Classification & Assessment in Psychiatry · Paper A
Brief Psychiatric Rating Scale (BPRS)
BPRS is the original one-number scorecard psychiatric drug trials have used since 1962: eighteen broad symptom items, each scored 1-7, added into a single total that moves as a patient gets better or worse. It is not a psychosis-only checklist — anxiety, depression, guilt and disorientation sit alongside hallucinationsSeeing, hearing, or otherwise sensing things that are not really there.Read more → and suspiciousness in the same eighteen items — because it was built to catch change across the whole range of presentations a drug trial might see. The exam trap is treating any similar-sounding, similarly-scored scale as BPRS: PANSSPositive and Negative Syndrome Scale, a scale used to score the severity of psychotic symptoms.Read more → shares many of the same item names but splits its result into three separate subscale totals, and that single structural difference, not the vocabulary, is what actually tells them apart.
Open a psychiatric drug trial's results section and you will usually meet the same eighteen-line table before you meet a single patient: an item list running from somatic concern through hostility to hallucinatory behaviour, each scored 1 to 7, added into one number that moves down as the drug works. This is the ghost that haunts almost every antipsychotic trial published for the last sixty years, yet most exam candidates can name its more famous descendant, , without being able to say what actually separates the two. This page builds BPRS from zero, then draws the exact line where it stops being BPRS and becomes something else.
The Brief Psychiatric Rating Scale (BPRS) is a clinician-rated instrument — meaning a trained rater scores it, not the patient themselves — first published by Overall and Gorham in 1962. Its standard form has 18 items (an expanded 24-item version also exists), each scored on a severity anchor (a fixed reference point a rater compares the patient's presentation against) running from 1 (not present) to 7 (extremely severe), based on what the rater observes and is told about over the preceding 2 to 3 days. Every item's score is added together into one combined total: there is no positive/negative split, no subscale structure (a portion of the total reported as its own separate score), just eighteen numbers and their sum.
Here is the trap almost every candidate falls into: assuming a scale this closely tied to antipsychotic trials must be a psychosis-symptom checklist. It is not. Look down the actual item list and you will find anxiety, depressive mood, guilt feelings, somatic concern and disorientation sitting next to suspiciousness, hallucinatory behaviour and unusual thought content. BPRS was built to catch symptom change across the whole range of presentations a psychiatric drug trial might see, not to diagnose psychosis or measure it in isolation — which is exactly why it still gets used, in various updated forms, well outside research.
Start from what a drug trial in 1962 actually needed: a single number, repeatable across raters, that moves in a predictable direction as an experimental drug works or fails. A subscale structure is a refinement that answers a more specific question — which domain improved — and that refinement did not exist until Kay, Fiszbein and Opler built on top of BPRS's own item list twenty-five years later. BPRS's one-total design is not a limitation so much as its whole design goal: maximum breadth, one comparable number, minimum structure standing between the rater and the result.
For example, imagine two trial reports land on your desk. The first says: "participants were rated using an 18-item scale, each item 1-7, yielding one combined severity score." The second says: "participants were rated using a 30-item scale, reported as three subscale totals — positive, negative and general psychopathology — alongside a combined score." Both trials might use several identical item names (hostility, suspiciousness and hallucinatory behaviour appear in both instruments), so item names alone will never tell you which is which. The number of items, and whether the result is reported as one total or three separate subscale totals, is what actually decides it: the first report describes BPRS, the second describes PANSS.
| Looks similar | How it actually differs |
|---|---|
| 30 items built on BPRS's own item set, reported as three separate subscale totals (positive/negative/general) rather than one combined total | |
| SANS / SAPS | Each rates only one symptom pole (negative-only or positive-only) in detail, never a broad general-psychopathology set alongside mood and hostility items |
| CGI | A single 1-7 global judgement with no itemised symptom list underneath it at all |
| YMRS / HAM-D / MADRS | Each is disorder-specific (, or depression only), with no role outside that one symptom cluster |
| HoNOS / GAF | Routine-care, whole-caseload functioning measures completed regardless of diagnosis, not research-trial symptom-severity instruments |
There is no diagnosis to establish with BPRS itself — it is an assessment tool, not a diagnostic criteria set, and it presupposes a diagnosis has already been made by other means (ICD-11 or DSM-5 criteria). What a candidate does need to identify correctly, from a stem, are BPRS's defining structural features: 18 items (or 24 in the expanded form), a 1-7 anchor, a 2-3 day recall window combining observation and interview, and one combined severity total rather than any subscale split. A stem that changes any one of those features — three subscale totals, a 0-100 single number, an 11-item -only set — is describing a different instrument, however similar the surrounding language sounds.
BPRS's own "guideline" is really its evidence base and its practical role: it remains a standard secondary or historical comparator outcome measure in and antipsychotic trials, cited alongside in the research literature, though PANSS has largely replaced it as the primary outcome measure in newer trials because of its subscale structure. In routine NHS clinical care, by contrast, structured outcome monitoring is far more likely to use HoNOS (as part of the Care Programme Approach) than either BPRS or PANSS, since neither trial instrument was designed for whole-caseload, cross-diagnosis routine use. NICE's guidance on schizophrenia and psychosis (CG178) does not mandate a specific symptom rating scale for routine practice, which is why you should expect BPRS and PANSS to appear in exam stems as research instruments, and HoNOS or GAF to appear as routine-care instruments.
Whenever a stem hands you a symptom-scale vignette, resist naming the first familiar scale you recognise. Count what actually distinguishes them: how many items, what anchor, and — the single most decisive clue — whether the result is one combined total (BPRS), three separate subscale totals (), a single domain in isolation (SANS, SAPS, YMRS, HAM-D, MADRS), or a routine whole-caseload functioning score (HoNOS, GAF, CGI). Before you move on, see if you can explain to a friend why sharing item names with PANSS does not make a scale PANSS.
BPRS is the original 18-item, one-combined-total general-psychopathology scale that later rebuilt into three separate subscales — the item count and, above all, whether the answer is one total or three, is what actually tells them apart.