Reporting aged care quality data monthly, quarterly, or annually? A cross-sectional study of 349 613 residents and...

The monitoring of long-term care facilities (LTCF) using quality indicators facilitates the identification of positive and negative performance, plus desirable benchmarks and targets for those requiring improvement. Choosing a quality indicator reporting period involves a trade-off between statistical power and timeliness of data for action. This trade-off has not been systematically examined for indicators employed in LTCFs. This study aimed to examine the prevalence and ability to detect outliers for seven indicators for four reporting periods (yearly, half-yearly, quarterly, and monthly) to characterize the impact of reporting period on the identification of performance variation.

A cross-sectional study was conducted using data from 349 613 non-Indigenous, LTCF residents (372 647 person-episodes) aged ≥65 years between 1 January 2017 and 31 December 2019 in 2803 Australian LTCFs. Seven quality indicators (antibiotic use, emergency department presentations, antipsychotic use, fall-related hospitalization, fractures, weight loss or malnutrition-related hospitalizations, and delirium or dementia-related hospitalizations) were calculated for each year, half-year, quarter, and month during the study period. Outliers (using 95% and 99.8% confidence limits) detected over different reporting periods are described, and sensitivities (relative to yearly reporting) are reported.

The prevalence of all quality indicators decreased with reduction of reporting period. The sensitivities for the detection of outliers at 95% confidence level for higher (>10%) prevalence indicators (antibiotic use, emergency department presentations, antipsychotic use, fall-related hospitalization) was 0.6 when using half-yearly reporting and 0.2-0.4 for low (<5%) prevalence indicators. For quarterly and monthly reporting, the sensitivities for high prevalence indicators were 0.3-0.5 and 0.1-0.13, respectively, and for low prevalence indicators were 0.1-0.13 and 0-0.07, respectively.

Reducing the LTCF quality indicator reporting period leads to lower ability to detect outliers in LTCF performance, especially for low prevalence indicators. Our results recommend against reporting periods shorter than one year for outlier detec--tion when aiming to identify the relative performance of specific quality indicators.