Use an MDADI near 74 to decide who gets a swallow study

Design

retrospective cross-sectional analysis of a prospective dysphagia surveillance programme, thresholds by Youden's J

Population

208 head and neck cancer patients, 470 paired MDADI and videofluoroscopy observations

Primary outcome

MDADI composite thresholds for videofluoroscopy-graded swallowing impairment

Effect

AUC 0.78 to 0.88; optimal threshold 74 to 78, about 74 for DIGEST grade 2 or above; 4-item short form AUC 0.85

Dysphagia surveillance after head and neck cancer treatment runs into a practical problem: videofluoroscopy cannot be done on everyone at every visit, and patient-reported scores have not had a threshold that means anything. This study took 208 patients from a prospectively maintained surveillance programme, contributing 470 paired MD Anderson Dysphagia Inventory and videofluoroscopy observations for the Penetration-Aspiration Scale and 443 for DIGEST grading, with a mean 56-day gap between questionnaire and study.

The composite score correlated inversely with DIGEST (r = -0.63) and with the Penetration-Aspiration Scale (r = -0.50), and discriminated impairment with area under the curve between 0.78 and 0.88. Optimal thresholds by Youden's statistic fell between 74 and 78, with about 74 best identifying DIGEST grade 2 or above - clinically significant impairment. An exploratory four-item short form agreed almost perfectly with the full composite (r = 0.96) and still discriminated DIGEST-defined impairment at AUC 0.85.

That converts a questionnaire people already complete into a triage rule. Score the MDADI at surveillance visits; a composite around or below 74 is the trigger for an instrumental swallow assessment rather than reassurance, and a score comfortably above it supports watching. The four-item short form is the version that will actually get done in a busy clinic, and its performance loss is small. Two cautions: the threshold identifies patients needing evaluation, not patients who are aspirating, and a patient with a high score and a clinical concern - weight loss, recurrent chest infection, a changed voice - still needs the study.

• Administer the MDADI at head and neck cancer surveillance visits and record the composite score
• Refer for instrumental swallow assessment at a composite around 74 or below
• Use the four-item short form where the full questionnaire will not be completed
• Override a reassuring score on clinical grounds - weight loss, chest infections, voice change
• Audit how many of your referred patients actually have DIGEST grade 2 or above, and adjust locally

Why it matters

It turns a questionnaire that was collecting data into a decision rule for who gets a swallow study.

Don't overread it

Retrospective, single-programme and cross-sectional; the short form was derived and tested in the same data, and the threshold flags who needs assessment rather than who is aspirating.

The statistics, in plain English

Area under the curve of 0.78 to 0.88 is good but not definitive discrimination - at a threshold of 74 a meaningful number of patients will be referred unnecessarily and some with impairment will be missed. Correlation of -0.63 with DIGEST means the questionnaire and the swallow study measure related but different things: one is how swallowing feels, the other is what it looks like. The four-item short form was derived in this same dataset, so its performance is optimistic until validated elsewhere.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.