Proactive primary care model for frail older people in New Zealand delays aged-residential care: A quasi-experiment.

BACKGROUND/OBJECTIVES: To determine the effect of a proactive primary care program on acute hospitalization and aged-residential care placement for frail older people.

DESIGN

Controlled before and after, and controlled after only quasi-experimental studies, with a comparison group created via propensity score matching. One-year follow-up.

SETTING

Nine general practices in Auckland, New Zealand.

PARTICIPANTS

Community-dwelling people aged 75 and older identified as at increased risk of hospitalization. One thousand and eighty five patients are compared with 3750 comparison patients matched by propensity score based on known risks.

INTERVENTION

Primary healthcare based, registered nurse-led, comprehensive geriatric assessment, goal-setting, care planning, and regular follow-up. Patients were also provided self-management education, health and social care navigation, and transitional care for hospital discharges. Practices received program support, workforce development, and mentoring of primary healthcare nurses by gerontology nurse specialists.

MEASUREMENTS

Outcomes from routinely collected administrative data. Primary: aged-residential care placement.

SECONDARY OUTCOMES

acute hospitalization, mortality, and other health service utilization.

RESULTS

Aged-residential care placement (odds ratio [OR] 0.66, 95% confidence interval (CI) = 0.48-0.91) and mortality (OR 0.66, 95% CI = 0.49-0.88) were significantly lower over the first year in Kare patients compared with matched controls. There was no difference in acute hospitalization (+0.06 admissions per year, 95% CI = -0.01-0.13). Support service use (allied health therapists and assessment for social support) was increased, and emergency department use decreased.

CONCLUSION

The Kare participants had lower aged-residential care placement and mortality in the first year, but no decrease in acute hospitalization. Because the design is nonexperimental caution is required in interpreting these results.