| 1. |
- Bruze, Gustaf, et al.
(författare)
-
Associations of Bariatric Surgery With Changes in Interpersonal Relationship Status Results From 2 Swedish Cohort Studies
- 2018
-
Ingår i: JAMA Surgery. - : American Medical Association. - 2168-6254 .- 2168-6262. ; 153:7, s. 654-661
-
Tidskriftsartikel (refereegranskat)abstract
- IMPORTANCE: Bariatric surgery is a life-changing treatment for patients with severe obesity, but little is known about its association with interpersonal relationships.OBJECTIVES: To investigate if relationship status is altered after bariatric surgery.DESIGN, SETTING, AND PARTICIPANTS: Changes in relationship status after bariatric surgery were examined in 2 cohorts: (1) the prospective Swedish Obese Subjects (SOS) study, which recruited patients undergoing bariatric surgery from September 1, 1987, to January 31, 2001, and compared their care with usual nonsurgical care in matched obese control participants; and (2) participants from the Scandinavian Obesity Surgery Registry (SOReg), a prospective, electronically captured register that recruited patients from January 2007 through December 2012 and selected comparator participants from the general population matched on age, sex, and place of residence. Data was collected in surgical departments and primary health care centers in Sweden. The current analysis includes data collected up until July 2015 (SOS) and December 2012 (SOReg). Data analysis was completed from June 2016 to December 2017.MAIN OUTCOMES AND MEASURES: In the SOS study, information on relationship status was obtained from questionnaires. In the SOReg and general population cohort, information on marriage and divorce was obtained from the Swedish Total Population Registry.RESULTS: The SOS study included 1958 patients who had bariatric surgery (of whom 1389 [70.9%] were female) and 1912 matched obese controls (of whom 1354 [70.8%] were female) and had a median (range) follow-up of 10 (0.5-20) years. The SOReg cohort included 29 234 patients who had gastric bypass surgery (of whom 22 131 [75.6%] were female) and 283 748 comparators from the general population (of whom 214 342 [75.5%] were female), and had a median (range) follow-up of 2.9 (0.003-7.0) years. In the SOS study, the surgical patients received gastric banding (n = 368; 18.8%), vertical banded gastroplasty (n = 1331; 68.0%), or gastric bypass (n = 259; 13.2%); controls received usual obesity care. In SOReg, all 29 234 surgical participants received gastric bypass surgery. In the SOS study, bariatric surgery was associated with increased incidence of divorce/separation compared with controls for those in a relationship (adjusted hazard ratio [aHR] = 1.28; 95% CI, 1.03-1.60; P =.03) and increased incidence of marriage or new relationship (aHR = 2.03; 95% CI, 1.52-2.71; P <.001) in those who were unmarried or single at baseline. In the SOReg and general population cohort, gastric bypass was associated with increased incidence of divorce compared with married control participants (aHR = 1.41; 95% CI, 1.33-1.49; P <.001) and increased incidence of marriage in those who were unmarried at baseline (aHR = 1.35; 95% CI, 1.28-1.42; P <.001). Within the surgery groups, changes in relationship status were more common in those with larger weight loss.CONCLUSIONS AND RELEVANCE: In addition to its association with obesity comorbidities, bariatric surgery-induced weight loss is also associated with changes in relationship status.
|
|
| 2. |
- Bruze, Gustaf, et al.
(författare)
-
Mental health from 5 years before to 10 years after bariatric surgery in adolescents with severe obesity: a Swedish nationwide cohort study with matched population controls
- 2024
-
Ingår i: LANCET CHILD & ADOLESCENT HEALTH. - : ELSEVIER SCI LTD. - 2352-4642 .- 2352-4650. ; 8:2, s. 135-146
-
Tidskriftsartikel (refereegranskat)abstract
- Background: The long-term effects of bariatric surgery on the mental health of adolescents with severe obesity remain uncertain. We aimed to describe the prevalence of psychiatric health-care visits and filled prescription psychiatric drugs among adolescents with severe obesity undergoing bariatric surgery in the 5 years preceding surgery and throughout the first 10 years after surgery, and to draw comparisons with matched adolescents in the general population. Methods: Adolescents with severe obesity and who underwent bariatric surgery were identified through the Scandinavian Obesity Surgery Registry. We included adolescents who had bariatric surgery between 2007 and 2017 and were younger than 21 years at time of surgery. Each adolescent patient was matched with ten adolescents from the general population by age, sex, and county of residence. Specialist psychiatric care and filled psychiatric prescriptions were retrieved from nationwide data registers. Findings: 1554 adolescents (<21 years) with severe obesity underwent bariatric surgery between 2007 and 2017, 1169 (75%) of whom were female. At time of surgery, the mean age was 19.0 years [SD 1.0], and the mean BMI was 43.7 kg/m(2) (SD 5.5). 15 540 adolescents from the general population were matched with adolescents in the surgery group. 5 years before the matched index date, 95 (6.2%) of 1535 surgery patients and 370 (2.5%) of 14 643 matched adolescents had a psychiatric health-care visit (prevalence difference 3.7%; 95% CI 2.4-4.9), whereas 127 (9.8%) of 1295 surgery patients and 445 (3.6%) of 12 211 matched adolescents filled a psychiatric drug prescription (prevalence difference 6.2%; 95% CI 4.5-7.8). The year before the matched index date, 208 (13.4%) of 1551 surgery patients and 844 (5.5%) of 15 308 matched adolescents had a psychiatric health-care visit (prevalence difference 7.9%; 95% CI 6.2-9.6), whereas 319 (20.6%) of 1551 surgery patients and 1306 (8.5%) of 15 308 matched adolescents filled a psychiatric drug prescription (prevalence difference 12.0%; 10.0-14.1). The prevalence difference in psychiatric health-care visits peaked 9 years after the matched index date (12.0%; 95% CI 9.0-14.9), when 119 (17.6%) of 675 surgery patients and 377 (5.7%) of 6669 matched adolescents had a psychiatric health-care visit. The prevalence difference in filled psychiatric drug prescription was highest 10 years after the matched index date (20.4%; 15.9-24.9), when 171 (36.5%) of 469 surgery patients and 739 (16.0%) of 4607 matched adolescents filled a psychiatric drug prescription. The year before the matched index date, 19 (1.2%) of 1551 surgery patients and 155 (1.0%) of 15304 matched adolescents had a health-care visit associated with a substance use disorder diagnosis (mean difference 0.2%, 95% CI -0.4 to 0.8). 10 years after the matched index date, the prevalence difference had increased to 4.3% (95% CI 2.3-6.4), when 24 (5.1%) of 467 surgery patients and 37 (0.8%) of 4582 matched adolescents had a health-care visit associated with a substance use disorder diagnosis. Interpretation: Psychiatric diagnoses and psychiatric drug prescriptions were more common among adolescents with severe obesity who would later undergo bariatric surgery than among matched adolescents from the general population. Both groups showed an increase in prevalence in psychiatric diagnoses and psychiatric drug prescriptions leading up to the time of surgery, but the rate of increase in the prevalence was higher among adolescents with severe obesity than among matched adolescents. With the exception of health-care visits for substance use disorders, these prevalence trajectories continued in the 10 years of follow-up. Realistic expectations regarding mental health outcomes should be set preoperatively. Funding: Swedish Research Council, Swedish Research Council for Health, Working Life and Welfare.
|
|
| 3. |
- Claesson, Andreas, et al.
(författare)
-
Defibrillation before EMS arrival in western Sweden
- 2017
-
Ingår i: American Journal of Emergency Medicine. - : Elsevier. - 0735-6757 .- 1532-8171. ; 35:8, s. 1043-1048
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Bystanders play a vital role in public access defibrillation (PAD) in out-of-hospital cardiac arrest (OHCA). Dual dispatch of first responders (FR) alongside emergency medical services (EMS) can reduce time to first defibrillation. The aim of this study was to describe the use of automated external defibrillators (AEDs) in OHCAs before EMS arrival.METHODS: All OHCA cases with a shockable rhythm in which an AED was used prior to the arrival of EMS between 2008 and 2015 in western Sweden were eligible for inclusion. Data from the Swedish Register for Cardiopulmonary Resuscitation (SRCR) were used for analysis, on-site bystander and FR defibrillation were compared with EMS defibrillation in the final analysis.RESULTS: Of the reported 6675 cases, 24% suffered ventricular fibrillation (VF), 162 patients (15%) of all VF cases were defibrillated before EMS arrival, 46% with a public AED on site. The proportion of cases defibrillated before EMS arrival increased from 5% in 2008 to 20% in 2015 (p<0.001). During this period, 30-day survival increased in patients with VF from 22% to 28% (p=0.04) and was highest when an AED was used on site (68%), with a median delay of 6.5min from collapse to defibrillation. Adjusted odds ratio for on-site defibrillation versus dispatched defibrillation for 30-day survival was 2.45 (95% CI: 1.02-5.95).CONCLUSIONS: The use of AEDs before the arrival of EMS increased over time. This was associated with an increased 30-day survival among patients with VF. Thirty-day survival was highest when an AED was used on site before EMS arrival.
|
|
| 4. |
- Neovius, Martin, et al.
(författare)
-
Risk of suicide and non-fatal self-harm after bariatric surgery: results from two matched cohort studies.
- 2018
-
Ingår i: The lancet. Diabetes & endocrinology. - : Elsevier. - 2213-8595 .- 2213-8587. ; 6:3, s. 197-207
-
Tidskriftsartikel (refereegranskat)abstract
- Bariatric surgery reduces mortality, but might have adverse effects on mental health. We assessed the risk of suicide and self-harm after bariatric surgery compared with non-surgical obesity treatment.Suicide and non-fatal self-harm events retrieved from nationwide Swedish registers were examined in two cohorts. The non-randomised, prospective Swedish Obese Subjects (SOS) study compared bariatric surgery (n=2010; 1369 vertical-banded gastroplasty, 376 gastric banding, and 265 gastric bypass) with usual care (n=2037; recruitment 1987-2001). The second cohort consisted of individuals from the Scandinavian Obesity Surgery Registry (SOReg; n=20256 patients who had gastric bypass) matched to individuals treated with intensive lifestyle modification (n=16162; intervention 2006-13) on baseline BMI, age, sex, education level, diabetes, cardiovascular disease, history of self-harm, substance misuse, antidepressant use, anxiolytics use, and psychiatric health-care contacts.During 68528 person-years (median 18; IQR 14-21) in the SOS study, suicides or non-fatal self-harm events were higher in the surgery group (n=87) than in the control group (n=49; adjusted hazard ratio [aHR] 1·78, 95% CI 1·23-2·57; p=0·0021); of these events, nine and three were suicides, respectively (3·06, 0·79-11·88; p=0·11). In analyses by primary procedure type, increased risk of suicide or non-fatal self-harm was identified for gastric bypass (3·48, 1·65-7·31; p=0·0010), gastric banding (2·43, 1·23-4·82; p=0·011), and vertical-banded gastroplasty (2·25, 1·37-3·71; p=0·0015) compared with controls. Out of nine deaths by suicide in the SOS surgery group, five occurred after gastric bypass (two primary and three converted procedures). During 149582 person-years (median 3·9; IQR 2·8-5·2), more suicides or non-fatal self-harm events were reported in the SOReg gastric bypass group (n=341) than in the intensive lifestyle group (n=84; aHR 3·16, 2·46-4·06; p<0·0001); of these events, 33 and five were suicides, respectively (5·17, 1·86-14·37; p=0·0017). In SOS, substance misuse during follow-up was recorded in 48% (39/81) of patients treated with surgery and 28% (13/47) of controls with non-fatal self-harm events (p=0·023). Correspondingly, substance misuse during follow-up was recorded in 51% (162/316) of participants in the SOReg gastric bypass group and 29% (23/80) of participants in the intensive lifestyle group with non-fatal self-harm events (p=0·0003). The risk of suicide and self-harm was not associated with poor weight loss outcome.Bariatric surgery was associated with suicide and non-fatal self-harm. However, the absolute risks were low and do not justify a general discouragement of bariatric surgery. The findings indicate a need for thorough preoperative psychiatric history assessment along with provision of information about increased risk of self-harm following surgery. Moreover, the findings call for postoperative surveillance with particular attention to mental health.US National Institutes of Health and Swedish Research Council.
|
|
| 5. |
- Stenberg, Erik, 1979-, et al.
(författare)
-
Comparison of Sleeve Gastrectomy vs Intensive Lifestyle Modification in Patients With a BMI of 30 to Less Than 35
- 2022
-
Ingår i: JAMA Network Open. - : American Medical Association. - 2574-3805. ; 5:7
-
Tidskriftsartikel (refereegranskat)abstract
- Importance: There is a lack of studies evaluating sleeve gastrectomy compared with intensive lifestyle treatment in patients with class 1 obesity (body mass index [BMI] 30 to <35 [calculated as weight in kilograms divided by height in meters squared]).Objective: To compare outcomes and safety of sleeve gastrectomy compared with intensive nonoperative obesity treatment in patients with class 1 obesity.Design, Setting, and Participants: This matched, nationwide cohort study included patients with class 1 obesity who underwent a sleeve gastrectomy or intensive lifestyle treatment between January 1, 2012, and December 31, 2017, and who were registered in the Scandinavian Obesity Surgery Registry or the Itrim health database. Participants with class 1 obesity were matched 1:2 using a propensity score including age, sex, BMI, treatment year, education level, income, cardiovascular disease, and use of antibiotic drugs, antidepressants, and anxiolytics.Interventions: Sleeve gastrectomy or intensive lifestyle treatment.Main Outcomes and Measures: Outcomes included weight loss after intervention, changes in metabolic comorbidities, substance use disorders, self-harm, and major cardiovascular events retrieved from the National Patient Register, Prescribed Drug Register, and Cause of Death Register as well as the Scandinavian Obesity Surgery Registry and the Itrim health database. Data were analyzed from December 1, 2021 until May 31, 2022.Results: The study included 1216 surgery patients and 2432 lifestyle participants with similar mean (SD) BMI (32.8 [1.4] vs 32.9 [1.4]), mean (SD) age (42.4 [9.7] vs 42.6 [12.7] years), and sex (1091 [89.7%] vs 2191 [90.1%] women). Surgery patients had greater 1-year weight loss compared with controls (22.9 kg vs 11.9 kg; mean difference, 10.7 kg; 95% CI, 10.0-11.5 kg; P < .001). Over a median follow-up of 5.1 years (IQR, 3.9-6.2 years), surgery patients had a lower risk of incident use of diabetes drugs (59.7 vs 100.4 events per 10 000 person-years; hazard ratio [HR], 0.60; 95% CI, 0.39-0.92; P = .02) and greater 2-year diabetes drug remission (48.4% vs 22.0%; risk difference 26.4%; 95% CI, 11.7%-41.0%; P < .001), but higher risk for substance use disorder (94 vs 50 events per 10 000 person-years; HR, 1.86; 95% CI, 1.30-2.67; P < .001) and self-harm (45 vs 25 events per 10 000 person-years; HR, 1.81; 95% CI, 1.09-3.01; P = .02). No between-group difference in occurrence of major cardiovascular events was observed (23.4 vs 24.8 events per 10 000 person-years; HR, 0.96; 95% CI, 0.49-1.91; P = .92).Conclusions and Relevance: In this cohort study, compared with intensive nonoperative obesity treatment, sleeve gastrectomy in patients with class 1 obesity was associated with greater weight loss, diabetes prevention, and diabetes remission but a higher incidence of substance use disorder and self-harm.
|
|
| 6. |
- Wallén, Stefan, 1978-, et al.
(författare)
-
Opioid Use After Gastric Bypass, Sleeve Gastrectomy or Intensive Lifestyle Intervention
- 2023
-
Ingår i: Annals of Surgery. - : Lippincott Williams & Wilkins. - 0003-4932 .- 1528-1140. ; 277:3, s. e552-e560
-
Tidskriftsartikel (refereegranskat)abstract
- OBJECTIVE: To compare opioid use in patients with obesity treated with bariatric surgery versus adults with obesity who underwent intensive lifestyle modification. SUMMARY OFBACKGROUND DATA: Previous studies of opioid use after bariatric surgery have been limited by small sample sizes, short follow-up, and lack of control groups.METHODS: Nationwide matched cohort study including individuals from the Scandinavian Obesity Surgery Registry and the Itrim health database with individuals undergoing structured intensive lifestyle modification, between August 1, 2007 and September 30, 2015. Participants were matched on Body Mass Index, age, sex, education, previous opioid use, diabetes, cardiovascular disease, and psychiatric status (n = 30,359:21,356). Dispensed opioids were retrieved from the Swedish Prescribed Drug Register from 2 years before to up to 8 years after intervention.RESULTS: During the 2-year period before treatment, prevalence of individuals receiving ≥1 opioid prescription was identical in the surgery and lifestyle group. At 3 years, the prevalence of opioid prescriptions was 14.7% versus 8.9% in the surgery and lifestyle groups (mean difference 5.9%, 95% confidence interval 5.3-6.4) and at 8 years 16.9% versus 9.0% (7.9%, 6.8-9.0). The difference in mean daily dose also increased over time and was 3.55 mg in the surgery group versus 1.17 mg in the lifestyle group at 8 years (mean difference [adjusted for baseline dose] 2.30 mg, 95% confidence interval 1.61-2.98).CONCLUSIONS: Bariatric surgery was associated with a higher proportion of opioid users and larger total opioid dose, compared to actively treated obese individuals. These trends were especially evident in patients who received additional surgery during follow-up.
|
|
| 7. |
- Backman, Olof, et al.
(författare)
-
Gastric Bypass Surgery Reduces De Novo Cases of Type 2 Diabetes to Population Levels : A Nationwide Cohort Study From Sweden
- 2019
-
Ingår i: Annals of Surgery. - : Lippincott Williams & Wilkins. - 0003-4932 .- 1528-1140. ; 269:5, s. 895-902
-
Tidskriftsartikel (refereegranskat)abstract
- OBJECTIVE: The aim of this study was to determine long-term changes in pharmacological treatment of type 2 diabetes after primary Roux-en-Y gastric bypass (RYGB) surgery, in patients with and without pharmacological treatment of diabetes preoperatively.SUMMARY OF BACKGROUND DATA: Several studies have shown that gastric bypass has good effect on diabetes, at least in the short-term. This study is a nationwide cohort study using Swedish registers, with basically no patients lost to follow-up during up to 7 years after surgery.METHODS: The effect of RYGB on type 2 diabetes drug treatment was evaluated in this nationwide matched cohort study. Participants were 22,047 adults with BMI ≥30 identified in the nationwide Scandinavian Surgical Obesity Registry, who underwent primary RYGB between 2007 and 2012. For each individual, up to 10 general population comparators were matched on birth year, sex, and place of residence. Prescription data were retrieved from the nationwide Swedish Prescribed Drug Register through September 2015. Incident use of pharmacological treatment was analyzed using Cox regression.RESULTS: Sixty-seven percent of patients with pharmacological treatment of type 2 diabetes before surgery were not using diabetes drugs 2 years after surgery and 61% of patients were not pharmacologically treated up to 7 years after surgery. In patients not using diabetes drugs at baseline, there were 189 new cases of pharmacological treatment of type 2 diabetes in the surgery group and 2319 in the matched general population comparators during a median follow-up of 4.6 years (incidence: 21.4 vs 27.9 per 10,000 person-years; adjusted hazard ratio 0.77, 95% confidence interval 0.67-0.89; P < 0.001).CONCLUSIONS: Gastric bypass surgery not only induces remission of pharmacological treatment of type 2 diabetes but also protects from new onset of pharmacological diabetes treatment. The effect seems to persist in most, but not all, patients over 7 years of follow-up.
|
|
| 8. |
- Bruze, Gustaf, et al.
(författare)
-
Hospital admission after gastric bypass : a nationwide cohort study with up to 6 years follow-up.
- 2017
-
Ingår i: Surgery for Obesity and Related Diseases. - : Elsevier. - 1550-7289 .- 1878-7533. ; 13:6, s. 962-969
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Several studies have addressed short-term admission rates after bariatric surgery. However, studies on long-term admission rates are few and population based studies are even scarcer.OBJECTIVE: The aim of this study was to assess short- and long-term admission rates for gastrointestinal surgery after gastric bypass in Sweden compared with admission rates in the general population.SETTING: Swedish healthcare system.METHODS: The surgery cohort consisted of adults with body mass index≥35 identified in the Scandinavian Obesity Surgery Registry (n = 28,331; mean age 41 years; 76% women; Roux-en-Y gastric bypass performed 2007-2012). For each individual, up to 10 comparators from the general population were matched on birth year, sex, and place of residence (n = 274,513). The primary outcome was inpatient admissions due to gastrointestinal surgery retrieved from the National Patient Register through December 31, 2014. Conditional hazard ratios (HR) were estimated using Cox regression.RESULTS: All-cause admission rates were 6.5%, 21.4%, and 65.9% during 30 days, 1 year, and 6 years after surgery, respectively. The corresponding rates for gastrointestinal surgery were 1.8%, 6.8%, and 24.4%. Compared with that of the general population, there was an increased risk of all-cause hospital admission at 1 year (HR 2.6 [2.5-2.6]) and 6 years (HR 2.7 [2.6-2.7]). The risk of hospital admission for any gastrointestinal surgical procedure was greatly increased throughout the study period (HR 8.6 [8.4-8.9]). Female sex, psychiatric disease, and low education were risk factors.CONCLUSION: We found a significant risk of admission to hospital over>6 years after gastric bypass surgery.
|
|
| 9. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
A Comparative Study of Machine Learning Algorithms in Predicting Severe Complications after Bariatric Surgery
- 2019
-
Ingår i: Journal of Clinical Medicine. - : MDPI. - 2077-0383. ; 8:5
-
Tidskriftsartikel (refereegranskat)abstract
- Background: Severe obesity is a global public health threat of growing proportions. Accurate models to predict severe postoperative complications could be of value in the preoperative assessment of potential candidates for bariatric surgery. So far, traditional statistical methods have failed to produce high accuracy. We aimed to find a useful machine learning (ML) algorithm to predict the risk for severe complication after bariatric surgery.Methods: We trained and compared 29 supervised ML algorithms using information from 37,811 patients that operated with a bariatric surgical procedure between 2010 and 2014 in Sweden. The algorithms were then tested on 6250 patients operated in 2015. We performed the synthetic minority oversampling technique tackling the issue that only 3% of patients experienced severe complications.Results: Most of the ML algorithms showed high accuracy (>90%) and specificity (>90%) in both the training and test data. However, none of the algorithms achieved an acceptable sensitivity in the test data. We also tried to tune the hyperparameters of the algorithms to maximize sensitivity, but did not yet identify one with a high enough sensitivity that can be used in clinical praxis in bariatric surgery. However, a minor, but perceptible, improvement in deep neural network (NN) ML was found.Conclusion: In predicting the severe postoperative complication among the bariatric surgery patients, ensemble algorithms outperform base algorithms. When compared to other ML algorithms, deep NN has the potential to improve the accuracy and it deserves further investigation. The oversampling technique should be considered in the context of imbalanced data where the number of the interested outcome is relatively small.
|
|
| 10. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Deep Learning Neural Networks to Predict Serious Complications After Bariatric Surgery : Analysis of Scandinavian Obesity Surgery Registry Data
- 2020
-
Ingår i: JMIR Medical Informatics. - : JMIR Publications. - 2291-9694. ; 8:5
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Obesity is one of today's most visible public health problems worldwide. Although modern bariatric surgery is ostensibly considered safe, serious complications and mortality still occur in some patients.OBJECTIVE: This study aimed to explore whether serious postoperative complications of bariatric surgery recorded in a national quality registry can be predicted preoperatively using deep learning methods.METHODS: Patients who were registered in the Scandinavian Obesity Surgery Registry (SOReg) between 2010 and 2015 were included in this study. The patients who underwent a bariatric procedure between 2010 and 2014 were used as training data, and those who underwent a bariatric procedure in 2015 were used as test data. Postoperative complications were graded according to the Clavien-Dindo classification, and complications requiring intervention under general anesthesia or resulting in organ failure or death were considered serious. Three supervised deep learning neural networks were applied and compared in our study: multilayer perceptron (MLP), convolutional neural network (CNN), and recurrent neural network (RNN). The synthetic minority oversampling technique (SMOTE) was used to artificially augment the patients with serious complications. The performances of the neural networks were evaluated using accuracy, sensitivity, specificity, Matthews correlation coefficient, and area under the receiver operating characteristic curve.RESULTS: In total, 37,811 and 6250 patients were used as the training data and test data, with incidence rates of serious complication of 3.2% (1220/37,811) and 3.0% (188/6250), respectively. When trained using the SMOTE data, the MLP appeared to have a desirable performance, with an area under curve (AUC) of 0.84 (95% CI 0.83-0.85). However, its performance was low for the test data, with an AUC of 0.54 (95% CI 0.53-0.55). The performance of CNN was similar to that of MLP. It generated AUCs of 0.79 (95% CI 0.78-0.80) and 0.57 (95% CI 0.59-0.61) for the SMOTE data and test data, respectively. Compared with the MLP and CNN, the RNN showed worse performance, with AUCs of 0.65 (95% CI 0.64-0.66) and 0.55 (95% CI 0.53-0.57) for the SMOTE data and test data, respectively.CONCLUSIONS: MLP and CNN showed improved, but limited, ability for predicting the postoperative serious complications after bariatric surgery in the Scandinavian Obesity Surgery Registry data. However, the overfitting issue is still apparent and needs to be overcome by incorporating intra- and perioperative information.
|
|
