Impact of automation on pharmacist interventions and medication errors in a correctional health care system.

The implementation and impact of an automated check-and-sortation system in a correctional health care system are described. In February 1998 the Managed Health Care Division of the University of Houston College of Pharmacy submitted a proposal to the Texas State Board of Pharmacy for installing an automated check-and-sortation device in its distributive system serving Texas Department of Criminal Justice offenders. The new system would eliminate final visual verification of medication orders by pharmacists and thereby give them more time for in-depth reviews of patient profiles and for recommending clinical interventions. A vendor interested in developing a prototypic automated system was identified, and a disease management training course was developed to enable staff pharmacists to provide evidence-based pharmacotherapy recommendations. Twelve months of data before automation (February 1997 to January 1998) were compared with 16 months of postautomation data (June 1998 to September 1999). The number of clinical interventions increased from 396 per 100,000 medication orders filled before automation to 1075 per 100,000 orders after automation. Dispensing errors decreased from 6.3 to 4.1 per 100,000 medication orders filled. Identified filling errors increased from 224 to 256 per 100,000 orders. The reliability of the automated system was evidenced by a dispensing error rate of 0.6 per 100,000 orders. Implementation of an automated check-and-sortation device in a correctional health care system appeared to reduce dispensing errors and give pharmacists more time to review patient profiles and recommend clinical interventions.

Influence of patient information leaflets on anticonvulsant drug compliance in prison.

The effect of patient information leaflets on anticonvulsant drug compliance in a prison system was studied. Starting in January 1996, patient information leaflets were given to inmates in 18 prison units within the Texas Department of Criminal Justice who were receiving anticonvulsant medications. Nine of the units had pharmacist-operated chronic care clinics (POCs); the other nine did not. The units were matched on the basis of demographic characteristics and patient chronicity levels. Leaflet distribution continued until two weeks into March 1996. Certified medication aides (CMAs) watched inmates take doses and documented administration. A report on the anticonvulsant medications prescribed, dosage frequencies, number of patients taking each drug, and cumulative rate of compliance for each prison unit was generated monthly. The leaflets appeared to have a positive effect on compliance. The largest percent increases in compliance occurred on POC units, possibly because of the reinforcement provided by pharmacist counseling. The findings may have been affected by the integrity of the CMAs, the possibility that some patients on POCs did not receive pharmacist counseling, exploitation of the system by inmates for secondary gain, and other factors. Patient information leaflets may be of some use in increasing drug compliance in the prison population, but expectations must be tempered by the realities of this setting.

Clinical training program for distributive pharmacists.

A training program to enable staff pharmacists in a state correctional system to shift their practice from distributive to clinical is described. Under a contract with the Texas Department of Criminal Justice (TDCJ), the University of Houston College of Pharmacy provides pharmaceutical services to 145,000 patients in 100 correctional facilities. An automated centralized distribution site was built in 1993 and four regional pharmacies were subsequently closed, resulting in the reallocation of eight distributive pharmacists positions to clinical services. Four of the eight pharmacists who would be displaced began a retraining program in preparation for decentralized clinical positions. The 18-month program, developed by university staff on contract to TDCJ and college faculty members, included 34 home-study modules, case studies, case presentations, seminars, and supervised practice experience. Three of the pharmacists completed the program, passed competency tests, and were certified for clinical practice privileges, including drug therapy assessment and written consultations. The three pharmacists now operate ambulatory care clinics at prison units. A clinical skills training program developed by a college of pharmacy correctional services division in cooperation with faculty members enabled staff pharmacists to convert their roles from distributive to clinical.

Impact of pharmacist-operated general medicine chronic care refill clinics on practitioner time and quality of care.

OBJECTIVE: To assess the impact of pharmacist-operated refill clinics on practitioner (physician/physician assistant) time. Secondary objectives included assessment of disease state control, drug acquisition costs, and human resource allocation. DESIGN: The study ran from October 1, 1993 through January 31, 1994. At one clinic the pharmacist performed chart review only and did not see the patient. At the second site the pharmacist also interviewed the patient. The pharmacist used no treatment algorithms. Practitioner and pharmacist time before and after implementation of the refill clinics was measured on 3 consecutive days of normal clinic operation. SETTING: Two state correctional facilities. PATIENTS: The study population consisted of patients receiving chronic care who were observed during regular clinic hours. MAIN OUTCOME MEASURES: We evaluated pharmacist and practitioner time before and after pharmacist involvement. Hypertension was the only disease state yielding sufficient numbers for statistical analysis. We used a two-tailed paired t-test with the priori level set at 0.05. We also evaluated the number of patients with disease state control before and after clinic implementation, drug acquisition costs, and human resource allocation. RESULTS: The refill clinic pharmacist reduced practitioner time commitment in both clinics. The greatest impact on practitioner time was found in the interview clinic. Quality of care was maintained equally between the two clinics, with a positive impact on human resource allocation. CONCLUSIONS: In the managed care environment, pharmacist-run refill clinics can decrease practitioner time commitment allowing reallocation of human resources while maintaining current quality of care.

Accuracy of dispensing in a high-volume, hospital-based outpatient pharmacy.

Accuracy of dispensing was studied in the outpatient pharmacy setting, and error rates were compared with workload. All prescriptions filled in an outpatient pharmacy over 12 weekdays were audited to determine the rate of dispensing errors. In this pharmacy, pharmacists filled prescriptions and technicians delivered the medications to the patients. Before the medication reached the patient, the auditors recorded any dispensing errors and determined whether they were potentially serious. Of the 9846 prescriptions filled, 1229 (12.5%) contained a total of 1371 errors. Of these errors, 155 (1.6%) were potentially serious. Statistical analysis of the data revealed differences between error rates and (1) the total number of prescriptions dispensed per hour and (2) the number of prescriptions filled per pharmacist hour. However, no correlation existed between the number of prescriptions dispensed per hour and the total number of errors made. No significant correlation was found between the rate of potentially serious errors and increasing work volume, suggesting that important factors in error avoidance are continuous quality improvement mechanisms and minimal interruption of dispensing. No association was found between work volume and the number of dispensing errors or potentially serious errors. Error rates were consistent with published estimates.

Antiulcer prescribing program in a state correctional system.

OBJECTIVE: To describe a formulary antiulcer agent prescribing program developed as the result of a drug use evaluation (DUE). Program implementation, methods, cost impact, and results of a follow-up DUE are provided. BACKGROUND: The institution is a 51,000-bed correctional system consisting of 40 separate units each containing an ambulatory care clinic. Medication orders are transmitted via mainframe computer system to one of four pharmacies, which collectively dispense an average of 4000 medication orders (30 days' supply) per day. METHODS: Results from the antiulcer agent (cimetidine, ranitidine, sucralfate) DUE revealed that the agents studied were prescribed in dosages and durations exceeding criteria developed by the Pharmacy and Therapeutics Committee. A program designed to reduce dosages to maintenance therapy after eight weeks at treatment dosage was developed by the Pharmacy and Therapeutics Committee with staff physician input. Antiulcer agent use and expenditures were followed and a follow-up DUE was completed seven months after program implementation. Antacid use and frequency of upper gastrointestinal studies that were ordered were followed. RESULTS: The follow-up DUE showed the mean daily dosage for prescribed histamine2-receptor antagonists decreased (cimetidine from 694 to 454 mg, ranitidine from 280 to 183 mg) and the mean duration of therapy decreased from 14 to 10 months. The percentage of patients with potentially significant drug interactions decreased from 14.2 to 6.5 percent. The mean number of antiulcer agents dosage units dispensed per month decreased by 24,461 units, resulting in a projected annual savings of $327,273. There were no identifiable clinically important changes in the use of antacid products or prescribing of upper gastrointestinal studies. CONCLUSIONS: A cost-savings program sponsored by the pharmacy and therapeutics committee decreased costs, corrected prescribing to more closely meet preset criteria, and produced no discernable unfavorable effect on patient care.

Multidisciplinary program for detecting and evaluating adverse drug reactions.

An adverse drug reaction (ADR)-reporting program involving detection of charted ADRs by quality assurance nurses and data collection and causality assessment by staff pharmacists is described. The voluntary ADR-reporting mechanism used in a 900-bed, university-based hospital complex produced less than one ADR report per month. The newly implemented system depends on nurses to detect and report documented ADRs through concurrent chart review. Staff pharmacists are then responsible for follow-up chart review, data collection, and causality assignment based on two published algorithms. An inservice education program designed to increase the awareness and understanding of ADRs was provided to the department of pharmacy and the quality assurance nurses. The clinical staff provides quality assurance through weekly ADR committee meetings. Drug information center personnel complete the causality algorithms by using the data collected by the staff pharmacists. The ADR committee then compares the algorithm results of the two assessors. Discrepancies in scoring are evaluated to determine whether a change in the system is necessary. An FDA report is generated if the staff pharmacist assessor and the drug information center assessor obtain results of "probable" for both algorithms. An ADR-reporting program that relies on quality assurance nurses to detect charted ADRs and on staff pharmacists to evaluate reported ADRs increased the average number of ADRs reported from 0.4 to 20 per month.

Potential use of bar codes to implement automated dispensing quality assurance programs.

Bar code-based systems have automated many counting, tracking, and sorting functions in health care delivery services. We designed and briefly tested the feasibility of a bar code-based dispensing quality assurance system for a hospital outpatient pharmacy. The use of bar codes to verify the identity of the dispensed product required an extra few seconds processing time for each prescription but did not increase markedly the processing time when compared to a control period. In addition to verifying product identity, the system checked the manufacture's expiration date to reduce the risk of delivering outdated medications to the patient. The on-site test period for this feasibility model was relatively short (one week) and no actual dispensing errors were detected. However, when the system was presented with 100 different prescription forms containing 50 randomly sequenced, precontrived dispensing errors, the system identified all medication errors and outdated products. As shown in other studies, bar code-based systems have the potential to capture information not effectively recorded using manual methods. We suggest they should be considered by pharmacists interested in automating inventory management and work flow functions or establishing automated dispensing quality assurance programs.