The overall mean was calculated using
| Cost Analysis of Solid Oral Medication Distribution and Administration: Applications for Chronic Nitrate Therapy | Edgar R. Gonzalez
Robert J. Chaponis R. K. Elswick, Jr. H. Edward Davidson Rhonda B. Liberto |
Objective: To determine the labor time and costs associated with dispensing and administering solid oral nitrate forms (i.e., immediate-release isosorbide dinitrate [ISDN], immediate-release isosorbide mononitrate [ISMN], and extended-release isosorbide mononitrate [ISMN]) to nursing facility residents.
Design: Trained observers measured the time required to complete dispensing and administration pathways for a 30-day bingo-card and for a 24-hour unit-dose (UD) system (nonprepackaged and prepackaged). Technicians', pharmacists', and nurses' salaries at participating sites were incorporated to determine the cost per administered dose. The costs of oral nitrate forms were obtained from 1994 Redbook average wholesale prices for bulk and UD products.
Setting: One long-term care provider and three nursing facilities in both California and Virginia.
Main Outcome Measures: Labor time and costs associated with dispensing and administering solid oral nitrate forms.
Results: Based on average California and Virginia salaries, the costs of the two 24-hour UD systems did not differ statistically (32.26 cents, nonprepackaged, versus 31.60 cents, prepackaged), but the 30-day bingo-card cost was significantly lower (22.39 cents; p = 0.001). Daily dispensing and administration costs for the 30-day bingo-card and 24-hour nonprepackaged and prepackaged UD systems, respectively, were as follows: Imdur 120 mg/day = $1.53, $1.63, $1.74; Ismo 20 mg B.I.D. = $1.78, $1.98, $2.08; Monoket 20 mg B.I.D. = $1.51, $1.72, $1.79; ISDN 20 mg T.I.D. = $0.82, $1.13, $1.44; and Isordil 20 mg T.I.D. = $1.89, $2.19, $2.25.
Conclusion: Even with a higher initial average wholesale price, the labor-time and cost-benefit of once-daily solid oral nitrate therapy equaled or exceeded those of twice- and thrice-daily brand-name therapies under the 30-day bingo-card system.
Abbreviations Used: AWP = Average wholesale price; ISDN = Isosorbide dinitrate; ISMN = Isosorbide mononitrate; LTC = Long-term care; NF = Nursing facility; UD = Unit dose; ANOVA = analysis of variance.
Key Words: Chronic oral nitrate therapy, Time-motion analysis, Drug administration, Oral solid dosage forms, Nurses, Costs.
Consult Pharm 1997; 12: 690-702.
Isosorbide dinitrate (ISDN) is an oral nitrate formulation frequently prescribed to treat coronary artery disease. Generic ISDN has the lowest per tablet cost of any available oral nitrate preparation.1 However, the requirement of multiple daily dosing and the measures needed to prevent the development of nitrate tolerance produce additional dispensing and administration costs. These factors should be considered when comparing the total costs of chronic nitrate therapy in an institutional setting.
Concerns about nitrate tolerance and loss of therapeutic effect during long-term therapy with ISDN2 led to the clinical introduction of isosorbide-5-mononitrate (ISMN). Immediate-release oral forms of ISMN (i.e., Ismo, Monoket) offer a twice-daily regimen less likely than ISDN preparations to produce nitrate tolerance.3 However, the eccentric dosing schedule (i.e., every seven hours) required to minimize the occurrence of tolerance3,4 may adversely affect compliance,5 compromise the effectiveness of immediate-release ISMN,6,7 and increase nursing time and drug-administration costs. Efforts to overcome compliance problems associated with the eccentric dosing of immediate-release ISMN led to the development of an extended-release ISMN preparation (Imdur) that provides once-daily dosing.6,8
Although Imdur is more expensive per tablet than forms of ISDN or immediate-release ISMN, the costs of dispensing and administering once-daily Imdur may be less than the overall costs associated with multiple daily doses of ISDN or ISMN. Unfortunately, no published comparative total costs have been available to guide the selection of cost-efficient chronic nitrate regimens in the institutional setting. Therefore, this study was conducted to determine the labor time and costs associated with dispensing and administering solid oral dosage forms to residents in nursing facilities (NFs). These forms included the following nitrate products: immediate-release ISDN, immediate-release ISMN, and Imdur.
Methods
Time-motion analyses were conducted to determine the time (in seconds) required to dispense and administer a single solid dosage form of medication via two systems: the 30-day bingo-card system and the 24-hour unit dose (UD) system. The 30-day bingo-card system was observed in three NFs and one LTC provider pharmacy in California, and the 24-hour UD system was observed in three NFs and one LTC provider pharmacy in Virginia.
Medication Dispensing by Pharmacy Providers
In the 30-day bingo-card system, a patient's 30-day drug requirement is packaged in a single dispensing unit (i.e., a bingo card), and an individual dose is extracted from the card at the time of drug administration. In the 24-hour UD system, a patient's 24-hour drug requirement is packaged in a single drawer, and the drawer is replaced after each 24-hour supply is administered. The critical steps for dispensing medications via each system were identified (Figures 1 and 2).
For the 30-day bingo-card system, the technician's prepackaging and procuring and the pharmacist's checking of medications were observed (Figure 1). The following data were collected for each process: (1) time to complete the process; (2) number of 30-day bingo cards filled; and (3) number of doses dispensed. The dispensing process included the pharmacist's quality assurance check of the technician's work. For the 24-hour UD system, the technician's prepackaging and the pharmacist's checking, procuring, and quality control checking of medications were observed (Figure 2). The time to complete a process and the number of dispensed doses were collected for each process.
Medication Administration in Nursing Facilities
We measured the time required to administer solid oral medications, including nitrate preparations, dispensed via the 30-day bingo-card system or the 24-hour UD system. Only NF patients 21 years of age or older who were receiving chronic oral therapy with ISDN, immediate-release ISMN, or Imdur were included in the study. Patients receiving only acute nitrate therapy (i.e., sublingual, translingual, transmucosal, or intravenous) and patients with special considerations (e.g., spinal cord injury or head trauma) were excluded.
Medication administration involved the following steps: (1) identifying the patient; (2) verifying the order from the medication administration record; (3) verifying the drug and dose; (4) preparing the dose, if necessary; (5) assessing vital signs, if necessary; (6) administering the medication; and (7) charting the medication as given. The timing of events began when the designated medication administration nurse left the nurses' station to administer medications and ended when the nurse returned to the nurses' station. Minor interruptions in this process by nurses (e.g., setting up a meal tray or turning a patient in bed) were included in time accrual; however, major interruptions (e.g., telephone calls from physicians) were excluded by stopping timing until the nurse returned to the medication cart. Research investigators trained consultant pharmacists to record the time needed to complete the medication administration "pass" and the number of doses administered.
Data Collection
The distribution and administration of solid oral dosage forms, including ISDN and ISMN, were observed and timed once during all eight- or 12-hour work shifts. During a work shift, trained observers measured, to the nearest one hundredth of a second, the time required to complete the critical steps. Three different observers conducted each time-motion study on three different weekdays to reduce potential biases introduced by interruptions, distractions, or changing workloads. Subjects knew they were being observed and timed but were not informed of their time scores to minimize any Hawthorne effect.9
Data Analysis
Mean times and standard errors were computed for each individually timed dispensing or administration process using weighted mean and variance formulas (Appendix 1). Salary data for technicians, pharmacists, and nurses at participating centers were used to incorporate the cost per administered dose into the time-motion analysis. The mean time for each process was multiplied by the appropriate cost, based on the practitioner's salary, and the processes' standard errors were adjusted to relate to the cost means. The means and standard errors were pooled across processes to obtain total labor costs for the 30-day bingo-card system and the 24-hour UD system. Data are presented as the mean plus or minus one standard error (SEM).
TABLE 1. Average Wholesale Prices for Nitrate Products (from Reference 8)
| Product | Bulka Price per 100 ($) | Unit Doseb Price per 100 ($) |
|---|---|---|
| Imdur 60 mg tablet | 64.80 | 71.28 |
| Ismo 20 mg tablet | 65.81 | 72.38 |
| Monoket 20 mg tablet | 52.65 | 57.90 |
| ISDN 10 mg tablet (Geneva Pharm) | 4.62 | 10.75 |
| ISDN 20 mg tablet (Geneva Pharm) | 4.27 | 16.25 |
| Isordil 10 mg titradose | 24.63 | 27.11 |
| Isordil 20 mg titradose | 39.73 | 43.31 |
The cost of oral nitrate forms was obtained from 1994 Redbook1 average wholesale prices (AWPs) for bulk and UD products (Table 1). The following equations were used to calculate the total daily cost ($/day) of dispensing and administering nitrates:
A single-factor analysis of variance (ANOVA) was conducted on study data. Tests of individual means were performed using Student's t test for paired data. The level of significance was set at p < 0.05.
Ethical Considerations
The study was conducted with an institutional review board-approved waiver of patient informed consent, as study participation did not pose an added risk to the NF patient. Approval for conducting the study was obtained from each facility's medical director, facility administrator, or director of nursing.
Results
The mean times for distribution processes A through C in the 30-day bingo-card system (Figure 1) are shown in Table 2. The morning and noon administrations of medications were timed in each of three NFs on multiple days. The mean nursing administration time in the 30-day bingo-card system for each NF is shown in Table 3. The mean time pooled across NFs was 69.86 ± 4.62 sec.
TABLE 2. Time (Seconds) for Processes A Through C in the 30-Day Bingo-Card System
| Single Card | Single Dose | |||||||
|---|---|---|---|---|---|---|---|---|
| Process Description | Mean | S.E.M. | No. Time Segments | Weight (no. cards) | Mean | S.E.M. | No. Time Segments | Weight (no. doses) |
| A. Technician prepack | 54.51 | 3.60 | 6 | 579 | 1.65 | 0.16 | 6 | 19,092 |
| B. Technician procure | 98.80 | 2.19 | 9 | 1,248 | 3.27 | 0.07 | 9 | 37,720 |
| C. Pharmacist check | 7.74 | 0.03 | 2 | 980 | 0.25 | 0.01 | 2 | 30,051 |
TABLE 3. Nursing Administration Times (Seconds)
| 30-Day Bing-Card System | 24-Hour Unit Dose System | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Nursing Facility | Mean | S.E.M. | N | Weight (no. doses) | Nursing Facility | Mean | S.E.M. | N | Weight (no. doses) |
| 1 | 59.35 | 4.86 | 10 | 773 | 1 | 96.04 | 10.20 | 13 | 606 |
| 2 | 71.52 | 8.85 | 6 | 659 | 2 | 82.22 | 11.32 | 9 | 431 |
| 3 | 87.91 | 9.70 | 7 | 389 | 3 | 98.26 | 8.34 | 15 | 574 |
| Pooled | 69.86 | 4.62 | 23 | 1,821 | Pooled | 93.27 | 5.62 | 37 | 1,611 |
The mean times for distribution processes A through D in the 24-hour UD system (Figure 2) are shown in Table 4. The mean nursing administration time in the 24-hour UD system for each NF is shown in Table 3. The mean time pooled across NFs was 93.27 ± 5.62 sec.
TABLE 4. Times for Processes A Through D in the 24-Hour Unit Dose System
| Time (seconds) | |||||
|---|---|---|---|---|---|
| Process | Description | Mean | S.E.M. | No. Time Segments | Weight (no. Doses) |
| A | Technician prepack | 2.72 | 0.16 | 3 | 16,340 |
| B | Pharmacist check | 0.19 | 0.07 | 11 | 15,309 |
| C | Technician procure | 9.24 | 0.18 | 12 | 6,764 |
| D | Technician check | 5.38 | 0.49 | 8 | 3,844 |
The mean costs (in cents) for dispensing and administering a single, solid oral dosage form in the NF setting, using actual salaries, were as follows: 26.31 ± 0.23; 26.97 ± 0.21; and 26.34 ± 0.43 for the 30-day bingo-card and 24-hour non-prepackaged and prepackaged UD systems, respectively. Because practitioner salaries were substantially higher in California than Virginia, the nursing dispensing and administering costs were recalculated using the average of California and Virginia salaries (Table 5). A single-factor ANOVA was used to compare the mean costs of the two 24-hour UD systems and the 30-day bingo-card system, using average salaries. The ANOVA was significant (F = 543.9; df = 125, 676; p < 0.0001).
TABLE 5. Cost of Dispensing and Administering a Single Solid Oral Dose Based on the Average of California and Virginia Salaries
| Costs (cents) | |||
|---|---|---|---|
| Pharmacy System | Mean | S.E.M. | 95% Confidence Interval |
| 30-day bingo-card | 22.39 | 0.20 | 22.00 - 22.79 |
| 24-hour unit dose nonprepackaged | 32.26 | 0.25 | 31.77 - 32.75 |
| 24-hour unit dose prepackaged | 31.60 | 0.52 | 30.57 - 32.63 |
To assess group differences, three t tests were run with alpha set at 0.0167 (0.05/3) to adjust for multiple testing. While the mean costs of the two 24-hour UD dispensing systems were not statistically different, the 30-day bingo-card cost was significantly lower than either 24-hour UD system cost (p = 0.001). Table 6 presents the total daily costs ($), derived from equations 1 and 2, of dispensing and administering regimens of given solid oral nitrate preparations using actual and average salaries.
TABLE 6. Total Daily Cost of Dispensing and Administering Nitrates in Nursing Facilities Based on Actual Versus Average Salaries
| Daily Cost ($) Salary Basis | ||||||
|---|---|---|---|---|---|---|
| 30-Day Bingo Card | 24-Hour Unit Dose Non-prepackaged | 24-Hour Unit Dose Prepackaged | ||||
| Nitrate Regimen | Actual | Average | Actual | Average | Actual | Average |
| Imdur 120 mg QD (two x 60 mg tablets) | $1.56 | $1.53 | $1.58 | $1.63 | $1.69 | $1.74 |
| Ismo 20 mg B.I.D. | $1.85 | $1.78 | $1.88 | $1.98 | $1.97 | $2.08 |
| Monoket 20 mg B.I.D. | $1.59 | $1.51 | $1.61 | $1.72 | $1.68 | $1.79 |
| ISDN 10 mg T.I.D.(Geneva Pharm) | $0.95 | $0.83 | $0.98 | $1.14 | $1.11 | $1.27 |
| ISDN 20 mg T.I.D.(Geneva Pharm) | $0.94 | $0.82 | $0.97 | $1.13 | $1.28 | $1.44 |
| Isordil 10 mg T.I.D. | $1.55 | $1.43 | $1.58 | $1.74 | $1.60 | $1.76 |
| Isordil 20 mg T.I.D. | $2.00 | $1.89 | $2.03 | $2.19 | $2.09 | $2.25 |
Discussion
Work measurement research assesses the productivity of personnel. When used in health care, such research provides information regarding the cost of a medical service or intervention to institutions and providers. Time-motion analysis is a work-measurement tool in which investigators directly observe and measure the time required by one or more subjects to perform a task. The tasks to be observed and compared should: (1) be similar, (2) be repetitive, (3) be accomplishable in a short time period, (4) have a predictable outcome, and (5) have identifiable starting and stopping points. Pharmacy researchers have used time-motion analysis to assess the time and cost of compounding admixtures.10
In this study, the principal driver of the cost per administered oral dose was the time required to administer the medication. Nursing administration time constituted 96% of this cost in the 30-day bingo-card system versus 90% in the 24-hour UD system. The average nursing time required to administer one dose was substantially different between the 30-day bingo-card and 24-hour UD systems, 69.86 and 93.27 seconds, respectively. Despite this difference, the higher salaries of health care professionals in California, which employed the 30-day bingo-card system, equalized the cost per administered dose compared with the 24-hour UD system. To eliminate the bias introduced by the salary differential between California and Virginia, we conducted a post hoc analysis using average salaries for the two states. With comparable salaries, the cost per administered dose was 9.54 cents higher, on average, in the 24-hour UD system than in the 30-day bingo-card system. The use of average salaries shifts costs in favor of the more rapid 30-day bingo-card system.
The average salaries for pharmacy technicians, pharmacists, and nurses at participating sites were combined with the time-motion analysis and 1994 AWP values to compare the cost per dose of dispensing and administering a once-daily regimen of Imdur, a two- or three-times daily regimen of ISDN, and a twice-daily regimen of ISMN. In this study, participating sites provided salary data. AWPs were used in computing the daily cost of nitrate therapy because AWP data are readily accessible to pharmacy providers. Practitioners may choose to apply different salaries and drug costs to the model to determine total daily costs in their settings.
Although ISDN was administered two or three times daily, it was the most affordable regimen because of its low AWP. We observed a potential cost savings with Imdur 120 mg once daily, compared with ISMN 20 mg twice daily or Isordil 20 mg three times daily, despite Imdur's higher AWP. By reducing the time and labor costs associated with drug distribution and administration, use of once-daily therapies may redirect valuable health care resources back to patient care. This redirection is especially important during times of fiscal constraint and personnel shortages.
Limitations
This study measured the time and costs of dispensing and administering one solid oral dose of medication as a surrogate for nitrates. Several limitations must be considered when extrapolating the results of this study. First, the health care practitioners knew they were being monitored and that their work was being timed. Although effort was made to minimize the potential for a Hawthorne effect, this threat could not be eliminated. Second, given the high prevalence of polypharmacy in NFs, nitrate product administration could not be isolated from other medication administration. The labor cost for a solid oral dosage form was then used by inference to estimate the cost of administering solid oral dosage forms of a nitrate preparation.
Averaging the salaries eliminated the cost bias introduced by the difference between California and Virginia salaries. Although the 30-day bingo-card system was less costly than the 24-hour UD system secondary to shorter nursing administration times, this difference cannot be attributed to the distribution system because a crossover comparison by site was not conducted. Finally, administration times for whole versus crushed medications were not compared. In another study, administration time was shorter for whole tablets (45.2 seconds) than for crushed tablets (80.3 seconds).11 However, this limitation does not readily apply to chronic nitrate therapy, as topical forms are preferred in patients who have difficulty swallowing solid oral dosage forms.
Conclusion
The productivity of health care personnel is important to fiscal administrators and third-party payers. To our knowledge, this is the first study to provide a model for determining the cost of processing a single oral dose of a medication from the pharmacy to the patient via a 30-day bingo-card system and a 24-hour UD system. In this application of the cost model to solid oral nitrate therapy, the data demonstrate that, even with a higher initial AWP, the labor-time and cost-benefit of once-daily therapy equal or exceed those of two- and three-times daily brand-name therapies under the 30-day bingo-card system.
References
1. Finkel AS, Hammond JM, Karlin DS et al., eds. 1994 Redbook. Montvale, NJ: Medical Economics Data Production Company, 1994.
2. Parker JO, Farrell B, Lahey KA et al. Effect of intervals between doses on the development of tolerance to isosorbide dinitrate. N Engl J Med 1987; 316: 1440-4.
3. Parker JO. Eccentric dosing with isosorbide-5-mononitrate in angina pectoris. Am J Cardiol 1993; 72: 871-6.
4. Thadani U, Maranda C, Amsterdam E et al. Lack of pharmacologic tolerance and rebound angina pectoris during twice-daily therapy with isosorbide-5-mononitrate. Ann Intern Med 1994; 120: 353-9.
5. Davidson HE, Gonzalez ER, Chaponis RJ et al. Evaluation of chronic nitrate therapy in nursing facilities. Consult Pharm 1996; 11: 477-84.
6. Parker JO. New insights into nitrate tolerance. J Myocardial Ischemia 1994; 6: 10-6.
7. Berg JS, Dischler J, Wagner DJ et al. Medication compliance: a healthcare problem. Ann Pharmacother 1993; 27: S5-19.
8. Wisenberg G, Roks C, Nichol P et al. Sustained effect of and lack of development of tolerance to controlled-release isosorbide-5-mononitrate in chronic stable angina pectoris. Am J Cardiol 1989; 64: 569-76.
9. Rascati KL, Kimberlin CL, McCormick WC. Work measurement in pharmacy research. Am J Health-Syst Pharm 1986; 43: 2445-52.
10. Gonzalez ER, Hruska BM, Elswick RK et al. Time-motion study of intravenous ranitidine admixtures. Drug Intell Clin Pharm 1990; 24: 590-4.
11. Ranz TT, Reinhart SP, Kelly PM et al. Nursing labor associated with the administration of oral medications in long-term care. Consult Pharm 1995; 10: 1251-64.
The following descriptive statistics were computed for each time-motion analysis: (1) mean time (seconds); (2) standard deviation; (3) standard error of the mean; (4) 95% confidence interval; (5) sample size (n); and (6) weights. Because there were unequal and varying number of events (e.g., cards filled, doses administered) per observation, the descriptive statistics were calculated using formulas that weigh the data based on the number of events per observation. Thus, the average associated with a larger number of events was given more weight than the average associated with a smaller number of events. As we were interested in determining the precision of the mean, with which to infer time for drug distribution or administration in the population, we focused on the standard error of the mean.
The following narrative describes the formulas used to compute
the descriptive statistics. Let xij be the number of seconds in
the ith recording interval for the jth process and wij be the
associated weights (e.g., card counts or number of doses) for
i = 1, 2, ..., nj and j = 1, 2, ..., m. The average per dose (or
card) for each recording interval was expressed as
The overall mean was calculated using
where wj = w1j + w2j + ... + wnjj. Finally, using the recording interval means and overall mean, the variance, the standard deviation, and the standard error of the mean are, respectively,
To account for the multiple tasks required for drug dispensing
or administration, micro-analyses were performed on the total
time-motion assessment. For each dosing method (30-day and 24-hour),
the average cost of dosing a tablet was estimated by incorporating
the cost for each process. The descriptive statistics were modified
as follows. The mean for a process, 
was modified to include
the cost in cents per second as 
. The associated variance
is
where sj2 is the variance of process j.
For a particular dosing method, the cost estimate was based on the sum of costs of the processes or
The associated variance using the formula for a pooled variance is
The standard deviation and standard errors, respectively, are
