Admission NarxCare Narcotic Scores Are Associated With Increased Odds of Readmission and Prolonged Length of Hospital Stay After Primary Elective Total Knee Arthroplasty : JAAOS Global Research & Reviews

Journal Logo

Research Article

Admission NarxCare Narcotic Scores Are Associated With Increased Odds of Readmission and Prolonged Length of Hospital Stay After Primary Elective Total Knee Arthroplasty

Galivanche, Anoop R. MD, MHS; Zhu, Justin BA; Mercier, Michael R. BA; McLean, Ryan BA; Wilhelm, Christopher V. MD; Varthi, Arya G. MD; Grauer, Jonathan N. MD; Rubin, Lee E. MD

Author Information
JAAOS: Global Research and Reviews 6(12):e22.00040, December 2022. | DOI: 10.5435/JAAOSGlobal-D-22-00040
  • Open

Abstract

In the United States, rates of prescription opioid utilization have increased substantially over the past several decades.1 Although alternatives are becoming increasingly widespread, opioids remain a common form of management for knee pain.2-4 As the opioid epidemic has unfolded, the use of such medications has become a focal point of medical and legislative reform in an effort to curb the rising prevalence of opioid-related mortality.1,5

Several studies have reported preoperative opioids to be associated with adverse perioperative outcomes after total knee arthroplasty (TKA).6-9 Rozell et al.7 and Chen et al.6 reported increases in length of stay, whereas Bedard et al8 reported greater risk for early revision. The study by Rozell et al. used data in the electronic medical record (EMR) to measure opioid consumption.10 Typically collected by patient self-reported opioid use, such data are potentially subject to response bias or erroneous reporting. Chen et al. and Bedard et al. used administrative claims data to measure opioid consumption, which can be limited by recording biases, clerical errors, and temporal changes in coding.11,12 These studies did not examine the relationship between preoperative narcotic use and differential patient satisfaction after joint surgery. Prior evidence has remained limited to spine surgery literature.13

Many institutions now use state pharmacy databases to monitor narcotic use by their patients. One such database is the Prescription Drug Monitoring Program, which requires all pharmacies to routinely report prescription data for scheduled medications, including opioids.13 NarxCare (Appriss Health) is a platform that facilitates the use of state databases and then quantifies a score for overall opioid and benzodiazepine medication use through a proprietary tool.

The NarxCare narcotic score is weighted for recent activity and is influenced by the following parameters: number of prescribers, number of pharmacies, morphine milligram equivalents, and overlapping prescriptions for each patient.13 Scores range from 0 (suggesting opioid naivete) to 999, with higher numbers for each aforementioned parameter, resulting in a higher NarxCare narcotic score.

The present study explores the predictive utility of the Narx Score system for complications after primary TKA. To accomplish this investigation, we examined a cohort of 1,136 elective TKA procedures performed at our institution between October 2017 and May 2020.

Methods

Study Population

An orthopaedic surgery registry was retrospectively compiled at a large academic medical center that included data on patient characteristics, 30-day adverse events (AEs), readmissions, revision surgeries, and mortality. This institutional registry additionally included Narx Score data, which is a component of the NarxCare platform. The NarxCare platform for Prescription Drug Monitoring Program analytics is integrated directly into the EMR system (Epic Systems) at our institution. Our institution's Human Investigations Committee exempted this study from further review based on its retrospective design that did not involve any further patient contact after the collection and deidentification of data.

Elective TKA patients from October 2017 to May 2020 were identified within the registry using the Current Procedural Terminology code 27447. Nonelective patients were excluded from the study. Patients with incomplete or missing demographic, comorbidity, or 30-day perioperative outcome data were further excluded from the study.

Cohort Characteristics

The following preoperative variables were extracted from the EMR: age, sex, weight and height (used to calculate the body mass index [BMI]), preoperative functional status, and American Society of Anesthesiology (ASA) classification.

The NarxCare narcotic score on admission was then extracted for each of the patients. Narx Scores range from 0 to 999, with the last digit of the score being the number of active narcotic scripts prescribed to a patient at a given time. These scores were binned into the following categories: 0, 1 to 99, 100 to 299, 300 to 499, and greater than 500.

Postoperative Outcomes

Postoperative variables included total length of hospital stay (LOS), 30-day AEs, readmissions, and revision surgeries. These were collected per the National Surgery Quality Improvement Program guidelines.

Perioperative AEs were classified as serious adverse events (SAEs) or minor adverse events (MAEs). An SAE was defined as the occurrence of any of the following: deep surgical site infection, sepsis, failure to wean, unplanned reintubation, postoperative renal failure, deep vein thrombosis, pulmonary embolism, cardiac arrest, myocardial infarction, or stroke. An MAE was defined as the occurrence of any of the following: superficial surgical site infection, wound dehiscence, pneumonia, urinary tract infection, or postoperative renal insufficiency. Patients that experienced either an SAE or MAE were further classified as having experienced any adverse event (AAE). Readmission and revision surgery within 30 days were collected as additional variables.

Statistical Analysis

The demographic characteristics, LOS, occurrence of AEs, readmission rate, revision surgery rate, and mortality of the different narcotic score cohorts were first compared. Demographic characteristics between the different cohorts were compared using chi-square tests. Total LOS was compared between the different cohorts using analysis of variance (ANOVA).

Multivariable logistic regression was used to compare the occurrence of AEs in the different narcotic score cohorts. Multivariable regressions controlled for age, sex, BMI, preoperative functional status, and ASA classification.

The 0 narcotic score cohort was used as the control group to identify opioid-naive TKA patients in these regressions. All statistical analyses were performed using STATA version 17 (StataCorp LP).

Results

Cohort Characteristics

In total, 1,136 patients met the criteria for inclusion in the study population. Of the total study population, 293 patients (25.79%) had an admission narcotic score of 0, 253 patients (22.27%) had an admission narcotic score between 001 and 099, 368 patients (32.39%) had an admissions narcotic score between 100 and 299, 161 patients (14.17%) had an admission narcotic score between 300 and 499, and 61 patients (5.37%) had an admission narcotic score of 500 or greater (Table 1 and Figure 1).

Table 1 - Demographic and Comorbid Characteristics of the Study Population
Narcotic Score on Admission Total Study Population 0 1-99 100-299 300-499 500+
Number Percent Number Percent Number Percent Number Percent Number Percent Number Percent P valuea
Total no. of patients 1136 100.00 293 25.79 253 22.27 368 32.39 161 14.17 61 5.37
Age (yr, mean ± SD) 65.2 ± 9.9 67.3± 10.0 65.8 ± 8.8 64.6 ± 9.6 64.0 ± 10.4 59.1 ± 11.2 <0.001
 ≤40 14 1.23 0 0.00 1 0.40 7 1.90 3 1.86 3 4.92
 41-50 51 4.49 11 3.75 9 3.56 16 4.35 6 3.73 9 14.75
 51-60 264 23.24 60 20.48 50 19.76 84 22.83 51 31.68 19 31.15
 61-70 461 40.58 106 36.18 115 45.45 161 43.75 58 36.02 21 34.43
 71-80 278 24.47 86 29.35 65 25.69 86 23.37 33 20.50 8 13.11
 >80 68 5.99 30 10.24 13 5.14 14 3.80 10 6.21 1 1.64
BMI (kg/m2, mean ± SD) 33.4 ± 7.3 32.3 ± 6.9 33.3 ± 6.4 34.2 ± 7.7 34.4 ± 7.4 31.2 ± 7.8 <0.001
 <18.5 6 0.53 2 0.68 0 0.00 1 0.27 2 1.24 1 1.64  
 18.5-25 102 8.98 34 11.60 16 6.32 33 8.97 7 4.35 12 19.67
 25-30 302 26.58 93 31.74 65 25.69 88 23.91 40 24.84 16 26.23
 30-40 531 46.74 121 41.30 137 54.15 174 47.28 76 47.20 23 37.70
 40-50 168 14.79 38 12.97 30 11.86 59 16.03 34 21.12 7 11.48
 ≥50 27 2.38 5 1.71 5 1.98 13 3.53 2 1.24 2 3.28
Sex 0.066
 Female 724 63.73 179 61.09 154 60.87 257 69.84 97 60.25 37 60.66
 Male 412 36.27 114 38.91 99 39.13 111 30.16 64 39.75 24 39.34
Functional status before surgery 0.225
 Independent 1105 97.27 289 98.63 248 98.02 356 96.74 154 95.65 58 95.08
 Part/total dependent 31 2.73 4 1.37 5 1.98 12 3.26 7 4.35 3 4.92
ASA 0.001
 1 17 1.50 8 2.73 4 1.58 3 0.82 1 0.62 1 1.64
 2 478 42.08 144 49.15 122 48.22 137 37.23 51 31.68 24 39.34
 3+ 641 56.43 141 48.12 127 50.20 228 61.96 109 67.70 36 59.02
Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD P value a
Length of stay (d) 2.6 2.1 2.4 1.1 2.6 1.4 2.6 1.1 3.0 4.7 3.1 1.5 0.030
aP < 0.05. Bolded p-values are statistically significant at p<0.05.

F1
Figure 1:
Histogram showing the distribution of NarxCare narcotic scores on admission for elective total knee arthroplasty (TKA) in the study sample.

The average age of patients in the study population was 65.2 ± 9.9 (mean ± SD) years. There was a statistically significant difference in age between patients in the different admission Narx Score groups (higher Narx Score groups associated with lower age, P < 0.001). The average BMI of patients in the study cohort was 33.4 ± 7.3 kg/m2, and there were statistically significant differences in BMI between patients in the different admission Narx Score groups (P < 0.001) by ANOVA. There were no statistically significant differences in sex or functional status between the admission Narx Score groups (Table 1).

In terms of ASA, there were differences between Narx Score groups by ANOVA (P = 0.001). The opioid-naive group had the lowest percentage of ASA 3+ patients (48.12%), and the 300 to 499 group had the highest percentage of ASA 3+ patients (67.70%).

Postoperative Outcomes

The mean total LOS of the study population was 2.6 ± 2.1 days. The mean length of stay increased with increasing admission Narx Score group (P = 0.030, Table 1).

The rate of AAEs in the total study population within 30 days of the index procedure was 5.63%, the rate of SAEs in the total study population was 3.35%, and the rate of MAEs in the total study population was 2.99%. These findings are summarized by the Narx Score group in Table 2. On univariate analysis, rates of AAE, SAE, and MAE increased with increasing admission Narx Score group (P = 0.001, P < 0.001, and P = 0.001, respectively).

Table 2 - Number of Adverse Events, Return to the Operating Room, Readmissions, and Mortality for Patients of Varying Narcotic Scores
Narcotic Score on Admission Total study population 0 1-99 100-299 300-499 500+
Total Number of Patients n = 1136 n = 293 n = 253 n = 368 n = 161 n = 61 a P value
Number Percent Number Percent Number Percent Number Percent Number Percent Number Percent
Any adverse event (AAE) 64 5.63 17 5.80 6 2.37 23 6.25 26 16.15 13 21.31 0.001
 Serious adverse event (SAE) 38 3.35 8 2.73 4 1.58 12 3.26 10 6.21 4 6.56 <0.001
  Deep infection 4 0.35 0 0.00 1 0.40 1 0.27 2 1.24 0 0.00
  Sepsis 7 0.62 2 0.68 0 0.00 0 0.00 4 2.48 1 1.64
  Failure to wean 1 0.09 0 0.00 0 0.00 0 0.00 1 0.62 0 0.00
  Reintubation 2 0.18 0 0.00 0 0.00 0 0.00 1 0.62 1 1.64
  Renal failure 1 0.09 0 0.00 0 0.00 0 0.00 1 0.62 0 0.00
  Thromboembolic events 23 2.02 5 1.71 3 1.19 9 2.45 4 2.48 2 3.28
  Cardiac arrest 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00
  MI 4 0.35 1 0.34 0 0.00 0 0.00 3 1.86 0 0.00
  Stroke 3 0.26 1 0.34 0 0.00 2 0.54 0 0.00 0 0.00
 Minor adverse event (MAE) 34 2.99 10 3.41 3 1.19 6 1.63 9 5.59 6 9.84 0.001
  Superficial infection 12 1.06 3 1.02 2 0.79 4 1.09 0 0.00 3 4.92
  Dehiscence 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00
  Pneumonia 8 0.70 1 0.34 1 0.40 1 0.27 4 2.48 1 1.64
  UTI 13 1.14 6 2.05 0 0.00 1 0.27 5 3.11 1 1.64
  Renal insufficiency 4 0.35 1 0.34 0 0.00 0 0.00 2 1.24 1 1.64
Return to the operating room within 30 d of operation 15 1.32 5 1.71 3 1.19 3 0.82 4 2.48 0 0.00 0.466
Readmission within 30 d of operation 70 6.16 12 4.10 10 3.95 24 6.52 16 9.94 8 13.11 0.009
 Mortality within 30 d of operation 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 1.000
aP < 0.05. MI = myocardial infarction, UTI = urinary tract infection.

On multivariable logistic regression assessing AEs, the highest Narx Score group had increased odds of AAE compared with the opioid-naive group (odds ratio [OR] = 2.96; 95% confidence interval [CI] = 1.18 to 7.40; P = 0.021). There were no other statistically significant differences in aggregated AEs between the different Narx Score groups (P > 0.05 for all). Table 3 and Figure 2 show the findings of the multivariable logistic regression analyses. Multivariable regressions that treated the admission Narx Score as a continuous variable did not find admission Narx Scores to be associated with postoperative complications.

Table 3 - Multivariate Odds Ratio for Adverse Events, Return to the Operating Room, Readmissions, and Mortality for Patients of Varying Narcotic Scores
Narcotic Score on Admission 0 1-99 100-299 300-499 500+
Total Number of Patients (n = 1136) n = 293 n = 253 n = 368 n = 161 n = 61
Odds Ratio a P value Odds Ratio a P value Odds Ratio a P value Odds Ratio a P value Odds Ratio a P value
Any adverse event (AAE) 1 1.000 0.39 0.051 0.80 0.543 1.39 0.399 2.96 0.021
 Serious adverse event (SAE) 1 1.000 0.57 0.371 1.21 0.689 2.24 0.103 2.75 0.122
 Minor adverse event (MAE) 1 1.000 0.31 0.084 0.41 0.103 1.41 0.487 3.06 0.056
 Revision surgery within 30 d of operation 1 1.000 0.72 0.656 0.50 0.355 1.39 0.634 N/A N/A
 Readmission within 30 d of operation 1 1.000 1.04 0.931 1.76 0.139 2.46 0.029 3.98 0.007
aP < 0.05.
Controlled for age, sex, BMI, functional status, and ASA classification.

F2
Figure 2:
Forest plot illustrating odds ratios for readmission within 30 days of the index procedure, using the NarxCare narcotic score 0 group as the referent.

Return to the operating room within 30 days of surgery was noted for 1.32% of the study population, and none of the patients died. There were no differences in the odds of revision surgery on multivariable regression by admission Narx Score group. There were insufficient events for the analysis of mortality on multivariable regression.

Readmission to the hospital within 30 days of surgery was noted for 6.16% of the study population. The rates of readmission increased with increasing admission Narx Score group (P = 0.009). On multivariable regression, patients in the following admission Narx Score groups had higher odds of readmission: 300 to 499 (OR = 2.46; 95% CI, 1.10 to 5.51; P = 0.029) and 500+ (OR = 3.98; 95% CI, 1.45 to 10.91; P = 0.007).

Discussion

Previous studies have noted that preoperative opioid use is associated with adverse perioperative outcomes after TKA, including in-hospital complications and greater risk for early revision.6,7 The current study probes whether preoperative opioid use (as measured by the NarxCare narcotic score) is associated with postoperative outcomes after TKA.

The current study found that younger patients tended to have higher preoperative NarxCare narcotic scores. This is consistent with the findings of Goesling et al.14 The current study also found that preoperative Narx Scores were higher for those with higher ASA classification. This is consistent with a previous study by Bedard et al, which found that several comorbidities (ie, anxiety/depression, low back pain, myalgia, and drug/alcohol/smoking dependence) were associated with an increased risk of opioid use.3 Taken together, these findings indicate that the groups with higher Narx Scores are younger and less healthy.

Univariable analyses found that the higher Narx Score groups had increased rates of any, serious, and minor AEs. The univariate ANOVA findings are consistent with those of Kim et al., which found that patients with consistent preoperative opioid use were at higher risk of several types of perioperative AEs compared with opioid-naive patients.9 However, once Kim et al adjusted their results for differences in demographics and preoperative comorbidity burden, they no longer observed differential risk of 30-day mortality. These results indicate that the association between preoperative opioid use and AEs may be confounded by a higher rate of preoperative comorbidities in narcotic users. These results suggest that further investigation of preoperative patient characteristics may be warranted to improve perioperative outcomes.

Both univariate and multivariable analyses found that the two highest narcotic score groups had a significantly higher risk of readmission (ORs of 2.46 and 3.98) within 30 days of TKA compared with the other groups. These findings are consistent with Weick et al., who found that preoperative opioid use increased 30-day readmission rates after TKA.15 The results of our study have implications for how narcotic scores can be used to assist and inform healthcare policy, as readmission rate is a critical factor in hospital reimbursement. Readmission rate became linked to hospital reimbursement in 2012 when the Affordable Care Act was passed with the Hospital Readmissions Reduction Program (HRRP).16 The HRRP financially penalizes hospitals for excessive readmissions for certain medical conditions and surgeries, specifically including TKA and total hip arthroplasty [B].17 The HRRP was aimed to incentivize hospitals to reduce costs and improve patient care, as 9% to 48% of readmissions are thought to be preventable.18–20 In response to these changes, hospitals have strived to reduce the readmission rate through several means.17 As they embark on these initiatives, hospitals can readily benefit from tools that can meaningfully identify patients' risk levels for readmission and personalize care. Our study indicates that the Narx Score has the potential to become one such tool and thus merits further exploration. Continued evaluation of this new scoring tool is warranted as it becomes increasingly adopted nationally.

The current study has several limitations. First, the NarxCare narcotic scores used to approximate opioid burden have not been previously validated and may not be available at all hospitals. In addition, these scores are proprietary, and the exact scoring algorithm is not publicly known. Second, Physician Drug Monitoring Program data can only track how often a prescription was filled but cannot track how much medication a patient took. Third, the present study was performed at a single institution, potentially limiting its generalizability.

The present study sheds light on the relationship between Narx Scores and both perioperative and postoperative outcomes. Patients with higher Narx Scores had increased odds of readmission, prolonged LOS, and aggregated AEs. In addition, this study identified key differences in demographics and comorbidities between Narx Score groups. These demographic and comorbid variations may account for the disparity in perioperative outcomes as initially identified by the univariate analysis. As a result, this may warrant further examination of preoperative patient characteristics to optimize patient outcomes through the identification of at-risk populations1 and personalizing treatment for these patients.

References

1. Blendon RJ, Benson JM: The public and the opioid-abuse epidemic. N Engl J Med 2018;378:407-411.
2. Lespasio MJ, Guarino AJ, Sodhi N, Mont MA: Pain management associated with total joint arthroplasty: a primer. Perm J 2019;23:18-169.
3. Bedard NA, Pugely AJ, Westermann RW, Duchman KR, Glass NA, Callaghan JJ: Opioid use after total knee arthroplasty: Trends and risk factors for prolonged use. J Arthroplasty 2017;32:2390-2394.
4. Zarling BJ, Yokhana SS, Herzog DT, Markel DC: Preoperative and postoperative opiate use by the arthroplasty patient. J Arthroplasty 2016;31:2081-2084.
5. Singh GK, Kim IE Jr, Girmay M, et al: Opioid epidemic in the United States: empirical trends, and a literature review of social determinants and epidemiological, pain management, and treatment patterns. Int J MCH AIDS 2019;8:89-100.
6. Chen EY, Lasky R, Dotterweich WA, Niu R, Tybor DJ, Smith EL: Chronic prescription opioid use before and after total hip and knee arthroplasty in patients younger than 65 years. J Arthroplasty 2019;34:2319-2323.
7. Rozell JC, Courtney PM, Dattilo JR, Wu CH, Lee GC: Preoperative opiate use independently predicts narcotic consumption and complications after total joint arthroplasty: J Arthroplasty 2017;32:2658-2662.
8. Bedard NA, DeMik DE, Dowdle SB, Owens JM, Liu SS, Callaghan JJ: Preoperative opioid use and its association with early revision of total knee arthroplasty. J Arthroplasty 2018;33:3520-3523.
9. Kim SC, Jin Y, Lee YC, et al: Association of preoperative opioid use with mortality and short-term safety outcomes after total knee replacement. JAMA Netw Open 2019;2:e198061.
10. Crawford DA, Andrews RL, Hurst JM, Berend KR, Lombardi AV, Morris MJ: Accuracy of patient-reported preoperative opioid use and impact on continued opioid use after outpatient arthroplasty. J Arthroplasty 2020;35:1504-1507.
11. Patel AA, Singh K, Nunley RM, Minhas SV: Administrative databases in orthopaedic research: pearls and pitfalls of big data. J Am Acad Orthop Surg 2016;24:172-179.
12. George J, Newman JM, Ramanathan D, Klika AK, Higuera CA, Barsoum WK: Administrative databases can yield false conclusions-an example of obesity in total joint arthroplasty. J Arthroplasty 2017;32:S86-S90.
13. Galivanche AR, Mercier MR, Adrados M, et al: Admission NarxCare Narcotics Scores are not associated with adverse surgical outcomes or self-reported patient satisfaction following elective spine surgery. Spine (Phila Pa 1976) 2019;44:1515-1522.
14. Goesling J, Moser SE, Zaidi B, et al: Trends and predictors of opioid use after total knee and total hip arthroplasty. Pain 2016;157:1259-1265.
15. Weick J, Bawa H, Dirschl DR, Luu HH: Preoperative opioid use is associated with higher readmission and revision rates in total knee and total hip arthroplasty. J Bone Joint Surg Am 2018;100:1171-1176.
16. Kelly EA, et al: Readmissions are not what they seem: incidence and classification of 30-day readmissions after orthopedic trauma surgery. J Orthop Trauma 2020:34e72-34e76.
17. Ramaswamy A, Marchese M, Cole AP, et al: Comparison of hospital readmission after total hip and total knee arthroplasty vs spinal surgery after implementation of the hospital readmissions reduction program. JAMA Netw Open, 2019;2:2e194634.
18. Weinberg DS, Kraay MJ, Fitzgerald SJ, Sidagam V, Wera GD. Are readmissions after THA preventable?. Clin Orthop Relat Res 2017;475:1414-1423.
19. Benbassat J, Taragin M: Hospital readmissions as a measure of quality of health care: advantages and limitations. Arch Intern Med 2000;160:1074-1081.
20. Vachon C.M., Aaland M., Zhu TH: Readmission of trauma patients in a nonacademic level II trauma center. J Trauma Acute Care Surg 2012;72:531-536.
Copyright © 2022 The Authors. Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Orthopaedic Surgeons.