The Boston questionnaire does not return one score, it returns two, and a Rasch analysis in 600 hands establishes that its nineteen items cannot be reliably added into a single construct. Its threshold is not single either: it is 1.04 after injection, 1.45 after surgery, and 1.55 in the diabetic patient. The tool below therefore asks for the context.

In 215 patients, no significant relationship was found between electrodiagnosis and these two scores, in mild to moderate forms. The detail is below.

Figures taken from the studies listed at the end of this page, every value carrying its source where it is written. The questionnaire itself is not reproduced.

What the BCTQ measures, and why it has no total

The Boston questionnaire is a self-report instrument made of two scales: a symptom severity scale, eleven items covering pain, numbness and weakness, and a functional status scale, eight items covering everyday manual tasks. Each question offers five answers of increasing severity, scored 1 to 5, and a scale's score is the mean of its items. A subscale therefore reads on the same range as its items, 1 to 5, where 5 is the worst.

The item count deserves a note, because the literature does not agree. An isolated abstract announces nine tasks for the functional scale, which would make twenty items in total; the Rasch analysis counts nineteen. The full text of one trial explicitly settles it: eleven items for symptoms, eight questions for function, that is nineteen, the only compatible count. Check the lines of the version you have in hand nonetheless.

And here is the rule that governs everything else: these two scales are not added together. The Rasch analysis, conducted in 600 hands confirmed by electrodiagnosis, establishes that the nineteen items cannot be reliably combined into a single linear construct, and concludes that a total BCTQ score is not psychometrically valid. The two are always reported separately, which is why the tool at the top of this page displays two.

The same analysis is harsher still with the functional scale, which alone accumulates problems of fit, targeting, differential item functioning and dimensionality. As for the symptom scale, a systematic review explicitly recommends revisiting its content, several works finding more than one factor in it.

What the BCTQ does not say about nerve conduction studies

Here is a result that should change how a report articulates the questionnaire and the electrophysiological examination.

In 215 patients, with and without adjustment for age, sex, body mass index, symptom duration, depression, somatisation and pain catastrophising, no statistically significant relationship appears between electrodiagnostic findings and functional status or symptom severity. A severely abnormal nerve conduction report says nothing about the score the patient sitting opposite will complete, and a high score prejudges no conduction velocity.

Its authors draw the operational conclusion: both must be measured, because they do not measure the same thing.

One caveat must accompany that citation, and it matters. Severe forms of carpal tunnel syndrome were among the exclusion criteria of that work. The dissociation is therefore demonstrated in mild to moderate forms, and has not been tested beyond.

There is no single threshold, there are four

There is no single minimal clinically important difference for this instrument, and using the wrong one turns a success into a failure, or the reverse.

The first of these thresholds is 1.04 point on the symptom scale. It comes from only 28 patients, treated by corticosteroid injection and reviewed at three weeks. No value is published for the functional scale in that work.

After surgery, the requirements rise markedly. Six months after carpal tunnel release, to declare themselves satisfied, 87 non-diabetic patients required 1.45 on symptoms and 1.6 on function. The same gain of 1.2 points is therefore a success after injection and a failure after surgery.

Metabolic status shifts these values further. In the same cohort, the 27 diabetic patients required 1.55 and 2.05, and improved less overall on final scores. Ignoring that line means announcing a failure to a patient who did what their physiology allowed.

Two warnings complete the picture. First, do not confuse the two quantities: the pooled estimates of a review of 34 articles put the minimal detectable change at 0.72 and 0.79, below the minimal important difference of 1.05 and 1.13. A difference can therefore exceed measurement error without the patient perceiving anything. Second, and most awkwardly: in 180 operated patients, it is relative changes, referred to the baseline score, that best match perceived improvement. The threshold should be computed individually, a highly symptomatic patient having to gain more to feel the difference. A single threshold in points is a shortcut the literature contradicts, and the tool on this page says so at every reading.

No validated French version, against fourteen other languages

No validated French version of the BCTQ is indexed in PubMed. The absence was searched for through seven distinct formulations, two of which return strictly no result. It is measured, not assumed.

Fourteen languages do have one: Romanian, Polish, Thai, Serbian, Danish, Bulgarian, Arabic, Dutch, Finnish, Greek, Spanish, Korean, Chinese and Portuguese. Not to be counted with them: the indexed Turkish version validates a short six-item scale derived from the instrument, not the BCTQ itself.

The contrast with what French-speaking teams publish is sharp. Several French works report BCTQ scores, in rehabilitation as in percutaneous surgery: they therefore use a translation whose measurement properties have never been published. And for want of a validated version, no French title is authoritative for this instrument.

Three administration pitfalls

Manufacturing a total score

The nineteen items cannot be reliably added into a single construct, and a total BCTQ score is not psychometrically valid. The two scales are always reported separately.

Reusing the 1.04-point threshold outside its context

It comes from 28 patients treated by injection and reviewed at three weeks. After surgery, 1.45 on symptoms and 1.6 on function are needed in non-diabetic patients, 1.55 and 2.05 in diabetic ones.

Applying the same threshold regardless of baseline

In 180 operated patients, it is relative and not absolute changes that match perceived improvement. A highly symptomatic patient must gain more to perceive the difference.

Frequently asked questions

What does the Boston questionnaire measure?

It rates carpal tunnel syndrome on two separate scales, symptom severity across eleven items and functional status across eight, each from 1 to 5 where 5 is the worst score.

Is there a total BCTQ score?

No. A Rasch analysis in 600 hands establishes that a total score is not psychometrically valid: the two scales are reported separately.

How much change counts?

It depends on the treatment and on the patient. The threshold is 1.04 after injection, 1.45 on symptoms after surgery in non-diabetic patients, and 1.55 in diabetic ones.

Does the BCTQ reflect nerve conduction studies?

No. In 215 patients, no significant relationship was found between electrodiagnosis and these scores, in mild to moderate forms. Both must be measured.

Is there a French version of the BCTQ?

No. Fourteen languages have a validated version, and French is not one of them, even though French teams publish scores from it.

References

8 sources, PMIDs included

1. Levine DW, Simmons BP, Koris MJ, Daltroy LH, Hohl GG, Fossel AH, Katz JN. A self-administered questionnaire for the assessment of severity of symptoms and functional status in carpal tunnel syndrome. J Bone Joint Surg Am 1993;75(11):1585-92. PMID 8245050. The original publication, fixing the instrument's form: two self-administered scales, one of symptom severity and one of functional status, each scored from 1 to 5, where 1 is the best score and 5 the worst. The direction is therefore that of a complaint, unlike the lower-limb function scales in this family.
2. Jerosch-Herold C, Bland JDP, Horton M. Is it time to revisit the Boston Carpal Tunnel Questionnaire? New insights from a Rasch model analysis. Muscle Nerve 2021;63(4):484-9. PMID 33455019. The Rasch analysis in 600 hands confirmed by electrodiagnosis, and the result that governs every reading of this instrument: its 19 items cannot be added reliably into a single linear construct. A total BCTQ score is not psychometrically valid. The functional scale alone accumulates problems of fit, targeting, differential item functioning and dimensionality.
3. Ozer K, Malay S, Toker S, Chung KC. Minimal clinically important difference of carpal tunnel release in diabetic and nondiabetic patients. Plast Reconstr Surg 2013;131(6):1279-85. PMID 23416439. The postoperative thresholds, and the demonstration that metabolic status shifts them. Six months after carpal tunnel release, to declare themselves satisfied, the 87 non-diabetic patients required 1.45 on symptoms and 1.6 on function, when the 27 diabetic ones required 1.55 and 2.05. Diabetic patients also improved less overall on final scores.
4. Ozyürekoğlu T, McCabe SJ, Goldsmith LJ, LaJoie AS. The minimal clinically important difference of the Carpal Tunnel Syndrome Symptom Severity Scale. J Hand Surg Am 2006;31(5):733-8; discussion 739-40. PMID 16713833. The threshold of 1.04 point on the symptom scale, and the context that goes with it: only 28 patients, treated by corticosteroid injection into the carpal tunnel and reviewed at three weeks. This threshold was not established after surgery, and there is no corresponding value for the functional scale in that work.
5. Mehta SP, Weinstock-Zlotnick G, Akland KL, Hanna MM, Workman KJ. Using Carpal Tunnel Questionnaire in clinical practice: A systematic review of its measurement properties. J Hand Ther 2020;33(4):493-506. PMID 32151499. The systematic review of 34 articles, giving the pooled estimates and the distinction practice skips. The minimal detectable change is 0.72 for symptoms and 0.79 for function; the minimal important difference is 1.05 and 1.13. A difference can therefore exceed measurement error without the patient perceiving anything. The same review recommends revisiting the content of the symptom scale, several works finding more than one factor in it.
6. De Kleermaeker FGCM, Boogaarts HD, Meulstee J, Verhagen WIM. Minimal clinically important difference for the Boston Carpal Tunnel Questionnaire: new insights and review of literature. J Hand Surg Eur Vol 2019;44(3):283-9. PMID 30463474. The work that shifts the threshold question. In 180 operated patients reviewed about eight months after release, it is relative changes, referred to the baseline score, and not absolute ones, that best match perceived improvement. Its authors conclude that the threshold must be computed individually from the initial score, the most symptomatic patients having to gain more to perceive a difference. Their abstract publishes no numerical value, and this page attributes none to it.
7. Chan L, Turner JA, Comstock BA, Levenson LM, Hollingworth W, Heagerty PJ, Kliot M, Jarvik JG. The relationship between electrodiagnostic findings and patient symptoms and function in carpal tunnel syndrome. Arch Phys Med Rehabil 2007;88(1):19-24. PMID 17207670. What the questionnaire does not say about nerve conduction studies, and vice versa. In 215 patients, with and without adjustment for age, sex, body mass index, symptom duration, depression, somatisation and pain catastrophising, no statistically significant relationship appears between electrodiagnostic findings and functional status or symptom severity. One caveat must accompany that citation: severe forms were excluded, so the dissociation is only demonstrated in mild to moderate forms.
8. Oh WT, Kang HJ, Koh IH, Jang JY, Choi YR. Morphologic change of nerve and symptom relief are similar after mini-incision and endoscopic carpal tunnel release: a randomized trial. BMC Musculoskelet Disord 2017;18(1):65. PMID 28158978. The reference that settles the item count, and its presence here deserves an explanation. An isolated abstract announces nine tasks for the functional scale, which would give twenty items in total; the Rasch analysis counts nineteen. The full text of this trial explicitly writes eleven items for the symptom scale and eight questions for the functional scale, that is nineteen. It also specifies that each question offers five answers scored 1 to 5 and that each scale's score is the mean of its items, which explains why a subscale reads on the same range as its items.