The Weekend effect and readmissions in the Helsinki and Uusimaa Hospital District : patient safety in secondary and tertiary health care

Background: Patients admitted to the hospital during the weekend have been found to have a higher risk of mortality than those admitted during the week. This weekend effect phenomenon has been widely investigated and the reasons for it extensively discussed. The existence of the weekend effect was examined by specialty at the university hospital and six secondary hospitals in the greater Helsinki area during a 14-year period. The majority of ear, nose and throat (ENT), as well as orthopedic and hand (OHS), day surgery procedures are performed under local or regional anesthesia in Finland. However, this is not true of many other countries. Associations and background factors for overstay, readmission and contacts were investigated at Helsinki University Hospital. Methods: Data for all inpatients during the years 2000-2013 in the Helsinki and Uusimaa Hospital District were selected retrospectively: for the university hospital, all those treated at some point at the university hospital; for the secondary hospitals, those inpatients only treated in secondary hospitals. Urgency and specialty of care were used to group the weekend effect study population. Associations between variables were analyzed. Patients undergoing ENT (n=1,011) or OHS (n=542) day surgery from January 1 to March 31, 2015 were collected retrospectively from the hospital’s surgery database (GE Healthcare Centricity Opera OR Management Software). The 15 most common procedures were selected for both ENT and OHS day surgery. Data was collected on all-cause overstays, readmissions and contacts during the 30-day follow-up period. For ENT patients, American Society of Anesthesiologists (ASA) class, age, sex, type of procedure and anesthesia comprised the multivariable logistic regression model. For OHS patients, diverse factors related to patient characteristics and anesthesia were scrutinized for their effect on these outcomes using Pearson chi-square test, Fischer’s exact test and multivariate logistic regression. For OHS patients, anesthesia charts were also examined and associations between variables were studied using risk profiles. Results: For the university hospital, data for 1,542,230 inpatients were collected retrospectively for the weekend effect study. Of these, 853,268 were emergency patients. Deaths in hospital or within 30 days of discharge numbered 47,122. In in-hospital mortality, a significant weekend effect was found in 7 of 12 specialties for emergency admissions and 4 of 12 specialties for elective admissions. In 30-day post-discharge mortality, a significant weekend effect was found in 1 of 12 specialties for emergency admissions and 2 of 12 specialties for elective admissions. For the six secondary hospitals, data for 456,676 inpatients were selected retrospectively for the weekend effect study. Of these, 292,399 were emergency patients. Deaths in hospital or within 30 days of discharge numbered 17,231. In in-hospital mortality, a significant weekend effect was found in 1 of 7 specialties for emergency admissions and 4 of 8 specialties for elective admissions. In 30-day post-discharge mortality, a significant weekend effect was found in 1 of 7 specialties for emergency admissions and 3 of 8 specialties for elective admissions. The specialties most sensitive to the phenomenon in both the university and the secondary hospitals were surgery, internal medicine, and gynecology and obstetrics. In addition, neurology was also sensitive to the phenomenon in the university hospital. For ENT patients, sex, age and type of procedure proved to be significant factors on the study outcomes of overstay, readmission and contact. General anesthesia patients had an overstay or readmission 3.2% (n=23) of the time, while local anesthesia patients only 1.4% (n=4) of the time. The majority of study outcomes occurred in tonsil surgery, which was only carried out under general anesthesia. For OHS patients, statistically significant factors related to outcomes of overstay, readmission or contact were female sex, total amount of fentanyl, use of remifentanil, other pain medication during procedure and administration of antiemetic medication. General anesthesia and plexus block, total amount of oxycodone and postoperative pain medication emerged as borderline significant factors on outcome after day surgery. Combination analysis was then performed to find risk profiles for outcomes. Conclusion: In the university hospital, a weekend effect for many specialties for both emergency and elective admissions was observed. In the secondary hospitals, elective patients had a weekend effect for many specialties. Weekend elective procedures must be restricted to only those necessary to be performed on the weekend and guidelines for patient selection are needed. Before staffing is adjusted, more disease-specific research is needed to find which patients would benefit most. ASA class and type of anesthesia did not affect the risk of outcomes in ENT day surgery but sex, type of procedure and age did. Female OHS patients with procedures under general anesthesia and requiring greater amounts of opioids in conjunction with surgery were undoubtedly linked with study outcomes. As type of anesthesia had no effect on study outcomes in ENT and OHS day surgery, local and regional anesthesia should be used when medically and procedurally possible. Overstay, readmission and contact rates were the same, or lower, than other international studies.