug not found in
the software and modify their medication list at any time. For each selected drug,
the pill strength was chosen from a list of available strengths. Every day, for each
drug, the patient entered the total number of pills taken.
ChronoRecord uses a 100-unit visual analog scale between the extremes of mania and
depression to rate mood. The mood categories were based on the validation studies
[21. 22], which compared patient self-ratings on ChronoRecord to clinical ratings on
the Hamilton Depression Rating Scale (HAMD) and Young Mania Rating Scale (YMRS)
[23 ]
[24 ]. A mood entry less than 40 was considered depression, 40–60
euthymia, and greater than 60 hypomania/mania. The depression ratings varied
from mild (entry of 20–39) to severe (entry of 0–19), and the mania
ratings varied from hypomania (entry of 61–80) to severe (entry of
81–100). During patient training, anchor points were set by having the
patient describe the most depressed and most manic states they had ever experienced.
Patients often described a mixed episode as the most manic experience. The
instructions given were to enter a single daily mood rating that best describes
their overall mood for the prior 24 hours, calibrate the mood rating to the
anchor points, and try not to let the previous day influence how the current day is
rated.
For this summary analysis, the descriptive statistics for the demographic and
clinical characteristics of the 609 patients were calculated using SPSS version
28.0.
Results
The demographic characteristics of patients who used ChronoRecord are shown in [Table 1 ]. The 609 patients had a mean age of
40.3±11.8 years, a mean age of onset of 22±11.2 years, and were
71.4% female. The average number of days of data was 227.
Table 1 Patient demographics
(N=609)*
Demographic
Category
N
Percent
Gender
Male
174
28.6
Female
435
71.4
Diagnosis
BP I
335
55.0
BP II
192
31.5
BP NOS
31
5.1
Unipolar
51
8.4
Disabled
No
424
76.3
Yes
132
23.7
Full-time employment
No
306
55.0
Yes
250
45.0
College graduate
No
270
46.8
Yes
307
53.2
Married
No
287
49.7
Yes
290
50.3
N
Min
Max
Mean
SD
Hospitalizations
572
0
30
2.2
4.0
Age of onset
576
3
75
22.0
11.2
Age
609
13
79
40.3
11.8
Years of illness
574
0
55
18.5
12.5
Number of mood days
609
1
4769
227.0
456.6
* 692 patients registered to use ChronoRecord. Eighty
patients did not enter any data. Three patients did not have a diagnosis
code and were deleted from all analyses.
The percent of time spent in the mood groups by diagnosis is shown in [Table 2 ]. When considering all patients,
70.8% of days were euthymic, 15,1% were mild depression,
6.6% were severe depression, 6.6% were hypomania, and 0.8%
were mania.
Table 2 Daily patient mood ratings by mood group and diagnosis
(N=138246 days)
Diagnosis
Bipolar I
Bipolar II
Bipolar NOS
Unipolar
All
Mood Group
Days
Percent
Days
Percent
Days
Percent
Days
Percent
Days
Percent
Severe depression
2939
4.4
5533
10.0
20
0.6
679
5.1
9171
6.6
Mild depression
10421
15.7
7704
14.0
376
11.1
2420
18.0
20921
15.1
Euthymia
47310
71.4
38846
70.4
2629
77.5
9075
67.5
97860
70.8
Hypomania
5076
7.7
2481
4.5
356
10.5
1258
9.4
9171
6.6
Mania
514
0.8
595
1.1
10
0.3
4
0.0
1123
0.8
All
66260
100.0
55159
100.0
3391
100.0
13436
100.0
138246
100.0
Author: please check if this is correctly spelled out? NOS: not
otherwise specified
The daily hours of sleep by mood group for all patients is presented in [Fig. 2 ], with 12.4% of days with less
than 6 hours of sleep, 51.7% between 6 and 9 hours of sleep,
and 35.8% of days with greater than 9 hours of sleep.
Fig. 2 Daily hours of sleep by mood group.
The number of daily psychotropic medications taken by the mood group is shown in
[Table 3 ]. When considering all mood
groups, 22.4% took 1–2 medications, 37.2% took 3–4
medications, 25.7% took 5–6 medications, 11.6% took
7–8 medications, and 3.1% took >8 medications. The most
frequently taken psychotropic medications are shown in [Fig. 3 ].
Fig. 3 Most frequently taken daily psychotropic medications.
Table 3 Daily number of psychotropic medications by mood group
(N=135647)
Mood Group
Number of Medications
Depressed
Euthymic
Manic
All
Days
Percent
Days
Percent
Days
Percent
Days
Percent
1 Medication
1066
3.6
7555
7.9
814
8.1
9435
8.0
2 Medications
3061
10.4
16640
17.3
1239
12.3
20940
15.4
3 or 4 Medications
9769
33.1
36754
38.3
3935
39.2
50458
37.2
5 or 6 Medications
9247
31.3
23131
24.1
2541
25.3
34919
25.7
7 or 8 Medications
4744
16.1
9795
10.2
1133
11.3
15672
11.6
> 8 Medications
1646
5.6
2201
2.3
376
3.7
4223
3.1
All
29533
100.0
96076
100.0
10038
100.0
135647
100.0
Discussion
For most patients, affective disorders are a recurrent, serious, life-long illness.
Given the substantial heterogeneity in the polarity, frequency, and severity of
episodes, a longitudinal approach is fundamental to understanding the course of
affective disorders. The daily self-reported data from patients using ChronoRecord
provided an overview of patient status and useful information about affective
disorders. The most frequent symptom that patients with bipolar disorder experienced
was depression ([Table 2) ], as found
previously [25 ]. Mild symptoms of both
depression [26 ]
[27 ] and hypomania [28 ] occurred frequently outside of episodes.
Subsyndromal symptoms may be associated with considerable functional impairment,
including unemployment [29 ]
[30 ].
The most useful sleep parameter for self-monitoring was sleep duration [31 ]. In some patients with bipolar disorder, a
change in sleep duration of more than 3 hours may signify an imminent mood
change [32 ]
[33 ]. Additionally, in patients with bipolar disorder, greater serial
irregularity in the mood time series was found before the onset of an episode [34 ].
Most patients take a unique drug regimen of polypharmacy for their affective disorder
([Table 3 ]); [35 ]
[36 ].
Considerable non-adherence with taking mood stabilizers was found, even among these
patients who were motivated to complete daily mood charting [37 ]. There was frequent irregularity in the
daily dosage of all psychotropic medications taken due to daily omissions and dosage
changes [38 ]
[39 ].
Other clinical knowledge was gained from the use of ChronoRecord. Patient recording
of the daily mood rating can provide important information about the course of
illness that is not available to the physician when the patient is only assessed at
routine office visits. The graphical illustration provided by ChronoRecord ([Fig. 1 ]) can help the clinician assess the
patterns of illness over time, determine the association between symptoms and events
such as hormonal changes, seasonal changes, and psychosocial stressors, and evaluate
treatment response, especially when multiple medications are involved. Automated
mood charting also provides immediate benefits for the patient. Patients who used
ChronoRecord frequently commented that the mood chart helps them communicate with
the psychiatrist. Instead of just focusing on how they feel the day of the
appointment, they can talk about the time between visits.
Challenges of automating mood charting
There are challenges when implementing any technology tool in clinical medicine.
The primary challenge of implementing ChronoRecord, or any automated system for
mood charting, is that a modification of the daily work routine is required.
Clear procedures for clinicians, administrative staff, and patients must define
the steps involved in registration, training, data collection, and ongoing
technical support. Regardless of the technology involved, both staff and
patients must receive adequate training to use a new technology, an adequate
budget must be allocated, and a staff member must be assigned overall
responsibility for the ongoing use of the new technology. Technology platforms
change over time, and products and procedures may need to be updated.
Other lessons were also learned from the development and implementation of
ChronoRecord. A product designed for the general public must be easy to use and
the output easy to understand. Initial training and ongoing human support for a
technology product should be available to both patients and physicians.
Encouragement from the physician is important to increase patient use of
technology. However, the individual patient must be interested in mood charting,
and not all patients are. The ChronoRecord data was collected as a convenience
sample which required testing for sample bias before publishing results.
ChronoRecord also showed that patient data entry could be collected and monthly
charts distributed securely using an encrypted attachment to an email.
In conclusion, many patients with affective disorders will accept and use a
technology product for recording daily mood ratings over long periods. The tool
is useful for increasing individual participation in their care, providing
detailed data to the physician, and providing data for research, including
longitudinal studies. Patient mood charting with a validated tool such as
ChronoRecord has helped confirm that daily self-reporting is useful in clinical
as well as research settings.
Author Contributions
MB and TG wrote the initial draft. All authors reviewed and approved the final
manuscript.
