The effect of polyallergy on ancillary service use among postoperative patients with gynecologic malignancies Authors: Kaitlyn Coppetti, Madeline Petty, Allison R. Walker, Luis Alvarado, Sarah P. Huepenbecker, Nicholas Taylor, Ashley Graul Published: 2026 ABSTRACT ObjectiveTo determine if polyallergy (four or more documented allergies in the electronic medical record [EMR]) is an independent predictor of ancillary service use.MethodsSingle institution, retrospective cohort study of patients who underwent surgery for gynecologic malignancy between 01/2016 and 05/2024. Demographics, comorbidities, surgical details, postoperative complications, and allergies were abstracted from the EMR. The primary outcome was ancillary service utilization for one year postoperatively including emergency room, psychologic services, primary care, urgent care visits, and hospitalizations. Descriptive statistics, univariate, and multivariate analyses were performed.Results398 patients met study criteria and 77 patients had polyallergy with a median of 5 allergies (IQR 4–7). Most patients were White (92%) non-smokers (56%) with a median age of 65 (IQR 37–51). Patients with polyallergy had higher rates of anxiety (38% vs 25%, p = 0.03), depression (45% vs 32%, p = 0.03), panic disorder (6% vs 2%, 0.04), and antipsychotic prescriptions (6% vs 2%, p = 0.03). There were more ancillary service encounters among those with polyallergy compared to non-polyallergy (median 4, IQR 2–6 vs 2 IQR 1–5, p = 0.001). On univariate analysis, polyallergy was significantly associated with ancillary service use (p = 0.002), which remained significant on multivariate analysis (RR = 0.21, CI 0.02–0.39, p = 0.03), as was any psychiatric diagnosis (RR = 0.34, CI 0.19–0.48, p < 0.001), and any postoperative complication (RR = 0.29, CI 0.08–0.50, p = 0.006).ConclusionsPolyallergy was an independent risk factor for ancillary service use. Patients with polyallergy may benefit from earlier primary care, case management, and possibly psychological intervention in an effort to prevent increased ancillary service use. Introduction Approximately 94,000 women are diagnosed with a gynecologic cancer each year and the median age of diagnosis varies by cancer type as well as race/ethnicity (National Center for Chronic Disease Prevention and Health Promotion (US), 2019). These patients often have high rates of healthcare utilization secondary to the nature of their disease. The number of annual emergency department visits these women have depends on the type of cancer diagnosis. Between 2010 and 2014, ovarian cancer patients had a mean of 112,456 emergency department visits; uterine cancer had a mean of 109,795 visits; cervical cancer had a mean of 133,226 visits; and vulvar/vaginal cancer had a mean of 4483 visits (Albright et al., 2021). Identifying patients who are at risk of increased healthcare utilization early on in their cancer journey may allow for earlier intervention to ease the burden on both patients and the healthcare system. Prior literature reflects that adult patients with a high number of documented allergies, which may be more prevalent among females, have higher rates of healthcare utilization including emergency department visits, primary care visits, and inpatient admissions (Jimenez et al., 2019, Blumenthal et al., 2018, Blumenthal et al., 2020). Patients with polyallergy also have a higher chance of having a coexisting mental health diagnosis (Jimenez et al., 2019, Blumenthal et al., 2018, Park et al., 2014). Katayama et al. found that patients who underwent a surgery for non-gynecologic cancer diagnoses and had a behavioral health disorder are more likely to have poor surgical outcomes, including postoperative complications, increased length of stay, and 90-day readmissions as well as worse long term postoperative survival. In addition, they are less likely to be discharged to their home and more likely to require additional support or facilities.5 Grossmann et al. analyzed the impact of patient reported penicillin allergies on postoperative complications in patients undergoing elective oncological surgery from different specialties including visceral, gynecological, urological, head and neck and thoracic cancers and found no significant difference in postoperative outcomes (Grossmann et al., 2023). The association between polyallergy, postoperative healthcare service use, and psychiatric diseases has been described in the general medical population, but it is unclear if this relationship persists in patients undergoing surgery for a gynecologic malignancy. Patients with gynecologic malignancies represent a complex population that has high healthcare utilization due to cancer-directed treatment/surveillance, postoperative follow-up, and complication management which may obscure other factors driving utilization. Determining if an association exists and if polyallergy increases the risk for utilization in this high-acuity population is clinically relevant for identifying a group of patients who would benefit from perioperative planning and additional psychosocial support. We aimed to determine if polyallergy is an independent predictor of ancillary service use among postoperative patients with gynecologic malignancies. We hypothesized that patients with polyallergy undergoing surgery for gynecologic malignancy would have increased ancillary service use and would be more likely to have a mental health diagnosis compared to patients without polyallergy. Methods We conducted a single institution, retrospective cohort study of patients who underwent surgery for gynecologic malignancy between January 2016 and May 2024. All adult gynecologic oncology patients with a pathologically confirmed diagnosis of cervical, uterine, vulvar, vaginal, or ovarian pathology who had been evaluated by a gynecologic oncologist in the St. Luke’s University Health Network and underwent surgery were included. Patients who died during the study period were excluded from the analysis. Based on definitions used in previous literature, polyallergy has been defined as the presence of 4 or more documented drug adverse reactions listed in the medical record regardless of if the allergy has been confirmed with medical testing (Jimenez et al., 2019, Park et al., 2014). Polyallergy was defined in this study as 4 or more documented allergies in the electronic medical record (EMR) including environmental allergies. This included patient reported allergies that had not been verified by clinical markers, regardless of severity (type of reaction i.e. hives or gastrointestinal upset versus anaphylaxis) or classification. Patients were classified into two cohorts: non-polyallergy (<4 documented allergies) and polyallergy (4 or more documented allergies). The primary outcome was ancillary service use between these cohorts including emergency department (ED) visits, urgent care visits, primary care visits, psychiatric and/or psychology visits, and inpatient admissions for one year following surgery. Secondary outcomes included the correlation between polyallergy and mental health diagnoses, postoperative complications, readmission rates, and the rates of polyallergy in patients with gynecologic malignancies. Study data were collected and managed using REDCap electronic data capture tools hosted at St. Luke’s University Health Network. Data were abstracted from the EMR via retrospective chart review and included patient demographics, comorbidities, psychiatric history, psychiatric medications, surgical and postoperative variables, cancer history, and documented allergies. Psychiatric diagnoses and functional somatic syndrome diagnoses of interest included generalized anxiety disorder, panic disorder, obsessive compulsive disorder, post-traumatic stress disorder, major depressive disorder, adjustment disorder, bipolar disorder (I/II), personality disorder, schizoaffective disorder, schizophrenia, delusional disorder, fibromyalgia, irritable bowel syndrome, somatization disorder, and conversion disorder. These diagnoses were included if they had a previous and active diagnosis in the EMR. Prescribed psychiatric medications were included if they were present at the time of the preoperative visit and included anti-depressants (including selective serotonin reuptake inhibitors, serotonin and norepinephrine reuptake inhibitors, and tricyclic antidepressants), anxiolytics, antipsychotics, mood stabilizers, and stimulants. These medications were not included if started in the postoperative period. Ancillary healthcare service utilization for one year postoperatively was collected from the EMR. Measures of healthcare utilization included emergency room visits, psychiatric service encounters, primary care visits, urgent care visits and hospitalizations including length of inpatient admissions. This included encounters that were able to be identified in the St. Luke’s University Health Care Network and in the Care Everywhere function in Epic. This does not include encounters that were not available for review in Epic. Continuous variables were analyzed using the median and interquartile range. Categorical variables were summarized using frequency and percentage. The main outcome variables of interest included a composite score of total healthcare utilization. Due to the distribution of the scores, the variable was log-transformed to ensure normality. Associations between the non-polyallergy and polyallergy cohorts and demographics, clinical characteristics, and ancillary service use were analyzed using Kruskal-Wallis, Wilcoxon Rank Sum, and Chi squared tests. Univariate and multivariate analyses were performed for variables associated with ancillary service use via stepwise regression. A univariate linear regression analysis was conducted to evaluate the relationship between total healthcare utilization (log-transformed) with all variables of interest. All variables that were considered significant in the univariate analysis were included in the final multivariable model. Each model was detailed using linear coefficients, 95% confidence interval (C.I.), and p-values. Results were considered statistically significant with a p-value of less than or equal to 0.05. All data management and analysis were conducted using Stata V.18. Results Three hundred ninety-eight patients met inclusion criteria, and the majority were White/Caucasian (88%, n = 349) non-smokers (87%, n = 347) with a median age of 65 (IQR 58–72) and median BMI of 33 (IQR 27–39). Most patients had a diagnosis of uterine cancer (66%, n = 261) and stage 1 disease (69%, n = 275). The most common comorbidities were hypertension (67%, n = 266) and hyperlipidemia (61%, n = 244). Of the surgeries performed, 65% (n = 260) were a robotic assisted total laparoscopic hysterectomy and bilateral salpingo-oophorectomy/adnexal surgery. Seventy-seven patients met criteria for polyallergy with a median of 5 allergies (IQR 4–7), whereas 321 patients were in the non-polyallergy group and had a median of 1 allergy (IQR 0–2) (Table 1). Severity of allergic reactions differed between the polyallergy and non-polyallergy cohort: low in 25% of the non-polyallergy cohort vs 3% of the polyallergy cohort, medium in 18% of the non-polyallergy cohort vs 32% in the polyallergy cohort, and high in 31% of the non-polyallergy cohort vs 45% in the polyallergy cohort (p < 0.001). The polyallergy cohort had significantly more diagnoses of diabetes (44% vs 31%, p = 0.033) and asthma (31% vs 20%, p = 0.047, n = 24). The polyallergy cohort had more mental health diagnoses compared to those without polyallergy including anxiety (38% vs 25%, p = 0.03), depression (45% vs 32%, p = 0.03), and panic disorder (6% vs 2%, 0.04) (Table 2). The two cohorts had no significant difference in other mental health diagnoses (Fig. 1). The polyallergy cohort had a significantly higher rate of antipsychotic medication use (6% vs 2%, p = 0.03) compared to the non-polyallergy cohort. There was no difference in the use of anti-depressants, anxiolytics, mood stabilizers, or stimulants between cohorts.Table 1Demographics, oncologic history, comorbidities, documented allergies and postoperative complications by allergy cohort.Overall (n = 398)Non-polyallergy (n = 321)Polyallergy(n = 77)p-valueAge, median (IQR)65 (58, 72)65 (58, 72)67 (60, 72)0.24BMI, median (IQR)33 (27–39)33 (27–39)33(29–39)Race0.97 White/Caucasian, n (%)349 (88%)278 (87%)71 (92%) Asian and Pacific Islander, n (%)3 (1%)3 (1%)0 (0%) Black or AA, n (%)16 (4%)15 (5%)1 (1%) Hispanic/Latinx, n (%)20 (5%)16 (5%)4 (5%) Mixed, n (%)6 (2%)6 (2%)0 (0%) Other, n (%)4 (1%)3 (1%)1 (1%)Marital status0.92 Single, n (%)74 (19%)62 (19%)12 (16%) Married, n (%)214 (54%)172 (54%)42 (55%) Widowed, n (%)58 (15%)47 (15%)11 (14%) Divorced, n (%)48 (12%)37 (12%)11 (14%) unknown, n (%)4 (1%)3 (1%)1 (1%)Tobacco use, n (%)51 (13%)44 (14%)7 (9%)0.34Type of Gynecologic Cancer0.75 Cervical cancer, n (%)13 (3%)10 (3%)3 (4%) Uterine Cancer, n (%)261 (66%)211 (66%)50 (65%) Ovarian Cancer, n (%)82 (21%)65 (20%)17 (22%) Vulvar cancer, n (%)40 (10%)34 (11%)6 (8%) Vaginal cancer, n (%)2 (1%)1 (0%)1 (1%)Stage0.11 I, n (%)275 (69%)230 (72%)45 (58%) II, n (%)23 (6%)16 (5%)7 (9%) III, n (%)75 (19%)55 (17%)20 (26%) IV, n (%)25 (6%)20 (6%)5 (6%)Procedure0.81 Laparoscopic adnexal surgery6 (2%)6 (2%)0 (0%) RA-BSO/adnexal surgery11 (3%)8 (2%)3 (4%) RA-TLH4 (1%)3 (1%)1 (1%) RA-TLH/BSO/adnexal surgery260 (65%)208 (65%)52 (68%) Total Abdominal Hysterectomy75 (19%)61 (19%)14 (18%) Vulvectomy42 (11%)35 (11%)7 (9%)Comorbidities T1 or T2 Diabetes, n (%)134 (34%)100 (31%)34 (44%)0.033 Tobacco use, n (%)51 (13%)44 (14%)7 (9%)0.34 Asthma, n (%)89 (22%)65 (20%)24 (31%)0.047 COPD, n (%)31 (8%)21 (7%)10 (13%)0.09 OSA, n (%)71 (18%)52 (16%)19 (25%)0.09 Hypertension, n (%)266 (67%)213 (66%)53 (69%)0.79 Hyperlipidemia, n (%)244 (61%)189 (59%)55 (71%)0.05 Congestive heart failure, n (%)27 (7%)19 (6%)8 (10%)0.20 Coronary artery disease, n (%)42 (11%)30 (9%)12 (16%)0.15 Myocardial infarction, n (%)15 (4%)10 (3%)5 (6%)0.18 Atrial fibrillation, n (%)42 (11%)31 (10%)11 (14%)0.30Documented allergies and reactionsNumber of allergies, median (IQR)1.0 (0.0, 3.0)1.0 (0.0, 2.0)5.0 (4.0, 7.0)<0.001Allergy type0.005 Drug class, n (%)33 (11%)31 (15%)2 (3%) Drug, n (%)189 (66%)127 (60%)62 (81%) Drug ingredient, n (%)3 (1%)3 (1%)0 (0%) Not documented, n (%)63 (22%)50 (24%)13 (17%)Severity of reaction for allergy<0.001 Low, n (%)55 (19%)53 (25%)2 (3%) Medium, n (%)64 (22%)39 (18%)25 (32%) High, n (%)100 (35%)65 (31%)35 (45%) Not documented, n (%)69 (24%)54 (26%)15 (19%)Type of reaction for allergy0.94 Allergy, n (%)199 (69%)145 (69%)54 (70%) Contradiction, n (%)2 (1%)1 (0%)1 (1%) Intolerance, n (%)53 (18%)40 (19%)13 (17%) Unspecified, n (%)3 (1%)2 (1%)1 (1%) Not documented, n (%)31 (11%)23 (11%)8 (10%)Postoperative Complications Pulmonary Embolism, n (%)1 (0%)1 (0%)0 (0%)1.00 Deep Vein Thrombosis, n (%)0 (0%)0 (0%)0 (0%) Surgical Site Infection, n (%)8 (2%)7 (2%)1 (1%)1.00 Wound Dehiscence, n (%)15 (4%)14 (4%)1 (1%)0.32 Pelvic Abscess, n (%)8 (2%)6 (2%)2 (3%)0.65 Hemoperitoneum/Bleeding, n (%)5 (1%)3 (1%)2 (3%)0.25 Pneumonia, n (%)0 (0%)0 (0%)0 (0%) UTI, n (%)6 (2%)3 (1%)3 (4%)0.09 Port-site hernia, n (%)0 (0%)0 (0%)0 (0%) Unanticipated ICU stay (during surgical admission), n (%)4 (1%)3 (1%)1 (1%)0.58 Other, n (%)20 (5%)14 (4%)6 (8%)0.24 Readmission, n (%)17 (4%)14 (4%)3 (4%)1.00 Any post-op complication, n (%)57 (14%)43 (13%)14 (18%)0.28RA-BSO: Robotic assisted bilateral salpingo-oophorectomy.RA-TLH: Robotic assisted total laparoscopic hysterectomy.RA-TLH/BSO: Robotic assisted total laparoscopic hysterectomy and bilateral salpingo-oophorectomy.Table 2Rates of mental health diagnoses and psychiatric medication summary by allergy cohort.Overall (n = 398)Non-polyallergy (n = 321)Polyallergy (n = 77)p-valueGeneral anxiety disorder, n (%)109 (27%)80 (25%)29 (38%)0.032Panic disorder, n (%)11 (3%)6 (2%)5 (6%)0.042Obsessive compulsive disorder, n (%)8 (2%)5 (2%)3 (4%)0.19Post traumatic stress disorder, n (%)6 (2%)5 (2%)1 (1%)1.00Major depressive disorder, n (%)139 (35%)104 (32%)35 (45%)0.034Adjustment disorder, n (%)18 (5%)13 (4%)5 (6%)0.36Bipolar disorder I/II, n (%)8 (2%)6 (2%)2 (3%)0.65Personality disorder, n (%)2 (1%)2 (1%)0 (0%)1.00Schizoaffective disorder, n (%)1 (0%)0 (0%)1 (1%)0.19Schizophrenia, n (%)1 (0%)0 (0%)1 (1%)0.19Delusional disorder, n (%)1 (0%)1 (0%)0 (0%)1.00Fibromyalgia, n (%)24 (6%)17 (5%)7 (9%)0.28Irritable bowel syndrome, n (%)19 (5%)14 (4%)5 (6%)0.39Somatization disorder, n (%)0 (0%)0 (0%)0 (0%)Conversion disorder, n (%)0 (0%)0 (0%)0 (0%) Psychiatric medications at time of preoperative visitAnti-depressants, n (%)96 (24%)71 (22%)25 (32%)0.07Anxiolytic, n (%)91 (23%)68 (21%)23 (30%)0.13Antipsychotic, n (%)10 (3%)5 (2%)5 (6%)0.027Mood stabilizer, n (%)1 (0%)1 (0%)0 (0%)1.00Stimulant, n (%)0 (0%)0 (0%)0 (0%)Fig. 1Shows the difference in psychiatric diagnoses between the polyallergy (orange) and non-polyallergy (blue) cohorts. Patients with polyallergy had significantly more diagnoses of major depressive disorder (p = 0.034) and general anxiety disorder (p = 0.032). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 1Demographics, oncologic history, comorbidities, documented allergies and postoperative complications by allergy cohort.Overall (n = 398)Non-polyallergy (n = 321)Polyallergy(n = 77)p-valueAge, median (IQR)65 (58, 72)65 (58, 72)67 (60, 72)0.24BMI, median (IQR)33 (27–39)33 (27–39)33(29–39)Race0.97 White/Caucasian, n (%)349 (88%)278 (87%)71 (92%) Asian and Pacific Islander, n (%)3 (1%)3 (1%)0 (0%) Black or AA, n (%)16 (4%)15 (5%)1 (1%) Hispanic/Latinx, n (%)20 (5%)16 (5%)4 (5%) Mixed, n (%)6 (2%)6 (2%)0 (0%) Other, n (%)4 (1%)3 (1%)1 (1%)Marital status0.92 Single, n (%)74 (19%)62 (19%)12 (16%) Married, n (%)214 (54%)172 (54%)42 (55%) Widowed, n (%)58 (15%)47 (15%)11 (14%) Divorced, n (%)48 (12%)37 (12%)11 (14%) unknown, n (%)4 (1%)3 (1%)1 (1%)Tobacco use, n (%)51 (13%)44 (14%)7 (9%)0.34Type of Gynecologic Cancer0.75 Cervical cancer, n (%)13 (3%)10 (3%)3 (4%) Uterine Cancer, n (%)261 (66%)211 (66%)50 (65%) Ovarian Cancer, n (%)82 (21%)65 (20%)17 (22%) Vulvar cancer, n (%)40 (10%)34 (11%)6 (8%) Vaginal cancer, n (%)2 (1%)1 (0%)1 (1%)Stage0.11 I, n (%)275 (69%)230 (72%)45 (58%) II, n (%)23 (6%)16 (5%)7 (9%) III, n (%)75 (19%)55 (17%)20 (26%) IV, n (%)25 (6%)20 (6%)5 (6%)Procedure0.81 Laparoscopic adnexal surgery6 (2%)6 (2%)0 (0%) RA-BSO/adnexal surgery11 (3%)8 (2%)3 (4%) RA-TLH4 (1%)3 (1%)1 (1%) RA-TLH/BSO/adnexal surgery260 (65%)208 (65%)52 (68%) Total Abdominal Hysterectomy75 (19%)61 (19%)14 (18%) Vulvectomy42 (11%)35 (11%)7 (9%)Comorbidities T1 or T2 Diabetes, n (%)134 (34%)100 (31%)34 (44%)0.033 Tobacco use, n (%)51 (13%)44 (14%)7 (9%)0.34 Asthma, n (%)89 (22%)65 (20%)24 (31%)0.047 COPD, n (%)31 (8%)21 (7%)10 (13%)0.09 OSA, n (%)71 (18%)52 (16%)19 (25%)0.09 Hypertension, n (%)266 (67%)213 (66%)53 (69%)0.79 Hyperlipidemia, n (%)244 (61%)189 (59%)55 (71%)0.05 Congestive heart failure, n (%)27 (7%)19 (6%)8 (10%)0.20 Coronary artery disease, n (%)42 (11%)30 (9%)12 (16%)0.15 Myocardial infarction, n (%)15 (4%)10 (3%)5 (6%)0.18 Atrial fibrillation, n (%)42 (11%)31 (10%)11 (14%)0.30Documented allergies and reactionsNumber of allergies, median (IQR)1.0 (0.0, 3.0)1.0 (0.0, 2.0)5.0 (4.0, 7.0)<0.001Allergy type0.005 Drug class, n (%)33 (11%)31 (15%)2 (3%) Drug, n (%)189 (66%)127 (60%)62 (81%) Drug ingredient, n (%)3 (1%)3 (1%)0 (0%) Not documented, n (%)63 (22%)50 (24%)13 (17%)Severity of reaction for allergy<0.001 Low, n (%)55 (19%)53 (25%)2 (3%) Medium, n (%)64 (22%)39 (18%)25 (32%) High, n (%)100 (35%)65 (31%)35 (45%) Not documented, n (%)69 (24%)54 (26%)15 (19%)Type of reaction for allergy0.94 Allergy, n (%)199 (69%)145 (69%)54 (70%) Contradiction, n (%)2 (1%)1 (0%)1 (1%) Intolerance, n (%)53 (18%)40 (19%)13 (17%) Unspecified, n (%)3 (1%)2 (1%)1 (1%) Not documented, n (%)31 (11%)23 (11%)8 (10%)Postoperative Complications Pulmonary Embolism, n (%)1 (0%)1 (0%)0 (0%)1.00 Deep Vein Thrombosis, n (%)0 (0%)0 (0%)0 (0%) Surgical Site Infection, n (%)8 (2%)7 (2%)1 (1%)1.00 Wound Dehiscence, n (%)15 (4%)14 (4%)1 (1%)0.32 Pelvic Abscess, n (%)8 (2%)6 (2%)2 (3%)0.65 Hemoperitoneum/Bleeding, n (%)5 (1%)3 (1%)2 (3%)0.25 Pneumonia, n (%)0 (0%)0 (0%)0 (0%) UTI, n (%)6 (2%)3 (1%)3 (4%)0.09 Port-site hernia, n (%)0 (0%)0 (0%)0 (0%) Unanticipated ICU stay (during surgical admission), n (%)4 (1%)3 (1%)1 (1%)0.58 Other, n (%)20 (5%)14 (4%)6 (8%)0.24 Readmission, n (%)17 (4%)14 (4%)3 (4%)1.00 Any post-op complication, n (%)57 (14%)43 (13%)14 (18%)0.28RA-BSO: Robotic assisted bilateral salpingo-oophorectomy.RA-TLH: Robotic assisted total laparoscopic hysterectomy.RA-TLH/BSO: Robotic assisted total laparoscopic hysterectomy and bilateral salpingo-oophorectomy. Demographics, oncologic history, comorbidities, documented allergies and postoperative complications by allergy cohort. RA-BSO: Robotic assisted bilateral salpingo-oophorectomy. RA-TLH: Robotic assisted total laparoscopic hysterectomy. RA-TLH/BSO: Robotic assisted total laparoscopic hysterectomy and bilateral salpingo-oophorectomy. Table 2Rates of mental health diagnoses and psychiatric medication summary by allergy cohort.Overall (n = 398)Non-polyallergy (n = 321)Polyallergy (n = 77)p-valueGeneral anxiety disorder, n (%)109 (27%)80 (25%)29 (38%)0.032Panic disorder, n (%)11 (3%)6 (2%)5 (6%)0.042Obsessive compulsive disorder, n (%)8 (2%)5 (2%)3 (4%)0.19Post traumatic stress disorder, n (%)6 (2%)5 (2%)1 (1%)1.00Major depressive disorder, n (%)139 (35%)104 (32%)35 (45%)0.034Adjustment disorder, n (%)18 (5%)13 (4%)5 (6%)0.36Bipolar disorder I/II, n (%)8 (2%)6 (2%)2 (3%)0.65Personality disorder, n (%)2 (1%)2 (1%)0 (0%)1.00Schizoaffective disorder, n (%)1 (0%)0 (0%)1 (1%)0.19Schizophrenia, n (%)1 (0%)0 (0%)1 (1%)0.19Delusional disorder, n (%)1 (0%)1 (0%)0 (0%)1.00Fibromyalgia, n (%)24 (6%)17 (5%)7 (9%)0.28Irritable bowel syndrome, n (%)19 (5%)14 (4%)5 (6%)0.39Somatization disorder, n (%)0 (0%)0 (0%)0 (0%)Conversion disorder, n (%)0 (0%)0 (0%)0 (0%) Psychiatric medications at time of preoperative visitAnti-depressants, n (%)96 (24%)71 (22%)25 (32%)0.07Anxiolytic, n (%)91 (23%)68 (21%)23 (30%)0.13Antipsychotic, n (%)10 (3%)5 (2%)5 (6%)0.027Mood stabilizer, n (%)1 (0%)1 (0%)0 (0%)1.00Stimulant, n (%)0 (0%)0 (0%)0 (0%) Rates of mental health diagnoses and psychiatric medication summary by allergy cohort. Shows the difference in psychiatric diagnoses between the polyallergy (orange) and non-polyallergy (blue) cohorts. Patients with polyallergy had significantly more diagnoses of major depressive disorder (p = 0.034) and general anxiety disorder (p = 0.032). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) The polyallergy cohort had more ancillary service encounters than the non-polyallergy cohort (median 4, IQR 2–6 vs 2, IQR 1–5, p = 0.001) (Table 3). Those with polyallergy had more ED visits (median 1, IQR 0–2 vs 0, IQR 0–1, p = 0.001), primary care visits (median 2, IQR 0–4 vs 1, IQR 0–3, p = 0.024), and inpatient admissions (median 0, IQR 0–1 vs 0, IQR 0–0, p = 0.040). The number of urgent care visits was not significant between the two cohorts (median 0, IQR 0–0 vs 0, IQR 0–0, p = 0.230). There was no significant difference in postoperative complications or readmission rates between the cohorts (Table 1).Table 3Healthcare utilization by allergy cohort.OverallNon-polyallergyPolyallergyp-valueNumber of ED visits, median (IQR)0 (0–1)0 (0–1)1 (0–2)0.001Number of urgent care visits, median (IQR)0 (0–0)0 (0–0)0 (0–0)0.23Number of primary care visits, median (IQR)1 (0–3)1 (0–3)2 (0–4)0.020Number of inpatient admissions, median (IQR)0 (0–1)0 (0–0)0 (0–1)0.040Total healthcare utilization, median (IQR)2 (1–5)2 (1–5)4 (2–6)0.001 Table 3Healthcare utilization by allergy cohort.OverallNon-polyallergyPolyallergyp-valueNumber of ED visits, median (IQR)0 (0–1)0 (0–1)1 (0–2)0.001Number of urgent care visits, median (IQR)0 (0–0)0 (0–0)0 (0–0)0.23Number of primary care visits, median (IQR)1 (0–3)1 (0–3)2 (0–4)0.020Number of inpatient admissions, median (IQR)0 (0–1)0 (0–0)0 (0–1)0.040Total healthcare utilization, median (IQR)2 (1–5)2 (1–5)4 (2–6)0.001 Healthcare utilization by allergy cohort. On univariate analysis, patients with polyallergy (RR = 0.31, CI 0.11–0.51p = 0.002), diabetes (RR = 0.3, CI 0.14–0.46, p < 0.001), asthma (RR = 0.48, 0.29–0.66, p < 0.001), COPD (RR = 0.39, CI 0.11–0.68, p = 0.01), hypertension (RR = 0.37, CI 0.21–0.53, p < 0.001), hyperlipidemia (RR = 0.45, CI 0.29–0.61, p < 0.001), CHF (RR = 0.52, CI 0.51–0.83, p < 0.001), CAD (RR = 0.34 0.08–0.59, p = 0.01), atrial fibrillation (RR = 0.41, CI 0.15–0.66, p = 0.002) and psychiatric diagnoses (RR = 0.42, CI 0.27–0.57, p < 0.001), were noted to have significantly higher ancillary service utilization (Table 4). The diagnosis of vaginal cancer and stage II disease were associated with significantly higher ancillary service utilization (p = 0.013 and p = 0.027). Patients with postoperative complications including pulmonary embolism (RR = 1.38, CI 0.18–2.94, p = 0.08), wound dehiscence (RR = 0.43, CI 0.02–0.85, p = 0.038), pelvic abscess (RR = 0.98, CI 0.43–1.53, p = 0.001), and complications listed as other (RR = 0.37, CI 0.01–0.73, p = 0.044) were more likely to have increased healthcare utilization (Table 4).Table 4Univariate regression of total healthcare utilization.Coeff95% C.I.p-valueAge0.002−0.0030.0080.45BMI−0.001−0.0090.0070.88Race OtherReference White/Caucasian−0.39−0.64−0.160.001Marital Status SingleReference Married−0.32−0.53−0.110.003 Widowed−0.17−0.440.100.22 Divorced−0.22−0.510.070.14T1 or T2 Diabetes0.30.140.46<0.001Tobacco use0.04−0.190.270.75Asthma0.480.290.66<0.001COPD0.390.110.680.008OSA0.08−0.130.280.46Hypertension0.370.210.53<0.001Hyperlipidemia0.450.290.61<0.001Congestive heart failure0.520.510.830.001Coronary artery disease0.340.080.590.008Myocardial infarction0.07−0.340.480.75Atrial fibrillation0.410.150.660.002Psychiatric Diagnosis0.420.270.57<0.001 Type of Gynecologic Cancer Uterine cancerReference Cervical cancer0.12−0.320.560.59 Ovarian cancer0.11−0.080.310.27 Vulvar cancer0.21−0.050.470.12 Vaginal cancer1.400.292.510.013Stage IReference II0.380.040.720.027 III0.15−0.050.350.15 IV0.18−0.150.500.28Polyallergy status Non-PolyallergyReference Polyallergy0.310.110.510.002Postoperative complications Pulmonary Embolism1.38−0.182.940.08 Deep Vein Thrombosis–––– Surgical Site Infection−0.14−0.690.420.63 Wound Dehiscence0.430.020.850.038 Pelvic Abscess0.980.431.530.001 Hemoperitoneum/Bleeding0.17−0.530.880.63 Pneumonia–––– UTI0.28−0.360.930.39 Port-site hernia–––– Unanticipated ICU stay (during surgical admission)0.54−0.241.330.18 Other0.370.010.730.044 Readmission0.990.621.37<0.0001 Any post-op complication0.350.120.570.002 Table 4Univariate regression of total healthcare utilization.Coeff95% C.I.p-valueAge0.002−0.0030.0080.45BMI−0.001−0.0090.0070.88Race OtherReference White/Caucasian−0.39−0.64−0.160.001Marital Status SingleReference Married−0.32−0.53−0.110.003 Widowed−0.17−0.440.100.22 Divorced−0.22−0.510.070.14T1 or T2 Diabetes0.30.140.46<0.001Tobacco use0.04−0.190.270.75Asthma0.480.290.66<0.001COPD0.390.110.680.008OSA0.08−0.130.280.46Hypertension0.370.210.53<0.001Hyperlipidemia0.450.290.61<0.001Congestive heart failure0.520.510.830.001Coronary artery disease0.340.080.590.008Myocardial infarction0.07−0.340.480.75Atrial fibrillation0.410.150.660.002Psychiatric Diagnosis0.420.270.57<0.001 Type of Gynecologic Cancer Uterine cancerReference Cervical cancer0.12−0.320.560.59 Ovarian cancer0.11−0.080.310.27 Vulvar cancer0.21−0.050.470.12 Vaginal cancer1.400.292.510.013Stage IReference II0.380.040.720.027 III0.15−0.050.350.15 IV0.18−0.150.500.28Polyallergy status Non-PolyallergyReference Polyallergy0.310.110.510.002Postoperative complications Pulmonary Embolism1.38−0.182.940.08 Deep Vein Thrombosis–––– Surgical Site Infection−0.14−0.690.420.63 Wound Dehiscence0.430.020.850.038 Pelvic Abscess0.980.431.530.001 Hemoperitoneum/Bleeding0.17−0.530.880.63 Pneumonia–––– UTI0.28−0.360.930.39 Port-site hernia–––– Unanticipated ICU stay (during surgical admission)0.54−0.241.330.18 Other0.370.010.730.044 Readmission0.990.621.37<0.0001 Any post-op complication0.350.120.570.002 Univariate regression of total healthcare utilization. On multivariate logistic regression, polyallergy remained a statistically significant independent predictor for increased total ancillary service utilization (RR = 0.21, CI 0.02–0.39, p = 0.03), as was psychiatric diagnosis (RR = 0.34, CI 0.19–0.48, p < 0.001) (Table 5). White/Caucasian race (RR = 0.44, CI − 0.65–−0.22, p < 0.001), stage II disease (RR = 0.33, CI 0.02–0.64, p = 0.036), and the comorbidities of hypertension (RR = 0.19, CI 0.03–0.36, p = 0.019) and hyperlipidemia (RR = 0.29, CI 0.14–0.46, p < 0.001) were also statistically significant independent predictors for increased total ancillary service utilization. Any postoperative complication was an independent predictor for total ancillary service utilization (RR = 0.29, CI 0.08–0.50, p = 0.006).Table 5Multivariate regression of variables associated with ancillary service utilization.Coef95% C.I.p-valueRace OtherReference White/Caucasian0.44−0.65−0.22<0.001Hypertension0.190.030.360.019Hyperlipidemia0.290.140.46<0.001Congestive heart failure0.280.000.570.05Psychiatric Diagnosis0.340.190.48<0.001Stage IReference II0.330.020.640.036 III0.08−0.110.260.42 IV0.16−0.140.460.29Polyallergy status Non-PolyallergyReference Polyallergy0.210.020.390.027Any post-op complication0.290.080.500.006 Table 5Multivariate regression of variables associated with ancillary service utilization.Coef95% C.I.p-valueRace OtherReference White/Caucasian0.44−0.65−0.22<0.001Hypertension0.190.030.360.019Hyperlipidemia0.290.140.46<0.001Congestive heart failure0.280.000.570.05Psychiatric Diagnosis0.340.190.48<0.001Stage IReference II0.330.020.640.036 III0.08−0.110.260.42 IV0.16−0.140.460.29Polyallergy status Non-PolyallergyReference Polyallergy0.210.020.390.027Any post-op complication0.290.080.500.006 Multivariate regression of variables associated with ancillary service utilization. Discussion In our study examining ancillary service use among postoperative patients with gynecologic malignancies for one year following surgery, polyallergy status was an independent risk factor for increased ancillary service use, including ED visits, primary care visits, and inpatient admissions. Additionally, the polyallergy cohort had significantly more mental health diagnoses and antipsychotic medications compared to the non-polyallergy cohort. Multiple prior studies have linked polyallergy to increased healthcare utilization. For example, Jimenez et al. reported that any patient with a clinical encounter who also had polyallergy was 2.5 times more likely to use ancillary services, specifically emergency visits and inpatient services. These patients had 2.5 times the odds of having at least one ED visit and 4 times the odds of having at least one inpatient admission (Jimenez et al., 2019). Similarly, a study completed by Blumenthal et al. reviewed antibiotic use and healthcare utilization in patients with a reported penicillin allergy found that the penicillin allergy cohort had significantly more emergency, primary care, and specialist outpatient visits per year, and those with multiple drug intolerance syndrome (≥3 reported drug allergies) had a higher incidence of all healthcare utilization including inpatient admissions (Blumenthal et al., 2018). Our study was consistent with these findings showing that postoperative patients with gynecologic malignancy and polyallergy have significantly higher rates of ancillary service utilization including ED and primary care visits as well as inpatient admissions. Any postoperative complication was an independent predictor for total ancillary service utilization. This contributes to the literature by examining a large cohort of patients at a small academic institution, specifically with a gynecologic malignancy after surgery, when healthcare utilization may be high due to both cancer and postoperative reasons. Polyallergy has also been linked to psychiatric diagnoses. A study performed by Park et al. assessed polyallergy in relation to non-epileptic seizures and found that it is predictive of non-epileptic seizures or any psychosomatic illness (Park et al., 2014). Another study performed by Katayama et al. found that patients with a behavioral health disorder who underwent surgery for a non-gynecologic cancer diagnosis were more likely to have poor surgical outcomes, postoperative complications, increased length of stay, and 90-day readmissions.5 In our study examining a gynecologic malignancy cohort, patients with polyallergy had a significantly higher chance of having a mental health diagnosis and were more likely to be on an antipsychotic medication. In this cohort, having a psychiatric diagnosis remained a significant independent predictor of total healthcare utilization. Our study, however, does not review postoperative complications or length of stay. Katayama et al. reviewed a 90-day readmission rate postoperatively; whereas we reviewed readmission rates for a year postoperatively (Katayama et al., 2024). It is difficult to compare our study to this 90-day readmission rate given our collection timeline spanned a year postoperatively. Our study findings confirm the previous literature findings of polyallergy as a predictor of healthcare utilization and expand on this work by demonstrating that polyallergy remains an independent risk factor for increased ancillary service utilization even in the complex gynecologic oncology surgical population. This association remains after adjusting for psychiatric diagnoses, suggesting it may function as an easily identifiable EMR-based screening tool for health care needs rather than just a surrogate for mental health diagnoses. Identifying polyallergy is easily performed in the electronic medical record, can be done prior to planned interventions, and based on our data may correlate with undiagnosed mental illness allowing extrapolation of polyallergy as a potential screening tool for additional postoperative needs after gynecologic oncologic surgery. Ensuring that these patients have resources available with behavioral health services could impact unnecessary utilization of healthcare services postoperatively. This study suggests that there is an opportunity to consider additional supportive resources for gynecologic cancer patients undergoing surgery. There may be a role for additional mental or behavioral health interventions. Our institution is currently identifying patients with polyallergy undergoing surgical intervention for gynecologic cancer and placing referrals for behavioral health services outpatient. We seek to determine if ancillary service use can be significantly decreased with this quality improvement intervention. Strengths of the study include a large sample size with 398 patients total and 77 patients meeting criteria for polyallergy. Postoperative ancillary service use and its association with polyallergy has not previously been assessed in the gynecologic cancer population. This study included a single electronic medical record with longitudinal data on patient follow-up and included a long follow-up period of one year. We also examined ancillary service use across a variety of settings including ED, urgent care, primary care, psychiatry/psychology, and inpatient admissions. Limitations of the study include those inherent to a single-institution retrospective cohort study. It is also limited by the accuracy of what is documented in the electronic medical record. Allergies can be reported by patients and documented but not confirmed by biochemical testing. We were unable to account for out of network encounters except for what was shown in the Care Everywhere feature in Epic, which could limit the amount of healthcare utilization that was assessed and bias our results. We also did not take into account patient telephone calls. These encounters take a significant amount of time for clinical staff that is currently not billable. Not including these encounters may be further underestimating utilization. Patients who died during the study were not included, potentially decreasing the amount of ancillary service use. Psychiatric medication use was collected at the time of preoperative visit and did not include medications that had been prescribed in the postoperative period. In addition, some of these medications could have been prescribed for other purposes outside of mental health diagnoses. In summary, this study demonstrates that polyallergy independently predicts greater use of ancillary healthcare services among patients with gynecologic cancer in the postoperative period. Patients with polyallergy were also significantly more likely to carry a behavioral health diagnosis, suggesting a meaningful intersection between allergy reporting and psychosocial factors. Together, these findings highlight an important opportunity for future research focused on behavioral interventions and their potential to reduce unnecessary utilization and improve postoperative care. CRediT authorship contribution statement Kaitlyn Coppetti: Writing – review & editing, Writing – original draft, Visualization, Investigation. Madeline Petty: Investigation. Allison R. Walker: Writing – review & editing, Writing – original draft, Visualization, Investigation, Conceptualization. Luis Alvarado: Formal analysis. Sarah P. Huepenbecker: Writing – review & editing, Supervision, Conceptualization. Nicholas Taylor: Supervision, Conceptualization. Ashley Graul: Writing – review & editing, Supervision, Conceptualization. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.