System for determining timing of therapy and method of timing therapy based on biological process indicators

Abstract

Therapy and / or treatment planning methods for administering substances to or performing procedures on a patient can be timed using a modified scheduling process, typically implemented as a computer process, where the non-linear treatment mapping is based on a circadian trajectory calculated from patient-specific inputs and uncertainty ranges.

Description

[Technical Field]

[0001] [Cross reference] This application claims the benefit of, and priority from, U.S. Provisional Patent Application No. 63 / 382,856, filed November 8, 2022, entitled "System for Determining Treatment Timing and Methods of Treatment Timed Based on Biological Process Indicators," and U.S. Patent Application No. 18 / 310,791, filed May 2, 2023, entitled "System For Determining Treatment Timing And Methods Of Treatment Timed Based on Biological Process Indicators." The entire disclosures of the above-listed applications are hereby incorporated by reference for all purposes as if fully set forth herein.

[0002] The present disclosure relates generally to methods of treatment, and more particularly to methods of treatment that are timed based on circadian trajectories calculated from biological process indicators. [Background technology]

[0003] It has long been known that the time of day a drug is taken can affect its efficacy and / or toxicity. More generally, biological therapies administered to human patients or other animals may be administered at specific times to maximize efficacy and minimize toxicity. Biological therapies may be, for example, the administration of a drug, nutrient, behavioral modification, or another substance, orally, intravenously, intramuscularly, transdermally, or via other routes. Biological therapies may be in addition to or instead of the administration of a substance. For example, biological therapies may be the administration of kidney dialysis, a surgical procedure, or another medical procedure. The terms "chronomedicine" and "chronomedical" are used in the literature to refer to considering the influence of the time of day on biological therapies, and "chronomodulation" and "chronomodulated" may refer to scheduling the timing of biological therapies based on the time of day, a process that may be referred to as "chronotherapy."

[0004] Examples of chronomedicine and chronomodulation in the literature include [

number

[2010] , which reported on rodent studies showing that over 40 anticancer drugs have different toxicity depending on the time of administration and / or different efficacy profiles over the course of a day. [Hrushesky] and [

number

number

[1997] reported up to a fivefold reduction in grade 3-4 mucositis and a halving of the incidence of neuropathy when drugs were given in a chronomodulated manner. More recently, the effect of time of day on overall treatment efficacy has been reported for temozolomide, a glioblastoma treatment [Damato], and for immunotherapy [Qian], with morning dosing being superior in both cases. Cancer is the condition with the most chronomedicinal research, but it is not the only one: morning versus evening effects have also been reported for treatment of conditions such as asthma, diabetes, and hypertension.

[0005] [Sato] describes how local and whole-body metabolic responses to exercise can vary based on the time of day.

[0006] Current methods of chronomedicine and chronomodulation are very limited and require improvement. References

[0007] [Adam] Adam, “Core Concepts: Emerging Science of Chronotherapy Offers Big Opportunities to Optimize Drug Delivery,” (2019) PNAS 116(44):21957-59 [doi: org / 10.1073 / pnas.1916118116].

[0008] [Damato] Damato et al., “A Randomized Feasibility Study Evaluating Temozolomide Circadian Medicine in Patients with Glioma,” (31 Jan 2022) Neuro-Oncology Practice 9(3):193-200 [doi: 10.1093 / nop / npac003].

[0009] [Damato2021] Damato et al., 「Temozolomide Chronotherapy in Patients with Glioblastoma: A Retrospective Single-Institute Study,」 (02 Mar 2021) Neuro-Oncology Advances 3(1):1-11, vdab041 [doi: org / 10.1093 / noajnl / vdab041]。

[0010] [Giacchetti] Giacchetti et al. 「Sex Moderates Circadian Chemotherapy Effects on Survival of Patients with Metastatic Colorectal Cancer: A Meta-Analysis,」 (2012) Annals of Oncology 23(12):3110-3116 [doi: org / 10.1093 / annonc / mds148]。

[0011] [Hill] Hill et al., 「Optimizing Circadian Drug Infusion Schedules towards Personalized Cancer Chronotherapy,」 (1985) PLoS Computational Biology 16(1):e1007218 (2020) [doi: org / 10.1371 / journal.pcbi.1007218]。

[0012] [Hrushesky] Hrushesky, 「Circadian Timing of Cancer Chemotherapy.」 (1985) Science 228(4695):73-75 [doi: 10.1126 / science.3883493]。

[0013] [Kramer] Kramer et al., 「Foundations of Circadian Medicine,」 (24 Mar 2022) PLoS Biology 20(3):e3001567 [doi: 10.1371 / journal.pbio.3001567; PMID: 35324893; PMCID: PMC8946668]。

[0014] [Lee] Lee et al., 「Time-of-Day Specificity of Anticancer Drugs May be Mediated by Circadian Regulation of the Cell Cycle,」 (12 Feb 2021) Science Advances 7(7): eabd2645 [doi: 10.1126 / sciadv.abd2645]。

[0015]

Number

Number

[0016]

Number

Number

[0017]

Number

Number

[0018]

Number

Number

[0019] [Montaigne] Montaigne et al., 「Daytime Variation of Perioperative Myocardial Injury in Cardiac Surgery and Its Prevention by Rev-Erbα Antagonism: A Single-Centre Propensity-Matched Cohort Study and a Randomised Study,」 (2018) The Lancet 391(10115):59-69 [doi: org / 10.1016 / S0140-6736(17)32132-3].

[0020] [Mormont] Mormont et al., 「Cancer Chronotherapy: Principles, Applications, and Perspectives,」 (2003) Cancer 97(1):155-169 [doi: org / 10.1002 / cncr.11040].

[0021]

Number

Number

[0022] [Peeples] Peeples, 「Medicine's Secret Ingredient - it's in the timing」 (2018) Nature News Feature 556(7700):290-292 [doi.org / 10.1038 / d41586-018-04600-8]。

[0023] [Qian] Qian et al., 「Effect of Immunotherapy Time-of-Day Infusion on Overall Survival among Patients with Advanced Melanoma in the USA (MEMOIR): A Propensity Score-Matched Analysis of a Single-Centre, Longitudinal Study,」 (12 Nov 2021) The Lancet Oncology 22(12):1777-1786 [doi: org / 10.1016 / S1470-2045(21)00546-5]。

[0024] [Sato] Sato et al., 「Atlas of Exercise Metabolism Reveals Time-Dependent Signatures of Metabolic Homeostasis,」 (1 February 2022) Cell Metabolism, 34(2):329-345, e1-e8 [doi: org / 10.1016 / j.cmet.2021.12.016]。

[0025] [Selfridge] Selfridge et al., 「Chronotherapy: Intuitive, Sound, Founded…But Not Broadly Applied,」 (2016) Drugs 76:507-1521 [doi: org / 10.1007 / s40265-016-0646-4].

[0026] [Wang] Wang et al., 「The Circadian Immune System,」 (3 Jun 2022) Science Immunology 7(72):eabm2465 [doi: 10.1126 / sciimmunol.abm2465].

[0027] [Wang] Wang et al., 「Time of Day of Vaccination Affects SARS-CoV-2 Antibody Responses in an Observational Study of Health Care Workers,」 (1 Feb 2022) Journal of Biological Rhythms 37(1):124-129 [doi: org / 10.1177 / 07487304211059315].

[0028] [Zhang] Zhang et al., 「A Circadian Gene Expression Atlas in Mammals: Implications for Biology and Medicine,」 (2014) PNAS 111(45):16219-16224 [doi: org / 10.1073 / pnas.1408886111].

Summary of the Invention

[0029] Therapy and / or treatment planning methods for administering substances to or performing procedures on a patient can be timed using a modified scheduling process, typically implemented as a computer process, where the non-linear treatment mapping is based on a circadian trajectory calculated from patient-specific inputs and uncertainty ranges.

[0030] This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter. A more extensive presentation of the features, details, utilities, and advantages of the methods and apparatus as defined in the claims is provided in the following written description of various embodiments of the present disclosure and illustrated in the accompanying drawings. [Brief explanation of the drawings]

[0031] Various embodiments according to the present disclosure will now be described with reference to the drawings.

[0032] [Figure 1] FIG.

[0033] [Figure 2] FIG. 1 is a diagram of the circadian trajectory.

[0034] [Figure 3] FIG. 1 is a diagram of a trajectory bundle.

[0035] [Figure 4] FIG. 1 is a diagram of the circadian trajectory bundle.

[0036] [Figure 5] FIG. 1 shows a set of trajectories for clock gene expression given light exposure history.

[0037] [Figure 6]FIG. 6 shows trajectory bundles, one for light history such as that shown in FIG. 5 and one for clock gene expression circadian trajectories such as that shown in FIG. 5.

[0038] [Figure 7] FIG. 1 illustrates an example of a trajectory with some uncertainty built in.

[0039] [Figure 8] FIG. 1 shows an exemplary circadian mapping profile, or efficacy profile, mapping from circadian state to drug efficacy.

[0040] [Figure 9] FIG. 10 shows another exemplary circadian mapping profile, i.e., toxicity profile, that maps from circadian state to drug toxicity.

[0041] [Figure 10] FIG. 1 is a diagram of a patient care system according to various embodiments.

[0042] [Figure 11] FIG. 1 is a diagram of a treatment allocation and coordination system according to various embodiments.

[0043] [Figure 12] FIG. 1 is a diagram of a treatment scheduling feedback system according to various embodiments.

[0044] [Figure 13] FIG. 1 illustrates components that may be used in a control system to vary environmental cues and behavioral triggers in response to fixed treatment times and user data, according to various embodiments.

[0045] [Figure 14] FIG. 1 illustrates components that may be used for digital twin simulation, according to various embodiments.

[0046] [Figure 15] FIG. 1 illustrates components that may be used for a profile identification system, according to various embodiments.

[0047] [Figure 16] FIG. 1 illustrates components that may be used for a treatment mapping module, according to various embodiments.

[0048] [Figure 17] FIG. 1 illustrates components that may be used for a treatment system module, according to various embodiments.

[0049] [Figure 18] FIG. 1 illustrates components that may be used for a control system, according to various embodiments.

[0050] [Figure 19] FIG. 1 illustrates components that may be used for a treatment allocation and coordination system, according to various embodiments.

[0051] [Figure 20] FIG. 10 illustrates another example of trajectory uncertainty, according to various embodiments.

[0052] [Figure 21] FIG. 1 illustrates an exemplary computer system memory structure that may be used to perform the methods described herein, according to various embodiments.

[0053] [Figure 22] FIG. 22 is a block diagram illustrating an example computer system on which the systems illustrated in FIGS. 1 and 21 may be implemented, according to various embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0054] In the following description, various embodiments are described. For purposes of explanation, specific configurations and details are set forth in order to provide a thorough understanding of the embodiments. However, it will also be apparent to those skilled in the art that the embodiments may be practiced without the specific details. Furthermore, well-known features may be omitted or simplified so as not to obscure the described embodiments.

[0055] The patient treatment system obtains various inputs, such as wearable device data from a wearable device worn by the patient, other sensor data from the patient, calendar data from the patient's calendar program, and inputs of location, movement, environment, etc. As described herein, a circadian trajectory is determined that maps between elapsed wall-clock time and circadian state from some or all of these inputs, and possibly also from pre-computed machine learning models and stored data not necessarily specific to the patient. The circadian trajectory may have different degrees of uncertainty from patient to patient and / or from time to time for a given patient, for example, if relatively few inputs are available, if the inputs are inconsistent, and / or for other reasons. From the circadian trajectory and the circadian mapping profile for a given treatment, the patient treatment system can determine the elapsed wall-clock time, time points, and / or time ranges for administering the treatment.

[0056] Highly accurate measurements of a patient's circadian state at a given time may be obtained with significant inconvenience to the patient and significant processing delays. For example, while multiple saliva samples can be collected while the patient is in a dark room for hours, this is typically impractical, and the patient's circadian state and circadian trajectory over that time are only known after the test results are completed, perhaps several days later, and therefore may be of limited use. Using the methods and devices described herein, good estimates of circadian state can be obtained in near-real time with minimal effort and inconvenience to the patient. In some cases, the methods and devices described herein may not be able to determine circadian state with high accuracy at some times (e.g., when the patient is mobile or engaged in irregular activities), while being more accurate at other times.

[0057] The time point, or time range, for signaling to administer or administer a therapy can be determined based on, for example, a circadian mapping profile that indicates which circadian states are associated with high efficacy and / or low toxicity. In a specific example, if a patient's circadian state at a snapshot at a certain point in time is represented in memory by a value between 0.0 and 1.0, and the patient treatment system determines that efficacy for drug D is highest at a state value of CS=0.4 and toxicity of drug D is also lowest at that state value, and further determines that CS=0.4 corresponds to 10 AM with an uncertainty of 3 minutes, the patient treatment system may send a message to the patient, perhaps on a wearable device, perhaps some time in advance, to the effect, "Take tablet D sometime between 9:57 and 10:03 AM for best results."

[0058] Other scales for circadian state are possible, such as circadian states that can be considered like circular phases and vary continuously from 0 to 2*π (in radians), which the patient care system may consider equivalent to 0. Circadian trajectories can represent past circadian state change as a function of elapsed real time and / or expected or predicted future circadian state change as a function of elapsed real time. Elapsed real time can be represented by clock circuits, time-keeping computer circuits, and / or computer elements that receive signals representing time measured independently of details such as physical activity or local time, and time can have different resolutions, such as {AM, day, noon, PM, night, midnight}, to HH hours, HH:MM on DD day, HH:MM on DD month, Y year, HH:MM:SS on DD day.

[0059] Elapsed real time can be represented in many ways, such as a stored value between 0.0 and 1.0 representing a time period, perhaps with a resolution of seconds or some other interval, a stored value between 0:00 and 23:59:59 representing a time period, or in epoch time, e.g., the number of seconds elapsed since some specified point in time and date. As an example of epoch time, a stored time value t, where t=1642724939, may represent the elapsed real time of 12:28:59 AM UTC on January 21, 2022.

[0060] In some cases, the patient treatment system may maintain multiple circadian trajectories for a patient and use one or more of the maintained circadian trajectories for a particular treatment management. An example of multiple maintained circadian trajectories may include a central circadian trajectory that maps the body's central circadian state to elapsed real time, where the central circadian state corresponds to the body's suprachiasmatic nucleus (SCN), such as a memory representation of how neuronal firing patterns change over time and / or instantaneous firing patterns in the SCN that correspond to patterns in changing concentrations of clock genes within cells.

[0061] The peripheral circadian state can reflect the concentration of circadian-related molecules in the body and a peripheral circadian trajectory that represents changes in that peripheral circadian state over a period of elapsed real time. For example, one particular peripheral circadian state can represent the current concentration of sodium proton exchange transporter (NHE3) in the intestinal lumen of a patient.

[0062] Examples of treatments may include administering a medication to the patient, outputting a message to be received by the patient for the patient to take the medication, and / or indicating an optimal time or estimate for some other treatment. If the treatment has a fixed elapsed real time, e.g., for a surgery scheduled for November 18th at 2:00 PM, the patient treatment system may calculate what inputs may affect the circadian trajectory, provide the patient with the inputs, and / or administer some pre-operative treatment that is expected to cause a shift in the patient's circadian trajectory so that the desired circadian state aligns with the fixed elapsed real time of the treatment.

[0063] Determining circadian trajectories can be a complex process because some input data may be missing and / or there may be some disruptions in the patient's normal schedule, such as travel across time zones. In some cases, circadian trajectories may need to be calculated on the fly, in real time, or near real time, and therefore some optimization in processing the data may be required. Furthermore, some of the calculations may be performed on low-power devices. In some implementations, some of the calculations are performed remotely, perhaps on a more powerful server, and some of the calculations are performed on a low-power, local, wearable, and / or portable device with communication capabilities.

[0064] In a typical configuration, a circadian trajectory is continuous in that two nearby points on the circadian trajectory map to nearby elapsed real-time. While a circadian trajectory may map strictly to elapsed real-time, e.g., a linear mapping, more typically, a circadian trajectory will compress or expand the unit of circadian time relative to elapsed real-time, which can be treated as a circadian clock and a wall clock running at different rates and ratios of different rates that vary over a day or other measurement period. Overall, a complete circadian trajectory can provide a mapping from circadian state to elapsed real-time, as well as a mapping from elapsed real-time to circadian state. Circadian trajectories are also recursive in nature, with previous estimates feeding into predictions for the next time step. Circadian trajectories also generate estimates for time scales significantly shorter than the oscillator period (approximately 24 hours). For example, a circadian trajectory may estimate the oscillator state every 6 minutes, generating 240 predicted or expected states per day.

[0065] Some treatments may be prescribed strictly relative to elapsed real time (e.g., "take one tablet early in the morning and a second at noon local time"), while other treatments may rely on circadian time, which represents some circadian state on a circadian trajectory. Such other treatments may benefit patients more if their circadian trajectory and / or current circadian state are more accurately and / or precisely determined.

[0066] The circadian trajectory may optionally be represented as a look-up table corresponding to a piecewise linear plot of circadian time versus elapsed actual time and stored in computer memory. The patient treatment system may store the circadian trajectory as a vector data structure that can be manipulated by computer instructions written to perform vector operations or by hardware capable of vector operations. The resolution of such a plot of circadian time versus elapsed actual time may be on the order of seconds or minutes. For example, a 24-hour plot of circadian time versus elapsed actual time may be stored as a vector of 288 values, each corresponding to a circadian state at a 5-minute interval.

[0067] A patient's body has biological states that can fluctuate over time. For example, a stored biological state value may reflect the concentration of melatonin in the patient's saliva, the concentration of caffeine in the patient's bloodstream, the concentration of the Bmal1 gene in the patient's brain, the current firing rate in the patient's suprachiasmatic nucleus, etc. For non-human patients, some biological states may not have a human counterpart. Circadian state is one particular example of a biological state.

[0068] A patient's circadian state is a biological state that corresponds to where the patient is in the circadian cycle. As described herein, a given body may have multiple circadian cycles, not all of which need to be aligned, such as central and peripheral circadian cycles. Melatonin concentration, Bmal1 gene concentration, SCN firing rate, and other biological states may be circadian state or thus proxies.

[0069] The time sequence of a biological state may be stored and / or represented in memory as a time series of state values, or a trajectory for the biological state. The representation of the trajectory in computer memory may be in the form of a list of records, each indicating a state value and a time point, a piecewise linear plot, coefficients of a fitting curve, or any other form of data structure that can be used to represent state values ​​and specified time points.

[0070] An example biological state may be a melatonin concentration of 4.3 picograms / ml at epoch time t=1642724939. Another example biological state may be a core body temperature of 97.3°F (36.278°C) at epoch time t=1642724939. The value R may be stored in memory as a representation of some unitless or unit-specific quantity corresponding to the biological state. For example, it may be the value of a biological oscillator, such as the concentration of melatonin in saliva in picograms per milliliter. Another example may be a unitless quantity corresponding to a measure of cohesion of neuronal firing in the suprachiasmatic nucleus, where R=0 represents neurons firing completely out of sync with one another and R=1 represents neurons firing completely in sync with one another.

[0071] An example of a circadian state is R a The trajectory may be represented as a sinusoidal curve defined by having an amplitude of Ψ = 0.4 and a phase of Ψ = π / 4 at an epoch time of t = 1642724939. The period of the trajectory, which may correspond to a circadian cycle, may not be constant and may vary from period to period (e.g., a 24.75 hour circadian cycle followed by a 25.4 hour circadian cycle, then a 24.9 hour circadian cycle). The effective period of the trajectory, the rate of change of state, may also change dynamically over the course of the trajectory as time contracts and expands.

[0072] In some cases, an intrinsic period may exist and / or be measured, which may correspond to the circadian cycle the body experiences throughout the day in the absence of external time-indicating signals such as changes in light. The patient treatment system may store values ​​for the intrinsic period, along with other parameters that may be of real importance to a person's circadian phenotype, such as light sensitivity, in defining trajectories, which may be used as parameters that can be adjusted to improve estimates of that person's circadian state. Varying such parameters can be used to generate trajectory bundles, and parameter ranges can be narrowed or widened if other information about those parameters is available, for example, by giving a person a test to see how much their pupils constrict to assess light sensitivity, or if demographic correlations that provide information about the parameters under consideration are discovered.

[0073] In another example, a circadian state may be stored as a representation of multiple gene expressions, such as a vector value [32.242, 484.23, 4994.2...] for t=16427249, representing expression levels for genes [Gene1, Gene2, Gene3,...]. For some such circadian states, there may be hundreds of genes represented.

[0074] FIG. 1 is a diagram of a trajectory 100. In this example, the biological state is the body's caffeine concentration over time following ingestion of caffeine at time t=0. In the plot shown, the scale of caffeine concentration ranges from A (which may or may not be 0) to A+B, which may be an arbitrary scale or the stored trajectory may include a specific scale. In this particular example, the caffeine concentration appears to decrease with an exponential decay. For some trajectories, the biological state may be measured at several time points and then fitted to a curve that exhibits a particular behavior, such as linear or exponential decay.

[0075] A trajectory is a mapping between real time and a biological state as either a past or future time series of biologically relevant data. An example may be a trajectory of adenosine concentration over time, a trajectory of TMZ concentration over time, etc. A trajectory may reflect the half-life of a drug concentration in the body, the interaction between a drug and other molecules in the body over time, etc.

[0076] 2 is a diagram of a circadian trajectory 200 representing a time series of circadian states. An example may be a biological state representing changes in melatonin in saliva over the course of a day, levels of firing patterns in the brain over the course of weeks, or expression rates of clock genes over any time scale. While this example of circadian trajectory 200 suggests a sinusoidal pattern over an approximately 24-hour period, that need not be the case for all circadian trajectories.

[0077] In addition to storing and manipulating trajectories of biological states, the patient care system may store, manipulate, and / or operate on bundles of multiple related trajectories. Different trajectories in a trajectory bundle may be created with different starting conditions and / or assumptions. A trajectory bundle may include a sampling of circadian trajectories, which should be considered as a finite representation obtained from a continuous trajectory probability distribution function (TPDF), or may encompass a representation of the entire distribution function (TPDF) or an approximation of this function. Thus, bundles may also be represented or obtained by an equation for the TPDF, and dynamics may be described by the expansion of any representation of this function.

[0078] FIG. 3 is a diagram of a trajectory bundle 300 of trajectories representing caffeine concentrations in the body. Different trajectories may represent different individuals, different measurements from a single person (or multiple people), and / or a set of measurements to which different assumptions or initial conditions apply. For example, the values ​​A+B shown in the plot in FIG. 3 may be normalized starting concentrations for different amounts of caffeine intake to indicate the probability ("fuzziness") of various biological states of the trajectories across multiple samples and across multiple caffeine intakes. In another example of a trajectory bundle, the trajectory bundle may represent trajectories for different average concentrations of adenosine receptors in the brain. For example, a trajectory bundle for four different coffee sizes consumed by a patient may be stored as a trajectory bundle of four trajectories, while a trajectory bundle for adenosine receptors may include a sampling of 10 trajectories, and a trajectory bundle of one trajectory for each combination of intake size and receptor concentration may include 40 trajectories.

[0079] FIG. 4 is a diagram of a circadian trajectory bundle 400. In this example, each trajectory is min ~Exp max For each trajectory, the expression levels at time t=0 are the same or normalized to be the same as shown, although this is not necessarily required for circadian trajectory bundles.

[0080] Trajectory bundles can be used to quantify uncertainty in wearable signals. A digital twin simulation can be run for 100,000 small modifications of the wearable history, where each wearable history corresponds to a different trajectory. The spread of the biological state shown across the trajectory bundles at every time point can provide the uncertainty for that time point.

[0081] Trajectory bundles can be used to quantify uncertainty in a user's circadian phenotype, such as the user's light sensitivity or endogenous tau. A digital twin simulation can be run with 10,000 selections of the circadian parameter tau, where each selection of tau corresponds to a different trajectory. The spread of trajectories at every time point can again provide the uncertainty for that time point. The uncertainty can be used to set the width of the dosing window. High uncertainty will correspond to a wide dosing window, while low uncertainty corresponds to a narrow dosing window.

[0082] Trajectory bundles can be used to patch missing data. In the absence of wearable data, for example, a digital twin simulation can be run over a distribution of possible wearable data histories. These wearable data histories can be selected from a uniform distribution or can be informed by historical data from the user (e.g., a "typical day"). Trajectory bundles can patch gaps during periods of missing data, reducing uncertainty in the circadian trajectory when wearable data resumes. For example, the circadian state at the moment wearable data returns can be obtained from the average state of the trajectory bundle, and the uncertainty at that moment can also be obtained from the trajectory bundle.

[0083] Trajectory bundles can be used to quantify the risk of drug interactions. For example, trajectory bundles can be used to calculate the risk of two interacting drugs, where different trajectories involve different effective half-lives of the drugs in the system. The risk of drug interactions calculated from trajectory bundles can be used to inform treatment mapping. A dosing window may not be recommended if the risk of a drug interaction is excessively high.

[0084] Trajectory bundles can be used to speed up expensive trajectory mapping studies. Optimization operations can often become prohibitively slow as complexity increases. For example, trajectory bundles can be calculated for the next two days using probabilistic wearable history (a "typical day"). Treatment mapping can then be applied to this hypothetical data prior to a user request for a dosing window. Treatment mapping will yield a dosing window for each trajectory in the bundle. When the user next asks for a dosing window, the trajectory that most closely resembles their actual trajectory will be identified, and the corresponding dosing window will be presented.

[0085] A dosing window is a recommended dosage or treatment with associated duration for dosing or treatment calculated from the circadian trajectory, where the condition may have associated uncertainty and one or more circadian mapping profiles, e.g., efficacy profile and / or toxicity profile. The patient treatment system may generate the dosing window by taking into account other biological trajectories as inputs, e.g., the trajectory of caffeine elimination from the body. The dosing window may be represented by a data structure indicating the timing window and dosage or other treatment suggestions and / or details.

[0086] An example of a dosing window may be data indicating a dosing window for M mg of Drug D between 2:00 and 2:30 PM based on the patient treatment system determining that the calculation of the expected circadian state during that time period has low uncertainty and that the expected circadian state, via the circadian mapping profile, maps to high efficacy and / or low toxicity for Drug D. Another example may be data indicating a dosing window for M mg of Drug D sometime during the afternoon based on the patient treatment system determining that the calculation of the expected circadian state during that time period has low uncertainty and that the expected circadian state, via the circadian mapping profile, maps to high efficacy and / or low toxicity for Drug D, where the expected circadian state has greater uncertainty.

[0087] The patient care system may provide messages to the patient that reflect uncertainty, for example: (a) "Take one TMZ tablet between 8:00 AM and 8:15 AM tomorrow" (low uncertainty), (b) "Take 45 mg of medication tomorrow afternoon" (high uncertainty), and (c) "Take a tablet tomorrow" (very high uncertainty).

[0088] Figure 5 shows a set of trajectories for clock gene expression given a light exposure history. The light exposure history may be obtained from a wearable device that periodically measures the amount or characteristics of light experienced by the wearer. The corresponding clock gene expression levels shown in Figure 5 may be determined from actual measurements of the wearer, or may be determined or estimated based on pre-computed volatile or light history to clock gene expression mappings.

[0089] As shown in Figure 5, each of the six examples shows a light history represented by a plot of light intensity over time (approximately 24-25 hours), which may be on some linear or nonlinear scale, with higher intensities (e.g., more light) being higher and lower light intensities (or no light) being closer to the x-axis in the plot. The corresponding clock gene expression plot over that same time span may be scaled based on measured or predetermined minimum and maximum clock gene expression values. While this is not always the case, in the examples shown in Figure 5, periods with greater light exposure generally correspond to higher clock gene expression. It should be understood that while the clock gene expression curve appears approximately sinusoidal, this need not necessarily be the case.

[0090] FIG. 6 shows trajectory bundles, one for a light history such as that shown in FIG. 5 and one for a clock gene expression circadian trajectory such as that shown in FIG. 5. A trajectory bundle may be a bundle of 10,000 trajectories, each representing a patient exposed to 10,000 distinct constant light levels over a one-hour period or some other time period. These may be from actual exposures measured or from models of such exposures. A circadian trajectory bundle may reflect the circadian trajectories that would result for a patient at 50 different levels of light sensitivity, i.e., for 50 trajectories. A circadian trajectory bundle may include one circadian trajectory for each combination of 50 different levels of light sensitivity and 10,000 light levels, i.e., a total of 500,000 circadian trajectories in the circadian trajectory bundle.

[0091] Trajectory bundles can include circadian trajectories or other biological trajectories. In the presence of missing data, trajectory bundles can be created by propagating forward with inputs drawn from a distribution from the last tracked circadian state and its associated uncertainty. The distribution can be uniform (e.g., uniform samples across all possible inputs) or informed by historical data (e.g., samples from the space of a "typical" day for this person, or samples from the space of likely chronotypes). Inputs can be wearable data inputs, such as hypothetical lighting or activity history. Inputs can be parameters that capture how the circadian system functions or is expected to function, such as higher or lower light sensitivity affecting the state. These can be averaged together to yield the most likely trajectory during the period of missing data.

[0092] In the presence of expensive optimization problems, such as the case of multiple interacting medications and inputs, trajectory bundles can be precomputed to save the patient time. For example, if a treatment decision is to be made or a message is to be communicated to the patient that needs to be made relatively soon after the patient makes a request (e.g., the patient presses a button on their wearable device and expects a message immediately thereafter), trajectory bundles can be precomputed with probabilistic future paths. Therefore, expensive optimization problems can be run over a set of trajectories in advance so that the data is accessible at the moment the patient makes the request. The updated trajectory can be compared to the precomputed trajectory, and if there is a good enough match, it is used. Otherwise, a relatively computationally expensive optimization can be performed.

[0093] Additionally, the trajectory bundle can be represented using the total trajectory density function, or equivalent coordinates for this density function that can be used to recover the probability density function, e.g., through using observable coordinates defined through Koopman operator theory, moments of probability distributions, Fourier coefficients, etc. The optimization problem under consideration can be solved more efficiently and / or accurately when framed under these coordinate transformations.

[0094] A patient care system may use digital twin modeling to generate a trajectory or inferred trajectory from input data. With digital twin modeling, the patient care system operates on a virtual model of the patient given several inputs. The inputs may be from body sensors, wearable devices, internal measurements, etc., which may have different degrees of accuracy and / or precision, and some inputs may have time gaps or missing data. The input data may also be physiological parameters such as light sensitivity. The input data may also be demographic data such as age. Digital twin modeling may include several simulations run to estimate what the actual trajectory may be under certain conditions. The input data may be pre-recorded or previously acquired data, or may be real-time data. Digital twin modeling may involve feedback of simulation results. Examples of hardware that may be used for digital twin modeling are described elsewhere herein.

[0095] A particular model for generating optimal times for taking medications by relating circadian time to times for taking medications may include raw model output of a core circadian pacemaker model connected to another model, such as a blood-brain barrier model and / or a liver model.

[0096] The state values ​​that result in a trajectory may have an associated uncertainty, and therefore the trajectory will have some uncertainty. The uncertainty may be in terms of the confidence level that the state at a given time is correct. High uncertainty may result when the data used to arrive at the state (e.g., wearable data used to predict circadian state) is sparse or highly irregular. Efficacy and toxicity profiles may look very different. An efficacy and / or toxicity profile identification system may identify an efficacy profile with a peak that is very different from a toxicity profile.

[0097] A computer-implemented method for administering a therapy, possibly under the control of one or more computer systems configured with executable instructions, can be provided for determining a set of patient inputs, determining a therapy, estimating a patient's circadian trajectory, determining one or more circadian mapping profiles, determining a preferred therapy duration from the circadian trajectory and one or more circadian mapping profiles, and then administering the therapy in response to an alert that the preferred therapy duration has occurred or will occur. The duration and trajectory can vary. The therapy can be administration of a substance, where the substance is one or more of a drug, nutrient, or medication. The set of patient inputs can include data derived from signals received from a patient wearing a wearable data system. The circadian trajectory can be derived by a scheduler using at least one biophysical model of the human circadian clock and at least one statistical model of the human circadian clock. The preferred therapy duration can be optimized based on relating circadian time to the time for taking the medication to generate raw model outputs.

[0098] The model output may be used by an environmental controller to adjust the environment such that circadian-related behaviors are adjusted toward target constrained times, which may correspond to treatments administered at an infusion clinic or scheduled surgery.

[0099] Figure 7 shows an example of a trajectory 700 that incorporates some uncertainty. The uncertainty for each state in the trajectory can be calculated in the trajectory bundle. In Figure 7, the uncertainty is indicated by the lightly shaded uncertainty region 702.

[0100] The digital twin simulation can be run for different inputs, each generating a different trajectory. The distribution of trajectories at a given time can be used to define the uncertainty for each circadian state. The different inputs given can be different wearable inputs (e.g., assuming the sensors on the wearable device are inaccurate or imperfect) or different model parameters (e.g., light sensitivity). The inputs can be selected randomly or systematically from a distribution. This distribution can be uniform, or it can be a distribution based on historical data from a single user or a population of users.

[0101] Uncertainty can result from a wide variance between trajectories. For example, if a patient care system has trajectories for several different input values, but lacks certainty about the input values ​​or across input values, the trajectories can be uncertain, as reflected in the plot shown in FIG. 7. For each circadian state, central or peripheral, uncertainty can reflect how reliable the estimate of that state is. For example, at time t x In , the concentration of a particular molecule can be between 3 and 4 picograms / ml, and the variance is x On the other hand, if the concentration is x If it were only known that the circadian pattern is between 3 and 300 picograms / ml at 2000 kcal, it would result in a higher uncertainty. Circadian uncertainty may increase in light of jet lag and travel, because the effect of jet lag / travel on circadian patterns for a particular individual may not be known. Uncertainty may increase in the presence of missing data and decrease in the presence of large amounts of input data.

[0102] Circadian mapping profiles that map variances in efficacy, toxicity, etc. to circadian states can be developed by collecting examples from a large number of patients and processing that data into circadian mapping profiles. These circadian mapping profiles can be stored as data structures processable by a patient care system and used along with trajectory data to make treatment timing decisions and / or recommendations.

[0103] 8 shows an example circadian mapping profile, or efficacy profile, that maps from circadian state to drug efficacy. The circadian mapping profile can be represented and stored in memory using various data structures that a processor can use to determine, for a given circadian state, what the efficacy value might be, or conversely, for a given efficacy value or range of values, what circadian state or range of states corresponds to that given efficacy value or range of values. In the example of FIG. 8, R is a value corresponding to how synchronized neurons in the SCN are, which in this example can range from R=0 to R=1, while Ψ is a value corresponding to a circadian state that can range from -π to +π and represent a state at a certain point in a circadian cycle or trajectory.

[0104] Efficacy is shown in a 10% range from 0% to 80%, with different shades of gray corresponding to specific ranges. Other ranges are possible. In this example of FIG. 8 , at point 802, corresponding to a circadian state of Ψ=+π / 3 and an SCN neuron synchronization value of R=0.5, efficacy would be between 10% and 20%. At another point 804, at a circadian state of Ψ=-2π / 3 and an SCN neuron synchronization value of R=0.84, efficacy would be between 40% and 50%. While FIG. 8 is a representation of what a circadian mapping profile stored in computer memory might look like, computer representations may have higher resolution, and the value represented in the circadian mapping profile for point 802 might be 14%. All other things being equal, this may indicate that treatment may benefit from being adjusted closer to Ψ=+π and R=0.98.

[0105] While Figure 8 shows a two-dimensional (2D) profile, a circadian mapping profile may have more than one dimension, with one or more of the dimensions being components of the circadian state or components of the central and peripheral circadian states. For example, an efficacy profile may have one dimension for the circadian state, which is the central circadian state of the expression level of the Bmal1 gene, and another dimension for drug efficacy for a given drug, the peripheral circadian state of the expression level of NHE3, the biological state of the concentration of a drug already in the system, etc. An N-dimensional circadian mapping profile stored in memory as a data structure may provide a processor with an indication of what efficacy of a drug is at a point or region in N-dimensional space corresponding to the current expression level of the Bmal1 gene, the current expression level of NHE3, the current concentration of a drug already in the system, etc.

[0106] FIG. 9 shows another example circadian mapping profile that maps circadian states to a drug's toxicity scale, which is a toxicity profile. The toxicity profile can be represented and stored in memory using various data structures that a processor can use to determine, for a given circadian state, what the toxicity value might be, or conversely, for a given toxicity value or range of values, what circadian state or range of states corresponds to that given toxicity value or range of values. This can be used by a patient management system to prompt a patient to begin treatment, such as taking a prescribed medication. As shown in FIGS. 8 and 9 , some circadian states may be more favorable than others for initiating treatment, and the methods described herein for inferring current or future circadian states may be useful.

[0107] In the example of FIG. 9, similar to FIG. 8, R is a value corresponding to how synchronized neurons in the SCN are, and Ψ is a value corresponding to the circadian state. For various values ​​of these two axes, a range of adverse events is shown, which may represent an indication of toxicity. The circadian mapping profile can be represented and stored in memory using various data structures that a processor can use to determine what the adverse event count value may be for a given circadian state, or conversely, which circadian state or range of states corresponds to a given adverse event count value or range of values. In the example of FIG. 9, R can range from R=0 to R=1, while the circadian state Ψ can range from -π to +π, representing a state at a certain point in a circadian cycle or trajectory.

[0108] The number of adverse events is shown in five-event ranges from 0 events to 40 events, with different shades of gray corresponding to specific ranges. Other ranges are possible. In this example of FIG. 9 , at point 902, corresponding to a circadian state of Ψ=+π / 3 and an SCN neuron synchronization value of R=0.5, the number of adverse events would be shown in the circadian mapping profile as being between 10 and 15. At another point 904, corresponding to a circadian state of Ψ=-2π / 3 and an SCN neuron synchronization value of R=0.84, the number of adverse events would be shown in the circadian mapping profile as being between 30 and 35. All other things being equal, this may indicate that treatment may benefit from being adjusted closer to Ψ=+π / 3 and R=0.95. While FIG. 9 is a representation of what a circadian mapping profile stored in computer memory might look like, a computer representation may have higher resolution, and the value represented in the circadian mapping profile for point 904 may be 32.

[0109] In some implementations, the selection of Ψ and R may be optimized taking into account multiple circadian mapping profiles, such as one or more efficacy profiles and one or more toxicity profiles, to find the best fit when the optimal regions of each circadian mapping profile are not all at the same values ​​of Ψ and R.

[0110] While Figure 9 shows a two-dimensional (2D) profile, a circadian mapping profile may have more than one dimension, with one or more of the dimensions being circadian components of circadian state or components of central and peripheral circadian state. For example, a toxicity profile may have one dimension for circadian state, which is the central circadian state of the expression level of the Bmal1 gene, and another dimension for drug toxicity for a given drug, the peripheral circadian state of the expression level of NHE3, the biological state of the concentration of a drug already in the system, etc. An N-dimensional profile stored in memory as a data structure may provide a processor with an indication of what toxicity of a drug is at a point or region in N-dimensional space corresponding to the current expression level of the Bmal1 gene, the current expression level of NHE3, the current concentration of a drug already in the system, etc.

[0111] As described elsewhere herein, the patient care system may convert multiple circadian trajectories and clinical outcome data into efficacy and / or toxicity profiles. The process for generating such a circadian mapping profile may take demographic data and other trajectories as inputs and output a circadian mapping profile that maps circadian states (and possibly other inputs) to efficacy, toxicity, or other profile-type values.

[0112] Therapy mapping can be represented in a data structure that provides rules for converting a circadian trajectory and circadian mapping profile with associated uncertainty into a set of dosing windows. The dosing window can be narrower when uncertainty is low (e.g., "take your medication between 5:00 PM and 5:30 PM"). The dosing window can be wider when uncertainty is high (e.g., "take your medication in the afternoon"). Therapy mapping can be calculated from the patient's circadian trajectory, which can be further derived from wearable data. For example, wearable data can be used to arrive at both the circadian trajectory and uncertainty point-by-point by simulating a digital twin of the suprachiasmatic nucleus (SCN). In this example, the circadian state can be represented by the firing rate and firing cohesion of neurons in the SCN or by the expression of genes in SCN cells at a given moment.

[0113] When a treatment outside of the user's control is scheduled, there may be a fixed treatment window for a certain period of time. This could be surgery or an infusion. A fixed treatment window may be addressed by initiating a control that can result in a change in the mapping from circadian state to elapsed real time. The control may be a behavior or activity that affects a biological state. The control may be represented in memory as a data structure that includes an action and a duration. The control may be a pulse of light at a specific time to change the light exposure history or a suggestion to do so (e.g., "Receive 10,000 lux light exposure for 45 minutes starting at 2:45 PM on October 3rd"), a signal to suggest exercise during a certain time window (e.g., "Do a moderate workout from 11:45 AM to 12:45 PM"), a signal to take medication during a certain time window, etc.

[0114] The control system may take into account the current circadian state with associated uncertainty and the fixed therapeutic window and prescribe a set of controls to move the recommended dosing window corresponding to the future circadian trajectory and efficacy and / or toxicity profile so that it overlaps with the fixed therapeutic window as much as possible.

[0115] Accurately determining an individual patient's circadian time so that it can be used in treatment mapping may involve wearable data augmented with data structures such as a study calendar data table. The inclusion of this data structure may reduce uncertainty.

[0116] Digital twin simulations for converting inputs into trajectories can address the problem of masking caused by external factors that obscure the true circadian signal to enable accurate and personalized treatment timing recommendations. For example, digital twin simulations can address the problem of noise and outliers in the signal that confound the estimation of true circadian time by imposing realistic limits on how much influence a signal of any duration can have on the output circadian state. Without addressing masking, administration window times can become highly volatile and imprecise.

[0117] Dosing windows may be presented as simple timing recommendations for dosing a fixed amount of medication, such as a tablet (e.g., "There is a good time period to take this tablet, and in your case specifically, that time period is tomorrow between approximately 10 AM and 11 AM local time"), or they can be presented as both dosing and timing recommendations (e.g., "Your morning doses should be X mg between approximately 10 AM and 11 AM local time, and X+Y mg between approximately 12 PM and 1 PM local time").

[0118] Inputs to the digital twin simulation used to arrive at the trajectory may be wearable device data streams, demographic data stored in user data records, or demographic details (e.g., gender, age, geographic residential location, etc.).

[0119] Inputs to the circadian mapping profile can include known optimal timing for the individual based on self-report or other records, a set of timing constraints represented in a constraint table or a stored constraint rule set (e.g., constraints on time intervals between medications), and the like.

[0120] A user interface may be provided to notify the patient of the administration window and alert them to the elapsed real-time corresponding to the administration window.

[0121] The digital twin simulation may also incorporate environmental factor inputs and outputs and use them to modify the trajectory. This may further change the inputs to the treatment mapping, resulting in new administration windows. For example, a patient room may contain known lighting levels in the digital twin simulation, reducing the uncertainty about the trajectory and narrowing the administration window presented to the user.

[0122] Environmental controls can also be integrated with a control system. For example, the control system may interact with the environmental controls to alter the environment based on a set of rules or data that shifts future administration windows to overlap fixed treatment times as much as possible. As an example, the environmental control system may change light levels, light colors, heating, cooling, etc. in a manner that alters the patient's circadian trajectory such that the resulting treatment mapping recommends administration windows near or overlapping fixed treatment times.

[0123] 10-12 show examples of hardware configurations that may be used to implement the patient care systems described herein.

[0124] The treatment system can provide a user interface to collect input, apply treatment mapping, and show the calculated administration window, possibly at or near a specific elapsed real-time. The administration window may initially cover a long time window (e.g., "all afternoon") and then become more specific as uncertainty decreases ("between 8 and 8:30 PM").

[0125] As described herein, the treatment system can incorporate patient and other inputs and process a treatment mapping that incorporates a biological state and at least one circadian state along with a circadian mapping profile, and generate a dosing window that corresponds to elapsed real time. A display of the dosing window can be triggered near the beginning of the dosing window, and this display can be provided through a user interface that will issue a treatment dosing message, such as a reminder to administer the corresponding treatment, to the patient or caregiver.

[0126] FIG. 10 is a diagram of a patient treatment system 1000 according to various embodiments. The patient treatment system 1000 can be used to generate optimal administration windows related to biological processes and provide patient information related to treatment and timing of the treatment. In one example of treatment mapping to determine the best timing, the best time to take a medication may be 7 hours after a circadian state corresponding to the onset of melatonin secretion. In another example, treatment mapping can include a complex molecular model with a formula that captures how different molecules in a medication bind to the body and interact with each other to select the best administration window given these drug interactions. In another example, treatment mapping can take demographic information such as race, gender, age, and other medications as input parameters. Logical gating can be used; in one example, treatment mapping ensures that medications are spaced at least 24 hours apart. In a further example, where molecules bind and interact in a formula, the variables represent the molecular formula and the parameters represent the binding rate. In many cases, indicating to a patient the optimal time to take a medication or perform another treatment may depend on determining their circadian state and possibly uncertainty about the patient's circadian state as well. Similar to the example above, if the optimal time to take a medication is 7 hours after the patient has experienced melatonin onset, the patient system 1002 of the patient treatment system 1000 may send a message around the time determined to be 7 hours after melatonin onset. If there is uncertainty as to when that will be, the patient may be notified of the message around the beginning of the uncertainty window and provided with an indication of the uncertainty (e.g., "The best time to take medication A is, for you, 7 hours after melatonin onset. Based on analysis of data measured from you and data you entered, your melatonin onset today was 11:25 AM with an uncertainty of approximately + / - 35 minutes").Thus, at the indicated dosing, Drug A should be taken at 6:25 PM, and given the uncertainty, the actual optimal time to take Drug A may be between 5:50 PM and 7:00 PM."

[0127] 10 , a user / patient 1006 may provide user data 1010 to a user interface 1008, which is stored in a user data storage 1012. The user data 1010 may include data received from the user 1006 via the user interface 1008, such as user demographics, user work information, user mobility information, indications of other medications or treatments, user emotions (how the user is feeling, how the user responded to past treatments, etc.), and the like, as well as self-reported or manually entered data, such as historical meal times, light exposure, exercise time, or a list of medications the patient is currently taking. Other user-related data may be acquired by a wearable sensor 1020 that converts sensor signals 1022 into digitized sensor data 1024 that is provided to the user data storage 1012. The user data may be provided to a therapy processing unit 1030. The digitized sensor data 1024 may include step counts, heart rate, temperature, etc.

[0128] The therapy processing unit 1030 may include a processor 1032 and program code / logic 1034 that may be used to incorporate user-specific data to determine the user's current circadian state, as well as state interaction information and data regarding efficacy and / or toxicity profiles from molecular stimulation of various medications to determine an optimal or preferred circadian state for administering therapy. The data provided to the therapy processing unit 1030 may be used to modify statistics 1040 and / or biophysical models 1042 to track circadian state. User details, such as a list of medications the patient reports taking, may be used to determine which circadian mapping profile to use from a set of profiles.

[0129] The user's circadian state may be passed as a circadian mapping profile for the patient, and the administration window may then be passed as output to the user. For example, as shown in FIG. 10 , the therapy processing unit 1030 may receive data 1046 related to the efficacy, toxicity, other drug facts, etc. of a particular treatment, such as the administration of "Drug A," from a drug dataset 1044. Taking into account user-specific data about the user / patient 1006, the therapy processing unit 1030 may output a record 1052 representing a relative time optimization, which would indicate the best time or time range for treatment for the user / patient 1006 expressed in circadian relative time. For example, the best time may be expressed as the number of hours before or after a defined circadian state, such as the circadian state corresponding to the onset of melatonin secretion. The model converter 1054 may convert the data in the records 1052 into wall clock optimized data records 1056, which may then be provided to the user interface 1008 for presenting a message or signal to the user / patient 1006 indicating the best, preferred, and / or optimal elapsed real time for administering a treatment, such as taking a medication.

[0130] As described in detail herein, a circadian cycle may pass through various circadian states, and the pattern of circadian states may be referred to as a circadian trajectory. A circadian trajectory may correspond to a particular physical state of a person's body (or perhaps across a population of people or other organisms that exhibit circadian cycles). While circadian states can be commonly represented as simple sine waves with frequencies on the order of a day or slightly longer, and thus thought of as corresponding to the phases of a simple sine wave, circadian states and circadian trajectories are often not so simply modeled. To account for this, the devices and methods described herein can be used to predict or estimate an organism's circadian trajectory and a mapping of circadian states along that circadian trajectory to elapsed real time. This can be useful for making treatment timing decisions where the optimal circadian state along the circadian trajectory for administering the treatment is known, where the user's / patient's current circadian state and / or future elapsed real time corresponding to the user's / patient's future circadian state need to be determined.

[0131] Circadian trajectories may be reflected in the concentrations of molecules that rise and fall with circadian patterns. These trajectories may be referred to as "peripheral circadian trajectories" if they occur in peripheral clocks, such as biological clock mechanisms associated with the stomach or liver. These trajectories may be referred to as "central circadian trajectories" if they occur in central clocks, i.e., the SCN. These trajectories, or estimates thereof, may be generated using digital twin simulations, as they are often not easily determined from direct measurements in living organisms. In one example, a digital twin simulation of a circadian trajectory is generated by a processor according to differential equations that describe the concentrations of molecules, allowing the concentrations to rise and fall dynamically, physiologically.

[0132] The concentration of molecules outside the body can be described by trajectories. These trajectories can be generated by digital twin simulations. In one example, a digital twin simulation of external molecular trajectories is generated by a processor according to a differential equation that describes the concentration of the molecule, allowing the concentration to rise and fall dynamically, physiologically. For example, the stimulation of caffeine on a system can be captured by a digital twin simulation in which caffeine is slowly eliminated from the bloodstream at a rate expressed as a differential equation.

[0133] Information provided to therapeutic mapping can include data about pill consumption, such as automatically tracking pills as they are retrieved, or detecting wrist movements via an accelerometer that are predicted to match the wrist movements when taking a pill. In this example, therapeutic mapping can use pill timing information to arrive at a more accurate trajectory of pill concentrations in the body. It can also use this information to apply logical gating rules, such as the interval between the intake of two consecutive doses of medication.

[0134] In yet another example, a treatment allocation and coordination system can be used to assist in clinic or hospital scheduling. In a treatment allocation and coordination system, treatment mapping is iteratively applied across multiple individuals to help the clinic or hospital schedule schedule people's appointments (infusions, surgeries) to arrive at recommendations for fixed treatment windows (e.g., clinical or surgical time slots for appointments) that best align with the individuals' treatment windows. For example, treatment mapping can be applied to determine treatment windows for all people in a clinic, and those with earlier treatment windows can be assigned to earlier fixed treatment windows. The joint raw treatment mapping output is converted into a human-interpretable representation of the best times for surgeries or clinic visits within the available time slots.

[0135] FIG. 11 is a diagram of a circadian mapping profile identification system. As shown therein, circadian mapping profiles, such as efficacy and / or toxicity profiles, are generated from circadian trajectories and outcome data from multiple individuals, who either self-report such data or have such data automatically detected. As shown in FIG. 11, there may be several patient systems 1102 for multiple patients. Each patient system 1102(i) may provide the timing profile determination module 1104 with the circadian trajectory and outcome data maintained for patient i, such as patient-specific data 1106 from each of the patient systems 1102. The timing profile determination module 1104 can then provide the medication profile data records 1110 to the treatment system 1108. An example of patient-specific data might be, "Adverse events for patient ABC12 included incident IN0045371, in which the patient took a dose at 7:34 PM, the patient's circadian state at the time of dosing was 0.243813 (along a trajectory scaled from 0.0 to 1.0), and the patient self-reported two headaches the next day." For example, if many patients reported next-day headaches and many did not, and the salient differentiator between these two groups is that those who took the dose in a circadian state between 0.2 and 0.3 at the time of dosing are the first group (the group with adverse effects), and those who took the dose in a circadian state between 0.6 and 0.7 at the time of dosing are the second group (the group not reporting adverse effects), then the drug profile data record 1110 might be expected to indicate that, for drug A, a circadian state of 0.6 to 0.7 is preferred for administering drug A.

[0136] The patient-specific data 1106 passed to the timing profile determination module 1104 can be circadian state data over time, as well as efficacy and / or toxicity outcomes. These can include efficacy outcomes such as remission rate, tumor shrinkage, overall survival rate, or changes in blood pressure, lipid profile, or other biological outcomes. They can also be toxicity outcomes, such as the number of adverse events or side effects reported. The timing profile determination module 1104 can aggregate data from multiple patients to arrive at a tool for translating circadian state and other inputs into efficacy and toxicity.

[0137] The cumulative data generates efficacy and / or toxicity profiles that incorporate circadian state and other possible inputs to efficacy and / or toxicity. For example, a circadian mapping profile identification system may identify that the reason an 8 AM dose reduces blood pressure for one person but has no effect for a different person is because these two individuals were in different circadian states at 8 AM in elapsed real-time. The system then learns the relationship between circadian state and efficacy obtained from the calculated trajectories, and in doing so, may reveal trends that are not obvious from the elapsed real-time of administration. For example, the system may generate an efficacy profile that converts circadian state (such as peak Bmal1 expression) to a percent efficacy (80% of peak efficacy at peak Bmal1 expression).

[0138] This circadian mapping profile identification system can identify sex differences and include them as optional inputs in the circadian mapping profile. For example, peak Bmal1 expression can correspond to 80% efficacy for males and 70% efficacy for females. It can also identify age-related effects and include them as optional inputs in the resulting circadian mapping profile. It can also identify demographic effects and include them as optional inputs in the resulting circadian mapping profile.

[0139] The circadian mapping profile identification system can identify dietary interaction effects and include them as optional inputs in the circadian mapping profile. The circadian mapping profile identification system can build a circadian mapping profile from self-reported data or from signals extracted directly from the user. For example, the system can identify circadian conditions associated with relatively low blood pressure without manual user input. The circadian mapping profile identification system can also be used to extract the best circadian timing for a treatment to reduce self-reported side effects of the treatment.

[0140] In addition to or instead of administering a therapy, the treatment system can include a control system, and a trigger from the control system can be an indicator of control. Control can be an action that a patient takes, such as performing an exercise routine at a specific time to help the individual move their administration window to overlap with a fixed treatment window. Control can also be used to reduce uncertainty, which can narrow the width of the administration window.

[0141] Efficacy profiles can be generalized to include non-drug and non-medical therapeutic activities. For example, efficacy profiles can relate exercise in different circadian states to its effectiveness in reducing diabetes risk.

[0142] The dosing window can be generalized to include non-drug and non-medical therapeutic activities. For example, the dosing window can recommend exercise, meal timing, and structure at specific times to target reducing diabetes risk according to an efficacy profile. Different efficacy profiles can result in dosing windows for exercise at different times, for example, if the goal is to improve high blood pressure.

[0143] The dosing window can also be generalized to include vaccine timing and other injections.

[0144] The dosing window may also be applied to recommended timing for infertility treatments such as in vitro fertilization (IVF).

[0145] The dosing window can also be generalized to include times to eat or avoid meals, as well as the composition and / or size of meals. For example, for glycemic control, the dosing window can provide a meal timing window and the protein, lipid, and carbohydrate composition of a meal.

[0146] The dosing window can be long, e.g., 6-12 hours in duration. In such cases, the specific timing of the dosing window may be most critical for shift workers, who have poorly defined concepts such as biological day and night, those who have recently traveled, and others who have recently experienced a significant shift in their circadian trajectory.

[0147] The application of this system is not limited to determining the timing of pill consumption. The system can be used to identify the best fixed treatment window for a user, for example for an infusion, or the best time for surgery. The system can instruct a person on the best time to go to the clinic from all times the clinic is available or from among the available appointment times at the clinic.

[0148] The system can inform people of the best time to schedule surgery. Alternatively, the system can inform people about which controls leading up to surgery would be best for them to use to allow for better recovery time from surgery. These controls can be actions the user should take. For example, if a patient's toxicity profile indicates that surgery would be easier if performed during a circadian state that typically occurs at night for that patient, but surgery is scheduled for daytime, the control system can recommend controls to shift the person's circadian trajectory to shift their circadian state earlier or later to better overlap with a fixed treatment window during daytime hours.

[0149] In another example, the treatment system can be coupled with digital therapeutics in medicine. The treatment system can be integrated into a feedback loop system where sensors built into medication or pill bottles can track medication consumption. The sensors can provide information about the actual time elapsed since the patient took the medication. This information can be provided to treatment mapping to inform logical gating rules for spacing medication consumption and modifying recommended dosing windows. This information can also be used to inform trajectories calculated using digital twin simulations that represent how much medication is in the body over time. These trajectories can then be used as inputs to treatment mapping to perform more advanced determination of dosing windows.

[0150] In yet another example, the system serves as a tool for pharmaceutical companies to identify circadian mapping profiles for their assets. Drug timing information can be retrieved directly from drug or patient self-reported timing in combination with a circadian mapping profile identification system such as that described herein to generate circadian mapping profiles for new assets, thereby enabling more new drugs to emerge from early-stage drug trials. These circadian mapping profiles will map circadian state, and potentially other inputs such as biological state or demographic information, to efficacy and / or toxicity.

[0151] 12 is a diagram of a control system 1200 according to various embodiments. In one example, the control system 1200 can be used to adjust a person's circadian trajectory through an environmental controller, such as light exposure via an LED device or behavioral prompts via a GUI, so that the person's recommended dosing window overlaps as much as possible with a fixed treatment window. The control system 1200 can include a patient system 1202 that provides patient-specific data 1214 to a feedback controller 1210. The feedback controller 1210 can retrieve constraint data 1212 from a treatment schedule constraint table 1216 to determine what timing constraints may exist for the treatment. The patient-specific data 1214 can include the patient's current circadian trajectory and condition, the patient's current prescription, etc.

[0152] A treatment schedule constraint table 1216 may be used to provide a list of constraint data 1212, such as times for light exposure and other environmental controls for a given efficacy / toxicity profile and circadian state, as well as when a particular procedure must be performed. The patient system 1202 may provide the predicted circadian state. The feedback controller 1210 can then combine the light exposure commands and the circadian state from the patient system 1202 and trigger changes in the environmental controls to move the patient's circadian state to better align with that prescribed by the treatment schedule constraint schedule. The feedback controller 1210 may provide control data 1208 to the environmental controller system 1206, which can control the environment to affect the circadian trajectory.

[0153] The control system 1200 can identify information such as a phase response curve, in which "phase advance" and "phase delay" regions are identified during which light pushes the clock forward or backward. The control system 1200 can identify the phase advance and phase delay regions for an individual and provide light during periods when the recommended administration window better aligns with fixed treatment times. The control data 1208 can be specific instructions such as "Turn on the lights at 4:45 AM" to shift the patient to an earlier part of their circadian trajectory.

[0154] The timing system can be integrated into a smart drug delivery system, for example, into a time-locked pill system that can automatically release drugs according to a dosing window generated by treatment mapping.

[0155] The timing system can be integrated into the artificial organ. The therapeutic system can identify a dosing window for optimal compound release from the artificial organ system to best match the function of the organ it is replacing. In this case, the efficacy profile is a circadian-mapped efficacy profile that best matches the original function of the organ (e.g., release in this circadian state has high efficacy because it best resembles the original organ).

[0156] The timing system can be integrated into an implantable pharmacy, which can release the drug according to a dosing window set by the treatment system.

[0157] The control system can be integrated into an implantable formulation that can release drugs to shift the circadian trajectory such that the administration window shifts to more overlap with a fixed treatment time. Example Components

[0158] FIG. 13 illustrates components that may be used in a control system 1300 to vary environmental cues and behavioral triggers in response to fixed treatment times and user data, according to various embodiments. FIG. 13 is a simplified diagram in which boxes may represent data structures, computational elements, sensors, and / or users. In some embodiments, elements shown that are similar to elements in FIG. 10 may operate similarly. As shown in FIG. 13, the control system 1300 may include a patient feedback system 1302 to provide output to adjust the trajectory to meet certain treatment schedule constraints.

[0159] The patient 1306 may have several treatment timing constraints, which may be represented in a dataset 1356 coupled to the feedback controller 1354. Input from the patient 1306, such as wearable sensor data and user-provided data, possibly provided via a user interface 1308 and stored in a user data storage 1312, may be provided to the therapy processing unit 1330. The sensor signal data may come from a wearable sensor device and / or environmental sensors on the patient 1306. The environmental sensors may be used to detect the lighting environment, the temperature environment, or other environmental details experienced by the user. The therapy processing unit 1330 may use the digitized sensor data along with a circadian mapping profile to design a series of environmental shifts that move the patient's circadian state so that the administration window better aligns with the fixed scheduled treatment time. The therapy processing unit 1330 may derive the circadian state from a machine learning subsystem.

[0160] The machine learning subsystem may include a processor, storage 1340 for a statistical model, and storage 1342 for a biophysical model, both trained by the processor using training data corresponding to the sensor data. The statistical model may be a machine learning model trained on wearable time series of data streams such as light, activity, temperature, and heart rate to predict gold-standard biomarkers of circadian rhythms, such as dim light melatonin onset time, melatonin concentration over time, or core body temperature. The biophysical model may be human-generated or data-derived and may be created to match the physical properties of the suprachiasmatic nucleus; for example, predicting firing rates in the ventral and dorsal SCN at any given time, which may then be mapped to an output such as the timing of dim light melatonin onset.

[0161] The therapy processing unit 1330 may output a circadian timing model 1352. The statistical model, biophysical model, and circadian mapping profile may be constructed as neural network weights or other data structures, possibly usable as trained machine learning models, and may be accessible to the therapy processing unit 1330 in computer memory. The feedback controller 1354 uses the circadian mapping profile as input and may be provided with a dataset 1356 indicative of scheduled treatment times. One operation of the patient feedback system 1302 is for the feedback controller 1354 to output signals and / or data corresponding to environmental outputs, behavioral reminders, and the like, that are adjusted to adjust the user's circadian state to better align with the scheduled treatment times. Signals may be output to an environmental settings controller, which may control room lighting by sending instructions to a room lighting and environment controller 1358. The feedback controller 1354 may send commands to any other environmental system that may control some aspect of the patient's environment.

[0162] For example, the control system may adjust room lighting to attempt to steer the user's circadian state toward alignment with the scheduled treatment time. The control system may also trigger the display or issuance of action reminders 1360 to the user. For example, the control system may cause the user's handheld device (e.g., via the user interface 1308) to display a message such as, "Your surgery is scheduled for tomorrow at 2:30 PM. Outcomes for this surgery are statistically improved for patients who have this surgery in the middle of the morning, so you should stay active later than usual tonight before going to bed. We will control the lighting to help you do this," or "You have an outpatient procedure tomorrow at 9:30 AM, including taking some medication two hours before the procedure. You should go to bed now so you can wake up around 6:30 AM, because better outcomes tend to occur when your body recognizes that an hour has passed since daytime awakening." In the latter example, the control system may cause the lighting to be altered at night to reduce the amount of blue light to help align the wall clock's nighttime hours with the nighttime portion of the user's circadian trajectory.

[0163] 14 illustrates components that may be used for a digital twin simulation, according to various embodiments. In the example digital twin simulation, a biophysical model representing a system of differential equations provides input to a machine learning model, which in turn produces a trajectory.

[0164] In the illustrated example, simulator 1400 obtains data structures corresponding to sensor ratings, activity, and / or user data. For example, the simulator may read data from the user's wearable device, read demographic data from a user database, read patterns of caffeine intake (perhaps based on the user's manual entry of caffeine intake, coordination with a purchasing app, coordination with a mapping app, or other sources), and then apply that data to a set of differential equations to generate a dataset with reduced dimensionality. The simulator may then use that dataset and the trained machine learning model to generate a circadian trajectory for the digital twin, and other portions of the control system may use the generated circadian trajectory as a proxy for the user's circadian trajectory.

[0165] 14, the distiller 1406 can receive the wearable data 1408, the demographic data 1412, and the caffeine intake record 1414 and distill it into data that can be provided to the model generator 1410. The model generator 1410 can then generate an ML model 1416, which can be provided to the trajectory generator 1420. The trajectory generator 1420 can then output and store a stored trajectory representation 1422.

[0166] FIG. 15 illustrates components that may be used for a profile identification system 1500, according to various embodiments. As shown there, a collection of wearable data from a large population, such as 1,000 people, 10,000 people, or more, may be provided to a digital twin simulator, which then generates a circadian trajectory for each member of the population. The profile identification system 1500 may read data from a database of self-reported side effects reported by members of the population. The digital twin simulation may include statistical models, biophysical models, or both to process the wearable data and produce a trajectory. Taking into account both the simulated trajectories and the reported side effects, the profile education system may be capable of outputting data indicating which circadian states are favorable for reducing side effects and which are not. The output data may be represented as a toxicity profile.

[0167] 16 illustrates components that may be used for a treatment mapping module 1600, according to various embodiments. In this example, treatment mapping module 1600 includes an optimization system that can load a data structure representing a circadian trajectory, which may have some uncertainty, such as that shown in FIG. 4, another data structure representing an efficacy profile dataset, and another dataset including a log of prior medication administration. Using that data, treatment mapping module 1600 may generate a dataset or transmit data indicating a preferred administration window. Another input to the optimization system is treatment mapping data.

[0168] 17 illustrates components that may be used for a treatment system module 1700, according to various embodiments. As shown therein, the treatment system module 1700 may provide functionality similar to that of the treatment mapping module 1600 in FIG.

[0169] 18 illustrates components that may be used for a control system 1800, according to various embodiments. As shown there, control system 1800, such as may be used as control system 1300 shown in FIG. 13, can take in a dataset representing a circadian state with uncertainty and a dataset representing a treatment time, and can output a dataset representing a list of controls, such as ambient light or behavioral changes, that may be selected to change a user's circadian state to align its best administration window with a fixed treatment time.

[0170] FIG. 19 illustrates components that may be used for a treatment allocation and coordination system 1900, according to various embodiments. In it, circadian state data for a number of people is mapped to a schedule that may assign each person to a fixed treatment time that best overlaps with their administration window, subject to constraints that assign each person to a treatment window and other desired parameters, such as the person's availability and timing preferences. The treatment allocation and coordination system 1900 may incorporate a dataset of circadian states or trajectories across a patient population (possibly with some uncertainty) of patients needing a particular treatment in a clinic with limited available appointments. The treatment allocation and coordination system 1900 may incorporate a dataset of available appointments. From that information, the treatment allocation and coordination system 1900 may generate a list of appointment assignments. Appointments may be assigned to patients to maximize or increase the overlap or alignment of treatment times and circadian state patients in the population based on what is known about the preferred mapping of treatments to circadian states.

[0171] FIG. 20 illustrates another example of trajectory uncertainty, according to various embodiments. In this specific example, plot 2000 shows a baseline estimate 2002 (thick line) of circadian state at any given time point with uncertainty 2004 (light shading). In this example, the circadian trajectory has a strong sinusoidal component and a large amount of uncertainty over the first four days. In this example, travel across several time zones is assumed prior to day 0 and may be a source of uncertainty. However, after several days, e.g., 7.5 days, the uncertainty significantly decreases. Various systems described herein can take into account how uncertainty can increase during periods of significant circadian discontinuity while decreasing during periods of circadian stability. In this example, the uncertainty in circadian state due to jet lag becomes more consistent over time as the person remains in one time zone. Hardware Components

[0172] FIG. 21 is a simplified functional block diagram of a storage device 2148 having applications that can be accessed and executed by a processor in a computer system, such as may be part of an embodiment of a patient treatment system and / or computer system that performs calculations required for treatment optimization. FIG. 21 also shows an example of memory elements that may be used by a processor to implement elements of embodiments described herein. In some embodiments, data structures are used by various components and tools, some of which are described in more detail herein. The data structures and program code used to operate on the data structures may be provided and / or carried by a transitory computer-readable medium, e.g., a transmission medium, such as in the form of a signal transmitted over a network. For example, when a functional block is referenced, it may be implemented as program code stored in memory. An application may be one or more of the applications described herein running on a server, client, or other platform or device, and may represent the memory of one of the clients and / or servers shown elsewhere.

[0173] Storage device(s) 2148 may be one or more memory devices that may be accessed by a processor, and storage device(s) 2148 may store application code 2150, which may be configured to store one or more processor-readable instructions in the form of write-only and / or writable memory. Application code 2150 may include application logic 2152, library functions 2154, and file I / O functions 2156 associated with the application. The memory elements of Figure 21 may be used for a server or computer that interfaces with a user, generates data, and / or manages other aspects of the processes described herein.

[0174] The storage device 2148 may also include application variables 2162, which may include one or more storage locations configured to receive input variables 2164. The application variables 2162 may include variables generated by the application or otherwise local to the application. The application variables 2162 may be generated, for example, from data retrieved from a user or an external source, such as an external device or application. The processor may execute the application code 2150 to generate the application variables 2162, which are provided to the storage device 2148. The application variables 2162 may include operational details required to perform the functions described herein.

[0175] The storage device 2148 can include storage for databases and other data described herein. One or more memory locations can be configured to store device data 2166. The device data 2166 can include data provided by a user or an external source, such as an external device. The device data 2166 can include, for example, records passed between servers before transmission or after reception. Other data 2168 can also be provided.

[0176] The storage device 2148 may also include a log file 2180 having one or more storage locations 2184 configured to store results of or inputs provided to an application. For example, the log file 2180 may be configured to store a history of actions, alerts, error messages, and the like.

[0177] According to some embodiments, the techniques described herein are implemented by one or more generalized computing systems programmed to execute the techniques according to program instructions in firmware, memory, other storage, or a combination thereof. Specialized computing devices such as desktop computer systems, portable computer systems, handheld devices, networking devices, or any other device incorporating hardwired and / or program logic that implements the techniques may be used.

[0178] An embodiment may include a carrier medium carrying data, including data processed by the methods described herein. The carrier medium may include any medium suitable for carrying data, including a storage medium, e.g., a solid-state memory, an optical or magnetic disk, or a transitory medium, e.g., a signal carrying data, such as a signal transmitted over a network, a digital signal, a radio frequency signal, an acoustic signal, an optical signal, or an electrical signal.

[0179] Figure 22 is a block diagram illustrating a computer system 2200 on which the computer systems described herein and / or the data structures shown in Figure 21 may be implemented. Computer system 2200 includes a bus 2202 or other communication mechanism for communicating information, and a processor 2204 coupled with bus 2202 for processing information. Processor 2204 may be, for example, a general-purpose microprocessor.

[0180] Computer system 2200 also includes a main memory 2206, such as a random-access memory (RAM) or other dynamic storage device, coupled to bus 2202 for storing information and instructions executed by processor 2204. Main memory 2206 may be used for storing temporary variables or other intermediate information during execution of instructions to be executed by processor 2204. Such instructions, when stored on non-transitory storage media accessible to processor 2204, render computer system 2200 a special-purpose machine customized to perform the operations specified in the instructions.

[0181] Computer system 2200 further includes a read only memory (ROM) 2208 or other static storage device coupled to bus 2202 for storing static information and instructions for processor 2204. A storage device 2210, such as a magnetic disk or optical disk, is provided and coupled to bus 2202 for storing information and instructions.

[0182] Computer system 2200 may be coupled via bus 2202 to a display 2212, such as a computer monitor, for displaying information to a computer user. An input device 2214, including alphanumeric and other keys, is coupled to bus 2202 for communicating information and command selections to processor 2204. Another type of user input device is a cursor control 2216, such as a mouse, trackball, or cursor direction keys, for communicating directional information and command selections to processor 2204 and for controlling cursor movement on display 2212. This input device typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), that allow the device to specify a position in a plane.

[0183] Computer system 2200 may implement the techniques described herein using customized hardwired logic, one or more ASICs or FPGAs, firmware, and / or program logic that, in combination with the computer system, causes or programs computer system 2200 to be a special-purpose machine. According to one embodiment, the techniques herein are performed by computer system 2200 in response to processor 2204 executing one or more sequences of one or more instructions stored in main memory 2206. Such instructions may be read into main memory 2206 from another storage medium, such as storage device 2210. Execution of the sequences of instructions stored in main memory 2206 causes processor 2204 to perform the process steps described herein. In alternative embodiments, hardwired circuitry may be used in place of or in combination with software instructions.

[0184] The term "storage medium," as used herein, refers to any non-transitory medium that stores data and / or instructions that cause a machine to operate in a specific fashion. Such storage media may include non-volatile media and / or volatile media. Non-volatile media include, for example, optical or magnetic disks, such as storage device(s) 2210. Volatile media include dynamic memory, such as main memory 2206. Common forms of storage media include, for example, floppy disks, flexible disks, hard disks, solid-state drives, magnetic tape, or any other magnetic data storage medium, CD-ROMs, any other optical data storage medium, any physical medium with a pattern of holes, RAM, PROM, EPROM, FLASH®-EPROM, NVRAM, or any other memory chip or cartridge.

[0185] Storage media are distinct from but may be used in conjunction with transmission media. Transmission media involves transferring information between storage media. For example, transmission media include coaxial cables, copper wire and fiber optics, including the wires that comprise bus 2202. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

[0186] Various forms of media may be involved in carrying one or more sequences of one or more instructions to processor 2204 for execution. For example, the instructions may initially be carried on a magnetic disk or solid state drive of a remote computer. The remote computer may load the instructions into its dynamic memory and send the instructions over a network connection. A modem or network interface local to computer system 2200 may receive the data. Bus 2202 carries the data to main memory 2206, from which processor 2204 retrieves and executes the instructions. The instructions received by main memory 2206 may optionally be stored on storage device 2210 either before or after execution by processor 2204.

[0187] Computer system 2200 also includes a communication interface 2218 coupled to bus 2202. The communication interface 2218 provides a two-way data communication coupling to a network link 2220 that is connected to a local network 2222. For example, communication interface 2218 may be a network card, modem, cable modem, or satellite modem to provide a data communication connection to a corresponding type of telephone or communication line. A wireless link may also be implemented. In any such implementation, communication interface 2218 sends and receives electrical, electromagnetic, or optical signals that carry digital data streams representing various types of information.

[0188] Network link 2220 typically provides data communication through one or more networks to other data devices. For example, network link 2220 may provide a connection through local network 2222 to a host computer 2224 or to data equipment operated by an Internet Service Provider (ISP) 2226. ISP 2226 in turn provides data communication services through the world-wide packet data communication network now commonly referred to as the “Internet” 2228. Local network 2222 and Internet 2228 both use electrical, electromagnetic, or optical signals that carry digital data streams. The signals through the various networks and the signals on network link 2220 and through communication interface 2218, which carry the digital data to and from computer system 2200, are exemplary forms of transmission media.

[0189] Computer system 2200 can send messages and receive data, including program code, through the network(s), network link 2220 and communication interface 2218. In the Internet example, a server 2230 might transmit a requested code for an application program through the Internet 2228, ISP 2226, local network 2222 and communication interface 2218. The received code may be executed by processor 2204 as it is received, and / or stored in storage device 2210, or other non-volatile storage for later execution.

[0190] The operations of the processes described herein can be performed in any suitable order unless otherwise indicated herein or clearly contradicted by context. The processes described herein (or variations and / or combinations thereof) may be performed under the control of one or more computer systems configured with executable instructions, and may be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) collectively executed on one or more processors. The code may be stored on a computer-readable storage medium, for example, in the form of a computer program including multiple instructions executable by one or more processors. The computer-readable storage medium may be non-transitory. The code may be provided conveyed by a transitory computer-readable medium, for example, a transmission medium, such as in the form of a signal transmitted over a network.

[0191] Conjunctive language, such as phrases of the form "at least one of A, B, and C" or "at least one of A, B, and C," is understood in the context where it is generally used to indicate that an item, term, etc. can be either A or B or C, or any non-empty subset of the set A, B, and C, unless specifically stated otherwise or clearly contradicted by context. For example, in the illustrative example of a set having three members, the conjunctive phrases "at least one of A, B, and C" and "at least one of A, B, and C" refer to any of the following sets: {A}, {B}, {C}, {A,B}, {A,C}, {B,C}, {A,B,C}. Thus, such conjunctive language is not generally intended to suggest that a particular embodiment requires that at least one of A, at least one of B, and at least one of C, each be present.

[0192] The use of examples or exemplary language (e.g., "e.g., "etc.") provided herein is intended merely to better clarify embodiments of the invention and does not impose limitations on the scope of the invention unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention.

[0193] In the foregoing specification, embodiments of the invention have been described with reference to numerous specific details that may vary from implementation to implementation. Accordingly, the specification and drawings should be interpreted in an illustrative rather than a restrictive sense. The sole and exclusive indicator of the scope of the invention, and what is intended by the applicant to be the scope of the invention, is the set of claims issuing from this application, together with the literal and equivalent scope of the specific form from which such claims are derived, including any subsequent amendments.

[0194] Further embodiments may occur to those skilled in the art after reading this disclosure. In other embodiments, combinations or subcombinations of the above-disclosed inventions may be advantageously made. Example arrangements of components are shown for illustrative purposes, and combinations, additions, rearrangements, and the like are contemplated in alternative embodiments of the invention. Thus, while the invention has been described with reference to exemplary embodiments, those skilled in the art will recognize that numerous modifications are possible.

[0195] For example, the processes described herein may be implemented using hardware components, software components, and / or any combination thereof. Accordingly, the specification and drawings are to be interpreted in an illustrative and not a restrictive sense. It will be apparent, however, that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the claims, and that the invention is intended to cover all modifications and equivalents within the scope of the following claims.

[0196] All references cited in this specification, including publications, patent applications, and patents, are hereby incorporated by reference to the same extent as if each reference was individually and specifically indicated to be incorporated by reference and was set forth in its entirety herein. (Other possible items) (Item 1) 1. A computer-implemented method for administering a therapy, comprising: Under the control of one or more computer systems configured with executable instructions, determining a set of patient inputs; Treatment decision stage; estimating a circadian trajectory of said patient; determining one or more circadian mapping profiles; determining a preferred treatment duration from the circadian trajectory and the one or more circadian mapping profiles; and administering said treatment in response to an alert that said preferred treatment period has occurred or will occur. A computer-implemented method comprising: (Item 2) Item 1. The computer-implemented method of item 1, wherein the period is one day. (Item 3) Item 10. The computer-implemented method of item 1, wherein the period is one minute. (Item 4) Item 1. The computer-implemented method of item 1, wherein the length of the period varies according to an uncertainty measure of the circadian trajectory, the length being longer when the uncertainty measure is higher and the length being shorter when the uncertainty measure is lower. (Item 5) Item 10. The computer-implemented method of item 1, wherein the treatment is administration of a substance. (Item 6) Item 6. The computer-implemented method of item 5, wherein the substance is one or more of a drug, a nutrient, or a medicine. (Item 7) Item 10. The computer-implemented method of item 1, wherein the set of patient inputs includes data derived from signals received from a patient wearing a wearable data system. (Item 8) Item 10. The computer-implemented method of item 1, wherein the circadian trajectory is derived by a scheduler using at least one biophysical model of the human circadian clock and at least one statistical model of the human circadian clock. (Item 9) Item 10. The computer-implemented method of item 1, wherein the preferred treatment period is optimized based on relating circadian time to the time for taking the medication to generate raw model outputs. (Item 10) 2. The computer-implemented method of claim 1, further comprising presenting the patient treatment output in a human-interpretable format. (Item 11) Item 10. The computer-implemented method of item 1, further comprising connecting the model output to an environmental control to adjust the environment such that circadian-related behavior is adjusted toward the target constrained time. (Item 12) Item 12. The computer-implemented method of item 11, wherein the treatment is administered in an infusion clinic. (Item 13) Item 12. The computer-implemented method of item 11, wherein the treatment is a scheduled surgery. (Item 14) Item 10. The computer-implemented method of item 1, further comprising filling gaps in one or more of the models at missing data according to a rule set. (Item 15) 2. The computer-implemented method of claim 1, further comprising providing a model for generating an optimal preferred duration of treatment, wherein the treatment is administration of a drug, the model relating circadian time to a time for taking the drug, and the model including rules describing how the drug and different molecules in the body bind and interact with each other. (Item 16) Item 16. The computer-implemented method of item 15, wherein the rules include variables representing molecular formulas and parameters representing binding rates. (Item 17) 2. The computer-implemented method of claim 1, further comprising providing a model that generates optimal times for taking medication by relating circadian time to times for taking medication to generate raw model outputs, and gating trigger alerts based on logical gating to provide optimal times for intervals between treatments. (Item 18) 2. The computer-implemented method of claim 1, further comprising providing a model that generates optimal times for taking medications by relating circadian time to times for taking medications to generate raw model outputs, wherein the model of the core circadian pacemaker can be connected to a model of the blood-brain barrier. (Item 19) 2. The computer-implemented method of claim 1, further comprising providing a model that generates optimal times for taking medications by relating circadian time to times for taking medications to generate raw model outputs, wherein the model of the core circadian pacemaker can be connected to a model of the liver. (Item 20) 2. The computer-implemented method of claim 1, further comprising providing a mechanism for converting the raw model output into a human-interpretable format, the method providing optimal times to take medications subject to a rule that a patient can only take one of these medications on a particular day. (Item 21) A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor of a computer system, cause the computer system to perform the method described in item 1. (Item 22) 1. A computer system comprising: one or more processors; and a storage medium storing instructions that, when executed by the at least one processor, cause the system to implement the method of claim 1; A computer system comprising:

Claims

1. 1. A computer-implemented method for administering a therapy, comprising: under the control of one or more computer systems configured with executable instructions, determining a set of patient inputs; determining a treatment for the patient; estimating a circadian trajectory of the patient, wherein the circadian trajectory is a mapping between the patient's measured circadian state and elapsed real time; determining one or more circadian mapping profiles of said treatments, wherein a particular treatment circadian mapping profile maps variance of said particular treatment to circadian states; determining a preferred treatment duration for the treatment from the circadian trajectory of the patient and the one or more circadian mapping profiles of the treatment; and administering said treatment in response to an alert that said preferred treatment period has occurred or will occur. A computer-implemented method comprising:

2. The computer-implemented method of claim 1 , wherein the preferred treatment duration is one day.

3. The computer-implemented method of claim 1 , wherein the preferred treatment duration is 1 minute.

4. 2. The computer-implemented method of claim 1, wherein the length of the preferred treatment period varies according to an uncertainty measure of the circadian trajectory, the length being longer when the uncertainty measure is higher and the length being shorter when the uncertainty measure is lower.

5. The computer-implemented method of claim 1 , wherein the treatment is administration of a substance.

6. The computer-implemented method of claim 5 , wherein the substance is one or more of a drug, a nutrient, or a medicine.

7. The computer-implemented method of claim 1 , wherein the set of patient inputs comprises data derived from signals received from a patient wearing a wearable data system.

8. The computer-implemented method of claim 1 , wherein the circadian trajectory is derived by a scheduler using at least one biophysical model of the human circadian clock and at least one statistical model of the human circadian clock.

9. 10. The computer-implemented method of claim 1, wherein the preferred treatment duration is optimized based on relating circadian time to times for taking medications to generate raw model outputs.

10. The computer-implemented method of claim 1 , further comprising presenting the patient treatment output in a human-interpretable format.

11. The computer-implemented method of claim 1 , further comprising connecting the model output to an environmental control to adjust the environment such that circadian-related behavior is adjusted toward the target constrained time.

12. The computer-implemented method of claim 11 , wherein the treatment is administered at an infusion clinic.

13. The computer-implemented method of claim 11 , wherein the treatment is a scheduled surgery.

14. The computer-implemented method of claim 1 , further comprising filling gaps in one or more of the models in missing data according to a rule set.

15. 10. The computer-implemented method of claim 1, further comprising providing a model for generating an optimal preferred duration of treatment, wherein the treatment is administration of a drug, the model relating circadian time to a time for taking the drug, and the model including rules describing how the drug and different molecules in the body bind and interact with each other.

16. The computer-implemented method of claim 15 , wherein the rules include variables representing molecular formulas and parameters representing binding rates.

17. 10. The computer-implemented method of claim 1, further comprising providing a model that generates optimal times for taking medications by relating circadian time to times for taking medications to generate raw model outputs, and gating trigger alerts based on logical gating to provide optimal times for intervals between treatments.

18. 10. The computer-implemented method of claim 1, further comprising providing a model that generates optimal times for taking medications by relating circadian time to times for taking medications to generate raw model outputs, wherein the core circadian pacemaker model can be connected to a model of the blood-brain barrier.

19. 2. The computer-implemented method of claim 1, further comprising providing a model that generates optimal times for taking medications by relating circadian time to times for taking medications to generate raw model outputs, wherein the core circadian pacemaker model can be connected to the liver model.

20. 20. The computer-implemented method of claim 19, further comprising providing a mechanism for converting the raw model output into a human-interpretable format, wherein the computer-implemented method provides optimal times to take medications subject to a rule that a patient can only take one of these medications on a particular day.

21. A computer program storing instructions which, when executed by at least one processor of a computer system, cause the computer system to perform the computer-implemented method of any one of claims 1 to 20.

22. 1. A computer system comprising: one or more processors; and A storage medium storing instructions that, when executed by the one or more processors, cause the computer system to implement the computer-implemented method of any one of claims 1 to 20. A computer system comprising:

Images