Most process-of-care quality measures are collected at hospital admission either in the emergency department or the operating room prior to exposure to Contact Isolation. None of the measures that are collected around the time of admission appeared impacted by exposure to Contact Isolation. However, the primary outcome of pneumonia care and smoking cessation advice and counseling were less likely to be achieved if a patient was on Contact Isolation. These elements of care generally occur while patients are on inpatient wards where Contact Isolation is more consistently used. Thus, our findings suggest that Contact Isolation, and not the underlying comorbidity associated with multidrug-resistant organism colonization, hinders delivery of care. If the organism colonization or the associated comorbidity hindered care, we would expect reduced compliance with process of care quality measures consistently throughout the patient admission.
Predictors of complete adherence with process-of-care composite measure differed by measure but were generally worse in patients with markers of chronic or severe illness such as having been admitted in the year prior to admission. Composite process-of care measures in AMI and SCIP were better with admission to specialized ICUs. Better AMI and SCIP process-of-care measures in patients admitted to an ICU was not expected. This may be explained by the fact that, in our hospital, there are ICUs dedicated to cardiac care and surgical care and both AMI and SCIP measures relate to processes that can be protocol driven. Patients with AMI who are admitted to a general ward may be on one of many wards, while those with a clear AMI syndrome are generally transferred from an outside hospital directly to the cardiac catheterization laboratory or the cardiac ICU where patient care managers round with the physician team, incorporating aspects of standard AMI quality measures. Patients with SCIP measures going to the surgical ICU also fit into protocols which likely improved adherence with measures.
The association between increased severity of illness (as measured by all-payer-refined mortality risk score, admission in the prior year or longer length of stay) and worse adherence to quality measures is not surprising given these patients are potentially sicker and therefore less likely to be evaluated and treated in a standard fashion 
. Medically complex patients have been associated with better process-of-care measures; however these studies were over many types of hospitals and not exclusively in a tertiary care center, where the average level of complexity tends to be higher 
Contact Isolation was negatively associated with process-of-care measures for pneumonia care and smoking cessation. Less documentation of smoking cessation advice and counseling was primarily identified in patients with myocardial infarction and pneumonia (AMI4, PNA4) and pneumococcal or influenza vaccination prior to discharge in patients with pneumonia (PNA2, PNA7). Contact Isolation may be a marker of more medically complex patients 
. However, patients on Contact Isolation also have a potential barrier to care. Because of the time required to don gowns and gloves as well as typically being in a private room, isolated patients have approximately half as many health care worker visits as non-isolated patients 
. Patients on Contact Isolation also have longer admissions (in part because of difficulty obtaining long-term care facility placement), which could affect the process related to delivering instruction on smoking cessation and vaccinations 
. The clinical impact of worse performance on these measures may be significant. Smoking cessation education has recently been questioned as a process-of-care quality measure because of difficulty in accurately measuring true delivery of cessation education 
. However, increasing a patient's association between smoking and recent illness event may increase motivation to stop smoking. Vaccinations against Pneumococcus and influenza can improve outcomes in high risk patients 
Those who provide care to hospital in-patients should be aware of potential negative associations with Contact Isolation. Hospital administrators and others involved in quality improvement should consider approaches to improving measures that appear to be affected by Contact Isolation. Future studies are needed to evaluate the impact of Contact Isolation on quality of care across multiple hospitals. These estimates need to be included in comparative effectiveness evaluation of infection prevention interventions that utilize Contact Isolation (e.g. MRSA active detection and isolation). In addition, as quality measures are rapidly changing, future measures that focus on care provided to inpatients should be evaluated for interactions with isolation or other concurrent performance improvement practices.
Our study has several potential limitations. These include being conducted at a single center, which may limit generalizability, as well as using retrospective administrative data. Only one primary outcome was statistically significant (pneumonia care) and differences in individual process measures were secondary outcomes. During the study period the MICU was using Contact Isolation for all patients, without regard to MDR bacteria status. In our analysis those patients without MDR bacteria in the MICU (but not the other 11 ICUs) were treated as not exposed which may have biased our findings towards the null.
As we continue to move towards greater accountability for process measures and greater use of interventions such as Contact Isolation as a means to prevent healthcare-associated infections, we must look carefully for unintended consequences of policy changes 
. During the past five years, compliance with quality measures has dramatically increased with some evidence of correlation with lower mortality 
. Hospital-associated infections have also decreased dramatically with widespread acceptance of the preventability of many infections 
. In order to maintain gains in overall quality of care delivered to an older and increasingly complex population of hospital inpatients, careful attention to unintended consequences of such interventions must be maintained. Interventions could be developed and tested to improve delivery of care for patients in Contact Isolation that retain the benefits of the intervention in preventing transmission of hospital-associated pathogens.