Minimum Detectable Activity for Contamination Surveys

Q: What is minimum detectable activity (MDA)?

Minimum detectable activity is the smallest true amount of radioactivity that a specific counting system — a defined detector, geometry, efficiency, and count time — will reliably detect above its background, with only a small, pre-set chance of a false negative. It is a property of the measurement setup decided before the sample is counted, not a reading taken from the sample. If a survey's MDA sits above the contamination limit it is supposed to enforce, the survey cannot demonstrate compliance no matter what number it produces.

Q: What is the difference between the critical level and the detection limit?

The critical level (Currie's L_C, also called the decision level or critical value) is a threshold applied after counting: if the net counts exceed L_C, you decide activity is present, accepting a small false-positive probability (typically 5 percent). The detection limit (L_D) is set before counting: it is the true activity large enough that it will exceed L_C most of the time, controlling the false-negative probability (also typically 5 percent). For a well-known background, L_C = 2.33*sqrt(B) counts and L_D = 2.71 + 4.65*sqrt(B) counts. The detection limit expressed in activity units (dpm or Bq) is the minimum detectable activity.

Q: Why are 2.71 and 4.65 the coefficients in the MDA formula?

They come from Currie's 1968 derivation for paired observations (sample and background counted for equal times) with a well-known background and 5 percent false-positive and false-negative risks. The decision-level coefficient is 1.645 times the square root of 2, which equals 2.33. The detection-limit constant 2.71 is 1.645 squared, and 4.65 is twice 2.33. So L_D = 2.71 + 4.65*sqrt(B) is simply k-squared plus 2*L_C. Changing the confidence levels changes the constants.

Q: How do I lower my MDA to meet a required detection limit?

Because a contamination survey is usually background-limited, MDA falls roughly in proportion to one over the square root of the count time, so quadrupling the count time roughly halves the MDA. You can also raise detector efficiency (a scintillation or well counter instead of a bare GM tube), lower the background (better shielding, a lower-background counter, energy discrimination), increase the counted or wiped area, or improve source geometry. Longer counting has diminishing returns, so instrument selection is often the bigger lever.

Q: What MDA does a contamination survey actually need?

The MDA must be comfortably below the limit the survey enforces. For removable contamination, medical programs commonly work to trigger levels on the order of 1,000 dpm per 100 square centimeters (beta-gamma removable) adopted as license conditions, and DOT sets package removable limits of 4 becquerels per square centimeter (about 240 dpm per square centimeter) for beta-gamma emitters. A common target is an MDA at or below 10 to 50 percent of the applicable limit, so that a clean surface reads well under the limit and a contaminated one is clearly flagged.

Q: Is MDA the same as the reading I get during a survey?

No. The reading is the measured result for a particular sample; the MDA is the detection capability of the measurement, computed from background, efficiency, and count time before or independent of that specific sample. A survey result reported as 'less than MDA' means the activity, if any, was too small for this setup to detect — it does not prove the surface is perfectly clean, only that any contamination is below the demonstrated detection floor.

Minimum detectable activity (MDA) is the smallest true amount of radioactivity a given counting setup can reliably distinguish from background, computed from counting statistics before you ever look at the sample result. It is the detection floor of a measurement — set by background, detector efficiency, geometry, and count time — and it is the single number that decides whether a contamination survey can actually demonstrate compliance with the limit it is supposed to enforce. If the MDA sits above the regulatory or license limit, the survey is not capable of the job, regardless of what value it prints. The framework comes from Lloyd Currie's 1968 formalism of the decision level and detection limit, carried into modern radiation-protection practice by NUREG-1507, MARSSIM, ANSI/HPS N13.30, and ISO 11929.¹²³⁴⁵

This post is the counting-statistics theory underneath the surveys DRPS clients run every day. It is the companion to two applied guides — package receipt and wipe testing, which converts a wipe count into removable activity, and survey meter calibration, which makes the instrument's efficiency trustworthy in the first place. Those articles cover how to perform the measurement; this one explains why the measurement can or cannot see the limit at all. DRPS supports detection-limit analysis, instrument selection, and survey-procedure design through its radiation safety officer and medical physicist consulting services.

Introduction

Every contamination survey is a decision made against a noisy background, and MDA is the mathematics of making that decision honestly. A detector counting a clean surface still registers counts — cosmic rays, ambient radionuclides, electronic noise, and nearby sources all contribute a background. Because radioactive decay is random, that background is not a fixed number but a fluctuating one, and any real contamination must be seen through those fluctuations. The question "is there contamination here?" is therefore a statistical hypothesis test, not a direct reading.¹²

Currie's insight, published in Analytical Chemistry in 1968, was to separate that problem into two distinct numbers with two distinct jobs.¹ The critical level ( $L_{C}$ ) is the threshold you apply after counting to decide whether a signal is real, controlling how often you cry wolf on a clean surface. The detection limit ( $L_{D}$ ) is the true activity you must have before counting for the measurement to reliably trip that threshold, controlling how often you miss real contamination. Translate $L_{D}$ into activity units — disintegrations per minute (dpm), becquerels (Bq), or dpm per 100 cm² — and you have the minimum detectable activity (or minimum detectable concentration, MDC, when normalized to area).²³

The practical stakes are direct. A nuclear medicine department, PET radiopharmacy, or radiochemistry lab must show that surfaces, packages, and released items are below contamination limits. If the counting system's MDA is above the limit, a "clean" result is meaningless. Conversely, chasing an MDA far below what regulation requires wastes count time. Getting MDA right is what makes a survey both defensible and efficient. This article gives the definitions, the worked math, the instrument trade-offs, and the regulatory context.

Topic Explanation

Three numbers people conflate

The critical level, the detection limit, and the MDA are three different things, and most survey mistakes come from confusing them. They answer different questions and are used at different moments in a measurement.¹²

Critical level ( $L_{C}$ ) — a threshold on the observed net signal. After you count a sample and subtract background, if the net counts exceed $L_{C}$ , you declare "detected." It governs the false-positive rate $α$ (calling a clean sample contaminated). It is an a posteriori decision rule.
Detection limit ( $L_{D}$ ) — a threshold on the true activity. It is the amount of activity that, if actually present, would produce a net signal exceeding $L_{C}$ with high probability. It governs the false-negative rate $β$ (missing real contamination). It is an a priori capability of the method.
Minimum detectable activity (MDA) — the detection limit $L_{D}$ converted from counts into activity units using efficiency, count time, geometry, and (for surfaces) area and pickup factors. It is the number you compare against a regulatory or license limit.

Background is the enemy, and randomness is why

The reason detection has a limit at all is that background counts follow Poisson statistics. If the mean background over a counting interval is $B$ counts, the standard deviation of a single measurement of it is $B$ . A net signal (gross minus background) can be positive purely by chance even when nothing is there, and it can be near zero even when a little activity is present. The MDA framework is the disciplined way of choosing thresholds that keep both kinds of error acceptably rare.¹⁵

This is why MDA depends on background, efficiency, and time as it does: more background means more noise ( $B$ grows), higher efficiency means each disintegration produces more signal, and longer counting averages down the relative fluctuation. Those three levers — plus geometry and area — are the entire toolkit for meeting a required MDA.

In a medical radioactive-materials program, MDA governs the credibility of nearly every contamination measurement: wipe (removable) surveys of hot labs and package surfaces, direct (fixed) surveys of floors and benches, release surveys of items leaving a restricted area, and the detection limits of bioassay counting such as thyroid screening for I-131 workers. The physics is identical; only the efficiency, geometry, and limit change, which is why the same Currie equations appear in package-receipt procedures, decommissioning surveys, and radiobioassay standards alike.³⁴ Detector choice can shift the achievable MDA substantially: in a published urine-bioassay comparison for I-131 and Cs-137, a NaI(Tl) well counter gave the lowest MDA for total-spectrum counting while a high-purity germanium well detector gave the lowest peak MDA, illustrating that "best" depends on the counting mode.¹¹

Key Technical Principles

The Currie decision level and detection limit

For a paired measurement (sample and background counted for the same time) against a well-known background, Currie's coefficients follow directly from a normal approximation to Poisson statistics with 5 percent false-positive and 5 percent false-negative risks.¹ The standard normal one-sided value for 5 percent is $k = 1.645$ .

The decision (critical) level, in net counts, is:

L_{C} = k 2 B = 1.645 2 B = 2.33 B

The detection limit, in net counts, is:

L_{D} = k^{2} + 2 L_{C} = 2.71 + 4.65 B

where $k^{2} = 1.64 5^{2} = 2.71$ and $4.65 = 2 \times 2.33$ . Here $B$ is the number of background counts accumulated in the counting interval (not a rate), and $2 B$ appears because the net signal's variance includes the uncertainty of both the gross count and the separately measured background. These are the coefficients adopted throughout NUREG-1507, MARSSIM, and instrument-vendor software.¹²³

Symbols used throughout:

$B$ = background counts in the count interval (counts); if $R_{B}$ is the background rate in cpm and $t$ the count time in minutes, $B = R_{B} \cdot t$ .
$L_{C}$ = critical/decision level (net counts) — the detect / don't-detect threshold.
$L_{D}$ = detection limit (net counts) — the true signal reliably above $L_{C}$ .
$ε$ = counting (detector) efficiency, counts per disintegration (dimensionless).
$ε_{i}$ = instrument efficiency; $ε_{s}$ = source efficiency; $ε_{t} = ε_{i} \cdot ε_{s}$ = total efficiency.
$t$ = count time (min); $f$ = wipe pickup (removal) efficiency (dimensionless, conventionally $\approx 0.1$ ).
$A_{p r o b e}$ = active probe area (cm²); $A_{w i p e d}$ = area wiped (cm²).

A quick illustration fixes the idea. If a background count gives $B = 100$ counts, then $B = 10$ , so $L_{C} = 2.33 \times 10 = 23.3$ net counts and $L_{D} = 2.71 + 4.65 \times 10 = 49.2$ net counts. You would declare "detected" for any sample whose net exceeds 23.3 counts; but you can only guarantee reliable detection for true signals producing about 49 net counts or more. The gap between the two is the price of controlling both error types.

From counts to activity: the MDA equation

MDA is $L_{D}$ divided by everything that converts a true disintegration into a counted event. In its most general contamination-survey form:

MDA = \frac{2.71 + 4.65 B}{ε \cdot t \cdot A _{s} \cdot 60}

The denominator collects each conversion factor: $ε$ turns disintegrations into counts, $t$ turns a count total into a rate, $A_{s}$ is any area/pickup factor needed to express the result per unit area, and the factor of 60 converts between per-minute and per-second (that is, between dpm and Bq, since $1 Bq = 60 dpm$ ). In practice health physicists usually work the pieces explicitly rather than lumping them, and they most often express the answer in dpm or dpm/100 cm², dropping the factor of 60 unless SI units are required.²³⁴

Worked example 1 — removable contamination by wipe (NaI well counter)

A technologist wipes a 100 cm² bench area, counts the smear in a NaI(Tl) well counter, and needs to know whether the setup can resolve the 1,000 dpm/100 cm² removable trigger level used as a license condition.⁹¹⁰

Given: background rate $R_{B} = 120$ cpm, count time $t = 2$ min, so $B = 240$ counts; $B = 15.49$ . Counter efficiency $ε = 0.40$ (counts per disintegration for this gamma emitter and geometry). Pickup fraction $f = 0.10$ .

Decision level and detection limit, in counts:

L_{C} = 2.33 240 = 36.1 net counts

L_{D} = 2.71 + 4.65 240 = 74.7 net counts

The activity detection limit on the wipe is $L_{D}$ divided by efficiency and time:

MDA_{w i p e} = \frac{74.7}{0.40 \times 2} = 93.4 dpm on the wipe

Correcting for the fact that a wipe collects only a fraction $f$ of the removable activity, over the 100 cm² wiped:

MDA_{r e m o v ab l e} = \frac{93.4}{0.10} = 934 dpm/100 cm^{2} \approx 15.6 Bq/100 cm^{2}

Because 934 dpm/100 cm² is below the 1,000 dpm/100 cm² trigger, this instrument and count time can just resolve the limit — though with little margin. A prudent program would count longer or use a lower-background counter to create headroom, since an MDA sitting right at the limit leaves no room for background drift.³¹⁰

Worked example 2 — fixed (total) contamination by direct scan (GM pancake)

Direct surveys of non-removable ("fixed") contamination use total efficiency, the product of instrument efficiency and source efficiency, following NUREG-1507 and ISO 7503-1. Source efficiency $ε_{s}$ accounts for self-absorption in the source: NUREG-1507 uses $ε_{s} = 0.5$ for beta emitters with maximum energy above 0.4 MeV and lower values for softer emitters.²⁶

Given a GM pancake probe: instrument efficiency $ε_{i} = 0.15$ , source efficiency $ε_{s} = 0.25$ (a low-energy emitter), so total efficiency $ε_{t} = 0.0375$ ; active probe area $A_{p r o b e} = 15.5$ cm²; background $R_{B} = 60$ cpm; static count $t = 1$ min, so $B = 60$ , $B = 7.75$ .

L_{D} = 2.71 + 4.65 60 = 38.7 net counts

The minimum detectable concentration, normalized to 100 cm²:

MDC = \frac{L _{D}}{ε _{i} \cdot ε _{s} \cdot t \cdot ( A _{p r o b e} /100 )} = \frac{38.7}{0.0375 \times 1 \times 0.155} = 6, 660 dpm/100 cm^{2}

This exceeds the 5,000 dpm/100 cm² total contamination trigger commonly adopted for release surveys.¹⁰ The culprits are the small probe area and low total efficiency — a one-minute count with a bare GM pancake simply cannot resolve the limit for this radionuclide. The fix is more count time, a larger-area or higher-efficiency detector, or both, as the next section shows.

Choosing count time to meet a required MDA

When background dominates, MDA falls approximately in proportion to $1/ t$ , so meeting a target is largely a question of how long you are willing to count. Holding the Example 2 geometry fixed and increasing the count time:

Count time $t$ (min)	Background $B$ (counts)	$L_{D}$ (net counts)	MDC (dpm/100 cm²)
1	60	38.7	6,660
2	120	53.6	4,610
5	300	83.3	2,860
10	600	116.6	2,010

Doubling the count from 1 to 2 minutes brings the MDC below the 5,000 dpm/100 cm² trigger; extending to 10 minutes cuts it to about 2,000. The relationship is not quite exact ( $1/ t$ because of the constant 2.71 term), but the scaling is clear: to halve the MDA you must roughly quadruple the count time. That diminishing return is precisely why instrument selection — a higher-efficiency, lower-background, larger-area detector — is often a better lever than counting longer.²³

The definitions side by side

Quantity	Question it answers	Set when	Units	Formula (well-known B)
Critical level $L_{C}$	Is this specific result real?	After counting	Net counts	$2.33 B$
Detection limit $L_{D}$	What true signal can we reliably see?	Before counting	Net counts	$2.71 + 4.65 B$
Minimum detectable activity	What activity can this setup detect?	Before counting	dpm, Bq, dpm/100 cm²	$L_{D} / (ε t \dots)$

Clinical Impact

Detection limits decide whether a radiation safety program can prove what it claims, and the failure mode is silent. A survey with an MDA above its limit does not throw an error; it returns comfortable "non-detect" results while being physically incapable of seeing the contamination it is meant to catch. An RSO who reports a bench as "clean" using a one-minute GM count that cannot resolve 5,000 dpm/100 cm² has documented nothing defensible, and neither the technologist nor an inspector may notice until the numbers are examined.²³

The consequences are practical: undetected removable contamination spreads and surfaces later as unexplained bioassay results or area-survey hits. In one published incident, initial thyroid screening after a leaking I-125 brachytherapy seed suggested contamination in 12 of 15 workers, but the readings hinged on how background was handled relative to each worker's own minimum detectable activity; more careful measurement using each person's thigh as background showed no true contamination.¹⁴ Get the background and MDA wrong and you can manufacture either false alarms or false reassurance.

MDA also sets the resolution of decommissioning and release surveys, where the stakes are regulatory closure. Peer-reviewed work shows how strongly technique matters: for gamma walkover surveys, realistic detector motion and height variation can raise the minimum detectable concentration by nearly 50 percent versus an idealized flat trajectory, meaning a survey that looks adequate on paper can miss small areas of elevated activity in practice.¹² The measurement layer beneath all of this is instrument trustworthiness — an efficiency or background that is wrong invalidates every MDA computed from it, which is why MDA and the survey meter calibration program are inseparable.

Practical Optimization Tips

A defensible detection-limit program treats MDA as something you design for, verify, and document — not something you compute once and forget.

Compute MDA before the survey, not after

Calculate the MDA for each survey type (wipe, direct, release, bioassay) using the actual instrument background, efficiency, geometry, and count time you will use.
Confirm the MDA is comfortably below the applicable limit — a common target is 10 to 50 percent of the limit — so a clean surface reads well under it and a contaminated one is unambiguous.
Document the MDA calculation as part of the procedure, so any result reported as "< MDA" is anchored to a stated, defensible detection floor.

Attack the right term

Background enters as $B$ : shield the counter, keep sources out of the counting area, and use energy discrimination to reject counts outside the region of interest.
Efficiency enters linearly: match the detector to the emission (thin-window or scintillation for low-energy betas, a well counter for small gamma samples, liquid scintillation for pure betas and alphas).
Count time enters as roughly $1/ t$ : extend it when needed, but recognize the diminishing return and prefer a better detector for large gains.
Geometry and area: maximize solid angle, use a larger probe, and normalize over the appropriate area.

Handle efficiency and source efficiency correctly

For direct (fixed) surveys, use total efficiency $ε_{t} = ε_{i} \times ε_{s}$ per NUREG-1507 and ISO 7503-1; treating instrument efficiency as the whole story understates the MDC.²⁶
Determine efficiency for each radionuclide and geometry you survey — a Cs-137 calibration does not give the right efficiency for a low-energy beta or an alpha.
For wipes, apply and document the pickup fraction (conventionally ~0.1); it is the largest correction and the one most often forgotten.
Periodically re-measure background and efficiency; a rising background silently raises MDA. Where a formal characteristic-limits treatment is required, follow ISO 11929 or ANSI/HPS N13.30.⁴⁵

Common pitfalls to avoid

Reporting "non-detect" without stating the MDA. A non-detect above the limit proves nothing.
Using background rate where counts are required. The formula uses $B$ in counts over the interval; mixing up cpm and counts corrupts $B$ .
Ignoring source efficiency on direct surveys. Omitting $ε_{s}$ can understate MDC by a factor of two to four.
Assuming a single efficiency across isotopes. Efficiency is radionuclide- and energy-specific.
Counting once and never re-checking. Background and efficiency drift; MDA must be reverified.

Regulatory Considerations

MDA is where the survey requirement of the regulations meets the counting statistics of the standards, and a defensible program documents both.⁷²⁴⁵

Key frameworks:

10 CFR 20.1501 requires surveys reasonable to evaluate radiological hazards and requires that instruments used for quantitative measurements be calibrated for the radiation measured. A survey whose MDA cannot resolve the relevant limit does not satisfy the intent of a "reasonable" survey.⁷
NUREG-1507, Revision 1 (2020) is the NRC's primary reference for minimum detectable concentrations with typical survey instruments, including the total-efficiency (instrument × source efficiency) treatment for surface contamination.²
MARSSIM (NUREG-1575, Revision 1) provides the consensus multi-agency methodology for planning and evaluating final-status contamination surveys, using the Currie detection-limit framework to demonstrate compliance with release criteria.³
ANSI/HPS N13.30 (2011, reaffirmed 2017) sets performance criteria for radiobioassay, including how to determine the minimum detectable amount for bioassay counting.⁴
ISO 11929 (2019 series) specifies the determination of the decision threshold, detection limit, and coverage-interval limits with full ISO-GUM uncertainty propagation — the modern, uncertainty-based generalization of Currie's approach, now widely implemented in analysis software.⁵¹³
ISO 7503-1 (2016, confirmed 2021) governs measurement and evaluation of surface contamination and source-efficiency selection.⁶
Contamination limits. DOT sets package removable limits of 4 Bq/cm² (beta-gamma) and 0.4 Bq/cm² (most alpha) in 49 CFR 173.443, and medical programs typically operate to release/trigger levels such as 5,000 dpm/100 cm² total and 1,000 dpm/100 cm² removable beta-gamma, adopted as license conditions and reflected in NRC materials-license guidance.⁸⁹¹⁰

Because contamination surveys in a medical program measure byproduct material, the licensee is regulated by the NRC under 10 CFR Parts 20 and 35 or by an Agreement State with equivalent rules. Of the states DRPS serves, Florida, Maryland, Virginia, California, Nevada, Pennsylvania, New York, and New Jersey are Agreement States administering their own equivalent survey and contamination requirements, while Washington, DC and Delaware are regulated directly by the NRC. Facilities should confirm which authority issues their license and which exact contamination limits, trigger levels, and recordkeeping periods apply. For how these surveys are performed and documented in practice, see package receipt and wipe testing and sealed source leak testing.

Frequently Asked Questions (FAQs)

What is minimum detectable activity (MDA)?

Minimum detectable activity is the smallest true amount of radioactivity that a specific counting system — a defined detector, geometry, efficiency, and count time — will reliably detect above its background, with only a small, pre-set chance of a false negative. It is a property of the measurement setup decided before the sample is counted, not a reading taken from the sample. If a survey's MDA sits above the contamination limit it is supposed to enforce, the survey cannot demonstrate compliance no matter what number it produces.

What is the difference between the critical level and the detection limit?

The critical level (Currie's $L_{C}$ , also called the decision level or critical value) is a threshold applied after counting: if the net counts exceed $L_{C}$ , you decide activity is present, accepting a small false-positive probability (typically 5 percent). The detection limit ( $L_{D}$ ) is set before counting: it is the true activity large enough that it will exceed $L_{C}$ most of the time, controlling the false-negative probability (also typically 5 percent). For a well-known background, $L_{C} = 2.33 B$ counts and $L_{D} = 2.71 + 4.65 B$ counts. The detection limit expressed in activity units (dpm or Bq) is the minimum detectable activity.

Why are 2.71 and 4.65 the coefficients in the MDA formula?

They come from Currie's 1968 derivation for paired observations (sample and background counted for equal times) with a well-known background and 5 percent false-positive and false-negative risks. The decision-level coefficient is 1.645 times the square root of 2, which equals 2.33. The detection-limit constant 2.71 is 1.645 squared, and 4.65 is twice 2.33. So $L_{D} = 2.71 + 4.65 B$ is simply $k^{2}$ plus $2 L_{C}$ . Changing the confidence levels changes the constants.

How do I lower my MDA to meet a required detection limit?

Because a contamination survey is usually background-limited, MDA falls roughly in proportion to one over the square root of the count time, so quadrupling the count time roughly halves the MDA. You can also raise detector efficiency (a scintillation or well counter instead of a bare GM tube), lower the background (better shielding, a lower-background counter, energy discrimination), increase the counted or wiped area, or improve source geometry. Longer counting has diminishing returns, so instrument selection is often the bigger lever.

What MDA does a contamination survey actually need?

The MDA must be comfortably below the limit the survey enforces. For removable contamination, medical programs commonly work to trigger levels on the order of 1,000 dpm per 100 square centimeters (beta-gamma removable) adopted as license conditions, and DOT sets package removable limits of 4 becquerels per square centimeter (about 240 dpm per square centimeter) for beta-gamma emitters. A common target is an MDA at or below 10 to 50 percent of the applicable limit, so that a clean surface reads well under the limit and a contaminated one is clearly flagged.

Is MDA the same as the reading I get during a survey?

No. The reading is the measured result for a particular sample; the MDA is the detection capability of the measurement, computed from background, efficiency, and count time before or independent of that specific sample. A survey result reported as "less than MDA" means the activity, if any, was too small for this setup to detect — it does not prove the surface is perfectly clean, only that any contamination is below the demonstrated detection floor.

Key Takeaways

Three distinct numbers. The critical level decides if a result is real (after counting), the detection limit is the signal you can reliably see (before counting), and the MDA is that detection limit in activity units. Confusing them is the root of most survey errors.¹
The coefficients are derived, not arbitrary. For a well-known background, $L_{C} = 2.33 B$ and $L_{D} = 2.71 + 4.65 B$ , from Currie's 1968 formalism at 5 percent error rates.¹
MDA is $L_{D}$ divided by every conversion factor — efficiency, count time, area, and pickup — and it is what you compare to the limit.²³
Background is the enemy. MDA scales with $B$ and roughly $1/ t$ ; halving the MDA by counting alone means quadrupling the time.²
Use total efficiency for direct surveys. $ε_{t} = ε_{i} \times ε_{s}$ per NUREG-1507 and ISO 7503-1; omitting source efficiency understates the MDC.²⁶
Design the MDA below the limit and document it. A "non-detect" above the applicable limit demonstrates nothing.³⁷

Conclusion

Minimum detectable activity is the honest accounting of what a contamination survey can and cannot see. It is not a reading; it is the statistical floor built from background, efficiency, geometry, and count time, and it decides whether every "clean" survey a facility records actually means anything. The framework is nearly six decades old — Currie's decision level and detection limit — but it remains the backbone of NUREG-1507, MARSSIM, ANSI/HPS N13.30, and ISO 11929, because the underlying problem never changes: detecting a small signal through the random fluctuations of a background.

For a medical radioactive-materials program, the discipline is easy to state and easy to neglect: compute the MDA for each survey before you rely on it, make sure it sits well below the limit it enforces, attack background and efficiency before reaching for longer count times, and document the calculation so a non-detect is anchored to a real detection floor. A program that does this turns its contamination surveys from reassuring rituals into defensible evidence. One that does not may be recording nothing at all — and never know it.

How DRPS Can Help

Diagnostic Radiation Physics Services helps medical and research facilities build contamination-survey programs that are statistically defensible, not just procedurally complete. This includes calculating minimum detectable activities and concentrations for wipe, direct, release, and bioassay counting; determining radionuclide- and geometry-specific instrument, source, and total efficiencies; selecting detectors matched to the required MDA and the emissions of interest; verifying that survey procedures can actually resolve the applicable regulatory and license limits; and applying the Currie, NUREG-1507, MARSSIM, ANSI/HPS N13.30, and ISO 11929 frameworks to your program. This work is part of our radiation safety officer, medical physicist consulting, and radiation safety training services.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware.

A contamination survey is only as trustworthy as its detection limit. DRPS makes sure the number on the page reflects what your instruments can actually see.

Related Resources

References

Currie LA. Limits for qualitative detection and quantitative determination: application to radiochemistry. Anal Chem. 1968;40(3):586-593. doi:10.1021/ac60259a007. pubs.acs.org
U.S. Nuclear Regulatory Commission. NUREG-1507, Revision 1: Minimum Detectable Concentrations with Typical Radiation Survey Instruments for Various Contaminants and Field Conditions. 2020. nrc.gov
U.S. Nuclear Regulatory Commission, EPA, DOE, DOD. NUREG-1575, Revision 1: Multi-Agency Radiation Survey and Site Investigation Manual (MARSSIM). 2000. nrc.gov
American National Standards Institute / Health Physics Society. ANSI/HPS N13.30-2011 (R2017): Performance Criteria for Radiobioassay. 2011 (reaffirmed 2017). hps.org
International Organization for Standardization. ISO 11929-1:2019: Determination of the characteristic limits (decision threshold, detection limit and limits of the coverage interval) for measurements of ionizing radiation — Fundamentals and application — Part 1: Elementary applications. 2019. iso.org
International Organization for Standardization. ISO 7503-1:2016: Measurement of radioactivity — Measurement and evaluation of surface contamination — Part 1: General principles. 2016 (confirmed 2021). iso.org
U.S. Nuclear Regulatory Commission. 10 CFR 20.1501: General (Surveys and Monitoring). ecfr.gov
U.S. Department of Transportation. 49 CFR 173.443: Contamination control. ecfr.gov
U.S. Nuclear Regulatory Commission. NUREG-1556, Volume 9, Revision 3: Consolidated Guidance About Materials Licenses — Program-Specific Guidance About Medical Use Licenses. nrc.gov
U.S. Nuclear Regulatory Commission. HPPOS-071: Control of Radioactively Contaminated Material (surface contamination guidance, 5,000 / 1,000 dpm per 100 cm²). nrc.gov
Burn AG, Haines DK, Khan AJ, et al. Gamma radioactivity detection limits and associated radionuclide intakes study in artificial human urine using sodium-iodide and high-purity germanium detectors. Health Phys. 2023;124(2):106-112. doi:10.1097/HP.0000000000001642. doi.org
King DA, Altic N, Greer C. Minimum detectable concentration as a function of gamma walkover survey technique. Health Phys. 2012;102(Suppl 1):S22-S27. doi:10.1097/HP.0b013e318237e757. doi.org
Zorko B, Korun M, Mora Cañadas JC, et al. Systematic influences of gamma-ray spectrometry data near the decision threshold for radioactivity measurements in the environment. J Environ Radioact. 2016;158-159:119-128. doi:10.1016/j.jenvrad.2016.04.009. doi.org
Caldwell C, Sankreacha R, Lightstone AW, O'Brien P, Matheson C. Contamination during a brachytherapy procedure. Health Phys. 2007;92(2 Suppl):S8-S12. doi:10.1097/01.HP.0000248114.00473.bc. doi.org