Risk Stratification for Opioid Misuse in Children, Adolescents, and Young Adults: A Quality Improvement Project
BACKGROUND: The Pediatric Palliative and Comfort Care Team (PACT) at Cincinnati Children’s Hospital Medical Center (CCHMC) provides opioids to a large population of patients in the ambulatory setting. Before this project, PACT had no reliable system to risk stratify patients for opioid misuse.
METHODS: The global aim was safe opioid prescribing by the palliative care team. The specific, measurable, achievable, realistic, and timely aim was as follows: “In patients who present for follow up with PACT, we will use the “opioid bundle” to increase risk stratification for opioid misuse from 0% to 90% over 5 months.” The opioid bundle includes a urine drug screen, Ohio Automated Rx Reporting System report, pill count, and screening history for drug abuse and mental health disorders. The setting was multiple CCHMC ambulatory clinics. Participants included all PACT members.
RESULTS: Since implementing the new system, we have increased risk stratification for opioid misuse among outpatients from 0% to >90%. Results have been sustained for 12 months. Key processes have become reliable: obtaining informed consent and controlled substance agreements for all new patients and obtaining the opioid bundle to enable risk stratification in a consistent and timely fashion. A total of 34% of patients have been stratified as high risk, and an additional 27% have been stratified as moderate risk.
CONCLUSIONS: A system to ensure safe opioid prescribing practices to all patients is critical for providers. Identifying key processes and executing them reliably has enabled the palliative care team at CCHMC to risk stratify >90% of patients receiving opioids in the ambulatory setting for opioid misuse.
- BMT —
- bone marrow transplant
- CCHMC —
- Cincinnati Children’s Hospital Medical Center
- HR —
- high risk
- ID —
- LOR —
- level of reliability
- MR —
- moderate risk
- OARRS —
- Ohio Automated Rx Reporting System
- PACT —
- Pediatric Palliative and Comfort Care Team
- UDS —
- urine drug screen
Prescription drug abuse is a national epidemic. In 2012, 4.9 million Americans admitted to the misuse of pain relievers, making them the second most commonly used illicit substance in the United States.1 The Centers for Disease Control and Prevention estimate that 44 people die in the United States every day from overdose of prescription painkillers.2 Palliative care physicians are often responsible for managing pain and other symptoms in patients with cancer and life-limiting diagnoses. Because opioids are the cornerstone of treatment of cancer-related pain, potential patient misuse of prescription medications and national regulatory responses should be of particular concern to palliative care physicians and oncologists.3 But general pediatricians, pediatric surgeons, and pediatric dentists may also be responsible for providing opioids to patients in the ambulatory setting. A model to enable safe opioid prescribing practices in patients with cancer should be accessible to general pediatric providers as well.
Pediatricians lack both data and expert guidelines to guide risk assessment for opioid misuse in children. To our knowledge, only 1 retrospective chart review study, from St Jude’s Children’s Research Hospital, has focused on this question. The medical records of nearly 400 adolescents and young adults receiving oncologic therapy were examined for “aberrant opioid-associated behavior.”4 A wide range of aberrant opioid-associated behaviors is reported in the literature, but these may include claiming to have lost pills or prescriptions or claiming that they were stolen, visiting multiple physicians or the emergency department to obtain pain medications, seeking and using medications for purposes other than pain relief (eg, relaxation, euphoria), stealing or selling drugs, concurrent use of alcohol or other illicit drugs, resisting changes to the medication regimen, and third-party concerns about opioid use.5–10 Of the nearly 400 patients included, researchers found that 94 received opioids, and 11.7% of these exhibited aberrant opioid-associated behaviors.5 Prospective studies among pediatricians evaluating patients for risk of opioid abuse or diversion, and consensus recommendations for management, are not available to our knowledge.
Research devoted to screening patients for opioid misuse most often focuses on adults with chronic noncancer pain. In 2009, the American Pain Society and American Academy of Pain Medicine formulated recommendations for the use of chronic opioid therapy in chronic noncancer pain, which include guidelines for obtaining informed consent for opioids, monitoring patients, and caring for high-risk (HR) patients.11 These guidelines state that “monitoring that involves regular, repeated evaluations and addresses a variety of domains is likely to be more informative than infrequent, narrowly focused evaluations,” but they do not suggest specific time intervals for such monitoring. Routine pill counts, urine drug screening, family or caregiver interviews, and use of prescription drug monitoring program data are noted to be helpful supplements in monitoring for potential misuse.11
An emerging body of literature has focused on screening for drug misuse in adult patients who have cancer or are receiving palliative care.3,12–14 A recent retrospective review of adults with cancer followed by a palliative care team found that 43% of patients were classified as having moderate risk (MR) or HR of opioid misuse.3 Yet a national survey of palliative care clinics found that only about a third (32.4%) used a screening tool to screen patients for opioid misuse, although half indicated that substance abuse and diversion were problems for their clinics.14 Similarly, a survey of hospices in Virginia found that less than half of respondents had policies for screening patients or family members for substance abuse.12
In September 2014, the Ohio state legislature enacted a bill requiring providers to obtain informed consent from parents or guardians before prescribing opioids to minors. The Pediatric Palliative and Comfort Care Team (PACT) at Cincinnati Children’s Hospital Medical Center (CCHMC) cares for a large population of children, adolescents, and young adults in the ambulatory setting and provides the bulk of pain management to patients with cancer who have moderate or severe pain. Before September 2014, PACT had no formal process to obtain informed consent for opioids, to contract patients for expectations surrounding opioid prescriptions, to risk stratify patients for opioid misuse, and to monitor patients for opioid abuse or diversion after risk stratification. A simplified failure mode and effects analysis15 identified inconsistent practices in screening PACT patients for opioid abuse or diversion, poor PACT provider knowledge of the importance of screening, discomfort in asking patients and families directly about drug abuse history, and an unreliable system for documenting and communicating concerns when they did occur. Therefore, we undertook a quality improvement project using an “opioid bundle” to risk stratify all patients for opioid misuse, with the global aim of safe opioid prescribing by our group. The “opioid bundle” included a prescription drug monitoring report, a urine drug screen (UDS), a questionnaire to screen for personal or family history of drug abuse or mental health disorders, and a pill count. The project’s specific, measurable, achievable, realistic, and timely aim16 was as follows: In patients who present for ambulatory care with the palliative care service, in the first encounter after informed consent and contracts for opioid prescriptions have been signed, we will use the “opioid bundle” to increase risk stratification for opioid abuse or diversion from 0% in September 2014 to 90% by January 21, 2015.
This study was reviewed by the CCHMC Institutional Review Board (study identification [ID] number 2016-2157) and determined not to be human subjects research.
The population of patients cared for by PACT in the ambulatory setting includes patients who may be enrolled in our hospice program, with varying life expectancies. Whether to risk stratify these patients was a source of controversy, and initially hospice patients were excluded. However, over the first several weeks of the project, we became aware that some patients or parents in this population were indeed at HR of opioid abuse, diversion, or misuse. Therefore, to standardize and provide the safest care, we began risk stratifying all patients followed in the ambulatory setting regardless of prognosis or enrollment in hospice.
CCHMC is a large tertiary-care pediatric hospital in Cincinnati, Ohio. The PACT at CCHMC is housed within the Pain Division of the Department of Anesthesia. In addition to providing palliative services, PACT is responsible for managing moderate to severe pain for all inpatients and outpatients in the oncology and bone marrow transplant (BMT) programs. Additionally, we follow all patients with a left ventricular assist device, patients who are evaluated for or receive a heart transplant, patients with Duchenne muscular dystrophy or other progressive neuromuscular disorders, and patients with myriad other chronic conditions. PACT physicians also are primarily responsible for managing pain and other symptoms in patients who are enrolled in the StarShine Hospice and Palliative Care program, a program within the umbrella of the CCHMC home care division. PACT sees >400 new inpatient consults annually and follows 15 to 20 ambulatory patients each week, typically in clinic appointments with the primary team (eg, oncology, cardiology).
Planning the Intervention
The improvement team consisted of 2 PACT physicians (R.T. and M.M.), 1 PACT nurse practitioner (K.P.), the PACT nurse coordinator (M.T.), and 1 quality improvement consultant (M.A.). R.T. primarily developed this project during a CCHMC-sponsored course on quality improvement methods; M.A. was an instructor whose time and investment in the project were part of this class. A key driver diagram (Fig 1) was developed based on the theory that improvement would be driven by PACT knowledge of the importance of risk stratification, how to risk stratify patients, and awareness of which patients needed bundle elements at the correct time. Key drivers included PACT providers being able to easily access updated patient information, providers being aware of which bundle elements were needed for patients, providers having the ability to easily and reliably obtain all bundle elements for all patients, providers being aware of the risk stratification algorithm (Fig 2) and being easily able to apply it, and buy-in from all team members.
The American Pain Society and American Academy of Pain Medicine guidelines11 and recommendations from the Centers for Disease Control and Prevention2 indicate that using a single test or screening measure is not adequate to risk stratify patients for opioid misuse. Therefore, we developed the “opioid bundle,” which included a urine drug screen, Ohio Automated Rx Reporting System (OARRS, or prescription drug monitoring) report, pill count, and screening history for drug abuse and mental health disorders.
Some PACT members expressed concern that using a detailed screening tool to review family history of drug abuse and mental health problems might contribute to the caregiver burden and psychological distress already associated with caring for a child with cancer or other serious illness.17–20 Failure to complete the history portion of the opioid bundle was also identified during our simplified failure mode and effects analysis. Therefore, our team developed a short screening questionnaire asking whether anyone in the child’s home or the patient had a history of drug abuse (yes or no) or mental health problems (yes or no). This screening was presented as part of the initial visit in the ambulatory clinic in which opioids were prescribed. If the parent or patient answered affirmatively to any of the questions, follow-up questions were asked verbally to obtain more detail regarding this history.
The CCHMC laboratory offers a urine screening test for illicit drugs and a urinary drug of abuse panel including mass spectrometry, which is a much more sensitive test for detecting specific illicit drugs. For young children, because the goal of the UDS was simply to check for positivity for prescription opioids and confirm the history of home administration, the screening test was used. For adolescents (children ≥12 years old) and adults, because our goal was to screen for illicit drugs in addition to those that had been prescribed, the more sensitive drug of abuse panel was used. The drug of abuse panel uses mass spectrometry to test for multiple benzodiazepines and metabolites, opioids and metabolites, cannabinoids, cocaine, amphetamines, barbiturates, nicotine, and phencyclidine.
Interventions were designed to increase the percentage of ambulatory patients risk stratified for opioid misuse by week. Multiple plan–do–study–act cycles were conducted to test interventions developed to address key drivers. The interventions were carried out sequentially in the order listed below, between September 2014 and December 2014.
In September 2014, the improvement team began extensive education of PACT providers during monthly business meetings and weekly patient care meetings regarding the prescription drug abuse epidemic in Ohio, the importance of screening all patients for opioid misuse, and the rationale behind the screening methods chosen. All PACT team members participated in ≥1 educational session.
Informed Consent (Intervention)
Before patients were eligible for risk stratification, informed consent was obtained according to Ohio law, and a controlled substances agreement was obtained. Our standard of care in obtaining UDS and reviewing OARRS reports was explained to patients during this process.
Patient Identification Cards (Intervention)
As we began the process of completing the bundle, it became clear that a driver was PACT provider knowledge of whether patients had already been consented and risk stratified. Quick creation of a reliable and easily accessible electronic medical record database was not feasible. Therefore, we developed patient ID cards, maintained in a central area of the office, for all ambulatory patients receiving opioids. The cards identified the dates of informed consent and bundle completion in addition to UDS, OARRS, and pill count results at each visit. Cards were then incorporated into clinic huddles.
Clinic Huddles (Intervention)
A clinic huddle lasting 5 to 10 minutes was incorporated into morning and afternoon PACT rounds. All team members participated in the huddle, and no additional staff or monetary resources were needed. In the morning, the outpatient list for the day was reviewed along with bundle elements necessary for each patient. The nurse coordinator (M.T.) gave providers a reminder checklist of necessary bundle elements and a copy of an OARRS report and ordered UDS tests. Providers reviewed patient ID cards to quickly review notes from previous visits, results of previous bundle elements (UDS results, OARRS findings, pill count findings), and elements needed that day. In the afternoon, outpatients and plans of care were reviewed and ID cards updated by the providers who saw each patient.
Family Screening Questionnaire (Intervention)
The initial plan was to ask patients directly for a history regarding drug abuse or mental health disorders, but as noted above, some providers expressed discomfort directly asking parents and patients for this history as part of their interview. To improve provider comfort, a printed screening questionnaire was developed. Parents or patients were given the opportunity to complete this questionnaire in private if they so chose, and any affirmative answers to screening questions were then explored in more detail by providers. We found that parents were often open about a positive history of drug use in family members, both living in the home and out of the home, and parents and patients were open about adolescent or young adult history of drug abuse. Additionally, many had concerns about the use of opioids for themselves or their child, and the screening history allowed us to begin a meaningful conversation about risks and benefits of these medications for individual patients.
The primary outcome was the percentage of patients eligible for risk stratification (ie, for whom informed consent and a controlled substances agreement had been completed at the immediately preceding visit) who were risk stratified for opioid abuse or diversion (ie, for whom all elements of the bundle were obtained), by week. Each ambulatory patient encounter, as documented in the electronic medical record, was reviewed by R.T. daily and then weekly to assess for eligibility for risk stratification and whether bundle elements had been appropriately obtained (yes or no). If all bundle elements were obtained, the patient encounter was considered successful; if some but not all elements were obtained, the patient encounter was considered a failure. If a patient had been seen before and bundle elements not obtained (failure), at the next visit they were considered eligible again. Patients who were enrolled in the CCHMC hospice program were considered eligible for risk stratification if they presented to clinic or were seen in the home by a PACT physician.
We initially constructed a weekly run chart (Fig 3)21 to track the impact of our interventions and to observe small changes over short time intervals. Because weekly numbers were small, after 6 months when variation had decreased, we converted to a p-chart (Fig 4)22 based on monthly data. Standard control chart rules were used to determine when to adjust the center line.22
The median proportion of patients who were risk stratified for opioid misuse (all bundle elements obtained) at baseline was 0%. As a result of improvement efforts, the median shifted to >90% by November 24, 2014 and was sustained through December 5, 2015. The control chart uses mean rather than median, and the shift of the mean to >90% occurred later than the shift in the median, in June 2015. This mean >90% has been sustained for >6 months. Between September 2014 and December 5, 2015, 106 patients were consented for opioid prescriptions or completed controlled substance agreements. Many of these patients were seen only once and did not present for follow-up. Sixty-two patients were subsequently risk stratified for opioid misuse. Patient demographic information, reasons for HR and MR stratification, and current prescription status are summarized in Table 1.
Twenty-one patients, or 34% of those who were risk stratified, were found to be at HR of opioid misuse. We are no longer prescribing opioids to 4 of these patients because they died of underlying disease, and we are no longer prescribing to 9 others for various reasons summarized in Table 1. Some of these patients were initially classified as low risk or MR, but over time risk stratification was changed based on new information or emerging behaviors surrounding opioid prescribing. Cases of diversion were made evident to our team through communication with pharmacies or because pill counts or prescription monitoring reports were inconsistent with one another and with the history provided. All HR patients are seen at least monthly by our team, UDS tests and pill counts are obtained at least monthly, and OARRS reports are reviewed at least monthly.
Seventeen patients, or 27% of those who were risk stratified, were classified as having MR of opioid misuse. Three of these patients have died because of underlying disease, 1 was weaned off of opioids, care was transferred back to the primary team for 2, and care was transferred to adult chronic pain teams in other institutions for 2 others (Table 1). MR patients are seen at least every 3 months by our team, UDS tests and pill counts are obtained at least every 3 months, and OARRS reports are reviewed at least every 3 months.
The remaining patients were classified as low risk. A provider sees these patients at least every 6 months, and OARRS reports and pill counts are requested at every visit. A UDS test is obtained at least annually.
We used quality improvement methods to sustainably increase the percentage of palliative care patients who are risk stratified for opioid misuse from 0% to >90% in 5 months. This improvement has been sustained for an additional 12 months.
Five interventions allowed us to affect our key drivers and to reach our goal: team education, obtaining informed consent from families for opioid prescriptions, patient ID cards, clinic huddles, and printed patient screening questionnaires. Reliability principles can be used to decrease the rate of system failure in health care.23 Interventions can be classified by their level of reliability (LOR), which translates to the expected failure rate that can be achieved with the intervention.23 Level 1 reliability, or 10−1, is achieved when there is an 80% to 90% success rate or 1 to 2 failures out of every 10 opportunities.24 We began with an LOR 1 intervention of educating our team regarding the importance of screening all patients for opioid misuse and the evidence behind using the elements of the opioid bundle to risk stratify patients. This intervention laid the groundwork for our higher-level interventions to succeed.
All other interventions we used were LOR 2, which would be expected to yield a failure rate of 10−2, or ≤5 failures per 100 opportunities.23 Ongoing review of failures reveals that the majority occurred in hospice patients. These ambulatory patient visits are typically scheduled in the oncology clinic, and PACT initially follows along for symptom management. Our role as hospice providers may evolve as the patient’s disease progresses. Often, our team is prepared to obtain bundle elements, but parents and children receive bad news or are involved in emotion-laden conversations about potential phase chemotherapeutic trials on that day. Thus most failures occur when emotion is heightened by such circumstances, and we choose to delay bundle completion out of sensitivity to the patient and family.
Before this project, no standard of care was in place to risk stratify patients or to monitor those who were identified to be at HR of opioid misuse. We believe the greatest impact of this project on our team is a heightened and earlier awareness of which patients are HR, which potentially allows earlier intervention in the care plan for patients who are known or believed to be diverting or abusing medications.
As team communication surrounding HR patients has improved, we have recognized an emerging problem in inconsistent management of patients whose ongoing behaviors necessitate concern. We have not developed standards regarding which patients to wean from opioids and for what reasons. At times, especially after difficult patient encounters, provider satisfaction in caring for these patients is anecdotally low. Therefore, as the next step to this project, we are developing a quality improvement project aimed at standardizing the care we provide to HR patients to decrease variation in care and to improve provider satisfaction in caring for HR patients.
A system to ensure safe opioid prescribing practices to all patients is critical for providers caring for patients with cancer. Through implementation of a quality improvement initiative, we were able to improve identifying key processes and execute them reliably, enabling PACT at CCHMC to risk stratify >90% of patients receiving opioids in the ambulatory setting for opioid abuse or diversion.
- Accepted June 30, 2016.
- Address correspondence to Rachel Thienprayoon, MD, MSCS, Pediatric Palliative and Comfort Care Team, Dept of Anesthesiology, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, ML-2001, Cincinnati, OH 45229-3039. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- US Department of Health and Human Services
- Centers for Disease Control and Prevention
- Barclay JS,
- Owens JE,
- Blackhall LJ
- Ehrentraut JH,
- Kern KD,
- Long SA,
- An AQ,
- Faughnan LG,
- Anghelescu DL
- Ehrentraut JH,
- Kern KD,
- Long SA,
- An AQ,
- Faughnan LG,
- Anghelescu DL
- Anghelescu DL,
- Ehrentraut JH,
- Faughnan LG
- Scholtes PR
- Doran GT
- Rosenberg AR,
- Dussel V,
- Kang T, et al
- Benneyan JC,
- Lloyd RC,
- Plsek PE
- Nolan T,
- Resar R,
- Haraden C,
- Griffin FA; Institute for Healthcare Improvement
- Copyright © 2017 by the American Academy of Pediatrics