Introduction The Institute of Medicine estimated that as

Introduction The Institute of Medicine estimated that as many as 98,000 patients die each year due to preventable medical errors, making this the sixth leading cause of death in the United States, and claiming more lives than diabetes or Alzheimer disease [1]. Additionally, the expense of medical care necessitated because of medical errors, lost income, and disability, results in a total cost between $17 billion and $29 billion per year [1]. There are many potential sources of error in patient care, such as medical prescriptions, transcriptions, dispensing and administration of drugs, and monitoring patient’s responses. However, among these, administration errors account for approximately 32% of morbidity and mortality cases in inpatient facilities [2]. Because of this, there is a pressing need to detect drug identity and concentration during administration, allowing for precise measurement of dosages, and preventing errors in real time before adverse effects take place. While monitoring of specific medications such as antidepressants and anticoagulants are important for compliance and toxicity checks [3], there are many other pharmaceuticals that when wrongly administered will result in dangerous consequences [2]. Currently, several specialized assay based techniques have been used to monitor medication errors in specific therapeutic treatments, known as therapeutic drug monitoring (TDM) [3–6]. However, assay and label based detection systems such as immunoassays, electrochemical assays, and lateral-flow assays (LFAs) have several limitations that prevent them from being used for point-of-care sensing [7]. These techniques are time consuming, and require large volumes of analyte to achieve the needed sensitivity. Furthermore, drug assays use complicated electrochemical measurements, suffer from background interference in complex solutions and have poor thermal purchase tropisetron [8–14]. Currently there are no available technologies for detection of overdose or incorrect drugs during administration. There are many sensors that can work in almost real time and do not require use of reagents, but they typically measure refractive index and are based on ring-resonators [23–26], photonic-crystals [15–21], whispering gallery-mode [22–24], plasmonic structures [25] and other optical components, that without additional modifications do not provide enough specificity for differentiation between multiple drugs. In addition, these techniques are prone to noise caused by any microscale particles present in the system and are very sensitive to small temperature changes. In general, traditional approaches to drug identification include color test [26], microscopic microcrystal analysis [27], thin layer chromatography [28], gas chromatography [29], mass spectroscopy [30], X-ray diffraction [31], and different types of spectroscopy that are most suitable for real time detection. By using traditional spectroscopic techniques (IR, Near-IR, NMR, Raman, etc.) it has been shown that drug identity can be determined [32–35]. Once a drug has been identified using traditional spectroscopy, information regarding concentration may be obtained by quantifying absorption at a specific wavelength. Furthermore, spectroscopy using a single optical fiber is becoming a powerful approach for analysis of biological samples [36–40]. However, for complex systems such as biological fluids that contain a variety of free floating particles and cells, the effectiveness in determining a drugs concentration is significantly reduced due to scattering and interference by these objects. As a result, additional reagents are required to amplify the spectral signatures of compounds of interest (e.g. biomarkers of a specific multiple diseases) [41–43]. Unfortunately, studies requiring reagents and labeling can only be performed in specialized laboratories, using large sample volumes, as well as extensive time for analysis ranging from several hours to multiple days. Because of this, a reliable, reagent and label-free, detection method that can run in real-time by using a small sample volume would greatly benefit point-of care drug monitoring.