Most guidelines for care of chronic medical conditions are disease-specific. However, adherence to such guidelines in the face of medical complexity may be impossible to carry out or may result in unintended consequences. Alternatively, guidelines may remain relevant for complex patients, but require alternate care strategies. These situations are not at all well defined.

We propose a model that describes how an event from a comorbid condition (such as an exacerbation of a coexisting chronic illness;or the diagnosis of a new comorbidity) can affect guideline-concordant care (GCC) for an index condition. Data on approximately 19,000 subjects will be extracted from electronic databases from Kaiser Permanente Colorado, a not-for profit HMO. We will assess the effect of the comorbidities of new-onset depression, new-onset treatable cancer, and exacerbations of chronic pulmonary disease on the index condition of type-2 diabetes. The primary outcome of interest will be Hemoglobin A1c, and secondary outcomes will be blood pressure control and lipid management. Time-varying multivariate regression techniques and latent class analysis will be used to a) assess both the initial and ongoing effects of each comorbid condition on GCC for the index condition over a 6-year time period;and b) identify a subgroup of patients at risk for poor health outcomes due to inability to achieve GCC. Additional analyses will assess medication adherence as a mediator of inability to achieve GCC. Based on results of these analyses, we will propose both specific adaptations to existing diabetes care guidelines, and alternate care management strategies for complex patients with diabetes, that will be amenable to further investigation. We expect that our recommendations will be relevant to other combinations of chronic conditions as well.

Using such a model to describe the attainment of GCC will help identify complex patients at risk for poor health outcomes so that a) guidelines can be appropriately altered to reflect the process-of-care needs of complex patients;b) these populations can be studied in interventional trials to determine the most appropriate care management;and c) we can define the target populations to which newly developed quality measures for complex patient populations may be applied. This information may provide the basis for significantly improving the quality and efficiency of care for a growing and vulnerable segment of the U.S. population.

R21HS017627

2008-09-01

2011-08-31