Key Takeaways

  • The Dynamic Gait Index has 8 walking tasks, each scored 0 to 3, for a total out of 24. The examiner marks the lowest category that applies[1].
  • Most papers use 19 or less as the line linked to falls[3][4][5]. It is a loose line. In a 1997 study of 44 older adults it correctly classified 59% of people with a fall history and 64% of those without[2][10].
  • The printed form says "less than 19", so a score of exactly 19 is where sources disagree. This page follows the papers and counts 19 as at the cutoff[1].
  • Between two tests, a change needs to reach 2.9 points in older adults at risk of falling[7] or 4 points in balance and vestibular disorders[8] before you can call it more than measurement error.

How the DGI Is Scored

Eight tasks, four categories each, one total.

The DGI watches how someone walks while the task keeps changing: speed, head turns, a pivot, obstacles and stairs. Each item is scored from 0 (severe impairment) to 3 (normal), and the examiner marks the lowest category that applies. The eight scores add up to a total out of 24[1].

The form lists a 20 foot walkway, 15 inches wide, a shoebox, two cones and a flight of stairs. It allows about 15 minutes[1].

There is no formula beyond the sum. The calculator adds the scores and keeps the cutoffs next to the total.

The Eight Items

Short descriptions in this page's own words. The official wording is on the free form.

ItemTask4-item
1. Gait on level surfaceWalk 20 feet at your normal speed.Yes
2. Change in gait speedWalk normally, then as fast as you can, then as slowly as you can, about 5 feet each.Yes
3. Gait with horizontal head turnsKeep walking while turning the head right, then left, then back to the centre on command.Yes
4. Gait with vertical head turnsKeep walking while tipping the head up, then down, then back to the centre on command.Yes
5. Gait and pivot turnWalk, then turn as quickly as you can to face the other way and stop on command.No
6. Step over obstacleStep over a shoebox in the path, not around it.No
7. Step around obstaclesWalk around two cones about 6 feet apart, the first on the right and the second on the left.No
8. StairsGo up and down a flight of stairs as you would at home, using the rail if you need it.No

Dynamic Gait Index items

The form is copyrighted, so the cues in the calculator are short summaries, not its wording. If you are scoring a real test, use the original from the University of Missouri Geriatric Examination Tool Kit: Dynamic Gait Index form[1].

One rule is easy to miss. On items 1 to 4 the form puts a person who uses a walking aid in category 2 at best, and on the stairs, using the rail caps the score at 2. The calculator cannot apply those rules, because it only adds the scores you enter[1].

What a Score Means

19 or less is the line most papers use, and it sorts fallers only modestly.

In a 1997 study, 44 community-dwelling adults aged 65 and over were compared by fall history. A score of 19 or less correctly classified 59% of the people who had fallen and 64% of those who had not[2][10].

I could only read that study through a later paper's summary, so the table below names both.

GroupSampleFalls measuredCutoffResult
Community-dwelling older adults[2][10]44 adults aged 65 and over (as summarised by Herman 2009)History of falls19 or lessSensitivity 59%, specificity 64%
Vestibular disorders[3]247 people, mean age 62.5Fall in the previous 6 months19 or lowerOdds ratio 2.58 (95% CI 1.47 to 4.53)
Balance and vestibular disorders[4]123 patients (34 fallers, 89 non-fallers), mean age 62.3At least one fall in the previous 6 months19 or lessSensitivity 68%, specificity 60%, AUC 0.67
Multiple sclerosis[5]81 people who could walk 100 m, mean age 49Fall in the previous 2 months19 or lessSensitivity 87%, positive predictive value 50%

Studies behind the DGI cutoff for falls

Sensitivity is the share of fallers the cutoff catches. Specificity is the share of non-fallers it correctly leaves above the line. The groups and the ways falls were counted differ, so read the results as a range.

Is 19 at risk or not?

The printed form says "< 19/24"[1]. The papers use "19 or less", and Whitney and colleagues write it as "19/24 or less"[3].

I found no source that explains the difference. The calculator follows the papers and flags 19. If your clinic works from the form, treat 19 as the borderline case.

Every row above compares a score with falls that had already happened. Marchetti and Whitney point out that without prospective data you cannot say how well either version of the test predicts a future fall[4].

A high score is not a clean bill of health either. In 326 balance and vestibular patients, close to half had optimal scores at discharge from therapy[8]. Among older adults with functional limitations, the DGI showed a larger ceiling effect than the Berg Balance Scale[9].

What do healthy older adults score? In 278 independent adults aged 70 to 90 who walked without an aid, men averaged 23.3 and women 22.5. A total below 24 was recorded in 65% of the women and 39% of the men[10].

The authors traced most of the women's lower scores to the stairs: 50.6% of women held the rail, against 19.6% of men. People with a fall history averaged 22.5 and those without 23.0[10]. That gap is tiny, which is part of why a single cutoff works poorly in healthy groups.

A study of 318 adults without symptoms found a ceiling up to about age 60 and a rapid decline in the seventies[11]. I only had its abstract, so this page gives no age bands.

Comparing Two Scores

A change smaller than the minimal detectable change could be measurement error.

The minimal detectable change at 95% confidence (MDC95) is the smallest change that measurement error alone is unlikely to produce. It does not tell you whether the change matters in daily life.

GroupMDC95 (points)How it was found
Older adults at risk of falling[7]2.942 community dwellers older than 65 with a history of falls or near falls; the assessments were videotaped and rescored by two experienced physical therapists, so the figure reflects differences in scoring.
Balance or vestibular disorders[8]4326 patients (mean age 60, range 18 to 95) tested at intake and discharge from therapy; the figure is an estimate of the change that exceeds chance variation, and a change of this size went with improvements in self-reported disability.

Minimal detectable change for the DGI

Scores are whole numbers, so with 2.9 a change of 3 or more clears the line, and with 4 it takes 4 or more. The calculator compares your two totals against whichever group you pick.

Another paper looked at how much change counts as important, using a statistical rule instead of patient feedback. In 100 community-dwelling older adults with functional limitations, who followed a 16-week exercise program, the estimates were 1.23 to 1.90 points for the whole sample[9].

They were 1.16 to 1.80 for people who started below 21 and 0.40 to 0.60 for those who started at 21 or higher[9]. These sit below the MDC values, and the authors could not calculate the patient-anchored version. The calculator uses the MDC to judge whether a change is real.

The 4-Item DGI

Items 1 to 4 on their own, for a score out of 12.

Marchetti and Whitney built the short form for people with balance and vestibular disorders. It keeps gait on a level surface, changes in gait speed, and horizontal and vertical head turns, and it needs no equipment because the stairs, shoebox and cones are gone[4].

In their data, 9 of 12 or less was the best cutoff for a fall in the previous 6 months, at 56% sensitivity and 62% specificity. At its best cutoff the 8-item total was more sensitive and the 4-item more specific[4].

The authors suggest a closer look at fall risk factors below 10 of 12, and the Missouri tool kit repeats that line[4][1]. The calculator shows the 4-item score as soon as items 1 to 4 are scored.

Other Gait and Balance Tests

Related tests, briefly.

The Functional Gait Assessment (FGA) is a related walking test scored out of 30. In 35 older adults who logged falls for 6 months, an FGA of 22 or less flagged everyone who fell, with 72% specificity[6]. The authors list the small sample as a limitation.

The Berg Balance Scale is a separate balance test. In the study of older adults with functional limitations mentioned above, it showed a smaller ceiling effect than the DGI[9].

A modified DGI adds time, level of assistance and gait analysis to the scoring, and its MDC95 is 4 points for its own total[1][13]. This page does not score it. It has its own scoring sheet, linked from the tool kit.

How Reliable Is It?

Total scores repeat well. Single items are less steady.

In 44 older adults, interrater reliability was .96 and test-retest reliability .98[2][10]. In 25 people with chronic stroke, total scores had an intraclass correlation of .96 for both, while single items ranged from .55 to .93[12].

Raters agree less on some items. Across six paired ratings of 39 people, item kappas ran from .54 for the pivot turn to .80 for level walking[4].

An earlier study in peripheral vestibular disorders, quoted in the same paper, reported a composite kappa of .64 and a Spearman correlation of .95 for the 8-item test with two raters[4].

The MDC of 2.9 came from two therapists rescoring videos, so some scoring difference is already built into it[7].

Related Tools

For other ways to look at walking and fitness.

Limits

What this page does not do.

  • It does not give the test. Scoring needs an examiner who watches each task, and the calculator only adds what you enter.
  • It does not tell you your risk of falling. The cutoff is a research line with modest accuracy, tested against past falls in specific groups: older adults, vestibular disorders and multiple sclerosis.
  • The earlier version of this tool showed a percentage of the maximum, four named bands and a note that 19 or above is good. I found no source for any of them, so they are gone.
  • I found no verified cutoffs for Parkinson disease, stroke or brain injury, so none appear here. Some websites list them without a source.
  • Several papers cited here were read as abstracts, and the 1997 study was read through a later paper's summary. Numbers from them match what those versions say, not the full original texts.
  • The wording of each item is this page's own summary. Check the official form before you score a real test[1].

Frequently Asked Questions

Sources & References

  1. Dynamic Gait Index (original 8-item test): form, scoring and the 4-item and modified versions. Shumway-Cook A, Woollacott M (instrument); hosted by the Geriatric Examination Tool Kit, University of Missouri. Free form. The instrument is copyrighted; the form cites Motor Control: Theory and Practical Applications (1995) and Herdman SJ, Vestibular Rehabilitation (2000)
  2. Predicting the probability for falls in community-dwelling older adults. Shumway-Cook A, Baldwin M, Polissar NL, Gruber W. Physical Therapy 77(8):812-819 (1997)
  3. The dynamic gait index relates to self-reported fall history in individuals with vestibular dysfunction. Whitney SL, Hudak MT, Marchetti GF. Journal of Vestibular Research 10(2):99-105 (2000)
  4. Construction and validation of the 4-item dynamic gait index. Marchetti GF, Whitney SL. Physical Therapy 86(12):1651-1660 (2006)
  5. Validity of the dynamic gait index in people with multiple sclerosis. Forsberg A, Andreasson M, Nilsagard YE. Physical Therapy 93(10):1369-1376 (2013)
  6. Functional gait assessment: concurrent, discriminative, and predictive validity in community-dwelling older adults. Wrisley DM, Kumar NA. Physical Therapy 90(5):761-773 (2010)
  7. Minimum detectable change of the Berg Balance Scale and Dynamic Gait Index in older persons at risk for falling. Romero S, Bishop MD, Velozo CA, Light K. Journal of Geriatric Physical Therapy 34(3):131-137 (2011)
  8. Responsiveness and minimal detectable change of the dynamic gait index and functional gait index in persons with balance and vestibular disorders. Marchetti GF, Lin CC, Alghadir A, Whitney SL. Journal of Neurologic Physical Therapy 38(2):119-124 (2014)
  9. Sensitivity to change and responsiveness of four balance measures for community-dwelling older adults. Pardasaney PK, Latham NK, Jette AM, Wagenaar RC, Ni P, Slavin MD, Bean JF. Physical Therapy 92(3):388-397 (2012)
  10. The Dynamic Gait Index in healthy older adults: the role of stair climbing, fear of falling and gender. Herman T, Inbar-Borovsky N, Brozgol M, Giladi N, Hausdorff JM. Gait & Posture 29(2):237-241 (2009)
  11. Clinical assessment of balance: normative data, and gender and age effects. Vereeck L, Wuyts F, Truijen S, Van de Heyning P. International Journal of Audiology 47(2):67-75 (2008)
  12. Reliability and validity of the dynamic gait index in persons with chronic stroke. Jonsdottir J, Cattaneo D. Archives of Physical Medicine and Rehabilitation 88(11):1410-1415 (2007)
  13. Expanding the scoring system for the Dynamic Gait Index (the modified DGI). Shumway-Cook A, Taylor CS, Matsuda PN, Studer MT, Whetten BK. Physical Therapy 93(11):1493-1506 (2013)
Manish Kumar
Manish Kumar

Certified Personal Trainer & Sports Nutritionist

NASM-CPTCertified Sports Nutritionist10+ Years Experience500+ Clients Coached

NASM-certified fitness and nutrition coach with over 10 years of hands-on experience helping people build strength, lose fat, and live healthier lives. Specializing in gym-based workouts with a strong focus on lifting technique, biomechanics, and practical exercise science. Through FitLifeRegime, sharing the tools, tips, and insights that have worked for hundreds of clients — helping you start your own fitness journey with confidence and clarity.

Learn more about Manish
NASM Certified Personal Trainer Certificate

NASM Certified

Certified Personal Trainer & Sports Nutritionist from the National Academy of Sports Medicine

Disclaimer: This page adds up Dynamic Gait Index scores that a trained examiner has already observed. It does not run the test, diagnose a condition or predict whether you will fall. Cutoffs come from studies of specific groups. Talk to a doctor or physical therapist about balance, dizziness or falls.