Coordination and continuity of intensive care unit patient care

Using Klein's model of team coordination, we explored the null hypothesis that intensive care unit (ICU) care coordination unfolds as a linear sequence. Our intent was to generate hypotheses for further research and to provide interim recommendations for improving care coordination. We also explored factors contributing to care coordination (e.g., role responsibilities, support tools).

Although the body of clinical communications research is considerable, few studies address broader team processes in real-world environments; hence, these processes are not well understood.

All bedside communications for 5 ICU patients were recorded for 5 days per patient and were coded using Klein's model. Markov analysis was used to describe the care coordination process. Multivariate contingency table analysis and standardized parameter estimates described important contributing factors, and support tools were described using descriptive statistics.

First-, second-, and third-order Markov analyses show that care coordination does not unfold as a linear sequence; however, Markov diagrams suggest some process structure. Standardized parameter estimates of factors contributing to care coordination were calculated from a statistically significant three-way model (chi2[df= 18] = 36.95, p < .005). Role-based differences depend on context, with important differences in contributions to care coordination occurring within rounds. Tools supported only 48% of conversations.

Three alternative research hypotheses were defined with at least a minimal level of support. Testing these hypotheses present substantial theoretical, methodological, and data analysis challenges.

Within a research framework, recommendations for change could achieve significant gains for understanding and for reducing breakdowns in care coordination.