Design: prospective national cohort study with linear mixed-effects modelling of pre- and post-fracture trajectories and formal interaction testing
Population: 25,873 participants in the China Health and Retirement Longitudinal Study 2011–2020, including 719 with incident hip fracture
Primary outcome: trajectory of activities-of-daily-living difficulty on the six-item Katz index
Effect: immediate increase of 0.580 difficulties (P < .001), rising to 0.908 in the 70–79 group; disability increased during recovery in those aged ≥80; male sex protective (β = −0.185, P = .023); difficulties 25% above pre-fracture level at six years

Hip fracture recovery is usually described as a partial return towards baseline over the first year. A prospective cohort using nine years of the China Health and Retirement Longitudinal Study — 25,873 participants, of whom 719 sustained an incident hip fracture — tracked activities-of-daily-living difficulty before and after the event using the six-item Katz index, with linear mixed-effects modelling.

The immediate loss was 0.580 additional Katz difficulties (P < .001), followed by gradual annual recovery that never completed: at six years, difficulties remained 25% above pre-fracture levels. Age modified the picture substantially (interaction β = 0.033, P < .001), and not in the direction that might be expected. The 70-to-79 group sustained the largest immediate loss, 0.908 difficulties. But those aged 80 and over showed increasing disability during the recovery period — that is, their trajectory after fracture pointed the wrong way rather than upward. Male sex was protective against worsening difficulty (β = −0.185, P = .023).

The clinical consequence is about what rehabilitation is for and how long it lasts. A patient in their seventies has the steepest fall and the most to regain, which makes intensive early rehabilitation an obvious investment. A patient over 80 is on a different trajectory, where the realistic goal is arresting decline and where the standard six-week rehabilitation window is measuring the wrong thing. And in every age group, the honest conversation with a family is that function at one year is not the endpoint — a quarter of the deficit is still there at six.

Tell families that function does not return fully; about a quarter of the deficit persists at six years.
Do not discharge rehabilitation at the point of plateau in an over-80; the trajectory may be downward, not flat.
Reassess function at 6 and 12 months rather than only at discharge from rehabilitation.
Target the 70–79 group for intensive early rehabilitation; they lose the most and have the most to regain.
Record pre-fracture function explicitly at admission, or recovery cannot be judged against anything.

Why it matters

Recovery after hip fracture is discussed as a curve that returns to baseline, and in the oldest patients it does not even point upwards.

The statistics, in plain English

The Katz index runs from 0 to 6, so an immediate increase of 0.580 difficulties is a shift of about half of one activity on average — modest as a mean, but averages hide the patients who lost several. The 25% residual at six years is the number to quote, because it is a proportion of the original deficit rather than an absolute score. Function was self-reported by participants, not measured, and a cohort study cannot separate the fracture's effect from the decline that would have happened anyway in people frail enough to fracture — though the pre-fracture trajectory in the same individuals goes some way towards addressing that.