At present, it is difficult for hospitals to assess the care of their multimorbid (MM) patients. Since multimorbidity is by definition individualized by many possible combinations of comorbidities and principal diagnoses, any single hospital cannot determine whether their care was optimal for their MM patients, since hospitals do not have access to closely matched controls from other hospitals that may provide insight into how their own MM patients may have fared at other institutions. Since multimorbid patients are far sicker than other patients, it may also be easy to dismiss individual poor outcomes as merely part of the natural pathway associated with multimorbidity. The goal of this proposal is to aid hospitals in observing how their MM patient outcomes differ from similar patients treated at other institutions. Using new definitions of MM in hospitalized surgical and medical patients that specifically define single, double and triple combinations of comorbidities that comprise qualifying comorbidity sets, this study examines whether hospitals vary in their ability to optimally treat MM patients. Results from this study will be directly actionable as an aid to inform hospitals how they fare with their specific MM patients as compared to how their patients may have fared elsewhere. The project will use Medicare claims from the entire country through the new CMS VRDC (Virtual Data Research Center), thereby facilitating a large control group to make close comparisons for any given hospital's patients, and apply new methods in multivariate matching (Indirect Standardization Matching ISM) developed by the investigators, to examine quality of care for MM patients. Using ISM, we will produce 10 ?copies? (or control patients matched from patients admitted to hospitals outside the index hospital) for each MM patient at an index hospital, allowing for close matching of specific qualifying comorbidity sets, and close examination of how well the hospital treats MM patients compared to the closest matched patients selected from other hospitals. Using multivariate matching, we will also examine hospital characteristics that may be associated with better outcomes in MM patients. The project has 4 aims:
AIM 1 : For every hospital performing study conditions or procedures, determine if their MM patient outcomes are significantly different than matched controls. Two control populations will be utilized, a representative control population and a control population selected from hospitals with a combination of characteristics associated with superior outcomes (derived from AIM 2).
AIM 2 : Identify types of hospitals that have especially good or especially poor outcomes when treating MM patients as compared to other closely matched patients derived from two control populations at other hospitals.
AIM 3 will examine specific patient qualifying comorbidity sets that present the greatest problems for a hospital (or type of hospital). Finally, AIM 4 will develop outcomes reports for hospitals, with the help of the American Hospital Association. Such reports will allow individual hospitals to gain insight, for the first time, into how their MM patient outcomes compare to extremely similar MM controls.
At present, hospitals do not have the ability to assess the care of their patients with multimorbidity. Since multimorbidity includes many different combinations of comorbidities, no single hospital can observe relevant controls that would inform them as to whether their specific multimorbid patients are doing as well as could be expected. The goal of this proposal is to allow hospitals, for the first time, to observe how their multimorbid patients? outcomes differ from very similar patients at other institutions, facilitating more targeted investigations concerning the care provided to this very fragile group of patients.
