Frequently Asked Questions
We love questions! Below are answers to the common ones people have as they start using our platform. Leave us a Message if you don't find what you need.
Individual accounts get free access to modeling methods that do not need PHI (Protected Health Information as defined by HIPAA). These include Probability of Target Attainment (PTA), Cumulative Fraction of Response (CFR), Empiric therapy/dosing for various antibiotics and (optimal) loading dose calculations for Phenytoin.
At this time, individual accounts can not access Bayesian adaptive dosing and other advanced models, due to the additional security and compliance infrastructure, and contracting needed for the use of PHI (e.g. past doses, measured blood concentrations, genomic data).
The validity, reproducibility and reliability of our models are key promises of our platform. To ensure this, we perform rigorous testing at four levels to validate that our models are implemented with integrity:
- Scientific & static code analysis: We ensure that the output values from models are exactly as defined in a peer-reviewed publication or in a pre-agreed specification for custom models. This validates that within the operating range of a model, our results will be exactly as expected from the model definition. All 3rd party libraries used to arrive at a computed outcome are also under strict source control.
- Unit tests: this ensures that each atomic function or transformation that contributes to a model is tested every time our code changes
- Integration tests: these ensure that functions interact reliably as they evolve, and
- Black-box/external tests: these validate that the overall output on our website and from our APIs are as expected, independent of how the underlying functions evolve.
The integrity of models is further supported by a strict internal Secure System Development policy, which requires code reviews of all substantive changes to all our source code repositories (modeling, application, backend services, staging and even homepage updates).
Our application interface is currently available in English, Portuguese, Spanish, French, Hungarian and Simplified Chinese, and can be further localized for 75 countries by fine-tuning your preferred date and time format, timezone, unit of measurements (height, weight, creatinine) and default paper size for printed reports.
If you are interested in using Rx Studio in other languages and locales, or would like to help with improving the current internationalization efforts, please contact us or consult our Translation Glossary pages.
Rx Studio is a Clinical Decision Support (CDS) platform. By design, our platform does not make any medical choices or decisions. Instead, given your choices and selections of models and input parameters, the platform simulates patient drug response and recommends an optimal dose to help you make an informed decision. The choice of models, regimens and target parameters are completely at the discretion of the medical professional. A trained health professional should always review the results of our platform and be the final decision maker.
While always deferring to the choices of the person running the simulation, our platform will occasionally provide warnings and guidance to mitigate the risk of anomalous data entry, improper model choice, poor model fit, (e.g. extreme PK parameters far from the related population's usual PK parameters), unexpectedly long times to reach steady-state (e.g. over 5 doses), or when the dosing recommendation output is outside known Clinical Practice Guidelines -- e.g. Rybak et al. (2020) for Vancomycin.
Some relevant controls are:
- All personal and health information is encrypted both in transit (e.g. when sharing information with our servers for simulation models) and at rest (e.g. when data is stored in our environment for audit purposes).
- We collect the minimal amount of patient information required for modeling.
- We conduct a risk assessment for all potential vendors, then execute a BAA with all vendors potentially dealing with PHI.
- Active monitoring of the cloud infrastructure and enterprise-grade application firewalls.