Audit & compliance

Every patient is tested twice a week. Can you show why each test was ordered?

Drug testing is one of the most routine things an addiction treatment program does, and one of the easiest to get wrong on paper. This reference covers the full path of a single test, from the order to the claim, and what the record has to show at each step.

Samuel Jean, Co-Founder17 November 202618 min read
The short version
  • A drug test is a clinical service, so it needs an individualized order with a reason, not just a schedule on the wall of the collection room.
  • Screening tests and confirmatory tests answer different questions, and each one needs its own justification in the record.
  • Testing frequency belongs in the treatment plan, tied to the patient's risk and stage of care, and it should change when the clinical picture changes.
  • A result nobody reviewed is a weak record. The chart should show who looked at it, when, and what the team did next.
  • Before billing anything, know which entity performed the test, which entity is allowed to bill it, and what the applicable payer policy says about frequency and panels.

Walk into most residential or IOP programs on a Monday morning and you will find the same routine: a list of names, a stack of cups, a staff member at the bathroom door, and a courier pickup at noon. The routine works. Patients get tested, results come back, and the clinical team has information it needs. The problem shows up months later, when a payer or a program integrity contractor pulls a sample of claims and asks a narrow set of questions. Who ordered this test? Why this patient, on this day, for this panel? Did anyone read the result? Did it change anything? If the honest answer to most of those is 'it was Monday,' the test may have been clinically reasonable and still be hard to defend. This article walks through drug testing as a sequence of decisions, each of which leaves, or fails to leave, a trace in the record.

Why drug testing draws scrutiny

Testing sits at an awkward intersection. It is cheap to collect, sometimes expensive to analyze, easy to repeat, and easy to automate. Those are exactly the characteristics that make a service attractive to bill in volume and attractive to review. Federal oversight bodies and payers have paid sustained attention to laboratory and toxicology billing in substance use treatment, and many payers publish specific coverage policies for drug testing that limit frequency, panel size or the circumstances under which confirmatory testing is covered.

None of that means testing is suspect. Toxicology is a legitimate and often necessary part of addiction treatment. It means the record has to carry the reasoning. A reviewer looking at a test claim is not evaluating whether testing is a good idea in general. They are asking whether this specific test, for this specific patient, was ordered for a clinical reason, performed by someone permitted to perform it, and used in the patient's care.

The useful frame for operators is simple: treat every test as a clinical service with four parts that must each be documented. An order. A collection. A result. A clinical response. Billing is a fifth step that depends on the first four being in place.

Two kinds of test, two different questions

Most programs use two broad categories of testing, and confusing them is the source of a large share of documentation problems.

Presumptive (screening) testing

Presumptive tests, such as point-of-care cups, dipsticks and immunoassay screens run on an analyzer, answer a yes-or-no question about whether a drug class is likely present above a cutoff. They are fast and inexpensive. They can also produce false positives and false negatives, and they generally identify classes rather than specific substances.

Definitive (confirmatory) testing

Definitive tests, typically chromatography with mass spectrometry, identify specific drugs or metabolites and can quantify them. They answer a different question: exactly what is present, and sometimes how much. They cost more, take longer and are the category most often limited by payer policy.

The documentation implication is that each category needs its own reason. A screen might be ordered because the treatment plan calls for monitoring at a set interval. A confirmation might be ordered because the screen was unexpectedly positive, because the patient disputes a result, because the screen cannot detect a substance the clinician is concerned about, or because a prescribed medication needs to be distinguished from an illicit one. 'Reflex everything to confirmation' is not a reason. Neither is 'the lab runs the full panel by default.'

A test the clinician did not choose is a test the clinician cannot explain, and a test nobody can explain is a test you will have trouble keeping paid.

Ask your reference lab how its requisition is configured. Some requisitions default to large definitive panels or automatic reflex testing unless someone unchecks a box. If the program's clinicians have never seen the requisition, the lab's defaults may be making ordering decisions on their behalf.

The order: who, what and why

Every test should trace back to an order from someone authorized under your state's rules and your payer contracts to order laboratory services. The order is where most programs are weakest, because testing often happens on a protocol rather than in response to an individual decision.

What a defensible order contains

  • The patient and the date or schedule. A one-time order, or a recurring order with a defined start and end.
  • The ordering practitioner, identified by name and credential, with a signature or electronic attestation that meets your payer and state requirements.
  • The test type and panel. Presumptive or definitive, and which substances or classes. 'Full panel' is not a description.
  • The clinical reason. Monitoring adherence to a prescribed medication, assessing recent use reported by the patient, evaluating a change in presentation, confirming an unexpected screen, admission baseline.
  • The link to the plan of care. Which problem or objective the testing supports.

Standing orders and protocols

Protocols are not inherently a problem. A program that tests every new admission on day one has a defensible clinical rationale for a baseline. The trouble comes when a protocol substitutes for individualized judgment indefinitely: every patient, same panel, same frequency, for the entire stay, regardless of how they are doing. Many payer coverage policies specifically address blanket or standing orders and expect testing to be individualized. Read the policy for each of your major payers and confirm what it says about standing orders before relying on one.

A workable middle ground is a protocol that sets a default at admission, combined with a required clinician review at each treatment plan review where testing frequency and panel are either confirmed or changed, with the decision documented. The protocol gets the patient tested on day one. The review makes the ongoing testing individual.

A hypothetical worth working through

Say a 24-bed residential program tests every patient twice a week with a presumptive cup and sends every specimen for a definitive panel regardless of the cup result. A patient stays 28 days, is consistently negative on the cups, and is on no controlled medications. At the end of the stay, the chart shows eight cups and eight definitive panels, a single standing order signed at admission, and no note anywhere that discusses a result. Ask yourself which of those sixteen tests you could explain to a reviewer one by one. The baseline, probably. A confirmation of a negative screen for a patient with no clinical concern, much harder. The program did not necessarily do anything clinically harmful. It produced a record that cannot speak for itself.

Frequency belongs in the treatment plan

The treatment plan is where testing frequency should be justified, because it is the document that already ties services to the patient's problems and goals. If the plan says nothing about toxicology, every test is floating free of the clinical rationale that should support it.

Useful plan language connects frequency to something observable: the patient's stage of care, recent use, the medications they are prescribed, their own goals, or the program's level of care. It also says when frequency will be reconsidered. Compare these two versions.

  • Weak: 'Patient will submit to random UDS.' No frequency, no panel, no reason, and phrased as a patient obligation rather than a clinical intervention.
  • Stronger: 'Presumptive urine drug testing twice weekly, randomized days, during the first two weeks of residential care to monitor early abstinence and inform level-of-care decisions; definitive testing when a presumptive result is unexpected or disputed. Frequency to be reviewed at the next plan review.'

The second version gives a reviewer what they need: a reason, a schedule, a rule for confirmatory testing and a point at which the decision will be revisited. When the plan is reviewed, the reviewer should see that testing was actually discussed. A patient who has been stable and negative for weeks may warrant reduced frequency. A patient whose presentation has changed may warrant more. Either way, the change should be a documented decision. For more on writing plan objectives that can be measured, see the treatment plan objectives rewrites.

Collection: observed, timed and attributable

Collection is where the specimen enters the record, and it is where integrity questions start if something later goes wrong. A result is only as good as the evidence that the specimen came from the right patient at the right time and was handled properly.

What the collection record should show

  • Date and time of collection, recorded when it happened rather than reconstructed at the end of a shift.
  • The staff member who collected or observed, identified individually, not as 'staff.'
  • Whether collection was observed, and if so, under what policy and by a staff member of an appropriate gender under that policy.
  • Specimen type (urine, oral fluid or other) and any validity checks performed, such as temperature on a cup.
  • Any irregularity: a refusal, a cold specimen, a dilute result, an inability to provide a specimen, a substituted or adulterated sample.
  • For specimens sent out, the requisition number and chain-of-custody handoff to the courier or lab.

Observed collection and dignity

Observed collection is sometimes clinically warranted and sometimes required by a referral source, but it is intrusive. Your policy should say when it is used, who may observe, how patient dignity is protected, and how the decision to observe is documented. If your program observes every collection by default, be prepared to explain why, and make sure staffing on every shift can actually comply with the policy as written. A policy that requires a same-gender observer and a night shift that cannot provide one is a finding waiting to happen.

Refusals and invalid specimens

A refusal or an invalid specimen is a clinical event, not a missing data point. Document it as what it is, note how the patient was approached, and route it to the clinician the same way you would route a positive result. Programs sometimes record a refusal as a positive by policy; if yours does, the policy should be written, known to patients, and applied consistently, and the record should still say what actually happened.

Who performs the test, and who bills it

This is the section where operational and billing mistakes most often turn into compliance problems, so it is worth being precise.

Laboratory certification

Under the Clinical Laboratory Improvement Amendments (CLIA), facilities that test human specimens for health assessment generally need a CLIA certificate appropriate to the complexity of the testing they perform. Many point-of-care drug screening devices are categorized as waived tests, which require a Certificate of Waiver and adherence to the manufacturer's instructions. Moderate- and high-complexity testing, which typically includes definitive testing, carries more extensive requirements. Confirm the categorization of each device you use and the certificate your site holds against the CMS CLIA program and 42 CFR Part 493. Also check your state, since some states have their own laboratory licensure on top of CLIA.

A practical check: list every test your program performs on site, the device or method, its categorization, and the CLIA certificate number of the location where it is performed. If you operate more than one site, each location's testing needs to line up with that location's certificate.

Who bills which part

When a program runs a point-of-care cup on site and sends the specimen to a reference lab for definitive testing, there are potentially two services, performed by two different entities. The general rule is that the entity that performed a test bills for it, subject to payer policy and the arrangements in your contracts. Problems arise when:

  • The program and the reference lab both bill for the same definitive test.
  • The program bills for a definitive test it did not perform, under an arrangement nobody reviewed against payer rules.
  • The program bills a presumptive test on the same day the lab bills one for the same specimen.
  • Results from a test the program did not bill for are still used to justify a confirmatory test someone else billed, with no order on file at either end.

Arrangements between treatment programs and laboratories have been an area of federal enforcement attention, including concerns about payments, free services or staff provided by labs to referring programs. If your lab provides anything of value beyond the testing itself, such as collection staff, supplies beyond what is needed for the specimens it receives, or software, have counsel review the arrangement. This is not a question to settle on a sales call.

Payer policy is the rulebook that actually applies

Coverage for drug testing varies widely by payer. Some limit the number of presumptive or definitive tests in a period, some limit panel size, some require a documented reason for every definitive test, and some bundle presumptive testing into a per-diem or program rate so that it cannot be billed separately at all. Medicare contractors publish local coverage determinations for drug testing; commercial plans and Medicaid managed care plans publish their own policies. Pull the current policy for each of your top payers, note the frequency and documentation rules, and confirm whether testing is billable separately under your contract or included in a bundled rate. If it is bundled, billing it separately is not a gray area.

Check the bundle before you bill the cup

If a payer reimburses your PHP, IOP or residential day at a per-diem or case rate, read the contract to see whether laboratory and toxicology services are included in that rate. Programs sometimes bill tests separately for years before anyone reads that clause, and the overpayment exposure grows every month it goes unnoticed. Confirm against the applicable contract and the payer's billing manual.

Results: received, reviewed and documented

A result that lands in a fax tray, or in a lab portal nobody logs into, has not entered the clinical record in any meaningful way. Reviewers look for evidence that someone qualified looked at it. Clinicians need it to make decisions. Both needs are served by the same workflow.

What review looks like on paper

  • The result is filed to the correct patient and episode, with the collection date and the result date both visible.
  • A named clinician acknowledges the result, with a timestamp.
  • Unexpected results, meaning a positive for a substance not prescribed or a negative for a medication that is prescribed, are flagged for the treating clinician rather than filed silently.
  • The interpretation is documented in a note, not just implied by the acknowledgment.
  • Where a confirmatory test was ordered, the note explains why, and the confirmation result is reviewed in turn.

Set an expectation for how quickly results should be reviewed after receipt, and measure it. If results routinely sit unreviewed for days, the testing is not informing care in the way the plan says it does. A weekly report of unacknowledged results by clinician is cheap to produce and quickly shows where the workflow is breaking.

Expected negatives and expected positives

A negative result for a patient on buprenorphine is not good news; it may mean the medication is not being taken. A positive for a benzodiazepine in a patient with a valid prescription is expected. Interpretation depends on the medication list, so the reviewing clinician needs to see the result next to the patient's current prescribed and administered medications, not in a separate system. This is also where definitive testing often has its clearest justification: distinguishing a prescribed substance from a non-prescribed one in the same class.

The clinical response: what changed because of the result

The most persuasive evidence that testing was medically necessary is evidence that it was used. A reviewer who sees a positive result followed by a note discussing it with the patient, a treatment plan adjustment and a change in testing frequency sees a service that did clinical work. A reviewer who sees twenty results and no reference to any of them sees a billing pattern.

Documenting the conversation

When a result is unexpected, the progress note should record that it was discussed with the patient, the patient's account in their own words where relevant, the clinician's assessment, and the plan. That plan might include increased support, a medication review, a change in level of care, or simply continued monitoring. Avoid documenting results in punitive language. Notes that read as disciplinary rather than clinical make it harder to argue the test served a treatment purpose, and they are not good care.

Closing the loop in the plan

If a result leads to a change, the treatment plan should reflect it at the next review, or sooner if your program's policy requires an update on a significant change. If testing frequency increases, say why. If it decreases because the patient has been stable, say that too. Over a stay, the plan should read as a running account of how testing informed decisions.

Results and 42 CFR Part 2

Drug test results held by a federally assisted substance use disorder program are generally Part 2 records, which means they carry the consent and disclosure protections of 42 CFR Part 2 on top of HIPAA. Test results tend to be requested by parties outside the treatment relationship more often than almost any other part of the chart, which makes this an area where programs over-disclose without realising it.

  • Reference laboratories. A lab performing testing for the program typically receives information under a qualified service organization agreement. Confirm you have one in place with every lab you send specimens to, and that it says what Part 2 requires.
  • Courts, probation and drug courts. Disclosures to the criminal justice system require a consent that meets Part 2's requirements, or a court order meeting the regulation's specific standards. A subpoena alone does not authorize disclosure.
  • Employers, family members and referral sources. Each needs a valid Part 2 consent naming the recipient and scoped to what is being released. A general release signed at intake that says 'all records' to 'anyone involved in my care' is unlikely to meet that bar.
  • Payers. Disclosures for payment purposes need to fall within the consent the patient signed. Since the 2024 final rule, a single consent can cover future treatment, payment and health care operations, but it still has to be a properly constructed Part 2 consent.

Programs that report results to referral sources on a routine schedule should audit a sample of those reports against the consents on file. The common failure is a consent that expired, was revoked, or never covered that recipient in the first place. For the consent elements in detail, see 42 CFR Part 2, explained for operators. Confirm current requirements against the regulation itself.

Opioid treatment programs: a separate rulebook

OTPs operate under 42 CFR Part 8, which includes its own requirements for toxicology testing as part of patient care, and testing results feed directly into decisions such as take-home eligibility. OTPs should read Part 8's current text on testing alongside their state opioid treatment authority's rules and their accreditor's standards, since state requirements can be more specific. Part 8 was substantially revised in 2024, so policies written before then deserve a fresh comparison against the current regulation.

Two practical points apply to most OTPs. First, the link between test results and take-home decisions should be documented explicitly: what the clinician reviewed, how the result factored into the decision, and who made it. Second, because methadone and buprenorphine are both prescribed and tested for, interpretation depends on an accurate, reconciled medication record. A test result read against an out-of-date medication list is a misinterpretation waiting to happen.

Common billing pitfalls, in one list

Most toxicology billing problems in behavioral health fall into a small number of patterns. Use this as a checklist against a sample of your own claims.

  1. No individual order on file for the date of service, or only a standing order signed months earlier with no review since.
  2. Definitive testing without a documented reason, particularly when every specimen is confirmed regardless of the screen.
  3. Panel size not justified, where a large panel is billed but the clinical concern involved only a few substances.
  4. Frequency above payer limits, where the treatment plan and the payer policy were never compared.
  5. Duplicate billing between the program and a reference lab for the same specimen.
  6. Billing separately for testing included in a bundled rate under a per-diem, case-rate or program contract.
  7. Testing performed outside the scope of the site's CLIA certificate, or at a site with no certificate on file.
  8. Results never reviewed, so the record cannot show the test informed care.
  9. Date mismatches, where the date of service on the claim does not match the collection date in the chart.
  10. Ordering practitioner not eligible under the payer's rules or not enrolled with that payer on the date of service.

Running an internal toxicology review

You do not need a consultant to find out where your testing records stand. A small, repeatable internal review will tell you more than a one-time project. Something like the following works for most programs.

  1. Each month, pull a sample of patients who had at least one billed drug test. Keep the sample small enough that it actually gets done.
  2. For each test in the sample, trace it end to end: order, plan reference, collection record, performing entity and certificate, result, review acknowledgment, clinical note, claim.
  3. Score each step as present, partial or missing. Do not score clinical quality at this stage, only whether the record exists.
  4. Compare billed frequency and panels against the current written policy for that patient's payer.
  5. Check every outside disclosure of results in the sample against the consent on file.
  6. Write down the two or three patterns that showed up most, assign each one an owner, and check the same patterns next month.

The point of the review is not to grade staff. It is to find where the workflow lets a step fall out, so you can fix the workflow. If orders are missing because collection staff start testing before the clinician signs, the fix is in scheduling, not in a reminder email. If results go unreviewed because they arrive by fax to a shared inbox, the fix is in how results get routed. For a broader approach to internal reviews that actually change things, see what a monthly internal chart audit should fix.

If the review turns up claims that should not have been billed, such as tests included in a bundled rate, or duplicates with a lab, involve compliance and counsel promptly. Identified overpayments carry obligations of their own, and the time to start working them out is when you find them, not when a payer does.

Where your systems help or hurt

Much of what is described above fails for structural reasons: the order is in one place, the collection log on a clipboard, the result in a lab portal, the medication list in another system, and the claim built by someone who sees none of them. Every handoff is a place where the trace breaks. When orders, results, medications and notes sit on the same chart, the reviewing clinician sees the result against the current medication list, the treatment plan shows the testing rationale, and the claim can be checked against an order that exists. In Unisona Health, toxicology results for MAT and OTP programs are ordered and returned electronically onto the chart beside a reconciled medication list, so the result is interpreted where the prescribing decisions are made.

When a reviewer does ask, the test-by-test trace described above is what you need to hand over: the order, the plan, the collection, the result, the note and the claim, together and in date order. That is the same structure an audit response packet is designed to compile, which is easier to assemble when those records were created on one record in the first place. Whatever system you use, the standard does not change. Every test should be able to explain itself.

Frequently asked questions

Do drug tests in addiction treatment need an individual order?

Every test should trace back to an order from an authorized practitioner that names the patient, the test type and panel, and the clinical reason. Protocols can set a default at admission, but many payer policies expect testing to be individualized, so frequency and panel should be reviewed and documented at each treatment plan review.

What is the difference between presumptive and definitive drug testing?

Presumptive tests, such as point-of-care cups and immunoassay screens, indicate whether a drug class is likely present above a cutoff. Definitive tests, typically chromatography with mass spectrometry, identify specific substances and can quantify them. Each answers a different question, so each needs its own documented reason in the record.

How often should patients be drug tested in treatment?

There is no single universal frequency. Testing frequency should be set in the treatment plan and tied to the patient's stage of care, recent use, prescribed medications and level of care, then reviewed as the clinical picture changes. Confirm frequency limits in each payer's coverage policy, and for OTPs, against 42 CFR Part 8 and state rules.

Does a treatment program need a CLIA certificate for urine drug screens?

Facilities that test human specimens for health assessment generally need a CLIA certificate appropriate to the complexity of the testing. Many point-of-care cups are categorized as waived and require a Certificate of Waiver, while definitive testing usually carries more extensive requirements. Confirm each device's categorization and your site's certificate with the CMS CLIA program.

Can a treatment program bill drug tests separately from a per-diem rate?

It depends on the contract. Some per-diem, case-rate and program contracts include laboratory and toxicology services in the rate, in which case billing tests separately is not permitted. Read the applicable contract and the payer's billing manual before billing any test separately.

Can drug test results be shared with probation or a court?

Results held by a federally assisted substance use disorder program are generally protected by 42 CFR Part 2. Disclosure to the criminal justice system requires a consent that meets Part 2's requirements or a court order meeting the regulation's specific standards; a subpoena alone does not authorize release.

Samuel Jean

Co-Founder at Unisona Health, the behavioral health platform built for the audit. More about us →

Get started

Start today, or take a look first.

Create your workspace in minutes and try the whole platform free for 14 days — no credit card. Or book a one-hour walkthrough first.

Start your free trial