Using preceding hospital admissions to identify children at risk of developing acute rheumatic fever.

AIMS: New Zealand (NZ) Māori and Pacific children have high rates of acute rheumatic fever (ARF). Around 150 new cases arise each year. As part of the national ARF prevention programme, funding is available to improve housing. To obtain maximum benefit from interventions, an effective tool is needed for targeting high-risk children. This study aimed to assess the effectiveness of using hospitalisations for identifying children at risk of subsequent ARF.

METHODS: Three potentially avoidable hospitalisation (PAH) groups were investigated, including diseases thought to be influenced by housing. All were developed using expert opinion or systematic reviews. These were: (i) the PAH conditions associated with the housing environment (PAHHE) group; (ii) the Crowding group; and (iii) the Ministry of Health (MoH) group. We analysed NZ public hospital discharge data (2000-2014). The prevalence of ARF among patients hospitalised in each group was calculated to estimate sensitivity and potential effectiveness. The number needed to screen (NNS) to identify one ARF case was estimated as a measure of efficiency.

RESULTS: Nearly one-third of ARF patients experienced a PAH as children (before developing ARF). Sensitivity for detecting future ARF ranged from <5% (MoH group) to 27% (PAHHE group). NNS ranged from 502.4 (PAHHE) to 707.5 (MoH).

CONCLUSIONS: Because ARF is relatively rare, observing hospitalisations is not particularly efficient for targeting prevention activities for this condition alone. However, housing interventions are likely to improve multiple outcomes; thus, the hospital setting is still useful for identifying at-risk children who could benefit from such programmes.