- Email: [email protected]
Evaluation of Cost Savings to a State Medicaid Program Following a Sertraline Tablet-Splitting Program
RESEARCH
Evaluation of Cost Savings to a State Medicaid Program Following a Sertraline Tablet-Splitting Program Phillip J. Vuchetich, Robert I. Garis, and Allison M.D. Jorgensen
Objectives: To evaluate the economic impact of implementing a sertraline (Zoloft—Pfizer) tablet-splitting program on the Nebraska Medicaid program based on the change in total and per-member-per-month (PMPM) prescription drug costs and to identify any real or perceived problems with tablet splitting using switches among selective serotonin reuptake inhibitors (SSRIs) as a proxy indicator.
Design: Retrospective study of prescription claims before and after the tablet-splitting program was implemented. Setting: Nebraska Medicaid. Patients: All 14,520 patients who received an SSRI during the study period, including 5,466 patients who received at least one prescription for sertraline. Interventions: The Nebraska Medicaid program implemented a mandatory tablet-splitting program for sertraline. Pharmacists were paid a supplemental fee to split tablets. Main Outcome Measures: Total costs, PMPM costs, and switches among SSRIs. Results: Using regression analysis, sertraline was the only SSRI that showed a downward slope in total cost per month, although the decrease was not statistically significant (P = .1156). Fluoxetine (Prozac—Eli Lilly) and paroxetine (Paxil— GlaxoSmithKline) both showed an upward slope, but the increases were not statistically significant (P = .1164 and .0671, respectively). Citalopram (Celexa—Forest) and fluvoxamine showed significantly positive upward slopes (P = .0001 and .0391, respectively). Sertraline was also the only SSRI that showed a downward slope in PMPM costs (P = .0093). Citalopram, fluvoxamine, fluoxetine, and paroxetine all showed an upward slope in PMPM costs (P = .4494, .0008, .0448, and .0482, respectively). The tablet-splitting program was not associated with a net change in patients being switched to or from sertraline. Conclusion: Implementing the sertraline tablet-splitting program significantly decreased the PMPM cost of sertraline prescriptions, but it did not significantly decrease total costs of sertraline, nor did it result in disproportionate numbers of patients switching from sertraline to other SSRIs. Total costs and PMPM costs of the other four SSRI drugs did not decrease.
Keywords: Prescription fees, drug costs, pharmaceutical economics, retrospective studies, insurance claim review, sertraline, paroxetine, fluoxetine, fluvoxamine, citalopram.
J Am Pharm Assoc. 2003;43:497–502.
In April 2000 Nebraska Medicaid instituted a cost-minimization program requiring pharmacists to halve 100 mg sertraline (Zoloft—Pfizer) tablets instead of using the 50 mg tablet for patients taking a 50 mg dose. The 50 mg dose was selected because it is a commonly prescribed dose of sertraline and because splitting the Received September 4, 2002, and in revised form December 3, 2002. Accepted for publication December 30, 2002. Phillip J. Vuchetich, PharmD, MS, is assistant professor; Robert I. Garis, PhD, is assistant professor, Department of Pharmacy Sciences, School of Pharmacy and Health Professions, Creighton University Medical Center, Omaha, Neb. Allison M.D. Jorgensen, PharmD, is director, Nebraska Medicaid Drug Utilization Review Board, Nebraska Pharmacists Association, Lincoln. Correspondence: Phillip J. Vuchetich, PharmD, MS, Department of Pharmacy Sciences, School of Pharmacy and Health Professions, Creighton University Medical Center, 2500 California Plaza, Science Complex, Room 153, Omaha, NE 68178. Fax: 402-280-1883. E-mail: [email protected].
Vol. 43, No. 4
100 mg tablet has been documented to be an effective means of reducing the cost of sertraline in the hospital setting.1,2 No other strengths of sertraline, nor any other drugs, were selected for this pilot program. Although common in institutional settings and for individual patients who want to reduce their medication expenses, tablet splitting may not be appropriate in all cases.3–6 The Nebraska Medicaid sertraline tablet-splitting program was initiated as a pilot program to help the state determine whether mandatory tablet splitting would be an appropriate and effective means of helping control escalating drug costs in Medicaid. A study reported in 1999 identified sertraline as the most economical selective serotonin reuptake inhibitor (SSRI) for a riskbased Medicare contract using split tablets for 50 mg doses, although no cost estimate was reported.7 The authors of another retrospective study, this one of 2,779 patients taking SSRIs, estimated that splitting tablets for 50 mg doses decreased the daily
July/August 2003
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
Journal of the American Pharmacists Association (www.japha.org)
497
RESEARCH
Tablet Splitting
cost by $0.70 per patient.8 Some tablet-splitting programs have required patients, not pharmacists, to split tablets to reduce costs for a managed care organization.3,9,10 One hospital study in which pharmacists split sertraline tablets documented a cost savings of over $260,000 per year, but this amount did not take into account the cost of having the pharmacist split the tablets.1 At least one state Medicaid program (Washington) had a voluntary tablet-splitting program for sertraline at the time of this study, but cost savings estimates had not been reported.11 In recognition of problems reported with tablet splitting,3,4,9,12 the Nebraska Drug Utilization Review (DUR) Board recommended that pharmacists dispense presplit tablets and counsel patients regarding the half-tablet dosage form. The DUR board anticipated that presplitting would ensure that the tablets were split evenly, which is often not the case when patients split tablets themselves, and that any special considerations for the split tablet would be clearly stated on the label and incorporated into patient counseling at the time the prescription was dispensed. Prescribers and pharmacists were permitted to request prior authorization to dispense 50 mg sertraline tablets to individual patients for whom split tablets were deemed therapeutically inappropriate. (Prior authorization was not required for 25 mg or 100 mg tablets or for any other SSRI.) Pharmacies in Nebraska were paid a supplemental professional fee of $0.15 per tablet split. The state set this fee after looking at a state-employed dispensing pharmacist’s base salary, which was $18.00/hour at the time of the study, and estimating that the pharmacist could split two tablets per minute. The claims data reviewed for this study do not include information on whether a pharmacist, pharmacist intern, or pharmacy technician split tablets for a given prescription or the usual and customary amount a pharmacy charged for splitting tablets for non-Medicaid patients.
Objectives The primary objective of this study was to quantify and compare costs to the Nebraska Medicaid program for SSRI prescriptions before and after implementation of a tablet-splitting program using total and per-member, per-month (PMPM) costs for SSRI prescriptions. In addition, we compared the change in number of patients switching between SSRIs in order to get a proxy indicator of real or perceived problems with the sertraline tablet-splitting program.
Methods We obtained approval from the Creighton University Institutional Review Board and the Nebraska Medicaid program to work with deidentified medical claims data. Nebraska Medicaid claims for services provided from January 1, 2000 through September 30, 2000 were used for evaluating drug
498
costs. Claims from October 1, 1999 through December 31, 1999 were used to establish baseline utilization patterns for comparing switch rates among SSRIs. All patients who received a prescription for an SSRI were included because the tablet-splitting program applied to all recipients. Medicaid eligibility and enrollment are dynamic confounding factors in any study involving Medicaid patients because recipients may be reviewed for eligibility on a monthly basis. The pharmacy claims data did not explicitly identify eligibility from month to month; therefore, we chose a short timetable for preintervention and postintervention so that we could minimize the effect of patient eligibility over time. All costs reported are the amount paid by the State of Nebraska to pharmacies, including the supplemental fee of $0.15 per split tablet for sertraline prescriptions. The amount paid does not include drug manufacturer rebates or a $1.00 co-payment from ambulatory adults. During the study period, citalopram (Celexa— Forest), fluoxetine (Prozac—Eli Lilly), fluvoxamine, paroxetine (Paxil—GlaxoSmithKline), and sertraline increased in average wholesale price (AWP) by 4.00%, 2.90%, 13.36%, 4.50%, and 3.10%, respectively.16 Increases in AWPs reflect increases in ingredient costs. Both the PMPM and total costs were evaluated because ingredient cost changes for an individual SSRI are reflected in PMPM costs, regardless of the number of patients who receive the drug, but changes in cost to the program are only evident in total prescription costs, the latter being a factor of the number of patients receiving a drug and the PMPM cost. During the study period, the pharmacy dispensing fee for Nebraska Medicaid prescriptions averaged $4.61, and this amount was included in the calculation of our dependent variables, gross monthly spending and PMPM cost.
Statistical M ethods We performed two analyses of SSRI costs to the Nebraska Medicaid program: (1) a regression analysis to determine significant increases or decreases in total spending and PMPM cost on each of the five SSRIs over the 9 months of the study and (2) an equality of slope analysis (a regression analysis with an interaction term) to determine whether the slopes of the total or PMPM costs of the five SSRI drugs were significantly different from each other. That is, if our “intervention” drug, sertraline, had a PMPM cost with a decreasing slope while the other four drugs had an increasing PMPM slope, we would have stronger evidence that sertraline was moving counter to the trend of the other four SSRIs. We also examined switches among SSRIs during the study period. The rationale here was to determine whether the sertraline tablet-splitting initiative was associated with real or perceived problems that result in switching from one SSRI to another. A statistically significant increase in switches from sertraline to any other SSRI would serve as a proxy indicator of real or perceived problems with the sertraline tablet-splitting program. For the trend analysis, we used the PROC GLM function in SAS 8.2 (SAS Institute, Inc., Cary, N.C.) and performed an equal-
Journal of the American Pharmacists Association (www.japha.org)
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
July/August 2003
Vol. 43, No. 4
Tablet Splitting
ity of slope analysis by including an interaction term in the equation for PMPM costs. An equality of slope analysis is a regression (GLM) analysis in which an interaction term is included to determine whether total or PMPM costs are all moving in the same direction (flat, up, or down). A significant interaction tells the analyst that at least one of the total or PMPM spending trends are significantly different from the others. The rationale for using the equality of slope analysis was to determine whether PMPM trends for all the SSRIs were equal. We set a priori significance at a = 0.05. Regression analysis determines whether a line has no slope (is flat), as determined by a nonsignificant slope, or whether the line has a significant slope (is not flat). If the slope is significantly positive, then the trend is increasing and vice versa. So, when we look at each SSRI, we are asking whether each of the five drugs’ plot of total cost and PMPM cost for each of the 9 months of the study is flat, increasing, or decreasing. All comparisons are looking at total monthly and PMPM spending compared with a flat line. To determine whether there was a net difference in switches among SSRI products before and after the tablet-splitting initiative, we used the nonparametric permutations test for paired replicates with an a = 0.05. Our null hypothesis for this test was that the net switches (to and from another SSRI) would be the same, whether or not there was an intervention. This was the case, with P > .05.17
Results
RESEARCH
the original SSRI. We examined the trend for total spending per SSRI product as the change in total spending for each SSRI per month and found the slopes of all SSRI except sertraline had a positive slope, as presented in Table 1. Using trend analysis, spending for citalopram and fluvoxamine had significantly positive slopes. Fluoxetine and paroxetine had positive, but not statistically significant, slopes. Sertraline had the only downward slope of the five SSRIs, but it was not statistically significant. The trends for PMPM spending demonstrated results similar to those for total spending. That is, PMPM spending for all SSRIs except sertraline had a positive slope, indicating an increase in PMPM costs over the study period. The PMPM trends analysis results are presented in Table 2. PMPM spending on fluoxetine, fluvoxamine, and paroxetine increased significantly over the study period. Citalopram showed a flat trend in PMPM costs, and PMPM spending on sertraline had a significant negative slope. We repeated the PROC GLM analysis with an interaction term to determine whether, in fact, there was an inequality of the slopes. Inequality of slope for the five SSRIs would indicate that the slope of the PMPM cost trend for one of the SSRIs was truly significant relative to those for the other SSRIs. This was the result, with the interaction term significant at P = .0148. One can see that for PMPM spending (Table 2), sertraline is the only SSRI with a significant downward trend. Therefore, sertraline had to be running counter to the other four SSRIs, which had either positive slopes or no slope.
During the study period, Nebraska Medicaid had an average monthly enrollment of 180,678 eligible recipients. Of these recipients, 14,520, or approximately 8.04% of the total eligible recipients, received at least one SSRI during the study period. Broken down by drug dispensed, 2,262 patients received citalopram, 3,587 received fluoxetine, 492 received fluvoxamine, 4,397 received paroxetine, and 5,466 received sertraline. The sum of patients exceeds 14,520 because 1,684 (11.59%) patients either switched among SSRIs or received augmented therapy with an additional SSRI. Russell et al.15 found switch rates among SSRIs of 12.4% to 21.3% over a 12-month period, but the researchers did not report the percentage of patients receiving an additional SSRI to augment
G ross M onthly Spending Total gross monthly spending on all SSRIs in the Nebraska Medicaid program is reported in Figure 1. Figure 2 shows total monthly spending for each SSRI. The change in total spending is represented by the trend of the five regression lines for the period from January 2000 through September 2000. Table 1 shows the slopes, F values, and P values for the spending lines. For the study period, sertraline was the only SSRI whose total monthly cost decreased, although the decrease was not statistically significant (P = .1156), indicating that the trend was flat, as demonstrated by the PROC GLM function in SAS.
Table 1. Total Spending Trends for Selective Serotonin Reuptake Inhibitors in Nebraska Medicaid
Table 2. Per-Member, Per-Month Spending Trends for Selective Serotonin Reuptake Inhibitors in Nebraska Medicaid
Slope/Month ($)
F
P
Slope/Month ($)
F
Citalopram
3,392
7.41
.0001 a
Citalopram
0.23
0.80
.4494
Fluoxetine
1,715
1.54
.1164
Fluoxetine
0.80
2.44
.0448
Fluvoxamine
775
2.53
.0391 b
Fluvoxamine
3.58
5.67
.0008 a
Drug
Drug
P
Paroxetine
1,678
2.16
.0671
Paroxetine
0.53
2.39
.0482 b
Sertraline
–4,429
–1.80
.1156
Sertraline
-1.43
-3.55
.0093 a
a P = .0001.
a P = .01.
b P = .05.
Vol. 43, No. 4
b P = .05.
July/August 2003
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
Journal of the American Pharmacists Association (www.japha.org)
499
RESEARCH
Tablet Splitting
Figure 1. Total Spending on All Selective Serotonin Reuptake Inhibitors, January–September 2000 $900,000
Figure 2. Total Spending on Each Selective Serotonin Reuptake Inhibitor, January–September 2000 $300,000
800,000 700,000
Citalopram Fluoxetine Fluvoxamine
Paroxetine Sertraline
250,000
600,000 200,000 500,000 150,000
400,000 300,000
100,000
200,000 50,000
100,000 0
Jan
Feb
Mar
Per-M em ber, Per-M onth Costs PMPM costs were determined for each SSRI. Patients who received two or more SSRIs were counted in the calculations for each SSRI. Sertraline was the only SSRI that showed a decrease in PMPM cost over the study period. Table 2 shows the change in PMPM spending, F values, and P values for the spending lines. Figure 3 shows the PMPM costs for each SSRI over the study period. Sw itches Am ong SS RIs Switches among SSRIs and augmentation with an additional SSRI are common in practice.15 We sought to learn whether there was a change in the number of patients switching to or from sertraline, compared with other SSRIs. An increase associated with the implementation of the program could indicate that patients might be having more problems with drug therapy. This is a proxy measure, however, because adverse effects, hospitalizations, lack of therapeutic effectiveness, or other factors contributing to the clinical management of SSRI therapy are not recorded in the pharmacy claims database from which the study data were extracted. Table 3 shows the switches among SSRIs. In the 3 months immediately preceding the tablet-splitting program (January–March), 112 patients switched from sertraline to another SSRI, and 81 patients switched from another SSRI to sertraline. Baseline data from the previous 3 months (October–December) were used to establish which SSRI the patient was taking to identify whether a switch occurred, whether the patient was augmenting existing SSRI therapy, or whether he or she had not taken an SSRI recently. In the first 3 months of the program (April–June), 132 patients switched from sertraline to another SSRI, and 108 patients switched from another SSRI to sertraline. The 3 months following this period (July–September) were used to differentiate
500
0
Apr May Jun July Aug Sep
Jan
Feb
Mar
Apr
May
Jun
July
Aug
Sep
Table 3. Switches Among Selective Serotonin Reuptake Inhibitors (SSRIs) Before and After Sertraline Tablet Splitting
Before
After
From Listed Drug to Another SSRI
From Another SSRI to Listed Drug
From Listed Drug to Another SSRI
Citalopram
58
98
75
128
Fluoxetine
70
90
89
92
9
9
13
14
117
88
133
100
Fluvoxamine Paroxetine
From Another SSRI to Listed Drug
Sertraline
112
81
132
108
Total
366
366
442
442
between patients who switched SSRIs and patients who augmented existing therapy with an additional SSRI. We used the nonparametric permutations test to determine whether there was a difference between the switches before and after tablet splitting. The tablet-splitting program was not associated with patients being switched to or from sertraline.
Discussion State Medicaid programs have seen a general increase in drug spending over the past several years, with annual double-digit increases. Unlike those of private insurers and self-insured employers, guidelines for Medicaid preclude the use of common cost containment tools, such as steeply tiered co-payments for generic, branded, or nonformulary products. Sertraline tablet splitting was implemented for Nebraska Medicaid beneficiaries
Journal of the American Pharmacists Association (www.japha.org)
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
July/August 2003
Vol. 43, No. 4
Tablet Splitting
because pharmacists have routinely split sertraline tablets to help self-paying customers save money and because several reports are available citing cost savings to institutions and other programs that split sertraline tablets.1,2,7,15 Pharmacists and patients routinely split tablets to adjust doses or save money. Self-paying patients may ask pharmacists to split tablets and may pay an additional fee for this service. However, the Nebraska Medicaid tablet-splitting program is novel because pharmacists are compensated by a third party payer for splitting tablets. We could find no other published reports of pharmacist compensation from a third party payer for this service. Therefore, it is not clear whether the $0.15 per tablet was appropriate compensation for splitting tablets. The incremental costs to the pharmacy include the time to split tablets, loss of tablets that are rendered unusable when not broken cleanly, and additional counseling time. Of note, since the supplemental fee is based only on the number of tablets dispensed and does not account for tablets that break unevenly or are otherwise rendered unusable, if just one tablet is unusable, the pharmacy may lose nearly the same amount as the supplemental fee. Studies on the true incremental costs and guidance in the pharmacy contract on appropriate billing practices for damaged tablets may be necessary as more and more drugs are considered for tablet splitting. The Nebraska Medicaid sertraline tablet-splitting program has continued for more than 2 years at the time of this writing, and the state is considering requiring tablet splitting for additional drugs. The United States Pharmacopeia (USP) Uniformity of Dosage Units Test13 requires that the relative standard deviation (RSD) of the weight of a selection of whole tablets have a minimal amount of deviation from the mean weight of the tablets, and that no individual tablet be far off of the mean. The uniformity test can be passed if a random selection of 10 tablets (from an original random sample of 30 or more tablets) has an RSD of no more than 6% from the mean weight of the tablets selected, at least 9 weigh withFigure 3. Per-Member, Per-Month Costs for Each Selective Serotonin Reuptake Inhibitor Citalopram Fluoxetine Fluvoxamine
$200 180
Paroxetine Sertraline
RESEARCH
in 15% of the mean weight, and none deviate by more than 25% from the mean weight. If the sample of 10 dosage units fails the test, an additional 20 are selected. The sample can still pass if all 30 tablets have an RSD of no more than 7.8% of the mean weight of all tablets, not more than 3 dosage units deviate from the mean weight by more than 15%, and none deviate from the mean weight by more than 25%. Two articles published after we completed our study present evaluations of the conformance of several commonly split tablets, including sertraline, to USP uniformity standards. One study of pharmacist-dispensed split tablets found an RSD of 7.2% and 6.4% in samples of 20 and 30 half-tables of 100 mg sertraline, respectively, with all of the samples weighing within 15% of the mean weight.5 A controlled laboratory study found an RSD of 7.5% in 60 half-tablets, and none of the samples deviated by more than 15% of the mean weight.6 Based on these two recent studies, it is reasonable to assume that pharmacistsplit dosages of 100 mg sertraline tablets will uniformly contain 50 mg of active drug. It should be noted that the USP standard is for whole dosage units and that no standard has been established for split tablets. Split tablets may be considered a compounded dosage form because they are not in the original form created by the manufacturer; therefore, pharmacists may also consider USP compounding standards that require each dosage unit to be within 10% of its theoretical weight.14 In each of the two uniformity studies, 100 mg sertraline tablets were in the minority of tablets that would pass the test or meet USP compounding standards. In our study, we focused only on prescription drug costs for the Nebraska Medicaid program within a single class of drugs following the initiation of a tablet-splitting program for one of the drugs in that class. Future studies might consider the direct and indirect economic costs, as well as noneconomic impact following changes in drug benefits. Economic studies might consider, if possible, all costs reasonably associated with an illness, depending on the perspective of the study (e.g., the individual, insurer, employer, or society). The impact on quality of life, as well as other noneconomic benefits, should be studied following changes to prescription drug benefits, especially if success of therapy is based, in part, on qualitative measures.
160
Limitations
140 120 100 80 60 40 20 0
Jan
Vol. 43, No. 4
Feb
Mar
Apr
May
Jun
July
Aug
Sep
Our study was a retrospective analysis of data extracted from a pharmacy claims database. The inherent limitations of these databases include the lack of corresponding physician and hospital claims and the lack of clinical histories for individual patients. Nor do pharmacy claims databases capture drug samples given to the patient by the physician. At the time of the study, none of the SSRIs were available from generic manufacturers, and some patients may have received drug samples as part of switching or augmenting therapy.
July/August 2003
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
Journal of the American Pharmacists Association (www.japha.org)
501
RESEARCH
Tablet Splitting
In this short-term study, we focused on the immediate impact of the tablet-splitting program. In the long term, increases in drug prices, the marketing of new drugs in the same class, the introduction of new dosage forms of existing drugs, and the availability of generic equivalents make analysis of and reporting on any program more complex. The supplemental fee paid to the pharmacy for splitting tablets was set by the state and based on an estimated time to split a tablet and a state-employed pharmacist’s hourly pay. Additional reimbursement models could be developed that take into consideration actual time spent by pharmacists, interns, and pharmacy technicians, as well as the potential loss associated with unevenly split tablets that cannot be dispensed. The uniformity of split 100 mg sertraline tablets cannot be generalized to all tablets. Some reports indicate that tablet splitting may not produce tablet fragments that consistently meet professional standards.4–6 Pharmacists, therefore, should be cautious in applying these results to other tablets and should also consider both the uniformity of the split tablets and the clinical impact of variations among doses. In addition, pharmacists must be involved in the tablet-splitting process because patients cannot be expected to be able to split tablets or resolve problems related to improperly split tablets.
Conclusion The implementation of the sertraline tablet-splitting program reduced the total costs and PMPM costs for sertraline and did not result in disproportionate numbers of patients switching from sertraline to other SSRIs. PMPM and total costs of the other four SSRI drugs did not decrease. Nebraska Medicaid has continued the sertraline tablet-splitting program and is considering other drugs for mandatory tablet splitting. Based on these results, we would submit that Medicaid administrators could consider a tabletsplitting initiative on SSRI drugs as a tool for holding the line on runaway Medicaid drug costs. This study was supported, in part, by an unrestricted grant from Pfizer Inc. and a software grant from Microsoft Corporation. Vuchetich owns 15 shares of Pfizer Inc. stock (estimated value: $500). The other authors declare no conflicts of interest or financial interests in any product or service mentioned in the article, including grants, employment, gifts, stock holdings, or honoraria.
502
The authors thank the staff at the Nebraska Department of Health and Human Services, especially Gary Cheloha Sr., Nebraska Medicaid pharmacy consultant, who has since taken a position as executive vice president of the Nebraska Pharmacists Association, and John Franklin, PharmD, Nebraska Medicaid clinical pharmacist, who managed the drug benefit program. The authors also acknowledge Kevin Conaway, information management services program analyst, and Lois Hanson, Medicaid Management Information System technical unit manager, who made the data available for analysis.
References 1. Bult J, Schiff G, Wisniewski M. Sertraline tablet splitting program. Hosp Pharm. 1999;34:996–9. 2. Panzarino PJ Jr., Xuan J. Economic evaluation of major selective serotonin reuptake inhibitors in a managed care population. Manage Care Interface. 2001;14:59–65. 3. Carr-Lopez SM, Mallett MS, Morse T. The tablet splitter: barrier to compliance or cost-saving instrument? Am J Health Syst Pharm. 1995;52:2707–8. 4. McDevitt JT, Gurst AH, Chen Y. Accuracy of tablet splitting. Pharmacotherapy. 1998;18:193–7. 5. Rosenberg JM, Joseph NP, Plakogiannis F. Weight variability of pharmacist-dispensed split tablets. J Am Pharm Assoc. 2002;42:200–5. 6. Teng J, Song CK, Williams RL, Polli JE. Lack of medication dose uniformity in commonly split tablets. J Am Pharm Assoc. 2002;42:195–9. 7. Valdez C, Grier D. Determining the most economical SSRI for a Medicare risk contract. Am J Health Syst Pharm. 1999;56:23–4. 8. Nurnberg HG, Hensley PL, Thompson PM, Paine SS. Modeling the pharmacoeconomic cost of three selective serotonin reuptake inhibitors. Psychiatr Serv. 1999;50:1351–3. 9. Fawell NG, Cookson TL, Scranton SS. Relationship between tablet splitting and compliance, drug acquisition cost, and patient acceptance. Am J Health Syst Pharm. 1999;56:2542–5. 10. Rindone JP, Arriola G. Conversion from fluvastatin to simvastatin therapy at a dose ratio of 8 to 1: effect on serum lipid levels and cost. Clin Ther. 1998;20:340–6. 11. Erdmann C. Drug costs. Double your measure. Hosp Health Netw. 1999;73:16, 18, 20. 12. Sedrati M, Arnaud P, Fontan JE, Brion F. Splitting tablets in half. Am J Hosp Pharm. 1994;51:548, 550. 13. U.S. Pharmacopeial Convention. Uniformity of dosage units. USP25–NF20: U.S. Pharmacopeia and National Formulary. Rockville, Md: United States Pharmacopeial Convention, Inc; 2002. 2082–4. 14. U.S. Pharmacopeial Convention. Pharmacy compounding. USP25– NF20: U.S. Pharmacopeia and National Formulary. Rockville, Md: United States Pharmacopeial Convention, Inc; 2002. 2053–7. 15. Russell JM, Berndt ER, Miceli R, et al. Course and cost of treatment for depression with fluoxetine, paroxetine, and sertraline. Am J Manage Care. 1999;5:597–606. 16. Price-Chek PC database. San Bruno, Calif: FirstDataBank; 2002. 17. Siegel S, Castellan NJ. The case of one sample, two measures or paired replicates. In: Nonparametric Statistics for the Behavioral Sciences. 2nd ed. New York, NY: McGraw-Hill, Inc; 1988:73–101.
Journal of the American Pharmacists Association (www.japha.org)
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
July/August 2003
Vol. 43, No. 4
Evaluation of Cost Savings to a State Medicaid Program Following a Sertraline Tablet-Splitting Program Phillip J. Vuchetich, Robert I. Garis, and Allison M.D. Jorgensen
Objectives: To evaluate the economic impact of implementing a sertraline (Zoloft—Pfizer) tablet-splitting program on the Nebraska Medicaid program based on the change in total and per-member-per-month (PMPM) prescription drug costs and to identify any real or perceived problems with tablet splitting using switches among selective serotonin reuptake inhibitors (SSRIs) as a proxy indicator.
Design: Retrospective study of prescription claims before and after the tablet-splitting program was implemented. Setting: Nebraska Medicaid. Patients: All 14,520 patients who received an SSRI during the study period, including 5,466 patients who received at least one prescription for sertraline. Interventions: The Nebraska Medicaid program implemented a mandatory tablet-splitting program for sertraline. Pharmacists were paid a supplemental fee to split tablets. Main Outcome Measures: Total costs, PMPM costs, and switches among SSRIs. Results: Using regression analysis, sertraline was the only SSRI that showed a downward slope in total cost per month, although the decrease was not statistically significant (P = .1156). Fluoxetine (Prozac—Eli Lilly) and paroxetine (Paxil— GlaxoSmithKline) both showed an upward slope, but the increases were not statistically significant (P = .1164 and .0671, respectively). Citalopram (Celexa—Forest) and fluvoxamine showed significantly positive upward slopes (P = .0001 and .0391, respectively). Sertraline was also the only SSRI that showed a downward slope in PMPM costs (P = .0093). Citalopram, fluvoxamine, fluoxetine, and paroxetine all showed an upward slope in PMPM costs (P = .4494, .0008, .0448, and .0482, respectively). The tablet-splitting program was not associated with a net change in patients being switched to or from sertraline. Conclusion: Implementing the sertraline tablet-splitting program significantly decreased the PMPM cost of sertraline prescriptions, but it did not significantly decrease total costs of sertraline, nor did it result in disproportionate numbers of patients switching from sertraline to other SSRIs. Total costs and PMPM costs of the other four SSRI drugs did not decrease.
Keywords: Prescription fees, drug costs, pharmaceutical economics, retrospective studies, insurance claim review, sertraline, paroxetine, fluoxetine, fluvoxamine, citalopram.
J Am Pharm Assoc. 2003;43:497–502.
In April 2000 Nebraska Medicaid instituted a cost-minimization program requiring pharmacists to halve 100 mg sertraline (Zoloft—Pfizer) tablets instead of using the 50 mg tablet for patients taking a 50 mg dose. The 50 mg dose was selected because it is a commonly prescribed dose of sertraline and because splitting the Received September 4, 2002, and in revised form December 3, 2002. Accepted for publication December 30, 2002. Phillip J. Vuchetich, PharmD, MS, is assistant professor; Robert I. Garis, PhD, is assistant professor, Department of Pharmacy Sciences, School of Pharmacy and Health Professions, Creighton University Medical Center, Omaha, Neb. Allison M.D. Jorgensen, PharmD, is director, Nebraska Medicaid Drug Utilization Review Board, Nebraska Pharmacists Association, Lincoln. Correspondence: Phillip J. Vuchetich, PharmD, MS, Department of Pharmacy Sciences, School of Pharmacy and Health Professions, Creighton University Medical Center, 2500 California Plaza, Science Complex, Room 153, Omaha, NE 68178. Fax: 402-280-1883. E-mail: [email protected].
Vol. 43, No. 4
100 mg tablet has been documented to be an effective means of reducing the cost of sertraline in the hospital setting.1,2 No other strengths of sertraline, nor any other drugs, were selected for this pilot program. Although common in institutional settings and for individual patients who want to reduce their medication expenses, tablet splitting may not be appropriate in all cases.3–6 The Nebraska Medicaid sertraline tablet-splitting program was initiated as a pilot program to help the state determine whether mandatory tablet splitting would be an appropriate and effective means of helping control escalating drug costs in Medicaid. A study reported in 1999 identified sertraline as the most economical selective serotonin reuptake inhibitor (SSRI) for a riskbased Medicare contract using split tablets for 50 mg doses, although no cost estimate was reported.7 The authors of another retrospective study, this one of 2,779 patients taking SSRIs, estimated that splitting tablets for 50 mg doses decreased the daily
July/August 2003
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
Journal of the American Pharmacists Association (www.japha.org)
497
RESEARCH
Tablet Splitting
cost by $0.70 per patient.8 Some tablet-splitting programs have required patients, not pharmacists, to split tablets to reduce costs for a managed care organization.3,9,10 One hospital study in which pharmacists split sertraline tablets documented a cost savings of over $260,000 per year, but this amount did not take into account the cost of having the pharmacist split the tablets.1 At least one state Medicaid program (Washington) had a voluntary tablet-splitting program for sertraline at the time of this study, but cost savings estimates had not been reported.11 In recognition of problems reported with tablet splitting,3,4,9,12 the Nebraska Drug Utilization Review (DUR) Board recommended that pharmacists dispense presplit tablets and counsel patients regarding the half-tablet dosage form. The DUR board anticipated that presplitting would ensure that the tablets were split evenly, which is often not the case when patients split tablets themselves, and that any special considerations for the split tablet would be clearly stated on the label and incorporated into patient counseling at the time the prescription was dispensed. Prescribers and pharmacists were permitted to request prior authorization to dispense 50 mg sertraline tablets to individual patients for whom split tablets were deemed therapeutically inappropriate. (Prior authorization was not required for 25 mg or 100 mg tablets or for any other SSRI.) Pharmacies in Nebraska were paid a supplemental professional fee of $0.15 per tablet split. The state set this fee after looking at a state-employed dispensing pharmacist’s base salary, which was $18.00/hour at the time of the study, and estimating that the pharmacist could split two tablets per minute. The claims data reviewed for this study do not include information on whether a pharmacist, pharmacist intern, or pharmacy technician split tablets for a given prescription or the usual and customary amount a pharmacy charged for splitting tablets for non-Medicaid patients.
Objectives The primary objective of this study was to quantify and compare costs to the Nebraska Medicaid program for SSRI prescriptions before and after implementation of a tablet-splitting program using total and per-member, per-month (PMPM) costs for SSRI prescriptions. In addition, we compared the change in number of patients switching between SSRIs in order to get a proxy indicator of real or perceived problems with the sertraline tablet-splitting program.
Methods We obtained approval from the Creighton University Institutional Review Board and the Nebraska Medicaid program to work with deidentified medical claims data. Nebraska Medicaid claims for services provided from January 1, 2000 through September 30, 2000 were used for evaluating drug
498
costs. Claims from October 1, 1999 through December 31, 1999 were used to establish baseline utilization patterns for comparing switch rates among SSRIs. All patients who received a prescription for an SSRI were included because the tablet-splitting program applied to all recipients. Medicaid eligibility and enrollment are dynamic confounding factors in any study involving Medicaid patients because recipients may be reviewed for eligibility on a monthly basis. The pharmacy claims data did not explicitly identify eligibility from month to month; therefore, we chose a short timetable for preintervention and postintervention so that we could minimize the effect of patient eligibility over time. All costs reported are the amount paid by the State of Nebraska to pharmacies, including the supplemental fee of $0.15 per split tablet for sertraline prescriptions. The amount paid does not include drug manufacturer rebates or a $1.00 co-payment from ambulatory adults. During the study period, citalopram (Celexa— Forest), fluoxetine (Prozac—Eli Lilly), fluvoxamine, paroxetine (Paxil—GlaxoSmithKline), and sertraline increased in average wholesale price (AWP) by 4.00%, 2.90%, 13.36%, 4.50%, and 3.10%, respectively.16 Increases in AWPs reflect increases in ingredient costs. Both the PMPM and total costs were evaluated because ingredient cost changes for an individual SSRI are reflected in PMPM costs, regardless of the number of patients who receive the drug, but changes in cost to the program are only evident in total prescription costs, the latter being a factor of the number of patients receiving a drug and the PMPM cost. During the study period, the pharmacy dispensing fee for Nebraska Medicaid prescriptions averaged $4.61, and this amount was included in the calculation of our dependent variables, gross monthly spending and PMPM cost.
Statistical M ethods We performed two analyses of SSRI costs to the Nebraska Medicaid program: (1) a regression analysis to determine significant increases or decreases in total spending and PMPM cost on each of the five SSRIs over the 9 months of the study and (2) an equality of slope analysis (a regression analysis with an interaction term) to determine whether the slopes of the total or PMPM costs of the five SSRI drugs were significantly different from each other. That is, if our “intervention” drug, sertraline, had a PMPM cost with a decreasing slope while the other four drugs had an increasing PMPM slope, we would have stronger evidence that sertraline was moving counter to the trend of the other four SSRIs. We also examined switches among SSRIs during the study period. The rationale here was to determine whether the sertraline tablet-splitting initiative was associated with real or perceived problems that result in switching from one SSRI to another. A statistically significant increase in switches from sertraline to any other SSRI would serve as a proxy indicator of real or perceived problems with the sertraline tablet-splitting program. For the trend analysis, we used the PROC GLM function in SAS 8.2 (SAS Institute, Inc., Cary, N.C.) and performed an equal-
Journal of the American Pharmacists Association (www.japha.org)
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
July/August 2003
Vol. 43, No. 4
Tablet Splitting
ity of slope analysis by including an interaction term in the equation for PMPM costs. An equality of slope analysis is a regression (GLM) analysis in which an interaction term is included to determine whether total or PMPM costs are all moving in the same direction (flat, up, or down). A significant interaction tells the analyst that at least one of the total or PMPM spending trends are significantly different from the others. The rationale for using the equality of slope analysis was to determine whether PMPM trends for all the SSRIs were equal. We set a priori significance at a = 0.05. Regression analysis determines whether a line has no slope (is flat), as determined by a nonsignificant slope, or whether the line has a significant slope (is not flat). If the slope is significantly positive, then the trend is increasing and vice versa. So, when we look at each SSRI, we are asking whether each of the five drugs’ plot of total cost and PMPM cost for each of the 9 months of the study is flat, increasing, or decreasing. All comparisons are looking at total monthly and PMPM spending compared with a flat line. To determine whether there was a net difference in switches among SSRI products before and after the tablet-splitting initiative, we used the nonparametric permutations test for paired replicates with an a = 0.05. Our null hypothesis for this test was that the net switches (to and from another SSRI) would be the same, whether or not there was an intervention. This was the case, with P > .05.17
Results
RESEARCH
the original SSRI. We examined the trend for total spending per SSRI product as the change in total spending for each SSRI per month and found the slopes of all SSRI except sertraline had a positive slope, as presented in Table 1. Using trend analysis, spending for citalopram and fluvoxamine had significantly positive slopes. Fluoxetine and paroxetine had positive, but not statistically significant, slopes. Sertraline had the only downward slope of the five SSRIs, but it was not statistically significant. The trends for PMPM spending demonstrated results similar to those for total spending. That is, PMPM spending for all SSRIs except sertraline had a positive slope, indicating an increase in PMPM costs over the study period. The PMPM trends analysis results are presented in Table 2. PMPM spending on fluoxetine, fluvoxamine, and paroxetine increased significantly over the study period. Citalopram showed a flat trend in PMPM costs, and PMPM spending on sertraline had a significant negative slope. We repeated the PROC GLM analysis with an interaction term to determine whether, in fact, there was an inequality of the slopes. Inequality of slope for the five SSRIs would indicate that the slope of the PMPM cost trend for one of the SSRIs was truly significant relative to those for the other SSRIs. This was the result, with the interaction term significant at P = .0148. One can see that for PMPM spending (Table 2), sertraline is the only SSRI with a significant downward trend. Therefore, sertraline had to be running counter to the other four SSRIs, which had either positive slopes or no slope.
During the study period, Nebraska Medicaid had an average monthly enrollment of 180,678 eligible recipients. Of these recipients, 14,520, or approximately 8.04% of the total eligible recipients, received at least one SSRI during the study period. Broken down by drug dispensed, 2,262 patients received citalopram, 3,587 received fluoxetine, 492 received fluvoxamine, 4,397 received paroxetine, and 5,466 received sertraline. The sum of patients exceeds 14,520 because 1,684 (11.59%) patients either switched among SSRIs or received augmented therapy with an additional SSRI. Russell et al.15 found switch rates among SSRIs of 12.4% to 21.3% over a 12-month period, but the researchers did not report the percentage of patients receiving an additional SSRI to augment
G ross M onthly Spending Total gross monthly spending on all SSRIs in the Nebraska Medicaid program is reported in Figure 1. Figure 2 shows total monthly spending for each SSRI. The change in total spending is represented by the trend of the five regression lines for the period from January 2000 through September 2000. Table 1 shows the slopes, F values, and P values for the spending lines. For the study period, sertraline was the only SSRI whose total monthly cost decreased, although the decrease was not statistically significant (P = .1156), indicating that the trend was flat, as demonstrated by the PROC GLM function in SAS.
Table 1. Total Spending Trends for Selective Serotonin Reuptake Inhibitors in Nebraska Medicaid
Table 2. Per-Member, Per-Month Spending Trends for Selective Serotonin Reuptake Inhibitors in Nebraska Medicaid
Slope/Month ($)
F
P
Slope/Month ($)
F
Citalopram
3,392
7.41
.0001 a
Citalopram
0.23
0.80
.4494
Fluoxetine
1,715
1.54
.1164
Fluoxetine
0.80
2.44
.0448
Fluvoxamine
775
2.53
.0391 b
Fluvoxamine
3.58
5.67
.0008 a
Drug
Drug
P
Paroxetine
1,678
2.16
.0671
Paroxetine
0.53
2.39
.0482 b
Sertraline
–4,429
–1.80
.1156
Sertraline
-1.43
-3.55
.0093 a
a P = .0001.
a P = .01.
b P = .05.
Vol. 43, No. 4
b P = .05.
July/August 2003
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
Journal of the American Pharmacists Association (www.japha.org)
499
RESEARCH
Tablet Splitting
Figure 1. Total Spending on All Selective Serotonin Reuptake Inhibitors, January–September 2000 $900,000
Figure 2. Total Spending on Each Selective Serotonin Reuptake Inhibitor, January–September 2000 $300,000
800,000 700,000
Citalopram Fluoxetine Fluvoxamine
Paroxetine Sertraline
250,000
600,000 200,000 500,000 150,000
400,000 300,000
100,000
200,000 50,000
100,000 0
Jan
Feb
Mar
Per-M em ber, Per-M onth Costs PMPM costs were determined for each SSRI. Patients who received two or more SSRIs were counted in the calculations for each SSRI. Sertraline was the only SSRI that showed a decrease in PMPM cost over the study period. Table 2 shows the change in PMPM spending, F values, and P values for the spending lines. Figure 3 shows the PMPM costs for each SSRI over the study period. Sw itches Am ong SS RIs Switches among SSRIs and augmentation with an additional SSRI are common in practice.15 We sought to learn whether there was a change in the number of patients switching to or from sertraline, compared with other SSRIs. An increase associated with the implementation of the program could indicate that patients might be having more problems with drug therapy. This is a proxy measure, however, because adverse effects, hospitalizations, lack of therapeutic effectiveness, or other factors contributing to the clinical management of SSRI therapy are not recorded in the pharmacy claims database from which the study data were extracted. Table 3 shows the switches among SSRIs. In the 3 months immediately preceding the tablet-splitting program (January–March), 112 patients switched from sertraline to another SSRI, and 81 patients switched from another SSRI to sertraline. Baseline data from the previous 3 months (October–December) were used to establish which SSRI the patient was taking to identify whether a switch occurred, whether the patient was augmenting existing SSRI therapy, or whether he or she had not taken an SSRI recently. In the first 3 months of the program (April–June), 132 patients switched from sertraline to another SSRI, and 108 patients switched from another SSRI to sertraline. The 3 months following this period (July–September) were used to differentiate
500
0
Apr May Jun July Aug Sep
Jan
Feb
Mar
Apr
May
Jun
July
Aug
Sep
Table 3. Switches Among Selective Serotonin Reuptake Inhibitors (SSRIs) Before and After Sertraline Tablet Splitting
Before
After
From Listed Drug to Another SSRI
From Another SSRI to Listed Drug
From Listed Drug to Another SSRI
Citalopram
58
98
75
128
Fluoxetine
70
90
89
92
9
9
13
14
117
88
133
100
Fluvoxamine Paroxetine
From Another SSRI to Listed Drug
Sertraline
112
81
132
108
Total
366
366
442
442
between patients who switched SSRIs and patients who augmented existing therapy with an additional SSRI. We used the nonparametric permutations test to determine whether there was a difference between the switches before and after tablet splitting. The tablet-splitting program was not associated with patients being switched to or from sertraline.
Discussion State Medicaid programs have seen a general increase in drug spending over the past several years, with annual double-digit increases. Unlike those of private insurers and self-insured employers, guidelines for Medicaid preclude the use of common cost containment tools, such as steeply tiered co-payments for generic, branded, or nonformulary products. Sertraline tablet splitting was implemented for Nebraska Medicaid beneficiaries
Journal of the American Pharmacists Association (www.japha.org)
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
July/August 2003
Vol. 43, No. 4
Tablet Splitting
because pharmacists have routinely split sertraline tablets to help self-paying customers save money and because several reports are available citing cost savings to institutions and other programs that split sertraline tablets.1,2,7,15 Pharmacists and patients routinely split tablets to adjust doses or save money. Self-paying patients may ask pharmacists to split tablets and may pay an additional fee for this service. However, the Nebraska Medicaid tablet-splitting program is novel because pharmacists are compensated by a third party payer for splitting tablets. We could find no other published reports of pharmacist compensation from a third party payer for this service. Therefore, it is not clear whether the $0.15 per tablet was appropriate compensation for splitting tablets. The incremental costs to the pharmacy include the time to split tablets, loss of tablets that are rendered unusable when not broken cleanly, and additional counseling time. Of note, since the supplemental fee is based only on the number of tablets dispensed and does not account for tablets that break unevenly or are otherwise rendered unusable, if just one tablet is unusable, the pharmacy may lose nearly the same amount as the supplemental fee. Studies on the true incremental costs and guidance in the pharmacy contract on appropriate billing practices for damaged tablets may be necessary as more and more drugs are considered for tablet splitting. The Nebraska Medicaid sertraline tablet-splitting program has continued for more than 2 years at the time of this writing, and the state is considering requiring tablet splitting for additional drugs. The United States Pharmacopeia (USP) Uniformity of Dosage Units Test13 requires that the relative standard deviation (RSD) of the weight of a selection of whole tablets have a minimal amount of deviation from the mean weight of the tablets, and that no individual tablet be far off of the mean. The uniformity test can be passed if a random selection of 10 tablets (from an original random sample of 30 or more tablets) has an RSD of no more than 6% from the mean weight of the tablets selected, at least 9 weigh withFigure 3. Per-Member, Per-Month Costs for Each Selective Serotonin Reuptake Inhibitor Citalopram Fluoxetine Fluvoxamine
$200 180
Paroxetine Sertraline
RESEARCH
in 15% of the mean weight, and none deviate by more than 25% from the mean weight. If the sample of 10 dosage units fails the test, an additional 20 are selected. The sample can still pass if all 30 tablets have an RSD of no more than 7.8% of the mean weight of all tablets, not more than 3 dosage units deviate from the mean weight by more than 15%, and none deviate from the mean weight by more than 25%. Two articles published after we completed our study present evaluations of the conformance of several commonly split tablets, including sertraline, to USP uniformity standards. One study of pharmacist-dispensed split tablets found an RSD of 7.2% and 6.4% in samples of 20 and 30 half-tables of 100 mg sertraline, respectively, with all of the samples weighing within 15% of the mean weight.5 A controlled laboratory study found an RSD of 7.5% in 60 half-tablets, and none of the samples deviated by more than 15% of the mean weight.6 Based on these two recent studies, it is reasonable to assume that pharmacistsplit dosages of 100 mg sertraline tablets will uniformly contain 50 mg of active drug. It should be noted that the USP standard is for whole dosage units and that no standard has been established for split tablets. Split tablets may be considered a compounded dosage form because they are not in the original form created by the manufacturer; therefore, pharmacists may also consider USP compounding standards that require each dosage unit to be within 10% of its theoretical weight.14 In each of the two uniformity studies, 100 mg sertraline tablets were in the minority of tablets that would pass the test or meet USP compounding standards. In our study, we focused only on prescription drug costs for the Nebraska Medicaid program within a single class of drugs following the initiation of a tablet-splitting program for one of the drugs in that class. Future studies might consider the direct and indirect economic costs, as well as noneconomic impact following changes in drug benefits. Economic studies might consider, if possible, all costs reasonably associated with an illness, depending on the perspective of the study (e.g., the individual, insurer, employer, or society). The impact on quality of life, as well as other noneconomic benefits, should be studied following changes to prescription drug benefits, especially if success of therapy is based, in part, on qualitative measures.
160
Limitations
140 120 100 80 60 40 20 0
Jan
Vol. 43, No. 4
Feb
Mar
Apr
May
Jun
July
Aug
Sep
Our study was a retrospective analysis of data extracted from a pharmacy claims database. The inherent limitations of these databases include the lack of corresponding physician and hospital claims and the lack of clinical histories for individual patients. Nor do pharmacy claims databases capture drug samples given to the patient by the physician. At the time of the study, none of the SSRIs were available from generic manufacturers, and some patients may have received drug samples as part of switching or augmenting therapy.
July/August 2003
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
Journal of the American Pharmacists Association (www.japha.org)
501
RESEARCH
Tablet Splitting
In this short-term study, we focused on the immediate impact of the tablet-splitting program. In the long term, increases in drug prices, the marketing of new drugs in the same class, the introduction of new dosage forms of existing drugs, and the availability of generic equivalents make analysis of and reporting on any program more complex. The supplemental fee paid to the pharmacy for splitting tablets was set by the state and based on an estimated time to split a tablet and a state-employed pharmacist’s hourly pay. Additional reimbursement models could be developed that take into consideration actual time spent by pharmacists, interns, and pharmacy technicians, as well as the potential loss associated with unevenly split tablets that cannot be dispensed. The uniformity of split 100 mg sertraline tablets cannot be generalized to all tablets. Some reports indicate that tablet splitting may not produce tablet fragments that consistently meet professional standards.4–6 Pharmacists, therefore, should be cautious in applying these results to other tablets and should also consider both the uniformity of the split tablets and the clinical impact of variations among doses. In addition, pharmacists must be involved in the tablet-splitting process because patients cannot be expected to be able to split tablets or resolve problems related to improperly split tablets.
Conclusion The implementation of the sertraline tablet-splitting program reduced the total costs and PMPM costs for sertraline and did not result in disproportionate numbers of patients switching from sertraline to other SSRIs. PMPM and total costs of the other four SSRI drugs did not decrease. Nebraska Medicaid has continued the sertraline tablet-splitting program and is considering other drugs for mandatory tablet splitting. Based on these results, we would submit that Medicaid administrators could consider a tabletsplitting initiative on SSRI drugs as a tool for holding the line on runaway Medicaid drug costs. This study was supported, in part, by an unrestricted grant from Pfizer Inc. and a software grant from Microsoft Corporation. Vuchetich owns 15 shares of Pfizer Inc. stock (estimated value: $500). The other authors declare no conflicts of interest or financial interests in any product or service mentioned in the article, including grants, employment, gifts, stock holdings, or honoraria.
502
The authors thank the staff at the Nebraska Department of Health and Human Services, especially Gary Cheloha Sr., Nebraska Medicaid pharmacy consultant, who has since taken a position as executive vice president of the Nebraska Pharmacists Association, and John Franklin, PharmD, Nebraska Medicaid clinical pharmacist, who managed the drug benefit program. The authors also acknowledge Kevin Conaway, information management services program analyst, and Lois Hanson, Medicaid Management Information System technical unit manager, who made the data available for analysis.
References 1. Bult J, Schiff G, Wisniewski M. Sertraline tablet splitting program. Hosp Pharm. 1999;34:996–9. 2. Panzarino PJ Jr., Xuan J. Economic evaluation of major selective serotonin reuptake inhibitors in a managed care population. Manage Care Interface. 2001;14:59–65. 3. Carr-Lopez SM, Mallett MS, Morse T. The tablet splitter: barrier to compliance or cost-saving instrument? Am J Health Syst Pharm. 1995;52:2707–8. 4. McDevitt JT, Gurst AH, Chen Y. Accuracy of tablet splitting. Pharmacotherapy. 1998;18:193–7. 5. Rosenberg JM, Joseph NP, Plakogiannis F. Weight variability of pharmacist-dispensed split tablets. J Am Pharm Assoc. 2002;42:200–5. 6. Teng J, Song CK, Williams RL, Polli JE. Lack of medication dose uniformity in commonly split tablets. J Am Pharm Assoc. 2002;42:195–9. 7. Valdez C, Grier D. Determining the most economical SSRI for a Medicare risk contract. Am J Health Syst Pharm. 1999;56:23–4. 8. Nurnberg HG, Hensley PL, Thompson PM, Paine SS. Modeling the pharmacoeconomic cost of three selective serotonin reuptake inhibitors. Psychiatr Serv. 1999;50:1351–3. 9. Fawell NG, Cookson TL, Scranton SS. Relationship between tablet splitting and compliance, drug acquisition cost, and patient acceptance. Am J Health Syst Pharm. 1999;56:2542–5. 10. Rindone JP, Arriola G. Conversion from fluvastatin to simvastatin therapy at a dose ratio of 8 to 1: effect on serum lipid levels and cost. Clin Ther. 1998;20:340–6. 11. Erdmann C. Drug costs. Double your measure. Hosp Health Netw. 1999;73:16, 18, 20. 12. Sedrati M, Arnaud P, Fontan JE, Brion F. Splitting tablets in half. Am J Hosp Pharm. 1994;51:548, 550. 13. U.S. Pharmacopeial Convention. Uniformity of dosage units. USP25–NF20: U.S. Pharmacopeia and National Formulary. Rockville, Md: United States Pharmacopeial Convention, Inc; 2002. 2082–4. 14. U.S. Pharmacopeial Convention. Pharmacy compounding. USP25– NF20: U.S. Pharmacopeia and National Formulary. Rockville, Md: United States Pharmacopeial Convention, Inc; 2002. 2053–7. 15. Russell JM, Berndt ER, Miceli R, et al. Course and cost of treatment for depression with fluoxetine, paroxetine, and sertraline. Am J Manage Care. 1999;5:597–606. 16. Price-Chek PC database. San Bruno, Calif: FirstDataBank; 2002. 17. Siegel S, Castellan NJ. The case of one sample, two measures or paired replicates. In: Nonparametric Statistics for the Behavioral Sciences. 2nd ed. New York, NY: McGraw-Hill, Inc; 1988:73–101.
Journal of the American Pharmacists Association (www.japha.org)
Downloaded From: http://japha.org/ by a Dalhousie University Library User on 05/18/2015
July/August 2003
Vol. 43, No. 4