The Impact of the Rhode Island Duplicate Prescription Law on Prescribing Practices for Schedule II Drugs

Multiple copy prescription laws have been implemented or proposed in several states to better track the distribution of drugs in Schedule II to the ultimate consumer. The states with such statutes report a 30 to 50 percent decrease in the number of Schedule II prescriptions written as well as a reduction in diversion of these drugs and forgeries. Risk factors were assessed for the effect of the Rhode Island Duplicate Prescription program on prescribers. A questionnaire was mailed to all prescribers (N=3016) registered with the Rhode Island Department of Health Division of Drug Control to prescribe Schedule I I drugs. The response rate was approximately 22%. The response group was evaluated for demographics, prescribing history, perception of the impact of the law on prescribing, and knowledge of the law. Associations were determined between prescriber characteristics and two outcomes (effect of ·the Rhode Island duplicate prescription form on decision-making and therapeutic preference to choose an alternative drug to a clearly indicated Schedule II drug) . Both the preand post-law groups were evaluated for the outcomes. Odds ratios were calculated for variables significant at the 0.15 level. Risk factors which explain some of the variation of the outcomes are age, sex, primary professional degree, specialty practice, practice type, number of years licensed, issuance of Schedule II prescriptions, and knowledge of Division of Drug


ACKNOWLEDGMENTS
There are many people to whom I owe much gratitude as the end of my journey as a graduate student in Pharmacy Practice comes to a close, for without their support, encouragement and creative motivation this project would be only an idea. Thank you to Dr. Jeffrey Jarrett whose vigor and passion for SAS inspired me.

A. Historical Perspectives of Drug Control
The United States has recognized the significance of drug abuse, illegal drug distribution and the impact of these on society at large. In terms of historical perspective, the first one hundred years of Federal law dealt primarily with quality control (Strauss and Sherman 1985). According to a 1985 National Household Survey on Drug Abuse by the National Institute on Drug Abuse (NIDA 1985), that of the estimated 15.7% of the United States population over age 12 who have reported using psychotherapeutic drugs, approximately 31% of these respondents have admitted to using these drugs for non-medical purposes. According to the 1985 Drug Abuse Warning Network (DAWN) statistics, legal controlled substances were involved in 53.5% of all drug-related emergency room visits and 49.6% of drug-related deaths (NIDA 1986). In the 1986 list of the top DAWN emergency room statistics, nearly 20% of those drugs listed were Schedule II substances (NIDA 1986). The Drug Enforcement Administration statistics reveal that 80 to 90 percent of drug diversion for non-medical use is at the practitioner and pharmacy levels (U.S. Department of Justice 1987). The concern over Schedule II drugs is that categorically these drugs have a high prevalence for non-medical use, and they result in large health consequences when used non-medically (AMA Department of Substance Abuse 1988) .

s. Multiple Copy Prescription Programs in the United States
To provide greater control in the area of distribution of schedule II drugs to the ultimate user, the multiple copy prescription was devised. According to DEA, the multiple copy prescription creates a closed

C. The Rhode Island Duplicate Prescription Law
In February 1979, Rhode Island became the sixth state to institute a multiple copy prescription program by statute. As defined in Title 21, Chapter 28, Section 3.18(d) of the Rhode Island Uniform Controlled Substances Act, Prescriptions for controlled substances in schedule II shall be filed separately and shall not be refilled. The form of record for prescription slips for controlled substances in schedule II shall consist of two (2) parts, an original and a duplicate which are required to be presented to the pharmacy by the ultimate user or his representative.
Pharmacies dispensing controlled substances in schedule II are required to deliver to the division of drug control all duplicate copies of such prescriptions on or before the fifth day of the month following the date of dispensing. The prescription slip shall be a form provided by the director of health.  . In an overview article about states with multiple copy prescription programs, the authors acknowledge the reduction of the prescribing of Schedule II drugs and offer several anecdotal reasons such as prescriber education or the utilization of Schedule III, IV or V drugs for the sake of convenience (Strauss and Bracelin 1984). Some manufacturers of Schedule II medications argue that multiple copy prescription programs are costly for the number of diversion cases convicted, patient confidentiality is compromised, the forms have street value thereby endangering the prescriber and his staff, and prescribers may alter their practices to avoid scrutiny by the law enforcement officials (Konnor, 1983).
A review of the literature reveals several theories for predicting physicians' prescribing Avorn 1987, Boreham 1989), the interaction among criteria when prescribing (Zelnio 1982) and means to improve physicians' decision making Avorn 1990, Peterson and. These theories do not provide an understanding of how a law such as the multiple copy prescription law interacts or affects prescribing practices. All responses were reviewed and hand-coded for specialty practice.

DATA ANALYSIS
The data collected from the questionnaire were analyzed using the Statistical Analysis System (SAS) Version 6.06 on the IBM mainframe computer at the University of Rhode Island.

(I exposed non-cases)*(# unexposed cases)
A 95% confidence interval (91% CI) was calculated on the odds ratio using the following formula: The adjusted risk odds ratios, ROR (adj.), for all independent variables in the final model as well as the model was calculated using the following formula : The ninety-five percent confidence intervals for the estimates of relative risk were calculated using the following formula:

Prescribinq of Schedule II Druqs
Regarding the total sample, 88.5% have prescribed drugs in Schedule II for their ambulatory, non-hospitalized patients.
Seventy percent of the sample also revealed that they prescribed between one and twenty-five Schedule II prescriptions per month, and analgesic narcotics in that schedule were prescribed by 76.6% of the group. (Table 3) . inconvenient the form appears to be for the prescriber, shows a skew to the right. (Figure 1).   1 column totals may not add up to 100% due to lack of response.

Perceptions of Prescribers' Willingness to Prescribe Agents in Schedule II Which are the Primary Drug(s) of Choice
Prescribers were asked several questions designed to measure their willingness and ability to prescribe Schedule II drugs. Nearly one-third (32.8%) of the respondents stated that they preferred to prescribe drugs other than Schedule II medications (i.e. Schedule I I I, IV or V or non-scheduled legend drugs) in situations where a Schedule II drug is clearly indicated. The availability of alternative medications to Schedule II drugs was a reason for not prescribing Schedule II drugs for 71% of the total group. The concern over malpractice litigation resulting from the use of an alternative drug to a Schedule II when the Schedule II was clearly indicated is apparent in 18.2% of the target group.
Although 51% and 54% of the study group agreed that the use of a scheduled alternative or a non-scheduled alternative respectively may have adverse consequences for the patient, 56.6% of the respondents believe that there is not less risk for the patient in using an alternative drug to the Schedule II (X 2 =7.02, p=0.01). Nearly 62% of the study group disagreed that there was better patient compliance in using an alternative drug to a Schedule II, and a significant relationship between that response and spatial time with respect to the passage of the Law (X 2 =3.34, p = 0.07).  Odds ratios and 95% confidence intervals were calculated.

The Bivariate Relationship Between the Independent Variables and If the Rhode Island Duplicate Prescription Form Affects Decision-making for the Study Group
The summary of the relationship between whether the Rhode Island duplicate prescription form affects the decision-making process in the creation of a therapeutic regimen and the independent variables for the entire study group is found in times as likely to agree that the form affects one's decision (95% CI 1.25, 3.29).
The risk that having a specialty practice was associated with agreeing that the form affects one's decision was 2.16 (95% CI 1.26, 3.70).
Two variables, sex and number of prescriptions issued per month, had odds ratios of less than 1 (OR=0.47 and 0.68 respectively).

Bivariate Association Between Independent Variables and If the Rhode Island Duplicate Prescription Form Affects Decision-making for the Pre-1979 Group
The analysis for the group of respondents who stated that they were in practice in Rhode Island prior to 1979 and the association of independent variables and if the Schedule II prescription form affects decision-making is presented in Table 9.
The bivariate analysis reveals that there are three variables which have cells too small to count, i.e., there were five or less respondents in a particular cell. These variables were the number of years licensed, presently practicing in Rhode Island and age. Of the remaining variables, the review variable had a Chi-square of 3.03 and was significant (p=0 . 082). The odds ratio associated with this variable was 0.61.

Bivariate Association Between Independent Variables and If the Rhode Island Duplicate Prescription Form Affects Decision-making for the Post-1979 Group.
The analysis for bivariate association for the post-1979 group revealed five statistically significant variables.
They were degree, specialty, number of years licensed, sex and number of schedule II prescriptions written per month .
The range of the Chi-squares for the independent variables was 0.00 (practice type) to 9.07 (specialty). The odds ratios for the significant variables ranged from 0.42 (sex) to 2.99 (specialty). The results are summarized in Table 10. in Rhode Island prior to 1979 (X 2 =0.00, p=0.994). Table 11 summarizes these statistics. The summary of the association between the independent variables and therapeutic preference for the pre-1979 group is found in Table 12. One variable had cells too small to count for Chi-square statistics (number of years licensed) .
The variable, issuance of a Schedule II prescription, was the only statistically significant variable (X 2 =3.41, p=0.065).
The remaining variables had Chi-squares ranging from 0.06 (age) to 1.74 (number of prescriptions issued per month).    The regression models were tested for collinearity and for interactions between two or more variables. Models were also designed for both dependent variables to determine the association between the independent variables for the preand post-1979 groups.

Collinearity
The PROC REG procedure with C04LIN option was invoked to assess collinearity problems. A summary of the collinearity diagnostics is found in Appendix D. A collinearity value of greater than 0.5 was used as an indicator of possible problems. All collinearities greater than 0.5 involved interaction terms.

Interactions
The models were assessed for the interactions between two or more of the independent variables. The interaction terms were evaluated in the models using the PROC LOGIST with STEPWISE option. No interaction terms were found to be significant in any of the models.

Evaluation of Interim Models
The interim models were evaluated for model Chi-squares            Table 27 summarizes these results.  group than the pre-1979 group work primarily in the hospital environment. As should be noted, many respondents who claimed to work in "other" sites extemporaneously stated to work in a health maintenance organization (HMO) .
More than 98 percent of the respondents in the pre-1979 group had been licensed in any jurisdiction for more than ten years, and more than 91 percent of this group also worked in Rhode Island for more than 11 years; whereas, 74.1% of the respondents in the post-1979 group had been licensed for 10 or less years, but 94.4% of this group had practiced in Rhode Island for up to 10 years.

Prescribing of Schedule II Drugs
Most respondents (88. 5%) had prescribed Schedule II drugs at some time since initial licensure for an ambulatory, nonhospitalized patient, and this was consistent for the pre-and

Prescribing Practices for Schedule II Drugs
Although more than one-third of the overall respondents (36.8%) state that the Rhode Island Duplicate Prescription form does not affect the decision-making process in the creation of a therapeutic regimen and 23.2% state that the form has no effect on the decision-making, another one-third of the respondents stated that the form does affect the decision-making process. There appears to be an unwillingness or concern to prescribe Schedule II drugs or a perception of an obstacle to prescribe Schedule II drugs for the 34.2% who state that the form does affect decision-making. The effect of the form on decision-making could be on several levels such as the unwillingness to prescribe a Schedule II drug or the unwillingness to prescribe certain quantities of drug, an awareness of the potential side effects and adverse consequences for the patient in utilizing a Schedule II drug or for the need to more closely monitor the patient who is treated with Schedule II drugs.
There was no statistical difference between the pre-and post-1979 groups with regard to the self-report of the effect of the form on prescribing (X 2 =1.14, p=0.29).
In attempting to ascertain if the Rhode Island Duplicate Prescription Law had an effect on prescribing of Schedule II drugs, the respondents were asked if knowledge of a colleague whose prescribing practices for Schedule II drugs caused the colleague licensure limitation, suspension or revocation or mandatory drug rehabilitation. If the response was yes, the respondent was asked if that knowledge had caused the respondent to curtail his/her prescribing. There was a highly significant association between knowledge of a colleague under ooc review and practicing before 197 9 (X 2 =12. 38, p=O. 00) .
While only 14.1% of the total group was aware of a colleague under DOC review, nearly one-quarter of these respondents (23.4%) stated that their prescribing practices had become limited as a result. This indicates that one of the outcomes of the law may be to have some inhibitory effects on prescribing.

Prescribers' Perceptions of the Law
Since the law was designed to accomplish several ends such as reducing abuse of Schedule II drugs, preventing forgeries and elevating awareness of the potential risks involved for the patient who uses a Schedule II drug, an attempt was made to ascertain how prescribers perceived the above stated variables.

B. Bivariate Association Between Independent Variables and the Dependent Variables
The relationship between the effect of the Rhode Island Duplicate Prescri ption form in decis i on-making for the creation of a therapeutic regimen and selected independent variables for the study group revealed several significant associations. The variables which were associated with therapeutic preference for those in practice after the passage of the law were similar to the profile of the entire group. A medical physician (MD), issuance of a Schedule I I prescription, and prescribing more than one Schedule I I prescription per month were associated with the dependent variable of interest, and each had odds ratios less than 1.0 (0.59, 0.50 and 0.50 respectively) . The influence on therapeutic preference is in part related to an effect by the law~

C. The Final Multivariate Models
Multivariate regression models were tested to best describe the relationship between the dependent variables of interest (effect of the form on decision-making and therapeutic preference to choose an alternative to a Schedule II drug) for the study group and for the pre-and post-1979 subgroups.

The Form Effects Models
The final form effects model for the study group is as The multivariate adjusted odds risk ratio, ROR (adj.), was calculated for each independent variable. Three variables have ROR (adj.) greater than 1.0 (an ROR which equals 1.0 indicated that the risk of the exposed versus the unexposed is equal thus there is no effect in the exposure) .
Those variables which are associated with a risk greater than 1.0 are specialty (1.68), degree (1.56) and the number of years licensed (1.37). In a multivariate model, all other variables are controlled, and in this model it indicated that those who have a specialty practice, those who are medical physicians and those who have been licensed for more than 10 years are at slightly higher risk for the affect of the form.
It is important to note that the 95% confidence intervals for these variables do contain 1.0 in the intervals; therefore, it is possible that these vari ables may not have an effect on the final outcome. The degree variable was highly associated with the e f fect of the form and had the largest ROR ( 1. 98) , thus medical physicians are at nearly twice the risk to report the form affects decision-making as compared to non-MD prescribers.
The ROR for number of years licensed is 1.42; thus, those prescribers in the post-1979 group who had been licensed for more than 10 years are nearly one and one-half times more likely to report that the form affects decision-making.

The Therapeutic Preference Models
The final therapeutic preference model for the entire study group was marginally significant (X 2 =15.084, p=0.0887) and is as follows:  In February of 1979, the Rhode Island Duplicate Prescription Lav vas passed in an effort to track the distribution of Schedule II controlled substances to the ambulatory, non-hospitalized patient via a tvo-part prescription blank iss~ed to the prescriber by the Department of Health.
Rhode Island is one of ten states vhich have similar statutes-.
The purpose of the enclosed survey is to gather information concerning this lav . There are no identifiers in the questionnaire thereby assuring your anonymity, and the data gathered vill be analyzed Jn aggregate form.
The analyses vlll be ' the basis of my Master of Science thesis in Pharmacy Administration.
Your responses and opinions are greatly appreciated. you for your time in completing this survey. 10.
In a situation vhere a Schedule II medication ls clearly indicated, I (31) prefer to prescribe a therapeutic alternative vhlch ls a Schedule !II, IV or V or non-controlled legend drug.
The availability of therapeutic alternatives to Schedule II medications (32) in certain situations is a factor in not prescribing the C-II medication .
In the creation of a therapeutic re gimen, us ln g a RI C-I I presc r i ption (3 3) form affects my decision process.
Vt111z1ng a therapeutic alternative rather than a schedule II med-(34) icatlon, vhlch may be the drug of choice, may be cause for malpractice litigation for the prescriber.
Utilizing a Schedule III, IV or V drug vs. a Schedule II medication (35) (vhen clearly indicated) aay have adverse consequences for the patient.
Ut111zlng a non-controlled legend drug vs. a schedule II ned!cat!on (36) (vhen clearly indicated) aay have adverse consequences for the patient.
There ls less medical risk to the patient by using a Schedule III , IV (37) or V or non-controlled legend drug vhen a Schedule II is appropriate.
There ls better patient compliance vith a Schedule III , IV or v or non-(38) controlled legend drug over a Schedule II medication .
The RI Duplicate Prescription Lav (RIDPL) helps reduce the abuse of legal controlled drugs ln Schedule II.
The RIDPL helps thvart •doctor shoppers• (individuals vho attempt to to obtain prescriptions fro• aultiple prescribers). , The RIDPL causes a decrease in the availability of Schedule· II drugs. The RIDPL has aade practitioners vho prescribe Schedule II 11edlcatio'ns (43) aore avare of the abuse potential/side effects/toxicities of these drugs.
The RIDPL is beneficial for decreasing potential overutillzatlon of Schedule II medications.
The RIDPL helps ln identifying drug dealers or abusers in the medical community.