Welcome to OpenForum. Join our global community: Sign up, Login or Learn more

Precision Clinical Trial Designs for Precision Drug Development and Precision Medicine

Accelerate precision medicine and productivity with clinical trial designs that yield Personal Quantitative Treatment Response Phenotypes.

Photo of Curtis A. Bagne

Written by

Focus of Innovation

  • Cost reduction
  • Data collection
  • Data interpretation
  • Dosing
  • Mobile technology
  • Monitoring or measurement
  • Patient-centric or decentralization
  • Regulation
  • Size of study
  • Target identification

Focus of Innovation - Other

Phenotype Phenomics Genotype Genomics Interaction Power Multivariate Time Series Individuality

Name of applicant

Curtis A. Bagne

Title of applicant

President and Founder

Affiliation of applicant

DataSpeaks, Inc.

Names, titles and affiliations of teammates

Curtis A. Bagne, the author of this idea, actively seeks collaborators, partners, and licensees for scientific and business development.

Introduction: Precision Randomized Controlled Trial (PRCT) designs promise to accelerate precision medicine for the prevention and treatment of chronic disorders that account for about 86% of annual health care expenditures of about $3 trillion in the United States alone. PRCTs will help make medicine, drug development and regulation, healthcare, and wellness more healthful and precise.Three PRCT designs, enabled by the Science of Individuality Measurement Algorithm (SIMA), build on each other to help users integrate and optimize:

  1. Individual patient care,
  2. Drug development and regulation,
  3. Advanced comparative effectiveness AND safety research.

PRCT designs for chronic disorder treatments:

  • Provide an alternative to classical RCT designs that permanently and often unnecessarily wash out the effects of individuality with group averages. This averaging convention is antithetical to the value of genomics and makes it hard to target the right drug to the right patient.
  • Will empower physicians to be better clinicians AND better researchers. Now physicians often are torn by having to use information gathered over time to improve treatment for individual patients versus group averages for research. PRCT designs use both types of evidence scientifically to help obviate the clinical research to clinical practice translation problem.
  • Capitalize on sensors, drug dispensing and monitoring devices, that collect multivariate time series data as well as cloud computing and the Internet of Things.
  • Investigate whole, unique, and valuable individual persons more like the Complex Adaptive Systems (CAS) they really are.
  • Accelerate advancement of population health and medicine. First, coordinate and conduct two or more single-patient PRCTs enabled by SIMA. Second, aggregate results with statistics. Single-patient (N of 1) PRCTs that improve individual health will improve group average or public health.
  • Can help empower individuals to take more responsibility for their own health and that of significant others. This can improve health and cut costs.

Most stakeholders – patients, physicians, providers, employers, the biosciences industry, regulators, investors, and payers stand to benefit from PRCT designs. Product liability attorneys would end up with less business.

SIMA computes Precision Quantitative Diagnostic Phenotypes and Personalized Precision Quantitative Treatment Response Phenotypes when applied to data about living systems such as patients. Distinguish quantitative phenotypes from categorical phenotypes.

Figure 1 shows that "Precision Diagnoses" and "Precision Treatment" mutually reinforce each other to form a closed-loop precision medicine learning system. This applies for chronic disorders. This same approach applies more broadly to precision health and healthcare.

Figure 1 focuses on applied science. Basic science would feed into Figure 1 especially when basic research is accelerated with SIMA.

Figure 1 also identifies characteristics that make both quantitative diagnostic and treatment response phenotypes precise for each person. Figure 1 also shows that these will lead to more and better-targeted products, more services, better outcomes, and lower costs.

BIG IDEA – SIMA computes Precision Quantitative Phenotypes.

Figure 1. The Closed Loop Patient-Centered Precision Medicine Learning System Enabled by SIMA

Image title

Learn more about Figure 1 and why it is important.

BIG IDEA – Both diagnostic and treatment response phenotypes need and can be single-patient or N of 1.

SIMA for PRCTs: The Science of Individuality Measurement Algorithm (SIMA) enables Precision RCT designs. Eric Topol championed “The Science of Individuality” on page 228 of his book, The Creative Destruction of Medicine. Eric inspired the SIMA name.

BIG IDEA – SIMA scores individuals. Statistics aggregates.

Figure 2 shows how “Apply SIMA” comes between “Collect Data” and “Aggregate” on the path to “Precision Medicine.” Figure 2 is for human health and medical applications of SIMA.

Figure 2. Introduction to the Science of Individuality Measurement Algorithm (SIMA)

Image title

Learn more about Figure 2 and why it is important.

SIMA is a discipline that processes multivariate time series broadly defined as two or more repeated measurements at the same times for each of two or more action variables. Unlike germline genetic characteristics used for forensic identification, action variables typically vary and fluctuate in level over time. SIMA requires at least one action variable time series to operate as an independent variable or predictor and one to be a dependent or predicted variable. Increase the number of repeated measurements to increase power for separating treatment effect and other signals from noise.

Unlike cross-sectional data and change scores, multivariate time series can provide orders of magnitude more information to understand individuals scientifically. By analogy, this comparison is like the difference between snapshots and movies. Table 1 identifies some advantages of data movies as distinct from data snapshots and change scores. By analogy, SIMA “develops” data movies.

Table 1. Some Advantages of “Data Movies” Compared to “Data Snapshots”______________

Apply SIMA to multivariate time series (data movies) whenever feasible to:

  • Understand individuals scientifically with data.
  • Measure signaling dynamics as in living systems.
  • Enable measures about individuals that increase reliability by using information from more repeated measurements to help separate signals from noise. (The current version of SIMA software can process data with from 2 to 500 repeats for up to about 100 time series.)
  • Enable measures about individuals that are more apt to be valid when doses and levels of other independent and predictor variables are randomized and blinded over time.
  • Be used to account for temporal phenomena such as episodes of independent and dependent events, delay of response, persistence of response, etc.
  • Separate signals from trends – disease progression, spontaneous recovery, assay drift.
  • Help quantify the temporal criterion of causal relationships when data are non-experimental.
  • Help elucidate how persons and other individuals develop, adapt, mature, and age.
  • Elucidate mechanisms as well as their up-regulation and down-regulation as by drugs.
  • Help quantify emergent system properties such as coordinated action. Pain and consciousness do not come in molecules.
  • Capture more of what distinguishes living people from dead people and inanimate objects.
  • Capitalize on being able to elucidate "work" as introduced just below.

Data movies are apt to prove more cost effective than data snapshots and change scores. People often prefer video, movies, and television to snapshots for good reasons.


Use SIMA to quantify how an individual person, investigated scientifically as a Complex Adaptive System, works over time as “work” is operationally defined in Figure 3. Figure 3 illustrates work in the context of health, disease, and wellness with a directed graph with nodes and edges. There is one action node for each time series with arrows for edges.

BIG IDEA – Each unique and valuable person is a Complex Adaptive System with emergent system properties. Investigate and treat each person accordingly.

Figure 3. SIMA Quantifies How an Individual Person "Works" Over Time

Image title

See more about Figure 3 and why it is important.

Benefit and Harm Scores Enable PRCTs: SIMA software users can select the benefit and harm scoring option to evaluate response (see Figure 3). Benefit and Harm Scores are a common metric for integrated and scientific evaluations of the benefits (effectiveness) and harms (safety). This common metric can help reduce the dimensionality of treatment evaluation problems from tens, hundreds, or thousands of variables that might be affected by treatment to one dimension of benefit and harm.

Benefit and Harm Scores are IoT scores for which the positive and negative signs are reversed as might be required to account for the fact that higher levels of a particular response variable action node can be either toward or untoward. As examples, SIMA users can specify both lower levels of LDL (“bad”) cholesterol and higher levels of HDL (“good”) cholesterol as being toward or beneficial as determined by clinical epidemiology. In addition, users can personalize PRCTs by using individual patient preferences for various health effects to help determine if drug effects, as measured with IoT scores, are beneficial or harmful. Assign 0 for weight when a treatment effect measured by an IoT score is considered to have no value.

Figure 4 introduces Benefit and Harm Person to represent use of SIMA for evaluative investigations of response as response over time is shown in Figure 3. Evaluations include patient or person-centered Precision RCT designs – the focus of this Forum idea.

Figure 4. Meet Benefit and Harm Person: Quantitative Treatment Response Phenotypes

Image title

All three PRCT designs introduced in Table 2 below start with advanced design single-patient, single-person, or N of 1 RCTs. PRCTs quantify and help evaluate drug effects, including causality, for each person before using statistics to aggregate results from two or more persons. Accordingly, Figure 4 illustrates a single person. Correspondingly, use statistics to aggregate and balance Benefit and Harm Scores from two or more persons. Then the person in Figure 4 would represent a group or population.

For simplicity, Figure 4 weighs benefits and harms against each other for only six response variables. Avoid missing potential safety problems and new indications by increasing the number of response variables to the extent feasible. Recognize how SIMA can provide detailed information about treatment effects. For example, a commonly used scale for antidepressant RCTs has 17 items. Profile benefit and harm across all 17 items separately for each person before computing Overall Benefit and Harm Scores. Improve identification of indications and contraindications with these capabilities.

SIMA separates the quantification of evidence for treatment effects from how these effects are valued. Treatment effects are facts quite distinct from how these facts are valued. Suppose there is uncertainty about how an IoT score with respect to a particular lipid or protein is valued in terms of something that matters to patients such as risk of heart attack, stroke, and death. Users can compute IoT scores for such effects. Except when testing prespecified primary hypotheses, update evaluations with new toward and untoward directionalities and weights as these become available. Furthermore, SIMA users need not guess about whether a particular response variable should be a part of either a safety or an effectiveness evaluation. Collect and process the data. SIMA will tell you about the effects. This illustrates the meaning of the company name – DataSpeaks.

Figure 4 illustrates how Benefit and Harm Person evaluates Overall Benefit and Harm over time and across response variables for individuals. Overall Benefit and Harm is 21.59 – the sum of weighted positive or beneficial effects and negative or harmful effects.

The key to Figure 4 shows the derivation of this result. All six Benefit and Harm Scores are possible with 16 repeated measurements of dose and the response variables. All six scores are in standard deviation units – (Bagnes) when computed with SIMA. Differential weights account for differences in clinical significance and patient preferences.

Benefit and Harm Person is very capable in ways that go beyond computing Overall Benefit and Harm Scores for each person as indicated in Figure 4. Learn more about the capabilities of Benefit and Harm Person.

BIG IDEA – Fit Precision Quantitative Treatment Response Phenotypes to Precision Quantitative Diagnostic Phenotypes whenever possible. 

Table 2 identifies three Precision Randomized Controlled Trial (PRCT) designs enabled by SIMA and Benefit and Harm Person (Figure 4).

Table 2. Three Precision Randomized Controlled Trial Designs__________________________

Here is a hierarchy of three PRCT designs. Each subsequent design in this list extends the previous design. Select the right design for your primary purpose and use.

  1. Single-patient (N of 1) PRCT designs: These have potential to become the new gold standard for the prevention and management of chronic disorders. N of 1 PRCTs are a scientific analogue to how clinicians often evaluate treatments over time for individual patients. Clinicians seeking to optimize the continued care of individuals often assess responses over time during drug challenge, de-challenge, and re-challenge with multiple doses. N of 1 PRCTs make this process more scientific by randomizing and masking doses, including any placebo as zero-dose, over time; collecting multivariate time series about both treatment and health; and using SIMA to compute Personal Precision Quantitative Treatment Response Phenotypes for each individual.
  2. Single-group, multiple N of 1 PRCT designs: These have potential to become the new gold standard for chronic disorder drug development and regulation. Single-group, multiple N of 1 PRCT designs are coordinated sets of N of 1 PRCTs introduced just above. Aggregate results from two or more patients with statistics to describe groups and make inferences from samples of patients to populations. 
  3. Multiple group, multiple N of 1 PRCT designs: These have potential to become a new gold standard for much of comparative safety AND effectiveness research. These are coordinated sets of single-group, multiple N of 1 PRCT designs introduced just above. Use double randomization. Randomize patients to groups defined by two or more types of treatment. Randomize two or more doses, including any placebo, of the same type of treatment over time for each patient.

Please note: Precision RCT designs apply only when both treatments AND response variables can vary and fluctuate in level over time. They do not apply for once in a lifetime treatments such as having a prostatectomy or once in a lifetime events such as death. “There is more to life than death.” Use SIMA to account for this "more."


BIG IDEA – Precision quantitative phenotypes, diagnostic and treatment response, will help capitalize on genomics by making phenomics more equally precise.

Demonstration: Figure 5 is a simple demonstration of a single-group, multiple N of 1 PRCT design – a coordinated set of three N of 1 PRCTs. Each N of 1 PRCT uses 16 repeated measurements to evaluate the effects of four randomized doses of one type of drug with respect to three response variables. See how IoT scores use “Direction” to become Benefit and Harm Scores and how the three response variable specific Benefit and Harm Scores for each patient are differentially weighted and averaged to compute an Overall Benefit and Harm Score. Next, the null hypothesis of no Overall Benefit and Harm across the three patients is evaluated with a two-tailed, single-group t-test on the mean. See how this null hypothesis is rejected in the positive or beneficial direction with only three patients. Capitalize on the power of repeated measurements.

BIG IDEA – Measure first with SIMA to make statistics easier.

Figure 5: A Single-group, Multiple N of 1 PRCT Design

Image title

Figure 6 shows response variable specific and overall benefit and harm as a function of dose for each of the three patients in Figure 5. Part D shows the aggregated results.

Figure 6: Get Doses Right

Image title

Protect patient safety as by processing data in Figure 5 iteratively as shown in Figure 7.

Figure 7: Protect Patient Safety and Reduce Liability

Image title

Inspired by: Many people deserve credit and definitely inspired this idea. Here are some of the most important sources of inspiration and encouragement over decades. I present these in roughly chronological order of first influence. SIMA does build on the shoulders of giants.

  1. Alvan Feinstein (https://en.wikipedia.org/w/index.php?title=Alvan_Feinstein&oldid=686009672) and more specifically his 1967 book, Clinical Judgment (http://www.amazon.com/Clinical-judgment-Alvan-R-Feinstein/dp/B0007GUEIU), and his long series on Clinical Biostatistics in Clinical Pharmacology and Therapeutics conditioned my thinking to help me invent SIMA.
  2. Lois Verbrugge (https://sites.google.com/a/umich.edu/verbrugg/) and especially her publication about health diaries (http://www.ncbi.nlm.nih.gov/pubmed/6986517) helped me recognize the need for better methods to extract information, knowledge, and value from repeated measurements data.
  3. Gordon Guyatt (http://fhs.mcmaster.ca/ceb/faculty_member_guyatt.htm) confirmed that I was on the right track when I happened upon his 1986 article, “Determining optimal therapy--randomized trials in individual patients” (http://www.ncbi.nlm.nih.gov/pubmed/2936958), on the new journals shelf in the library where I worked. Gordon continued to inspire and encourage me through a joint pharmaceutical company sponsored project and more recent contacts.
  4. Helena Kraemer (https://med.stanford.edu/profiles/helena-kraemer) was an inspired and very helpful peer-reviewer for the 1992 publication that included core computations for what I now call SIMA: http://dataspeaks.com/resources/APA-JCCP-1992-Vol60-No2-P225-239.pdf.
  5. Jill Goldberg (https://www.linkedin.com/in/jill-goldberg-82a472) engineered a robust version of what I now call SIMA software. She did so with expertise and inspiring style.
  6. Gil Omenn (http://www.hg.med.umich.edu/faculty/gilbert-s-omenn-md-phd) and Brian Athey (http://www-personal.umich.edu/~bleu/) have inspired and encouraged me for many years. I have been informed and inspired by many seminars at the University of Michigan.
  7. In 2005 Robert Temple of the FDA (http://www.fda.gov/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CDER/ucm374560.htm) wrote the following as part of an extended series of email exchanges: “I like the idea of multiple doses to an individual and have been suggesting n of 1 designs (but with no takers).” SIMA helps enable this approach.
  8. Leroy Hood (https://www.systemsbiology.org/bio/leroy-hood/) provided my effective introduction to systems biology and the importance of “interactions between/among elements” that “give rise to the system’s Emergent properties.” SIMA quantifies Interactions-over-Time and coordinated action as an emergent property of individual Complex Adaptive Systems. In addition, Lee is “pioneering new approaches to P4 medicine—predictive, preventive, personalized and participatory….” SIMA can accelerate P4 medicine. I presented the following at an international symposium at the Institute for Systems Biology in 2008: http://dataspeaks.com/resources/bagne_handout.pdf.
  9. Roger Newton (http://www.esperion.com/meet-esperion/the-esperion-team/) was quick to understand and appreciate what I now call SIMA. Roger has provided inspiration, advice, and encouragement over many years.
  10. Both Francis Collins (https://www.nih.gov/about-nih/who-we-are/nih-director/biographical-sketch-francis-s-collins-md-phd) and Thomas Insel (http://www.nimh.nih.gov/about/dr-tom-insel-to-step-down-as-nimh-director.shtml) provided inspiration and were quick to understand and appreciate what I had to say during brief conversations that led to some follow up.
  11. David Sahner (https://www.sri.com/about/people/david-sahner) recognizes SIMA’s potential. This led to a licensing agreement: https://www.sri.com/newsroom/press-releases/sri-international-licenses-dataspeaks-software-development-virtual-clinical.
  12. Eric Topol (http://www.stsiweb.org/translational_research/research_faculty/topol/) published The Creative Destruction of Medicine (http://creativedestructionofmedicine.com/?p=3). Eric champions “THE SCIENCE OF INDIVIDUALITY” on page 228 of his book. SIMA is a new discipline that helps enable the science of individuality. Beginning on page 229, Eric discusses: “THE N OF 1 TO THE N OF BILLIONS.” Achieve this by using statistics to analyze scores from SIMA together with other data about individuals when there are data on two or more individuals as shown in Figure 2.
  13. I heard and met Keith Yamamoto (http://profiles.ucsf.edu/keith.yamamoto) and learned of this Harvard Forum at the Personalized Medicine World Conference 2016 (http://2016sv.pmwcintl.com/). Keith coauthored “Precision medicine: Beyond the inflexion point:” http://stm.sciencemag.org/content/scitransmed/7/300/300ps17.full.pdf. SIMA does need to be evaluated for its potential to help drive precision medicine beyond the inflexion point.
  14. Google led me to Nicholas Schork (http://www.jcvi.org/cms/about/bios/nschork/) and his publication, “Time for one-person trials” (http://www.nature.com/polopoly_fs/1.17411!/menu/main/topColumns/topLeftColumn/pdf/520609a.pdf) on 2/12/2016. Nicholas provides an inspiring description of why clinical trials need to begin with “one-person trials.”

The time is ripe, both in terms of thought leadership and technological capabilities, to test and further develop Precision Randomized Controlled Trial designs and Precision Quantitative Phenotypes.


DataSpeaks, Inc. is an intellectual property out-licensing and development company. For more information see DataSpeaks.com. 

Why would your idea have a significant impact?

The pharmaceutical industry and its regulators have been muddling through a productivity crisis for decades with most stakeholders paying a price. This crisis, known as Eroom’s Law (Moore’s Law spelled backwards), has included:
1. Multiple rounds of corporate mergers, acquisitions, and downsizings resulting in the elimination of several hundred thousand high-paying jobs
2. Huge lost opportunity and product liability costs
3. Damaged reputations and political attacks
4. Emergence of business models that involve acquiring approved drugs, huge price increases, and slashing research and development.
This crisis is occurring despite huge advances in science and technology such as decoding genomes, high throughput screening, omic sciences, sensors and devices that collect multivariate time series, computing, and connectivity.
Limitations and failings of classical RCT designs appear to be a primary scientific and technical root cause of this productivity crisis.
PRCT designs, as introduced by this Forum idea, offer an opportunity to help reverse this productivity crisis.

What are the biggest hurdles to implementing your idea?

The biggest hurdle is that this idea disrupts current practice, regulatory guidance as by the FDA, and cultural habits of thought. Here are some relevant observations and comments.
1. Precision RCT designs would disrupt regulatory guidance as by the FDA. For example, this page from the FDA appears to focus on the “conduct of clinical trials” as distinct from the design of clinical trials: http://www.fda.gov/RegulatoryInformation/Guidances/ucm122046.htm.
2. This guidance about “Drug Study Designs” focuses on the use of group comparisons to assess causality as if there is no other option: http://www.fda.gov/RegulatoryInformation/Guidances/ucm126501.htm. There is no mention of conducting RCTs in individual patients or N of 1 RCT designs. There is no mention of evaluating safety and efficacy by measuring the amount of evidence for benefit and harm over time and across response variables for each individual before statistical aggregation for group results.
3. Results from the Critical Path Initiative (http://www.fda.gov/ScienceResearch/SpecialTopics/CriticalPathInitiative/ucm076689.htm) and the Clinical Trial Transformation Initiative (http://www.ctti-clinicaltrials.org/) as well as methodology standards from the Patient-Centered Outcomes Research Institute (http://www.pcori.org/research-results/research-methodology/pcori-methodology-standards) appear to be subject to essentially the same limitations just identified for the FDA. Classical RCT designs are not precise, sufficiently quantitative, or patient-centric. Asking patients for input about what matters for response variables and outcomes is a good idea. However, this does not overcome limitations of classical RCT design.
4. The great majority of thought leaders seem to presume that scientific generalization must occur across individuals in groups and populations as distinct from over time within individuals as with SIMA. Statisticians drive RCT design. Primary hypotheses are framed in terms of primary response variables as distinct from Overall Benefit and Harm scores as illustrated by Figure 4. Baseline and endpoint thinking largely precludes a signal processing approach such as SIMA. We are burdened by categories as distinct from quantitative phenotypes. Reductionists favor molecules, focus on biology to the relative neglect of psychological and social levels of investigation, focus on efficacy in ways that neglect safety, and neglect emergent system properties such as coordinated action. Many seem to think that answers will come from mathematics and mathematical models as distinct from measurement. (SIMA can inform model development by offering new measures to model.)

How difficult will it be to overcome those obstacles?

Precision Randomized Controlled Trial designs will be difficult to achieve quickly and widely. However, time is ripe for PRCTs. Thought leaders have taken us to the cusp of PRCTs as documented in the “Inspired by” section above. More and more connected sensors and devices are collecting the required multivariate time series data. These include devices that can control, adjust, and monitor doses over time as well as implanted drug pumps. Computers process big data faster than ever before. Books such as The Master Algorithm (http://www.amazon.com/The-Master-Algorithm-Ultimate-Learning/dp/0465065708) document the importance and impact of algorithms. SIMA is an algorithm. President Obama stated, “Doctors have always recognized that every patient is unique,” and announced and continues to support The Precision Medicine Initiative (https://www.whitehouse.gov/precision-medicine). PMI appears to be focusing on genomics. PRCTs offer a complementary approach that focuses on phenomics. Precision Quantitative Phenotypes can help reveal what genetic differences mean. This promises to help capitalize on genomics.

And now the Harvard Business School offers the Precision Trials Challenge. Time is ripe for PRCT designs.
Here are some low-hanging fruit opportunities to apply, develop and externally validate SIMA in the context of evaluating the effects of environmental exposures including drugs and for RCTs.
1. Many classical RCTs collect more repeated measurements than are used to test primary hypotheses. Often SIMA could process more repeated measurements to gain power and mine new insights.
2. Some clinical trials collect data on levels of drug and drug metabolites in bodily fluids repeatedly together with data for other laboratory measures, signs, symptoms, and electrophysiological variables. SIMA can help process such data to mine new insights.
3. Drug rescue and repurposing, post-marketing surveillance
4. SIMA can process data about independent or predictor variables such as diet, exercise, allergens, and pollutants in addition to drugs.
5. SIMA could help empower members of quantified-self community to conduct clinical trials, including PRCTs, on themselves and their significant others. Groups of like-minded individuals could cooperate so that results can be aggregated and shared.

How would you test for the impact of this idea and how would you quantify the impact?

Long term, the impact of PRCT designs can be evaluated in terms of how PRCTs improve productivity and make better outcomes more affordable.

In the meantime, it might be possible to evaluate impact with a classical RCT that randomizes patients to either classical RCT design or PRCT design groups. Consider evaluating impact with an approved drug with rather fast temporal dynamics as for pain or blood pressure. Evaluate impact with measures of health outcomes and cost. Such a trial could raise ethical concerns about randomizing patients to the classical RCT only arm, especially for any life threatening condition.


Join the conversation:

Photo of Jim

Dear Curtis Bagne ,

Following our conversation on the @Centralized Electronic Medical Record System with Research Friendly Capabilities and De-Identified Remote Monitoring forum I would like to discuss a couple topics with you if you don't mind to:  

1. First of all, I want to tell you that our @Living labs to link together researcher teams and practitioners don't include any kind of statistical tools for data analysis: The reason being that we focus exclusively on fostering the relation between researchers, specialists and primary care practitioners to allow close collaborative interaction for precision clinical trial enrolment, best practice translational adoption and collaborative clinical decision support. Out view is that once clinical data is collected, there must but specialised third parties that would analyse the data and extract the right conclusions.I mean this because I would love to discuss a possible collaboration with DataSpeaks and Linkcare beyond the "Precision Trials Challenge" if you think is worth spending some time discussing such possibility.  

2. On the other hand, I want to share with you that following our strategy on being agnostic on "statistical data processing" I am not sure if I have mentioned to you that we are experiencing a similar approach to SIMA's algorithm with a complete different purpose:The local catalan health system (the province in Spain to which Barcelona belongs to) has been able to successfully develop a methodology called "Adjusted Morbidity Groups" (AMG) which is an improved version of 3M's "Clinical Risk Groups" (CRGs). The different being that AMGs use real life data to assess individual patient's risk based in the information collected by HC3.The system uses also a multivariate annalists. Some early results can be found at: http://es.slideshare.net/jimroldan/01-gma-en-and-zh  

3. Finally, I want to point out that Linkcare has an open architecture that allows third parties such as DataSpeaks to create external services that provide Clinical Demission Support Services (CDSS) to interact with other programs and protocols. So it seems quite simple to:a) Retrieve clinical data from patients enrolled in Linkcare's Open Living Labsb) Provide "turnkey" CDSS services in what we call "Knowledge as a Service" (KaaS) mode.Such services may be provided on free Creative Commons basis or as a subscription or pay-per-use fee similar to iTunes, AppStore or GooglePlay models.

What do you think about it?

Best regards,


Photo of Curtis A.

Hi, Jim:
I definitely "would love to discuss a possible collaboration with DataSpeaks and Linkcare...."

Please consider continuing the conversation through http://www.dataspeaks.com/contact

I hope your travels went well. I delayed my response knowing you had an equally long trip home.

Best Regards,

View all comments