Differences in Thirty Day Healthcare Encounters Following Hospitalization in an Insured and Uninsured Population: A Retrospective Study

Objective: To describe the incidence of 30 day healthcare encounters following an index hospitalization in an uninsured and insured population. Methods: This was a retrospective cohort study conducted at an urban, tertiary care hospital. Patients admitted between 7/1/2010 6/30/2011 that were between the ages of 18-64 were included if they had a financial code for no insurance or commercial insurance provider. The primary outcome assessed between groups was incidence of 30 day healthcare encounter following hospital disharge to either the emergency department or hospital. Results: 5,489 patients were admitted over the study period 2,093 uninsured and 3,396 insured. The number of 30 day healthcare encounters for the uninsured and insured groups were 348 (16.6%) and 566 (16.7%), respectively (p = 0.99). The uninsured population was more likely to be have a 30 day healthcare encounter in the emergency department 237 (11.3%) vs 246 (7.2%), p <0.001. The uninsured group with a 30 day healthcare encounter were younger and more likely reside in a urban zip code. Conditions most frequently cited in 30 day healthcare encounter in the uninsured were abdominal pain, infections, diabetes, chronic pain and angina. Conclusion: This analysis demonstrated no difference in incidence of 30 day healthcare encounters between an insured and uninsured population. However, the uninsured population was more likely to be seen in the emergency department at 30 days.


Introduction
Disease states frequently cited as a cause of 30 day hospital readmission in a Medicare population are heart failure, pneumonia, and chronic obstructive pulmonary disease (COPD) [1]. Poor follow up after hospital discharge, race of the patient and demographics a hospital serves have been linked to an increase rate of patient readmission [2,3].
Strategies to prevent readmission in the aforementioned medical conditions after discharge have been inconsistent in their effectiveness [4]. Strategies effective at preventing readmission are specific to the individual patient and involve early follow up, medication reconciliation and patient education of warning signs of the respective disease states that signifies worsening condition [5,6].
In 2011, it was noted that approximately, 49 million non-elderly patients in the United States are uninsured and comprise 22.0% of the total population aged 19-64 [7]. It is well described that this population has chronic conditions that do not receive adequate follow up care or proper preventive measures [8,9]. This has been subsequently linked to an increase morbidity and mortality for the same chronic conditions when compared to an insured population [10,11].
The readmission rate in the uninsured is poorly described in the medical literature with respect to frequency and reasoning for hospital readmission. Extrapolating interventions targeted at specific co-morbidities cited as a frequent cause of 30 day readmission in a Medicare population may not meet the needs of this population. The goal of this study is to identify the demographic characteristics, 30 day healthcare encounter rate and reason for 30 day healthcare encounter in the uninsured and insured populations at a large, urban, teaching institution.

Methods
This was a retrospective cohort study conducted at a tertiary care, urban, teaching institution in Detroit, MI. All patients hospitalized between the ages of 18 to 64 with an insurance billing code status of commercial insurance or uninsured between 7/1/10 to 6/30/11 were included. The commercially insured population will be referred to as the insured population for the remaining portion of the article. This study was approved by the institutional review board.
Patients admitted to hospital were identified by their insurance status as either uninsured or insured. Patients were identified using Corporate Data Stores, an electronic patient medical database that is utilized by the respective health organization. Baseline characteristics that were collected included age, sex, race and zip code. Zip code was stratified by urban or suburban zip code. The primary outcome was the incidence of 30 day healthcare encounter.
This was collected by identifying all hospital admissions, emergency department visits and hospital discharge dates.
Index hospitalization was defined as a hospital admission that was followed by a separate hospitalization or emergency department visit within 30 days of the initial hospitalization discharge date. An emergency department visit was defined as an acceptance of a patient to the emergency department that was then discharged directly from the emergency department to the community rather than being admitted to the hospital. A 30 day healthcare encounter was defined as either a hospitalization or emergency department visit that was preceded within 30 days of Continuous variables were reported as mean + SD. Categorical data was expressed as frequency and assessed utilizing chi squared test. Students-t test was used to analyze continuous variables. All tests were two-tailed, and a p-value <.05 was determined to represent statistical significance.

Results
During the one year study time period, 5,489 patients were admitted; 2,093 uninsured and 3,396 insured. Average age of the uninsured and insured populations was 42.6 + 11.6 and 48.9 + 12.2 years (p < 0.001), respectively. Male gender was more predominant in the uninsured population when compared to the insured (55.1 vs 41.3%, p < 0.001). The uninsured population were more likely to reside in an urban Zip-code when compared to the insured population (91.2 vs 39.9%, p < 0.001). Additional demographic characteristics can be seen in Table 1.
Overall, there was no difference in the incidence of 30 day healthcare encounter between the uninsured and insured groups (348 (16.6%) vs 566 (16.7%), p = 0.99) (

Discussion
Compared to the insured, the uninsured population were younger, more likely to be male and resided in an urban Zip code. When looking at the insured and uninsured population that had a 30 day healthcare encounter, the uninsured population was significantly younger and more likely to reside in an urban Zip code. There were no differences in the incidence of 30 day healthcare encounter between the insured and uninsured populations. The uninsured population were more likely to be seen in the emergency department for 30 day healthcare encounter. The rate of 30 day healthcare encounters in the both groups was similar to what has reported in previous literature with readmission rates ranging from 11.9 -24.8% [1][2][3][4][5][6]. This was an interesting finding given the younger population that was analyzed in this study compared to previous literature with average age ranging from 75 -81 years [1][2][3][4][5][6]. Patients in the uninsured group utilized emergency room resources more frequently than the insured group. This was an interesting finding and previous media reports suggest frequent use of emergency department by uninsured patients as a source of primary care [12,15]. Emergency department care has identified as being significantly more costly to manage conditions that could have been managed or prevented by seeing a primary care physician [13,14,16]. Findings in this study suggest high use of emergency department resources in the uninsured group this population would benefit from interventions previously described that focus on improving transitioning of care [5,6]. These interventions would need to be tailored to the demographic group and comorbidities seen in the present study to be successful [4].
Medical conditions that were most frequently involved healthcare visits were abdominal pain, infection related complications, diabetes, chronic pain and chest pain. This was different than what has previously been reported in literature in Medicare populations [1]. A possible explanation for these differences in 30 day healthcare encounter in this analysis is the population was younger and not typically affected by conditions such as heart failure and COPD.
However, long term chronic conditions such as unstable angina and diabetes were a common source of a 30 day healthcare encounter. This study presents information on a population in which not much is known in regards to demographics and conditions causing healthcare encounters. It is interesting to see that the overall rate of 30 day healthcare encounter is 16.6% (n=348) and was primarily driven by emergency department visits. These findings suggests that this population would benefit from improved transitioning of care to set patients up for early outpatient clinic visits; an intervention that has been shown to decreased rehospitalization in a Medicare population [2]. With the high volume of emergency department visits utilized per patient in the uninsured population compared to in the insured group, it is likely that this population would benefit from an increase access to primary care physicians to improve patient's care and prevent costly emergency department visits. Conditions that were seen in this population that were 30 day healthcare encounters similar to what has been seen when utilizing systematic education prior to patient discharge for other disease states [5,6].
The uninsured population had a similar rate of 30 day healthcare encounters compared to the insured group and utilized emergency department more frequently. The population was younger and more likely to reside in an urban Zip code. Future studies should focus on developing a systematic process for caring for hospitalized uninsured patients, focusing on the disease states encountered in this population. These patients should be provided with early follow up care to a primary care doctor, systematic education about their disease state and what should be done if their condition worsens to prevent unnecessary emergency department and hospital visits.