- Email: [email protected]
The false vital sign: When pain levels are not predictive of discharge opioid prescriptions
International Journal of Medical Informatics 129 (2019) 69–74
Contents lists available at ScienceDirect
International Journal of Medical Informatics journal homepage: www.elsevier.com/locate/ijmedinf
The false vital sign: When pain levels are not predictive of discharge opioid prescriptions
T
Jennifer A. Villwock , Mark R. Villwock, Jacob New, Gregory Ator ⁎
University of Kansas Medical Center, Departments of Otolaryngology-Head and Neck Surgery and Clinical Informatics, 3901 Rainbow Blvd, Mailstop 3010, Kansas City, KS 66160, United States
ARTICLE INFO
ABSTRACT
Keywords: Opioid Prescribing patterns Opioid prescriptions Pain scale Opioid prescriptions Opioid prescribing patters Narcotic prescriptions Emergency department Inpatient pain Post-operative pain
Background: Pain gained recognition as a vital sign in the early 2000s, underscoring the importance of accurate documentation, characterization, and treatment of pain. No prior studies have demonstrated the utility of the 0–10 pain scale with respect to discharge opioid prescriptions, nor characterized the most influential factors in discharge prescriptions. Methods: Inpatient and emergency department(ED) encounters from July 1, 2012 to April 1, 2018 resulting in a discharge prescription for tablet opioid medications were identified. The primary outcome was to determine if pain levels in 24 h prior to discharge correlated with opioids (in milligrams of morphine equivalents (MME)) prescribed. Secondary outcomes included the impact of patient and prescriber demographics, demographics. A generalized linear model was created to investigate factors affecting the quantity of prescribed opioids. Results: n = 78,691 patient encounters. Overall mean adjusted MME for non-ED visits was 378 versus 197 for ED visits. Whites received the highest quantities; those identifying as non-white and non-black received the lowest. Women received significantly fewer discharge MMEs in both the ED and inpatient cohorts. Provider prescribing patterns exhibited the most profound effect on discharge MMEs. The most prolific (≥300 prescriptions over the study period) writing the largest amount. In the ED, there was a significant negative correlation between documented pain levels and discharge MMEs(ρ = 0.074,p < 0.001). Conclusions: Pain scale was significantly negatively correlated with discharge MMEs in the ED and positively correlated in the inpatient population. Individual prescriber characteristics were the more influential variable, with prolific high prescribers writing for the largest MME amounts. The inverse association of pain and MMEs at discharge in the ED, and the large effect pre-existing prescriber patterns exhibited, both improved methodology for assessing and appropriately treating pain, and effective prescriber-targeted interventions, must be a priority.
1. Introduction In the early 2000s, the idea of pain as the “fifth vital sign” was introduced. The impetus for recording pain as a vital sign was to ensure routine assessment and avoid the undertreatment of pain. However, this increased focus on pain, in combination with other societal factors, likely contributed to the boom in opioid prescriptions, dependency, addiction, and overdose that comprises today’s “Opioid Epidemic.” As a result, in 2016, Physicians for Responsible Opioid Prescribing (PROP) have asked for pain-treatment questions to be removed from satisfaction surveys that may be linked to hospital ratings and individual provider quality metrics [1,2]. Examples of these questions include “During this hospital stay, how often did the hospital staff do everything they could to help you with your pain?” (bold emphasis added
⁎
by the authors). Depending on the patient and their reason for admission, complete elimination of pain – as implied by this question – may not be possible or desired. Seventy-five percent of all opioid misuse begins with people using medication that was not prescribed for them [3–5]. Given that, in 2012, 289 million prescriptions for opioids were dispensed in the United States [6], proper opioid stewardship, such as prescribing medications in proportion to patient pain levels and minimizing overprescribing, is of paramount importance. It is equally important to note that failure to utilize opiates in appropriate circumstances causes suffering and creates a barrier to effective pain care [7]. Despite the clear need to balance these competing priorities with meaningful and clinically relevant pain assessments, there have been no published studies validating the utility of the most commonly used 0–10 pain scale or delineating the factors
Corresponding author. E-mail address: [email protected] (J.A. Villwock).
https://doi.org/10.1016/j.ijmedinf.2019.05.005 Received 7 September 2018; Received in revised form 18 April 2019; Accepted 8 May 2019 1386-5056/ © 2019 Elsevier B.V. All rights reserved.
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
that most highly impact prescribing patterns. This study sought to address both these knowledge gaps by retrospectively analyzing opioid discharge prescriptions during a five-year period at an academic, tertiary care hospital. The primary objective was to determine correlations between the 0–10 pain scale and prescribing patterns. Secondary objectives included characterizing the impact of patient and provider factors on prescriptions. This included subset analysis comparing prescribing patterns for inpatient patients versus those who were treated, and ultimately discharged, in the emergency department setting.
Table 1 Patient demographics.
2. Methods The methodology for this quality improvement study was reviewed by the University of Kansas Institutional Review Board and deemed to be non-human subject research. 2.1. Patient cohort All patients who were (1) admitted for > 24 h or (2) treated in our ED between July 1, 2012 and April 1, 2018 were identified in our electronic medical record (EMR). These data were queried to identify all discharges containing a discharge prescription for hydrocodoneacetominophen, oxycodone, or oxycodone-acetominophen were identified by a trained informatics specialist. These medications were selected as they were the three most common opioid in tablet form prescribed during the study time period. Eligible encounters were divided into groups of patients who had undergone procedures – as determined by a recorded trip to the operating room/procedural suite and an associated operative/procedure note in the EMR – and those who had not. The following additional data points were also collected: quantity of discharge opioids; ordering and authorizing provider; number of prescriptions per trainee (resident or fellow), attending physician, or advanced practice practitioner (APP); patient length of stay; and pain level in the 24 h preceding discharge. Morphine milligram equivalents (MME) were determined by the following drug-specific multipliers: hydrocodone, 1; oxycodone, 1.5 to control for the strength of the medication prescribed [8]. The Area Deprivation Index (ADI), a composite score generated from seventeen socioeconomic variables, was derived from patient primary residence zip code and used as a proxy for socioeconomic status [9].
Non-Emergency Department Visits
Emergency Department Visits
Age, y [median (IQR)]
41 (56–66)
40 (29–54)
Insurance Commercial Medicare Medicaid Other No insurance
20243 (37.0%) 17072 (31.2%) 4950 (9.1%) 1561 (2.9%) 10812 (19.8%)
7390 (30.7%) 3468 (14.4%) 3205 (13.3%) 790 (3.3%) 9200 (38.2%)
Race White Black Other
41413 (75.8%) 7284 (13.3%) 5711 (10.5%)
13102 (54.5%) 6858 (28.5%) 4028 (16.7%)
Gender Male Female
25726 (47.1%) 28912 (52.9%)
11254 (46.8%) 12799 (53.2%)
Area Deprivation Index < 20 20–39 40–59 60–79 > = 80
2056 (3.8%) 8612 (15.8%) 15777 (28.9%) 18497 (33.9%) 7940 (14.5%)
571 (2.4%) 2552 (10.6%) 5492 (22.8%) 7336 (30.5%) 7333 (30.5%)
Patient distance from hospital, miles < 10 16979 (31.1%) 10–29 17352 (31.8%) 30–59 7490 (13.7%) > = 60 12792 (23.4%)
16102 (66.9%) 5859 (23.5%) 990 (4.1%) 1292 (5.4%)
Pain (last 24 h) 0–3 4–6 7–10
885 (3.7%) 4397 (18.3%) 15705 (65.3%)
17164 (31.4%) 25765 (47.2%) 5749 (10.5%)
3. Results Between July 1, 2012 and April 1, 2018, 78,691 encounters were analyzed at a single, tertiary care, academic medical center. 3.1. Non-emergency department visits Non-emergency department related visits constituted 69% of opioid encounters (54,638 cases). Patient and provider demographics are shown in Tables 1 and 2. There was a statistically significant increase in admission (τ = .672, p < .001) and operative procedures (τ = .787, p < .001) and decrease in overall length of stay (τ = −0.640, p < .001) during the study period. The average patient length of stay significantly decreased from 137 h in July 2012 to 111 in April 2018. Operating room procedures increased from 29% to 78% during the
2.2. Statistical analyses SPSS version 25 (Armonk, NY) was used for the data analyses. Group comparisons were performed using Chi-squared, Kruskal-Wallis, and Mann-Whitney U tests, as appropriate. Temporal trends in the monthly number of admissions, number of operative procedures, and length-of-stay were assessed via Mann-Kendall tests. Spearman’s rho was used to analyze correlation between pain-level and quantity of prescribed opioids. To further investigate factors affecting the quantity of prescribed opioids, a generalized linear model using Gamma as the distribution and Log as the link function was created. Separate models were constructed for non-emergency department visits and emergency department visits. The analysis considered age, gender, insurance status, prescriber type (attending/trainee/APP), ordering prescriber frequency (stratified based on the number of scripts written over the study period), specialty (medicine/surgical) [non-emergency model only], race, pain in the 24 h preceding discharge, patient distance from hospital, ADI, and whether the patient had an operating room procedure [non-emergency model only]. Estimated marginal means and 95% wald confidence intervals for each factor were reported. When appropriate, pairwise comparisons are reported using least significant difference adjustment for multiple comparisons. Statistical significance was set at p < 0.01.
Table 2 Provider demographics.
Provider Type Attending Trainee APP Specialty Medicine Surgical
Non-Emergency Department Visits
Emergency Department Visits
21111 (38.6%) 21765 (39.8%) 11762 (21.5%)
10288 (42.8%) 3956 (16.4%) 9809 (40.8%)
24615 (45.1%) 30023 (54.9%)
a
Ordering provider frequency (#scripts) < 100 15772 (28.9%) 100–199 15682 (28.7%) 200–299 12785 (23.4%) > = 300 10399 (19.0%) a
70
Not applicable.
a
4156 (17.3%) 2488 (10.3%) 5054 (21.0%) 12355 (51.4%)
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
same time frame. The overall mean adjusted MME on discharge for a non-emergency department visit was 462 (95% CI: 455–469). Patients identifying as Caucasian/white comprised 75.8% of the study population. Whites received the highest quantities of discharge MMEs (adjusted mean 497, 95% CI: 490–504), while those identifying as non-black and non-white received the lowest mean MMEs (adjusted mean 419, 95% CI: 409–429). Males received 10% more discharge MMEs (adjusted means, Table 1) than females (p < .001). 89% of encounters documented pain levels in the 24 h preceding admission. There was a statistically significantly positive correlation (ρ = 0.051, p < 0.001) between documented pain levels and number of milligrams of opioids prescribed at discharge. Those with a pain of 7 or more received on average 37% more (adjusted mean) opioids than patients reporting a pain of 3 or less (Table 1). Socioeconomic status, using zip code and ADI as a proxy, impacted MMEs prescribed, with the most disadvantaged group (≥80) receiving the smallest amount (Table 1). Yet, a pairwise comparison with mean MME from the least disadvantaged group (< 20) was statistically insignificant (p = .260). Patients living 30–60 miles from the hospital received the highest discharge MMEs (p < 0.001) when compared to other groups. Provider prescribing patterns exhibited the most profound effect on discharge opioid amounts. The most prolific script writers (≥300 scripts over the study period) tended to write the largest amount of opioids (34% more than providers that wrote < 100 prescriptions – Table 2). Trainees (residents and fellows) wrote significantly higher amounts of discharge opioids than their attending and APP counterparts (Table 2, p < .001).
Table 3 Adjusted means of discharge opioids in milligrams of morphine equivalents (MME) for non-emergency department and emergency department-related visits. Unadjusted means along with adjusted means from multivariate analysis are presented. Non-Emergency Department Visits Mean
95% CI Lower Upper
Mean
95% CI Lower
Upper
451 443 471 504 443
441 433 460 494 435
462 453 482 515 452
137 135 141 148 156
133 131 136 143 151
142 140 145 153 162
445 479
438 470
453 489
a
a
a
a
a
a
465 452 501 439 455
456 441 481 432 446
475 463 522 447 464
135 136 148 142 156
131 131 141 138 152
140 141 156 146 161
M F
484 441
476 434
492 449
151 136
147 132
155 140
Black Other White
474 419 497
463 409 490
485 429 504
139 142 148
135 138 145
144 147 152
463 461
455 452
471 470
a
a
a
a
a
a
479 493 418
470 484 408
488 502 428
159 163 114
154 158 110
163 168 118
Patient area deprivation index < 20 20–39 40–59 60–79 > = 80
454 471 467 475 444
437 460 459 467 435
471 481 476 483 453
136 150 144 146 140
128 145 140 142 137
145 155 149 150 144
Patient pain as reported in 24 hrs preceding discharge < =3 4-6 > =7
397 458 543
390 451 530
404 466 555
164 144 124
156 140 121
172 148 127
Patient distance from hospital, miles < 10 10–29 30–59 > = 60
448 459 481 460
441 451 470 450
456 467 492 471
124 130 163 161
121 126 156 153
127 133 171 170
Ordering provider frequency (#scripts) < 100 100–199 200–299 > = 300 Overall
432 407 449 578 462
423 399 440 565 455
441 415 458 590 469
137 152 147 138 143
132 147 142 134 140
141 158 151 142 147
Age, y
< 30 30–39 40–49 50–59 > = 60
Operative case Yes No Insurance Medicare Medicaid Other Commercial No insurance Gender
Race
Provider specialty Medicine Surgical
3.2. Emergency department visits Emergency department visits constituted 31% of opioid encounters (24,053 cases). Patient and provider demographics are shown in Tables 1 and 2. The number of admissions increased for the first two years of the study period but the trend reversed during the subsequent years, resulting in a lack of a monotonic trend (τ = −0.042, p = .605). The median length of stay was 5.7 h (95% CI: 5.1–6.3) and there was a statistically significant decrease in overall length of stay over the study period (τ = −0.222, p = .006). The overall mean adjusted MME on discharge for an emergency department visit was 143 (95% CI: 140–147). Patients identifying as caucasian/white comprised 54.5% of the ED population, a proportion significantly lower than represented within the inpatient cohort (p < .001). Whites received significantly more MMEs (adjusted mean 148, 95% CI: 145–152) than blacks (adjusted mean 139, 95% CI: 135–144) (p < .001). Males received 11% more discharge MMEs (Table 3) than females (p < .001). Patients residing > 30 miles from the hospital received a significantly higher mean MMEs on discharge (p < 0.001). 87% of encounters documented pain levels in the 24 h preceding admission. There was a statistically significantly negative correlation (ρ = −0.074, p < 0.001) between documented pain levels and number of milligrams of opioids prescribed at discharge. Those with a pain of 7 or more received on average 24% less (adjusted mean) opioids than patients reporting a pain of 3 or less (Table 3). Unlike providers treating non-emergency visits, those within the ED with a high prescription frequency (≥300 scripts over the study period) did not prescribe more MMEs than providers that wrote less than 100 scripts (Table 3, p = .512). APPs wrote significantly lower amounts of discharge opioids than their attending and trainee counterparts (Table 3, p < .001).
Provider type Attending Trainee Advanced Practice
a
71
Emergency Department Visits
Not applicable.
J.A. Villwock, et al.
International Journal of Medical Informatics 129 (2019) 69–74
4. Discussion
proportions of prescriptions. APPs represented the remaining fifth of the prescriptions. Conversely, in the ED, attending and APPs represented the bulk of prescriptions (˜40% each) with trainees responsible for only 16% of opioid prescriptions. This indicates that educational interventions may need to be specifically tailored to different prescribing groups, depending on the location. Another important finding of this study is the impact of provider prescribing patterns on opioid prescriptions. Database studies of prescription drugs have similarly shown that such high-risk prescribers – in the top 5th percentile among all prescribers – can account for a large proportion of total opioids dispensed. For example, in Florida, such high-risk providers account for only 4% of prescribers, but are responsible for 67% of the total opioid volume and 40% of total opioid prescriptions. This study also found that these prescribers are particularly responsive to prescription drug monitoring policies [16]. Given that we found that individual prescribing patterns had a highly significant impact on discharge MMEs for inpatients, this indicates interventions targeted at high prescribers may be of particular benefit.
4.1. Summary Patient pain levels, most commonly on a 0–10 scale, have been routinely documented for nearly two decades. Despite widespread use, the utility of pain levels – and their impact on clinical practice – had not previously been well-studied. Patient and provider factors that impact prescribing patterns have also not been completely elucidated. Given the current opioid epidemic – and the fact that most improperly used opioids are from “leftover” or excess prescriptions [3–5] – ensuring pain medication prescriptions are appropriate for the pain being treated is of paramount importance. This study of 78,691 inpatient encounters from July 1, 2012 to April 1, 2018 demonstrated a statistically significantly poor, and inverse, correlation between documented pain levels and opioid prescription quantity in the emergency department. Patient factors predictive of higher opioid prescriptions in the inpatient setting included white race – self-identification as non-black and non-white resulted in significantly smaller prescriptions – and male gender. Interestingly, provider characteristics exerted the most influence on prescribing patterns.
4.4. Study limitations This study is not without limitations. This is a retrospective single institution study at an academic medical center. As such, results may not be externally valid. However, given the large sample size, large patient catchment area, and diverse clinical faculty, we do not believe significant bias was introduced. We were unable to determine what proportion of the prescribed medication in the investigated encounters were filled. Additionally, our results are limited to the data available within our retrospective review. Due to the retrospective nature of the study, we also do not know the clinical decision making that occurred to justify the MMEs prescribed. Furthermore, unless admitting care teams select diagnoses related to chronic pain, we are unable to know what proportion of our study population suffer from chronic pain. This is particularly important since episodic care settings, such as inpatient admissions, should refrain from supplying opioids to chronic pain patients whenever possible [17].
4.2. Impact of patient characteristics Given the wealth of searchable information electronic health records afford, it behooves us to review and ensure that high-quality, data-driven care is being provided [10]. We found that, while documented pain levels correlated with discharge prescription amounts for the inpatient cohort, this was not true of patients who received their treatment in the ED. In the latter group, there was a statistically significantly negative correlation between pain level with discharge opioids. In fact, those with higher levels of pain (> 7) received an average of 24% less opioids than patients with pain levels of 3 or less. This data may underscore the frustrations of many clinicians who feel that current pain assessments are devoid of meaning because they are not linked to function or other outcomes of interest. Patient demographics did impact prescribing patterns, especially those of race and gender. For example, our results show that blacks received significantly less opioids than whites in both the inpatient and ED cohorts. This is consistent with the existing literature on racial disparities in pain control. For example, it has previously been shown that blacks are at elevated risk for undertreatment and underestimation of pain by their health care team [11,12]. However, interestingly, these differences have been shown to be significantly impacted by the specialty of the prescribing provider [6]. For example, in academic primary care, it has been found that Hispanic, Asian, and older patients are less likely to be prescribed higher opiate doses [13]. We similarly found that patients identifying as non-black and non-white received the lowest mean MMEs. In our study, women also received significantly less MMEs than men in both the inpatient and ED cohorts. This is in agreement with prior literature describing that male physicians are more likely to prescribe higher opioid amounts to male patients [14]. Similarly, women with metastatic cancer have previously been found to be significantly more likely than men (OR 5) to be undertreated for pain [15]. Distance from the hospital also impacted opioid prescribing patterns. In the ED setting, patients living > 30 miles away from the hospital received significantly more MMEs than those living closer. Among inpatients, those living 30–60 miles from the hospital received the highest MMEs on discharge. This is in agreement with a study conducted by Bauer et al reporting that patients with nonlocal home addresses were more likely to be prescribed higher opiate doses [13].
4.5. Future directions We identified multiple factors that impacted the quantity of opioids prescribed upon discharge from acute hospital care. The impact of prescriber factors highlight the need for targeted research and interventions on educational interventions targeted at these groups. Factors leading to the differences in patient care settings (inpatient versus emergency department care) and the fact that higher pain levels are associated with smaller opioids quantities upon discharge from the emergency department should also be elucidated. For patients living > 30 miles from the hospital, research targeting ways to ensure appropriate outpatient follow-up and management of pain is needed to avoid reflex prescribing of larger quantities. 5. Conclusions There were multiple factors associated with differential opioid prescribing patterns. Prolific prescription writers were responsible for the largest adjusted MMEs on discharge for inpatients, indicating that targeted interventions for these high-risk prescribers may be of particular benefit. In the ED, the 0–10 pain scale poorly correlates with discharge opiate prescriptions. Improved methods for assessing pain should also be a priority. Author statement
4.3. Impact of prescriber characteristics
All authors have made substantial contributions to all of the following: (1) the conception and design of the study, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or
With respect to provider characteristics, in the inpatient setting, trainees and attendings were responsible for approximately equal 72
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
revising it critically for important intellectual content, (3) final approval of the version to be submitted.
Acknowledgement The authors would like to thank Jennifer B. Wilson RN, BSN for her instrumental support in the data acquisition necessary to complete this work.
Funding None.
Appendix 1 Unadjusted means of discharge MMEs for non-emergency department and emergency department visits
Age, y
Operative case Insurance
Gender Race
Provider specialty Provider type
Non-Emergency Department Visits 95% CI Mean Lower
Upper
Emergency Department Visits 95% CI Mean Lower
Upper
< 30 30-39 40-49 50-59 > =60
434 430 462 504 446
420 416 446 489 439
448 443 479 519 454
113 109 115 122 133
110 105 112 118 128
116 112 118 126 138
Yes No
452 468
447 455
457 481
* *
* *
* *
Medicare Medicaid Other Commercial No insurance
469 456 542 445 450
459 436 514 438 438
480 476 569 453 462
121 101 126 116 121
116 98 118 113 118
125 104 135 120 123
M F
480 438
473 431
488 445
127 108
124 107
129 110
Black Other White
451 381 469
435 367 463
468 395 475
106 112 124
103 108 122
108 115 127
Medicine Surgical
461 454
451 449
472 460
* *
* *
* *
450 468 450
441 459 441
459 478 459
134 139 89
131 133 88
136 145 90
428 455 461 472 427
406 441 452 462 414
449 469 470 482 440
116 128 121 120 107
108 123 118 116 104
123 133 125 123 109
406 470 556
400 462 533
412 477 579
169 137 108
154 132 107
184 142 110
439 457 487 466
429 447 474 455
449 466 501 478
109 123 163 165
108 119 150 147
111 127 176 183
454 409 449 543 445
439 402 441 532 440
469 416 458 553 450
134 132 131 102 116
129 128 126 100 114
139 137 136 103 117
Attending Trainee Advanced Practice Patient area deprivation index < 20 20–39 40–59 60–79 > = 80 Patient pain as reported in 24 hrs preceding discharge < =3 4–6 > =7 Patient distance from hospital, miles < 10 10–29 30–59 > = 60 Ordering provider frequency (#scripts) < 100 100–199 200–299 > = 300 Overall
* Not applicable.
within 180 days, after allowing a reasonable period for public comment. If you have any questions about our requested changes, please contact the office of Dr. Andrew Kolodny at (347) 396-0371.,” p. 4. [3] “Results from the 2014 National Survey on Drug Use and Health: Detailed Tables, SAMHSA, CBHSQ.” [Online]. Available: https://www.samhsa.gov/data/sites/ default/files/NSDUH-DetTabs2014/NSDUH-DetTabs2014.htm#tab6-47b. [Accessed: 27-May-2018]. [4] M. Szalavitz and M. Szalavitz, “Opioid Addiction Is a Huge Problem, but Pain Prescriptions Are Not the Cause,” Scientific American Blog Network. [Online]. Available: https://blogs.scientificamerican.com/mind-guest-blog/opioid-addiction-
References [1] “Johnson, Senators Introduce Bipartisan PROP Act - Press Releases - Ron Johnson.” [Online]. Available: https://www.ronjohnson.senate.gov/public/index.cfm/pressreleases?ID=CC5F9714-9A1A-4B24-AE50-A86D5AAB7F70. [Accessed: 27-May2018]. [2] H. Chen, R. J. Goldsmith, and K. M. Murphy, “For the reasons discussed in this petition, the undersigned Petitioners request that CMS issue a proposal for removing the pain questions listed above within 90 days, and finalize the rule change
73
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
[5] [6] [7] [8] [9] [10] [11]
is-a-huge-problem-but-pain-prescriptions-are-not-the-cause/. [Accessed: 27-May2018]. N.D. Volkow, A.T. McLellan, Opioid abuse in chronic pain — misconceptions and mitigation strategies, N. Engl. J. Med. 374 (March (13)) (2016) 1253–1263. B. Levy, L. Paulozzi, K.A. Mack, C.M. Jones, Trends in opioid analgesic-prescribing rates by specialty, US., 2007-2012, Am. J. Prev. Med. 49 (September (3)) (2015) 409–413. K.K. Dineen, J.M. DuBois, Between a rock and a hard place: can physicians prescribe opioids to treat pain adequately while avoiding legal sanction? Am. J. Law Med. 42 (no. 1) (2016) 7–52. D. Dowell, T.M. Haegerich, R. Chou, CDC guideline for prescribing opioids for chronic pain–United States, JAMA 315 (15) (2016) 1624–1645. Apr. 2016. A.J.H. Kind, W.R. Buckingham, Making neighborhood-disadvantage metrics accessible — the neighborhood atlas, N. Engl. J. Med. 378 (June (26)) (2018) 2456–2458. D. Backenroth, H.S. Chase, Y. Wei, C. Friedman, Monitoring prescribing patterns using regression and electronic health records, BMC Med. Inform. Decis. Mak. 17 (December) (2017). Institute of Medicine, US) committee on understanding and eliminating racial and
[12] [13] [14] [15] [16] [17]
74
ethnic disparities in health care, Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care, National Academies Press (US), Washington (DC), 2003. A. Cintron, R.S. Morrison, Pain and ethnicity in the United States: a systematic review, J. Palliat. Med. 9 (December (6)) (2006) 1454–1473. S.R. Bauer, L. Hitchner, H. Harrison, J. Gerstenberger, S. Steiger, Predictors of higher-risk chronic opioid prescriptions in an academic primary care setting, Subst. Abuse 37 (January (1)) (2016) 110–117. C.S. Weisse, P.C. Sorum, K.N. Sanders, B.L. Syat, Do gender and race affect decisions about pain management? J. Gen. Intern. Med. 16 (April (4)) (2001) 211–217. D.E. Hoffmann, A.J. Tarzian, The Girl Who Cried Pain: a Bias Against Women in the Treatment of Pain, Social Science Research Network, Rochester, NY, 2001 SSRN Scholarly Paper ID 383803. H.-Y. Chang, et al., Impact of prescription drug monitoring programs and pill mill laws on high-risk opioid prescribers: a comparative interrupted time series analysis, Drug Alcohol Depend. 165 (01) (2016) 1–8. T.L. Skaer, A.C. Nwude, Opioid prescribing laws and emergency department guidelines for chronic non-cancer pain in Washington State, Pain Pract. Off. J. World Inst. Pain 16 (5) (2016) 642–647.
Contents lists available at ScienceDirect
International Journal of Medical Informatics journal homepage: www.elsevier.com/locate/ijmedinf
The false vital sign: When pain levels are not predictive of discharge opioid prescriptions
T
Jennifer A. Villwock , Mark R. Villwock, Jacob New, Gregory Ator ⁎
University of Kansas Medical Center, Departments of Otolaryngology-Head and Neck Surgery and Clinical Informatics, 3901 Rainbow Blvd, Mailstop 3010, Kansas City, KS 66160, United States
ARTICLE INFO
ABSTRACT
Keywords: Opioid Prescribing patterns Opioid prescriptions Pain scale Opioid prescriptions Opioid prescribing patters Narcotic prescriptions Emergency department Inpatient pain Post-operative pain
Background: Pain gained recognition as a vital sign in the early 2000s, underscoring the importance of accurate documentation, characterization, and treatment of pain. No prior studies have demonstrated the utility of the 0–10 pain scale with respect to discharge opioid prescriptions, nor characterized the most influential factors in discharge prescriptions. Methods: Inpatient and emergency department(ED) encounters from July 1, 2012 to April 1, 2018 resulting in a discharge prescription for tablet opioid medications were identified. The primary outcome was to determine if pain levels in 24 h prior to discharge correlated with opioids (in milligrams of morphine equivalents (MME)) prescribed. Secondary outcomes included the impact of patient and prescriber demographics, demographics. A generalized linear model was created to investigate factors affecting the quantity of prescribed opioids. Results: n = 78,691 patient encounters. Overall mean adjusted MME for non-ED visits was 378 versus 197 for ED visits. Whites received the highest quantities; those identifying as non-white and non-black received the lowest. Women received significantly fewer discharge MMEs in both the ED and inpatient cohorts. Provider prescribing patterns exhibited the most profound effect on discharge MMEs. The most prolific (≥300 prescriptions over the study period) writing the largest amount. In the ED, there was a significant negative correlation between documented pain levels and discharge MMEs(ρ = 0.074,p < 0.001). Conclusions: Pain scale was significantly negatively correlated with discharge MMEs in the ED and positively correlated in the inpatient population. Individual prescriber characteristics were the more influential variable, with prolific high prescribers writing for the largest MME amounts. The inverse association of pain and MMEs at discharge in the ED, and the large effect pre-existing prescriber patterns exhibited, both improved methodology for assessing and appropriately treating pain, and effective prescriber-targeted interventions, must be a priority.
1. Introduction In the early 2000s, the idea of pain as the “fifth vital sign” was introduced. The impetus for recording pain as a vital sign was to ensure routine assessment and avoid the undertreatment of pain. However, this increased focus on pain, in combination with other societal factors, likely contributed to the boom in opioid prescriptions, dependency, addiction, and overdose that comprises today’s “Opioid Epidemic.” As a result, in 2016, Physicians for Responsible Opioid Prescribing (PROP) have asked for pain-treatment questions to be removed from satisfaction surveys that may be linked to hospital ratings and individual provider quality metrics [1,2]. Examples of these questions include “During this hospital stay, how often did the hospital staff do everything they could to help you with your pain?” (bold emphasis added
⁎
by the authors). Depending on the patient and their reason for admission, complete elimination of pain – as implied by this question – may not be possible or desired. Seventy-five percent of all opioid misuse begins with people using medication that was not prescribed for them [3–5]. Given that, in 2012, 289 million prescriptions for opioids were dispensed in the United States [6], proper opioid stewardship, such as prescribing medications in proportion to patient pain levels and minimizing overprescribing, is of paramount importance. It is equally important to note that failure to utilize opiates in appropriate circumstances causes suffering and creates a barrier to effective pain care [7]. Despite the clear need to balance these competing priorities with meaningful and clinically relevant pain assessments, there have been no published studies validating the utility of the most commonly used 0–10 pain scale or delineating the factors
Corresponding author. E-mail address: [email protected] (J.A. Villwock).
https://doi.org/10.1016/j.ijmedinf.2019.05.005 Received 7 September 2018; Received in revised form 18 April 2019; Accepted 8 May 2019 1386-5056/ © 2019 Elsevier B.V. All rights reserved.
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
that most highly impact prescribing patterns. This study sought to address both these knowledge gaps by retrospectively analyzing opioid discharge prescriptions during a five-year period at an academic, tertiary care hospital. The primary objective was to determine correlations between the 0–10 pain scale and prescribing patterns. Secondary objectives included characterizing the impact of patient and provider factors on prescriptions. This included subset analysis comparing prescribing patterns for inpatient patients versus those who were treated, and ultimately discharged, in the emergency department setting.
Table 1 Patient demographics.
2. Methods The methodology for this quality improvement study was reviewed by the University of Kansas Institutional Review Board and deemed to be non-human subject research. 2.1. Patient cohort All patients who were (1) admitted for > 24 h or (2) treated in our ED between July 1, 2012 and April 1, 2018 were identified in our electronic medical record (EMR). These data were queried to identify all discharges containing a discharge prescription for hydrocodoneacetominophen, oxycodone, or oxycodone-acetominophen were identified by a trained informatics specialist. These medications were selected as they were the three most common opioid in tablet form prescribed during the study time period. Eligible encounters were divided into groups of patients who had undergone procedures – as determined by a recorded trip to the operating room/procedural suite and an associated operative/procedure note in the EMR – and those who had not. The following additional data points were also collected: quantity of discharge opioids; ordering and authorizing provider; number of prescriptions per trainee (resident or fellow), attending physician, or advanced practice practitioner (APP); patient length of stay; and pain level in the 24 h preceding discharge. Morphine milligram equivalents (MME) were determined by the following drug-specific multipliers: hydrocodone, 1; oxycodone, 1.5 to control for the strength of the medication prescribed [8]. The Area Deprivation Index (ADI), a composite score generated from seventeen socioeconomic variables, was derived from patient primary residence zip code and used as a proxy for socioeconomic status [9].
Non-Emergency Department Visits
Emergency Department Visits
Age, y [median (IQR)]
41 (56–66)
40 (29–54)
Insurance Commercial Medicare Medicaid Other No insurance
20243 (37.0%) 17072 (31.2%) 4950 (9.1%) 1561 (2.9%) 10812 (19.8%)
7390 (30.7%) 3468 (14.4%) 3205 (13.3%) 790 (3.3%) 9200 (38.2%)
Race White Black Other
41413 (75.8%) 7284 (13.3%) 5711 (10.5%)
13102 (54.5%) 6858 (28.5%) 4028 (16.7%)
Gender Male Female
25726 (47.1%) 28912 (52.9%)
11254 (46.8%) 12799 (53.2%)
Area Deprivation Index < 20 20–39 40–59 60–79 > = 80
2056 (3.8%) 8612 (15.8%) 15777 (28.9%) 18497 (33.9%) 7940 (14.5%)
571 (2.4%) 2552 (10.6%) 5492 (22.8%) 7336 (30.5%) 7333 (30.5%)
Patient distance from hospital, miles < 10 16979 (31.1%) 10–29 17352 (31.8%) 30–59 7490 (13.7%) > = 60 12792 (23.4%)
16102 (66.9%) 5859 (23.5%) 990 (4.1%) 1292 (5.4%)
Pain (last 24 h) 0–3 4–6 7–10
885 (3.7%) 4397 (18.3%) 15705 (65.3%)
17164 (31.4%) 25765 (47.2%) 5749 (10.5%)
3. Results Between July 1, 2012 and April 1, 2018, 78,691 encounters were analyzed at a single, tertiary care, academic medical center. 3.1. Non-emergency department visits Non-emergency department related visits constituted 69% of opioid encounters (54,638 cases). Patient and provider demographics are shown in Tables 1 and 2. There was a statistically significant increase in admission (τ = .672, p < .001) and operative procedures (τ = .787, p < .001) and decrease in overall length of stay (τ = −0.640, p < .001) during the study period. The average patient length of stay significantly decreased from 137 h in July 2012 to 111 in April 2018. Operating room procedures increased from 29% to 78% during the
2.2. Statistical analyses SPSS version 25 (Armonk, NY) was used for the data analyses. Group comparisons were performed using Chi-squared, Kruskal-Wallis, and Mann-Whitney U tests, as appropriate. Temporal trends in the monthly number of admissions, number of operative procedures, and length-of-stay were assessed via Mann-Kendall tests. Spearman’s rho was used to analyze correlation between pain-level and quantity of prescribed opioids. To further investigate factors affecting the quantity of prescribed opioids, a generalized linear model using Gamma as the distribution and Log as the link function was created. Separate models were constructed for non-emergency department visits and emergency department visits. The analysis considered age, gender, insurance status, prescriber type (attending/trainee/APP), ordering prescriber frequency (stratified based on the number of scripts written over the study period), specialty (medicine/surgical) [non-emergency model only], race, pain in the 24 h preceding discharge, patient distance from hospital, ADI, and whether the patient had an operating room procedure [non-emergency model only]. Estimated marginal means and 95% wald confidence intervals for each factor were reported. When appropriate, pairwise comparisons are reported using least significant difference adjustment for multiple comparisons. Statistical significance was set at p < 0.01.
Table 2 Provider demographics.
Provider Type Attending Trainee APP Specialty Medicine Surgical
Non-Emergency Department Visits
Emergency Department Visits
21111 (38.6%) 21765 (39.8%) 11762 (21.5%)
10288 (42.8%) 3956 (16.4%) 9809 (40.8%)
24615 (45.1%) 30023 (54.9%)
a
Ordering provider frequency (#scripts) < 100 15772 (28.9%) 100–199 15682 (28.7%) 200–299 12785 (23.4%) > = 300 10399 (19.0%) a
70
Not applicable.
a
4156 (17.3%) 2488 (10.3%) 5054 (21.0%) 12355 (51.4%)
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
same time frame. The overall mean adjusted MME on discharge for a non-emergency department visit was 462 (95% CI: 455–469). Patients identifying as Caucasian/white comprised 75.8% of the study population. Whites received the highest quantities of discharge MMEs (adjusted mean 497, 95% CI: 490–504), while those identifying as non-black and non-white received the lowest mean MMEs (adjusted mean 419, 95% CI: 409–429). Males received 10% more discharge MMEs (adjusted means, Table 1) than females (p < .001). 89% of encounters documented pain levels in the 24 h preceding admission. There was a statistically significantly positive correlation (ρ = 0.051, p < 0.001) between documented pain levels and number of milligrams of opioids prescribed at discharge. Those with a pain of 7 or more received on average 37% more (adjusted mean) opioids than patients reporting a pain of 3 or less (Table 1). Socioeconomic status, using zip code and ADI as a proxy, impacted MMEs prescribed, with the most disadvantaged group (≥80) receiving the smallest amount (Table 1). Yet, a pairwise comparison with mean MME from the least disadvantaged group (< 20) was statistically insignificant (p = .260). Patients living 30–60 miles from the hospital received the highest discharge MMEs (p < 0.001) when compared to other groups. Provider prescribing patterns exhibited the most profound effect on discharge opioid amounts. The most prolific script writers (≥300 scripts over the study period) tended to write the largest amount of opioids (34% more than providers that wrote < 100 prescriptions – Table 2). Trainees (residents and fellows) wrote significantly higher amounts of discharge opioids than their attending and APP counterparts (Table 2, p < .001).
Table 3 Adjusted means of discharge opioids in milligrams of morphine equivalents (MME) for non-emergency department and emergency department-related visits. Unadjusted means along with adjusted means from multivariate analysis are presented. Non-Emergency Department Visits Mean
95% CI Lower Upper
Mean
95% CI Lower
Upper
451 443 471 504 443
441 433 460 494 435
462 453 482 515 452
137 135 141 148 156
133 131 136 143 151
142 140 145 153 162
445 479
438 470
453 489
a
a
a
a
a
a
465 452 501 439 455
456 441 481 432 446
475 463 522 447 464
135 136 148 142 156
131 131 141 138 152
140 141 156 146 161
M F
484 441
476 434
492 449
151 136
147 132
155 140
Black Other White
474 419 497
463 409 490
485 429 504
139 142 148
135 138 145
144 147 152
463 461
455 452
471 470
a
a
a
a
a
a
479 493 418
470 484 408
488 502 428
159 163 114
154 158 110
163 168 118
Patient area deprivation index < 20 20–39 40–59 60–79 > = 80
454 471 467 475 444
437 460 459 467 435
471 481 476 483 453
136 150 144 146 140
128 145 140 142 137
145 155 149 150 144
Patient pain as reported in 24 hrs preceding discharge < =3 4-6 > =7
397 458 543
390 451 530
404 466 555
164 144 124
156 140 121
172 148 127
Patient distance from hospital, miles < 10 10–29 30–59 > = 60
448 459 481 460
441 451 470 450
456 467 492 471
124 130 163 161
121 126 156 153
127 133 171 170
Ordering provider frequency (#scripts) < 100 100–199 200–299 > = 300 Overall
432 407 449 578 462
423 399 440 565 455
441 415 458 590 469
137 152 147 138 143
132 147 142 134 140
141 158 151 142 147
Age, y
< 30 30–39 40–49 50–59 > = 60
Operative case Yes No Insurance Medicare Medicaid Other Commercial No insurance Gender
Race
Provider specialty Medicine Surgical
3.2. Emergency department visits Emergency department visits constituted 31% of opioid encounters (24,053 cases). Patient and provider demographics are shown in Tables 1 and 2. The number of admissions increased for the first two years of the study period but the trend reversed during the subsequent years, resulting in a lack of a monotonic trend (τ = −0.042, p = .605). The median length of stay was 5.7 h (95% CI: 5.1–6.3) and there was a statistically significant decrease in overall length of stay over the study period (τ = −0.222, p = .006). The overall mean adjusted MME on discharge for an emergency department visit was 143 (95% CI: 140–147). Patients identifying as caucasian/white comprised 54.5% of the ED population, a proportion significantly lower than represented within the inpatient cohort (p < .001). Whites received significantly more MMEs (adjusted mean 148, 95% CI: 145–152) than blacks (adjusted mean 139, 95% CI: 135–144) (p < .001). Males received 11% more discharge MMEs (Table 3) than females (p < .001). Patients residing > 30 miles from the hospital received a significantly higher mean MMEs on discharge (p < 0.001). 87% of encounters documented pain levels in the 24 h preceding admission. There was a statistically significantly negative correlation (ρ = −0.074, p < 0.001) between documented pain levels and number of milligrams of opioids prescribed at discharge. Those with a pain of 7 or more received on average 24% less (adjusted mean) opioids than patients reporting a pain of 3 or less (Table 3). Unlike providers treating non-emergency visits, those within the ED with a high prescription frequency (≥300 scripts over the study period) did not prescribe more MMEs than providers that wrote less than 100 scripts (Table 3, p = .512). APPs wrote significantly lower amounts of discharge opioids than their attending and trainee counterparts (Table 3, p < .001).
Provider type Attending Trainee Advanced Practice
a
71
Emergency Department Visits
Not applicable.
J.A. Villwock, et al.
International Journal of Medical Informatics 129 (2019) 69–74
4. Discussion
proportions of prescriptions. APPs represented the remaining fifth of the prescriptions. Conversely, in the ED, attending and APPs represented the bulk of prescriptions (˜40% each) with trainees responsible for only 16% of opioid prescriptions. This indicates that educational interventions may need to be specifically tailored to different prescribing groups, depending on the location. Another important finding of this study is the impact of provider prescribing patterns on opioid prescriptions. Database studies of prescription drugs have similarly shown that such high-risk prescribers – in the top 5th percentile among all prescribers – can account for a large proportion of total opioids dispensed. For example, in Florida, such high-risk providers account for only 4% of prescribers, but are responsible for 67% of the total opioid volume and 40% of total opioid prescriptions. This study also found that these prescribers are particularly responsive to prescription drug monitoring policies [16]. Given that we found that individual prescribing patterns had a highly significant impact on discharge MMEs for inpatients, this indicates interventions targeted at high prescribers may be of particular benefit.
4.1. Summary Patient pain levels, most commonly on a 0–10 scale, have been routinely documented for nearly two decades. Despite widespread use, the utility of pain levels – and their impact on clinical practice – had not previously been well-studied. Patient and provider factors that impact prescribing patterns have also not been completely elucidated. Given the current opioid epidemic – and the fact that most improperly used opioids are from “leftover” or excess prescriptions [3–5] – ensuring pain medication prescriptions are appropriate for the pain being treated is of paramount importance. This study of 78,691 inpatient encounters from July 1, 2012 to April 1, 2018 demonstrated a statistically significantly poor, and inverse, correlation between documented pain levels and opioid prescription quantity in the emergency department. Patient factors predictive of higher opioid prescriptions in the inpatient setting included white race – self-identification as non-black and non-white resulted in significantly smaller prescriptions – and male gender. Interestingly, provider characteristics exerted the most influence on prescribing patterns.
4.4. Study limitations This study is not without limitations. This is a retrospective single institution study at an academic medical center. As such, results may not be externally valid. However, given the large sample size, large patient catchment area, and diverse clinical faculty, we do not believe significant bias was introduced. We were unable to determine what proportion of the prescribed medication in the investigated encounters were filled. Additionally, our results are limited to the data available within our retrospective review. Due to the retrospective nature of the study, we also do not know the clinical decision making that occurred to justify the MMEs prescribed. Furthermore, unless admitting care teams select diagnoses related to chronic pain, we are unable to know what proportion of our study population suffer from chronic pain. This is particularly important since episodic care settings, such as inpatient admissions, should refrain from supplying opioids to chronic pain patients whenever possible [17].
4.2. Impact of patient characteristics Given the wealth of searchable information electronic health records afford, it behooves us to review and ensure that high-quality, data-driven care is being provided [10]. We found that, while documented pain levels correlated with discharge prescription amounts for the inpatient cohort, this was not true of patients who received their treatment in the ED. In the latter group, there was a statistically significantly negative correlation between pain level with discharge opioids. In fact, those with higher levels of pain (> 7) received an average of 24% less opioids than patients with pain levels of 3 or less. This data may underscore the frustrations of many clinicians who feel that current pain assessments are devoid of meaning because they are not linked to function or other outcomes of interest. Patient demographics did impact prescribing patterns, especially those of race and gender. For example, our results show that blacks received significantly less opioids than whites in both the inpatient and ED cohorts. This is consistent with the existing literature on racial disparities in pain control. For example, it has previously been shown that blacks are at elevated risk for undertreatment and underestimation of pain by their health care team [11,12]. However, interestingly, these differences have been shown to be significantly impacted by the specialty of the prescribing provider [6]. For example, in academic primary care, it has been found that Hispanic, Asian, and older patients are less likely to be prescribed higher opiate doses [13]. We similarly found that patients identifying as non-black and non-white received the lowest mean MMEs. In our study, women also received significantly less MMEs than men in both the inpatient and ED cohorts. This is in agreement with prior literature describing that male physicians are more likely to prescribe higher opioid amounts to male patients [14]. Similarly, women with metastatic cancer have previously been found to be significantly more likely than men (OR 5) to be undertreated for pain [15]. Distance from the hospital also impacted opioid prescribing patterns. In the ED setting, patients living > 30 miles away from the hospital received significantly more MMEs than those living closer. Among inpatients, those living 30–60 miles from the hospital received the highest MMEs on discharge. This is in agreement with a study conducted by Bauer et al reporting that patients with nonlocal home addresses were more likely to be prescribed higher opiate doses [13].
4.5. Future directions We identified multiple factors that impacted the quantity of opioids prescribed upon discharge from acute hospital care. The impact of prescriber factors highlight the need for targeted research and interventions on educational interventions targeted at these groups. Factors leading to the differences in patient care settings (inpatient versus emergency department care) and the fact that higher pain levels are associated with smaller opioids quantities upon discharge from the emergency department should also be elucidated. For patients living > 30 miles from the hospital, research targeting ways to ensure appropriate outpatient follow-up and management of pain is needed to avoid reflex prescribing of larger quantities. 5. Conclusions There were multiple factors associated with differential opioid prescribing patterns. Prolific prescription writers were responsible for the largest adjusted MMEs on discharge for inpatients, indicating that targeted interventions for these high-risk prescribers may be of particular benefit. In the ED, the 0–10 pain scale poorly correlates with discharge opiate prescriptions. Improved methods for assessing pain should also be a priority. Author statement
4.3. Impact of prescriber characteristics
All authors have made substantial contributions to all of the following: (1) the conception and design of the study, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or
With respect to provider characteristics, in the inpatient setting, trainees and attendings were responsible for approximately equal 72
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
revising it critically for important intellectual content, (3) final approval of the version to be submitted.
Acknowledgement The authors would like to thank Jennifer B. Wilson RN, BSN for her instrumental support in the data acquisition necessary to complete this work.
Funding None.
Appendix 1 Unadjusted means of discharge MMEs for non-emergency department and emergency department visits
Age, y
Operative case Insurance
Gender Race
Provider specialty Provider type
Non-Emergency Department Visits 95% CI Mean Lower
Upper
Emergency Department Visits 95% CI Mean Lower
Upper
< 30 30-39 40-49 50-59 > =60
434 430 462 504 446
420 416 446 489 439
448 443 479 519 454
113 109 115 122 133
110 105 112 118 128
116 112 118 126 138
Yes No
452 468
447 455
457 481
* *
* *
* *
Medicare Medicaid Other Commercial No insurance
469 456 542 445 450
459 436 514 438 438
480 476 569 453 462
121 101 126 116 121
116 98 118 113 118
125 104 135 120 123
M F
480 438
473 431
488 445
127 108
124 107
129 110
Black Other White
451 381 469
435 367 463
468 395 475
106 112 124
103 108 122
108 115 127
Medicine Surgical
461 454
451 449
472 460
* *
* *
* *
450 468 450
441 459 441
459 478 459
134 139 89
131 133 88
136 145 90
428 455 461 472 427
406 441 452 462 414
449 469 470 482 440
116 128 121 120 107
108 123 118 116 104
123 133 125 123 109
406 470 556
400 462 533
412 477 579
169 137 108
154 132 107
184 142 110
439 457 487 466
429 447 474 455
449 466 501 478
109 123 163 165
108 119 150 147
111 127 176 183
454 409 449 543 445
439 402 441 532 440
469 416 458 553 450
134 132 131 102 116
129 128 126 100 114
139 137 136 103 117
Attending Trainee Advanced Practice Patient area deprivation index < 20 20–39 40–59 60–79 > = 80 Patient pain as reported in 24 hrs preceding discharge < =3 4–6 > =7 Patient distance from hospital, miles < 10 10–29 30–59 > = 60 Ordering provider frequency (#scripts) < 100 100–199 200–299 > = 300 Overall
* Not applicable.
within 180 days, after allowing a reasonable period for public comment. If you have any questions about our requested changes, please contact the office of Dr. Andrew Kolodny at (347) 396-0371.,” p. 4. [3] “Results from the 2014 National Survey on Drug Use and Health: Detailed Tables, SAMHSA, CBHSQ.” [Online]. Available: https://www.samhsa.gov/data/sites/ default/files/NSDUH-DetTabs2014/NSDUH-DetTabs2014.htm#tab6-47b. [Accessed: 27-May-2018]. [4] M. Szalavitz and M. Szalavitz, “Opioid Addiction Is a Huge Problem, but Pain Prescriptions Are Not the Cause,” Scientific American Blog Network. [Online]. Available: https://blogs.scientificamerican.com/mind-guest-blog/opioid-addiction-
References [1] “Johnson, Senators Introduce Bipartisan PROP Act - Press Releases - Ron Johnson.” [Online]. Available: https://www.ronjohnson.senate.gov/public/index.cfm/pressreleases?ID=CC5F9714-9A1A-4B24-AE50-A86D5AAB7F70. [Accessed: 27-May2018]. [2] H. Chen, R. J. Goldsmith, and K. M. Murphy, “For the reasons discussed in this petition, the undersigned Petitioners request that CMS issue a proposal for removing the pain questions listed above within 90 days, and finalize the rule change
73
International Journal of Medical Informatics 129 (2019) 69–74
J.A. Villwock, et al.
[5] [6] [7] [8] [9] [10] [11]
is-a-huge-problem-but-pain-prescriptions-are-not-the-cause/. [Accessed: 27-May2018]. N.D. Volkow, A.T. McLellan, Opioid abuse in chronic pain — misconceptions and mitigation strategies, N. Engl. J. Med. 374 (March (13)) (2016) 1253–1263. B. Levy, L. Paulozzi, K.A. Mack, C.M. Jones, Trends in opioid analgesic-prescribing rates by specialty, US., 2007-2012, Am. J. Prev. Med. 49 (September (3)) (2015) 409–413. K.K. Dineen, J.M. DuBois, Between a rock and a hard place: can physicians prescribe opioids to treat pain adequately while avoiding legal sanction? Am. J. Law Med. 42 (no. 1) (2016) 7–52. D. Dowell, T.M. Haegerich, R. Chou, CDC guideline for prescribing opioids for chronic pain–United States, JAMA 315 (15) (2016) 1624–1645. Apr. 2016. A.J.H. Kind, W.R. Buckingham, Making neighborhood-disadvantage metrics accessible — the neighborhood atlas, N. Engl. J. Med. 378 (June (26)) (2018) 2456–2458. D. Backenroth, H.S. Chase, Y. Wei, C. Friedman, Monitoring prescribing patterns using regression and electronic health records, BMC Med. Inform. Decis. Mak. 17 (December) (2017). Institute of Medicine, US) committee on understanding and eliminating racial and
[12] [13] [14] [15] [16] [17]
74
ethnic disparities in health care, Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care, National Academies Press (US), Washington (DC), 2003. A. Cintron, R.S. Morrison, Pain and ethnicity in the United States: a systematic review, J. Palliat. Med. 9 (December (6)) (2006) 1454–1473. S.R. Bauer, L. Hitchner, H. Harrison, J. Gerstenberger, S. Steiger, Predictors of higher-risk chronic opioid prescriptions in an academic primary care setting, Subst. Abuse 37 (January (1)) (2016) 110–117. C.S. Weisse, P.C. Sorum, K.N. Sanders, B.L. Syat, Do gender and race affect decisions about pain management? J. Gen. Intern. Med. 16 (April (4)) (2001) 211–217. D.E. Hoffmann, A.J. Tarzian, The Girl Who Cried Pain: a Bias Against Women in the Treatment of Pain, Social Science Research Network, Rochester, NY, 2001 SSRN Scholarly Paper ID 383803. H.-Y. Chang, et al., Impact of prescription drug monitoring programs and pill mill laws on high-risk opioid prescribers: a comparative interrupted time series analysis, Drug Alcohol Depend. 165 (01) (2016) 1–8. T.L. Skaer, A.C. Nwude, Opioid prescribing laws and emergency department guidelines for chronic non-cancer pain in Washington State, Pain Pract. Off. J. World Inst. Pain 16 (5) (2016) 642–647.