OBJECTIVE:

To identify relationships between variations in team structure and risk-adjusted adverse events across 86 teams in a large US home health care organization.

METHODS:

Patient episode data were collected for two 6-month periods, January-June 2002 (N = 54,732 episodes) and January-June 2003 (N = 51,560 episodes). An adverse event was defined as having 1 or more events defined by the Centers for Medicare and Medicaid Services for home health care episodes. Events were risk adjusted using 2 alternative approaches-a Z-score and a Fixed Effects (FE)-score, for each team in each period. These scores (1 for each team in each period) were then regressed against objective measures of team structure.

RESULTS:

The regressions based on the FE-score as the measure of quality performed better than the traditional Z-score. Based on these regressions we find that volume (number of episodes) (P = 0.03), number of weekend visits (P = 0.02), and workload distribution (P = 0.02) were negatively associated with the occurrence of adverse events, whereas higher weekend admissions (P = 0.01) were positively associated with adverse events.

CONCLUSIONS:

Our analysis identifies a number of key team-level organizational variables that influence adverse events in home health care services. We also have demonstrated that the FE-score is a more accurate measure of team quality, as opposed to the Z-score, given that it focuses only on "team attributable" adverse events by isolating and excluding random variation from the quality score.