Team Learning for Healthcare Quality Improvement

Abstract

In organized healthcare quality improvement collaboratives (QICs), teams of practitioners from different hospitals exchange information on clinical practices with the aim of improving health outcomes at their own institutions. However, what works in one hospital may not work in others with different local contexts because of nonlinear interactions among various demographics, treatments, and practices. In previous studies of collaborations where the goal is a collective problem solving, teams of diverse individuals have been shown to outperform teams of similar individuals. However, when the purpose of collaboration is knowledge diffusion in complex environments, it is not clear whether team diversity will help or hinder effective learning. In this paper, we first use an agent-based model of QICs to show that teams comprising similar individuals outperform those with more diverse individuals under nearly all conditions, and that this advantage increases with the complexity of the landscape and level of noise in assessing performance. Examination of data from a network of real hospitals provides encouraging evidence of a high degree of similarity in clinical practices, especially within teams of hospitals engaging in QIC teams. However, our model also suggests that groups of similar hospitals could benefit from larger teams and more open sharing of details on clinical outcomes than is currently the norm. To facilitate this, we propose a secure virtual collaboration system that would allow hospitals to efficiently identify potentially better practices in use at other institutions similar to theirs without any institutions having to sacrifice the privacy of their own data. Our results may also have implications for other types of data-driven diffusive learning such as in personalized medicine and evolutionary search in noisy, complex combinatorial optimization problems.

Tunably Rugged Landscapes With Known Maximum and Minimum

“…For example, this study was originally motivated by some of our previous research in comparing search strategies for healthcare improvement [2], [24]. In the context of clinical fitness landscapes, it is not reasonable to assume that all features have only main effects (corresponding to K = 0 in NK landscapes) as there are many known interactions between various practices and/or treatments in the real world (e.g., [25], [26]).…”

“…9). In [2], [24] we used logistic transforms of general parametric interaction models with unknown maxima to model search on clinical fitness lanscapes with varying numbers of second order interactions. While the logistic function successfully bounds the transformed fitnesses to the open interval (0, 1), it also has the side effect of compressing high fitness values to the degree that there is very little difference between the fitnesses of the optimal peak and many suboptimal peaks.…”

“…However, in some applications it would be desirable to have one type of model with tunable ruggedness that could be applied to binary alphabets, integer alphabets, real-valued alphabets, or heterogeneous alphabets. For example, in real clinical fitness landscapes, decision variables can have a variety of arities ranging from binary (e.g., whether or not a certain practice is performed) to real-valued (e.g., the amount or duration of application of a particular treatment) [24].…”