The Journal of Primary Prevention

, Volume 36, Issue 2, pp 131–137 | Cite as

The Effects of North Carolina’s Prescription Drug Monitoring Program on the Prescribing Behaviors of the State’s Providers

  • Chris Ringwalt
  • Mariana Garrettson
  • Apostolos Alexandridis
Report from the field

Abstract

State-level prescription drug monitoring programs (PDMPs) show promise as a key strategy to respond to the epidemic of the misuse and abuse of controlled substances (CS), particularly opioid analgesics, in the United States. Undocumented concerns have been expressed that these PDMPs may have a “chilling effect” on providers’ willingness to prescribe these substances to their patients. Using data from North Carolina’s PDMP for the 3-year period from 2009 through 2011, we examined whether rapid increases in (1) the number of providers who queried the system, and (2) the number of days on which they queried it, would be related to their prescribing practices in regards to CS. We hypothesized that neither marker of PDMP utilization would be associated with a decrease in either patients receiving CS prescriptions or CS prescriptions filled. We found no association between either of these variables and the number of patients who filled prescriptions for CS or the number of prescriptions for CS filled. However, we did find a slight positive relationship between the growth in the utilization of the PDMP and the number of prescriptions filled for opioid analgesics. Concerns that PDMPs may constrain prescribing behavior with regards to CS are not supported.

Keywords

Prescription drug monitoring program Controlled substances Pain management Physician prescribing behavior 

Introduction

The misuse, abuse, and diversion of controlled substances (CS) in the United States constitute a public health epidemic [Centers for Disease Control (CDC), 2011; Hernandez & Nelson, 2010]. Deaths attributable to drug overdoses have increased by a factor of three since 1990 (CDC, 2013). In 2008, opioid analgesics accounted for three out of four overdoses attributable to prescription drugs (CDC, 2011). In 2010, the quantity of all prescription analgesics sold in the United States exceeded that of 1999 by a factor of four (Garcia, 2013). In the same year, 12 million people, or 2.7 % of the nation’s population 12 or over, reported that they were current nonmedical users of prescription drugs (Substance Abuse and Mental Health Services Administration, 2011).

As of fall 2013, every state but Missouri has passed legislation to establish a prescription drug monitoring program (PDMP), and 42 have fully operational programs. PDMPs are designed to collect data on prescription drugs as they are dispensed and to make these data available for various uses, including patient care, public health surveillance, and law enforcement investigations relating to drug diversion and insurance fraud. The data generally collected by PDMPs contain information specific to the patient: the name, strength, quantity, and duration of substances prescribed; the provider; and the dispenser (Perrone & Nelson, 2012). North Carolina’s PDMP resembles those of other states insofar as it comprises an electronic data depository that is designed to help providers and dispensers reduce inappropriate prescribing and dispensing of CS, by requiring all community-based pharmacies to report information concerning dispensed CS into a central database. Once they register with the PDMP, providers and dispensers are encouraged to access and query the system to download a history of CS dispensed to their active patients. As specified in the act establishing the program, the legislative mandate of the NC PDMP is clear: improve the State’s ability to identify CS misusers or abusers; refer them for treatment; and identify and stop the illegal use of prescription drugs in an efficient and cost effective manner, without impeding the appropriate medical utilization of licit CS (North Carolina Legislation). In addition, the law specifies that a provider who manifests unusual prescribing patterns may be referred to the State Attorney General’s office, which may (at its discretion) refer the provider to the State Bureau of Investigation. Further, the reporting law excludes hospital and long-term care pharmacies that dispense CS for the purpose of administration in an in-patient setting. Results of several (Reifler et al., 2012; Reisman, Shenoy, Atherly, & Flowers, 2009; Simeone & Holland, 2006; Worley, 2012), but not all (Paulozzi, Kilbourne, & Desai, 2011) studies of the effects of PDMPs suggest that these programs have at least a moderately positive effect on reductions in the possible diversion of CS.

The inclusion of the clause to protect appropriate medical utilization of CS emphasizes the extent to which there was significant concern among those who established the program in NC in 2005 of its potential negative effect on medical care. Because 44 of the states’ PDMPs now sanction access to their databases by licensing and regulatory boards (National Alliance for Model State Drug Laws, 2013), concerns have been expressed that the fear of intrusive regulatory oversight may have a “chilling effect” [Brennan, Carr, & Cousins, 2007; Garcia, 2013; Institute of Medicine (IOM), 2011] if it induces providers to reduce prescriptions for CS to levels that fail to meet their patients’ needs (Perrone & Nelson, 2012). To avoid a potentially damaging investigation, the literature expresses an undocumented concern that some physicians may “play it safe” by decreasing the strength of the medications they prescribe, switching to alternative medications that may be less appropriate, or either purging their practices of patients suffering from chronic pain or declining to accept new pain patients (Fishman et al., 2004; Joranson et al., 2002; Wang & Christo, 2009). Regardless, there is genuine fear that the result will be to curtail legitimate practice (Simoni-Wastila, 2011) and thus to fail to treat pain adequately, which is increasingly being seen as unethical practice (Brennan et al., 2007). Indeed, the failure to treat pain adequately may be potentially as destructive as over-treatment (IOM, 2011), and the war on drugs may be seen as directly conflicting with the war on pain (Fishman et al., 2004).

To date, outside of surveys of providers’ and dispensers’ attitudes towards PDMPs (Fass & Hardigan, 2011; Feldman, Williams, Coates, & Knox, 2011; Turk, Brody, & Okifuji, 1994), there has been little empirical research into the potential of these registries to have a chilling effect on prescribing practices (Simoni-Wastila, 2011). In one study of prescriptions in the Automation and Reports and Consolidated Orders System (ARCOS), Twillman (2006) found that PDMPs have resulted in a reduction in Schedule II analgesics that was accompanied by an increase in Schedule III prescriptions for analgesics. He concluded that the registries may thus be inducing providers to substitute higher- for lower-scheduled opioid analgesics. However, a subsequent study by Reisman et al. (2009) failed to confirm this finding. Questions concerning the potential suppressant effect of PDMPs on patients’ legitimate access to analgesics for treatment of pain remain unresolved (Garcia, 2013).

The purpose of this study is to examine three consecutive years of data from North Carolina’s PDMP to determine if increased usage of the system by the State’s prescribers is associated with a decline in prescriptions either for CS, in general, or opioid analgesics, in particular. We hypothesized that, over time, the total number of patients filling prescriptions for CS, and the total number of prescriptions they filled for CS in general, and opioid analgesics in particular, would not be associated with either the number of providers who queried the PDMP, or the number of days on which they queried it.

Methods

Data

We obtained 3 years of data pertaining to North Carolina’s PDMP for the years 2009–2011 inclusive, from Health Information Designs, which serves as a data repository and manager for several states’ programs. The dataset required considerable manipulation and cleaning before it could be converted into an analysis file. Cleaning the data included removal of all non-CS that had been mistakenly reported by participating pharmacies. We then divided the 3 years of PDMP data into six 6-month blocks, as a means to determine trends over time in our outcomes of interest. Since our dataset included no identifying information, our institutional review board declared it exempt from review.

Independent Variables

For each 6-month block, we calculated the total number of providers who used or queried the PDMP, and the mean number of days on which those providers queried the system as a function of the number of days in the period between the prescriber’s first and last day of access that block. Thus, if a provider had accessed the PDMP for 10 days in the first 6 months of a given year, over a period beginning on February 1 and ending on June 20 (the provider’s first and last days of access within the 6-month period), the provider’s utilization ratio would have been 0.07. This figure represents the ratio of 10 (the number of days in the 6-month period in which the provider accessed the system) to 140 (the total number of days between February 1 and June 20, inclusive). We developed this measure to control for the likelihood that many providers would enter and exit the PDMP pool throughout the 3-year period under study, as they initially registered to access the system or moved in or out of the state.

Dependent Variables

Our three dependent variables were the total number, in each 6-month block, of: (1) patients who filled prescriptions for any CS, (2) prescriptions for any CS they filled, and (3) prescriptions for opioid analgesics they filled. We defined “opioid analgesics” as full mu-opioid receptor agonists used for pain management, of which there were 67 discrete drug names in the database we examined, the most salient of which were fentanyl, oxycodone, morphine, hydromorphone, oxymorphone, tapentadol, and solid oral methadone formulations.

Analysis

For purposes of analyses we divided each of the three study years (2009–2011 inclusive) into two discrete spring and fall 6-month blocks. We analyzed the number of providers as a proportion of all licensed providers in the State (National Center for Health Statistics), and the number of prescriptions for CS as a proportion of the total population of the State (North Carolina Health Professions Data System). We independently assessed the time trend of all variables using ANOVAs, and tested the association between each independent and dependent variable using linear regression using Stata 13.1 (StataCorp, College Station, TX).

Results

Figure 1 displays trends in the independent variables over the 3 study years, in six discrete (i.e., non-cumulative) 6-month intervals. As this figure indicates, over this period the proportion of providers querying the PDMP increased by a factor of 2.0, and the mean number days for which they queried the program increased by a factor of 2.4. In contrast, the trends in the dependent variables over the 3-year study period shown in Fig. 2 were generally more stable. The incidence of prescriptions for all CS and patients filling prescriptions for CS remained fairly constant over the six temporal study blocks. In contrast, there was a slight increase in the incidence of prescriptions for opioids.
Fig. 1

Use of prescription drug monitoring program (PDMP). Note “S” indicates spring (January–June) and “F” indicates fall (July–December)

Fig. 2

Total proportions of CS recipients and count of CS prescriptions filled and opioid prescriptions filled. Note “S” indicates spring (January–June) and “F” indicates fall (July–December)

Linear regressions (results not shown) of each of the independent and dependent variables confirmed that neither PDMP registration nor use was associated with a significant decline in the incidence of either CS prescriptions in general or opioid prescriptions in particular.

Discussion

We found no evidence to support the continuing concern that the use of North Carolina’s PDMP might have a suppressing or “chilling” effect on providers’ willingness to prescribe CS. As hypothesized, we found no association over the period 2009–2011 between either the number of providers who accessed North Carolina’s PDMP or the frequency with which they accessed the database, and the number of patients who filled prescriptions for CS. Our findings thus contribute to the scant empirical literature that suggests that concerns about the potentially chilling effect of PDMPs on prescriptions for CS are unwarranted (Garcia, 2013). However, we were surprised to find a strong positive relationship between the growth in the utilization of the PDMP and the opioid analgesic prescriptions filled over this period.

We can think of three explanations for this association. First, despite the high rate of growth of providers’ utilization of the PDMP over this period, providers consulting the system may have had little reason to suspect their patients of drug-seeking behavior. Second, in 2011, at the end of the study period, the total proportion of providers registered with the program constituted only 27 % of all the State’s providers who were authorized to dispense CS (William Bronson, personal communication, September 3, 2013). Any decline in prescribing behaviors among the providers who consulted the PDMP may have been masked by increases in these behaviors by providers who had not registered with the system, and were thus unable to query it. Third, utilizing the PDMP may have increased providers’ confidence that they could reliably track their patients’ full prescription histories over time by monitoring the registry.

We note the steady increase in filled prescriptions for opioid analgesics, and stable trend in all CS prescriptions. This finding mitigates the potential concern that providers are turning to non-opioid analgesics to meet their patients’ pain needs.

The results of our study provide compelling support for the argument that the presence of North Carolina’s PDMP has not impeded overall prescribing of CS for medical care. However, we suspect that our study’s results will not eliminate all skepticism concerning the potential for PDMPs to inappropriately suppress prescribing, especially of opioid analgesics. We suggest that studies of PDMPs should continue to be conducted to determine their effects on prescribing behaviors as greater proportions of providers register with and query their states’ PDMPs. It would also be helpful to disaggregate overall rates of filled prescriptions for CS by the registration status of the providers who wrote these scripts. Future studies should also examine differences in trends, for registered versus unregistered providers, in the strength and duration of scripts for CS in general, and opioid analgesics in particular.

Our study’s findings do not suggest that the use of North Carolina’s PDMP is failing to reduce deaths related to the misuse and abuse of prescription opioids. Indeed, the State’s unintentional deaths attributable to prescription opioids plateaued starting in 2008, following a steady rise since 1999 (Poisoning Overdose Prevention, 2013). Providers’ registration with the program during the period studied was relatively low (27 %), and the modest increases noted in filled opioid prescriptions could well be attributed to providers who were not registered with, or did not consult, the PDMP. Further, our data do not speak to any trends in either the strength or duration of opioids prescribed; future studies should examine these outcomes as a function of providers’ utilization of the PDMP.

Limitations

We acknowledge several limitations to this study. First, our findings should be extrapolated to other states only with caution; the laws and regulations that govern the administration and use of the PDMP in each state differ considerably (NAMSDL, 2013a, 2013b), and are likely to differentially affect trends in providers’ prescribing behaviors. Second, not all pharmacies in the State report into its PDMP; the law establishing the registry exempted several types of pharmacies from reporting data to the State’s PDMP, including those associated with hospitals and with long-term care facilities. Further, during the study period pharmacies associated with the Veterans Administration and with the State’s military bases also did not report CS filled by their patients. Third, our selection of 6-month blocks limited our data points to two per study year, or six in total. While we could have increased the number of data points by selecting smaller units, we believe that the utility of our calculations of the mean number of days in which providers queried the system, as a function of the number of days in the period between their first and last day of access in any particular block, would have been reduced. Fourth, as noted, only slightly more than a quarter of the State’s providers were registered with the PDMP during the study period. We would have had greater confidence in our findings if registration levels were higher. Fifth, because our study reports findings related to providers’ population, but not individual-level behaviors, it does not speak to changes in individual providers’ patterns of prescriptions following PDMP consultation. We note, however, that the PDMP has the capacity to support an investigation of this nature, which may be an appropriate topic for a future study. Indeed, to the extent that concerns about the potential chilling effects of PDMP persist, however poorly documented they may be, more studies are needed.

Conclusions

Although the use of PDMPs as a tool in the effort to reduce the epidemic of prescription drug overdose is widely encouraged, the evidentiary body describing the positive and negative effects of such programs is still emerging. This study adds to the current literature in its finding that a potential negative unintended consequence of implementing a PDMP, namely the suppression of overall prescribing rates, does not appear to be occurring at the population level. We believe, however, that individual providers’ utilization per se of the database is unlikely to be a determining factor in any eventual declines we may find in prescription rates; what may drive these declines could be the fear of investigation, rather than particular patients’ use profiles. It is also possible that as information in the lay press becomes more salient regarding the risks of addiction and fatal overdose associated with prescription opioid use, patients may begin to request non-opioid analgesics. We fully agree with the IOM (2011) that clinicians have an ethical responsibility to treat pain, and that it would be highly unfortunate if reductions in prescribing rates for these analgesics were ever found to adversely affect patients in need of relief from pain. If used responsibly, the PDMP constitutes a powerful clinical tool to prevent the misuse and abuse of CS.

Notes

Conflict of interest

The authors have no conflict of interest.

References

  1. Brennan, F., Carr, D. B., & Cousins, M. (2007). Pain management: A fundamental human right. Anesthesia and Analgesia, 105, 205–221.CrossRefPubMedGoogle Scholar
  2. CDC. (2013). Policy impact: Prescription painkiller overdoses. Retrieved from http://www.cdc.gov/homeandrecreationalsafety/rxbrief/
  3. Centers for Disease Control (CDC). (2011). Vital signs: Overdoses of prescription opioid pain relieversUnited States, 19992008 [Vital Signs No. 60(43)] (pp. 1487–1492). Atlanta, GA: Centers for Disease Control. Retrieved from http://www.cdc.gov/mmwr/preview/mmwrhtml/mm6043a4.htm
  4. Fass, J. A., & Hardigan, P. C. (2011). Attitudes of Florida pharmacists toward implementing a state prescription drug monitoring program for controlled substances. Journal of Managed Care Pharmacy, 17(6), 430–438.PubMedGoogle Scholar
  5. Feldman, L., Williams, K. S., Coates, J., & Knox, M. (2011). Awareness and utilization of a prescription monitoring program among physicians. Journal of Pain and Palliative Care Pharmacotherapy, 25(4), 313–317.CrossRefPubMedGoogle Scholar
  6. Fishman, S. M., Papazian, J. S., Gonzalez, S., Riches, P. S., & Gilson, A. (2004). Regulating opioid prescribing through prescription monitoring programs: Balancing drug diversion and treatment of pain. Pain Medicine, 5(3), 309–324.CrossRefPubMedGoogle Scholar
  7. Garcia, A. M. (2013). State laws regulating prescribing of controlled substances: Balancing the public health problems of chronic pain and prescription painkiller abuse and overdose. The Journal of Law, Medicine & Ethics, 41(s1), 42–45.Google Scholar
  8. Hernandez, S. H., & Nelson, L. S. (2010). Prescription drug abuse: Insight into the epidemic. Clinical Pharmacology and Therapeutics, 88(3), 307–317.CrossRefPubMedGoogle Scholar
  9. Institute of Medicine (IOM). (2011). Relieving pain in America: A blueprint for transforming prevention, care, education, and research. Washington, DC: The National Academies Press.Google Scholar
  10. Joranson, D. E., Carrow, G. M., Ryan, K. M., Schaefer, L., Gilson, A. M., Good, P.,… & Dahl, J. L. (2002). Pain management and prescription monitoring. Journal of Pain and Symptom Management, 23(3), 231–238.Google Scholar
  11. National Alliance for Model State Drug Laws (NAMSDL). (March, 2013a). Prescription drug abuse, addiction, and diversion: Overview of state legislative and policy initiatives. Part 1: State prescription drug monitoring programs (PMPS). http://www.iowa.gov/odcp/docs/Rx%20Mon%20Programs%20NAMSDL%20March%202013.pdf
  12. NAMSDL. (July, 2013b). Compilation of state prescription monitoring program maps. http://www.namsdl.org/library/13D46B1B-1372-636C-DD8A80A2928024DF/
  13. National Center for Health Statistics. Bridged-race population estimates 19902012. Public-use data file, CDC WONDER online database. http://wonder.cdc.gov/
  14. North Carolina Health Professions Data System. Annual profiles: 20092011 health professions tables. Cecil G. Sheps Center for Health Services Research, The University of North Carolina at Chapel Hill. http://www.shepscenter.unc.edu/hp/index.html
  15. Paulozzi, L. J., Kilbourne, E. M., & Desai, H. A. (2011). Prescription drug monitoring programs and death rates from drug overdose. Pain Medicine, 12(5), 747–754.CrossRefPubMedGoogle Scholar
  16. Perrone, J., & Nelson, L. S. (2012). Medication reconciliation for controlled substances—An “ideal” prescription-drug monitoring program. New England Journal of Medicine, 366(25), 2341–2343.CrossRefPubMedGoogle Scholar
  17. Poisoning Overdose Factsheet. (January, 2013). North Carolina Division of Public Health: NC Injury and Violence Prevention Branch. http://injuryfreenc.ncdhhs.gov/About/PoisoningOverdoseFactSheet2013.pdf
  18. Reifler, L. M., Droz, D., Bailey, J. E., Schnoll, S. H., Fant, R., Dart, R. C., & Bucher Bartelson, B. (2012). Do prescription monitoring programs impact state trends in opioid abuse/misuse? Pain Medicine, 13(3), 434–442.CrossRefPubMedGoogle Scholar
  19. Reisman, R. M., Shenoy, P. J., Atherly, A. J., & Flowers, C. R. (2009). Prescription opioid usage and abuse relationships: An evaluation of state prescription drug monitoring program efficacy. Substance Abuse: Research and Treatment., 3, 41–51.Google Scholar
  20. Simeone, R., & Holland, L. (2006). An Evaluation of Prescription Drug Monitoring Programs. Simeone Associates, Inc., No. 2005PMBXK189, September 1, 2006. http://www.simeoneassociates.com/simeone3.pdf
  21. Simoni-Wastila, G. E. L. (2011). Prescription monitoring programs: Striking the balance between medical use and diversion. Journal of Addictions Nursing, 22(1–2), 77–82.Google Scholar
  22. Substance Abuse and Mental Health Services Administration. (2011). Results from the 2010 National Survey on Drug Use and Health: Summary of National Findings [No. HHS Publication No. (SMA) 11-4658]. Rockville, MD: Substance Abuse and Mental Health Services Administration. Retrieved from http://oas.samhsa.gov/NSDUH/2k10NSDUH/2k10Results.htm#2.16
  23. Turk, D. C., Brody, M. C., & Okifuji, E. A. (1994). Physicians’ attitudes and practices regarding the long-term prescribing of opioids for non-cancer pain. Pain, 59(2), 201–208.CrossRefPubMedGoogle Scholar
  24. Twillman, R. (2006). (324/995): Impact of prescription monitoring programs on prescription patterns and indicators of opioid abuse. The Journal of Pain, 7(4), S6.CrossRefGoogle Scholar
  25. Wang, J., & Christo, P. J. (2009). The influence of prescription monitoring programs on chronic pain management. Pain Physician, 12(3), 507–515.PubMedGoogle Scholar
  26. Worley, J. (2012). Prescription Drug Monitoring Programs, a response to doctor shopping: Purpose, effectiveness, and directions for future research”. Issues in Mental Health Nursing, 33(5), 319–328.CrossRefPubMedGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2014

Authors and Affiliations

  • Chris Ringwalt
    • 1
  • Mariana Garrettson
    • 1
  • Apostolos Alexandridis
    • 1
  1. 1.Injury Prevention Research CenterUniversity of North CarolinaChapel HillUSA

Personalised recommendations