A modality-choice page, not a fall-prevention page. That exercise improves balance and prevents falls is settled and high-certainty -> Exercise for Preventing Falls in Older Adults; the active ingredient there is safely challenging balance, and water is one place to do it. The question here is narrow and peripheral: for an older adult, does doing that work in water (AE) beat doing it on land (LE) — and when is the aquatic modality worth choosing? Single gold SR+MA (Deng 2024, 29 RCTs).
The honest verdict — AE is roughly equivalent to, not clearly better than, LE
The source’s own headline (“AE offers better benefits than LE”) overstates its data. Of four pooled outcomes, only one robustly favours water once fragility and heterogeneity are accounted for (Deng et al., 2024):
| Outcome | AE vs LE, SMD (95% CI) | After sensitivity analysis | Read |
|---|---|---|---|
| BBS (overall balance) | 1.13 (0.25 to 2.00), I²=94% | 0.82 (0.17 to 1.48) — survives | the one robust AE edge |
| 30 CST (chair-stand) | 2.02 (0.50 to 3.54), I²=96% | 0.58 (-0.20 to 1.36), p=0.15 — flips NULL | fragile; do not rely on |
| TUGT (dynamic balance) | 0.44 (-0.04 to 0.91), I²=85% | NS | no modality difference |
| 6 MWT (walk capacity) | 0.13 (-0.16 to 0.43), I²=62% | NS | no modality difference |
- Only BBS survives the sensitivity analysis as an AE advantage; the chair-stand result collapses to null, and the two dynamic-balance / walking outcomes show no modality difference at all. The SMDs sit on I² = 94-96% heterogeneity, so even the surviving magnitude is not a stable number — report the direction (AE ≥ LE on overall balance), not the point estimate.
- This is weak evidence (confidence: low): a single SR+MA, extreme heterogeneity, three of four outcomes null or fragile, no durability data, and the authors themselves note “there are also studies with the oppo- site conclusion, suggesting that LE may be more effective than AE” (Deng et al., 2024). The defensible claim is equivalence with a possible small overall-balance edge for water, not superiority.
Why the modality can still matter — a route-(e) constraint lever, not a route-(b) effect edge
The decision-change is not “prefer water for the balance effect.” The relative effect is close to a wash. The lever is feasibility and safety for the stratum that cannot train well on land (Deng et al., 2024):
- «AE provides a safer training environment for participants than LE», and «buoy- ancy counteracts the effects of gravity, which reduces the joint burden, alleviates pain, and promotes relaxa- tion» (Deng et al., 2024).
- For a joint-limited, deconditioned, or weight-bearing-painful older adult (knee/hip OA, obesity, high fall-anxiety on uneven ground), buoyancy removes the pain and fall-hazard ceilings that cap land-based dosing — so water lets them do the balance work at all. Under adherence-is-part-of-the- effect, a sustained aquatic programme can beat a land one abandoned for pain. This is the same low-impact-substitution logic as Stationary Cycling for Knee Osteoarthritis, applied to balance. (inferred from Deng et al., 2024)
- The subgroup signals do NOT license effect-modification targeting. AE “won” in different health subgroups on different outcomes (BBS in nervous-system/MSK; TUGT only cardiopulmonary; 30CST in healthy/cardiopulmonary), with no coherent pattern and “insufficient research to confirm this heterogeneity” (Deng et al., 2024). These are the false-positive- prone route-(b) subgroup claims -> Baseline Risk and the Relative-Absolute Split; treat them as hypothesis-generating, not as a map of who benefits more from water.
Where this sits in the ranking — peripheral, by construction
For the healthy, mobile older adult, land exercise is cheaper, more accessible, and equally effective, so this is a low-rank refinement (attention-is-an-anti-signal: water-vs-land is a heavily-discussed small-effect question). It rises only for the specific constraint stratum above, where it is not a marginal gain but the difference between training and not training. It does not transport to the middle-aged CRF/function decision — older-adult balance is a different support-factor profile — and aquatic exercise is relevant there only as one low-impact modality option, not a general recommendation. The loop stays open (R1): no operation here grades this against a realized fall or function trajectory. (inferred from Deng et al., 2024)