The cost of pneumonia after acute stroke

OBJECTIVE

To determine the incremental costs of pneumonia occurring during hospitalization for stroke.

METHODS

We reviewed hospital records of all Medicare patients admitted for ischemic or hemorrhagic stroke to 29 hospitals in a large metropolitan area, 1991 through 1997, excluding those who died or had do not resuscitate orders written within 3 days of admission. Hospital costs of patients with stroke were determined using Medicare Provider Analysis and Review data after adjustment for baseline factors affecting cost and propensity for pneumonia. Secondary analyses examined the risk-adjusted relationship of pneumonia to discharge disposition.

RESULTS

Pneumonia occurred in 5.6% (635/11,286) of patients with stroke, and was more common among patients admitted from nursing homes and those with greater severity of illness (p < 0.001). Mean adjusted costs of hospitalization for patients with stroke with pneumonia were $21,043 (95% CI $19,698 to 22,387) and were $6,206 (95% CI $6,150 to 6,262) for patients without pneumonia, resulting in an incremental cost of $14,836 (95% CI $14,436 to 15,236). Patients with pneumonia were over 70% more likely to be discharged with requirements for extended care (adjusted OR 1.73, 95% CI 1.32 to 2.26).

CONCLUSION

Extrapolated to the over 500,000 similar patients hospitalized for stroke in the United States, the annual cost of pneumonia as a complication after acute stroke is approximately $459 million.

The effect of pneumonia on mortality among patients hospitalized for acute stroke

OBJECTIVE

To determine the effect of pneumonia on 30-day mortality in patients hospitalized for acute stroke.

METHODS

Subjects in the initial cohort were 14,293 Medicare patients admitted for stroke to 29 greater Cleveland hospitals between 1991 and 1997. The relative risk (RR) of pneumonia for 30-day mortality was determined in a final cohort (n = 11,286) that excluded patients dying or having a do not resuscitate order within 3 days of admission. Clinical data were obtained from chart abstraction and were merged with Medicare Provider Analysis and Review files to obtain deaths within 30 days. A predicted-mortality model (c-statistic = 0.78) and propensity score for pneumonia (c-statistic = 0.83) were used for risk adjustment in logistic regression analyses.

RESULTS

Pneumonia was identified in 6.9% (n = 985) of all patients and in 5.6% (n = 635) of the final cohort. The rates of pneumonia were higher in patients with greater stroke severity and features indicating general frailty. Unadjusted 30-day mortality rates were six times higher for patients with pneumonia than for those without (26.9% vs 4.4%, p < 0.001). After adjusting for admission severity and propensity for pneumonia, RR of pneumonia for 30-day death was 2.99 (95% CI 2.44 to 3.66), and population attributable risk was 10.0%.

CONCLUSION

In this large community-wide study of stroke outcomes, pneumonia conferred a threefold increased risk of 30-day death, adding impetus to efforts to identify and reduce the risk of pneumonia in patients with stroke.

Managed care for the Medicaid disabled: effect on utilization and costs

The objective of this study was to describe the effect on health care utilization and costs of a program of managed care for the Medicaid disabled. The study was designed as a pre/post enrollment cohort comparison and was carried out in three Ohio counties. The subjects were disabled Medicaid-insured patients who voluntarily enrolled in a managed care program for at least 6 months between July 1, 1995 and December 31, 1997, and who had (1) at least one Medicaid claim in the 24-months pre-enrollment period and (2) overall satisfactory postenrollment encounter-level data. Ohio Medicaid provided claims and reimbursements (costs) for the pre-enrollment period and encounter-level data for the postenrollment period. Postenrollment costs were estimated by applying category-specific average pre-enrollment costs to postenrollment utilization data. We measured the following per patient-month: (1) trends in category-specific utilization and costs for up to 24 months before and after enrollment, (2) differences in overall and category-specific costs 1 year before and after enrollment, and (3) changes in the distribution of services 1 year before and after enrollment. Utilization categories included inpatient care, outpatient hospital (including emergency department) care, physician services, prescription medications, durable medical equipment and supplies, and home health care. We found that satisfactory encounter data were available in two of three counties. Of 1,179 enrollees, 592 met all inclusion criteria. Before enrollment, utilization and costs were increasing significantly in four of six categories and were unchanging in two. Postenrollment, decreasing utilization was observed for three categories, one remained unchanged, and two were increasing, but from a lower "baseline." Except for physician services and home health care, there were lower utilization and estimated costs in all categories in the year after enrollment. Estimated inpatient and total costs declined by $155/patient-month (44.9%) and $210/patient-month (37.1%), respectively. Findings were similar across sites. Inpatient care, outpatient hospital care, and prescription medications accounted for 97% of the reductions in estimated costs in the postenrollment period. Among patients voluntarily enrolled for at least 6 months, managed care for the Medicaid disabled was associated with striking decreases in health care utilization and estimated costs. The effect of managed care on these patients' satisfaction, access to specialized services, quality of care, and health outcomes are understood incompletely.

Comparison of risk-adjustment systems for the medicaid-eligible disabled population

OBJECTIVE

The objective of this study was to compare 2 approaches for subjecting capitation rates for disabled Medicaid-eligible patients in managed care plans to risk adjustment, the Disability Payment System (DPS) and the Ohio Prior Expenditure System (OPES).

DESIGN

This was a retrospective cohort.

SETTING AND SUBJECTS

The subjects were 157,142 nonelderly disabled individuals eligible for > or =1 month during state fiscal year 1995 (SFY95) for a 3-county Ohio Medicaid managed care demonstration project.

DATA SOURCE

Data were from the Ohio Medicaid eligibility and fee-for-service claims files.

ANALYSIS

As per OPES policy, individuals were classified by the duration of their eligibility in SFY93 as "old" eligibles (> or =6 months) or "new" eligibles (<6 months). Published relative payment weights for each system were adjusted and used to predict SFY95 expenditures in a budget-neutral comparison. Measures were variance in SFY95 expenditures explained by predicted payments (R2) and predictive ratios (predicted payment/actual SFY95 expenditure). Individuals with HIV/AIDS and hematological conditions, who enrolled disproportionately across the demonstration counties, were analyzed separately.

RESULTS

Of the 157,142 individuals, 56.4% were new eligibles; 40.1% of the old eligibles had no claims-documented chronic disease diagnosis in the baseline year. The overall R2 was 0.091 with OPES and 0.057 with DPS. Neither system predicted >1% of individual-level expenditures for new eligibles. OPES severely underpaid for eligibles in the top percentile of predicted expenditures; DPS had mixed results. DPS predicted SFY95 expenditures substantially better than OPES for the enrollment bias categories.

CONCLUSIONS

Before Medicaid programs move to full-risk capitation for disabled populations, better risk-adjustment methods are needed, especially for eligible patients with little claims experience, high predicted expenditures, or enrollment-bias conditions.

The association between hospital type and mortality and length of stay: a study of 16.9 million hospitalized Medicare beneficiaries

OBJECTIVES

To examine the association between hospital type and mortality and length of stay using hospitalized Medicare beneficiaries for a 10-year period.

METHODS

The retrospective cohort study included 16.9 million hospitalized Medicare beneficiaries > or = 65 years of age admitted for 10 common medical conditions and 10 common surgical procedures from 1984 to 1993. A total of 5,127 acute-care hospitals in the United States were grouped into 6 mutually exclusive hospital types based on teaching status and financial structure (for-profit [FP], not-for-profit [NFP], osteopathic [OSTEO], public [PUB], teaching not-for-profit [TNFP], and teaching public [TPUB]) as reported in the 1988 American Hospital Association database. Logistic and linear regression methods were used to examine risk-adjusted 30-day and 6-month mortality and length of stay.

RESULTS

During the 10-year study period, 10.6 million patients were admitted with 1 of the 10 selected medical conditions, and 6.3 million patients were hospitalized for 1 of the 10 selected surgical procedures. Patients at TNFP hospitals had significantly lower risk-adjusted 30-day mortality rates than patients at other hospital types when all diagnoses or procedures were combined (combined diagnoses: RR(TNFP) = 1.00 [reference], RR(TPUB) = 1.40, RR(OSTEO) = 1.14, RR(PUB) = 1.07, RR(FP) = 1.03, RR(NFP) = 1.02; combined procedures: RR(TNFP) = 1.00 [reference], RR(OSTEO) = 1.36, RR(TPUB) = 1.30, RR(PUB) = 1.16, RR(FP) = 1.13, RR(NFP) = 1.08). The results were mostly consistent when diagnoses and procedures were examined separately. After adjustment for patient characteristics, patients at other hospital types had 10% to 20% shorter lengths of stay (LOS) than patients at TNFP hospitals for most diagnoses and procedures studied.

CONCLUSION

As measured by the risk-adjusted 30-day mortality, TNFP hospitals had an overall better performance than other hospital types. However, patients at TNFP hospitals had relatively longer LOS than patients at other hospital types, perhaps reflecting the medical education and research activities found at teaching institutions. Future research should examine the empirical evidence to help elucidate the adequate LOS for a given condition or procedure while maintaining the quality of care.

Rapid growth of rehabilitation services in traditional community-based nursing homes

OBJECTIVE

To describe the changes in rehabilitation therapy services in nursing homes based in the community during a period of rapid escalation of Medicare payments to nursing homes.

SETTING

All Medicaid-certified nursing homes in Ohio.

SUBJECTS

The 52,705 residents newly admitted to nursing homes in 1994 and 1995.

DESIGN

Retrospective trend analysis of administrative data.

MAIN OUTCOME MEASURES

For newly admitted residents receiving 90 or more minutes of rehabilitation therapy per week, the trends in percentage and in the amount and type of therapy received were determined for eight quarters.

RESULTS

Of all newly admitted residents, 50.5% received 90 or more minutes of therapy. When they received such therapy it averaged 412 minutes per week (SD = 259). Those residents who received rehabilitation services increased by 2.2% each quarter (p<.001), and the amount of therapy they received increased by 6.4 minutes each quarter (p<.0001). All three types of rehabilitation therapy-physical, occupational, and speech-increased (p<.015) over the study period.

CONCLUSIONS

The traditional nursing home is an important site for the provision of rehabilitation therapy services. Rehabilitation specialists should be aware of these trends as they plan for the future of rehabilitation. The effectiveness of this increased provision of therapy service in terms of measurable outcomes needs to be evaluated.

Patient satisfaction with time spent with their physician

BACKGROUND

We examined the variables related to patient satisfaction with the time spent with their family physician during the office visit.

METHODS

Research nurses directly observed consecutive patient visits to 138 family physicians in 84 practices. Analyses examined sequential models of the association of patient and physician characteristics, visit type and length, and time use during visits, with patients' satisfaction with the amount of time spent with their physician.

RESULTS

Among 2315 visit by adult patients returning questionnaires, patient satisfaction with the time spent with their physician was high and strongly linked to longer visits (P < 001). After controlling for visit duration, greater patient satisfaction with time spent was associated with older patient age, white race, better perceived health status, visits for well care, and visits with a greater proportion of the visit spent chatting. The physician's discussion of test results or findings from the physical examination was associated with greater satisfaction with time spent for visits longer than 15 minutes, but with less satisfaction with time spent for shorter visits.

CONCLUSIONS

Physicians can enhance patient satisfaction with the amount of time spent during an office visit by spending a small proportion of time chatting about nonmedical topics, and by allowing sufficient time for exchange with the patient is feedback is necessary.

Indications, outcomes, and provider volumes for carotid endarterectomy

CONTEXT

While trials have demonstrated that carotid endarterectomy is superior to best medical therapy, most recently among asymptomatic patients, uses and outcomes of the procedure in more representative settings have not been established.

OBJECTIVES

To profile the use and outcomes of carotid endarterectomy in a representative sample of Ohio's Medicare beneficiaries and to examine the relationships between provider-specific procedural volumes and patient outcomes.

DESIGN

Retrospective cohort using Medicare Provider Analysis and Review files supplemented by detailed reviews of medical records on a random sample of patients.

SETTING

Ohio hospitals performing carotid endarterectomy.

PATIENTS

A random sample of 678 charts of the 4120 non-health maintenance organization Medicare beneficiaries who underwent carotid endarterectomy between July 1, 1993, and June 30, 1994.

MAIN OUTCOME MEASURES

Nonfatal stroke or death within 30 days of surgery.

RESULTS

The reviewed patients were similar to all eligible patients in sociodemographic characteristics and 30-day mortality rates. Among the 678 patients, indications for surgery were asymptomatic carotid stenosis in 167 (24.6%), transient ischemic attack in 294 (43.4%), completed stroke in 62 (9.1%), and nonspecific symptoms in 155 (22.9%). Thirty-two patients (4.7%) died or suffered nonfatal strokes by 30 days postoperatively. In univariate analyses, rates varied by hospital volume (P=.004) but not surgeons' volume (P=.47), although power to detect this difference was limited. Patients at higher- and lower-volume hospitals had similar indications and distributions of comorbidities. In analyses controlling for indications, comorbid conditions, and surgeon's volume, being operated on in a higher-volume hospital conferred a 71% reduction in risk for 30-day stroke or death (odds ratio, 0.29; 95% confidence interval, 0.12-0.69; P=.006).

CONCLUSIONS

Almost half (47.5%) of the carotid endarterectomies among Ohio's Medicare population are performed on persons who are asymptomatic or who have nonspecific symptoms. These results highlight the importance of identifying patients and providers having the most favorable outcome profiles. The higher rate of adverse outcomes observed in lower-volume hospitals deserves further investigation, as it does not appear to be due to differences in patient selection.

Barriers to adequate delivery of hemodialysis

Mortality rates among American hemodialysis patients are the highest in the industrialized world. Measures of delivered dialysis (Kt/V) correspond strongly with survival and are estimated to be inadequate in one third of patients. We sought to determine the importance of potential barriers to adequate dialysis, including patient-related and technical factors. Using a cross-sectional study design, we abstracted the charts of 721 randomly selected patients from all 22 chronic hemodialysis units in northeast Ohio. For each of 1,836 treatments provided to these patients, we assessed delivered dialysis (Kt/V) and patient-related factors (ie, hypotension, intradialytic symptoms, and treatment time missed due to noncompliance or transportation problems) and technical factors (ie, dialysis prescription, type of vascular access, clotting, and dialyzer reuse). We used hierarchical regression analysis to determine which potential barriers were independently related to delivered dialysis after adjustment for patient demographic and medical characteristics. Barriers independently related to dialysis delivery (all P values < 0.001) included patient noncompliance, present in 3% of treatments; low dialysis prescription, 14%; use of a catheter for vascular access, 11%; and clotting, 1%. The prevalence of identified barriers varied dramatically across facilities (eg, the prevalence of low dialysis prescription ranged from 0% to 37%, while the prevalence of catheter use ranged from 3% to 28%). In conclusion, patient noncompliance, low dialysis prescription, catheter use, and clotting are the most important barriers to dialysis delivery. Further work is needed to develop interventions to overcome these barriers and to determine the effect of such interventions on dialysis adequacy and patient survival.

Atrial fibrillation as a risk factor for stroke: a retrospective cohort study of hospitalized Medicare beneficiaries

OBJECTIVES

This study examined the relationship between atrial fibrillation and (1) stroke and (2) all-cause mortality.

METHODS

All eligible Medicare patients older than 65 years of age hospitalized in 1985 were followed up for 4 years. Kaplan-Meier and Cox proportional hazards models were used for assessment of risk of stroke and mortality.

RESULTS

A total of 4,282,607 eligible Medicare patients were hospitalized in 1985. The mean age was 76.1 (+/- 7.7) years; 58.7% were female; 7.2% were Black; and 8.4% had a diagnosis of atrial fibrillation. During the follow-up period, 66,063 patients (32.6/1000 person-years) developed nonembolic stroke and 7285 (3.6/1000 person-years) developed embolic stroke. After adjustment for age, race, sex, and comorbid conditions, atrial fibrillation remained a significant risk factor for both nonembolic stroke (relative risk [RR] = 1.56) and embolic stroke (RR = 5.80) and for mortality (RR = 1.31). Approximately 4.5% of nonembolic and 28.7% of embolic strokes among hospitalized Medicare patients aged 65 years and older were attributable to atrial fibrillation.

CONCLUSIONS

This study demonstrates that atrial fibrillation is associated with an appreciable increase in the risk of stroke (both embolic and nonembolic) and in the risk of mortality from all causes.

Failure to prescribe warfarin to patients with nonrheumatic atrial fibrillation

OBJECTIVE

To determine how often warfarin was prescribed to patients with nonrheumatic atrial fibrillation in our community in 1992 when randomized trials had demonstrated that warfarin could prevent stroke with little increase in the rate of hemorrhage, and to determine whether warfarin was prescribed less frequently to older patients-the patients at highest risk of stroke but of most concern to physicians in terms of the safety of warfarin.

DESIGN

Cross-sectional study. Appropriateness of warfarin was classified for each patient based on the independent judgments of three physicians applying relevant evidence and guidelines.

SETTING

Two teaching hospitals and five community-based practices.

PATIENTS

Consecutive patients with nonrheumatic atrial fibrillation (n = 189).

MEASUREMENTS AND MAIN RESULTS

Warfarin was prescribed to 44 (23%) of the 189 patients. Warfarin was judged appropriate in 98 patients (52%), of whom 36 (37%) were prescribed warfarin. Warfarin was prescribed to 11 (14%) of 76 patients aged 75 years or older with hypertension, diabetes mellitus, or past stroke, the group at highest risk of stroke. In a multivariable logistic regression model controlling for appropriateness of warfarin and other patient characteristics, patients aged 75 years or older were less likely than younger patients to be treated with warfarin (odds ratio 0.25; 95% confidence interval 0.10, 0.65).

CONCLUSIONS

Warfarin was prescribed infrequently to these patients with nonrheumatic atrial fibrillation, especially the older patients and even the patients for whom warfarin was judged appropriate. These findings indicate a substantial opportunity to prevent stroke.

Effect of a nurse case manager on postdischarge follow-up

OBJECTIVE

To examine whether use of a nurse case manager to coordinate postdischarge care would improve rates of follow-up, emergency department utilization, and unexpected readmission for general medicine patients.

DESIGN

Prospective cohort trial.

SETTING

Publicly supported, tertiary-care teaching hospital.

PATIENTS

Four hundred seventy-eight patients admitted to the general medicine service.

INTERVENTIONS

Use of a nurse case manager to provide discharge planning before hospital discharge and to arrange for postdischarge outpatient follow-up. Patients in the control group had discharge planning in the traditional ("usual care") manner.

MEASUREMENTS AND MAIN RESULTS

The proportion of patients with scheduled outpatient appointments in the medical clinic and the proportion making clinic visits, emergency department visits, or with readmission to the hospital within 30 days following discharge. A significantly greater proportion of patients assigned to the nurse case manager intervention had appointments scheduled at the time of hospital discharge (63% vs 46%, p < .001), and made scheduled visits in the outpatient clinic (32% vs 23%, p < .03). Intervention group patients were especially more likely than control group patients to have definite follow-up appointments if they were discharged on weekends. Intervention and control group patients did not differ, however, in the rates of emergency department utilization (p = .52) or unexpected readmissions within 30 days of discharge (p = .11).

CONCLUSIONS

Use of a nurse case manager to coordinate outpatient follow-up prior to discharge improved the continuity of outpatient care for patients on a general medical service. The intervention had no effect on unexpected readmissions or emergency department utilization.

Atrial fibrillation: restoration and maintenance of sinus rhythm and indications for anticoagulation therapy

PURPOSE

To review the efficacy and safety of electrical and pharmacologic conversion of atrial fibrillation, strategies for maintenance of sinus rhythm, and the importance of antithrombotic therapy.

DATA SOURCES

English-language trials were identified from the MEDLINE database through 1995, bibliographic references, Current Contents, textbooks, and recent abstracts.

STUDY SELECTION

Randomized trials (including abstracts) were selected. Cohort studies were used if randomized trials were not available.

DATA EXTRACTION

Study design and data were extracted from clinical trials. Statistical analysis of combined data was not appropriate, given the marked variations in study designs and study populations.

DATA SYNTHESIS

Cardioversion restores sinus rhythm in more than 80% of patients. In atrial fibrillation of recent onset, pharmacologic regiments have a success rate of 40% to 90%. Sinus rhythm at 1 year is maintained in 30% of patients without antiarrhythmic therapy but in 50% of patients with such therapy. The efficacy and safety of antiarrhythmic drugs relative to one another are not established because trials done to date have been small and cases vary. Successful cardioversion and maintenance of sinus rhythm are most predictable when the duration of atrial fibrillation is brief. Warfarin reduces the incidence of ischemic strokes and emboli in patients with nonvalvular atrial fibrillation from 4.5% to 1.4% per year. Aspirin (325 mg/d) appears to be equally effective in patients younger than 75 years of age who do not have hypertension or a history of thromboembolism or recent heart failure. Although warfarin is more effective than aspirin in preventing embolic strokes in patients older than 75 years of age, it may increase the incidence of hemorrhagic stroke and result in a similar rate of disabling stroke.

CONCLUSION

Cardioversion remains the preferred method with which to re-establish sinus rhythm. Long-term antiarrhythmic therapy is warranted only if recurrences or initial clinical instability are seen; pro-arrhythmic concerns and potential side effects should guide drug selection. Antithrombotic therapy is indicated for all patients older than 60 years of age and in all patients younger than 60 years of age who have clinical evidence of a primary cardiac disorder.

A cost analysis of alternative treatments for duodenal ulcer

OBJECTIVE

To compare the costs of alternative strategies for the treatment of duodenal ulcer.

DESIGN

A cost comparison using decision analysis.

METHODS

A decision model was used to compare the costs per cure of an endoscopically documented duodenal ulcer for three initial treatment strategies: 1) H2-receptor antagonist therapy for 8 weeks, 2) antibiotic therapy for Helicobacter pylori infection plus H2-receptor antagonist therapy, and 3) urease test-based treatment. For symptomatic recurrences, secondary treatment strategies included empiric retreatment with the same or other regimen, and treatment based on repeat endoscopy-guided urease test or biopsy, with an assumption of subsequent cure. The cohort modeled for this analysis consisted of patients at low risk for a malignant ulcer. Probability estimates were derived from published clinical trials, cohort studies, and expert opinion. Side effects from combination therapy with antibiotics and H2-receptor antagonists and resulting costs were included from the perspective of a group practice model health maintenance organization.

RESULTS

For all secondary treatment strategies, initial therapy with antibiotics for H. pylori infection plus an H2-receptor antagonist resulted in the lowest average costs per symptomatic cure when the prevalence or likelihood of H. pylori infection exceeded 66% to 76%; the costs ranged from $284 for secondary (re)treatment with empiric antibiotic and H2-receptor antagonist therapy to $398 for endoscopy-guided secondary treatment. Initial treatment with an H2-receptor antagonist resulted in the highest costs, ranging from $372 for secondary treatment with empiric antibiotic and H2-receptor antagonist therapy to $679 for endoscopy-guided secondary treatment. The results were not sensitive to the rates of duodenal ulcer recurrence after either treatment, to the cost of either treatment, or to prevalence of H. pylori.

CONCLUSIONS

This cost analysis indicates that, regardless of the secondary treatment used for ulcer recurrence, initial therapy with antibiotics for H. pylori infection plus an H2-receptor antagonist provides the lowest costs per symptomatic cure. These cost savings and the lower recurrence rates associated with this treatment favor eradication of H. pylori as part of the initial treatment of duodenal ulcer.

Improving compliance with breast cancer screening in older women. Results of a randomized controlled trial

BACKGROUND

To compare three approaches for improving compliance with breast cancer screening in older women.

METHODS

Randomized controlled trial using three parallel group practices at a public hospital. Subjects included women aged 65 years and older (n = 803) who were seen by residents (n = 66) attending the ambulatory clinic from October 1, 1989, through March 31, 1990. All provider groups received intensive education in breast cancer screening. The control group received no further intervention. Staff in the second group offered education to patients at their visit. In addition, flowsheets were used in the "Prevention Team" group and staff had their tasks redefined to facilitate compliance.

RESULTS

Medical records were reviewed to determine documented offering/receipt of clinical breast examination and mammography. A subgroup of women without previous clinical breast examination (n = 540) and without previous mammography (n = 471) were analyzed to determine the effect of the intervention. During the intervention period, women without a previous clinical breast examination were offered an examination significantly more often in the Prevention Team group than in the control group, adjusting for age, race, and comorbidity and for physicians' gender and training level. The patients in the Prevention Team group were offered clinical breast examination (31.5%) more frequently than those in the patient education or control groups, but this was not significant after adjusting for the above covariates. Likewise, mammography was offered more frequently to patients in the Prevention Team and in the patient education group than to patients in the control group, after adjusting for the factors above using logistic regression.

CONCLUSIONS

The results provide support for patient education and organizational changes that involve nonphysician personnel to enhance breast cancer screening among older women, particularly those without previous screening.

You can lead a horse to water--improving physicians' knowledge of probabilities may not affect their decisions

OBJECTIVES

To determine whether improving physicians' judgments of the probability of streptococcal pharyngitis for patients with sore throats would affect their use of antibiotics and affect the variation in such use.

DESIGN

Post-hoc retrospective analysis of data previously collected as part of a controlled trial.

SETTINGS

University student health services in Pennsylvania and Nebraska.

PATIENTS

Sequential patients with pharyngitis seen before and after the time clinicians received either an experimental educational intervention designed to improve probabilistic diagnostic judgments (at the Pennsylvania site) or a control intervention, a standard lecture (at the Nebraska site). The clinician-subjects were the primary case physicians practicing at either site.

MEASUREMENTS

Clinical variables prospectively recorded by the clinicians, probability assessments, and treatment decisions.

RESULTS

At the experimental site, despite marked decreases in clinicians' overestimations of disease probability after the intervention, the proportion of patients prescribed antibiotics showed a trend toward increasing: 100/290 (34.5%) pre-intervention, 90/225 (40%) post-intervention. The intervention did not decrease practice variation between individual doctors. Univariable and multivariable analyses showed no major change in the relationships between clinical variables and treatment decisions after the intervention. At the control site there was no major change in probability judgments or treatment decisions after the intervention.

CONCLUSIONS

Teaching physicians to make better judgments of disease probability may not alter their treatment decisions.

Limitations of transesophageal echocardiography in the risk assessment of patients before nonanticoagulated cardioversion from atrial fibrillation and flutter: an analysis of pooled trials

Recent studies have proposed that the exclusion of an atrial thrombus by transesophageal echocardiography (TEE) would allow for the safe cardioversion from atrial fibrillation or flutter without the need of prophylactic anticoagulation. Because all of the TEE trials have been small and descriptive and have lacked randomized, conventionally treated control groups, the pooled risk of embolic events from TEE trials was compared with that of a control group pooled from the literature on cardioversion both with and without conventional anticoagulation. Studies were identified from a MEDLINE search, references in review articles, and recent cardiology abstracts and were included if there were > 10 patients and if atrial fibrillation or flutter was of > 48 hours' duration. Where > 1 study had been published by the same group only the largest study was used. Studies were not selected by cause of arrhythmia, by predisposing risk factors for atrial fibrillation and flutter, or by method of cardioversion. The only patients excluded from TEE reports were those with atrial thrombi diagnosed on the precardioversion TEE or those documented to have adequate standard precardioversion anticoagulation. Seven TEE and 18 control studies met the inclusion criteria. More patients in the control studies had rheumatic valvular disease. Embolic events were significantly more frequent in the TEE group than in the anticoagulated control group (1.34% vs 0.33%, respectively; p = 0.04), whereas there was no significant difference between the TEE group and the nonanticoagulated control group (2.00%; p = 0.26). Thus the use of TEE screening to exclude patients with atrial thrombi before cardioversion does not identify patients who can safely undergo this procedure without anticoagulation.

Improving compliance with immunization in the older adult: results of a randomized cohort study

OBJECTIVE

To compare three approaches for improving compliance with influenza and pneumococcal vaccination of elderly patients.

DESIGN

Randomized controlled trial using three parallel group practices at a public urban teaching hospital.

SETTING

Public teaching hospital.

SUBJECTS

All patients 65 years of age and older (n = 1202) seen by resident physicians (n = 66) attending three ambulatory medical practices from October 1, 1989 to March 31, 1990.

INTERVENTIONS

All three provider groups received intensive education in immunization standards. The control group received no further intervention. Staff in the second group offered education to patients at their visits. In the third group, the prevention team, a flowsheet was used, patient education offered, and staff had their tasks redefined to facilitate compliance; for vaccinations, eg, nurses could vaccinate independent of MD initiative.

MEASUREMENTS AND MAIN RESULTS

Medical records were reviewed for the 1202 patients seen, including 756 patients seen during both the 1988-89 and 1989-90 influenza seasons, to determine documented offering and receipt of vaccinations. During the intervention period (1989-90), influenza vaccinations were offered significantly more frequently to prevention team patients (68.3%) than to patients in either the patient education (50.4%) or control (47.6%) groups (P = 0.006), even after adjusting for the patients' prior vaccination status, age, gender, race, and high-risk co-morbidity and for physicians' level of training. Likewise, pneumococcal vaccinations were offered more frequently to previously unvaccinated prevention team patients (28.3%) than to patient education (6.5%) or control (5.4%) group patients (P = 0.001), even after adjusting for the factors using multivariate analysis. Compliance rates did not differ between patient education and control subjects for either vaccine. Pre-intervention physician surveys documented higher perceived than actual compliance for both vaccines, with 89.0% and 52.8% of physicians believing that they complied with influenza and pneumococcal vaccination guidelines, respectively.

CONCLUSIONS

The results of this trial provide strong support for organizational changes that involve non-physician personnel to enhance vaccination rates among older adults.

Practice variation in the management of pharyngitis: the importance of variability in patients' clinical characteristics and in physicians' responses to them

The objective of this study was to assess whether geographic differences in antibiotic-prescribing rates for patients with pharyngitis could be explained by intersite differences in patients' clinical characteristics and in how physicians responded to these clinical cues when making decisions. As part of the initial phase of a prospective controlled trial to improve physicians' diagnostic ability, the authors enrolled cohorts of consecutive patients seen at staff-model--HMO student health services in Pennsylvania and Nebraska. Physicians' decisions whether to prescribe antibiotics for 310 consecutive patients presenting with pharyngitis to the former and 214 such patients presenting to the latter at the time of the initial visit were examined. There was a large discrepancy between the antibiotic-prescribing rates at the student health services in Pennsylvania, 106/310, 32.4%, and Nebraska, 156/214, 72.9%. The clinical variables significantly independently associated with treatment at both sites in a logistic regression model were fever, adjusted odds ratio = 2.1 (95% CI = 1.1, 3.8); exudates, 5.4 (2.8, 10); palatine petechiae, 6.5 (1.5, 28); rhinorrhea, 0.46, (0.25, 0.85); and high risk of complications, 3.8 (1.04, 14). There was a significant interaction between site and anterior cervical adenopathy, 5.5 (1.6, 19); and a borderline interaction between site and rhinorrhea, 2.4 (0.89, 6.7). Site was not a significant independent predictor of treatment, 1.8 (0.45, 6.6.). Practice variation was related to geographic differences in patients' clinical characteristics and in how physicians responded to these factors when prescribing antibiotics. How physicians weight patients' clinical characteristics when making decisions may be an important element of their "practice styles."

Randomized, controlled trials using the Metro Firm System

The Firm System at MetroHealth Medical Center was begun almost two decades ago to foster improved continuity of patient care and teaching of medical students and residents in Internal Medicine. For the past 8 years, these parallel teams of providers and patients also have been used to conduct clinical, educational, and health care delivery research. Randomized, controlled trials are made possible by ongoing random assignment of patients and providers to the three teams, or small group practices. Each group practice has equivalent inpatient and outpatient areas supported by nonrotating nursing, paramedical, and clerical staff. The system's current relationships were established after a controlled trial established both decreased costs and increased effectiveness of the "group practice model" as compared to more traditional approaches to patient care by residents in an academic medical center. Other trials, both on the inpatient and outpatient settings, have been used to guide ongoing institutional change. The unique advantages of the randomized controlled trial are high-lighted by contrasting the results of within-group changes during an intervention with results that incorporate control group changes. A variety of methodologic and logistical issues must be addressed when conducting controlled trials that use ongoing randomization within a single institution. These include determination that the groups are equivalent for all important parameters preintervention, choosing an analytic approach that accounts for potential differences among providers and patients, and, in trials designed to affect behavior, assurance that a "steady state" exists prior to initiating another trial designed to affect similar behavior. Consideration also must be given to the possibilities of cross-team contamination, the Hawthorne effect, the "dominant attending effect," and ethical issues related to informed consent. Clinical trials in a single institution may be performed for common problems and those with well defined stages of severity. Since the costs of maintaining the ongoing randomization are relatively small, the incremental costs of conducting randomized controlled trials are low by comparison to traditional approaches. The establishment of analogous systems at other institutions will permit multisite trials for less common clinical problems, enhance the generalizability of findings, and permit new types of health services research, including the rigorous evaluation of practice guidelines.

Pursuit of abnormal coagulation screening tests generates modest hidden preoperative costs

To measure the follow-up costs of preoperative coagulation screening tests, the authors studied 829 consecutive patients undergoing inpatient orthopedic surgery. The results of the initial prothrombin and activated partial thromboplastin time tests were divided into three groups: normal; abnormal above the hospital laboratory's upper limit of normal but below an "action limit"; and abnormal above an action limit. Patients with abnormal preoperative coagulation screening test results were matched on the basis of operative procedure and age with patients who had normal results. The matched groups of patients were compared according to preoperative length of stay and the cost of subsequent related preoperative testing. The average cost of follow-up preoperative testing for patients with abnormal screening test results was $5.05, compared with $0.58 for patients with normal screening results. The difference in average preoperative lengths of stay was not statistically significant. The attributable cost of evaluating an abnormal result added 3% to the cost of the initial coagulation screening program. This represents an average preoperative cost of $0.36 per patient in addition to the cost of the screening tests themselves.

Carotid endarterectomy

STUDY OBJECTIVE

To evaluate the usefulness of doing carotid endarterectomy in patients with symptomatic or asymptomatic carotid artery disease.

DESIGN

Synthesis and summarization of data on the untreated course of cerebrovascular disease, and review of selected evidence and expert opinion on the risks and benefits of endarterectomy and medical therapy.

MAIN RESULTS

There is a large body of data related to the untreated course of cerebrovascular disease, the efficacy of aspirin in patients with transient ischemic attacks, and the comparative responsiveness to surgery of symptomatic patients with different presentations. Randomized trials in progress will increase the knowledge about the effects of aspirin in asymptomatic patients, the comparative efficacy of aspirin and endarterectomy in asymptomatic and symptomatic patients, and the factors that influence surgical risk. Clinicians and investigators wish to define a more limited and precise set of indications for carotid endarterectomy. In setting standards, attention has properly focussed on "maximum acceptable complication rates," native stroke risk, and surgical efficacy for patients with different clinical presentations. Illustrative, acceptable surgical mortality rates are less than 1%, and stroke-related morbidity is less than 3% for patients who have had a transient ischemic attack; surgical mortality and stroke-related morbidity are less than 2% for patients with asymptomatic carotid stenosis. Medical comorbidity and angiographic findings are important factors affecting specific recommendations.

CONCLUSIONS

Data from ongoing clinical trials will provide a more scientific foundation for recommendations about when to do carotid endarterectomy. Current evidence suggests that fewer endarterectomies should be done, and those on the basis of a more precise set of indications. Clinicians should consider the clinical presentation, the patient's comorbidity and native stroke risk, angiographic findings, and the experience of the surgical team.

Evaluating physicians' probabilistic judgments

Physicians increasingly are challenged to make probabilistic judgments quantitatively. Their ability to make such judgments may be directly linked to the quality of care they provide. Many methods are available to evaluate these judgments. Graphic means of assessment include the calibration curve, covariance graph, and receiver operating characteristic (ROC) curve. Statistical tools can measure the significance of departures from ideal calibration, and measure the area under ROC curve. Modeling the calibration curve using linear or logistic regression provides another method to assess probabilistic judgments, although these may be limited by failure of the data to meet the model's assumptions. Scoring rules provide indices of overall judgmental performance, although their reliability is difficult to gauge for small sample sizes. Decompositions of scoring rules separate judgmental performance into functional components. The authors provide preliminary guidelines for choosing methods for specific research in this area.

Guiding individual decisions: a randomized, controlled trial of decision analysis

In early 1983, all 1,280 faculty and resident physicians at one hospital who were eligible to be vaccinated against hepatitis B were divided randomly into three groups: Group 1 physicians received general information about the risks and benefits of alternative vaccine decisions; Group 2 physicians were additionally invited to provide personal information for an individualized decision analysis (12.6 percent responded); and Group 3 physicians, who served as controls, were not contacted. In one year's follow-up, 20 percent of physicians were screened for hepatitis B antibody or vaccinated. More Group 2 physicians whose decision analyses recommended screening or vaccination took these actions (39 percent) than any other group. Group assignment remained significantly associated with vaccine decisions after analyzing results by the "intention to treat" principle, and after adjusting for training status, exposure to blood and blood products, and pre-study intentions about the vaccine. Despite the low overall vaccine acceptance rate, it is concluded that individualized decision analysis can influence the clinical decisions taken by knowledgeable and interested patients.

The importance of considering single testing when two tests are available

When considering two dichotomous tests in combination for reaching a treatment decision, the choice between single and multiple testing depends, in part, on the pretest probability of disease. The authors show that two tests are never preferred to a single test for all disease probabilities, regardless of whether the two tests are performed in parallel or in series.

Biochemical profiles. Applications in ambulatory screening and preadmission testing of adults

Evidence bearing on the utility of biochemical profiles for ambulatory care screening and preadmission testing was reviewed, and recommendations were formulated. Certain biostatistical principles relevant to this analysis include the concepts of regression to the mean, the meaning of biochemical "normality" (particularly as it relates to multiple testing), and the effects of disease prevalence on the post-test probability of disease for individual test results. Applying these principles to a typical 12-test battery in asymptomatic adults, one would expect a low yield of unanticipated diagnoses and a high proportion of false-positive results. For the typical test battery, the empirical evidence supports these expectations in ambulatory care screening. In addition, routine preadmission use of biochemical profiles is not supported by studies examining their impact on patient care, hospital costs, or lengths of stay. Whereas components of these profiles can be endorsed for screening or preadmission testing, we recommend that the routine use of the entire profile be abandoned in these settings.

The comparative cost-effectiveness of statistical decision rules and experienced physicians in pharyngitis management

We examined whether probability-based decisions for streptococcal pharyngitis, using probabilities derived from predictive models along with Tompkins' decision rules, could be more cost-effective than the actual decisions of ten physicians. We retrospectively calculated the probability of a positive throat culture ("disease") for each of 310 patients using four different models based on discriminant analysis (1), a branching algorithm (2), and logistic regression (3 and 4). "Projected decisions" were based on these probabilities and Tompkins' rules. We calculated direct medical and indirect costs per correct action taken (diseased patient-treated or nondiseased patient-not-treated). Two models' projected decisions were more cost-effective than the physicians'. Model 1 primarily would have reduced treatment costs (leaving no diseased patient untreated); model 4 primarily would have reduced throat culture costs (with 15% projected undertreatment). While using statistical decision rules may be cost-effective in this setting, their adoption should be consistent with physician and patient priorities.

The importance of disease prevalence in transporting clinical prediction rules. The case of streptococcal pharyngitis

Because clinical prediction rules often are applied in new settings to calculate the probability of a disease, we evaluated the accuracy of three rules for predicting streptococcal pharyngitis in 310 patients. Use of the rules led to overestimations of disease probability in 47%, 82%, and 93% of the patients. When we used receiver-operating characteristic curve analysis, no rule lost power to discriminate streptococcal from nonstreptococcal causes of pharyngitis. The overestimations in disease probability likely were caused by differences in disease prevalence between our setting (5%) and the settings in which they were developed (15% to 17%). All rules led to accurate predictions when they were adjusted for the disease prevalence found in our setting using a likelihood ratio formulation of Bayes' theorem. The value of prediction rules, like that of other diagnostic tests, is affected by differences in disease prevalence in different settings. Failure to recognize and adjust for these differences may cause poor decision making or the premature dismissal of valid rules.

Utility of rheumatoid factor in the diagnosis of juvenile rheumatoid arthritis

Rheumatoid factor is commonly used by clinicians to assess children with possible juvenile rheumatoid arthritis. To assess its usefulness, we reviewed the case histories of patients in whom latex agglutinating rheumatoid factor was determined during 1981 to 1982 at our institution. A total of 437 charts were available for review. There were 11 patients with positive tests for rheumatoid factor, of whom five had juvenile rheumatoid arthritis, all polyarticular. A total of 426 children had negative results, of whom 100 had juvenile rheumatoid arthritis. This yields a sensitivity of 4.8% and a specificity of 98%. We then estimated the prevalence of juvenile rheumatoid arthritis in three clinical settings: a primary practitioner's office, a tertiary children's hospital walk-in clinic, and a pediatric rheumatology center. The predictive values and marginal benefits for rheumatoid factor were then calculated in those settings using Bayes' theorem. In the two general outpatient settings, the primary practitioner's office and tertiary walk-in clinic, the positive predictive values were 0.7% and 0.5%, respectively; marginal benefits were 0.4% and 0.3%, respectively. Rheumatoid factor testing appeared to be of some benefit in the pediatric rheumatology center with a positive predictive value of 72.5% and marginal benefit of 22.5%. In no case was rheumatoid factor testing helpful in establishing a diagnosis of juvenile rheumatoid arthritis or in ruling it out. Testing for rheumatoid factor is a poor screening procedure for juvenile rheumatoid arthritis in the general situations in which it is more likely to be requested and of supportive diagnostic value only in the highly restricted population of older children with polyarticular arthritis.

Multiple test analyzer (MTA): a microcomputer program for determining preferred strategies with two diagnostic tests

There are ten distinct management strategies in clinical situations that involve two diagnostic tests with dichotomous outcomes. The authors describe a microcomputer program, based on a previously described model, that can be used to identify test and test-treatment thresholds and to compute preferred strategies. The program provides tables and graphs of the results, which can be viewed or printed, and there is an optimization routine that facilitates comprehensive analysis. It can be used by decision-analytic researchers and policy analysts, medical educators who teach decision analysis, and clinicians who use decision analysis in their practices.

Acute cholecystitis. Evaluation of factors influencing common duct exploration

The clinical records of 216 patients with proven acute cholecystitis treated by cholecystectomy form the basis of this retrospective study. Common bile duct stones were present in 12.4 per cent of these patients. Thirty per cent of the patients with elevated SGOT values, 26.2 per cent of the patients with elevated alkaline phosphatase, and 23.1 per cent of the patients with elevated amylase had common duct stones. The authors found that 17.6 per cent of patients with bilirubin between 1.5 and 2.9 mg/dl had common duct stones, whereas 71.4 per cent of common bile ducts with a bilirubin greater than 5 mg/dl contained stones. Six of 28 patients with common duct stones had normal bilirubin. Cholangiograms were normal in 115 of the 154 cholangiograms performed; six of these common ducts were explored, and no common duct stones found (false-negative cholangiograms 0.0%). Cholangiograms showed stones in 24 patients; common bile duct stones were recovered from 20 of these patients (accuracy rate 83%, false-positive cholangiograms 17%). Wound infections occurred in seven patients (3.7%). Sepsis resulted in death of three patients, and the other two deaths resulted from multi-system failure. This study demonstrates operative cholangiograms to be the most accurate method of detection of common duct stones, and its routine use in patients undergoing cholecystectomy is recommended.

Clinical guidelines for using two dichotomous tests

Building on the threshold model developed by Pauker and Kassirer for a single test, the authors describe a decision analytic model for two tests with dichotomous outcomes. The model includes ten decision strategies that differ depending on which tests are performed, whether the tests are performed together or in sequence, and the definition of a positivity criterion used to make the treatment decision when the test results disagree. Formulas derived from the model are used to compute the preferred option as a function of disease probability and to calculate test and test-treatment thresholds. General guidelines developed from the model can be used without calculation to identify relative preferences for alternative options and to predict threshold effects.

The accuracy of experienced physicians' probability estimates for patients with sore throats. Implications for decision making

Ten physicians recorded their treatment decisions and estimated probabilities of streptococcal infection for patients with sore throats. Of 308 throat cultures, 15 (4.9%) were positive for group A streptococci. The physicians overestimated the probability of a positive culture for 81% of their patients and their estimates and treatment decisions were strongly associated. Of 104 patients treated before culture results were available, only eight had positive cultures. Probability overestimation may have been due to neglect of the low culture-positive rate, assignment of undue importance to weakly predictive or highly intercorrelated clinical features, and a value-induced bias, occurring when features important for treatment are erroneously linked to the likelihood of disease. Cognitive limitations in information processing may limit the effectiveness of pharyngitis management protocols that require subjective estimates of disease probability.

'A Look at the Chief Complaints' revisited. Current obstacles and opportunities for decision analysis

In 1979, Dr. William Schwartz wrote a "Sounding Board" article in the New England Journal of Medicine entitled: "Decision Analysis. A Look at the Chief Complaints". In it, he listed the major concerns of practicing physicians, academicians, and students about decision analysis. As an educator and an advocate of the technique, he then replied to each of these concerns and objections. Dr. Schwartz's article is an excellent and thoughtful discussion and one that we routinely recommend to our students to prepare them for real-world reactions to the zeal with which they often leave our classrooms. Now, more than four years later, I would like briefly to review the "chief complaints" outlined by Dr. Schwartz and to update them to 1984. My purpose is to consider the cognitive and practical obstacles to the dissemination of decision analysis, and to discuss alternative approaches to surmounting these obstacles. To do this, I have engaged four "consultants" from different fields.

A course in clinical decision making adaptable to diverse audiences

Faculty at the University of Pennsylvania have developed a prototype course in clinical decision making that can be adapted to the diverse backgrounds of a variety of medical audiences. The course was offered in its entirety to third and fourth-year medical students and in abbreviated form to two postgraduate audiences (community and university-based physicians) during 1982. Methods were developed for content, process, and outcome evaluation for the courses; the latter consisted of pretest and posttest comparisons of performance on a written examination. Ninety-four individuals attended one or more sessions of the three courses. All courses were very favorably received, although the postgraduate audiences perceived less clinical relevance than educational relevance in the material (p less than 0.05). The medical students performed better on the pretest than either group of physicians, with the student-university physician difference reaching statistical significance (p less than 0.01). Nevertheless, all groups performed better on the posttest than on the pretest (p less than 0.001) and the degree of improvement was no different among the groups (p greater than 0.29). We conclude that our course's concepts and skills can be effectively adapted to and assimilated by physicians at all levels of training and experience.

Noninvasive neurovascular tests for carotid artery disease

Indirect noninvasive neurovascular tests provide information on hemodynamic changes cephalad to the carotid bifurcation, whereas direct tests measure anatomic or physiologic changes at the bifurcation itself. Batteries of tests are often done and should combine both indirect and direct methods. Results with two ultrasonic techniques and digital subtraction intravenous angiography suggest a larger role for these methods. Carotid arteriography is the definitive procedure for evaluating the carotid artery, although it should only be done when carotid endarterectomy is contemplated. The relative risks and benefits of other diagnostic and therapeutic management strategies should guide the decision to do noninvasive neurovascular tests, or to proceed directly to arteriography.

Antitumor effects of doxorubicin against a virally-induced rat osteosarcoma with minimal immunosuppression

Intratibial inoculation of a Moloney strain of Murine Sarcoma Virus (MSV-M) in neonatal Wistar-Lewis rats produced osteosarcoma in 96% of animals and resulted in a median survival of 20 days. Intraperitoneal (i.p.) administration of doxorubicin (adriamycin) (1-2 mg/kg/d, on day 10-12) resulted in reduced tumor growth and prolonged median survival to 95+ and 64 days, respectively. Higher dose doxorubicin (3-4 mg/kg/d, on day 10-12) caused early lethal toxicity. Autopsy data revealed a characteristic sarcomatous tumor producing osteoid. Gross pulmonary nodules appeared in 30% of both treated and untreated animals. Microscopic evaluation of lung tissue revealed anaplastic tumors without osteoid in as many as 90% of rats. Hepatosplenomegaly was usually present but microscopic sections of the spleen did not reveal tumor. Long bone metastases were increased in frequency in those animals receiving doxorubicin. Cell mediated immunity (CMI) to osteosarcoma cells by peripheral blood lymphocytes of tumor-bearing animals was detectable between days 21-48. This was bimodal with an early peak at day 21 (CMI = 56%) and a late peak at day 39 (CMI = 48%). CMI in rats given 1 mg/kg/d x 3d of doxorubicin was similar, with peak cytotoxicity (CMI = 61%) on day 26. Two mg/kg/d x 3d of doxorubicin did not significantly suppress either the early response (CMI = 50% on day 22) or the second peak (CMI = 38% and 50% on day 40 and 46, respectively). Thus, doxorubicin was effective in decreasing the growth of an MSV-M induced osteosarcoma and prolonging survival in the rat while usually failing to suppress CMI against rat osteosarcoma cells.

Induction of anaemia in monkeys

20 healthy young adult male rhesus monkeys were bled 3 times a week in an attempt to induce an anaemia and a depletion of iron stores. A mean haemoglobin of 6·1 g/100 ml and a haematocrit of 23% were obtained by 30 days. The iron saturation had fallen to a level of 11% and the reticulocyte response to 4·6%. Of note was the short period required to induce these haematological changes, and the minimal impact on the animals.