Cost-effectiveness of an electronic medication ordering system (CPOE/CDSS) in hospitalized patients.

INTRODUCTION: Prescribing medication is an important aspect of almost all in-hospital treatment regimes. Besides their obviously beneficial effects, medicines can also cause adverse drug events (ADE), which increase morbidity, mortality and health care costs. Partially, these ADEs arise from medication errors, e.g. at the prescribing stage. ADEs caused by medication errors are preventable ADEs. Until now, medication ordering was primarily a paper-based process and consequently, it was error prone. Computerized Physician Order Entry, combined with basic Clinical Decision Support System (CPOE/CDSS) is considered to enhance patient safety. Limited information is available on the balance between the health gains and the costs that need to be invested in order to achieve these positive effects. Aim of this study was to study the balance between the effects and costs of CPOE/CDSS compared to the traditional paper-based medication ordering. METHODS: The economic evaluation was performed alongside a clinical study (interrupted time series design) on the effectiveness of CPOE/CDSS, including a cost minimization and a cost-effectiveness analysis. Data collection took place between 2005 and 2008. Analyses were performed from a hospital perspective. The study was performed in a general teaching hospital and a University Medical Centre on general internal medicine, gastroenterology and geriatric wards. Computerized Physician Order Entry, combined with basic Clinical Decision Support System (CPOE/CDSS) was compared to a traditional paper based system. All costs of both medication ordering systems are based on resources used and time invested. Prices were expressed in Euros (price level 2009). Effectiveness outcomes were medication errors and preventable adverse drug events. RESULTS: During the paper-based prescribing period 592 patients were included, and during the CPOE/CDSS period 603. Total costs of the paper-based system and CPOE/CDSS amounted to €12.37 and €14.91 per patient/day respectively. The Incremental Cost-Effectiveness Ratio (ICER) for medication errors was 3.54 and for preventable adverse drug events 322.70, indicating the extra amount (€) that has to be invested in order to prevent one medication error or one pADE. CONCLUSIONS: CPOE with basic CDSS contributes to a decreased risk of preventable harm. Overall, the extra costs of CPOE/CDSS needed to prevent one ME or one pADE seem to be acceptable.

A randomized, double-blind, placebo-controlled pilot study of i.v. immune globulins in combination with i.v. methylprednisolone in the treatment of relapses in patients with MS.

BACKGROUND: Some patients with multiple sclerosis (MS) do not show a clear improvement of acute relapses after treatment with intravenous methylprednisolone (IVMP). We compared the efficacy of the combination of intravenous immunoglobulins (IVIg) and IVMP with the standard treatment of IVMP alone in promoting recovery from moderate to severe acute relapses in MS. METHODS: Patients with clinically definite MS having a relapse with at least a one point increase in Kurtzke's expanded disability status scale (EDSS) in comparison to the preattack EDSS were randomized to IVMP-IVIg or IVMP-placebo treatment. The primary outcome criterion was the EDSS grade at four weeks. A preplanned interim analysis was performed after inclusion of 19 consecutive MS patients to evaluate the sample size necessary for a larger trial. FINDINGS: Both groups had improved one point on the EDSS four weeks after start of treatment (P = 0.81) and one of the stopping rules of the interim analysis was fulfilled. There were also no differences in secondary outcomes: EDSS at eight and 12 weeks, time to improve > or = 1 EDSS points, difference in Scripps score and ambulation index. Five patients in the IVMP-IVIg group and two in the IVMP group had a new relapse in the six month follow-up. INTERPRETATION: Our study could not show superiority of IVMP-IVIg in the treatment of moderate to severe acute relapses in MS.

Validity of hospital discharge International Classification of Diseases (ICD) codes for identifying patients with hyponatremia.

Medical diagnosis can be studied using various sources of information, such as medical and hospital discharge records and laboratory measurements. These sources do not always concur. The objective of the present study was to assess the sensitivity, specificity, and positive and negative predictive values of hospital discharge diagnosis compared with clinical laboratory data for the identification of hyponatremia. Patients with hyponatremia were selected from a hospital information system determined by the International Classification of Diseases, 9th edition (ICD-9). The validity parameters for hyponatremia (ICD code 276.1) were estimated by comparison with accurate serum sodium (Na+) levels. A total of 2632 cases of hyponatremia were identified using laboratory measurements (Na+ < or =135 mmol/L). The sensitivity of ICD coding for hyponatremia was maximally about 30% for patients with very severe hyponatremia (Na+ < or =115 mmol/L). Corresponding specificities were high (>99%). In 87% of the cases with severe hyponatremia (Na+ < or =125 mmol/L), other discharge ICD codes reflecting severe morbidity were found. This study suggests that ICD codes for hyponatremia represent only one third of the patients admitted to the hospital and experiencing hyponatremia. About two thirds of the patients with hyponatremia were classified as hospitalized for other reasons. To assess the validity of case finding of patients with hyponatremia, the use of analytical techniques, such as certain laboratory measurements, is advisable.

Selective serotonin reuptake inhibitor-induced urinary incontinence.

PURPOSE: Irrespective of its cause, urinary incontinence is a medical condition seriously affecting quality of life and is increasingly recognized. In this study, we examined the association between the use of selective serotonin reuptake inhibitors (SSRIs) and urinary incontinence. METHODS: A retrospective follow-up study among starters with an SSRI was performed to estimate the relative and absolute risk for urinary incontinence associated with SSRI use. Data came from the PHARMO database, which includes information on drug dispensing for approximately 450,000 residents living in eight Dutch cities. All patients initially using an SSRI between 1994 and 1998 were selected. The frequency measures for urinary incontinence were estimated by using prescription sequence analysis, where initiation of spasmolytic drugs or absorbent products was used as a measure for urinary incontinence. Besides crude incidence density calculations, Andersen-Gill's model was used in order to control for possible confounding factors and time varying covariates. RESULTS: A total of 13,531 were identified as first time users of an SSRI. Compared to non-exposure, the incidence density ratio for urinary incontinence during SSRI exposure was 1.75 (95% CI 1.56-1.97). Overall, compared to baseline, SSRI use caused 14 extra cases of urinary incontinence per 1000 patients treated per year; the elderly were more at risk resulting in 60 extra cases per 1000 patients per year. The adjusted relative risk for urinary incontinence due to SSRI use was 1.61 (95% CI 1.42-1.82); the risk for sertraline users was 2.76; 95% CI 1.47-5.21). CONCLUSIONS: Exposure to SSRIs is associated with an increased risk for developing urinary incontinence, which can be explained pharmacologically. Approximately 15 out of 1000 patients treated per year with an SSRI developed urinary incontinence. The elderly and users of sertraline are at the highest risk.

Serotonergic antidepressants associated with an increased risk for hyponatraemia in the elderly.

BACKGROUND: Hyponatraemia may have serious clinical consequences. Several reports of hyponatraemia associated with the use of antidepressants have been published. However, it remains unclear whether a specific class or individual antidepressants are associated with an increased risk for hyponatraemia. OBJECTIVES: To investigate the association between the use of serotonergic antidepressant drugs and the occurrence of hyponatraemia compared with non-users of these agents and to determine the time-to-admission rate after initiation of these drugs. METHOD: A matched case-control study was conducted. Data were obtained from the PHARMO database including information on drug dispensing and hospital admission indications for 320,000 inhabitants of eight Dutch cities. Data from 1990 to 1998 were used. Case patients ( n=203) were all patients who were admitted to a hospital for hyponatraemia. Community controls ( n=608), matched by age and gender, were sampled within the same living area and calendar (index) date as the case patients. All patients were 18 years of age or older. Exposure to antidepressant drugs, classified as serotonergic versus non-serotonergic agents, and potential confounding factors were determined on the index date. Time-to-admission was defined as the period between start of the antidepressant drug and hospital admission. Conditional logistic regression model was used to estimate odds ratios (ORs) and to adjust for potential confounding factors. RESULTS: Ten (5%) case patients used serotonergic antidepressants compared with eight (1%) in the control group; compared with non-use, the risk for hyponatraemia was fourfold higher [OR 3.96; 95% confidence interval (CI) 1.33, 11.83] due to serotonergic antidepressant drug use. Risk for developing hyponatraemia was greatest in the first 2 weeks of serotonergic drug therapy. CONCLUSION: Use of serotonergic antidepressants is associated with the development of hyponatraemia. Hyponatraemia occurred during the first 2 weeks of treatment, which justifies blood-sodium monitoring during the first weeks after initial treatment with a serotonergic antidepressant.

Assessment of indicators for hospital drug formulary non-adherence.

BACKGROUND: Translation of rational drug therapy into practice remains an international problem. Although pharmacotherapeutic treatment guidelines (PTGs) as managerial tools are favoured over hospital drug formularies (HDFs), the latter are still applied in most hospitals. HDF enforcement often leads to time-consuming consultation from the perspective of both pharmacy staff and prescriber. So far, research on HDFs has only been conducted outside Europe. Moreover, this research has only been descriptive. Straightforward indicators qualitatively characterising HDF non-adherence have never been assessed. METHODS: A retrospective 1:1 case-control study was conducted across three general teaching hospitals. Non-HDF requests were compared with HDF requests. Data were multivariably analysed, considering patient, prescriber, drug, and HDF characteristics as possible indicators for non-adherence. RESULTS: HDF adherence was almost universal across characteristics. Non-adherence was characterised by newly marketed drugs, drugs that were part of patients' pre-admission drug therapy, drugs with many fellow drugs within the drug group on the market, and drugs originating from a drug group for which the HDF was highly restrictive. Contrary to common perception, non-adherence was independent of medical specialty, therapeutic area, and patient characteristics. CONCLUSION: This research provides an epidemiological framework for hospitals (drug and therapeutics committees) for evaluating pharmacy data on HDF non-adherence. It can be used for educational tailor-made feedback to prescribers and for drug selection when the inclusion of newly marketed drugs is considered or HDF restrictiveness for certain drug groups is reconsidered. Moreover, it demonstrates the importance of a regional approach involving secondary and primary health care to establish continuity in seamless care of drug therapy.

Selective prescribing of spasmolytics.

BACKGROUND: Daily clinical practice often differs largely from the clinical trial setting, so extrapolation of outcomes from trial data, such as safety, effectiveness, and economic outcomes, can be deceptive. Prescribers may intend to treat a selected group of patients with new drugs; this practice could result in significant bias in assessing outcomes of these agents during their use in daily clinical practice. OBJECTIVE: To evaluate what type of patient received tolterodine compared with the spasmolytic drugs previously marketed (oxybutynin, flavoxate, emepronium). DESIGN: An observational, follow-up study. SETTING: Eighteen collaborating community pharmacies. PATIENTS: Aged > or = 18 years, noninstitutionalized; initial therapy with tolterodine, oxybutynin, flavoxate, or emepronium. RESULTS: Tolterodine was often used as a second-line and even as a third-line treatment, and was prescribed to a "polluted" population in terms of concomitant psychotropic medication. Tolterodine users were 7.5 times more likely to have received another spasmolytic drug (RR 7.5, 95% CI 4.8 to 11.9). In addition, these patients more frequently used antiparkinsonian drugs (RR 4.1, 95% CI 1.6 to 10.4) as well as antipsychotic drugs (RR 2.9, 95% CI 1.4 to 6.2). There was a small difference in concomitant use of antidepressants and benzodiazepines between patients receiving tolterodine versus those taking other spasmolytic drugs. CONCLUSIONS: Tolterodine is prescribed for a population differing from that receiving previously marketed spasmolytic drugs. Selective prescribing should recognized when evaluating new drugs in daily clinical practice. Policy makers, such as pharmacy and therapeutics committees, should consider this aspect in their formulary decisions since selective prescribing can lead to unjustified conclusions about a drug's therapeutic effects (e.g., efficacy, safety, cost-effectiveness).

Risk factors for the development of adverse drug events in hospitalized patients.

Adverse drug events in hospitalized patients lead to increased morbidity, mortality and costs. Early detection of adverse drug events could aid in the prevention of these adverse outcomes. A cost-effective system for the early detection of adverse drug events should focus on high risk patients. A study was set up with the primary aim to identify characteristics that are associated with the development of adverse drug events (ADEs) in hospitalized patients. ADE reports were gathered from physicians and nurses (spontaneous reports) and from patients after intensive ward interviews by hospital pharmacists. All patients admitted to the internal medicine wards of two Dutch hospitals, during a two month period, were included. The following characteristics were analyzed for their potential relationship to the occurrence of ADEs: age (categorized), gender, number of drugs prescribed during hospital stay, types of drugs used and changes in drug use on admission. Age was found to be inversely associated with the development of ADEs (OR 0.36, CI 0.21-0.61 for age category > 80 years; OR 0.56; CI 0.31-1.02 for age category 75-80 years and OR 0.69; CI 0.42-1.11 for age category 60-74 years). Furthermore, statistically significant associations were found for the number of drugs prescribed per hospitalized patient (for the class of 4-6 drugs per patient OR 2.61, CI 1.32-5.18), for newly prescribed drugs (OR 6.65, CI 2.63-16.81) and for the cessation of drugs on hospital admission (OR 1.50, CI 1.02-2.20). The use of gastrointestinal drugs (OR 2.13, CI 1.32-3.45), central nervous system drugs (OR 1.66, CI 1.07-2.57) and antibiotics (OR 2.44, CI 1.65-3.60) were associated with the development of ADEs, when compared to all other drugs taken by the patients. In this study, the most important risk factors are the number of drugs used per patient and the starting of a new drug during hospitalization. As most hospitalized patients start new drug therapies while in hospital, this seems an inappropriate focus. However, careful monitoring of patients using more than 7 drugs at a time may be possible in a cost-effective system for the early detection of ADEs.

Adverse drug events in hospitalized patients. A comparison of doctors, nurses and patients as sources of reports.

OBJECTIVE: This study investigated the relative value of adverse drug events reported by doctors, nurses and patients. METHODS: The study was conducted on a total of four wards: the paediatric and internal medicine wards (including geriatric patients) of two peripheral hospitals in The Netherlands. Adverse drug events were collected by spontaneous reporting (doctor and nurse reports) and by daily ward visits, during which the patients were interviewed by a hospital pharmacist (patient reports). Criteria for relative value of the reported adverse drug events were the number of potentially serious reactions, the number of reactions not mentioned in the patient information leaflet and the number of reactions reported to new drugs (5 years or less on the Dutch market). No formal causality assessment was applied. RESULTS: Over a period of 2 months in 1996 (Hospital I) and 2 months in 1997 (Hospital II) a total of 620 patients were included in the study and adverse drug events were reported in 179 (29%) of these cases. Doctors reported a statistically significant larger number of serious (26% of all doctor reports; odds ratio (OR) 3.2; confidence interval (CI) 1.2-8.7) and unknown (39%; OR 2.5; CI 1.0-6.0) adverse drug events than patients themselves during the daily ward visit. Doctors also reported more serious and unknown adverse drug events than nurses. Adverse reactions to new drugs were reported during the daily ward visit only (8% of all daily ward visit reports). CONCLUSION: This study reconfirms that doctors are the main source for reports of serious and unknown adverse drug events in hospitalized patients. However, patients themselves seem to report more adverse reactions to new drugs (during the daily ward visit). By focusing on patients using new drugs, the daily ward visit might become cost-effective. This needs to be explored in future studies.

Incidence of discontinuation of highly active antiretroviral combination therapy (HAART) and its determinants.

OBJECTIVE: To determine the incidence and determinants for discontinuation of initial highly active antiretroviral therapy (HAART). DESIGN: In this retrospective follow-up study from hospital files and pharmacy dispensing data, a standard dataset was collected including patient characteristics, therapy characteristics, and HIV-monitoring parameters (e.g., CD4+ lymphocyte counts, viral load determinations). Kaplan-Meier estimates of the cumulative probability of discontinuation of initial HAART were calculated. Cox proportional hazard analysis was used to identify determinants for discontinuation of initial HAART. PATIENTS: All patients starting HAART (n = 99) during June 1996 to February 1997 at our regional AIDS center. MAIN OUTCOME MEASURES: Incidence and determinants for discontinuation of HAART. RESULTS: During the mean follow-up of 450+/-10 days, 27 patients switched initial HAART, 3 patients stopped any antiretroviral therapy. Reasons for switching were increasing viral load (18x), insufficient decrease of viral load (3x), and adverse events (6x). Nonnaivete for antiretroviral therapy and a lower CD4+ lymphocyte count at start were identified as determinants for discontinuation of initial HAART. CONCLUSIONS: The overall incidence density for discontinuation of initial HAART was 25 per 100 patients/year. The main reason for switching was an increasing viral load. CD4+ lymphocyte counts at start and nonnaivete for antiretroviral therapy were identified as determinants for discontinuation.

Channeling of three newly introduced antidepressants to patients not responding satisfactorily to previous treatment.

The demand for knowledge about differences in effectiveness, tolerability, safety, and economic outcomes between and within groups of antidepressant drugs when used in routine daily clinical practice is growing. For gaining this knowledge, observational pharmacoepidemiologic studies are often the most feasible option. However, the results of such studies can only be valid if either patient baseline characteristics associated with the outcome under study are similar or if differences can be adjusted for in the analysis. The aim of this study was to evaluate to what extent and for what type of patients three antidepressant drugs recently introduced in The Netherlands (mirtazapine, sertraline, and venlafaxine) were prescribed during the first year after their introduction and whether there were differences compared with longer-available antidepressant drugs. For this purpose, prescription drugs histories from 20 pharmacies serving a population of approximately 200,000 persons were analyzed. One year after their introduction, the newly introduced antidepressant t drugs accounted for approximately 6% of new uses of all antidepressant drugs. In comparison to longer-available antidepressant, the newly introduced antidepressant drugs were more often prescribed for patients with prior prescriptions of another antidepressant drug (rate ratio [RR] 2.7 [95% confidence interval [CI], 2.3-3.0]), for patients with prior prescriptions of other psychotropic medicines (RR 1.3 [95% CI, 1.1-1.4), and by psychiatrists (RR 1.9 [95% CI, 1.6-2.2]). In addition, the newly introduced antidepressant drugs seemed to be more often, although not significantly, prescribed for patients who had been hospitalized on a psychiatric ward (RR 1.5 [95% CI, 0.9-2.5]). No differences were observed with regard to age and gender distribution, the total number of medicines prescribed, and prescriptions of any cardiovascular or gastrointestinal medicine. These finding suggest that a significant proportion of the patients receiving one of the newly introduced antidepressant drugs did not respond satisfactorily to previous pharmacologic treatment. This channeling phenomenon may have important consequences for the interpretation of observational comparisons between different antidepressant drugs after their introduction.

Tetracaine versus lidocaine-prilocaine for preventing venipuncture-induced pain in children.

The efficacy of tetracaine cream versus that of lidocaine-prilocaine cream for the prevention of pain in children undergoing venipuncture was studied. Hospital inpatients 1-15 years of age received, on the back of each hand, a 30-minute application of tetracaine 4% cream or a 60-minute application of lidocaine-prilocaine cream (EMLA, Astra) before undergoing scheduled venipuncture. The phlebotomists in this open, randomized trial evaluated the efficacy of the cream at the moment of venipuncture as adequate, inadequate, or inconclusive. Blood samples were taken immediately after venipuncture from 10 patients one to five years of age to measure the serum concentrations of tetracaine and its metabolite, N-butyl-p-aminobenzoic acid. Lidocaine-prilocaine cream was significantly more efficacious in preventing pain than tetracaine 4% cream (97% of the former group [n = 32] had adequate pain relief, compared with 76% of the latter [n = 34]. The only adverse effects observed were mild local erythema in the tetracaine group and local skin blanching in the lidocaine-prilocaine group. No tetracaine could be detected in serum, and the serum concentrations of N-butyl-p-aminobenzoic acid ranged from 0 to 1.8 mg/l. Statistically, lidocaine-prilocaine cream was more efficacious than tetracaine 4% cream, but the difference is of minor clinical significance and is outweighed by the practical advantages of tetracaine 4% cream, namely the shorter application time, vasodilation and lower cost.

Pentamidine aerosols and environmental contamination: health-care workers at risk.

In our hospital safety guidelines are available for the handling of pentamidine in the day-care department, but no safety ventilation cabin is used because only one patient a day has been treated, The number of patients to be treated, however, is growing, resulting in the need to treat more than one patient a day. To determine the environmental contamination and exposure of health-care workers during and after aerosolised pentamidine treatment of more than one patient with the acquired immunodeficiency syndrome a day a high-performance liquid chromatographic method is used for the detection and quantification of pentamidine. High volume air samples were taken before, immediately after and the day after a treatment session of up to three patients. Also, sediment samples and personal air samples close to the mouth of the health-care workers were taken. Immediately after a treatment session the air in the room contains 1.0-99.7 micrograms pentamidine per m3 of air. Before and the morning after treatment no pentamidine could be detected in the air. Sediment samples vary in detectable amounts of pentamidine from < 5 to 1165 micrograms. pentamidine/cm2. The personal air samples also show a large variation in quantities of pentamidine: < 5-170 ng a filter. When large amounts of pentamidine in the high volume air samples are found high amounts of pentamidine on the sediment samples and the personal air samples are found as well. This means that the patients treated should be instructed well on how to use the nebulizer correctly and be monitored during treatment. Additional safety measures (for example the use of a safety ventilation cabin) should be taken when more than one patient is treated a day.

Single-dose v. short-term antibiotic therapy for prevention of wound infection in general surgery. A prospective, randomized double-blind trial.

To investigate the effectiveness of a single-dose antibiotic regimen for preventing postoperative wound infection, a prospective, randomized double-blind trial was carried out in patients undergoing "clean-contaminated", "contaminated" or "clean" (vascular) surgery. Both elective and emergency operations were included. Single-dose (preoperative) prophylaxis was compared with short-term prophylaxis (1 dose preoperatively and 2 doses postoperatively). The antibiotics were penicillin, tobramycin and metronidazole in various combinations, and comparisons between single-dose and short-term prophylaxis were made with all the regimens. The incidence of wound infection was 5/277 (1.8%) in the short-term group and 9/287 (3.1%) in the single-dose group. The difference was not statistically significant. Nor was statistically significant difference found when the type of operation and the degree of contamination were considered. Single-dose antibiotic prophylaxis thus gave a low incidence of postoperative wound infection, even in "clean-contaminated" or "contaminated" cases.