Barriers to Increasing Prescription Drug Monitoring Program Use: A Multidisciplinary Perspective

Abstract

Prescription drug monitoring programs are implemented through individual state policies and are one solution to curb the opioid crisis. The objectives of this study are to 1) describe the multidisciplinary experiences using this program in practice, 2) identify limitations of the program and the desired features for improvement, and 3) characterize expectations for improved access when prescription drug monitoring programs are embedded in the electronic health record. A qualitative descriptive study design used semi-structured interviews of 15 multidisciplinary healthcare providers. Textual data were analyzed using content analysis. Results showed the prescription drug monitoring program was helpful to decision-making processes related to opioid prescribing and referral to treatment, there were barriers limiting healthcare providers’ use of the prescription drug monitoring program, preferences were delineated for integrating prescription drug monitoring program into electronic health record, and recommendations were provided to improve the program and increase use. In conclusion, the prescription drug monitoring program was viewed as useful in making strides to reduce the impact of inappropriate opioid prescribing in our country. By engaging a multidisciplinary group of healthcare providers, solutions were offered to improve the interface and function of the prescription drug monitoring program to assist in increasing use.

Prescription Drug Monitoring Programs (PDMP) are one solution to curb the opioid crisis and have been implemented through individual state policies. Currently there are 54 operational PDMPs in the United States.¹ The positive effects of implementing this program on outcomes are well documented. This includes lower rates of prescription opioid-related hospitalizations and emergency department visits,^2,3 reduction in “doctor-shopping,”⁴ and reductions in prescribing high doses and overprescribing.⁵

PDMPs help healthcare providers (HCP) and pharmacists identify potential misuse, monitor prescribing, and analyze prescription opioid dispensing for potential risk of harm. Two characteristics of these programs have been identified to improve outcomes: monitoring four or more drug schedules, and data entry that updates at least weekly. These characteristics are associated with greater reductions in opioid-related overdose deaths than programs without these features.⁵

Despite positive outcomes, PDMPs continue to be underutilized by prescribers for a variety of reasons.^6,7 One national survey of 1,000 practicing primary care physicians in 2014 showed that 72% were aware of PDMP however, only 53% reported using the programs.⁸ State-wide mandates for using this program have had an inconsistent effect on use.^9,10 Barriers were cited as a user interface that lacked an intuitive format, had time-consuming information retrieval, privacy concerns, lacked interoperability, and lack of a national program connection.^7–11 A cross sectional study from 2017 to 2018 indicated that integration of PDMP into hospital Electronic Health Records (EHR) could reduce some of these barriers, yet only 22% to 30% of hospitals integrated.¹² Integration increased in 2019 with 84% of hospitals and 61% of clinicians adopting and implementing application programming interface technology to allow for integration of data across systems.^13,14 Integration of PDMP into EHR is a logical step to overcome barriers, however, we still need to determine how to increase use by HCPs. The purpose of this study is to investigate how HCPs and pharmacists use the PDMP in practice, what its limitations to using are, and to understand interventions to increase use.

Objectives

The objectives of this qualitative study were to describe multidisciplinary prescriber and pharmacist experiences using the PDMP in practice, including ways they use the program to enhance care. The participants were to identify limitations of the PDMP and desired features for improvement. Participants were also asked about their expectation for improved access when integrating the program into the EHR.

Method

The [blinded] Prescription Monitoring Program has been in existence since 2009 and is maintained by the [blinded] Board of Pharmacy. The [blinded] Department of Public Health and the [blinded] Board of Pharmacy partnered with a Midwest academic center to trial the integration of the state PDMP into the hospital EHR. Single-participant qualitative interviews were used for data collection prior to integration. The interviews occurred between February 2020 and June 2020. A semi-structured interview guide was developed (see Figure 1) with the intent to spur participants to share their practice experiences and perspectives on the strengths and weaknesses of the use of the PDMP in their clinical role prior to integration of the program into the EHR, and how use could be improved. Interviews were conducted in-person, and in March were conducted over the phone or via video conference due to COVID-19 pandemic restrictions. This study was deemed quality improvement by the institutional review board and is adherent to guidelines for the ethical conduct of research.

Figure 1.

Semi-structured Interview Guide

Recruitment

The multidisciplinary participants included physicians, nurse practitioners, physician assistants, and pharmacists from multiple specialties within the health system. Inclusion criteria were, 1) performing direct patient care, 2) state licensure to practice, 3) experience using the state PDMP, and 4) fluent in English. Convenience sampling occurred to recruit participants through contacts in family practice, emergency medicine, pain medicine, and participant referral. The investigators contacted potential participants by email to explain the study and if interested, set up a time for the 30–40-minute interview. No monetary incentive was used.

Data Collection

Demographic data of participants were collected to describe the sample. Interviews were conducted by one of two interviewers experienced in conducting qualitative research ([blinded]). In-person interviews were conducted in a private location, however with pandemic restrictions, telephone or video conference interviews were conducted for most of the study. Participants were asked permission to voice record the interviews. Notes also were taken by the investigators and used for comparison and back up of data. The audio files were downloaded and transcribed using a secure transcription service and checked by one of the authors ([blinded]) for accuracy.

Data Analysis

The transcripts were imported into MAXQDA^™ (version 18, VERBI, Berlin Germany), a software program used in qualitative data management. A descriptive qualitative approach¹⁵ was used for content analysis of the textual data from interview transcripts. Transcripts were reviewed by two authors [blinded] to get a general sense of the overall dataset and record initial impressions of potential patterns, compelling findings, and differing perspectives. These investigators engaged in initial coding separately by applying descriptive codes to text segments at the idea level which was typically a phrase or sentence in length. Two authors [blinded] met to discuss and compare qualitative coding to establish consensus and sort participant statements into similar topics based on a set of descriptive codes and create a code list. One investigator [blinded] applied the code list to the transcripts which were then checked by the other two investigators [blinded]. Once the codes were applied to the transcripts, text segments were sorted by code and the investigators independently developed summaries for each code. This process led to the consolidation of codes into sub-categories. Exemplary quotes were selected to accompany the descriptions of the sub-categories to provide in-vivo support for the descriptions inferred by the investigators. Following this process, the sub-categories were further consolidated into descriptive categories and summaries of these categories were written to present the key findings.

Several approaches were used to promote rigor in the data analysis. These approaches were independent coding and consensus reaching through meetings with discussions of data during analysis. Descriptions were developed for the codes as well as identifying and presenting representative quotes from different participants. Credibility was established through the research team discussions which occurred through video communication due to COVID-19 restrictions. Through the creation of a description of prescribers’ and pharmacists’ experiences, the component of transferability was demonstrated. Dependability was assured by the stability of participants’ codes, sub-categories, and categories, with no contradictory statements occurring within the time frame.¹⁶ Finally an audit trail detailing decisions made by the research team on data collection, analysis, and presentation was maintained to establish dependability and confirmability.

Results

Fifteen HCPs were interviewed for this study. Six were physicians, four were advanced practice registered nurses, one was a physician assistant, four were pharmacists, and seven participants were female. The recording file of one of the pharmacists interviews was corrupt and this interview was not used, instead the notes taken by the investigator as part of procedure, was used for analysis. There were major topical categories from the participant interviews: 1) PDMP is helpful to clinical practice by assisting with decision-making related to opioid prescribing and referral to treatment; 2) there are barriers limiting providers’ use of the PDMP when providing care; 3) preferences for integrating program into EHR were delineated; and 4) recommendations were provided to improve the PDMP and to increase use.

PDMP is Helpful to Clinical Practice

PDMP was helpful in clinical practice encounters described by participants. These interactions included when the patients were not known to the provider, when the clinical assessment indicated necessary treatment, or when feedback was provided to the patient. Participants identified several ways this resource assisted them in improving their quality of care when opioids were prescribed.

Participants described situations when the patient was not known to them. Medical records were not always available when a new patient was encountered. However, the PDMP will have data on the patient’s scheduled medications, dates of refill and prescribers that may inform an assessment. Participants also found this content helpful when they were cross-covering for another prescriber.

Following clinical assessment, the HCP developed necessary treatment. In some cases, the treatment may include initiating medication for opioid use disorder. The PDMP can help the prescriber to verify the initial dose for opioid taper when initiating medication for opioid use disorder. One participant in the emergency department stated difficulty in admitting patients into a Medication Assisted Treatment (MAT) clinic, so having this information helped emergency department prescribers to initiate medication for opioid use disorder in that setting.

…emergency departments are obviously open 24/7… patients all of a sudden gets interest and they decide, … “Today is the day, or today is the hour that I need treatment” … these patients will come in and we can initiate it early, figured out the dose they need, … then we have great resources to get them into the MAT clinics. Medication for opioid use disorder is … the most lifesaving [resource] we actually do in medicine.

The information contained in the PDMP can be used to promote open and respectful conversations between the HCP and the patient about the treatment plan. Changes in the treatment plan may be indicated, such as referral to treatment or taper from opioids.

A major safety benefit to using the PDMP to improve the quality of care is the verification of medications and feedback on the participant’s own prescribing patterns. One participant described these advantages.

… the safety piece I think is nice. … verifying and double checking that patients are being safe with their controlled substance pills from the pharmacy… to look back for historically what you did. That provider report they give you … compares you to your counterparts … and [you can] see how you fit with other similar providers.

This transparency was not seen as punitive by most participants, rather this resource provided information that assisted in keeping patients safe and created accountability that may influence prescribing behaviors.

Existing PDMP Limitations

Several limitations in user interface and functional features of the PDMP were identified. User interface is defined as how the end-user (e.g. the HCP) interacts with the system (e.g. the PDMP).^17,18 Functional features are the designs that capitalize on the unique capabilities of people and systems and strengthen the behavior of the system.¹⁹ User interface and functional features were the lens used to explain the barriers affecting the use of the PDMP.

User Interface

Three user interface issues were identified. User interface difficulties were related to the need for multiple websites, direct messaging capabilities, and graphic displays.

Multiple Websites.

Accessing PDMP from various states is time-consuming due to repetitive and variable input processes including login. “Every click is sacred,” said a participant when referring to the increased time in accessing. One participant stated, “Right now, there’s so many steps to access P[D]MP, people los[e] interest or run out of time.” Another stated the program was not helpful for patients out of state. Recommendations from participants were to have the states linked so a separate login was not necessary, or to have one PDMP nation-wide because patients move states back and forth in the year and need prescriptions filled.

One participant stated bordering a state without a statewide PDMP was a problem in quickly obtaining opioid prescription information from that state. As policy changes occur in that state, HCPs and pharmacists may be able to use this resource to make decisions in prescribing and treating patients in ways that will allow reductions in opioid misuse. This would alleviate the problem this participant was having.

Direct Messaging.

The PDMP has a feature for direct messaging. However, this feature has low utilization and does not provide timely communication. One participant waited four months to hear back from a pain management physician.

Graphics Display.

Prescribers found the graphic displays busy, and the crowded layout made information difficult to find. One participant described the layout of the display was like a scoreboard at a sporting event. Additionally, a proprietary scoring system summarized risk for opioid overdose which was confusing with participants not knowing how to interpret the score and concerns about how it was derived.

Functional Features

Participants found three functional features burdensome. Challenging functional features included delay of data entry, accuracy of patient names, and clinicians with limited to no access.

Delay of Data Entry.

While data entry and reporting from pharmacies has improved over the years, it does not provide real-time content. One participant wanted real-time data and used pseudoephedrine sales as an example of successful real-time data. Participants explained that a delay in data entry can result in incomplete data impeding safe and effective decision-making.

Accuracy of Patient Name.

When prescribers or their delegates access a patient’s name, the accuracy of the name is important for obtaining PDMP information. If there is a hyphenated name or a name with a period in it, the program displays “no match.” Additionally, when patients use different names or nicknames, the problem of “no match” continues. One participant explained their experience and offered a solution.

I have a few patients who have five or six different accounts just because of name variations. Sometimes they use their legal name, sometimes they use their nickname in the P[D]MP. And I’ve never really understood how that is, because I just don’t know why it’s not their legal name.

Accuracy of the patient’s name is vital to reducing delays and obtaining accurate information. This could be resolved by only entering their legal name to ensure accurate patient names and fulfill the integrity of these programs.

Clinicians With Missing Functionalities or Limited Access.

Pharmacists were missing important functionalities for using the PDMP and medical residents had limited access reported in this study. These disciplines are vital to the care of patients who have risk and they need access and checks to the resource to reduce opioid misuse or dangerous drug interactions.

Pharmacists.

Pharmacists have independent access but cannot convey that the PDMP has been checked, therefore cannot satisfy the state regulation that the resource is checked for all prescriptions. Some physicians reported relying on pharmacists to check the resource on their behalf as part of a team-based clinic.

Medical Residents.

The participants who teach medical residents reported that medical residents did not have access to the PDMP. The medical resident needed to ask the pharmacist or another delegate to obtain information from this resource. Having no access prohibits medical residents from learning about its use and applying medical decision-making from the information acquired.

Integrating the PDMP Into the EHR

The participants stated they anticipated partial or total mitigation of some of the PDMP limitations when integrated into the EHR. Integration means the program will be embedded into the EHR allowing for one login to the EHR without requiring a separate login and password for access. This direct connection between EHR and PDMP will improve the ease of access, accuracy of patient’s name, and documentation in the EHR that this resource was checked. The participants were very positive and could not foresee a downside to the integration. One participant said, “…logging in and typing their stuff in is a hassle, but I’m not sure there’s a better way other than if the app [integration] automatically puts in that data …” The word “auto-populate” was used by participants to describe how nice it would be to not type in patient information with every access. Participants proposed that auto-populate may free up time in the appointment for providing care or valuable patient education. One participant suggested, “If you had sufficient time in clinics to check P[D]MP, use it as an educational [program], I think that as a profession, health care providers could drastically impact the opioid epidemic by not prescribing inappropriately.”

Improvement Recommendations to Increase Use

Participants had many suggestions to improve the conditions of the PDMP that create barriers to use. These suggestions are thought to increase use and are summarized in Table 2 entitled Recommendations to Identified Problems. This table is organized by area needing improvement, the specific problems, and the solution. The creation of this table demonstrates the desire of healthcare providers and pharmacists to improve PDMP capabilities and can be used to guide future versions.

Discussion

Key findings were determined from the experiences of 15 prescribers and pharmacists and identified a range of positive and negative perspectives on using the PDMP in clinical practice. HCPs generally found the resource to be a useful support for practice. Prescribers and pharmacists used the PDMP for information about opioid doses, information on patients not known to them, and to communicate and make decisions about the potential need for treatment. Improving treatment decisions and prescribing practices were identified as goals for accessing the PDMP.^20,21 Prior literature demonstrated healthcare providers and pharmacists felt this resource was extremely helpful in identifying “doctor shopping,” misuse of opioids and identifying those at risk for opioid overdose.^4,22

There were also multiple shortcomings and barriers to optimal use. A commonly voiced problem was the password requirement and entry of patient details to access the PDMP information. Participants also noted time-consuming access to multiple websites when needing to access the programs for multiple states. In a systematic review by Martin and others²³ found when passwords need resets frequently, the login process is complicated, and multiple websites need log-in so these were access problems. Other studies also noted time-consuming processes to login, difficulty accessing information, and these issues created significant barriers to use.^5,8,21,24,25

Our study found that displays and graphics made information difficult to find. Yet another study by Weiner and others²⁶ found enhanced displays and graphics helpful in allowing physicians to correctly identify patients with multiple providers, overlapping opioid/benzodiazepine or overlapping opioid prescriptions, high daily opioid dosages, and multiple pharmacies. However, the enhanced graphics did not decrease data interpretation time, nor did it change prescribing behavior for obvious cases of aberrant drug-related behavior (e.g., multiple prescribers) or obvious need for opioids (e.g., fracture).²⁶

Direct messaging was not addressed in other studies but the need for improved communication of health information was described.²⁵ To improve this interface, the difficulties and delays of direct messaging will require further study to determine what would improve usability and acceptability.

Our participants drew upon their own experiences and perceptions when they commented on the functional features of the PDMP. They revealed time lags in data entry of prescriptions, barriers when pharmacists were designated as delegates, and difficulties when medical residents did not have easy access or any access. The findings of pharmacists not being able to denote they had checked the resource for a prescription and medical residents without access were unique to our study and could be a result of a study in an academic care center. However, the time lags were comparable to those found in other studies. In our study, prescribers relied on timely data entry to facilitate prescribing decisions. In a systematic review by Martin and others,²³ real-time data updates were a measure of quality information and determined how quickly the prescriber had access to PDMP data. Lack of timely data was identified as a cause of prescribing and dispensing decisions delays at the point of care.^21,23,27

Some of these findings herald a need for change. As EHR systems such as Cerner, EPIC, and Allscripts integrate PDMP for their customers, it is important to encourage the end-user (e.g., prescribers and pharmacists) to become involved with the optimization and subsequent adoption so their needs and expectations are heard. This will enhance successful outcomes, as studies show successful integration significantly increases access to this data, improves satisfaction in use, and potentially guides safe and effective decision-making in opioid prescribing.^20,28,29

There appeared to be an overall lack of training on the PDMP system received by the disciplines who prescribe and dispense.^10,22,23,30 More training may facilitate an increase in use. A possible solution is deliberate educational outreach to produce awareness of this resource, how to register for access, strategies for access, interpretation of the data, communicating with patients on results of their report, how the report can be used to teach prescribers and improve prescribing behaviors, how to minimize bias in interpretation, and improve knowledge about updates to the system.

Limitations of the Study

These findings were exploratory to discover PDMP facilitators to quality care and gaps in care that changes in the program could rectify. Other providers, health care systems, and geographic areas may find different findings or have valuable experiences to add to understanding how to increase PDMPs use in practice. This study occurred prior to integrating the state PDMP into the EHR. Some issues may resolve when the health system moves to integrate the PDMP into the EHR. However, other issues may persist related to accessing this resource from neighboring states, use of patient’s legal name in the program records, disciplines with unique accessing challenges.

Conclusion

The PDMP is viewed as a useful program in making strides to reduce the impact of inappropriate opioid prescribing in our country. As each state works to increase the use of this resource to improve safety and efficacy of prescribing opioids, a critical piece when adapting the PDMP is to engage the prescribers and pharmacists and understanding their perspectives. This manuscript describes benefits and barriers to use of the PDMP and offers solutions to improving the interface and function of the program and increase use.

Table 1.

Recommendations to identified problems

Area	Problem	Solution
Access	Separate website, username, password needed to login to PDMP.	Integrate PDMP into EHR for only one login for both.
	Not all state PDMP are integrated into EHR.	Include all states in integrated EHR, or at least neighboring states.
	Patient with multiple names and PDMP profiles.	Create smooth ability to merge accounts to link to a single legal name.
	Access PDMP only at certain times when there is a healthcare encounter.	Access PDMP in EHR anytime, not just when there is an encounter.
Documentation	Not every PDMP access resulted in documentation of the access for a specific prescription.	Require documentation that PDMP was accessed for a given prescription more universally. This includes by pharmacists acting as delegates. Run through this to ensure pharmacists can do this given they have their own NPI.
	It is unclear if a patient has been or should be referred to substance use disorder treatment.	The PDMP could provide information that a patient was referred to treatment and the date.
	An opioid use disorder diagnosis is listed in EHR but not in the PDMP.	Relevant diagnoses should be included in the PDMP.
	Providers in other health systems may not have access to relevant clinical notes related to controlled substance use.	Relevant clinical notes from EHR could be copied into PDMP so other prescribers have access.
Safety alerts	Co-prescribing of opioids and benzodiazepines may not be obvious.	Create robust alerts that acknowledge risk of drug interactions.
Resources	Evidence-based practice (EBP) guidelines and assessment instruments not easily accessible.	Integrate a health system’s evidence based prescribing guidelines and psychometrically sound opioid use disorder assessment instrument into the PDMP. A prompt for EBP could occur when prescribing history indicates multiple prescribers, multiple pharmacy use, high doses of opioids, and/or a combination of opioids and benzodiazepine prescriptions.