Chapter 7: START-IT Tool for Automated UDT Interpretation

Let’s take a closer look at how your clinic might be able to use the START-IT Tool to simplify urine drug testing (UDT) in your clinical setting…

In clinical use since early 2018, START-IT is applied at the bedside on a tablet PC by a non-medical clinical employee (i.e. no special medical designation is required). The START-IT Tool was designed to simplify UDT, including automating as much of the process as possible.

Since START-IT was initially created, it has evolved quickly to suit the needs of frontline clinicians and will soon simplify nearly every aspect of UDT in the office. START-IT collects all of the required information to interpret a UDT immunoassay (IA) test, including but not limited to: prescribed opioid and last dose, self-reported non-prescribed medications/drugs, and UDT IA testing results. START-IT then automatically interprets the test using best-evidence, and the report is uploaded into the EMR with the click of a button. START-IT gives explanations about limitations for that specific UDT result, including reasons for false positives and false negatives as applicable. START-IT also makes tailored recommendations about when to send the sample for further “confirmatory” testing (LC-MS), and facilitates tracking data digitally for quality improvement and research. Additional options in development also include having START-IT obtain patient consent, apply a validated risk stratification tool, and apply and calculate a Brief Pain Inventory. It is planned that it will also have the capacity to apply a one-time opioid treatment agreement (if desired). Reports will even be customized depending on whether the UDT was being conducted for chronic pain, opioid agonist treatment, or as an employee requirement. START-IT does all of this with no paper, no scanning, no delays, no errors in interpretation, and no time from the physician.

While START-IT has always been free (we have never charged a fee), a membership with OCEAN was previously required because its platform was needed. This however had recently changed and we are able to host START-IT on our website. We are grateful to the team at CognisantMD (makers of OCEAN) for their help in making it available for free from any internet browser, without an OCEAN subscription. Of course, the EMR integration is not available with the browser option, so in this case there will be a few extra steps (copying and pasting into the EMR), and automated data collection would not be possible.

As we delve more into how START-IT works and what it does, let’s start with an example of a final report. This gives a “big picture” perspective on how the information in START-IT can be used clinically. We’ll then cover the various steps of setting it up in your office and applying it in patient encounters.

Sample report for START-IT 2019:This assumes that a patient is prescribed morphine, reports last dose <24h ago, does not report any other drug use, has IA tested with a standard 5-panel IA (in this case – morphine panel, oxycodone panel, EDDP (methadone) panel, cocaine panel and benzodiazepine panel), and all panels are negative.

START-IT Automated IA Urine Drug Test Interpretation
OVERALL INTERPRETATION: UNEXPECTED. Morphine is prescribed and should be detected on today’s test (last dose reported <24h ago), however it was not detected (unexpected result). Given that patient reports are inconsistent with the result obtained here, confirmatory testing (LC-MS) would be strongly recommended. For detailed IA interpretation see “Interpretation Details” below.

TESTING CAPABILITIES and LIMITATIONS: With the IA panels used today, the following drugs would typically be DETECTABLE: morphine/codeine/heroin, oxycodone, methadone, benzodiazepines (with the exception of clonazepam and lorazepam), cocaine. The following drugs are VARIABLY DETECTABLE: hydromorphone, hydrocodone, clonazepam and lorazepam. The following is a list of drugs that would be UNDETECTABLE: fentanyl, tramadol, buprenorphine, barbiturates, amphetamines, methamphetamines, MDMA, cannabis, phencyclidine (PCP). Explanations for the more common examples of false positives and false negatives relevant to today’s test are explained under “Interpretation Details” below.

PATIENT BACKGROUND: Patient initials: RM. Pain Diagnosis: Chronic Back pain. Risk Category: LOW. Initial Selection Method: Random. Notification period: <36h. From recall list? No.

PRESCRIBED MEDICATIONS: Morphine (last reported <24h ago).


UDT IA PANEL RESULTS: Morphine panel NEGATIVE; EDDP (methadone) panel NEGATIVE; Cocaine panel NEGATIVE; Oxycodone panel NEGATIVE; Benzodiazepine panel NEGATIVE

MORPHINE (UNEXPECTED): Patient is prescribed morphine which should be detected with the IA panels used on this sample. The negative result is therefore unexpected, and is also inconsistent with patient’s last self-reported dose (patient reports last taking <24h ago which would typically be detected on this test). This raises concern that patient has actually not been taking the prescribed morphine (detection window approximately 3-4d). Explanations for a true negative result include that patient is taking too much of the medication early in a refill and so hasn’t taken it recently, or result may indicate that patient is diverting the medication. False negatives (negative results when the patient is actually taking the medication) are rare for morphine, however can be caused by: very dilute urine, dose being too low and/or medication being metabolized so rapidly that it is not detected (note that morphine IA tests have very low detectability thresholds so these reasons are unlikely).

ACTING ON UDT RESULT: Before acting on any UDT result, it is important to understand the limitations of the testing, and that they can inform – but are not a replacement for – the Clinical Gestalt. The intent of UDT is not to punish patients, but to inform the risk/benefit balance and facilitate open discussion. Discussion with the patient about today’s results should be used to guide any changes that need to be made to the morphine prescription.

If you are still unsure of the significance of this UDT result – and/or if you plan on making significant changes to patient’s management based on this result – then strongly consider incorporating LC-MS results and/or the input of a clinical biochemist prior to discussing with the patient.

DISCLAIMER: The authors of START-IT have done their best to incorporate best-evidence into this tool. Beta-testing has been done to ensure accuracy however errors can occur. If you do note an erroneous interpretation, please notify us at and we will work to correct the problem. Furthermore, the authors take no responsibility for the actions you take based on START-IT reports– clinical judgement and your own verification of what results mean supersede any recommendations given here.

Explanations of each of these sections are as follows:

  • Overall interpretation: The overall interpretation is expected, unexpected or equivocal with a brief explanation about each medication/drug relevant for this test. A recommendation is given on how important confirmatory testing is for this specific test, largely based on degree of discordance between patient’s reports and test results, but also addressing general limitations of IA.
  • Testing Capabilities and Limitations: Summary of what drugs/medications are detectable, undetectable, or variably detectable based on the specific combination of panels used on today’s test. In general, we used a sensitivity of >= 75% for detectable, <10% for undetectable, and 10-74% for variably detectable.
  • Patient background: relevant demographics, risk category, selection method, etc. This section is mostly relevant for research but also gives a very brief clinical summary of the patient. The initials are there for safety assurance to make sure that the report isn’t accidentally uploaded into the wrong patient’s file.
  • Prescribed medications: relevant medications (opioids, other controlled substances) that patient is prescribed and last dose
  • Self-reported non-prescribed medications: medications/drugs that a patient is not prescribed but wants to self-report.
  • UDT IA Panel results: the actual results of IA testing. This is also a safety assurance in case there’s a bug in START-IT with the interpretation, because you will be able to see the actual test result.
  • Interpretation details: the details of why UDT was interpreted this way, with specific explanations, are provided here.
  • Acting on UDT Result: reminders about the importance of applying this result to your patient, and that UDT informs your clinical picture but is not to be used in isolation. In general, if you’re not sure, verify what results actually mean, and then discuss with your patient in a non-judgmental manner.
  • Disclaimer: a reminder about the limitations of IA in general, and the START-IT tool


START-IT Initial Set-up

If planning on using START-IT with the OCEAN platform, then the process is really three parts: tablet set-up, OCEAN set-up and START-IT set-up.

  • For set-up of hardware and software, if using with OCEAN platform we would recommend you check out Once you have OCEAN and the tablet, simply add the form “START-IT” (publicly available) within OCEAN.
  • It is worth testing on a dummy patient in the EMR chart first to make sure that everything goes according to plan. Little glitches may come up in the beginning. If you are a subscriber to the OCEAN platform, feel free to contact them at:


Note that we are not affiliated in anyway with OCEAN/CognisantMD.


Using START-IT for patient encounters

Now that START-IT is set-up, you’re ready to start applying it with real patients. Check out our website for screenshots that show the latest version and examples of how the tool is applied at the bedside.

Let’s give a brief explanation of how the interpretation algorithm built into START-IT works. START-IT looks at a variety of factors to make a final interpretation of either “Expected”, “Unexpected”, or “Equivocal”. The distinction is based on what medication(s) is prescribed and claims of the last dose, any self-reported non-prescribed medication(s) and claims of the last dose, the detectability of these drugs of interest for this particular IA (detectable, undetectable, variably detectable), the detection window (how long a drug is expected to be detected in the urine) and the actual IA result (positive or negative).

  • An expected result means that the patient’s UDT result is consistent with prescribed medication(s) and last dose, and there were no non-prescribed drugs of concern either self-reported or detected on IA. Note that START-IT never considers cannabis unexpected (even if it is not prescribed).
  • An unexpected result means that either i) there was a drug(s) of concern that was self-reported and/or detected on the IA that is not prescribed, or ii) the prescribed medication that should have been detected based on detectability of the test and patient’s self-report was not detected.
  • An equivocal result describes that grey area where there is a soft concern but not enough to call it unexpected. An example of this would be someone who is prescribed a detectable medication, but claims not to have taken it recently, and not surprisingly it is not found in the urine. This result could be concerning (perhaps the person is diverting his medication), or alternatively perhaps the person has intermittent pain and has good reason to have not been taking it.

There are over 200 different “results” for the specific interpretations – four types of “expected”, nine types of “unexpected”, and five types of “equivocal” based on prescribed vs non-prescribed, detectable vs non-detectable vs variably detectable, last dose reported, and actual IA result. The different types of results can be stored in the OCEAN database (if you subscribed for “studies”) and then analyzed for QI and research.

Comments are customized based on reported sensitivities and specificities in the literature as much as possible. Note that there is significant variability in the literature with these numbers. There is also significant variability with detection windows for the various drugs, and causes of false positives for the various panels. Comments in the START-IT report are meant to acknowledge this ambiguity as much as possible and provide the information that a clinician would want to know when trying to ascertain what a result actually means.

Note that we do not attempt to evaluate urine quantitatively. While it seems logical to look at the strength of the dose and the specific time since that last dose, and evaluate what concentration is in the urine to determine if the person is taking more or less than prescribed/claimed – this approach is limited by significant variability between people and resultant room for error. There is so much variability between people in drug absorption, metabolism and excretion that the practice of quantitative evaluation, as we see it, is over-complicating the issue and potentially leading to drawing erroneous conclusions. That said, on a case-by-base basis, use common sense. If a patient is prescribed oxycodone at a low dose infrequently, and oxycodone panel on IA is negative, then there is a very good chance this is a false negative. As START-IT will tell you, the sensitivity of the oxycodone panel for oxycodone is on-the-whole only about 75% and the test would have lower sensitivities for lower, infrequent doses.

Information for the person administering START-IT:
What to do with the following results:

  • When considering what to do with the result “unexpected” – see Appendix II
  • “Equivocal” – suggest messaging the physician as depending on the specific situation and physician, some
    may be alarmed and want to discuss with the patient while others may not
  • “Expected” – typically no action needed unless your clinic has a system built-in to bill for these in which case physician typically has to acknowledge the interpretation (Appendix IV).


Data Compilation
If using START-IT with the OCEAN platform, then it is possible to easily track the information in START-IT digitally for QI and research purposes. While we do not charge anything to use START-IT (if using on the OCEAN platform then the company that runs it – CognisantMD – does have a monthly fee), we ask that clinics using START-IT give us permission to use their de-identified data for evaluative purposes. Any research will of course be conducted with Research Ethics Board (REB) approval. There are a few simple steps to setting up data tracking.

Please contact us if you will be using START-IT at your clinic so that we can give you the password, and help you get set up with data tracking for your QI purposes (QI does not require REB approval). As mentioned, if we are going to have access to the numbers (for research purposes) then we would need REB approval beforehand and would obtain it as applicable.

Chapter Pearls

  • START-IT can greatly simplify the entire UDT process.
  • START-IT reduces consumption of human resources in numerous ways: no printing or scanning of consents, self-reports or results; no time on the part of the physician looking up specific results for sensitivities, specificities and false positives/ negatives; data on UDT is stored digitally so can be easily analyzed for QI and research purposes.
  • By automating UDT interpretation, START-IT is expected to prevent errors in interpretation and subsequent errors in management.

While START-IT can do most of the work for you when it comes to IA interpretation, the next chapter will – in addition to covering the basics of IA interpretation – cover the nuts and bolts of interpreting a LC-MS result…