Use of conventional antipsychotics and the cost of treating schizophrenia.

This study documents the drug therapy patterns and 1-year treatment costs for 18,833 Medicaid patients with schizophrenia treated with conventional antipsychotic medications in Michigan, Kentucky, Alabama, and Georgia. One in four patients used no antipsychotic, but had total costs that were less than for treated patients (-$2,576, p < .0001); 18 percent of treated patients delayed therapy for at least 1 month and had significantly higher total costs of $3,994 (p < .0001); 41 percent of treated patients changed therapy with similar results (+$4,067, p < .0001). Only 20 percent of patients were compliant with drug therapy but this had no significant impact on total treatment costs.

Antipsychotic drug use patterns and the cost of treating schizophrenia.

This study investigated the relationships between antipsychotic drug use patterns and direct costs for 3,321 Medi-Cal patients with schizophrenia. Ordinary least-squares regression models were used to estimate the impact on costs of receiving antipsychotic drug treatment, delays in treatment, changes in therapy, and continuous therapy. Average costs were $25,940 per year per patient. Having used an antipsychotic drug was correlated with lower psychiatric hospital costs ($2,846 less) but higher nursing home costs. Completing one year of uninterrupted drug therapy was correlated with higher nursing home costs. Delayed drug treatment and changes in therapy increased the cost by $9,418 and $9,719, respectively.

The use of conventional antipsychotic medications for patients with schizophrenia in a medicaid population: therapeutic and cost outcomes over 2 years.

OBJECTIVE: To evaluate the association between drug therapy patterns achieved with conventional antipsychotics and direct healthcare costs over 2 years. METHODS: Paid claims data from the California Medicaid (Medi-Cal) program were used to identify 2476 patients with schizophrenia for whom 2 years of data were available. Ordinary least squares (OLS) regression models were used to estimate the association between lack of antipsychotic drug therapy, delayed therapy, changes in medications, and continuous therapy on healthcare costs over a 2-year period. RESULTS: Nearly 99% of Medi-Cal patients with schizophrenia were treated with conventional antipsychotics. Patients with schizophrenia consumed nearly $48,000 in direct costs over 2 years. Over 16% of patients did not use any antipsychotic medication for 2 years. Untreated patients used more healthcare resources than treated patients did ($10,833, P = .0422), especially psychiatric hospital care ($8,027, P = .0004). However, treated patients frequently experienced suboptimal drug use patterns. Nearly 33% of treated patients delayed antipsychotic therapy for up to 2 years. Delayed therapy was associated with increased costs of $12,285 (P = .070). Over 56% of patients experienced changes in therapy that were associated with higher total direct costs ($17,644, P < .0001). Finally, only 3.2% of treated patients used an antipsychotic medication consistently for 2 years. However, continuous drug therapy was not associated with lower costs. CONCLUSION: Suboptimal drug use patterns are common and costly in Medi-Cal patients with schizophrenia who initiated therapy with conventional antipsychotics.

Use patterns for antipsychotic medications in medicaid patients with schizophrenia.

OBJECTIVE: We investigated the use patterns for antipsychotic medications generated by Medicaid patients with schizophrenia. METHOD: Paid claims data from the California Medicaid program (Medi-Cal) were used to identify 2655 patients with schizophrenia. Data from 1987-1996 were used, during which time Medi-Cal maintained prior authorization restrictions on second generation antipsychotic drugs. Prescription records were used to identify 3 patterns of antipsychotic drug use: no drug therapy for over 1 year; delayed onset of antipsychotic drug therapy; and switches in antipsychotic drugs within 1 year. Multiple logistic regression models were used to identify factors affecting these antipsychotic drug use patterns. RESULTS: Conventional antipsychotic medications account for over 98% of all patient treatment episodes. Over 24% of patients with schizophrenia do not use any antipsychotic medication for periods lasting up to 1 year. Over 24% of treated patients delayed the use of antipsychotic medications at least 30 days. For those patients who did not delay their use of antipsychotic medications, over 47% switched or augmented their initial antipsychotic medication during the first treatment year. Only 11.6% of treated patients achieved 1 year of uninterrupted antipsychotic drug therapy. The mean duration of uninterrupted therapy was 142 days. DISCUSSION: Antipsychotic drug use patterns suggest that conventional antipsychotic medications do not meet the therapeutic needs of patients with schizophrenia.

Treatment cost of acute exacerbations of chronic bronchitis.

In 1994, the National Center for Health Statistics estimated that more than 14 million people (54 per thousand) had chronic bronchitis and sought treatment for 90.9% of their acute episodes. However, few studies have been done on the treatment cost of chronic bronchitis using national data. We conducted a retrospective analysis of claims for patients treated for acute exacerbations of chronic bronchitis (AECB) to assess the frequency of services rendered and the costs to the health care system. Records were selected for the study based on a primary diagnosis of AECB according to the International Classification of Diseases, Ninth Revision, code. Medicare was the primary source of data on patients aged > or =65 years; data from the National Healthcare and Cost Utilization Project, the National Ambulatory Medical Care Survey, and the National Hospital Ambulatory Medical Care Survey were used for patients aged <65 years. We calculated a total treatment cost of $1.2 billion for patients aged > or =65 years and $419 million for patients aged <65 years. These calculations were based on the following: 280,839 hospital discharges resulting in hospital costs of $1.1 billion for the 207,540 patients aged > or =65 years, and $408 million for the 73,299 patients aged <65 years. The mean hospital length of stay was 6.3 days with a mean cost of $5497 for patients aged > or =65 years and 5.8 days with a mean cost of $5561 for younger patients. Room and board represented the largest percentage of the mean hospital costs of AECB. Inpatient physician services cost $32 million and $11 million for the 2 age groups, respectively. Diagnosis-specific data for outpatient services were found to be less reliable than inpatient data, possibly due to diagnostic coding omissions; 331,000 outpatient office visits for AECB were found for those aged > or =65 years and 237,000 visits for those aged <65 years, resulting in respective total outpatient costs of $24.9 million and $15.1 million. If the number of outpatient visits remain consistent with 1994 levels, there would be 5.8 million visits annually for those aged > or =65 years and 4.2 million visits for those aged <65 years; total outpatient costs would be $452 million and $317 million, respectively. Because the treatment costs of AECB are largely the costs of hospitalization, any new therapy that allows more patients to be treated in the outpatient setting is likely to generate significant savings.

Compliance with sulfonylureas in a health maintenance organization: a pharmacy record-based study.

OBJECTIVE: To determine which factors affect compliance with sulfonylureas in a population served by a health maintenance organization in Southern California. METHODS: Retrospective analysis of pharmacy records and healthcare utilization data for two years (April 1993-March 1995), and a survey mailed to patients. Patients treated with sulfonylureas were selected for analysis on the basis of their prescription profile. Compliance was measured from the pharmacy records as the proportion of days the patient was in possession of the prescribed medications. Patient compliance with sulfonylureas was modeled as a function of four clusters of determinants: patient-related attributes, drug regimen characteristics and complexity, health status and disease-related variables, and characteristics of the interaction with healthcare providers. RESULTS: 786 patients were identified for analysis (49.1% women, mean age 59 y). The mean compliance rate was 83% +/- 22% SD. Compliance was significantly positively related with age and self-reported level of medication-taking compliance at baseline. Factors shown to have an inverse relationship with compliance were treatment complexity, perception of general health, and being a newly treated patient (adjusted R2 for the final model = 0.148). CONCLUSIONS: Our results suggest that factors found to be associated with noncompliant behavior (e.g., being a newly treated patient, self-reported compliance, regimen complexity) can be assessed by physicians and pharmacists as a routine practice.

Medication use in residential care facilities for the elderly.

OBJECTIVE: To describe medication use by residents of residential care facilities for the elderly (RCFEs). DESIGN: A cross-sectional survey of medication use. SETTING: Licensed, private RCFEs recruited from a roster of all licensed RCFEs in the Los Angeles area. SUBJECTS: Residents who were > or =60 years of age and whose medications were centrally stored in the facility. MEASURES: Age, gender, race, health insurance coverage, dietary restrictions, ambulation status, medical diagnoses, and medication profile. RESULTS: A total of 818 residents were surveyed. Residents were primarily white women who were >80 years. The average number of medications per resident was five; 94% of the sample took at least one medication. Cardiovascular drugs, central nervous system drugs, analgesics, diuretics, and potassium supplements were most commonly used. Use of multiple drugs within a therapeutic class was also common, with means ranging from 1.46 to 1.81 per resident for the most commonly prescribed classes. Diagnoses supporting the use of many medications were not documented in the residents' health records. CONCLUSIONS: This RCFE sample was medically frail and took many medications. The frequent use of cardiovascular medication reflected the prevalence of cardiac disease in the elderly. The frequency of psychotropic drug use without a corresponding indication suggested prescribing for symptoms rather than documented medical conditions. Lack of recorded diagnoses limited the ability to evaluate drug therapy. Improved record keeping; periodic medication review; and resident, staff, and prescriber education are necessary to ensure appropriate medication use in this setting.

Selective contracting and patient outcomes: a case study of formulary restrictions for selective serotonin reuptake inhibitor antidepressants.

OBJECTIVES: Many health maintenance organizations (HMOs) have selected 1 or 2 selective serotonin reuptake inhibitors (SSRIs) as their preferred drug for treating depression. This study investigated the effect of "single-drug" formulary restrictions on the likelihood of drug therapy completion for new patients, controlling for initial SSRIs used and other factors. METHODS: Prescription drug and medical record data for 187 patients who were newly prescribed SSRIs were retrieved from a single California group practice consisting of 22 board-certified primary care physicians. The group practice contracted with 2 independent practice association-model HMOs with different SSRI formulary restrictions. A multivariate analysis of drug therapy completion was conducted and 2 sensitivity analyses were performed. Completed therapy was based on the patient having achieved 6 months of uninterrupted therapy at a minimum therapeutic dose. RESULTS: Patients from the HMO with a single preferred SSRI (paroxetine) were 80% less likely to complete therapy than were patients from the HMO with 2 preferred SSRIs (fluoxetine and paroxetine) (odds ratio [OR] = 0.200, 95% confidence interval [CI] = 0.083-0.430). This formulary effect was independent of the initial drug used to treat the patient. Drug selection was also found to affect completion rates. Patients treated with sertraline were significantly less likely to complete therapy than were patients treated with fluoxetine (OR = 0.319, 95% CI = 0.105-0.968). Similar results were found for patients taking paroxetine relative to fluoxetine (OR = 0.357, 95% CI = 0.149-0.853). CONCLUSION: These results suggest that the use of single-product formularies may have unintended consequences on patient completion rates, independent of whether or not the most effective product is selected for preferred formulary status.

The role of SSRI antidepressants for treating depressed patients in the California Medicaid (Medi-Cal) program.

This research used paid claims data to investigate the likelihood that patients achieved an adequate course of antidepressant drug therapy and the impact of completed therapy on health care costs. Completed therapy was defined as six or more months of uninterrupted therapy at an adequate dose as determined by AHCPR treatment guidelines. Apparent average daily dose for each prescription filled was calculated from data on the prescription paid claim and allowances were made for titration of dose to therapeutic levels and changes in antidepressant therapy. A total of 1648 new episodes of antidepressant therapy were identified for analysis. The likelihood of achieving an adequate course of antidepressant therapy was 22%. Completion rates varied significantly across antidepressants with fluoxetine achieving the highest completion rate at nearly 51%. Total health care costs were significantly lower for patients who completed therapy (-dollar 1487; P = .0487) due primarily to lower ambulatory care costs (-dollar 1296; P = .0110). Fluoxetine was the only antidepressant therapy which exhibited significantly lower total health care cost per patient relative to the older tricyclic antidepressants (-dollar 3524; P = .0024). The total costs of treating depression in the ambulatory setting were found to vary widely across alternative antidepressants. Most of the cost-savings associated with fluoxetine use were associated with the increased likelihood of completed therapy. Further research is needed to verify if fluoxetine achieves better rates of completed therapy relative to other SSRI antidepressants using data from other settings.

The Kaiser Permanente/USC Patient Consultation Study: change in use and cost of health care services.

The impacts of three alternative models of pharmacist consultation on the use and cost of health care services were studied. Two studies were conducted concurrently in an HMO over two years. In one, 6000 patients were randomly assigned to one of three consultation models; in the other, the three models were implemented in six geographic regions of California (4600 patients). The models were (1) consultation about new or changed prescriptions as mandated by state law (state model), (2) consultation focused on selected high-risk ambulatory care patients (Kaiser Permanente [KP] model), and (3) a control model. The patients were surveyed three times about their health status and satisfaction, and computerized data on health care use and cost were collected. The effect of the consultation models on the use and cost of health care services was examined across five risk groups that were based on drug-use profiles. An additional 37,750 patients (10% of the patients residing in the areawide study sites) were included in a supplemental analysis of the use and cost of health care services. There was no indication in the random-assignment study that pharmacist consultations affected either drug costs or the cost of office visits. Similar results were found in the areawide study, with the exception that the KP model was associated with lower drug costs than the control model. In the 10% sample, the KP model appeared to be associated with lower office visit costs but higher drug costs. Both models were associated with a lower likelihood of a hospital admission and with lower total health care costs for some high-risk patients compared with the control model. Counseling patients about their medications may be unlikely to reduce medication costs or the cost of office visits but may reduce the likelihood of hospital admissions and the overall costs of health care services; a combination of counseling patients at high risk for drug-related problems and counseling all patients about any new or changed prescription should be considered.

The cost of treating community-acquired pneumonia.

Community-acquired pneumonia (CAP) is responsible for an average of 4.5 million visits annually to physicians' offices, emergency departments, and outpatient clinics. However, there have been few studies using national data on the costs of treating CAP. Without such data, it is difficult to assess whether new therapies and treatment strategies are needed to improve patient outcomes. We conducted a retrospective analysis based on national incidence data and paid claims data for patients treated for CAP to assess the frequency of services rendered and costs to the health-care system. Records were selected for the study based on a primary diagnosis of CAP according to the International Classification of Diseases, 9th Revision. Incidence data were derived from the National Health and Nutrition Examination Survey III. Medicare was the primary source of data for patients aged > or =65 years. Data from the National Healthcare Cost and Utilization Project, the National Ambulatory Medical Care Survey, and the National Hospital Ambulatory Medical Care Survey were used to determine the cost of treating patients aged <65 years. We arrived at a total cost of $4.8 billion for treating patients aged > or =65 years and $3.6 billion for treating patients aged <65 years. These calculations were based on the following: 1.1 million hospital discharges resulting in inpatient costs of $4.4 billion (52.4% of the $8.4 billion) for the 0.6 million patients aged > or =65 years and $3.1 billion (36.9% of the $8.4 billion) for the 0.5 million patients aged <65 years. The average hospital length of stay was 7.8 days with an average cost of $7166 for patients aged > or =65 years and 5.8 days with an average cost of $6042 for younger patients. Room and board represented the largest percentage of the average hospital bill for patients with CAP. Inpatient physician service costs were $305 million and $192 million for the > or =65 and <65 groups, respectively. Based on 1.1 million outpatient office visits for those aged > or =65 years and 3.3 million visits for those aged <65, total outpatient costs were $119 million and $266 million, respectively. Given the overwhelming cost burden for CAP in the hospital setting, any new therapy that allows patients to be treated in the outpatient setting could result in significant savings, especially for patients aged > or =65 years.

Pharmacoeconomics: what is it and where is it going?

The use of economic analyses as they pertain to the practice of medicine has become increasingly commonplace. Driven by both changes to reimbursement incentives and the clinical practice of medicine, these analyses have the potential to answer many of the questions posed by the delivery of health care in a fiscally responsible era. As such, the health economist is faced with the increasingly important role of providing data and insights key to the deliberative process of comparing the costs and benefits of medical care. To this end, health economists rely on a number of tools in the practice of pharmacoeconomics. The most common tool is a cost-effectiveness analysis, which compares the clinical effects of alternative therapies (mortality, morbidity) to their net costs. A relatively new tool to the field of clinical economics, the cost per quality-adjusted life-year has the potential of becoming the gold standard in clinical economics for measuring clinical effects if further refinement to measurement technique is achieved. All of these tools have as their outcome the development of a medical decision model or decision tree. These decision trees have four components: disease outcome states; the probability of transition between disease states; the costs, financial benefits, and effects incurred by patients in each outcome state; and risk groups. Data from a wide range of sources, including randomized clinical trials and retrospective analyses of data collected from clinical practice settings, are drawn upon in the development of decision trees. Although economists can help design these models and document the costs and effects of alternative therapies, it is physicians and consumers who will ultimately have to make the decisions that recognize that spending resources on one disease or risk group decreases the resources available to treat other diseases or risk groups.

Impact of pharmacist consultations provided to patients with diabetes on healthcare costs in a health maintenance organization.

We conducted a study to assess the impact on healthcare utilization and costs of pharmacist consultations provided to patients with diabetes. Data for this study were derived from a larger study conducted by Kaiser Permanente and the University of Southern California that evaluated three alternative models of pharmacist consultations (control, state, and Kaiser). Computerized data were available for patient demographic characteristics and healthcare utilization. We used medication data to classify patient cohorts as insulin only or oral antidiabetics +/- insulin. We estimated hospitalization costs based on diagnostic related group and medication costs based on average wholesale price; office visits were estimated at $70 each. In the insulin only cohort, total costs for patients who had their prescriptions filled at a state model pharmacy were 7.8% less than those for patients filling prescriptions at a control model pharmacy (P = 0.008). In the oral +/- insulin cohort, total costs for patients filling new prescriptions at a Kaiser model pharmacy were 21.9% less than those for patients using a control model pharmacy (P = 0.0001). The state model also was negatively correlated (beta coefficient, -0.0997) with total costs (P = 0.0001). These data suggest that pharmacist consultations provided to patients with diabetes can decrease total healthcare costs in a health maintenance organization.

The Kaiser Permanente/USC Patient Consultation Study: change in quality of life. University of Southern California.

The impacts of two models of pharmacist consultation on patient function and health-related quality of life (HRQOL) relative to a control model were studied. Patients in the random-assignment study and the areawide study of the Kaiser Permanente/USC Patient Consultation Study were surveyed three times over a two-year period. The patients were receiving pharmaceutical services under the Kaiser Permanente model of consultation (KP model), a state model of consultation, or a control model and were stratified according to prescription drug use. A global visual-analogue scale and Short Form-36 were used to assess HRQOL. In the areawide study, only 2 of 42 comparisons of the effects of the KP or state model on HRQOL were significant; both were associated with the state model and involved only small increases. In the random-assignment study, 3 of 21 estimated effects of the KP model on HRQOL were positive and significant; 2 of these were significantly different between the KP model and the state model. Although the KP and state models of consultation were associated with some changes in HRQOL, the overall influence was not consistent and not clinically important.

The Kaiser Permanente/USC Patient Consultation Study: patient satisfaction with pharmaceutical services. University of Southern California.

The impacts of two models of pharmacist consultation on patient satisfaction with pharmaceutical services relative to a control model were studied. Patients in the random-assignment and areawide studies of the Kaiser Permanente/USC Patient Consultation Study were asked to rate their satisfaction with six aspects of the service they received at the pharmacy. The patients were receiving pharmaceutical services under the Kaiser Permanente (KP) model of consultation, a state-mandated model, or a control model. They were assigned to risk categories according to their pattern of prescription drug use. Differences in satisfaction among the three consultation models for all patients and for each risk group and correlations between satisfaction items were determined. The KP and state models of consultation were associated with better satisfaction than the control model on five of the six satisfaction items. In the high-risk patient groups, the KP and state models were associated with higher levels of satisfaction than the control model. In the low-risk group, the state model was associated with higher levels of satisfaction than the control model; results for low-risk patients assigned to the KP model were inconsistent between the random-assignment study and the areawide study. Overall satisfaction appeared to be a suboptimal measure of satisfaction with consultation. Overall satisfaction was more closely related to satisfaction with waiting time and with whether the pharmacy staff was helpful and caring than with the three items reflecting satisfaction with the content of the consultation. Patients were more satisfied under the state model of pharmacist consultation or the KP model of consultation, which resulted in more-intensive counseling for fewer patients, than under a control model.

Factors affecting patient compliance with antihyperlipidemic medications in an HMO population.

OBJECTIVE: To identify factors that influence compliance in patients taking antihyperlipidemic medications. STUDY DESIGN: This was a retrospective cohort study in which computerized pharmacy records were used to estimate medication compliance in patients in a Health Maintenance Organization from 1993 to 1995. PATIENTS AND METHODS: Data on 772 patients on antihyperlipidemic medications were obtained from pharmacy and healthcare utilization claims and from a cross-sectional survey. The medication compliance ratio for each patient was calculated from the prescription profile. Patient compliance was modeled as a function of four clusters of determinants: patient characteristics, complexity of drug regimen, health status, and patient-provider interaction. Correlation between specific characteristics and compliance was estimated by logistic regressions. RESULTS: Approximately 37% of patients complied with at least 90% or more of their antihyperlipidemic medications. The following variables had a significant influence on compliance: female gender (odds ratio [OR], 0.64), baseline compliance (high: OR, 3.42; medium: OR, 1.86), perceived health status (SF-36 bodily pain score: OR, 1.02; SF-36 vitality score: OR, 0.97), comorbidity (OR, 0.90), and number of daily doses of antihyperlipidemic medications (OR, 0.60). CONCLUSIONS: The findings suggest that women, patients with comorbidities, patients reporting high SF-36 vitality scores, and patients with multiple doses of antihyperlipidemic medications are less likely to be compliant. Patients who self-report good compliance with previous medications are more likely to comply. This information may be used to target interventions at patients who are likely to be noncompliant with their medication regimens.

Measuring the impact of patient counseling in the outpatient pharmacy setting: the research design of the Kaiser Permanente/USC patient consultation study.

This article describes the research method used to measure the impact of three alternative models of patient counseling in the outpatient pharmacy setting. The study was conducted in pharmacies operated by the Southern California region Kaiser Permanente Medical Care Program. Both random assignment and large-scale geographic area research designs were used. The presentation of the research design includes discussions of data collection and patient sampling methods; the measurement of patient outcomes, including measures of health care costs and utilization, patient functional status, and quality of life. Demographic data are presented for the study population, including an analysis of potential biased selection of patients electing to participate in random assignment. Data are also presented documenting potential selection bias across geographically determined treatment groups in the geographic area design arm. Finally, the article presents the analysis plan for the study and discusses study limitations.