PCL-5: administration, scoring and interpretation

Much of what circulates about the PCL-5 prints a severity table — mild, moderate, severe — that the instrument’s owner never published. Its own scoring document says so in one line: “There are currently no empirically derived severity ranges for the PCL-5.” If your source handed you bands, it invented them. This guide covers what the PCL-5 actually supports, why its cutoff does not travel, and how to obtain the Spanish translation, which exists but cannot be downloaded.

What the PCL-5 is and what backs it

The PCL-5 is a 20-item self-report measure of the 20 DSM-5 PTSD symptoms, developed by staff at the US Department of Veterans Affairs’ National Center for PTSD and cited as Weathers et al. (2013).

The owner names three purposes without closing the list — monitoring symptom change, screening for PTSD, making a provisional diagnosis — and its user guide adds using an intake administration to guide treatment planning. The original study (Blevins et al., 2015) reported internal consistency of .94 and test–retest reliability of .82 in 278 trauma-exposed college students, replicated in a further 558 — undergraduates, with no diagnostic interview. Its scores are not interchangeable with the DSM-IV PCL, and DSM-5-TR made no changes to the adult PTSD criteria.

PCL-5 and CAPS-5 are not the same instrument

The PCL-5 is a screen; the diagnostic standard is something else. From the owner: “The gold standard for diagnosing PTSD is a structured clinical interview such as the Clinician-Administered PTSD Scale (CAPS-5),” and “The PCL-5 should not be used as a stand-alone diagnostic tool.”

	PCL-5	CAPS-5
Format	Self-report, 20 items	Structured interview, 30 items
Time	5–10 minutes	45–60 minutes
Access	Free download	Email request self-certifying credentials

The PCL-5’s public-domain status says nothing about the CAPS-5.

Administering and scoring it

It covers the past month, takes 5 to 10 minutes, and is preferably self-administered, though the manual allows reading it to the patient in person or by phone. Interpretation “should be made by a clinician.”

Each item is rated 0 to 4, anchored verbatim as: 0 = Not at all, 1 = A little bit, 2 = Moderately, 3 = Quite a bit, 4 = Extremely. The anchor of 2 is the hinge of the DSM-5 rule, so clinicians working in another language should use the owner’s confirmed translation, not an improvised gloss.

The total runs 0 to 80, and cluster scores can be summed: B (items 1–5), C (6–7), D (8–14) and E (15–20). The alternative is the DSM-5 diagnostic rule: any item rated 2 or higher counts as a symptom endorsed, requiring at least 1 B item, 1 C, 2 D and 2 E. The owner states its preference: “use of a cutoff score tends to produce more reliable results than the DSM-5 diagnostic rule.”

Two conditions get skipped. Without an index event the score cannot be interpreted — the form opens with “Keeping your worst event in mind…” and a line to write it in. And only the one-month window is validated; other timeframes are not, even though the owner also publishes a past-week form.

Four PDFs are free, which the owner groups into three formats: past month, past week, with Criterion A and with LEC-5 and Criterion A. The LEC-5 in the fourth is not another PCL-5: it lists 16 events, with no total score and no psychometrics available.

Interpretation: no bands, and no single cutoff

Start with what does not exist. The PHQ-9 and GAD-7 have severity bands published by their authors; the PCL-5 has none. The mild/moderate/severe tables in circulation are the PHQ-9 template stamped onto an instrument that does not support it: you cannot write “moderate PTSD” on the strength of a score, or call a move from 55 to 42 a band crossing.

The cutoff is not a single number either. The owner’s wording is a range with a warning attached: “Initial research suggests that a PCL-5 cutoff score between 31-33 is indicative of probable PTSD across samples. However, because the population and the purpose of the screening may warrant different cutoff scores, users are encouraged to consider both of these factors when choosing a cutoff score.” It does not settle whether the threshold is 31, 32 or 33, so anyone offering “≥ 33” as the official figure is closing a boundary the owner left open.

The range comes from Bovin et al. (2016), who derived it in a subsample of 140 US veterans against CAPS-5. Since then, a 2023 systematic review (Forkus et al., 64 samples) found recommended cutoffs ranging from 22 to 49 and concluded that a universal cutoff applicable across diverse samples and settings does not exist for the PCL-5. A 2025 German clinical study (Pettrich et al., N = 443, with CAPS-5) then made the cutoff depend on purpose — 34 for clinical use, 38 for prevalence estimation, 42 or 43 for clear-cut cases — and found that 31–33, while yielding acceptable to excellent diagnostic utility, was not optimal for any of them.

None of that licenses you to move the cutoff on your own authority. Choose one on grounds you can state — lower for screening, higher when moving toward a provisional diagnosis — and record which one you used and why.

Measuring change: this is where the useful numbers are

The owner suggests a 10-point decrease as an indicator of response, while noting the evidence for judging clinically meaningful improvement is limited, and cites a post-treatment score below 28 as an indicator of clinically significant change (Marx et al., 2022), the study that also set reliable change at 15 to 18 points depending on sample. Read the fine print: those authors excluded participants who identified as female, and their samples were US veterans. The VA/DoD clinical practice guideline (2023) suggests the PCL-5 for detecting change in symptom severity over time — a weak recommendation — and warns that its evidence samples “consisted predominantly of White male Veterans.”

A change threshold turns a measure into follow-up rather than a snapshot: “reduce PCL-5 by 10 points or more over 12 sessions” is an evaluable objective for your treatment plan, and the series of scores feeds the work of measuring therapy progress.

What a score does not tell you

It is not a diagnosis, however high. Diagnosis rests on clinical interview.
It does not cover risk. Unlike the PHQ-9, the PCL-5 contains no suicidal ideation item; do not assume the topic has been screened. Any risk disclosure calls for an immediate clinical risk assessment under your own protocol and training. This guide will not hand you a step-by-step protocol — that comes from professional guidelines and supervised training — and documenting that assessment has a guide of its own. Make sure the client leaves knowing the crisis line serving their location, plus local emergency services.
It is self-report, anchored to the event the client chose as worst; the CAPS-5 additionally rates overall response validity.

The Spanish evidence: translation is not validation

This matters for any clinician serving Spanish-speaking clients, and it is where the material in circulation is most confidently wrong.

A translation is not a validation. The National Center for PTSD has a Spanish PCL-5 that its own staff edited and confirmed. That is a translation, and its translations page is explicit: “These assessments have not been validated in published research. As such, use of the translated measures requires acknowledgement of this limitation.”

A reliability study is not a diagnostic-accuracy study. The Spanish-language literature is real and growing, and almost all of it is internal consistency and factor structure. The Mexican adaptation by Durón-Figueroa and colleagues (2019) used 204 UNAM distance-learning students, with internal consistency of .97, on a translation of their own. Valencia and colleagues (2025) derived a brief 7-item version in 1,337 help-seeking adults. Nuño-Gutiérrez (2025) studied 744 Mexican adolescents and retained only 12 items. The pattern repeats in a Colombian study and a Peruvian preprint.

The exception, and what is still missing. One Mexican study did measure diagnostic accuracy: Martínez-Levy and colleagues (2021), in mental health service users seen after the September 2017 Mexico City earthquake and with CAPS-5 as the reference, found an optimal cutoff of 27 (sensitivity .96, specificity .73) — below 31–33, in a post-disaster clinical sample, and the paper does not say which translation it used. Beyond that we could not locate another Spanish-language study establishing sensitivity, specificity or a cutoff against a diagnostic interview; none of the studies above administered one, and the 2019 adaptation’s authors said as much about their own work — future studies, they wrote, should explore the scale’s diagnostic characteristics and propose a cutoff suited to assessing PTSD in Mexico. Each Mexican factor-structure study also used a translation of its own, none of them the National Center’s.

That does not make it unusable: use it as a within-client change signal and a screen that triggers an interview, and record the version you administered and its unvalidated status.

Licensing and how to obtain it

The PCL-5 is unusually clean here: “This measure was developed by staff at VA’s National Center for PTSD and is in the public domain and not copyrighted. In accordance with the American Psychological Association’s ethical guidelines, this instrument is intended for use by qualified health professionals and researchers.”

All four PDFs download free, no registration, no permission. They carry no copyright notice and no permissions line, only a version stamp, so you may reproduce the items with the citation.
The Spanish version does not download. You request it by email at ncptsd@va.gov, and using it obliges you to acknowledge in writing that it has not been validated. The page lists only the “Standard version” — apparently the past-month form without Criterion A; confirm that in the same email.
Public domain does not extend to everything else. The CAPS-5 is requested by email, self-certifying graduate-level training, a clinical master’s degree or research credentials, and every instrument in your battery has its own terms — sorted out in the guide to psychological test licensing.

The copies we have seen circulating do not state their provenance and are not the owner’s confirmed translation.

Where the score lives in the note

Most jurisdictions require clinical records to capture the results of the diagnostic aids the clinician relied on, and few name psychometric instruments explicitly; verify the rules where you practice. Mexico’s NOM-004-SSA3-2012 is a workable example: the progress note records the relevant results of diagnostic aids, and numeral 5.16 requires psychological records to be adapted to the nature of the service provided, as Mexico’s clinical-record standard sets out. An administered PCL-5 is such a result.

“Administered the PTSD Checklist for DSM-5 (PCL-5), standard past-month version, English. Total: 42/80 (prior, 12 Mar: 55/80; change: −13 points, above the 10-point threshold suggested as an indicator of response). Cluster scores: B 12, C 6, D 14, E 10. Above the 31–33 range the National Center describes as indicative of probable PTSD. Screening and monitoring measure; not a diagnosis.”

That paragraph drops into the O section of a SOAP note unedited. Recording “PCL-5: 42” and nothing else is the opposite error: no version, no window, no stated cutoff.

Make the number land where it is useful

Administering, scoring and writing it down with its version and its limits is easy with one client and heavy with thirty. That is where gesell.ai helps: it includes built-in validated scales with automatic scoring — PHQ-9 and GAD-7 among them — inside each client’s structured chart, and drafts SOAP, DAP, BIRP or GIRP notes from your session records. Always as decision support you evaluate: the clinical judgment stays yours.