Effectiveness of a Therapeutic<i>Tai Ji Quan</i>Intervention vs a Multimodal Exercise Intervention to Prevent Falls Among Older Adults at High Risk of Falling

The researchers compared three exercise programs head-to-head in older adults who had already fallen or had mobility problems:

**Tai Ji Quan: Moving for Better Balance (TJQMBB):** A therapeutic adaptation of traditional tai chi. This was not a generic tai chi class. It used modified forms and specific therapeutic movement exercises designed to challenge balance, weight shifting, postural control, and ankle/knee/hip coordination. Participants attended two 60-minute classes per week for 24 weeks.

**Multimodal Exercise (MME):** A conventional "best practice" exercise program that combined balance exercises, aerobic conditioning (walking, stepping), strength training (using resistance bands, body weight), and flexibility/stretching. This was the active comparator — the kind of program a doctor might typically prescribe. Same dose: two 60-minute classes per week for 24 weeks.

**Stretching Exercise (control):** A light stretching program that involved seated and standing stretches. This was the control group — not expected to improve balance or prevent falls. Same dose: two 60-minute classes per week for 24 weeks.

**Primary outcome:** Number of falls (incidence of falls) over the 6-month intervention period. Falls were defined as "unintentionally coming to rest on the ground, floor, or other lower level."

**Secondary outcomes:** Proportion of participants who fell (fallers), number of injurious falls, time to first fall, and changes in physical performance measures (balance, gait speed, lower limb strength).

**Sample size:** 670 participants (randomized from 1,147 screened).

**Age:** Mean 77.7 years (SD 5.6 years; range 70–96).

**Sex:** 65% women (436 women, 234 men).

**Race/ethnicity:** 92.1% White, 4.6% African American, 3.3% other.

**Health status:** All were "high risk" for falls. Inclusion required either: (a) at least one fall in the previous 12 months, OR (b) impaired mobility (scored ≤9 on the Short Physical Performance Battery, a standard test of lower extremity function). Many had multiple chronic conditions.

**Setting:** Community-dwelling (living independently at home, not in nursing homes or assisted living) across 7 urban and suburban cities in Oregon, USA.

**Exclusions:** People with severe cognitive impairment (Mini-Mental State Exam <23), progressive neurological disorders (Parkinson's, multiple sclerosis), severe vision loss, or conditions that would prevent safe exercise participation.

**Falls surveillance:** Participants were given monthly fall calendars to record any falls. Research staff called participants monthly to collect calendar data. If a fall was reported, a structured telephone interview was conducted to gather details (location, activity, injuries, medical care sought).

**Physical performance:** Assessed at baseline and 6 months using:

- **Short Physical Performance Battery (SPPB):** A composite test of balance (standing with feet side-by-side, semi-tandem, tandem), gait speed (4-meter walk), and chair stand time (5 repeated stands from a chair). Score 0–12, higher = better function.

- **Functional Reach:** How far forward a person can lean while standing with feet fixed (measure of dynamic balance).

- **Timed Up-and-Go (TUG):** Time to stand from a chair, walk 3 meters, turn, walk back, and sit down.

- **30-Second Chair Stand:** Number of full stands from a chair in 30 seconds (lower body strength).

**Fall-related injuries:** Classified as "injurious falls" if the participant reported seeking medical attention (doctor visit, emergency department, hospitalization) or reported fractures, sprains, bruises, or cuts that required treatment.

**Adherence:** Class attendance was recorded by instructors. Participants were asked to practice at home (recorded in logs).

**Design:** Single-blind, 3-arm, parallel-design, randomized clinical trial (RCT). This is the gold standard for testing causal effects of an intervention.

**Randomisation:** Participants were randomly assigned (1:1:1 ratio) to TJQMBB, MME, or stretching using a computer-generated random sequence, stratified by study site and sex. This ensures that known and unknown confounding factors (age, health status, fall history) are balanced across groups at baseline.

**Blinding:** This was *single-blind*. The outcome assessors (staff who collected fall data and performed physical performance tests) were blinded to group assignment. Participants and instructors could not be blinded — you cannot hide whether someone is doing tai chi vs. stretching. This is a common limitation in exercise trials. However, the use of objective fall calendars (rather than subjective satisfaction ratings) reduces bias from unblinding.

**Duration:** The intervention lasted 24 weeks (6 months). Falls were tracked during this period. There was no follow-up after the intervention ended — so we don't know if effects persist once people stop classes.

**Statistical approach:** Intention-to-treat (ITT) analysis — all participants were analyzed in the group they were randomized to, regardless of how many classes they actually attended. This preserves the benefits of randomisation and gives a real-world estimate of effectiveness (since some people will always drop out or skip sessions). The primary analysis used negative binomial regression to model fall counts, adjusting for site and sex. Results are reported as incidence rate ratios (IRR) with 95% confidence intervals.

**What this design can prove:** Because of randomisation, blinding of assessors, and ITT analysis, this study can convincingly show that TJQMBB *causes* fewer falls than stretching or MME in this specific population over 6 months. The causal claim is strong.

**What this design cannot prove:**

**Long-term effects:** No data beyond 6 months. We don't know if benefits persist, fade, or require ongoing practice.

**Mechanism:** The study shows *that* TJQMBB works, but not *why* it works better than MME. Is it the specific balance challenges? The weight-shifting? The mind-body component? The study cannot disentangle these.

**Generalizability to other populations:** Results apply to community-dwelling, mostly White, high-risk older adults aged 70+. Not tested in nursing homes, younger adults, or people with specific neurological conditions.

**Dose-response:** Only one dose was tested (2x/week, 60 min). Would 1x/week work? 3x/week? Unknown.

**Superiority over no intervention:** The control group did stretching, not "nothing." So we cannot say how TJQMBB compares to doing absolutely nothing — though the 58% reduction vs. stretching is compelling.

**Major methodological strengths:** Large sample, rigorous randomisation, blinded outcome assessment, ITT analysis, active comparator (MME) representing real-world alternatives, high retention (86% completed the trial).

**Major methodological weaknesses:** No blinding of participants or instructors, no long-term follow-up, relatively homogeneous sample (mostly White), reliance on self-reported falls (though monthly calls improve accuracy), and the stretching control group may not be a true "no effect" control (stretching might have some minor benefits or harms).

**Primary outcome — Fall incidence (6 months):**

**TJQMBB group:** 152 falls among 85 individuals (out of ~223 per group).

**MME group:** 218 falls among 112 individuals.

**Stretching group:** 363 falls among 127 individuals.

**TJQMBB vs. Stretching:** Falls reduced by 58% (IRR 0.42; 95% CI 0.31–0.56; p < 0.001). This means the tai chi group had less than half the fall rate of the stretching group.

**MME vs. Stretching:** Falls reduced by 40% (IRR 0.60; 95% CI 0.45–0.80; p = 0.001). MME was also effective, but less so.

**TJQMBB vs. MME:** Falls reduced by 31% (IRR 0.69; 95% CI 0.52–0.94; p = 0.01). The tai chi program was significantly better than the multimodal exercise program.

**Secondary outcomes:**

**Proportion of fallers:** 38% in TJQMBB, 50% in MME, 57% in stretching. Fewer people fell at all in the tai chi group.

**Injurious falls:** 67 injurious falls in TJQMBB, 103 in MME, 153 in stretching. TJQMBB reduced injurious falls by 56% vs. stretching (IRR 0.44; 95% CI 0.29–0.66; p < 0.001) and by 35% vs. MME (IRR 0.65; 95% CI 0.44–0.96; p = 0.03).

**Time to first fall:** TJQMBB had a significantly longer time before the first fall compared to both MME and stretching (hazard ratio 0.54 vs. stretching, p < 0.001).

**Physical performance:** Both TJQMBB and MME improved SPPB scores, gait speed, and chair stand performance compared to stretching. TJQMBB showed greater improvements in balance-specific measures (functional reach, tandem stance time) compared to MME.

**Adherence:** Average class attendance was ~75% across all groups (about 36 of 48 sessions). No significant differences in adherence between groups.

**Adverse events:** No serious exercise-related adverse events (fractures, hospitalizations) were reported. Minor muscle soreness was common across all groups.

Let's translate these numbers into plain English:

If you are a high-risk older adult (aged ~78, with a recent fall or mobility problems) and you do stretching for 6 months, you can expect about **1.6 falls per person** over that period (363 falls / 223 people).

If you do the multimodal exercise program instead, you can expect about **1.0 falls per person** — a meaningful reduction.

If you do the therapeutic tai chi program, you can expect about **0.7 falls per person** — meaning roughly 1 in 3 people in this group will fall once, and most will not fall at all.

The 31% reduction of TJQMBB over MME means that for every 100 people treated with tai chi instead of multimodal exercise, about **15 fewer falls** would occur over 6 months. For every 100 people treated with tai chi instead of stretching, about **55 fewer falls** would occur.

In terms of "number needed to treat" (NNT): To prevent one additional fall compared to stretching, you would need to treat about 2 people with tai chi for 6 months. To prevent one additional fall compared to multimodal exercise, you would need to treat about 7 people.

The effect on injurious falls is particularly striking: tai chi reduced falls that required medical attention by more than half compared to stretching. This is not just about preventing minor stumbles — it's about preventing the kind of fall that sends someone to the emergency room.

**What the authors acknowledge:**

No blinding of participants or instructors (inherent to exercise trials).

No long-term follow-up beyond the 6-month intervention period.

Falls were self-reported, though monthly calls improved accuracy.

The sample was predominantly White and well-educated, limiting generalizability.

The stretching control group may have had some unintended benefits (social interaction, mild activity), potentially underestimating the true effect of tai chi vs. no exercise.

**What a critical reader would note:**

**No dose-response data:** We don't know if 1x/week would work, or if 3x/week would work better. The 2x/week dose was chosen based on prior work, but it's arbitrary.

**No mechanistic data:** The study doesn't measure *why* tai chi works better. Is it the specific balance challenges? The weight-shifting? The cognitive demands? The mind-body component? This limits our ability to optimize the intervention.

**High-risk population only:** These results apply to people who have already fallen or have mobility problems. For healthy older adults without fall risk, the relative benefits might be smaller (or different).

**Home practice not controlled:** Participants were asked to practice at home, but adherence was self-reported and likely variable. Some of the benefit might come from home practice, not just classes.

**Instructor effects:** Different instructors taught different classes. While they were trained and certified, instructor quality could influence outcomes. The study didn't analyze this.

**Multiple comparisons:** Several secondary outcomes were tested. While the primary outcome was pre-specified and significant, some secondary findings could be chance.

**Industry funding:** The study was funded by the National Institute on Aging (NIH) — no industry conflicts. This is a strength, not a limitation, but worth noting.

**Attrition:** 14% of participants dropped out. While ITT analysis mitigates this, dropouts could bias results if they differed systematically between groups.

For someone running their own n=1 experiment (or a small group experiment):

### What to test

**Intervention:** Therapeutic tai chi (specifically, the TJQMBB protocol or a close approximation). This is NOT a generic "tai chi for relaxation" class. It must emphasize:

- Weight shifting from foot to foot

- Controlled, slow movements that challenge single-leg stance

- Ankle, knee, and hip coordination

- Turning and reaching while maintaining balance

- Mindful attention to body position

**Dose:** Two 60-minute sessions per week, plus optional home practice (10–20 minutes on non-class days).

**Comparator options:**

- Option A (best): Compare tai chi to a multimodal exercise program (balance + strength + aerobics) at the same dose.

- Option B (simpler): Compare tai chi to your current routine (or no exercise).

- Option C (minimal): Just do tai chi and track whether your fall risk changes over time.

### Minimum meaningful duration

**At least 12 weeks** to see measurable changes in balance and strength.

**24 weeks (6 months)** to see a clear reduction in falls (based on this study).

Falls are rare events, so you need enough time to observe a meaningful number of them. For an n=1, you might need 6–12 months to get reliable data.

### What to measure

**Primary metric:** Number of falls (defined as unintentionally landing on the ground, floor, or lower surface). Use a daily log or calendar.

**Secondary metrics:**

- **Timed Up-and-Go (TUG):** Time to stand, walk 3m, turn, walk back, sit. Test monthly. A change of ~1–2 seconds is meaningful.

- **30-Second Chair Stand:** Number of full stands in 30 seconds. Test monthly. A change of 2–3 stands is meaningful.

- **Functional Reach:** How far forward you can lean without moving your feet. Test monthly. A change of ~2–3 inches is meaningful.

- **Single-leg stance time:** How long you can stand on one leg (eyes open, hands on hips). Test monthly. A change of 5–10 seconds is meaningful.

- **Fall-related injuries:** Track any injuries from falls (bruises, sprains, fractures, ER visits).

- **Fear of falling:** Use the Falls Efficacy Scale-International (FES-I, 16 items, 16–64 scale). Lower = less fear.

**Frequency of measurement:** Falls daily (log), physical tests monthly, FES-I at baseline and end.

### Key confounds to control for

-