WO2015003247A1

WO2015003247A1 - Systems and methods relating to subject monitoring, assessment and treatment

Info

Publication number: WO2015003247A1
Application number: PCT/CA2014/000560
Authority: WO
Inventors: Yik Chau LUI; Martin KATZMAN; Jonathan POLAK
Original assignee: 1Datapoint Diagnostics Inc.
Priority date: 2013-07-10
Filing date: 2014-07-10
Publication date: 2015-01-15

Abstract

A mobile system includes processing structure configured to electronically transmit collected data about ambient factor and/or usage of the mobile system to a remote processing structure; receive, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data; at each of the provided dates/times, automatically trigger execution of one or more user-interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and electronically transmit the received user data to the remote processing structure, A system includes a plurality of mobile systems and a remote processing structure in electronic communication with each of the mobile systems, the remote processing structure comprising at least one database for storage and aggregation of respective collected data and user data received from each of the plurality of mobile devices.

Description

SYSTEMS AND METHODS RELATING TO SUBJECT MONITORING, ASSESSMENT

AND TREATMENT

Field of the Invention

[0001] Embodiments of the present invention relate generally to tools for use in monitoring subjects such as medical patients in connection with medical assessment and treatment, and more particularly to systems and methods for use in monitoring and erring behaviour of a subject.

Background of the Invention

[0002] Comparison of treatment to placebo - specifically for psychiatric drugs - is much contested. Kirsch & Sapirstein (1998) reporting on 19 placebo controlled trials of antidepressants, noted placebo groups averaged a 1.5 standard deviation (s.d.) units of improvement, which was 75% of the overall progress shown by the drug groups. Furthermore, they noted that the superiority of the drag group over the placebo groups was only 0.5 s.d. Further support for this view comes from earlier studies. Others (Quality Assvurance Project, 1983; Joffe et al, 1996) reported that the size of the progress attributed to the placebo group in depression trials was greater than the additional advantages attributed to the drugs over the placebo. Furthermore, irsh and Sapirstein (1998) also showed that with a correlation of 0.9 between the placebo effect and drug effect, virtually all the variation between the improvements in the drug-treated groups in the different trials could be predicted by the response in the subjects randomized to the placebo groups.

[0003] Further, Kirsch & Sapirstein (1998) in reviewing their data, concluded that one-quarter of the improvement observed in the drug-treated group was due to the active medication, one-quarter to natural history and half to the placebo effect They then raised the possibility that the improvement attributed to the drug could even be a non-specific response to the side-effects generated by the medication. Further support for this view comes from Moncrieff et al (1998), who noted that the superiority of drug over the active placebo atropine was reduced from an effect size of 0.50 in non- active placebo trials to an effect size of 0.21 with active placebos, consistent with the Kirsch & Sapirstein suggestion that people in trials respond more positively if they experience side-effects (Andrews; 2001).

10004] Still, others have disagreed. Quitkin et al (2000) after systematically reviewing the methodological issues raised by Kirsch & Sapirstein, concluded that, despite the large response in the placebo group, antidepressants are treatments that produce specific, additional benefit. Furthermore, Enserink (1999). noted that difficulties facing drug trials when compared with a placebo group is disproportionately large compared with that of the drug-treated group. As a result, there remains a great deal of discussion in the general media as to the value of antidepressants. The public have views about the benefits of antidepressants that are quite negative (Jorrn et al, 2000) and such negative press can only further restrict the low rates of depression treatment. Poor coverage is an important public health problem (Andrews et al, 2000), given that depression ranks fourth in the world in terms of the global burden of disease (Murray & Lopez, 1 96). It is important that health professionals and the public have correct information about the small but definite benefit that antidepressants can offer.

[0005] One of the problems that conveys the message of limited efficacy of antidepressants is the abnormally high placebo response rate in major depressive disorder. The Quality Assurance Project (QAP) conducted a series of meta-analyses across the major mental disorders in the mid-1980s (Quality Assurance Project, 1982, 1983, 1984, 1985a,b), with similar methodology of effect size estimation for all studies allowing comparison of all the results can be compared. The authors reported that in depression, the placebo groups improved by 0.93 s.d. units, the active treatment groups by 1.54 s.d. units, with the placebo groups making 60% of the progress recorded in the drug groups. They also noted that in no other disorder was improvement while on placebo this large, either in absolute terms or as a proportion of the change in the treatment group, as depicted in Table 1 below (Adapted from the Quality Assurance Project, 1982, 1983, 1984, 1985a, b):

Table 1

[0006] In depression, therefore, the extent of the placebo response is unusually large.

[0007] Andrews (2001), stated that change in any placebo group occurs for three main reasons.

These are: the encouraging effect of being in treatment; the effect of spontaneous remission while in treatment; and that people with chronic symptoms normally seek help when their symptoms are worst and, through natural fluctuations in severity, are likely to be improved when next assessed.

[0008] Andrews (2001) reports that the encouraging effect of being in treatment, and the effect of spontaneous remission while in treatment are particularly true of studies in depression, in part related to the self-defeating nature of depression being sensitive to the encouragement that comes from being in treatment.

[0009] Good clinical care (Andrews, 1 93) consists of a review of what the patient did and did not do, with encouragement to resolve problems and resume positive activity. Structured problem- solving and activity scheduling are systematic approaches to achieve these goals (Mynors-Wallis et al, 1995; Andrews & Jenkins, 1999) that have been demonstrated in randomized controlled trials to be effective. They are easily taught to general practitioners who like using these techniques, but it is somewhat harder to convince psychiatrists to do such simple things.

[0010] Furthermore, spontaneous remission accounts for a considerable amount of the improvement observed. There are two naturalistic studies in which people have been interviewed on two occasions and the duration of intervening depressive episodes noted. McLeod et al (1992) reported from a sample of married persons that the median duration of DS -11I-R {American Psychiatric Association, 1987) episodes of depression was 10 weeks, with 75% having episodes of under 22 weeks. Kendler et al (1997) studied a population sample of women and found a median time to recovery of 6 weeks, with 75% recovering in 12 weeks. If the population time to recovery were a median of 8 weeks and 75% recovered within 16 weeks, then people recruited into a trial after being depressed for 8 weeks would have a 50% chance of remitting during the conduct of the usual 8-week trial. These two factors, response to encouragement and a 50% probability of spontaneous remission during the trial, could account for the considerable progress of placebo control groups in depression trials.

[0011] Thus we are left with the concern as to how to deal with this high placebo response rate. Summary of the Invention

[0012] In accordance with an aspect, there is provided a mobile system comprising processing structure configured to: electronically transmit collected data about ambient factor and/or usage of the mobile system to a remote processing structure; receive, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data; at each of the provided dates/times, automatically trigger execution of one or more user-interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and electronically transmit the received user data to the remote processing structure.

[0013] In accordance with another aspect, there is provided a system comprising a plurality of mobile systems as described above; and a remote processing structure in electronic comraunication with each of the mobile systems, the remote processing structure comprising at least one database for storage and aggregation of respective collected data and user data received from each of the plurality of mobile devices.

[0014] In accordance with another aspect, there is provided a non-transitory computer readable medium embodying a computer program executable on a processing structure of a mobile system, the computer program comprising computer program code for electronically transmitting collected data about ambient factor and/or usage of the mobile system to a remote processing structure; computer program code for receiving, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data; computer program code for, at each of the provided dates/times, automatically triggering execution of one or more user-interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and computer program code for electronically transmitting the received user data to the remote processing structure.

[0015] In accordance with another aspect, there is provided a method in a mobile system, the method comprising: electronically transmitting collected data about ambient factor and/or usage of the mobile system to a remote processing structure; receiving, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data; at each of the provided dates/times, automatically triggering execution of one or more user-interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and electronically transmitting the received user data to the remote processing structure.

[0016] In accordance with another aspect, there is provided a computing system comprising processing structure configured to: automatically receive data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system; automatically process the collected data to construct an electronic behaviour profile for the or each subject; and categorize each subject based on their electronic behaviour profile.

[0017] In accordance with another aspect, there is provided a non-transitory processor-readable medium embodying a computer program executable on a computing system, the compxiter program comprising: computer program code for automatically receiving data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system; computer program code for automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and computer program code for categorizing each subject based on their electronic behaviour profile.

[0018] hi accordance with another aspect, there is provided a processor-implemented method comprising: automatically receiving data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system; automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and categorizing each subject based on their electronic behaviour profile.

[0019] In accordance with another aspect, there is provided a computing system comprising processing structure configured to: automatically receive data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system; automatically process the collected data to construct an electronic behaviour profile for the or each subject; and display one or more graphical representations of the electronic behaviour profile. [0020] Γη accordance with another aspect, there is provided a non-transitory processor-readable medium embodying a computer program executable on a computing system, the computer program comprising: computer program code for automatically receiving data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system; computer program code for automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and computer program code for displaying one or more graphical representations of the electronic behaviour profile,

[0021] In accordance with another aspect, there is provided a processor-implemented method comprising: automatically receiving data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system; automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and disp laying one or more graphical representations of the electronic behaviour profile.

[0022] In accordance with another aspect, there is provided a computing system comprising processing structure configured to: receive, from each of at least one mobile system, data collected by the mobile system about ambient factor and/or usage of the mobile system; process the collected data to determine, for each of the at least one mobile system, a plurality of dates/times at which one or more user-interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data; and transmit a respective plurality of dates/times to each of the mobile systems.

[0023] In accordance with another aspect, there is provided a non-transitory processor-readable medium embodying a computer program executable on a computing system, the computer program comprising: computer program code for receiving, from each of at least one mobile system, data collected by the mobile system about ambient factor and/or usage of the mobile system; computer program code for processing the collected data to determine, for each of the at least one mobile system, a plurality of dates/times at which one or more user- interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data; and computer program code for transmitting a respective plurality of dates/times to each of the mobile systems.

[0024] In accordance with another aspect, there is provided a processor-implemented method comprising: receiving, from each of at least one mobile system, data collected by the mobile system about ambient factor and/or usage of the mobile system; processing the collected data to determine, for each of the at least one mobile system, a plurality of dates/times at which one or more user- interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data; and transmitting a respective plurality of dates/times to each of the mobile systems,

[0025] Other aspects and advantages will become apparent from the following. Brief Description of the Drawing

[0026] Embodiments of the invention will now be described with reference to the appended drawings in which:

[0027] Figure 1 is a graph showing an example of an inferred distribution of psychiatric state, for example emotional state or mood, according to assumptions made in existing methodologies;

[0028] Figure 2 is a graph showing the inferred distribution (repeated from Figure 1) and its deviance from an actual underlying mood distribution;

[0029] Figure 3 is a graph showing the inferred mood distribution as determined according, to embodiments of the invention;

[0030] Figure 4 is a flow chart illustrating a process for conducting a clinical trial according to embodiments of the invention, The example given is shown for the factor 'location' but it generalizes to other factors as well;

[0031] Figure 5 is a block diagram illustrating logical components of a mobile system that may be carried by a subject such as a patient, and that may be used to implement embodiments of the present invention;

[0032] Figure 6 is a block diagram illustrating logical components of a system including a plurality of mobile systems in communication with a remote processing structure and providing a medical professional interface for gathering data of a subject such as a patient or participant in a clinical trial setting;

[0033] Figure 7 is a graph showing the convergence of inferred mood distribution of placebo and treatment groups (as an example) according to assumptions made in existing methodologies;

[0034] Figure S is a graph showing the inferred mood scores as determined according to embodiments of the present invention;

[0035] Figure 9 is a graph showing conditional and marginal distributions of mood measured under differing conditions thus illustrating the concept of fair and unfair measurement paradigms;

[0036] Figure 10 is a graph showing an example of mood score variation around the baseline represented by the x-axis. The graph illustrates ideal sampling times;

[0037] Figure 11 is a drawing depicting panel organization of collected samples;

[0038] Figure 12 is a flowchart depicting steps in a method conducted by a mobile system, according to an embodiment, to receive dates/times for triggering execution of user-interactive programs and to provide user data received via the user-interactive programs;

[0039] Figure 13 is a schematic block diagram of components of a system according to an embodiment, including the mobile system in further detail and a remote processing structure; [0040] Figure 14 is a flowchart depicting steps in a method conducted by the remote processing structure in order to determine dates/times for triggering execution of user-interactive programs by the mobile systems based on collected data from the mobile systems;

[0041] Figure 15 is a flowchart depicting steps in a method conducted by the remote processing structure relating to the use of subject behaviour profiles; and

[0042] Figure 16 is a schematic block diagram of the system including a plurality of mobile systems, with the remote processing structure suitable for executing the methods described herein shown in further detail.

Detailed Description

[0043] It may be possible that high placebo response rates can be understood, given good clinical care (Andrews, 1993) (review of what the patient did and did not do), encouragement to resolve problems and resume positive activity, with associated structured problem-solving, activity scheduling and systematic approaches to achieve these goals (Mynors-Wallis et al, 1995; Andrews & Jenkins, 1999) as having their maximal effects while the patient is in the clinical setting. As such, when in the clinical setting, the patient is at their peak level over the time period since their last visit. Ironically, that is when we assess function and symptomatic severity of the patient. As such, assessing the patient at the time of their feeling their best over the previous period (i.e., the outlier time period as to how they are doing overall since their last visit) would increase the placebo response rate and decrease the likelihood of noting a difference between the active treatment and placebo,

[0044] Current mood measuring methodology for evaluation of results in clinical trials involves clinicians measuring mood while the participant is in the clinic. We formulate it quantitatively as follows.

[0045] The actual mood score measured may be understood as the combined effect of a person's real mental baseline b(T) and some other incident ambient effects that affect the underlying true status b(T), Without loss of generality we can illustrate this with an additive model:

m(T) = b(T + e_prior (1) where:

• T is a random vector that incorporates time and observable factors that have an impact on mood score;

• m(T) is the mood score as measured by a psychiatric validated questionnaire.

An example of one is a depression rating scale (DRS) such as PHQ-9, given the set of factors T;

• b(T) represents the baseline emotional state (in our working example it is

mood) - but is itself not directly measurable; and • ⁶prior is unobservable and thus untestable randomness

[0046] There are apparent drawbacks in the model. For instance, one factor that may generally affect mood is the location of measurement An example is in clinic or at work or at home. The latter may better reflects the person's real life and closer to the baseline b(T). However, the person's mood is often measured in clinic. As discussed by Mynors-Wallis et al, 1995; Andrews & Jenkins, 1 99), this leads to bias in mood measurement.

[0047] Referring initially to FIG. 1, there is shown a graph 100 illustrating a baseline mood distribution 105 that would be inferred by clinicians according to existing methodologies, and which is categorized by mood as sampled when the patient is visiting clinic. In graph 100, the sampled moods are shown as data points 110. Now referring also to FIG. 2, graph 200 illustrates the same baseline mood distribution 105 inferred from data points 110, but superimposed with an actual distribution 205 representing an example of a patient's true mood variance throughout the day over a period of days. Note that baseline mood distribution 105 is the inferred distribution according to existing methodologies. Note also that in both FIGS. 1 and 2, the x-axis is broken in two places representing an arbitrary gap in time between measurements 110 taken at clinic.

[0048] In the following description, the notation is fixed as follows. Capital letters are used to denote random variables/vectors, such as X, Y and T. When specifying a realization of their values, we use lower case letters, such as x, y, t. In the deterministic setting, lower case letters are also used. For example, by s(T~), we mean the random quantity s(T) that depends on T, which is also random. It is to be noted, however, that after s(7) has been estimated, and optimization has been performed on it, it becomes a deterministic quantity. As such, s(t) notation is used.

[0049] Definition: Consider the random vector W = (Z, T), where Z is a random variable and T is a random vector of arbitrary finite dimension. We call a random variable Z is measured fairly in W, when marginal distribution of Z is considered. Otherwise, if some components of T are not marginalized, we are working with conditional distribution of Z given some particular T value or range. In the latter case, we say Z is measured unfairly.

[0050] Assume Z is a random variable and T is a random vector. They both possess probability density functions .as shown in Equations 2 through 4, below:

Z~f(z) (2)

(3)

(4) where:

f(z) describes a marginal distribution - that is measured fairly in the first case, and unfairly in the second case. The last equation shows the connection between the two. [0051] Further details on Equations 2 through 4 may be found in Feller 1968, and Ash 1999.

[0052] For example, FIG. 9 is a graph 900 illustrating differences between fair and unfair measurements from a distribution. In graph 900, curve 905 represents the aggregate mood distribution of a patient cohort when measured onsite (e.g., at clinic), while curve 910 represents the aggregate mood distribution of a patient cohort when it is measured offsite (e.g., not at clinic). Curve 915 may represent the marginal distribution— which weighs all possible scenarios - and is therefore a better approximation of the true aggregate patient mood score baseline b (Γ) in at least some contexts.

[0053] Referring now to FIG. 3, embodiments of the present invention provide method(s) and apparatus that permit estimation of the marginal distribution of m(T) against different ambient factors that may be collected. For example, as seen in FIG. 3, mood measurements 110 taken onsite at clinic may be consistently biased - and, when averaged, lead clinicians to an inferred baseline mood distribution 105, as seen also in FIGS. 1 and 2. Whereas, the apparatus according to the invention has the capability to determine more ideal times at which to make mood measurements 310, thereby leading clinicians to infer a more accurate baseline mood distribution 305 (which is not the same as baseline mood distribution 105 in FIGS. 1 and 2). Actual mood distribution 205 is also shown on FIG. 3. Note how on-site mood measurements 110 as well mood measurements 310 are each points on actual mood distribution 205.

[0054] Definition:

[0055] Suppose Z≡ fairly (or Z = unfairly). Since Z denotes a fixed measurement method it is omitted from the definition of tfbelow.

[0056] Note that Z itself has a distribution Z~f(z) , when the notation Z = fairia used, it is to mean that its marginal distribution is considered with respect to T; when the notation Z≡ unfairis used, it is meant that its distribution is measured conditioning on some T = t value: Z~f{z\t) .

[0057] Similarly, in Equation 5 below:

T = (T₁, T₂, ... , T_n) (5)

[0058] In order to define δ we consider only one T_j at a time, where 1 < j < nas i Equation 6 below:

m(7}) = m(Z≡ fair or unfair , = ¾ , ... , Tj, ... , T₇ tn) (6)

[0059] As Tj is the only variable, all which are fixed as constants are omitted. Consider any two values of Tj, for example a or β, we define:

[0060]

[0061] If we add Z to the framework, we now have; fair, Tj - a)— m(Z≡ fair, 7} (8)

Z≡f ir

where:

{unfair

• m(Z, Tf) is the measurement of mood with method Z and conditions 7).

[0062] Example: Assume we are interested in the location effect on mood and mood is measured by method Z≡ fairly with respect to time of the day. Thus marginal distribution of mood is considered with respect to time. If Tj is location that can take on values of either at home or at clinic then ^ftome^ciinic ^{me sures} &e location effect on the representative participant's mood score. The representative participant's mood score is interpreted as a group-wise response. Fairness with respect to time can be realized if, when we compare locations, we also sample τη(Τ) at different times during a day, biasing towards neither morning nor evening.

[0063] Example: Assume we are interested in drug effect on mood and it is measured by method Z≡ unfair with respect to locations of measurement. Suppose by unfair, we mean we measure only

i d-VZlG

placebo ^0UT ^^m *^{s to se}P^afate ^^{e e}^^ec^ of drug from placebo. The unfairness is realized because clinic is just one place that the patient goes to and does not reflect his mood at other locations (which is where he spends most of his time). We should consider mood scores in all locations and weigh them according to proportion of time spent by the person.

[0064] It is to be noted that one of the problems with current practice of measuring δ^τί | when

Z≡ unfair, is that we are not working with a marginal distribution but a conditional one. If a participant's mood score is measured in clinic, we bias towards their mood in clinic - which does not reflect their real life - the patient does not stay at clinic in bis typical daily life. Using the above terminology, dru effect against placebo effect is measured unfairly, conditioning on location clinic. Embodiments of the present invention at least partially address the issue of placebo effect.

[0065] In accordance with embodiments of the present invention, there is provided a method of remotely administering psychiatric diagnostic tests, for example as part of a clinical trial, using a mobile device to achieve reduced placebo effect.

{0066] Within the described embodiments, at least Some of the previously non-quantifiable effects are quantified. For example, this is done by a data collection device (described in more detail with reference to FIG. 5 below) and working with marginal probability distribution - thus measuring fairly with respect to the mood score data m(T) we collect. This then leads to a less biased measure by choosing T = t* so that the conditioning better reflects the participant's life reality, We decompose the currently untreated effect:€p_rj_or further to newly quantified component s(T) and other unobservable quantity e_new . Uncertainty in our method is generally less than the current practice, as long as s(T) is not negligible and it does not depend on€_new too much. This leads to better understanding of mood score measurement, allowing us to lessen placebo effect and potentially better drug discovery. More quantitatively, compared to the current method, we decompose m(T') from that shown in Equation 1 above into that shown in Equation 9, below:

m(T) = b(T) + s(T) + e_new (9) where:

• m(T) represents mood as measured by a psychiatric assessment scale such as PHQ-9 at any given time;

• b (T) represents the baseline (mood);

• s(T) represents observable and testable stimuli (e.g., clinic vs. home) and other short-term observational influences; and

• e_new is unobservable and thus untestable randomness.

[0067] The framework of Equation 9 is not uncommon in current literature in statistics, and is illustrated pictorially in FIGS. 7 and 8, which show graphs 700 and 800, respectively. In FIG. 7, the treatment group is represented by curve 705, while the placebo group is represented by curve 710. The two curves 705 and 710 generally cannot be measured except when sampled, shown as data points 715 and 720, respectively. Note that curve 705 for the treatment group and curve 710 for the placebo group tend to converge when sampled in a clinical setting at data points 715 and 720. Averaging onsite measurements leads clinicians to an inferred baseline 725 for the treatment group and an inferred baseline 730 for the placebo group. The difference between the two baselines 725 and 730 is δ, and is depicted in FIG. 7 as 735. This 5 is one of the factors, in many cases an important if not the most important factor, in determining a drug's effectiveness in a placebo controlled trial. The larger this δ, the fewer participants may generally be needed in the study. Note that all these curves are now referring to the aggregate mood distributions of participants in the study.

[0068] In FIG. 8, however, mood scores are measured in non-clinical environment. As shown, curve 805 represents the treatment group, while curve 810 represents the placebo group. On curves 805 and 810, mood measurements are represented by data points 815 and 820, respectively. The baseline mood score inferred from observations 81 for the treatment group is depicted by line 825, while the baseline mood score inferred from observations 820 for the placebo group is depicted by line 830. Again, the difference, 8, between the treatment and placebo groups is shown as 835. Note that in FIG. 8, die S (i.e., 835) is larger than the S (i.e., 735) in FIG 7, at least in part due to mood scores being measured at more optimal times. [0069] For the sake of clarity in presentation, assume s( ) and e_new are statistically independent, or almost statistically independent - in the sense that dependency is outweighed by uncertainty of s{T) - we have the following.

[0070] Under the variance / L² paradigm (Ash 1999), let VA be the variance of a random variable, as shown in Equation 10 below:

VAR(e_prlm.) = VAR(s(T) + e_new) = VAR (S(T)) + VAR (e_new)≥ VAR(e_new) 00

[0071] Under an information theory framework (McKay 2003), let H be the entropy of a random vector, as shown in Equation 11 below:

Sprier) = HQsQO. ) = H(s(T)) + H(e_new)≥ H(e_new) (11)

[0072] Embodiments of the present invention may reduce the source(s) of uncertainty by reducing either entropy or variance of the error term in the system as defined above.

[0073] Using the above terminology, we may be able to measure drug effect versus placebo effect fairly against the ambient data collected.

[0074] Altering the typical routine of the patient, as discussed by (Mynors-Wallis et al, 1995; Andrews & Jenkins, 1999), may tend to create measurement bias. For example, such measurement bias may be seen in 420 of FIG. 4 (discussed further below). More specifically, the shaded area labeled δ₀ in 420 represents the difference between what is inferred by clinicians as baseline mood (dashed line, circular data points) and what actual mood would be if it were measured in real-time (solid line, square data points).

[0075] According to the described embodiments, a psychiatric assessment like a depression rating scale (where the PHQ-9 is a specific example) may be administered when the chances of measuring the patient's baseline response are determined to be highest. More generally, referring again to FIG. 3, the described embodiments may be operable to determine optimal times, as well as other optimal ambient factors t* , to administer depression rating scales so as to most accurately measure the true baseline mood score, b T . As ised herein throughout, t* may be used to represent a particular realization of relevant (and measurable) ambient factors.

[0076] For clarity of presentation, assume T represents only time for the moment. The argument below works for all other factors combined together as well.

[0077] Define b(T) = E(m(T)) over an experimental period. We then study how this deviation depends on time, T.

[0078] More formally, let us define Y( ) to be a mean-adjusted mood score, as shown in Equation 12 below:

[0079] and thus, following Equation 12: [0080] Y(T) = s(T) + e_new

[0081] The new S^T is computed on the Y(T) as opposed to m(T) . The problem then becomes, how to choose optimum t* such that the deviation of mean-adjusted mood score, Y(T), from 0 is minimized so as to find the least biased measure of mood score m(T). As noted above, curve 960 in FIG. 10 represents mood score m(r). In practice, £"(τη(Γ)) is never observed, it is instead computed by its sample counterpart, m(t).

[0082] Definition: The set of optimal times for measuring mood score m(t) :

[0083] Time* = {t: \s(t)\ < threshhold}

[0084] Where, t*, are the individual optimal times and t* E Time"

[0085] Note the change of notation from T to t.

[0086] Referring now to FIG. 10, there is shown a graph 950 illustrating ideal and non-ideal measurement times for administration of a depression rating scale. In graph 950, the x-axis 980 now represents a baseline mood score, defined as h(t) = m(t) is shown as curve 960 and represents the instantaneous value that m(t) could take at any given time. Curve 960, Y(t), is defined at least partially by non-ideal measurement times 965, as well as ideal measurement times 970, each of which is located on the x-axis 980 (as opposed to non-ideal measurement times 965). Since the x-axis 980 in graph 950 is taken to represent the baseline, finding the ideal measurement times 970 corresponds to finding the zero sets of Y(t).

[0087] In some -embodiments, the set containing the optimal sampling times, Time* , may exclude all times when the patient is in the clinic, and cannot be administered more frequently than once every 5 days, etc. Embodiments of the present invention may utilize one or more of the following criteria to determine when such optimal sampling times may occur:

[0088] 1. We provide all participants with behavioral tracking apparatus that is implemented within their mobile device (see FIG. 5, described further below), which may be a smartphone, tablet, or other computing device the participant may be likely or elect to carry around on a regular basis.

[0089] A mobile device software stack is built for use across multiple platforms to record one or more of the following ambient factors (referred to interchangeably herein as features'):

• Location: L

• Distance From Important Landmarks (e.g., Home/Work/School): D

• Ambient light levels

• Ambient background sounds

[0090] A mobile device software stack is built for use across multiple platforms to record one

[0091] The same mobile device(s) are used record one or more of the or more of the following user data representing factors/features: [0092]

• SMS Log: S

• Call Log: C

• Analysis of Written Language Used In SMS Messages: W

• Application Usage Log: P

• Users' voice conversation samples

[0093] 2. The patient's mobile device remotely administers the accepted psychiatric diagnostic tests, i.e. PHQ-9, GAD-7, as well as cognitive assessment games or other programs that measure the participants' working memory (i.e., 2-back game).

[0094] 2. Mathematical modeling of data and computation of Time * .

[0095] PROCESS FOR FINDING Time*

[0096] The method for Time^* is achieved by

[0097] (Experimental Design)

[0098] I. Partition the experimental period into two parts or periods. First period is 1/3 of study period.

[0099] Π. Setup patient background data: when onboarding patients onto the system, we setup a database which lists the geo-coordinates of patient-relevant landmarks, such as home, work, school. Additional tables in this database include the telephone numbers of the relevant relationships in his life, like mom, children, boss, etc. An additional table lists the times of day when the patient is active, for example, M-Thurs 7:00am to 11 :00pm, Fri-Sun 10am to 2am. This database is part of the input to the algorithm

[00100] ΠΙ. During the first third of the study period, we survey patients randomly with constraints such as no surveys after midnight, none before 7:00am. The patients' responses are associated with the timestamps of when he was prompted, and the timestamps associated with the completion of each question. Hence, if the patient ignores a prompt, or, dismisses them by randomly clicking on responses, his reaction time will be noted. The samples can be organized as panel data, as shown in Figure 11.

[00101] IV. Choose a baseline function. For example, we may choose the average of all the mood scores over a period for a group described above. We do not claim that this is the only appropriate choice. Other functions that reflect seasonality, prior-knowledge, or background information may also be used. We represent the baseline, b(T), m E (m(T)) - the expected value of the mood scores.

[00102] Y T) = m(T) - E(m(T)) = s(T) + e_new

[00103] This is a standard regression model which we use to model s (T),

[00104] (Regression Analysis) [00105] V. Perform regression to estimate s(T) . Details on the computation of this step is provided in steps 1 to 21 below.

[00106] (Inverse/Optimization Problem)

[00107] VI. Assume s T is estimated in the above sense. From now on, we switch our perspective to an optimization perspective. If the model is analytically/numerically tractable, we find the set:

[00108] Time* = { t: \s(t)\ < threshold]

[00109] The inverse image of s such that s(t) is close to zero. The interpretation is that we choose the factors t" G Time* so that the mood score measurement is least affected by the ambient environment. Note the change of notation from s(T) to s(t). In the latter, t is not a random vector, it is a specific t chosen in optimization.

[00110] VII. At this point, randomize all participants into two groups, traditionally PLACEBO and TREATMENT groups. Proceed with usual clinical protocol with the exception that we now send psychiatric questionnaires according to Time^*

[00111] Algorithmic and Computational Details to V above:

[00112] (Feature Selection) Without loss of generality, we will describe feature-selection in one particular way— in order to preserve clarity of presentation.

[00113] 1, Generate a features from the raw data (T) from, e.g., prior medical knowledge and / or experience or other factors and considerations. Additional features may be generated from correlation and exploratory data analysis. We describe in steps a through to step f below exactly how we deal with each factor.

[00114] a. Location:

[00115] From M the list of landmark coordinates from experimental setup described in III above, are recoded into numeric representation {1, 2, 3, ... , m} to facilitate automatic iteration during computatioa

[00116] b. Distance:

[00117] Since each sample of mood includes its location, L, for each mood measurement, calculate the distances of the patient to each landmarks found in list as a vector of dimension equal to number of mood scores sampled. One possible example, is home = (1km, 0km, 1km, 3km, ... ) and work = (0km, 1km, 2km, 2.5km, ... ) every time we sample moocl score we record the distances. So m landmarks will gives us m features from Distance factor. Thus we have m features chosen.

[00118] c. Time period selection:

[00119] Let S be the span of the study period. Let P be a period of time. Then P can be: one day, one week, and any time period within S. Let the total number of periods chosen be K. This is an abstract step that is reused by algorithm to compute particular features. It is also a user specified step. For example, P consists of all weeks in study period S and K is the total number of weeks.

[00120] d. SMS log:

[00121] THESE ARE THE STATISTICAL DISTRIBUTIONS FROM WHICH FEATURES ARE TO BE EXTRACTED

[00122] Applying c above.

• SMS traffic distribution

- Distribution of count of SMSs, over contact numbers, in P

• SMS time density distribution

- Distribution of hour of day (in 24hr format) when the SMSs are sent, normalized to a 24 hour basis

• SMS type distribution

- Distribution of call types: incoming, outgoing, or missed text, in P

[00123] Set SMSjthreshold > 3 this is the minimum to compute statistics defined below. SMSjthreshold is a number chosen so as to better utilize the information in the data. Increasing SMS_th.resh.old leads to fewer features included, while at the same time, decreasing SMSjthreshold too much, will allow noise to contaminate the statistics computed below. For each of the distributions above, if the sample_size ≤ SMS _t/ires/io£d, treat this entry of the feature as a missing value.

[00124] As a guide, SMS threshold, Calljhreshold, and ScreenJchreshold (the latter two are discussed below) are bounded by:

3≤ threshold≤ 25^th percentile of daily observations in period P.

[00125] FEATURES:

• SMS span of counterparties

- Proportion of texts expended to/from the most frequently contacted counterparty (defined by creating a sorted list of contacts where the SMS count is the primary key over a time period, P ). The output of this step is in the form of a ratio

• of SMS to most frequently contacted counterparty „. . ,

— _^ _ ,„ ,„ „ _,,— : — . Smce we have K tune periods, K are total tt of SMS during period P ^r

features generated.

• SMS type proportion

- Proportion of incoming, outgoing, or missed texts, in P, 3 X K features.

• SMS late activity

- Whether there is late night SMS texts, as defined on a patient by patient basis. For example, a late night period might be defined as (lam - 4 am). This generates K features. [00126] For each statistical distribution described above (Traffic, Density, Type) that have within each of periodssampie_size > SMSjhreshold transform them into densities using kernel density estimation or functional data analysis (FDA) regularized smoothing (Ramsay and Silverman 2005).

[00127] Compute, for each density generated above: mean, standard deviation, median, interquartile range, entropy, Gini-index, skewness, kurtosis, minimum and maximum,

[00128] Thus there are 13 features chosen for Traffic and Density distributions. Using the combinations, we have selected 13 X 2 X K features.

[00129] We have generated 13 x 2 x ii + 5 x A: features.

[00130] e. Call log:

[00131] STATISTICAL DISTRIBUTIONS FROM WHICH FEATURES ARE TO BE EXTRACTED

[00132] Let P be generated by applying c above. For each P within the study period S:

• Call traffic distribution

- Distribution of number of calls, over contact numbers, in P

• Call duration distribution

- Distribution of call duration, over contact numbers, in P

• Call time density distribution

- Distribution of time when the calls are made, normalized to a 24 hour basis

• Call type distribution

- Distribution of call types: incoming, outgoing, or missed call, in P

[00133] Set Call hreshold > 3 this parameter serves the same purpose as SMSjhreshold discussed in d, and is computed similarly. However, on the assumption that people send text messages more often than making calls, this C lljhreshold should be set lower than SMSjhreshold, as we have less data than in calls than text messages. This aligns with the general principle of using information in data in d above.

[00134] FEATURES:

[00135] Let P be generated by applying c above. For each P within the study period S:

• Call span of counterparties by frequency

- Proportion of calls to/from the topmost counterparty. The topmost counterparty is represented as the 1^st item in a sorted list of counterparties with call frequency within period P as the primary key. The output of this feature is a ratio, defined similarly d above. features are generated.

• Call span of counterparties by duration - Proportion of calls to/from the topmost counterparty, This is defined identically to, "Span of call counterparties by frequency," except that sum of call duration is used. K features are generated.

» Call type proportion

- Proportion of incoming, outgoing, or missed calls, inP, 3 X K features generated; « Call late activity

- Whether there is late night calling activity. As an example, the output to this feature is a vector of length containing the number of calls made or received between (0 - 4 am). K features are generated.

[00136] For each statistical distribution originating from call log (Call Traffic Distribution, Call Duration Distribution, Call Type Distribution) that have within each of K periods sample_size ≥ Call_threshold, transform them into probability densities using kernel density estimation or functional data analysis (FDA) regularized smoothing (Ramsay and Silverman 2005).

[00137] mean, standard deviation, median, interquartile range, entropy, Gini-index, skewness, kurtosis, minimum and maximum.

[00138] Thus there are 13 features chosen for traffic, duration, and density distributions. Assume there are K periods chosen. Using the combinations, we have selected 13 X 3 X K features.

[00139] We have generated 13 x 3 x K + 6 x K features.

[00140] F. Analysis of language used in SMS messages

[00141] Let P be generated by applying c. For each P within the study period S:

[00142] i, Extract the contents of all SMS messages and index them by timestamp in period P.

[00143] ii. Collapse all the text from SMS log for each period of time into a list of words, for each period P

[00144] iii. Score the overall optimism of language use, for each period P, with the technique described by Seligman (1986) that assigns a positive score to an optimistic word or phrase, and a negative score to a pessimistic one.

[00145] iv. Other natural language optimism metrics can be used; among them, Acerbi, Lampos, Bentley 2013 describe the Joy-Sadness z-score, for each period P.

[00146] v. The output of this step is a vector with two optimism scores for each period P in the K time periods

[00147] We thus have 2 X K features, for each mood score and over K periods.

[00148] Screen Usage:

[00149] Let P be generated by applying c above. For each P within the study period S:

• Screen traffic distribution

- Distribution of phone usage, in period P • Screen time density distribution

- Distribution of time when the usages are made, normalized to a 24 hour basis.

• Screen late activity

- Whether there is late phone usage (0 - 4 am), K features

[00150] Set screen_threshold > 3 this is the minimum to compute statistics defined below: Finding screen_threshold follows a process identical to that described in d above. If the sample _size > screen threshold, transform them to density by kernel density estimation or functional data analysis regularized smoothing, compute: mean, standard deviation, median, interquartile range, entropy, Gini-index, skewness, kurtosis, minimum and maximum,

[00151] Thus there are 13 features chosen for each distribution. Assume there are K periods chosen. Using the combinations, we have selected (13 X 2 + 1) X K features.

[00152] (Model Estimation)

[00153] Let Q be an empty set to begin with. For each of the steps 3 to 17 below that produce statistical models, we are going to insert each model into the set Q, described in more detail below.

[00154] Below, we will compare model performances at various places. Recall the definitions of Y from TV above.

[00155] Let Y be the actual observations corresponding to the inputs T.

[00156] Let Ϋ', Ϋ" be model predictions from two models, referred to as, M_t and M₂ over the same inputs, Γ.

[00157] The two models are similar, if:

[00158] I] Ϋ'— Ϋ" I] < y, where y is a constant set by the user to capture two models' similarity.

[00159] When we evaluate performance of models, we use k-fold cross validation with L² errors ( g 201 ) as our performance measure criterion. By better best performance, we mean a lower/lowest L² in a k-fold cross-validation procedure.

[00160] Let Y be the actual observations corresponding to the k-fold cross validated subset of inputs T.

[00161] Let Ϋ be model prediction from a statistical model over the inputs on the k-fold cross validated subset of inputs T.

[00162] We measure model performance by:

[00163] [I?— y|j² < y₂, where y₂ is a constant set by the user to capture two models' similarity.

|[Y

[00164] 2. Setting γ and y₂ : As a rule, y can is set to— and 0 = 10. Thus it is ten percent of the total sum of squares of the actaal output. If Y has higher standard deviation, D should be decreased; if it has lower standard deviation, D should be increased. [00165] 3. Fit the raw data on the features described in step l.a to l.g by random forest (Breiman 2001), using either entropy or Gini-index as criterion. The output is a model - a random forest - and it is inserted into the set Q.

[00166] 4. Rank the features by variable importance (as described in Breiman 2001) in the random forest model. We use the model selection procedure as described in Genuer, Poggi, and Tuleau- Malot, 2010. We rank the variables by RF score. Extract the important variables into vector V. Record the forest model described by V, as T. Insert 7 into the set Q .

[00167] 5. Many possible forests can be generated from the variable importance table that is outputted from 4 above, because we can have different threshold level for the RF score. For more details, see Genuer, Poggi, and Tuleau-Malot, 2010. Generate other possible random forests and insert them into Q also.

[00168] 6. We move to the simplest category: linear models. Building on V (derived in 4), we parametrically impose some features of T as our linear model's regressors to fit the data based on the features in 7 above in step 4 - selected by random forest's variable importance RF score (Genuer, Poggi, and Tuleau-Malot, 2010).

[00169] 7, Use different cost functions (A¹, L² , regularized, etc) to perform linear regressions on the regressors in step 6 and ranli the model performances (by k-fold cross-validation). Record the best linear model as initiai where best means performance in the k-fold cross-validation procedure on the test data set.

[00170] 8. Many different linear models are produced in step 7, as different cost functions/estimation procedures are used. Insert all these,

, into the set Q.

[00171] 9. Fit data by Support Vector Machine (SVM). Different kernels and parameter choices lead to different SVMs. Among them, choose the one with best k-fold cross validation performance, named as SVM*.

[00172] 10. Compare L initial performance to SVM*

• If the performances are similar design a new SVM kernel based on the features in ^initial-

* If SVM and linear models are still similar, stop and record the kernel as KernelsvM and the corresponding SVM model as SVM_Unear.

• Otherwise iterate until maximum attempt (e.g. 10 attempts - it is a constant set by the user) is reached.

• If SVM* is different from L_initiai tune Li_ni_tiai to approximate SVM" (Rahimi and Recht 2007) and (Vedali and Zisserman 2010) and (Vempati, Vedaldi, Zisserman, Jawahar 2010. Record the retimed linear model separately H&L_SVM - If the performances are never similar, continue to next. Among the SVMs, record the ones similar to random forest, 5 Mf_orest5;_mijarthe least similar ones as 5V ^_ores£:D^_erent. Note there could be more than one model in SV M _0restSi_mn_ar and as SVMf_0restD[ff_erent. In this case, index them separately.

[00173] 11. Note in step 10, many models may be produced. Among them, SVM*, SVM _near, LSVM, SVMforestsimiiar, SVM_{forestDifferent}. Insert these and all other SVMs produced into the set

Q.

[00174] 12. Fit data by different layers of neural networks (Hinton, G; University of Toronto 2013). Record the networks that performs best in k-fold cross validation as Jf.

[00175] 13. Record the neural networks that are most similar to the random forest, , as Nforestsimilar- Record the ones that are least similar as ^NforestDifferent- If there is a tie, record and index them all (as in step 10). Record the networks that are most similar to SVM (in step 10 above), as

Each of the four sets: NforestSimilarl ^forestDiff event ¹ ^SVMSimllar ^! ^SVMDifferent ^can contain more than one model. Thus we index them all.

[00176] 14. In step 13, many models may be produced. Insert JV ,

» tyarestDlfferent ^SVMSimilar ^ ^SVMDif erent ^to the set Q. The latter four models are used to hedge risks of favoring a particular model. Also insert other neural nets whose k-fold cross-validation performances are satisfied by γ_ζ set by the user.

[00177] 15. Compare the performance of linear model L_lni_ti_ai to the neural networks in step 14: J\f / Neuralf_0rgst$i_mi_lar / Neuralf_0restDiff ^ent I Neural_SVMSi_mu_ar I and N eural_SVMDiff_erent:

a. l£ Linitiai is similar to a neural net, look at the regressors selected in initial an also the neural net's neuron/hidden variables. If it is similar to more than one, search through all the hidden variables and treat them as additional regressors.

b. If we can further improve L^^'s k-fold cross validation performances by using additional features from a neural net, we record the new linear model as as Ne ral- Insert Ltf_eurai from either step above to the set Q.

c. If they are not similar, terminate, continue to next step.

[00178] 16. Fit data by nonparametric Bayesians, via Gaussian/Dirichlet processes (Hjort, Holmes, MiiUer, and Walker 2010) and (Rasmussen and Williams 2006). Repeat the comparison steps just as was done in neural net step 15 above. Record the basis functions estimated by Bayesian model ®&BayeSf_0restSi_mii_ar ^ayes^_orestoiff_erent,BayessvMsi U r > B y^essvMdifferent > ^m^ the linear model as L_Bay_M , within step 15 above and insert them all into Q.

[00179] 17. Perform boosting algorithms, such as AdaBoost (Schapire, 2001) and Gradient Boosting Machine Algorithm (Friedman, 2001 and 2002) in all models recorded above in the set Q; we refer to this model as Boosted_AAa , Boosted_GBM, corresponding to each of the boostedmodels respectively. Insert these boosted models into the set Q.

[00180] 18, Use k-fold cross-validation with L² error to rank each of the different model performances in the set Q.

[00181] 19. Isolate to quintile of k-fold cross validation performers (step 18), call this M_tov, If there is a linear model i this range, label it as L_tov. There could be many linear models that meet this criterion.

[00182] 20. If there exist linear models in the quintile, choose the most computable one as our s*(t). Solve the inverse problem

100183] Time" = { t: \s*(t) \ < threshold }

a. If the Bayesian model in step 16 above is tractable and its performance lies in top 10%, we may use it to compute Time* as well. If mis fails, proceed to 0 below.

b. If it is not analytically invertible, use brute-force approach on the top models in M_tov by discretizing and computing the discretized models, Specifically, call the estimated model as Y(t). Compute F(t) for different values of t: (t_1; t₂, etc.), and select the one that gives us the smallest as

[00184] 21. This finishes the description of estimating the optimum Time* to measure mood score.

End of algorithm description

[00185] Another way to use these techniques is to apply it to the personalized care of a particular patient. For example, a patient who uses the embodiments of this invention for long enough, would allow for the generation of Time^* - that best reflect Ms baseline mood measurement. The algorithmic procedures remain the same, it's just that the panel data is for only one patient.

[00186] Referring now to Figure 4, there is shown a flow chart of an example process 400 for administering a clinical trial according to embodiments of the invention. In 405, a patient cohort is recruited for participation in the clinical trial. Such patient cohort may comprise generally any arbitrary number of patients exhibiting symptoms of a mood disorder, but in some cases will comprise at least 50 different patients. In 410, a particular target factor is selected for investigation. For example, as described herein, the impact of location on mood score measurements may be selected. In 415, a depression mood score may be administered to each participating patient both at clinic (on-site) as well as off-site using a patient's mobile device (see Figure 5) by determining optimal measurement times as described herein according to the embodiments. For example, the optimal measurement times may be determined using the method(s) presented herein above. In 420, the difference 5₀ between measurements taken according to the embodiments of the invention (in this case, offsite) and measurements taken at clinic is taken. As noted, the shaded area in 420 depicts the difference between baseline mood inferred from at clinic measurements and those measurements taken at ideal times according to the embodiments of the invention.

[00187] Referring now to Figure 5, there is shown a block diagram showing logical components of a mobile system 1000 that may be used in embodiments of the invention. The mobile system 1000 may be any suitable computing device or coordinated devices carried by a subject such as a patient or study participant on which one or more different modules may be implemented.

[00188] In some embodiments, mobile device 1000 may include at least a set of patient interface modules 1001 and a set of ambient sensing modules 1005. The set of patient interface modules 1001 may generally be used for providing one- or two-way interaction with the patient, such as to administrate depression mood scores, memory games, etc. as described herein. Thus, patient interface modules 1001 may include at least a surveys module that can be programmed with established diagnostic questionnaires, such as PHQ9, GAD7 for administration to the patient at determined optimal measurement times.

[00189] Patient interface modules 1001 may further include a games module that is programmed with and to implement one or more different cognitive games that may be indicative of a patient's mental state. Examples of such cognitive games may include the "two back" game, or connect the dots, as well as others not specifically named. However, patient interface modules 1001 are not limited just to a surveys module and games module, and may include other customizable modules, in modular fashion, and generally without limitation.

[00190] Ambient sensing modules 1005 may also include one or more different customizable modules that may be used for passive data collection relating to the patient. For example, ambient sensing modules 1005 may include a location sensor configured to passively collection data relating to the location of the patient and/or mobile device 1000- The data collected by location sensor may include, without limitation, GPS- coordinates, MAC addresses of nearby Wi-Fi devices, and/or nearby Bluetooth transmitters, which are indicative of the ambient conditions, like, how many different people are nearby.

[00191] Ambient sensing modules 1005 may further include a communications logging module that is configured to collect one or more different types of meta-data related to patient communications, including but not limited to, number(s) dialed from the mobile system 1000, number(s) from which a call to the mobile system 1000 is received, duration of call(s), time of call(s), etc. Communications logging module may further be configured to collect one or more different types of meta-data from SMS messages sent from or received by the mobile system 1000, including but not limited to, number of words in SMS message(s), length of word(s) used, number(s) the SMS message is sent to or received from, etc. Again, similar to patient interface modules 1001, the ambient sensing modules 1005 are not limited just to the particular modules depicted in Figure 5, and may include other customizable modules, in modular fashion, and generally without limitation.

[00192] Referring now to Figure 6, there is shown a block diagram showing logical components of an example system 600 for administering a clinical trial according to embodiments of the invention. In some embodiments, system 600 may comprise a remote processing structure 3000 including a central database and a plurality of different mobile systems 1000, each such mobile system 1000 associated with a participating patient in the clinical trial. Thus, each mobile system 1000 shown in Figure 6 may have the general logical configuration as shown in Figure 5 and described above. The number of participating, patients in the clinical trial may be any number suitable for the clinical trial.

[00193] Each mobile device 1000 of a participating patient may be configured to establish a corresponding communication link with the remote processing structure 3000, Such communication link may be established using any suitable network communication protocol, currently existing or later developed, such as but not limited to different configurations of public network(s), private network(s), or combinations thereof. For example, the communication link may include an SSH channel, an SSL channel, or incorporate another form of encryption. In some cases, the communication link may implement data-sanitization before transmission in order to provide enhanced security or privacy. The data collected from patients' mobile systems 1000 may be archived in the central database of remote processing structure 3000 for storage and further processing.

[00194] Preferably, system 600 further includes or is at least accessible via a user interface 3001 that is connected to remote processing structure 3000 using a communication link. Such user interface 3001 may provide a data analysis pack for clinical staff or other eligible personnel, which includes a custom electronic medical record. Alternatively, user interface may integrate custom or off-the shelf statistical data-visualization programs in order to provide a data analysis pack.

[00195] Figure 12 is a flowchart depicting steps in a method 5 conducted by the mobile system 1000 in order to receive dates/times for triggering execution of user-interactive programs and to provide user data received via the user-interactive programs. During the method 5, ambient factor and/or usage data is collected at the mobile system 1000 (step 10). The collected data is electronically transmitted to a remote processing structure (step 12). Pursuant to the bransmitting, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data is received from the remote processing structure (step 14). At each of the provided dates/times, execution of one or more user-interactive programs is automatically triggered on the mobile system thereby to receive user data inputted by a user of the mobile system (step 16), and the received user data is electronically transmitted to the remote processing structure (step 18). This process may continue for an assessment period as established by a treatment or assessment professional, and as such may be re-iterated (step 20). In the event that the assessment period is ended or the treatment or assessment professional wishes to end the assessment, the data collection is ended (step 22),

[00196] In an embodiment, the user-interactive programs are one or more user-interactive programs such as a psychiatric assessment questionnaire or a cognitive assessment game.

[00197] In this embodiment, the received user data is transmitted in association with concurrent data relating to ambient factor and/or usage of the mobile device. Furthermore, in addition to transmitting user data relating to actual user input, in this embodiment user data relating to one or more failures of the user to input user data when requested is transmitted. The user data relating to the one or more failures is transmitted in association with concurrent data relating to ambient factor and/or usage of the mobile device so that further information as to the behaviour profile of the user can be gleaned.

[00198] In an embodiment, the received user data is transmitted in association with data relating to the dates/times at which the user data was inputted.

[00199] In this description, ambient factor data comprises data relating to the conditions of the mobile system in. its environment when the ambient factor data is collected. Such ambient factor data may be used as a proxy to infer actual behaviour of the mobile system user at a particular time, at a particular place, and so forth. Such ambient factor data may include physical location of the mobile system, other devices in proximity to the mobile system, background light levels, background audio levels, recorded sound levels, video and/or still images captured by the mobile system, and other such factors.

[00200] In an embodiment, the data relating to physical location of the mobile system includes one or more of global positioning data and location data derived from local wireless networks. The data relating to other devices in proximity to the mobile system may inchide data relating to proximate wireless networks,, such as Wi-Fi networks and/or Bluetooth networks.

[00201] The usage data may include data relating to one or more of; call activity on the mobile system, messaging activity on the mobile system, and screen usage of the mobile system. Data relating to call activity may include data relating to one or more of: duration of calls, time of calls, frequency of calls, counterparties on calls, whether calls are incoming or outgoing, whether calls are unanswered, and recorded voice samples.

[00202] The data relating to messaging activity may include data relating to one or more of: time of messages, counterparties on messages, frequency of messages, whether messages are incoming or outgoing,, whether messages are unresponded to, and language used in messages. [00203] The data relating to screen usage may include data relating to one or more of: times of screen use, application use.

[00204] In an embodiment, the data collected in step 10 above is produced by selectively capturing and storing at least a portion of all ambient factor and/or usage data from one or more data streams being automatically produced during operation of the mobile system 1000. For example, the operating system or system applications of mobile system 1000 may include call and message logging that may be accessed through a respective application programming interface (API) or other mechanism and selectively sampled and stored on mobile device 1000 for subsequent transmission to the remote processing structure 3000.

[00205] In this embodiment, collected ambient factor and usage data, as well as user data collected b user-interactive programs, is periodically stored and sent in batches to the remote processing structure 3000.

[00206] Figure 13 is a schematic block diagram of components of system 600 including the mobile system 1000 in further detail, and a remote processing structure 3000, according to an embodiment. In this embodiment, mobile system 1000 is configured to execute the method 5. In this embodiment, mobile system 1000 is a single mobile device in the form of a smartphone powered by an internal power supply such as a battery (not shown), which provides power to a main board 1012, which in turn converts the power as required for logic circuitry, and provides the power to various other components. Each of a touch screen 1020, physical buttons 1022, cellular transceiver 1024, Bluetooth transceiver 1028, Wi-Fi transceiver 1030, speaker 1032, microphone 1034, and USB interface 1036, as well as other components not shown, is operably connected to the main board 1092 to receive power and to communicate with a central processor 1016. In this embodiment, central processor 1016 is a single microcontroller and is in communication with onboard processor-readable memory 1014 configured to collect and store various pieces of data including operational data and processor-readable program code for programming the central processor 1016 to operate various user- interactive programs on the mobile system 1000 as well as to operate the various components of the mobile system 1000. In alternative embodiments, central processor 1016 may be a plurality of coordinated processors,

[00207] USB interface 1018 can receive an external USB cable 1019 for enabling data communications with remote processing structure 3000 via one of its own USB ports in its communications interface 3020 such that data can be received by and sent to remote processing structure 3000. More typically, however, communications between mobile system 1000 and remote processing structure 3000 is encrypted and conveyed via a wireless connection through communications network 2000 such as the Internet including a Wi-Fi base station in communication with Wi-Fi transceiver 1030. [00208] In this embodiment, the mobile system 1000 performs processing steps described herein in response to the central processor 1016 executing one or more sequences of one or more instructions contained in a memory, such as the memory 1014. Such instructions may be read into the memory 1014 from another computer readable medium, such as a hard disk or a removable media drive. One or more processors in a multi-processing arrangement may also be employed to execute the sequences of instructions contained in. memory 1014. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions. Thus, embodiments are not limited to any specific combination of hardware circuitry and software.

[00209] The mobile system 1000 includes at least one computer readable medium or memory for holding instructions programmed according to the teachings of the invention and for containing data structures, tables, records, or other data described herein. Examples of computer readable media are compact discs, hard disks, floppy disks, tape, magneto-optical disks, P OMs (EPROM, EEPROM, flash EPROM), DRAM, SRAM, SDRAM, or any other magnetic medium, compact discs (e.g., CD- ROM), or any other optical medium, punch cards, paper tape, or other physical medium with patterns of holes, a carrier wave (described below), or any other medium from which a computer can read.

[00210] Stored on any one or on a combination of computer readable media, the present invention includes software for controlling the mobile system 1000, for driving a device or devices for implementing the invention, and for enabling the mobile system 1000 to interact with a human user, Such software may include, but is not limited to, device drivers, operating systems, development tools, and applications software. Such computer readable media further includes the computer program product of aspects of the present invention for performing all or a portion (if processing is distributed) of the processing performed in implementing aspects of the invention.

[00211] The computer code devices of the present invention may be any interpretable or executable code mechanism, including but not limited to scripts, interpretable programs, dynamic link libraries (DLLs), Java classes, and complete executable programs. Moreover, parts of (lie processing of the present invention may be distributed for better performance, reliability, and/or cost.

[00212] A computer readable medium providing instructions to a central processor 1016 may take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Non-volatile media includes, for example, optical, magnetic disks, and magneto-optical disks, such as a hard disk or the removable media drive. Volatile media includes dynamic memory, such as the memory 1014. Transmission media includes coaxial cables, copper wire and fiber optics, including the wires that make irp a bus. Transmission media also may also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

[00213] Various forms of computer readable medi may be involved in carrying out one or more sequences of one or more instructions to central processor 1016 for execution. For example, the instructions may initially be carried on a magnetic disk of another remote computer. The other remote computer can load the instructions for implementing all or a portion of the present invention remotely into a dynamic memory and send the instructions over a communication line using a modem. A modem local to the mobile system 1000 may receive the data on the communication line and use an infrared transmitter to convert the data to an in&ared signal. An infrared detector coupled to a bus can receive the data carried in the infrared signal and place the data on the bus. The bus carries the data to the memory 1014, from which the central processor 1016 retrieves and executes the instructions. The instructions received by the memory 1014 may optionally be stored on an external or selectively couplable storage device either before or after execution by central processor 1016.

[00214] It will be understood that additional components, such as status lights and audible indicators, though not shown in the drawings, may also be connected to central processor 1016 for use in operation of mobile system 1000.

[00215] While mobile system 1000 is a single mobile device in this embodiment, mobile system 1000 may alternatively be implemented as multiple mobile devices in close-range communication with one another (such as a wired USB or other connection or alternatively as a wireless Wi-Fi, Bluetooth, Zigbee, ANT, IEEE 802.15.4, or Z-Wave connection, for examples).. For example, mobile system 1000 may include a smartphone and a wrist-mountable computing device, each having respective microcontrollers that work in concert via the wired or wireless connection to achieve a desired result as described herein. Mobile system 1000 may alternatively be implemented in the form of a laptop computing device either alone or in combination with another device, a head-mountable computing device such as a Google Glass™ device either alone or in combination with another device, or alternatively some other suitable system that can be carried with a user during typical daily activities.

[00216] It will be understood that it is not necessary that all mobile systems 1000 being used in an overall system such as is described herein must be of the exact same construction. In one embodiment, a subject may make use of his or her own smartphone, which may employ an operating system from Android, Apple, Blackberry or some other producer, provisioned as described herein, However, the nature of the ambient factor data and usage data being collected for transmission to remote processing structure 3000 should be either the same, be amenable to normalizing, or at least include a known variable for use to ensure that a behaviour profile is not unduly sensitive to the particular mobile system being used by the subject,

[00217] Figure 14 is a flowchart depicting steps in a method 55 conducted by the remote processing structure 3000 in order to determine dates/times for triggering execution of user-interactive programs by the mobile systems 1000 based on collected data from the mobile systems 1000. During the method 55, data collected by the mobile system about ambient factor and/or usage of the mobile system is received from each of at least one mobile system (step 60). In this embodiment, the collected data is automatically processed to construct an electronic behaviour profile for each subject (step 62), and a plurality of dates/times to correspond to dates/times at which baseline behaviour for the respective user can be inferred from the electronic behaviour profile is determined (step 64). This plurality of dates/ times is intended to indicate the dates/times at which one or more user-interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data. Each plurality of dates/times is transmitted to respective mobile systems (step 66) for use by the mobile system in determining when to trigger the user-interactive programs as described above. As collected data is received on the ambient factors and/or usage of the mobile systems, the method 55 may continually be reiterated (step 68) so as to continually refine the behaviour profiles. Such refinement may result in the determination of new dates/times that correspond to new or better-understood baseline behaviour.

[00218] Figure 15 is a flowchart depicting steps in a method 70 conducted by the remote processing structure 3000 relating to the use of subject behaviour profiles. During the method 70, data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system is automatically received (step 72). The received collected data is automatically processed data to construct an electronic behaviour profile for the or each subject (step 74). Based on the electronic behaviour profile, a subject may be categorized (step 76) such as by segmenting the subject as qualified or unqualified for a clinical trial or as suitable or unsuitable for a particular medication or mode of treatment. For example, the behaviour profile may reveal a particular type of behaviour that may infer that the subject behaves in a way that would not be amenable to treatment with a particular medication, or that the subject would not be a suitable participant for a given clinical trial. Based on the electronic behaviour profile, one or more graphical representations of the behaviour profile may be generated and displayed on a display device for assessment by a treatment professional.

[00219] Figure 16 is a schematic block diagram of the system including a plurality of mobile systems 1000, with the remote processing structure 3000 suitable for executing the methods 55 and 70 shown in further detail. Any number of mobile systems 1000, corresponding to the number of subjects whose behaviour is being monitored, can be used during an assessment period. It will be noted that remote processing structure 3000 is remote only in the sense that it is not carried by a subject along with the respective mobile system 1000. In this embodiment, remote processing structure 3000 is a powerful computing system that is incorporated into a data or processing centre or device, and includes database storage for storage and aggregation of collected data and user data received from a number of different mobile systems 1000 carried by a number of subjects. While a description of an embodiment of remote computing structure 3000 is described below, it will be understood that alternative configurations capable of implementing the herein described systems and methods may be employed. [00220] In this embodiment, remote processing structure 3000 includes a bus 3010 or other communication mechanism for communicating information, and a processor 3018 having a plurality of parallel processors coupled with the bus 3010 for processing the information. Remote processing structure 3000 also includes a main memory 3004, such as a random access memory (RAM) or other dynamic storage device (e.g., dynamic RAM (DRAM), static RAM (SRAM), and synchronous DRAM (SDRAM)), coupled to the bus 3010 for storing information and instructions to be executed by processor 3018. In addition, the main memory 3004 may be used for storing temporary variables or other intermediate information during the execution of instructions by the processor 3018. Processor 3018 may include memory structures such as registers for storing such temporary variables or other intermediate information during execution of instructions. The remote processing structure 3000 further includes a read only memory (ROM) 3006 or other static storage device (e.g., programmable ROM (PROM), erasable PROM (EPROM), and electrically erasable PROM (EEPROM)) coupled to the bus 3010 for storing static information and instructions for the processor 1018.

[00221] The remote processing structure 3000 also includes a disk controller 3008 coupled to the bus 3010 to control one or more storage devices for storing information and instructions, such as a magnetic hard disk 3022, and a removable media drive 3024 (e.g., floppy disk drive, read-only compact disc drive, read/write compact disc drive, compact disc jukebox, tape drive, and removable magneto-optical drive). The storage devices may be added to the remote processing structure 3000 using an appropriate device interface (e.g., small computing system interface (SCSI), integrated device electronics (IDE), enhanced-IDE (E-IDE), direct memory access (DMA), or ultra-DMA).

[00222] The remote processing structure 3000 may also include special purpose logic devices (e.g., application specific integrated circuits (ASICs)) or configurable logic devices (e.g., simple programmable logic devices (SPLDs), complex programmable logic devices (CPLDs), and field programmable gate arrays (FPGAs)).

[00223] The remote processing structure 3000 may also include a display controller 3002 coupled to the bus 3010 to control a display 3012, such as a liquid crystal display (LCD) screen, for displaying information to a user of the remote processing structure 3000. In this embodiment, the remote processing structure 3000 includes input devices, such as a. keyboard 3014 and a pointing device 3016, for interacting with a computer user and providing information to the processor 3018. The pointing device 3016, for example, may be a mouse, a trackball, or a pointing stick for communicating direction information and command selections to the processor 3018 and for controlling cursor movement on the display 3012. In addition, a printer may provide printed listings of data stored and/or generated by the remote processing structure 3000.

[00224] In this embodiment, the- remote processing structure 3000 performs processing steps described herein in response to the processor 3018 executing one or more sequences of one or more instructions contained in a memory, such as the main memory 3004. Such instructions may be read into the main memory 3004 from another computer readable medium, such as a hard disk 3022 or a removable media drive 3024. One or more processors in a multi-processing arrangement may also be employed to execute the sequences of instructions contained in main memory 3004. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions. Thus, embodiments are not limited to any specific combination of hardware circuitry and software.

[00225] The remote processing structure 3000 includes at least one computer readable medium or memory for holding instructions programmed according to the teachings of the invention and for containing data structures,, tables, records, or other data described herein. Examples of computer readable media are compact discs, hard disks, floppy disks, tape, magneto-optical disks, PROMs (EPROM, EEPROM, flash EPROM), DRAM, SRAM, SDRAM, or any other magnetic medium, compact discs (e.g., CD-ROM), or any other optical medium, punch cards, paper tape, or other physical medium with patterns of holes, a carrier wave (described below), or any other medium from which a computer can read.

[00226] Stored on any one or on a combination of computer readable media, the present invention includes software for controlling the remote processing structure 3000, for driving a device or devices for implementing the invention, and for enabling the remote processing structure 3000 to interact with a human user. Such software may include, but is not limited to, device drivers, operating systems, development tools, and applications software. Such computer readable media further includes the computer program product of aspects of the present invention for performing all or a portion (if processing is distributed) of the processing performed in implementing aspects of the invention.

[00227] The computer code devices of the present invention may be any interpretable or executable code mechanism, including but not limited to scripts, interpretable programs, dynamic link libraries (DLLs), Java classes, and complete executable programs. Moreover, parts of the processing of the present invention maybe distributed for better performance, rehability, and/or cost.

[00228] A computer readable medium providing instructions to a processor 3018 may take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Nonvolatile media includes, for example, optical, magnetic disks, and magneto-optical disks, such as the hard disk 3022 or the removable media drive 3024. Volatile media includes dynamic memory, such as tlie main memory 3004. Transmission media includes coaxial cables, copper wire and fiber optics, including the wires that make up the bus 3010. Transmission media also may also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

[00229] Various forms of computer readable media may be involved in carrying out one or more sequences of one or more instructions to processor 3018 for execution. For example, the instructions may initially be carried on a magnetic disk of another remote computer. The other remote computer can load the instructions for implementing all or a portion of the present invention remotely into a dynamic memory and send the instructions over a communication line using a modem. A modem local to the remote processing structure 3000 may receive the data on the communication line and use an infrared transmitter to convert the data to an infrared signal. An infrared detector coupled to the bus 3010 can receive the data carried in the infrared signal and place the data on the bus 3010. The bus 3010 carries the data to the main memory 3004, from which the processor 3018 retrieves and executes the instructions. The instructions received by the main memory 3004 may optionally be stored on storage device 3022 or 3024 either before or after execution by processor 3018.

[00230] The remote processing structure 1000 also includes a communication interface 3020 coupled to the bus 3010. The communication interface 3020 provides a two-way data communication coupling to a network link that is connected to, for example, a local area network (LAN) 3500, or to the communications network 2000, or to another device via, for example, a USB connection such as device 1000. The communication interface 3020 may include a network interface card to attach to any packet switched LAN. As another example, the communication interface 3020 may include an asymmetrical digital subscriber line (ADSL) card, an integrated services digital network (ISDN) card or a modem to provide a data communication connection to a corresponding type of communications line. Wireless links may also be implemented. In any such implementation, the communication interface 3020 sends and receives electrical, electromagnetic or optical signals that carry digital data streams representing various types of information and, in the case of USB, electrical power.

[00231] The network link typically provides data communication through one or more networks to other data devices. For example, the network link may provide a connection to another computer through a local network 3500 (e.g., a LAN) or through equipment operated by a service provider, which provides communication services through a communications network 2000. The local network 3500 and the communications network 2000 use, for example, electrical, electromagnetic, or optical signals that carry digital data streams, and the associated physical layer (e.g., CAT 5 cable, coaxial cable, optical fiber, etc). The signals through the various networks and the signals on the network link and through the communication interface 3020, which carry the digital data to and from the remote processing structure 3000 may be implemented in baseband signals, or carrier wave based signals. The baseband signals convey the digital data as unmodulated electrical pulses that are descriptive of a stream of digital data bits, where the term "bits" is to be construed broadly to mean symbol, where each symbol conveys at least one or more information bits. The digital data may also be used to modulate a carrier wave, such as with amplitude, phase and/or frequency shift keyed signals that are propagated over a conductive media, or transmitted as electromagnetic waves through a propagation medium. Thus, the digital data may be sent as unmodulated baseband data through a "wired" communication channel and/or sent within a predetermined frequency band, different than baseband, by modulating a carrier wave. The remote processing structure 3000 can transmit and receive data, including program code, through the network(s) 3500 and 2000, the network link and the communication interface 3020. Moreover, the network link may provide a connection through a LAN 3500 to another mobile device 3300 such as a personal digital assistant (PDA) laptop computer, or cellular telephone.

[00232] Although embodiments have been described with reference to the drawings, those of skill in the art will appreciate that variations and modifications may be made without departing from the spirit, scope and purpose of the invention as defined by the appended claims.

[00233] REFERENCES:

1. Feller, William; An Introduction to Probability Theory and Applications; 1968

2. Ash, Robert; Probability and Measure Theory; 1999

3. McKay, David; Information Theory, Inference, and Learning Algorithms; 2003

4. Lampos A and Bentley, Garnett; The Expressions of Emotions of in 20th Century Books

5. Seligman, Martin; Explanatory style as a predictor of productivity and quitting among life insurance sales agents; 1986

6. Genuer, and Poggi, and Tuleau-Malot; Variable Selection using Random Forests; 2010;

Available here: http://eoo.gl/thNZT

7. Breiman, Leo; Random Forest; Machine Learning; vol 45 p. 5 - 32; 2001

8. Rahimi, A. and Recht, B; Random features for large-scale kernel machines; Advances in neural information processing; 2007

9. Vedaldi, A. and Zisserman, A.; Efficient additive kernels via explicit feature maps; Computer Vision and Pattern Recognition; 2010

10. Vempati, S. and Vedaldi, A. and Zisserman, A. and Jawahar, CV; Generalized RBF feature maps for Efficient Detection; 2010

11. Nils Lid Hjort, Chris Holmes, Peter Mtiller and Stephen G. Walker. Bayesian Nonparametrics. Cambridge University Press; 2010

12. Bishop, Christopher; Pattern Recognition and Machine Learning; 2007

13. Friedman, J.H. and Hasite, T. and Tibshirani R.; Elements of Statistical Learning; 2009

14. Schapire, Robert E; The Boosting Approach to Machine Learning; AT&T Labs Research;

Shannon Laboratory; 2001 http://goo.gl/iFVli

15. Andrew Ng - Lecture Notes for CS229 - Machine Learning; Department of Computer Science, Stanford University; 2013 http:// goo. gl/OanEC

16. C. E. Rasmussen and C. K. I. Williams; Gaussian Processes for Machine Learning, MIT Press; 2006

17. Acerbi A, Lampos V, Garnett P, Bentley RA (2013) The Expression of Emotions in 20th Century Books. PLoS ONE 8(3); e59030. doi:10.1371/journal.pone.0059030

18. Friedman, JH; Greedy function approximation: a gradient boosting machine; Annals of Statistics; 2001 Friedman, JH; Stochastic Gradient Boosting; Computational Statistics and Data Analysis; 2002

American Psychiatric Association (1980) Diagnostic and Statistical Manual of Mental Disorders (3rd edn) (DSM-ΠΙ). Washington, DC: APA.

American Psychiatric Association (1987) Diagnostic and Statistical Manual of Mental Disorders (3rd edn, revised) (DSM-III-R). Washington, DC: APA.

American Psychiatric Association (1 94) Diagnostic and Statistical Manual of Mental Disorders (4th edn) (DSM-IV). Washington, DC: APA.

Andrews, G. (1993) The essential psychotherapies. British Journal of Psychiatry, 162, 447 - 451.

Andrews, G. & Jenkins, R. (eds) (1999) The Management of Mental Disorders. London: World Health Organization Collaborating Centre in Mental Health.

Andrews, G. Sanderson, K., Slade, T., et al (2000) Why does the burden of disease persist? Relating the burden of anxiety and depression to effectiveness of treatment. Bulletin of the World Health Organization, 78, 446 -454.

Andrews G, Placebo response in depression: bane of research, boon to therapy; BJP March 2001 178:0; doi:10.1192/bjp,178.3.0

Cochrane, A. L. (1989) Effectiveness and Efficiency: Random Reflections on Health Services. London: British Medical Journal

Enserink, M. (1999) Can the placebo be the cure? Science, 284, 238 -240.

Harrison, Virginia. "Mobile Mental Health: Review of the Emerging

Field and." Journal of Mental Health 20.6 (2011): 509-24. Print.

Hofniann SG, Sawyer AT, Witt AA, Oh D. The effect of Mindfulness-based Therapy on anxiety and depression: A meta-analytic review. Journal of Counseling and Clinical Psychology 2010; 78(2): 169-183.

Jaeggi, S. M., Buschkuehl, M., Perrig, W. J., & Meier, B. (2010). The concurrent validity of the N-back task as a working memory measure. Memory, 18(4), 394-412.

JofTe, R., Sokolov, S. & Streiner, D. (1996) Antidepressant treatment of depression: a metaanalysis. Canadian Journal of Psychiatry, 41, 613 -616.

Jorm, A. F., Angermeyer, M. & atschnig, H. (2000) Public knowledge of and attitudes to mental disorders: a limiting factor in the optimal use of treatment services. In Unmet Need in Psychiatry (eds G. Andrews & A. Henderson), pp. 399-416. Cambridge: Cambridge University Press.

Kendler, . S., Walters, E. E, & Kessler, R. C. (1997) The prediction of length of major depressive episodes: results from an epidemiological survey of female twins. Psychological Medicine, 27, 107 -117. CrossRefMedline Khan, A,, Warner, H. A. & Brown, W. A. (2000) Symptom reduction and suicide risk in patients treated with placebo in antidepressant clinical trials. Archives of General Psychiatry, 57, 311 -330. CrossRefMedline

Kirsch, I. & Sapirstein, G. (1998) Listening to prozac but hearing placebo: a meta-analysis of antidepressant medication. Prevention & Treatment, 1, Article 0002a. http://journals.apa.org/prevention/volume 1/preOO 10002a.html.

Kraemer, H. C. (2000) Statistical analysis to settle ethical issues? Archives of General Psychiatry, 57, 327 -328. CrossRefMedline

Lang,. Ariel J. What mindfulness brings to psychotherapy for anxiety and depression; Depression and Anxiety 2013; 30:409-412.

Kroenke K, Spitzer RL, Williams JB, Monahan PO, Lowe B. Anxiety disorders in primary care: prevalence, impairment, comorbidity, and detection. Ann Intern Med. 2007 Mar 6; 146(5):317-25.

Leon AC, Olfson M, Portera L, Farber L, Sheehan DV (1997). Assessing Psychiatric Impairment in Primary Care ith the Sheehan Disability Scale. The International Journal of Psychiatry in Medicine Vol27, N 2 P 93-105.

Lowe B, Decker O, Mailer S, Brahler E, Schellberg D, Herzog W, Herzberg PY. Validation and standardization of the Generalized Anxiety Disorder Screener (GAD-7) in the general population. Med Care. 2008 Mar; 46(3):266-74.

McLeod, J. D., Kessler, R, C. & Landis, K. R. (1992) Speed of recovery from major depressive episodes in a community sample of married men and women. Joii nal of Abnormal Psychology, 101, 277 -286.

Moncrieff, J., Wessely, S. & Hardy, R. (1998) Meta-analysis of trials comparing antidepressants with active placebos. British Journal of Psychiatry, 172, 227 -231.

Morris, Margaret E. "Mobile Therapy: Case Study evaluations of a Cell Phone Application for Emotional Self Awareness." Journal of Internet Medical Research 12.2 (2010): 1-10. 20 Apr. 2010. Web. 26 Apr. 2013.

Murray, C. J. L. & Lopez, A. D. (1996) Global Burden of Disease. Cambridge, MA; Harvard University Press.

Mynors- Walks, L. M., Gath, D. H., Lloyd-Thomas, A. R., et al (1995) Randomised controlled trial comparing problem solving treatment with amitriptyline and placebo for major depression in primary care. British Medical Journal. 310, 441 -445.

Palmier-Claus, Jasper E, "Integrating Mobile-phone Based Assessment for." BMC Psychiatry 13.34 (2013): 1-12.

Quality Assurance Project (1982) A treatment outline for agoraphobia. Australian and New Zealand Journal of Psychiatry, 16, 25 -33. Quality Assurance Project (1983) A treatment outline for depressive disorders. Australian and New Zealand Journal of Psychiatry, 17, 129 -146.

Quality Assurance Project (1984) A treatment outline for the management of schizophrenia. Australian and New Zealand Journal of Psychiatry, 18, 19 -38.

Quality Assurance Project (1985a) Treatment outlines for the management of anxiety states. Australian and New Zealand Journal of Psychiatry, 19, 138 -151.

Quality Assurance Project (1985b) Treatment outlines for the management of obsessive- compulsive disorders. Australian and New Zealand Joivrnal of Psychiatry, 19, 240 -253. Quitkin, F. M., Rabkin, J. G., Gerald, J., et al (2000) Validity of clinical trials of antidepressants. American Journal of Psychiatry, 157, 327 -337.

Rizzo R, Piccinelli M, Mazzi MA, Bellantuono C, Tansella M. (2000)

The Personal Health Questionnaire: a new screening instrument for detection of ICD-10 depressive disorders in primary care. Psychol Med. 2000 Jul; 30(4):831 -40.

Sanghara H, Kravariti E, Jakobsen H, Okocha C, (2010) Using short message services in mental health services assessing feasibility. ", Mental Health Review Journal, Vol. 15 Iss: 2, pp.28 - 3.

Spitzer RL, roenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006 May 22; 166(10): 1092-7.

Stone A, Briggs J, Smith C. (2002) SMS and Interactivity - Some Results from the Field, and its Implications on Effective Uses of Mobile Technologies (2002). In: IEEE International Workshop on Wireless and Mobile Technologies in Education; 29-30 Aug 2002, Vaxjo, Sweden. ISBN 0769517064

Thase, M. E. (1999) How should efficacy be evaluated in randomized clinical trials of treatments for depression? Journal of Clinical Psychiatry, 60 (suppl. 4), 23 -31

Vanderplas, J. M., & Garvin, E. A. (1959). The association value of random shapes. Journal of Experimental Psychology, 3, 147-154.

World Health Organization (1 92) International Classification of Diseases and Related Disorders (ICD-10). Geneva: World Health Organization.

Hinton, G; Machine Learning Coxirse Notes; Department of Computer Science, University of Toronto; 2013. Available here: http://goo.gl/6S2Ee

Hinton G; Advanced Topics in Machine Learning; Department of Computer Science, University of Toronto; 2013. Available here: http://goo.gl VQ243

Hinton G; Special Topics in Machine Learning; Department of Computer Science, University of Toronto; 2013. Available here: http://goo.gl/isdbp

Ramsay, J.O. and Silverman, B.W; Functional Data Analysis; Springer Series in Statistics; 2005

Claims

What is claimed is:

1. A mobile system comprising processing structure configured to:

electronically transmit collected data about ambient factor and/or usage of the mobile system to a remote processing structure;

receive, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data; at each of the provided dates/times, automatically trigger execution of one or more user- interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and

electronically transmit the received user data to the remote processing structure.

2. The mobile system of claim 1, wherein the mobile system is a single mobile device.

3. The mobile system of claim 1, wherein the mobile system is a plurality of mobile devices, wherein the mobile devices are in close-range communication with one another, and further wherein the processing structure comprises at least one computer processor on each mobile device.

4. The mobile system of claim 3, wherein at least one of the mobile devices is one of: a smartphone, a laptop computing device, a wrist-mountabie computing device, a head-mountable computing device.

5. The mobile system of claim 3, wherein the mobile devices are in close-range communication with one another via a wireless connection.

6. The mobile system of claim 5, wherein the wireless connection is established based on one of: Wi-Fi, Bluetooth, Zigbee, ANT, IEEE 802.15.4, Z-Wave.

7. The mobile system of any one of claims 1 to 6, wherein the one or more user-interactive programs are selected from the group consisting of: a psychiatric assessment questionnaire, a cognitive assessment game.

8. The mobile system of any one of claims 1 to 7, wherein the processing structure is configured to transmit the received user data in association with concurrent data relating to ambient factor and/or usage of the mobile device.

9. The mobile system of any one of claims 1 to 7, wherein the processing structure is configured to transmit user data relating to one or more failures of the user to input user data when requested, wherein the user data relating to the one or more failures is transmitted in association with concurrent data relating to ambient factor and/or usage of the mobile device.

10. The mobile system of any one of claims 1 to 9, wherein the processing structure is configured to transmit the received user data in association with data relating to the dates/times at which the user data was inputted.

11. The mobile system of any one of claims 1 to 10, wherein ambient factor data comprises data relating to at least one of: physical location of the mobile system, other devices in proximity to the mobile system, background light levels, background audio levels, recorded sound levels, video and/or still images captured by the mobile system.

12. The mobile system of claim 11, wherein the data relating to physical location of the mobile system comprises at least one of: global positioning data, location data derived from local wireless networks.

13. The mobile system of claim 11, wherein the data relating to other devices in proximity to the mobile system comprises data relating to proximate wireless networks.

14. The mobile system of claim 13, wherein the data relating to proximate wireless networks comprise data relating to at least one of: proximate Wi-Fi networks, proximate Bluetooth networks.

15. The mobile system of any one of claims 1 to 14, wherein usage data comprises data relating to at least one of: call activity on the mobile system, messaging activity on the mobile system, screen usage of the mobile system.

16. The mobile system of claim 15, wherein the data relating to call activity comprises data relating to at least one of: duration of calls, time of calls, frequency of calls, counterparties on calls, whether calls are incoming or outgoing, whether calls are unanswered, recorded voice samples.

17. The mobile system of claim 15, wherein the data relating to messaging activity comprises data relating to at least one of: time of messages, counterparties on messages, frequency of messages, whether messages are incoming or outgoing, whether messages are unresponded to, language used in messages.

18. The mobile system of claim 15, wherein the data relating to screen usage comprises data relating to at least one of: times of screen use, application use.

19. The mobile system of any one of claims 1 to 18, wherein the processing structure is configured to:

produce the collected data by selectively capturing and storing at least a portion of all ambient factor and or usage data from one or more data streams being automatically produced during operation of the mobile system.

20. The mobile system of any one of claims 1 to 19, wherein the processing structure is configured to:

electronically transmit each of the collected data and the received user data periodically in batches.

21. A non-transitory computer readable medium embodying a computer program executable on a processing structure of a mobile system, the computer program comprising:

computer program code for electronically transmitting collected data about ambient factor and/or usage of the mobile system to a remote processing structure;

computer program code for receiving, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data;

computer program code for, at each of the provided dates/times, automatically triggering execution of one or more user-interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and

computer program code for electronically transmitting the received user data to the remote processing structure.

22. The non-transitory computer readable medium of claim 21, wherein the mobile system is a single mobile device.

23. The flon- transitory computer readable medium of claim 21, wherein the mobile system is a plurality of mobile devices, wherein the mobile devices are in close-range communication with one another, and further wherein the processing structure comprises at least one computer processor on each mobile device.

24. The non-transitory computer readable medium of claim 23, wherein at least one of the mobile devices is one of: a smartphone, a laptop computing device, a wrist-mountable computing device, a head-mountable computing device.

25. The non-transitory computer readable medium of claim 23, wherein the mobile devices are in close-range communication with one another via a wireless connection.

26. The non-transitory computer readable medium of claim 25, wherein the wireless connection is established based on one of: Wi-Fi, Bluetooth, Zigbee, ANT, IEEE 802.15.4, Z-Wave.

27. The non-transitory computer readable medium of any one of claims 21 to 26, wherein the one or more user-interactive programs are selected from the group consisting of: a psychiatric assessment questionnaire, a cognitive assessment game.

28. The non-transitory computer readable medium of any one of claims 21 to 27, comprising: computer program code for transmitting the received user data in association with concurrent data relating to ambient factor and/or usage of the mobile device.

29. The non-transitory computer readable medium of any one of claims 21 to 27, comprising: computer program code for transmitting user data relating to one or more failures of the user to input user data when requested, wherein the user data relating to the one or more failures is transmitted in association with concurrent data relating to ambient factor and/or usage of the mobile device.

30. The non-transitory computer readable medium of any one of claims 21 to 29, comprising: computer program code for transnutting the received user data in association with data relating to the dates/times at which the user data was inputted.

31. The non-transitory computer readable medium of any one of claims 21 to 30, wherein ambient factor data comprises data relating to at least one of; physical location of the mobile system, other devices in proximity to the mobile system, background light levels, background audio levels, recorded sound levels, video and/or still images captured by the mobile system.

32. The non- transitory computer readable medium of claim 31, wherein the data relating to physical location of the mobile system comprises at least one of: global positioning data, location data derived from local wireless networks.

33. The non-transitory computer readable medium of claim 31, wherein the data relating to other devices in proximity to the mobile system comprises data relating to proximate wireless networks.

34. The non-transitory computer readable medium of claim 33, wherein the data relating to proximate wireless networks comprise data relating to at least one of: proximate Wi-Fi networks, proximate Bluetooth networks.

35. The non-transitory computer readable medium of any one of claims 21 to 34, wherein usage data comprises data relating to at least one of: call activity on the mobile system, messaging activity on the mobile system, screen usage of the mobile system.

36. The non-transitory computer readable medium of claim 35, wherein the data relating to call activity comprises data relating to at least one of: duration of calls, time of calls, frequency of calls, counterparties on calls, whether calls are incoming or outgoing, whether calls are unanswered, recorded voice samples.

37. The non-transitory computer readable medium of claim 35, wherein the data relating to messaging activity comprises data relating to at least one of: time of messages, counterparties on messages, frequency of messages, whether messages are incoming or outgoing, whether messages are unresponded to, language used in messages.

38. The non-transitory computer readable medium of claim 35, wherein the data relatmg to screen usage comprises data relating to at least one of: times of screen use, application use.

39. The non-transitory computer readable medium of any one of claims 21 to 38, comprising: computer program code for producing the collected data by selectively capturing and storing at least a portion of all ambient factor and/or usage data from one or more data streams being automatically produced during operation of the mobile system.

40. The non-transitory computer readable medium of any one of claims 21 to 39, comprising; computer program code for electronically transmitting each of the collected data and the received user data periodically in batches.

41. A method in a mobile system, the method comprising:

electronically transmitting collected data about ambient factor and/or usage of the mobile system to a remote processing structure; receiving, from the remote processing structure, electronic date/time data including a plurality of dates/times determined by the remote processing structure based on the transmitted collected data; at each of the provided dates/times, automatically triggering execution of one or more user- interactive programs on the mobile system thereby to receive user data inputted by a user of the mobile system; and

electronically transmitting the received user .data to the remote processing structure.

42. The method of claim 41 , wherein the mobile system is a single mobile device.

43. The method of claim 41, wherein the mobile system is a plurality of mobile devices, wherein the mobile devices are in close- range communication with one another.

44. The method of claim 43, wherein at least one of the mobile devices is one of: a smartphofle, a laptop computing device, a wrist-mountable computing device, a head-mountable computing device.

45. The method of claim 43, wherein the mobile devices are in close-range communication with one another via a wireless connection.

46. The method of claim 45, wherein the wireless connection is established based on one of: Wi- Fi, Bluetooth, Zigbee, ANT, IEEE 802.15.4, Z-Wave.

47. The method of any one of claims 41 to 46, wherein the one or more user-interactive programs are selected from the group consisting of: a psychiatric assessment questionnaire, a cognitive assessment game.

48. The method of any one of claims 41 to 47, comprising:

transmitting the received user data in association with concurrent data relating to ambient factor and/or usage of the mobile device.

49. The method of any one of claims 41 to 47, comprising:

transmitting user data relating to one or more failures of the user to input user data when requested, whereiu the user data relating to the one or more failures is transmitted in association with concurrent data relating to ambient factor and/or usage of the mobile device.

50. The method of any one of claims 41 to 49, comprising:

transmitting the received user data in association with data relating to the dates/times at which the user data was inputted.

51. The method of any one of claims 41 to 50, wherein ambient factor data comprises data relating to at least one of: physical location of the mobile system, other devices in proximity to the mobile system, background light levels, background audio levels, recorded sound levels, video and/or still images captured by the mobile system.

52. The method of claim 51, wherein the data relating to physical location of the mobile system comprises at least one of: global positioning data, location data derived from local wireless networks.

53. The method of claim 51, wherein the data relating to other devices in proximity to the mobile system comprises data relating to proximate wireless networks.

54. The method of claim 53, wherein the data relating to proximate wireless networks comprise data relating to at least one of: proximate Wi-Fi networks, proximate Bluetooth networks.

55. The method of any one of claims 41 to 54, wherein usage data comprises data relating to at least one of: call activity on the mobile system, messaging activity on the mobile system, screen usage of the mobile system.

56. The method of claim 55, wherein the data relating to call activity comprises data relating to at least one of: duration of calls, time of calls, frequency of calls, counterparties on calls, whether calls are incoming or outgoing, whether calls are unanswered, recorded voice samples.

57. The method of claim 55, wherein the data relating to messaging activity comprises data relating to at least one of: time of messages, counterparties on messages, frequency of messages, whether messages are incoming or outgoing, whether messages are unresponded to, language used in messages,

58. The method of claim 55, wherein the data relating to screen usage comprises data relating to at least one of: times of screen use, application use,

59. The method of any one of claims 41 to 58, comprising:

producing the collected data by selectively capturing and storing at least a portion of all ambient factor and/or usage data from one or more data streams being automatically produced during operation of the mobile system.

60. The method of any one of claims 41 to 59, comprising: electronically transmitting each of the collected data and the received user data periodically in batches.

61. A system comprising:

a plurality of mobile systems as claimed in one of claims 1 to 20;

a remote processing structure in electronic communication with each of the mobile systems, the remote processing structure comprising at least one database for storage and. aggregation of respective collected data and user data received from each of the plurality of mobile devices.

62. The system of claim 61, wherein the remote processing structure comprises at least one computer processor configured to generate the electronic date/time data including a plurality of dates/times by processing the transmitted collected data.

63. The system of claim 61 , wherein the at least one computer processor is configured to process at least the transmitted collected data to identify the plurality of dates/times as dates/times when baseline conditions for receiving user data at respective mobile devices are likely to be present.

64. A computing system comprising processing structure configured to:

automatically receive data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system;

automatically process the collected data to construct an electronic behaviour profile for the or each subject; and

categorize each subject based on their electronic behaviour profile.

65. The computing system of claim 64, wherein the categorizing comprises segmenting each subject as qualified or unqualified for a clinical trial.

66. A non-transitory processor-readable medium embodying a computer program executable on a computing system, the computer program comprising:

computer program code for automatically receiving data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system;

computer program code for automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and

computer program code for categorizing each subject based on their electronic behaviour profile.

67. The non-transitory processor-readable medium of claim 66, wherein the categorizing comprises segmenting each subject as qualified or unqualified for a clinical trial.

68. A processor-implemented method comprising:

automatically receiving data collected from at least one mobile system carried by a respective subject about respective ambient factor and/or usage of the mobile system;

automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and

categorizing each subject based on their electronic behaviour profile.

69. 'The processor-implemented method of claim 68, wherein the categorizing comprises segmenting each subject as qualified or unqualified for a clinical trial.

70. A computing system comprising processing structure configured to:

display one or more graphical representations of the electronic behaviour profile.

71. The computing system of claim 70, wherein at least one of the graphical representations, of each electronic behaviour profile depict clusters of behaviour.

72. A non-transitory processor-readable medium embodying a computer program executable on a computing system, the computer program comprising:

computer program code for displaying one or more graphical representations of the electronic behaviour profile.

73. The processor-readable medium of claim 72, wherein at least one of the graphical representations of each electronic behaviour profile depict clusters of behaviour.

74. A processor-implemented method comprising; automatically receiving data collected from at least one morale system carried by a respective subject about respective ambient factor and/or usage of the mobile system;

displaying one or more graphical representations of the electronic behaviour profile.

75. The processor-implemented method of claim 74, wherein at least one of the graphical representations of each electronic behaviour profile depict clusters of behaviour.

76. A computing system comprising processing structure configured to:

receive, from each of at least one mobile system, data collected by the mobile system about ambient factor and/or usage of the mobile system;

process the collected data to determine, for each of the at least one mobile system, a plurality of dates/times at which one or more user-interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data; and

transmit a respective plurality of dates/times to each of the mobile systems.

77. The computing system of claim 76, wherein the processing structure is configured to:

determine each plurality of dates/times to correspond to dates/times at which baseline behaviour for the respective user can be inferred from the electronic behaviour profile.

78. A iion- transitory processor-readable medium embodying a computer program executable on a computing system, the computer program comprising:

computer program code for receiving, from each of at least one mobile system, data collected by the mobile system about ambient factor and/or usage of the mobile system;

computer program code for processing the collected data to determine, for each of the at least one mobile system, a plurality of dates/times at which one or more user-interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data; and

computer program code for transmitting a respective plurality of dates/times to each of the mobile systems.

79. The processor-readable medium of claim 78, comprising:

computer program code for automatically processing the collected data to construct an electronic behaviour profile for the or each subject; and computer program code for determining each plurality of dates/times to correspond to dates/times at which baseline behaviour for the respective user can be inferred from the electronic behaviour profile.

80. A processor-implemented method comprising:

receiving, from each of at least one mobile system, data collected by the mobile system about ambient factor and/or usage of the mobile system;

processing the collected data to determine, for each of the at least one mobile system, a plurality of dates/times at which one or more user-interactive programs should advantageously be triggered on the mobile system to request the user of the mobile system to input user data; and

transmitting a respective plurality of dates/times to each of the mobile systems.

81. The method of claim 80, wherein the processing comprises:

determining each plurality of dates/times to correspond to dates/times at which baseline behaviour for the respective user can be inferred from the electronic behaviour profile,