Background Musculoskeletal pain research in South African is scant. Hence this study to determine the percentages of nurses with musculoskeletal pain at baseline + at follow-up interview a year later (persistent pain) and occurrence of incident pain (no pain at baseline but pain at follow-up); and to identify risk factors associated with persistent pain.

Methods CUPID methodology was used. Nurses in wards with patient handling tasks were selected from two large public-sector hospitals in Gauteng, South Africa. They were interviewed at baseline about the one-month prevalence of site-specific musculoskeletal pain and about a large number of physical, social and psychological variables. Of the 251 baseline nurses, 189 (75%) completed a similar follow-up interview. The baseline nurses and follow-up nurses did not differ by study variables, and having pain at baseline did not influence follow-up participation. Explanatory variables associated with musculoskeletal pain were identified using multivariate logistic regression; a backward hierarchical stepwise approach was used to build final models.

Results One-month prevalences of site-specific pain were consistently higher at follow-up than at baseline. For example, low back pain and neck pain were 33.9% and 42.3% and 20.7% and 27.5% respectively. Persistence of pain was common: 77% of nurses reported persistence of neck pain and 70% knee pain with the lowest being wrist/hand pain at 35%. Nearly half, 39/82, of nurses who did not have any pain during the 1-month prior to baseline, developed pain in the 1-month before the follow-up interview. Regarding site specific pain, incident low back pain had the highest (32.8%) and incident elbow pain had the lowest occurrences (8.5%). Only past injuries [OR 7.0, 95% CI: 3.0–16.2] and ≥2 distressing somatic symptoms were associated with persistent pain.

Conclusions Persistence of pain was common. A large number of potential determinants of persistent pain were considered, but few identified.