O003. Validity and acceptability of the Swahili pGALS screening tool at a tertiary referral hospital in Kenya

Abstract Background A common reason implicated in the severity of presentation and mortality caused by most rheumatic conditions among children is the delay of presentation to a paediatric rheumatologist. Evidence suggests poor practice of paediatric musculoskeletal clinical skills could be a cause for delay and that history suggestive of musculoskeletal disease alone has poor sensitivity in detecting rheumatological conditions. The paediatrics gait, arms, legs and spine (pGALS) musculoskeletal screening tool was therefore developed to encourage early detection of paediatric rheumatological conditions. It has been validated and translated successfully in other countries. Swahili is well understood and spoken in Kenya and other East African countries and for the pGALS to be incorporated in our setting, we needed to translate and asses its validity in Swahili. The objective is todetermine the sensitivity and specificity of the translated Swahili pGALS screening tool in detecting abnormal joints among children aged 5-16 years. To measure the acceptability of the Swahili pGALS screening tool with reference to time taken performing the examination and the discomfort as perceived by the children and their guardians. Methods We conducted a diagnostic utility study. The three English pGALS screening questions were translated into Swahili according to the WHO standard of translation of a tool and the manoeuvres were performed by a ‘copy me’ approach. The ‘copy me’ approach entailed asking the participants to imitate the research assistants as various manoeuvres were done. Children between the ages of 5–16 years seen at the Aga Khan University Accident and emergency department and family medicine clinic who spoke Swahili and had symptoms suggestive of musculoskeletal conditions were enrolled using systematic random sampling. Children who were already on follow up with the Aga Khan paediatric rheumatologist were excluded. Children had the Swahili pGALS and gold standard review (with the paediatric rheumatologist) on the same day. Guardians were asked to answer questions on acceptability with regards to time taken to perform the tool and level of discomfort. Those with possible rheumatological condition as per the gold standard were offered follow up. Results We enrolled one hundred children with a median age of the children was nine years (IQR 7–11). The sensitivity and specificity of the Swahili pGALS screening tool were 76.8% (CI 63.6–87.0%) and 40.0% (CI 23.9–57.9%). The diagnostic accuracy was 62.7% (CI 52.1–72.1%) and the ROC area was 0.58 (CI 0.48–0.68). The median time taken to perform the Swahili pGALS was 5.0 minutes (IQR 3.5-6.0 minutes). Ninety percent of the guardians found the practice of pGALS to have ‘no’ to ‘some’ discomfort. Conclusions This study had a lower sensitivity and specificity than what was seen with similar studies. Although the sensitivity seen was low, it was still within acceptable range for a screening tool and may still be applicable in Kenya and other Swahili speaking countries if used in conjunction with good rheumatological history to refer early the 76% of children picked up by the screening tool. However, the low specificity implies that relatively large number of false positives would still need to be reviewed by a rheumatologist if the tool is adapted for use. The median time taken to perform the tool was also longer than observed in previous studies but the tool was overall deemed acceptable by majority of the guardians/ parents. Keywords pGALS, Kenya, Swahili, Paediatric, Rheumatology