1
Abstract Identifying Patients at Risk for Postsurgical Opioid-Related Adverse Events 35 Kathy Nipper-Johnson, BSN, RN, CCM 1 ; Harold S. Minkowitz, MD 1 ; Richard Scranton, MD, MPH 2 ; Aditya Raju, MS, BPharm 3 ; Akash Dandappanavar, PharmD, MPH, MA 3 ; Laura Menditto, MPH, MBA 4 1 Memorial Hermann Memorial City Medical Center, Houston, TX; 2 Pacira Pharmaceuticals, Inc., Parsippany, NJ; 3 Xcenda ® AmerisourceBergen Consulting Services, Palm Harbor, FL; 4 Laura A. Menditto, LLC, Newtown, PA Purpose: Identify opioid-related adverse event (ORADE) risk factors, derive a risk score to identify high-risk patients, and evaluate potential benefits of targeting high-risk patients for strategies aimed at reducing ORADEs. Methodology: Administrative claims data were analyzed to identify adults who received opioids following gastro-intestinal or orthopedic surgeries. Logistic regression was used to stratify patients according to risk. Generalized linear and binomial regression models were used to compare cost and length of stay (LOS) according to risk. Results: Overall, ORADEs occurred in 551 (11.3%) of the 4,888 patients who received postsurgical opioids and were more frequent in the high-risk group (22%) than the low-risk group (6.9%). Significant risk factors included age, gender, pre-surgical opioid use and several comorbidities such as diabetes and benign prostatic hypertrophy. Higher risk patients had a longer LOS (mean — 7.2 vs. 4.1 days, P<0.0001) and greater costs (mean— $21,292 vs. $14,849, p<0.0001). Alternative pain management strategies intended to decrease ORADE incidence from 25%–100% could reduce LOS by 74 – 294 days and decrease costs by $255,811– $1,023,243 per 1000 patients. Findings are limited by the clinical data available in a claims database from a single hospital system and examination of orthopedic and gastrointestinal procedures only. Perioperative Nursing Implications: Postsurgical ORADEs are common, have higher LOS and costs and can threaten patient safety. Alternative postoperative pain management strategies should be explored in higher risk patients. ■ More than 90 million surgeries are performed annually in the United States 2 ; 80% of patients experience pain after surgery. 3-5 ■ Opioids have demonstrated efficacy for pain relief after surgery and are often the analgesics of choice for postoperative pain, but their use is frequently accompanied by opioid-related adverse drug events (ORADEs) and other negative consequences, including increased mortality. 6,7 ■ Previous studies have reported the occurrence of ORADEs in post-surgical patients with associated higher costs, readmissions, and longer length of stay (LOS). 8-15 ■ Previously demonstrated patient-specific ORADE risk factors include age, gender, race/ethnicity, smoking status, obesity, and comorbid diagnoses such as obstructive sleep apnea (OSA), chronic obstructive pulmonary disease (COPD), renal and hepatic function, cardiac dysrhythmia, degenerative joint disease (DJD), and benign prostatic hyperplasia (BPH). 6,12,13,16,17 Efforts to identify specific segments of the surgical population at increased risk of ORADEs may provide an opportunity to avoid ORADEs that negatively impact clinical and economic outcomes. Data Source ■ This retrospective cohort study utilized administrative data from the Memorial Hermann Hospital System, the largest non-profit healthcare system in Texas, USA, comprising 11 hospitals accounting for approximately 3,500 inpatient beds. Study Design and Sample Selection ■ Patients at least 18 years old receiving opioids after common orthopedic and soft tissue surgeries (Table 1) between 01/01/2010 through 12/31/2010 were followed from admission date through 30 days post-discharge. ■ Post-surgical opioid pain management was defined by the administration of parenteral or oral opioid analgesics on or after the procedure date but before discharge. Opioid analgesics included morphine, oxycodone, hydromorphone, fentanyl, meperidine, codeine, methadone, propoxyphene, and hydrocodone. ■ Respiratory (resp), gastrointestinal (GI), and genitourinary (GU) ORADEs were identified using ICD-9 CM diagnosis codes (Table 2) and patients were placed into cohorts based on whether they experienced any ORADEs or not. ■ Putative ORADE risk factors considered in this analysis included age, gender, opioid use prior to surgery, and specific comorbidities considered plausible based on previous research 6,12,13,16,17 and biologic plausibility. Primary Endpoints ■ ORADE risk factors and impact of ORADE risk on subsequent outcomes, specifically: – Length of Stay (LOS), defined as the time in days from index admission date to discharge date; – Total hospitalization costs, as measured from admission index date to discharge date. Statistical Analysis ■ Gender-specific multivariate logistic modeling was used to evaluate risk factors and a risk-scoring model was developed that was composed of the final male- and female-specific models. A risk score was calculated for each patient by summing the β-coefficients for each risk factor present in the patient. ■ A threshold for identifying high-risk from low risk patients was determined based on the risk score with maximum sum of sensitivity + specificity and the utility of the final model for predicting adverse events was evaluated using receiver operator characteristics (ROC) analysis and by calculating sensitivity, specificity and positive predictive value (PPV). ■ Differences in total hospitalization cost and LOS between high-risk and low-risk patients were assessed using generalized linear and binomial regression models, respectively. ■ Of the 4,888 patients identified for analysis in Table 3, 11.3% experienced an ORADE with GI ORADEs being the most common (7.1%). The incidence of ORADEs was almost twice as high in males compared to females (16.3% vs. 8.5%). ■ The distribution of risk factors across patients with and without an ORADE is presented in Table 4. Factors found to have a statistically significant association with ORADE incidence in univariate analysis were selected for the final risk score models. ■ The best risk score model, illustrated in Figure 1, was a composite of gender by procedure specific models and had an AUC of 0.726, which is considered an acceptable accuracy level. This model correctly identified 56.6% of patients who experienced ORADEs and 74.5% of patients who did not. Only 22.2% of patients classified as high-risk had a corresponding ORADE, however this value reflects not only the utility of the risk score but is also partly a function of the overall ORADE incidence in the patient population. OVERALL SURGICAL POPULATION Age ≥65 1.13 0.1181 Prior opioid use 1.26 0.2294 Obesity 0.46 – 0.7717 DJD 0.47 – 0.7632 COPD 2.10 0.743 CHF 2.17 0.7726 BPH 5.99 1.7907 Atherosclerosis 0.84 – 0.1767 Cardiac dysrhythmia 5.20 1.6492 Diabetes 1.09 0.0883 Regional enteritis 0.57 – 0.566 Diverticulitis 2.21 0.7946 Ulcerative colitis 5.25 1.6587 OSA 1.84 0.6081 Male GI Soft Tissue Odds Component Risk Factor Ratio Score High Risk Threshold ≥ 0.4358 AUC 0.727 Sensitivity 51.8% Specificity 79.8% PPV 35.8% Age ≥65 2.36 0.8607 Prior opioid use 0.98 – 0.0156 Obesity 1.19 0.1732 DJD 1.75 0.5621 COPD 3.30 1.1925 CHF 1.53 0.4242 BPH NA NA Atherosclerosis 1.02 0.0231 Cardiac dysrhythmia 3.21 1.1664 Diabetes 1.24 0.2133 Regional enteritis 7.96 2.0745 Diverticulitis 2.12 0.75 Ulcerative colitis 1.43 0.3563 OSA 1.34 0.2935 Female GI Soft Tissue Odds Component Risk Factor Ratio Score High Risk Threshold ≥ 0.4511 AUC 0.698 Sensitivity 58.8% Specificity 70.4% PPV 17.2% Age ≥65 0.80 – 0.2276 Prior opioid use 1.20 0.1832 Obesity 0.73 – 0.3109 DJD 0.70 – 0.3619 COPD 4.61 1.5284 CHF 3.05 1.1148 BPH 4.51 1.5057 Atherosclerosis 1.59 0.4668 Cardiac dysrhythmia 9.30 2.2295 Diabetes 1.45 0.3742 Regional enteritis – 0 Diverticulitis – 0 Ulcerative colitis – 0 OSA 1.05 0.0512 Male Orthopedic Odds Component Risk Factor Ratio Score High Risk Threshold ≥ 0.8872 AUC 0.779 Sensitivity 56.9% Specificity 83.9% PPV 33.9% Age ≥65 0.81 – 0.2107 Prior opioid use 1.44 0.3671 Obesity 0.61 – 0.5006 DJD 0.71 – 0.3401 COPD 2.64 0.9717 CHF 2.70 0.9927 BPH 4.36 1.4734 Atherosclerosis 1.61 0.4739 Cardiac dysrhythmia 8.72 2.1651 Diabetes 1.25 0.2216 Regional enteritis – 0 Diverticulitis – 0 Ulcerative colitis – 0 OSA 1.91 0.6449 Female Orthopedic Odds Component Risk Factor Ratio Score High Risk Threshold ≥ 0.3671 AUC 0.754 Sensitivity 71.4% Specificity 73.8% PPV 12.7% Population AUC Sensitivity Specificity PPV Male 0.741 53.0% 81.1% 35.3% Female 0.707 60.5% 71.1% 16.3% GI Soft Tissue 0.702 55.5% 73.3% 22.4% Orthopedic 0.783 62.0% 77.9% 20.3% All Patients 0.726 56.6% 74.5% 22.0% Figure 1. Final Risk Score Model High Risk (N=1,420) GU ORADE GI ORADE Respiratory ORADE Any ORADE Low Risk (N=3,468) 0% 10% 20% 30% 40% All Patients (N=4,888) 4.1% 0.7% 12.7% 4.9% 7.8% 1.9% 22.0% 6.9% High Risk (N=425) GU ORADE GI ORADE Respiratory ORADE Any ORADE Low Risk (N=1,309) 0% 10% 20% 30% 40% Males (N=1,734) 8.9% 1.4% 21.2% 6.7% 11.5% 3.1% 32.3% 10.2% High Risk (N=995) GU ORADE GI ORADE Respiratory ORADE Any ORADE Low Risk (N=2,159) 0% 10% 20% 30% 40% Females (N=3,154) 2.0% 0.2% 9.1% 3.8% 6.2% 1.1% 16.3% 4.9% Figure 2. Incidence of AEs by High vs. Low Risk Model Characterization Mean Length of Stay (Days) Low Risk No ORADE 0 2 4 6 8 10 12 14 High Risk No ORADE Low Risk ORADE High Risk ORADE Mean LOS by Risk × ORADE Status High Risk Low Risk 3.6 5.9 10.8 12.0 4.1 7.2 Mean Hospitalization Cost Low Risk No ORADE $0 $5,000 $10,000 $15,000 $20,000 $25,000 $30,000 $35,000 $40,000 High Risk No ORADE Low Risk ORADE High Risk ORADE Mean Hospitalization Costs by Risk × ORADE Status High Risk Low Risk $13,412 $17,776 $34,158 $33,781 $14,849 $21,292 Figure 3. Length of Stay (LOS) and Mean Hospitalization Costs by High vs. Low Risk Model Characterization LOS Reduction per 1,000 Patients (days) 100% 0 50 100 150 200 250 300 350 75% Effectiveness of High-risk Intervention 50% 25% 294 221 147 74 Savings per 1,000 Patients 100% $0 $200,000 $400,000 $600,000 $800,000 $1,000,000 $1,200,000 75% Effectiveness of High-risk Intervention 50% 25% $1,023,243 $767,432 $511,621 $255,811 Figure 4. Benefit of High-risk Intervention (per 1,000 patients in the overall surgical population) ■ A total of 1,420 patients (29.1%) were classified as high-risk (24.5% of men and 31.5% of women). Overall, 22.0% of high-risk patients experienced one or more ORADEs compared to only 6.9% of low-risk patients. In addition to predicting overall AEs as designed, the final composite risk score model was effective at predicting specific ORADEs (Figure 2) and 30-day readmissions (12.3% for high risk versus 9.1% for low risk population; p=0.0006). ■ Furthermore, high-risk patients tended to have higher hospitalization costs and longer LOS than low-risk patients [mean LOS=7.2 days vs. 4.1 days (p<0.0001) and mean cost=$21,292 vs. $14,849 (p<0.0001) for high-risk vs. low-risk patients, respectively (Figure 3)]. ■ Alternative pain management strategies intended to prevent ORADEs among high-risk patients have the potential to reduce LOS by 74 – 294 days per 1,000 surgical patients, assuming a decrease in ORADE incidence ranging from 25% – 100%, with accompanying cost savings of $255,811 – $1,023,243 (Figure 4). 01. The Joint Commission. Sentinel Event Alert: Safe Use of Opioids in Hospitals. 2012; issue 49. 02. Cullen KA, et al. Ambulatory surgery in the United States, 2006. Natl Health Stat Report. 2009(11):1-25. 03. Apfelbaum JL, Chen C, Mehta S, et al. Postoperative pain experience: results from a national survey suggest postoperative pain continues to be undermanaged. Anesth Analg. 2003;97:534-40. 04. Owen H, McMillan V, Rogowski D. Postoperative pain therapy: a survey of patients’ expectations and their experiences. Pain. 1990;41:303-7. 05. Warfield CA, Kahn CH. Acute pain management: programs in U.S. hospitals and experiences and attitudes among U.S adults. Anesthesiology. 1995;83:1090-4. 06. Adamson RT, Lew I, Beyzarov E, et al. Clinical and economic implications related to postsurgical analgesic devices. Hosp Pharm. 2011;46 (6Suppl1):S12-20. 07. Shang AB, Gan TJ. Optimizing postoperative pain management in the ambulatory patient. Drugs. 2003;63:855-67. 08. Caroll NV, Miederhoff PA, Cox FM, et al. Costs incurred by outpatient surgical centers in managing post operative nausea and vomiting. J Clin Anesth. 1994;6:363-9. 09. Coley KC, Williams BA, DaPos SV, et al. Retrospective evaluation of unanticipated admissions and readmissions after same day surgery and associated costs. J Clin Anesth. 2002;14:349-53. 10. Oderda GM, Said Q, Evans RS, et al. Opioid-related adverse drug events in surgical hospitalizations: impact on costs and length of stay. Ann Pharmacother. 2007;41:400-7. 11. Taillefer MC, Carrier M, Bélisle S, et al. Prevalence, characteristics, and predictors of chronic nonanginal postoperative pain after a cardiac operation: a cross-sectional study. J Thorac Cardiovasc Surg. 2006;131:1274-80. 12. Kessler ER, Shah M, Gruschkus SG et al Cost and Quality Implications of Opioid-Based Postsurgical Pain Control Using Administrative Claims Data from a Large Health System: Opioid-Related Adverse Events and Their Impact on Clinical and Economic Outcomes. Pharmacotherapy. 2013;33(4):383-91. 13. Minkowitz HS, Gruschkus SK, Shah M et al: Risk Factors and Outcomes of Adverse Drug Events among Patients Receiving Opioid-Based Pain Management within a Large Health System. AJHP. Accepted with revisions March 19, 2013. 14. Oderda GM, Evans RS, Loyld A et al: Cost of opioid-related adverse drug events in surgical patients. J Pain Symptom Manage. 2003;25(3):276-83. 15. Oderda GM; Gan TJ; Johnson BH et al: Effect of Opioid-Related Adverse Events on Outcomes in Selected Surgical Patients. J Pain Palliat Care Pharmacother. 2013;27(1):62-70. 16. Craft J. Patient controlled analgesia: is it worth the painful prescribing process? Proc (Bayl Univ Med Cent). 2010;23(4):434-8. 17. Classen DC, Pestotnik SL, Evans RS, et al. Adverse drug events in hospitalized patients. Excess length of stay, extra costs, and attributable mortality. JAMA. 1997;277:301-6. ■ Opioids and their related adverse events threaten patient safety, lead to prolonged hospital stays, and increase the economic burden to hospital systems. ■ A risk score model can be used to identify high risk patients who are not only pre-disposed to opioid related adverse drug events based on their risk profile but who are also more likely to have additional negative downstream consequences such as longer hospital length of stay, higher readmission rates, and greater hospitalization costs. ■ Targeting high-risk patients for non-opioid pain management strategies, including locally acting, non-systemic medications and surgical interventions, may reduce opioid requirements and offer benefits to both patients and hospital systems. ■ Additional research is needed to further quantify the clinical and economic benefits of targeted interventions directed at high-risk populations and to validate risk models in surgical populations with different demographic and clinical profiles and among patients receiving other surgical procedures. Adverse Event ICD-9 Diagnosis Code Respiratory (Resp) Bradypnea 786.09 Pulmonary insufficiency following surgery and trauma 518.5 Respiratory complications 997.3 Hypoxemia 799.02 Gastrointestinal (GI) Constipation 564.09 Constipation – Narcotic E937.9 Ileus, postoperative (following surgery) 997.4 Paralytic Ileus 560.1 Nausea/vomiting 787.01 Nausea/vomiting following GI surgery 564.3 Genitourinary System (GU) Urinary retention 788.2 Oliguria 997.5 Presented at AORN Surgical Conference & Expo 2014, March 29 – April 2, 2014, Chicago, IL. Without Any With Any With Respiratory With With N ORADE ORADE ORADE GI ORADE GU ORADE Patients N % N % N % N % N % Overall Population All Patients 4,888 4,337 88.7% 551 11.3% 176 3.6% 349 7.1% 81 1.7% Male 1,734 1,451 83.7% 283 16.3% 90 5.2% 177 10.2% 56 3.2% Female 3,154 2,886 91.5% 268 8.5% 86 2.7% 172 5.5% 25 0.8% GI Soft Tissue Procedures All Patients 3,684 3,233 87.8% 451 12.2% 121 3.3% 321 8.7% 46 1.3% Male 1,221 1,003 82.2% 218 17.9% 52 4.3% 158 12.9% 32 2.6% Female 2,463 2,230 90.5% 233 9.5% 69 2.8% 163 6.6% 14 0.6% Orthopedic Procedures All Patients 1,204 1,104 91.7% 100 8.3% 55 4.6% 28 2.3% 35 2.9% Male 513 448 87.3% 65 12.7% 38 7.4% 19 3.7% 24 4.7% Female 691 656 94.9% 35 5.1% 17 2.5% 9 1.3% 11 1.6% ■ Majority of population are women explained by relative distribution of bariatric surgery in men vs. women. ■ Results from a single hospital system in Texas may not be fully generalizable to the broader US surgical population. Similarly, our study focused only on patients receiving orthopedic and gastrointestinal procedures and results may not be applicable to patients receiving other types of surgeries. ■ Limitations common to retrospective studies using administrative data include the potential for missing data or coding errors, possibility of under-coding of ICD-9 CM diagnosis codes for ORADEs and ORADE risk factors. ■ A lack of detailed data on opioid dosage/morphine equivalency and clinically defined patient characteristics such as body mass index (BMI), body surface area (BSA), and medical history limited our ability to evaluate other potentially significant ORADE risk factors. RESULTS INTRODUCTION METHODS CONCLUSIONS REFERENCES limitations Surgical Procedures N Total 4,888 GI Soft Tissue Procedures 3,684 Laparoscopic Cholecystectomy 1,724 Laparoscopic Gastric Bypass 663 Open Colectomy – Partial excision of large intestines 482 Other Partial Gastrectomy 285 Laparoscopic Colectomy – Partial excision of large intestines 251 Open Cholecystectomy 111 Open Gastric Bypass 68 Ileostomy Reversal 55 Open Colectomy –Total excision of large intestines 40 Laparoscopic Colectomy –Total excision of large intestines 5 Orthopedic Procedures 1,204 Hip Replacement–Total hip 634 Hip Fracture–Op Fx reduct w/int fix, femur 522 Hip Replacement–Partial hip 48 Table 1. Surgical Procedures Table 2. Definition of ORADEs Patients with Patients without All Patients ORADE ORADE Risk Factor N % N % N % p-value Age (mean, SD) 54.0 (18.8) 59.0 (18.3) 53.4 (18.8) <.0001 Gender (female) 3154 (64.5%) 268 (48.6%) 2886 (66.5%) <.0001 Opioid use prior to surgery 2108 (43.1%) 274 (49.7%) 1834 (42.3%) 0.0009 Obesity 1455 (29.8%) 124 (22.5%) 1331 (30.7%) <.0001 Degenerative Joint Disease (DJD) 743 (15.2%) 56 (10.2%) 687 (15.8%) 0.0005 Chronic obstructive pulmonary disease (COPD) 65 (1.3%) 20 (3.6%) 45 (1.0%) <.0001 Asthma* 234 (4.8%) 32 (5.8%) 202 (4.7%) 0.2336 Pulmonary hypertension* 0 (0.0%) 0 (0.0%) 0 (0.0%) – Congestive heart failure 278 (5.7%) 68 (12.3%) 210 (4.8%) <.0001 Benign prostatic hypertrophy 94 (1.9%) 45 (8.2%) 49 (1.1%) <.0001 Coronary atherosclerosis 366 (7.5%) 61 (11.1%) 305 (7.0%) 0.0007 Cardiac dysrhythmias 170 (3.5%) 66 (12.0%) 104 (2.4%) <.0001 Hypertension* 2211 (45.2%) 249 (45.2%) 1962 (45.2%) 0.9830 Dementia* 168 (3.4%) 21 (3.8%) 147 (3.4%) 0.6087 Depression* 56 (1.2%) 8 (1.5%) 48 (1.1%) 0.4733 Diabetes 1062 (21.7%) 144 (26.1%) 918 (21.2%) 0.0077 Irritable Bowel Syndrome (IBS)* 15 (0.3%) 1 (0.2%) 14 (0.3%) 0.5721 Regional enteritis (large and small intestine) 22 (0.5%) 7 (1.3%) 15 (0.4%) 0.0023 Diverticulitis 340 (7.0%) 82 (14.9%) 258 (6.0%) <.0001 Ulcerative colitis 28 (0.6%) 8 (1.5%) 20 (0.5%) 0.0037 GERD* 792 (16.2%) 91 (16.5%) 701 (16.2%) 0.8326 Obstructive sleep apnea 294 (6.0%) 43 (7.8%) 251 (5.8%) 0.0607 *Not included in final risk score model Table 4. Evaluation of Potential ORADE Risk Factors Table 3. Distribution of ORADEs by Gender and Procedure
