Healthcare use diverged from the general population ten years before rheumatoid arthritis was diagnosed

Design

nationwide register-based group-based multitrajectory modelling with matched population controls

Population

6,824 patients with first-ever seropositive rheumatoid arthritis diagnosed 2016-2020 in Sweden

Primary outcome

patterns of healthcare use across nine components in the ten years before diagnosis

Effect

five trajectories - healthy 60%, high sick leave 18%, chronic pain 10%, high infection 9%, high burden 3%; divergence from controls from 10, 3 and 2 years before diagnosis

Swedish nationwide registers were used to trace the ten years of healthcare before a first rheumatoid arthritis diagnosis in 6,824 people with seropositive disease diagnosed between 2016 and 2020, against matched population controls, using group-based multitrajectory modelling across nine components of care.

Five trajectories emerged. Sixty percent were 'healthy' with low healthcare use throughout. Eighteen percent had high sick leave without musculoskeletal or infection-related inpatient care. Ten percent followed a chronic pain pattern, with heavy analgesic use and frequent musculoskeletal admissions. Nine percent had frequent infection-related inpatient care. Three percent had high use across everything. Crucially, the divergence from matched controls began as early as ten years before diagnosis in three of the five groups, three years before in the high-infection group, and two years before in the high sick leave group.

The interpretation needs care. That 60% of future rheumatoid arthritis patients look like everybody else until close to diagnosis is a caution against expecting to find them early; the trajectories are descriptions of groups, not a screening test, and reverse causation is everywhere - a decade of analgesic use before a rheumatoid arthritis diagnosis may be undiagnosed disease rather than a risk factor for it. The clinically useful residue is the high-infection group. Infection-related admissions rising three years before diagnosis is consistent with immune dysregulation preceding clinical arthritis, and it is a pattern a physician might actually notice.

• Do not expect a pre-clinical signature in most patients - 60% looked like the general population
• Take a specific joint history in a patient with recurrent infections and new musculoskeletal symptoms
• Treat long-standing analgesic use with musculoskeletal admissions as a reason to check inflammatory markers and serology rather than to escalate analgesia
• Read sick leave without a diagnosis as a symptom worth exploring, not an administrative fact
• Remember this is seropositive rheumatoid arthritis in a Swedish system with universal access - the trajectories reflect that healthcare, not a biology that transfers directly

Why it matters

It shows that for a minority of patients, rheumatoid arthritis was visible in the healthcare record years before anyone called it that.

Don't overread it

Descriptive trajectory modelling in registry data - it cannot separate prodromal disease from risk factors, and most future patients showed no distinctive pattern at all.

The statistics, in plain English

Group-based trajectory modelling always produces groups; the number and shape are chosen by the analyst from fit statistics, so the five clusters described here are a useful summary rather than five real types of person. The timing of divergence - ten years, three years, two years - is descriptive and cannot distinguish a cause of rheumatoid arthritis from an early undiagnosed manifestation of it. With 60% of patients in the 'healthy' trajectory, any screening strategy built on these patterns would miss most cases.

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.