Background: Agency of Healthcare Research and Quality reported an increase in Opioid-related hospitalizations from 2005 to 2014. Previous studies in patients undergoing surgery have shown an increased risk of readmission with opioid dependence. This study was intended to identify the all-cause 7 and 30-day readmission rates of hospitalizations with opioid-related disorders, and potential predictors of readmissions.
Methods: National Readmission Database (NRD year 2016) maintained and made available by the Agency of Healthcare Research and Quality was analyzed. It consists of about 17 million discharges which can be weighted to produce national estimates.Inclusion criteria were a primary diagnosis of any Opioid-related disorder (ORD). Exclusion criteria were age <18, admissions in month of December and inpatient mortality. Index admission was defined as any admission with a diagnosis of ORD meeting the above criteria. Index admission did not need a preceding period of no admissions. Admission could be a readmission for the previous and index for the subsequent admission. Descriptive statistics were applied to estimate the total number of index admissions, 7 & 30-day all-cause readmissions rates, and the most common readmission diagnoses. Multivariate regression analysis using the variables in the dataset was performed to identify the predictors of readmissions.
Results: An estimated 130,774 admissions met the criteria for index admissions. 4.9% (6364/130,774) & 13.5% (17,672/130,7740) of the index admissions were readmitted within 7 & 30-days respectively. The most common primary diagnoses associated with 7-day readmissions were opioid withdrawal (12%), opioid dependence (11%), depression (5%) and sepsis (3.6%). For 30-day readmissions, these were opioid withdrawal (15%), opioid dependence (9%), depression (4.5%) and sepsis (3.6%).Odds of 7-day readmissions were higher with higher Charlson comorbidity index (OR 1.1; P <0.01), discharges with home health care (OR 1.26; P <0.01), transfer to short term hospital (OR 3.2; P <0.01) and against medical advice (AMA) (OR 1.77; P <0.01) v/s routine discharges, metropolitan teaching v/s non-teaching hospitals (OR 1.29; P <0.01), and lower in females (OR 0.86; P <0.01) as well as private insurance (OR 0.72; P <0.01) and self-pay (OR 0.69; P <0.01) v/s Medicare. In the case of 30-day readmissions, odds were higher with higher Charlson comorbidity index (OR 1.1; P <0.01), home health care (OR 1.18; P <0.01), transfer to short term hospital (OR 1.79; P <0.01) and against medical advice discharges (OR 1.46; P <0.01) v/s routine discharges, metropolitan teaching v/s non-teaching hospitals (OR 1.22; P <0.01), and lower in females (OR 0.87; P <0.01), private insurance (OR 0.71; P <0.01) and self-pay (OR 0.69; P <0.01) v/s Medicare.
Conclusions: The analysis revealed about 1 in 20 and 1 in 10 admissions with a primary diagnosis of ORD end up being readmitted within 7 & 30 days respectively. Predictors and diagnoses for readmissions were similar for both 7 & 30-day readmissions. Nearly 1/4th of the readmissions had a primary diagnosis related to opioid use. Depression was identified as another common reason for readmissions indicating the coexistence of underlying psychiatric comorbidities in these patients. AMA discharges have a higher likelihood of readmissions likely because of a lack of resources and poor follow-up at discharge. Readmissions were lower in case of private insurance and self-pay compared to Medicare conceivably because of the more robust follow-ups.