Search

Objective: To establish a baseline for the incidence of sepsis by severity and presence on admission in acute care hospital settings before implementation of a broad sepsis screening and response initiative. Methods: A retrospective cohort study using hospital discharge abstracts of 5672 patients, aged 18 years and above, with sepsis-associated stays between February 2012 and January 2013 at an academic medical center and 5 community hospitals in Texas. Results: Sepsis was present on admission in almost 85% of cases and acquired in-hospital in the remainder. The overall inpatient death rate was 17.2%, but was higher in hospital-acquired sepsis (38.6%, medical; 29.2%, surgical) and Stages 2 (17.6%) and 3 (36.4%) compared with Stage 1 (5.9%). Patients treated at the academic medical center had a higher death rate (22.5% vs. 15.1%, P

The purpose of this study was to determine whether stress myocardial perfusion (SPECT) optimized with stress-only (SO) imaging is comparable to cardiac CT angiography (CTA) for evaluating patients with acute chest pain (ACP).This was a prospective randomized observational study in 598 ACP patients who underwent CTA versus SPECT. The primary endpoint was length of hospital stay, and secondary endpoints were test feasibility, time to diagnosis, diagnostic accuracy, radiation exposure, and overall cost. Median follow-up was 6.5 mo, with a 3.8% cardiac event rate defined as death or an acute coronary syndrome.Of 2,994 patients screened, 1,703 (56.9%) were not candidates for CTA because of prior cardiac disease (41%) or imaging contraindications (16%). Time to diagnosis (8.1 ± 8.5 vs. 9.4 ± 7.4 h) and length of hospital stay (19.7 ± 27.8 vs. 23.5 ± 34.4 h) were significantly shorter with CTA than with SPECT (P = 0.002). However, time to diagnosis (7.0 ± 6.2 vs. 6.8 ± 5.9 h, P = 0.20), length of stay (15.5 ± 17.2 vs. 16.7 ± 15.3 h, P = 0.36), and hospital costs ($4,242 ± $3,871 vs. $4,364 ± 1781, P = 0.86) were comparable with CTA versus SO SPECT, respectively. SO was also superior to conventional SPECT regarding all of the above metrics and significantly reduced radiation exposure (5.5 ± 4.4 vs. 12.5 ± 2.7 mSv, P0.0001).Stress SPECT when optimized with SO imaging is similar to CTA in time to diagnosis, length of hospital stay, and cost, with improved prognostic accuracy and less radiation exposure. Our results emphasize the importance of SO imaging, particularly in low-intermediate-risk emergency room patients who are a population likely to have a normal test result.

The hypothesis that spondylolysis (SL) and/or isthmic spondylolisthesis (IS) cause low back pain (LBP) is widely accepted representing surgical indication in symptomatic cases. If SL/IS cause LBP, individuals with these conditions should be more prone to LBP than those without SL/IS. Therefore, the goal of the study was to assess whether the published primary data demonstrate an association between SL/IS and LBP in the general adult population. Systematic review of published observational studies to identify any association between SL/IS and LBP in adults. The methodological quality of the cohort and case–control studies was evaluated using the Newcastle-Ottawa scale. Fifteen studies met inclusion criteria (one cohort, seven case–control, seven cross-sectional). Neither the cohort study nor the two highest-quality case–control studies detected an association between SL/IS and LBP; the same is true for the remaining studies. There is no strong or consistent association between SL/IS and LBP in epidemiological studies of the general adult population that would support a hypothesis of causation. It is possible that SL/IS coexist with LBP, and observed effects of surgery and other treatment modalities are primarily due to benign natural history and nonspecific treatment effects. We conclude that traditional surgical practice for the adult general population, in which SL/IS is assumed to be the cause of non-radicular LBP whenever the two coexist, should be reconsidered in light of epidemiological data accumulated in recent decades.

To test the hypothesis that compared with daily soap and water bathing, 2% chlorhexidine gluconate bathing every other day for up to 28 days decreases the risk of hospital-acquired catheter-associated urinary tract infection, ventilator-associated pneumonia, incisional surgical site infection, and primary bloodstream infection in surgical ICU patients.This was a single-center, pragmatic, randomized trial. Patients and clinicians were aware of treatment-group assignment; investigators who determined outcomes were blinded.Twenty-four-bed surgical ICU at a quaternary academic medical center.Adults admitted to the surgical ICU from July 2012 to May 2013 with an anticipated surgical ICU stay for 48 hours or more were included.Patients were randomized to bathing with 2% chlorhexidine every other day alternating with soap and water every other day (treatment arm) or to bathing with soap and water daily (control arm).The primary endpoint was a composite outcome of catheter-associated urinary tract infection, ventilator-associated pneumonia, incisional surgical site infection, and primary bloodstream infection. Of 350 patients randomized, 24 were excluded due to prior enrollment in this trial and one withdrew consent. Therefore, 325 were analyzed (164 soap and water versus 161 chlorhexidine). Patients acquired 53 infections. Compared with soap and water bathing, chlorhexidine bathing every other day decreased the risk of acquiring infections (hazard ratio = 0.555; 95% CI, 0.309-0.997; p = 0.049). For patients bathed with soap and water versus chlorhexidine, counts of incident hospital-acquired infections were 14 versus 7 for catheter-associated urinary tract infection, 13 versus 8 for ventilator-associated pneumonia, 6 versus 3 for incisional surgical site infections, and 2 versus 0 for primary bloodstream infection; the effect was consistent across all infections. The absolute risk reduction for acquiring a hospital-acquired infection was 9.0% (95% CI, 1.5-16.4%; p = 0.019). Incidences of adverse skin occurrences were similar (18.9% soap and water vs 18.6% chlorhexidine; p = 0.95).Compared with soap and water, chlorhexidine bathing every other day decreased the risk of acquiring infections by 44.5% in surgical ICU patients.

Duration of Initiative 48 months and currently ongoing. Setting The Houston Methodist Hospital System and affiliated hospitals (3 facilities with 2 hospital-run skilled nursing facilities in and around Houston), St. Joseph's Regional Health Center (1 acute care hospital and 2 skilled nursing facilities in Bryan, Texas), Hospital Corporation of America (2 acute care facilities in Houston, 1 acute care facility in McAllen, Texas [Rio Grande Valley]), Kindred Healthcare (2 long term acute care facilities in Houston), Select Medical Specialty Hospitals (2 long term acute care facilities in Houston). Whom This Should Concern Hospital administrators, quality and safety officers, performance improvement and patient safety professionals, clinic managers, infection control and prevention staff, and other physicians, nurses, and clinical staff.

Racial/ethnic health care disparities are well described in people living with HIV/AIDS, although the processes underlying observed disparities are not well elucidated.A retrospective analysis nested in the University of Alabama at Birmingham 1917 Clinic Cohort observational HIV study evaluated patients between August 2004 and January 2007. Factors associated with appointment nonadherence, a proportion of missed outpatient visits, were evaluated. Next, the role of appointment nonadherence in explaining the relationship between African American race and virologic failure (plasma HIV RNA50 copies/mL) was examined using a staged multivariable modeling approach.Among 1221 participants, a broad distribution of appointment nonadherence was observed, with 40% of patients missing at least 1 in every 4 scheduled visits. The adjusted odds of appointment nonadherence were 1.85 times higher in African American patients compared with whites [95% confidence interval (CI) = 1.61 to 2.14]. Appointment nonadherence was associated with virologic failure (odds ratio = 1.78, 95% CI = 1.48 to 2.13) and partially mediated the relationship between African American race and virologic failure. African Americans had 1.56 times the adjusted odds of virologic failure (95% CI = 1.19 to 2.05), which declined to 1.30 (95% CI = 0.98 to 1.72) when controlling for appointment nonadherence, a hypothesized mediator.Appointment nonadherence was more common in African American patients, associated with virologic failure, and seemed to explain part of observed racial disparities in virologic failure.

OBJECTIVES: Percutaneous endoscopic gastrostomy (PEG) tube placement is a widely used method for long-term enteral feeding of demented patients unable to take sufficient food by mouth. National time trends in PEG tube use over the last decade have not been previously reported. The objective of this study was to determine whether use of PEG tubes for patients with dementia has changed over time and by race. DESIGN: Retrospective cohort study. SETTING: All Veterans Affairs hospitals. PARTICIPANTS: Using an administrative database of the Veterans Health Administration, all veterans with dementia and all veterans who received a PEG tube were identified between fiscal years 1990–2001. MEASUREMENTS: Proportion of PEG tube placement for dementia patients over time and by race. RESULTS: Four hundred thirteen thousand six hundred twenty-seven dementia patients aged 60 and older were identified, of whom 6,464 (1.6%) received a PEG tube. Use of PEG tubes for dementia patients increased during the first half of the decade but subsequently decreased almost to baseline after peaking in 1996 (1990: 1.2%, 1996: 1.8%, 2001: 1.3%). Time trends in the use of PEG tube feeding for dementia patients varied by race. Specifically, the relative risk for PEG tube placement in African-American dementia patients increased from 1.65 (95% confidence interval (CI)51.25–2.17, FY 1990) to 1.97 (95% CI51.62–2.4, FY 2001). CONCLUSION: Although the overall use of PEG tube feeding for dementia patients decreased over time, rates in use and changes in use over time varied significantly by race. Reasons for the differential use of this procedure should be explored.

To determine whether patients' decisions are an important determinant of nonuse of invasive cardiac procedures and whether decisions vary by race.Observational prospective cohort.Patients (N= 681) enrolled at the exercise treadmill or the cardiac catheterization laboratories at a large Veterans Affairs hospital.Doctors' recommendations and patients' decisions were determined by both direct observation of doctor and patient verbal behavior and by review of medical charts. Performance of coronary angiography, angioplasty, and bypass surgery were determined by chart review for a minimum of 3 months follow-up.Coronary angiography was recommended after treadmill testing for 83 of 375 patients and 72 patients underwent angiography. Among 306 patients undergoing angiography, recommendations for coronary angioplasty or bypass surgery were given to 113 and 45 patients and were completed for 98 and 33 patients, respectively. Recommendations were not significantly different by race. However, 4 of 83 (4.8%) patients declined or did not return for recommended angiograms and this was somewhat more likely among black and Hispanic patients, compared with white patients (13% and 33% vs 2%; P =.05). No patients declined angioplasty and 2 of 45 (4.4%) patients declined or did not return for recommended bypass surgery. Other recommended procedures were not completed after further medical evaluation (n = 32). There was no difference (P.05) by race/ethnicity in doctor-level reasons for nonreceipt of recommended invasive cardiac procedures.Patient decisions to decline recommended invasive cardiac procedures were infrequent and may explain only a small fraction of racial disparities in the use of invasive cardiac procedures.

PURPOSE: To examine variation in men’s long-term regret of treatment decisions, ie, surgical versus chemical castration, for metastatic prostate cancer and its associations with quality of life. METHODS: Survey of previously treated patients to assess treatment decisions and quality of life, supplemented with focus groups. Two items addressing whether a patient wished he could change his mind and the belief that he would have been better off with the treatment not chosen were combined in classifying survey respondents as either satisfied or regretful. χ2 and t tests were used to test associations between regret and treatment history, complications, and quality of life. RESULTS: Survey respondents included 201 men aged 45 to 93 years (median, 71 years), who had begun treatment (71% chemical castration, 29% orchiectomy) a median of 2 years previously. Most reported complications: hot flashes (70%), nausea (34%), and erectile dysfunction (81%). Most were satisfied with the treatment decision, but 23% expressed regret. Regretful men more frequently reported surgical (43%) versus chemical (36%) castration (P = .030) and nausea in the past week (54% v 32%; P = .010) but less frequently reported erectile dysfunction (56% v 72%; P = .048). Regretful men indicated poorer scores on every measure of generic and prostate cancer–related quality of life. Qualitative analyses revealed substantial uncertainty about the progress of their disease and the quality of the decisions in which patients participated. CONCLUSION: Regret was substantial and associated with treatment choice and quality of life. It may derive from underlying psychosocial distress and problematic communication with physicians when decisions are being reached and over subsequent years.

BACKGROUND: The Department of Veterans Affairs (VA) established six mobile clinics to provide care for rural veterans. Each was operated by a parent VA Medical Center (VAMC). OBJECTIVE: To describe the use of a cost-accounting system which does not provide costs at the service or patient level to determine the costs of the mobile clinics. RESEARCH DESIGN: Costs per visit were compared among the mobile clinics with the parent VAMCs and with simulated fixed-location clinics. Cost data came from VA's Centralized Accounting for Local Management (CALM) data. Utilization data came from VA's outpatient file. RESULTS: Information was obtained from the VAMCs' fiscal services to reallocate costs among the CALM subaccounts to generate cost data that was comparable among the mobile clinics. Costs per visit for the mobile clinics were twice as high as those of the parent VAMCs. Costs per visit would be lower at fixed-location clinics unless the volume were substantially less than that provided by the mobile clinics. CONCLUSION: Differences between cost allocations for accounting purposes and research are likely to necessitate adjusting cost accounting data for research purposes. Fortunately, information from the accountants or primary data can lead to a cost database which is appropriate for research evaluations. In the mobile clinics study, the analysis of cost accounting data led to the conclusion that mobile clinics were not a cost-effective way in which to provide care to rural veterans.

BACKGROUND Little data exist supporting the association of quality of care and nonfatal adverse outcomes in hospitalized patients, yet those outcomes are routinely scrutinized in quality assessment efforts. OBJECTIVE To determine whether measurable differences in quality of care are associated with the occurrence of non-fatal, in-hospital, and treatment-related complications. DESIGN Retrospective cohort study. SUBJECTS A total of 2,268 patients who were discharged alive from 9 Southwestern Veterans Affairs Medical Centers with congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD) or diabetes mellitus. MEASURES Retrospective chart review was performed to collect information on patient severity of illness, in-hospital complication occurrence, and process quality of care. Process quality was assessed as the adherence scores for admission work-up and for treatment during the hospital stay. Process quality represents the proportion of applicable admission or treatment criteria that were met by that patient's care providers. Once severity of illness was taken into account Cox proportional hazards regression was used to assess the independent contribution of process quality of care to complication occurrence. RESULTS Higher admission work-up adherence scores for COPD patients and higher treatment adherence scores for COPD and diabetes patients were associated with a lower risk of complication occurrence. The adjusted risk ratios of complications for higher versus lower adherence scores (with 95% CI) were 0.64 (0.43, 0.97) and 0.52 (0.33, 0.80) for admission and treatment, respectively, in COPD patients, and 0.51 (0.31, 0.83) for treatment in diabetics. No significant association was found in CHF patients. CONCLUSION Better admission work-up and treatment quality in COPD patients, as well as treatment quality in diabetic patients, are associated with lower risk of nonfatal treatment-related complications in the study population.

In the United States, geographic variation in hospital use is common. It is uncertain whether there are similar geographic variations in the health care system of the Department of Veterans Affairs (VA), which differs from the private sector because it predominantly serves men with annual incomes below $20,000, has a central system of administration, and uses salaried physicians. Thus, it might be less likely to have geographic variations.We used VA data bases to obtain information on patients treated for eight diseases (chronic obstructive pulmonary disease, pneumonia, congestive heart failure, angina, diabetes, chronic renal failure, bipolar disorder, and major depression). We analyzed their use of hospital and outpatient services by assessing the risk-adjusted numbers of hospital days (the average number of days a patient spent in the hospital per 12 months of follow-up, regardless of the number of hospital stays), hospital-discharge rates, and clinic-visit rates from 1991 through 1995 for the entire system and within the 22 geographically based health care networks.We found substantial geographic variation in hospital use for all eight cohorts of patients and all the years studied. Variations in the numbers of hospital days per person-year among the networks were greatest among patients with chronic obstructive pulmonary disease (ranging from a factor of 2.7 to a factor of 3.1) during a given year and smallest among patients with angina (ranging from a factor of 1.5 to a factor of 2.1). Levels of hospital use were highest in the Northeast and lowest in the West. The variation in the rates of clinic visits for principal medical care among the networks ranged from a factor of approximately 1.6 to a factor of 4.0; variations in the rates were greatest among patients with chronic renal failure and smallest among patients with chronic obstructive pulmonary disease. There was no clear geographic pattern in the rates of outpatient-clinic use.There are significant geographic variations in the use of hospital and outpatient services in the VA health care system. Because VA physicians are unable to increase their income by changing their patterns of practice, our findings suggest that their practice styles are similar to those of other physicians in their geographic regions.

This article and the following article (Parts I and II) report the development of two clinical staging systems for HIV-infected individuals. The objective of the research reported here (Part I) was to construct a clinical staging system to predict progression to AIDS. We analyzed data from VA Cooperative Study Number 298, a multicenter, double-blind, randomized trial that compared immediate versus deferred zidovudine therapy in 338 HIV-infected individuals who did not have AIDS at enrollment. Baseline variables were tested in univariate Cox regression for their relationship to progression to AIDS, and those that appeared predictive were examined in multivariable analysis. Based on these analyses, we constructed a new clinical staging system based on CD4+ cell count, age, hemoglobin, oral hairy leukoplakia or oral thrush, and fever. The stages of the system were significant predictors of progression to AIDS (p = 0.0001, log-rank test). In conclusion, simple, valid, clinical staging systems for HIV-infected patients can be constructed using information that is readily available in clinical practice settings. Such systems provide better prognostic distinction than CD4+ cell count alone by taking into account the known prognostic effects of other variables.

Following a 4-year controlled trial comparing early and later zidovudine treatment, we conducted an additional 3-year follow-up. Of the original 338 patients, 275 participated. Clinical outcome measures were AIDS and death. In the early therapy group (n = 170), 67 patients progressed to AIDS compared with 85 in the later therapy group (n = 168); the relative risk (RR) comparing early with later therapy was 0.72% (95% confidence interval [CI] 0.52-0.99; p = 0.044). The early therapy group had 74 deaths compared with 73 in the later therapy (RR = 0.98; 95% CI, 0.71-1.36; p = 0.91). The early group had a peak CD4+ count increase at 1-2 months and a delay of 1 year before CD4+ counts fell below baseline. For patients who received zidovudine for more than the median duration (20.3 months) before their first AIDS diagnosis, the RR for death was 2.08 (95% CI, 1.36-3.19, p = 0.001). Additional factors independently associated with poor prognosis following AIDS were a CD4+ count of < 100 cells/mm3 and increased severity of the first AIDS diagnosis, whereas use of another antiretroviral agent was associated with improved survival. We conclude that early zidovudine therapy delays progression to AIDS but does not affect survival. Patients who progress to AIDS while on prolonged zidovudine monotherapy many benefit from a change to other antiretroviral therapy(ies).

The reasons why many patients seemingly benefit from arthroscopic treatment of osteoarthritis of the knee remain obscure. The purpose of this pilot study was to determine if a placebo effect might play a role in arthroscopic treatment of this condition. After giving full informed consent, including full knowledge of the possibility and nature of a placebo surgery, five subjects were randomized to a placebo arthroscopy group, three subjects were randomized to an arthroscopic lavage group, and two subjects were randomized to a standard arthroscopic debridement group. The physicians performing the postoperative assessment and the patients remained blinded as to treatment. Patients who received the placebo surgery reported decreased frequency, intensity, and duration of knee pain. They also thought that the procedure was worthwhile and would recommend it to family and friends. Thus, there may be a significant placebo effect for arthroscopic treatment of osteoarthritis of the knee. The small numbers in this preliminary study preclude a valid statistical analysis, and no conclusions can be drawn regarding the superiority of one treatment over another. A larger study is needed to evaluate fully the efficacy of an arthroscopic procedure for this condition and to decide if it is reasonable to expend health care resources for this treatment; the larger study should include a placebo control group.

Background. Zidovudine is recommended for asymptomatic and early symptomatic human immunodeficiency virus (HIV) infection. The best time to initiate zidovudine treatment remains uncertain, however, and whether early treatment improves survival has not been established. Methods. We conducted a multicenter, randomized, double-blind trial that compared early zidovudine therapy (beginning at 1500 mg per day) with late therapy in HIV-infected patients who were symptomatic and had CD4+ counts between 0.2×109 and 0.5×109 cells per liter (200 to 500 per cubic millimeter) at entry. Those assigned to late therapy initially received placebo and began zidovudine when their CD4+ counts fell below 0.2×109 per liter (200 per cubic millimeter) or when the acquired immunodeficiency syndrome (AIDS) developed. Results. During a mean follow-up period of more than two years, there were 23 deaths in the early-therapy group (n = 170) and 20 deaths in the late-therapy group (n = 168) (P = 0.48; relative risk [late vs. e...

This study examines the intravenous use of two thrombolytic agents [streptokinase (SK) and tissue plasminogen activator (tPA)] in the acute phase of myocardial infarction (MI). These two agents have very different costs and offer an excellent opportunity to study both the impact of economic factors on clinical decisionmaking and the potential for cost savings by limiting the use of expensive new therapeutic agents. A nationwide survey of the 5,792 acute care general hospitals listed in the American Hospital Association's 1988 data file was responded to by 2,651 hospitals (46%) and revealed that 2,384 of these responding hospitals (90%) were using thrombolytic therapy. For 2,200 of these 2,384 hospitals (92%), the respondent was a physician who primarily used one of the two drugs. Eight hundred eighty-six of these 2,200 physicians (40%) primarily used SK while 1,314 (60%) primarily used tPA. SK users were more concentrated in federal public hospitals (69% used SK) than in nonfederal public hospitals (47% used SK), and were least concentrated in private hospitals (36% used SK). There was no difference between the rate of SK vs tPA use in investor-owned and not-for-profit private hospitals. SK users most often (62%) cited various economic factors as the reason for their choice. The users of tPA primarily (73%) cited clinical preferability as the reason for their choice even though trials are still ongoing to see which drug is preferable. Several multivariate analyses shed light upon the association between choice of thrombolytic agent and various additional physician and hospital characteristics. These data clearly indicate that while new therapies are rapidly implemented by the medical community, considerations of cost have a substantial impact upon the pattern of implementation and reflect a desire to implement cost savings in the use of new drugs.

No national data exist regarding racial/ethnic differences in the use of interventions for patients at the end of life.To test whether among 3 cohorts of hospitalized seriously ill veterans with cancer, noncancer or dementia the use of common life-sustaining treatments differed significantly by race/ethnicity.Retrospective cohort study during fiscal years 1991-2002.Hospitalized veterans55 years, defined clinically as at high-risk for 6-month mortality, not by decedent data.Utilization patterns by race/ethnicity for 5 life-sustaining therapies. Logistic regression models evaluated differences among Caucasians, African Americans and Hispanics, controlling for age, disease severity and clustering of patients within Veterans Affairs (VA) medical centers.Among 166,059 veterans, both differences and commonalities across diagnostic cohorts were found. African Americans received more or the same amount of end-of-life treatments across disease cohorts, except for less resuscitation [OR = 0.84 (0.77-0.92), p = 0.002] and mechanical ventilation [OR = 0.89 (0.85-0.94), por = 0.0001] in noncancer patients. Hispanics were 36% (cancer) to 55% (noncancer) to 88% (dementia) more likely to receive transfusions than Caucasians (p0.0001). They received similar rates as Caucasians for all other interventions in all other groups, except for 161% higher likelihood for mechanical ventilation in patients with dementia. Increased end-of-life treatments for both minority groups were most pronounced in the dementia cohort. Differences demonstrated a strong interaction with the disease cohort.Differences in level of end-of-life treatments were disease specific and bidirectional for African Americans. In the absence of generally accepted, evidence-based standards for end-of-life care, these differences may or may not constitute disparities.

BackgroundChronic obstructive pulmonary disease (COPD) affects 14 to 20 million Americans and is associated with increased prevalence of affective disorders, contributing significantly to disability. This study compared cognitive behavioral therapy (CBT) group treatment for anxiety and depression with COPD education for COPD patients with moderate-to-severe anxiety and/or depressive symptoms.MethodA randomized controlled trial (RCT) was conducted between 11 July 2002 and 30 April 2005 at the Michael E. DeBakey VA Medical Center, Houston, TX. Participants were 238 patients treated for COPD the year before, with forced expiratory value in 1 second (FEV)1/forced vital capacity (FVC)1ResultsBoth treatments significantly improved QoL, anxiety and depression (pConclusionsCBT group treatment and COPD education can achieve sustainable improvements in QoL for COPD patients experiencing moderate-to-severe symptoms of depression or anxiety.

Obstructive lung diseases are associated with high rates of depression and anxiety, yet many patients are never screened or treated. This study evaluated the five-question Depression and Anxiety modules of the Primary Care Evaluation of Mental Disorders (PRIME-MD) Patient Questionnaire as a telephone screen in 1,632 patients with chronic breathing disorders at a Veterans Affairs Medical Center in Houston, TX. Subsequent testing of 828 patients with the Beck Depression Inventory-II and the Beck Anxiety Inventory showed that the sensitivity and specificity, respectively, of the Depression and Anxiety modules of the PRIME-MD Patient Questionnaire screening were 94.6% and 49.5% (Depression); 93.7% and 32.2% (Anxiety); and 97.7%, and 36.0% (combined screen), with an overall accuracy of 80.7%. In such populations, these two modules of the PRIME-MD Patient Questionnaire may be a useful screening tool.

Article-at-a-Glance Background Sepsis is a leading cause of death, but evidence suggests that early recognition and prompt intervention can save lives. In 2005 Houston Methodist Hospital prioritized sepsis detection and management in its ICU. In late 2007, because of marginal effects on sepsis death rates, the focus shifted to designing a program that would be readily used by nurses and ensure early recognition of patients showing signs suspicious for sepsis, as well as the institution of prompt, evidence-based interventions to diagnose and treat it. Methods The intervention had four components: organizational commitment and data-based leadership; development and integration of an early sepsis screening tool into the electronic health record; creation of screening and response protocols; and education and training of nurses. Twice-daily screening of patients on targeted units was conducted by bedside nurses; nurse practitioners initiated definitive treatment as indicated. Evaluation focused on extent of implementation, trends in inpatient mortality, and, for Medicare beneficiaries, a before-after (2008–2011) comparison of outcomes and costs. A federal grant in 2012 enabled expansion of the program. Results By year 3 (2011) 33% of inpatients were screened (56,190 screens in 9,718 unique patients), up from 10% in year 1 (2009). Inpatient sepsis-associated death rates decreased from 29.7% in the preimplementation period (2006–2008) to 21.1% after implementation (2009–2014). Death rates and hospital costs for Medicare beneficiaries decreased from preimplementation levels without a compensatory increase in discharges to postacute care. Conclusion This program has been associated with lower inpatient death rates and costs. Further testing of the robustness and exportability of the program is under way.

Objective: Screening for mental illness in primary care is widely recommended, but little is known about the evaluation, treatment, and long-term management processes that follow screening. The aim of this study was to examine and describe the quality of mental health care for persons with chronic obstructive pulmonary disease (COPD) and anxiety/depressive disorders, as measured by adherence to practice guidelines. Method: This retrospective chart review examined data for 102 primary care and mental health care patients with COPD who were diagnosed, using Structured Clinical Interview for DSM-IV criteria, with major depressive disorder, dysthymia, depression not otherwise specified, generalized anxiety disorder, or anxiety not otherwise specified. Data were gathered from primary care progress notes from the year prior to enrollment in a randomized controlled trial (enrollment was from July 2002 to April 2004). We compared the care received by these patients over 1 year with that recommended by practice guidelines. Charts were abstracted using a checklist of recommended practice guidelines for diagnostic evaluation, acute treatment, and long-term management of anxiety and depressive disorders. Results: Fifty (49%) of the 102 patients were recognized during the review year as having an anxiety or depressive disorder. Eighteen patients were newly assessed for depressive or anxiety disorders during the chart review year. Patients followed in primary care alone, compared with those who were comanaged by mental health care providers, were less likely to have guideline-adherent care. Conclusion: Depressive and anxiety disorders are recognized in about half of patients; however, guideline-supported diagnostic evaluation, acute treatment (except for medications), and long-term management rarely occur in the primary care setting. To improve the treatment of depressive and anxiety disorders in primary care, the process of care delivery must be understood and changed.

Background: Many healthcare organizations use diagnosis-based risk adjustment systems for predicting costs. Health self-report may add information not contained in a diagnosis-based system but is subject to incomplete response. Objective: The objective of this study was to evaluate the added predictive power of health self-report in combination with a diagnosis-based risk adjustment system in concurrent and prospective models of healthcare cost. Research Design: This was a cohort study using Department of Veterans Affairs (VA) administrative databases. We tested the predictive ability of the Adjusted Clinical Group (ACG) methodology and the added value of SF-36V (short form functional status for veterans) results. Linear regression models were compared using R2, mean absolute prediction error (MAPE), and predictive ratio. Subjects: Subjects were 35,337 VA beneficiaries at 8 VA medical centers during fiscal year (FY) 1998 who voluntarily completed an SF-36V survey. Measures: Outcomes were total FY 1998 and FY 1999 cost. Demographics and ACG-based Adjusted Diagnostic Groups (ADGs) with and without 8 SF-36V multiitem scales and the Physical Component Score and Mental Component Score were compared. Results: The survey response rate was 45%. Adding the 8 scales to ADGs and demographics increased the crossvalidated R2 by 0.007 in the prospective model. The 8 scales reduced the MAPE by $236 among patients in the upper 10% of FY 1999 cost. Conclusions: The limited added predictive power of health self-report to a diagnosis-based risk adjustment system should be weighed against the cost of collecting these data. Adding health self-report data may increase predictive accuracy in high-cost patients.

Although rarely studied, physician-patient interactions immediately following diagnostic tests are significant medical events because during these encounters the physician and patient often make decisions about major and sometimes invasive treatment. This investigation analyzed patterns of physician-patient communication following coronary angiography with particular attention to behaviors important to decision-making: physician information-giving, physician use of partnership-building, and active forms of patient participation (e.g., asking questions, being assertive, expressing concerns). We were particularly interested in effects related to the patient's race in light of documented evidence of racial disparities in cardiac care and outcomes. From audiotape recordings, 93 physician-patient interactions after coronary angiogram in a catheterization laboratory in a large US Veterans Affairs Medical Center were coded to measure the frequency of physicians' information-giving and partnership-building and the frequency of active patient participation. We also stratified these behaviors according to whether the behavior was prompted (e.g., physician information in response to a patient's question; a patient's opinion solicited by the doctor) or self-initiated. Several findings were noteworthy. First, these interactions were very brief and dominated by the physician. Second, although physician information-giving increased with more active patient participation, which in turn was correlated with physicians' use of partnership-building, proportionally little of the physicians' information (8%) and active patient participation (9%) was directly prompted by the other interactant. Finally, there was a tendency for physicians to self-initiate less information giving to black patients and for black patients to self-initiate less active participation than white patients. Although these differences were attenuated when other variables (e.g., the physician's training, disease severity) were included in the analysis, the pattern suggests a potential cycle of passivity where certain patients tend to receive fewer informational resources and these patients in turn do less to prompt the doctor for more.

Disease-specific registries have many important applications in epidemiologic, clinical and health services research. Since 1989 the Department of Veterans Affairs has maintained a national HIV registry. VA's HIV registry is national in scope, it contains longitudinal data and detailed resource utilization and clinical information. To describe the structure, function, and limitations of VA's national HIV registry, and to test its accuracy and completeness. The VA's national HIV registry contains data that are electronically extracted from VA's computerized comprehensive clinical and administrative databases, called Veterans Integrated Health Systems Technology and Architecture (VISTA). We examined the number of AIDS patients and the number of new patients identified to the registry, by year, through December 1996. We verified data elements against information obtained from the medical records at five VA sites. By December 1996, 40,000 HIV-infected patients had been identified to the registry. We encountered missing data and problems with data classification. Missing data occurred for some elements related to the computer programming that creates the registry (e.g., pharmacy files), and for other elements because manual entry is required (e.g., ethnicity). Lack of a standardized data classification system was a problem, especially for the pharmacy and laboratory files. In using VA's national HIV registry we have learned important lessons, which, if taken into account in the future, could lead to the creation of model disease-specific registries.

Chronic heart failure (CHF) is a highly prevalent condition associated with serious morbidity, intense levels of health services use, and shortened survival. It is also a condition for which ameliorative therapies exist. The evidence indicates that there is substantial need to change clinical practice and health care delivery for people with CHF and thereby improve their outcomes. The goal of the Veterans Affairs (VA) Quality Enhancement Research Initiative in CHF (CHF QUERI) is to create measurable, rapid, and sustainable improvements in quality of care and health outcomes of veterans with heart failure. This article describes the current state of knowledge and practice in care for people with CHF. Using the framework of the 5 steps of the QUERI process, we point out the gaps in research and practice that must be filled if the CHF QUERI is to achieve its goal. We relate our recommendations for how the VA can put its research and administrative infrastructure to work to fill the gaps. Lessons learned about CHF in the course of the CHF QUERI will be applicable to all people with heart failure and to all health care systems--VA as well as non-VA--that care for them.

The objective of this study was to describe patterns of hospital and clinic use and survival for a large nationwide cohort of patients with heart failure. A retrospective cohort study of patients treated in the Veterans Affairs medical care system was conducted using linked administrative databases as data sources. In 1996, the average heart failure cohort member had 1-2 hospitalizations, 14 inpatient days, 6-7 visits with the primary physician, 15 other visits for consultations or tests, and 1-2 urgent care visits per 12 months. The overall risk-adjusted 5-year survival rate was 36%. Hospital use rates in the cohort fell dramatically between 1992 and 1996. One-year survival rates increased slightly over the period. Patients with heart failure are heavy users of services and have a very poor prognosis. Utilization and outcome data indicate the need for major efforts to assure quality of care and to devise innovative ways of delivering comprehensive services.

The authors describe the role the Veterans Affairs (VA) medical system plays as a provider of clinic and hospital services by examining utilization levels and users' characteristics.The Veterans Affairs hospital discharge database, the Veterans Affairs outpatient clinic files, and the veteran population files were used to estimate the number of persons using the Veterans Affairs medical care system in 1994 and the intensity of their clinic and hospital use. Demographic and clinical characteristics of users were tabulated.In 1994, 2.7 million veterans, 10.3% of all US veterans, and approximately 23% of veterans who would have met the statutory eligibility requirements for Veterans Affairs care, used the hospital and/or clinic components of the Veterans Affairs medical system. Sixty-three percent of the system's users were younger than age 65, and 10.5% were women. These 2.7 million veterans had 901,665 Veterans Affairs hospital stays, 15.5 million bed-days, and 31.2 million outpatient visits in fiscal year 1994. The average number of hospitalizations per hospital user was 1.71; the average number of visits per clinic user was 11.7. Medical, surgical, and psychiatric diagnosis-related groups (DRGs) accounted for 56%, 21%, and 23%, respectively, of hospitalizations, but psychiatric diagnosis-related groups accounted for 43% of all inpatient days. Principal medicine clinic visits and psychiatry clinic visits accounted for 21% and 16% of Veterans Affairs ambulatory care.Because the patient population served by the Veterans Affairs system is skewed in a number of ways, its contribution as a provider of health services in the United States varies by gender, age, socioeconomic status, and diagnosis.

This article (Part II) and the preceding article (Part I) report the development of two clinical staging systems for HIV-infected individuals. The objective of the research reported here (Part II) was to construct a clinical staging system to predict survival in patients with AIDS. We analyzed data from VA Cooperative Study Number 298, a multicenter, double-blind, randomized trial that compared immediate versus deferred zidovudine therapy in HIV-infected individuals. Baseline variables obtained at the onset of AIDS in 204 individuals were tested in univariate Cox regression for their relationship to survival, and those that appeared predictive were examined in multivariable analysis. Based on these analyses, we constructed a new AIDS Clinical Staging System. The system is based on age, CD4+ cell count, type of first AIDS-defining condition, and functional status. The stages of the system were significant predictors of survival (p = 0.0001, log-rank test). In conclusion, valid, simple clinical staging systems for patients with AIDS can be developed based on a few variables that are readily available in clinical settings.

This study evaluated the predictive validity of two clinical staging systems for HIV infection (the Rabeneck and Royce systems) using data obtained from the Department of Veterans Affairs Cooperative Study Number 298, a randomized clinical trial involving 335 symptomatic patients with CD4 counts of 200 to 500/mm3. The relation between the HIV clinical stages and progression to AIDS was examined using Kaplan-Meier estimates, and Cox models were used to determine if the stages remained predictive after controlling for CD4 count. Both systems were significant independent predictors of progression to AIDS. This work demonstrates that simple, valid staging systems for HIV infection can be developed that provide greater prognostic distinction than the CD4 count alone.

To determine whether the quality of care during a hospital stay is associated with unplanned readmission within 14 days.Case-control study.12 Veterans Affairs hospitals.Men discharged after a hospitalization for diabetes (n = 593), chronic obstructive lung disease (n = 1172), or heart failure (n = 748). The ratio of controls (men without an unplanned readmission within 14 days to any Veterans Affairs hospital) to cases (men with such a readmission) was 3:1.Unplanned readmission to any of the 159 Veterans Affairs hospitals within 14 days of discharge. Quality of care during the index stay was assessed by chart review using disease-specific explicit criteria for the process of inpatient care, which were developed by panels composed of expert physicians. Adherence scores (the percentage of applicable criteria that were met) were calculated for the admission workup, evaluation and treatment, and readiness for discharge.After adjustment was made for demographic characteristics, severity of illness, and need for care, adherence scores correlated with early unplanned readmission (P0.05). For patients with diabetes and heart failure, decreased readiness-for-discharge adherence scores correlated with increased risk for readmission (P = 0.001 and P = 0.016, respectively). In patients with obstructive lung disease, decreased admission-workup scores correlated with increased risk for readmission (P = 0.013). One of 7 readmissions in patients with diabetes, 1 of 5 readmissions in patients with heart failure, and 1 of 12 readmissions in patients with obstructive lung disease were attributable to substandard care.Lower quality of inpatient care increases the risk for unplanned early readmission in patients with heart failure, diabetes, or obstructive lung disease. Under certain circumstances, readmission is associated with remediable deficiencies in the process of inpatient care.

Health care payors and providers are increasingly monitoring hospital discharge data bases for adverse events as markers for quality of care. The principal criticisms of these analyses have focused on the impediments to risk adjustment posed by the incompleteness and inaccuracy of the data bases. However, efforts to address the inadequacies of the data bases will not correct deficiencies of the analytic process. These deficiencies arise from the application of one adverse outcome to all disease states. Instead, analysis should be restricted to comparisons of subgroups of patients in which a close fit exists between the quality of care for the disease state and the expected outcome. Furthermore, these disease-outcome pairs should be minimally subject to measurement error. The authors present a conceptual framework for developing such meaningful disease-outcome pairs, and using the hospital discharge data base of the Department of Veterans Affairs, show how the framework can be used to devise a monitoring strategy for re-admission.

To determine the incidence of and risk factors for perioperative myocardial infarction with noncardiac surgery and to test the accuracy of a risk stratification system.Prospective cohort study.A large urban Veterans Affairs hospital.A total of 1487 men older than 40 years undergoing major, nonemergent, noncardiac operations.Infarction was established by at least two of the following: development of new Q waves, typical change in creatine kinase MB, and positive technetium pyrophosphate scintigraphy. Patients were stratified preoperatively into high-, intermediate-, low-, and negligible-risk strata based on clinical markers corresponding to different levels of coronary artery disease prevalence.Patients with coronary disease (high-risk stratum) had a 4.1% incidence of infarction (13 of 319; 95% CI, 1.8% to 6.4%); patients with peripheral vascular disease but no evidence of coronary disease (intermediate-risk stratum) had a 0.8% incidence (2 of 260, upper bound of CI, 2.0%); patients with high atherogenic risk factor profiles but no clinical atherosclerosis (low-risk stratum) had a 0% incidence (0 of 256, upper bound of CI, 1.2%). No cardiac deaths occurred in 652 men who had no atherosclerosis and low atherogenic risk factor profiles (the negligible-risk stratum). Factors independently associated with infarction included age more than 75 years (adjusted odds ratio, 4.77; CI, 1.17 to 19.41), signs of heart failure on the preoperative examination (adjusted odds ratio, 3.31; CI, 0.96 to 11.38), coronary disease (adjusted odds ratio, 10.39; CI, 2.27 to 47.46), and a planned vascular operation (adjusted odds ratio, 3.72; CI, 1.12 to 12.37).Coronary artery disease is the major risk factor for perioperative infarction. The stratification scheme identifies subsets of patients with different risks, and finer within-stratum distinctions can be made using additional variables.

Objective: To develop a simple clinical staging system based on CD4 count and clinical variables that predicts progression to AIDS in HIV-infected non-AIDS patients. Design: Retrospective cohort study. Setting: A primary care outpatient clinic for HIV-infected patients at a VA Medical Center. Patients: One hundred seventy-six HIV-infected non-AIDS patients seen at the Houston VA Special Medicine Clinic between January 1986 and December 1990 and followed for a mean of 22 months. Fifty-four patients (31%) progressed to AIDS during follow-up. Measurements: The medical records were reviewed, and data corresponding to the initial (baseline) clinic visit and subsequent sixmonth visits were extracted. Main results: “Predictive” baseline variables (i.e., those associated with progression to AIDS) were first identified and then examined in Cox proportional hazards modeling. In the final model, CD4 category, oral thrush, and night sweats made significant independent contributions. A three-stage prognostic system was constructed by assigning points to the three variables: CD4>500 cells/mm3=0; 500≥CD4≥200=1; CD4

Although many clinicians routinely recommend a base-line preoperative electrocardiogram (ECG) and obtain frequent postoperative ECGs to screen for myocardial infarction or ischemia, the diagnostic utility of screening perioperative ECGs is unknown. The present analysis evaluates the sensitivity and specificity of the perioperative ECG and examines its value as a predictor of early postoperative cardiac events and outcomes during the postoperative year. ECGs obtained preoperatively and on the first 3 postoperative days in 206 men undergoing transurethral prostate resection were analyzed using the Minnesota Code. The occurrence of cardiac events during the operative stay was assessed by measurement of the cardiospecific MB creatine kinase isoenzyme on the first 3 postoperative days and review of the entire clinical course. Twenty-one percent of patients developed postoperative ECG changes, mostly involving the T wave; none had cardiac symptoms or sustained creatine kinase MB elevation. Changes were not significantly more common in men known to have coronary disease. The single patient who had a perioperative myocardial infarction confirmed by enzymes had no codable ECG changes. The specificity of any ECG change for perioperative infarction was 78%; of ST segment changes only, 95%. Only one of the patients (2%) who had postoperative ECG changes had a cardiac event in the year after surgery. Routine perioperative ECGs is of little diagnostic/predictive utility in situations in which the incidence of perioperative myocardial infarction is low.

To The Editor: We appreciate Dr. Fowler's positive comments on our study ("Arthroscopic Lavage or Debridement Did Not Reduce Pain More Than Placebo Did in Patients with Osteoarthritis" [ N Engl J Med. 2002;347:81-8]) in Evidence-Based Orthopaedics ( J Bone Joint Surg Am. 2002;84:387) regarding the internal validity of the study, and we wish to address his concerns regarding its external validity. We believe that all of his concerns are due to our inability to include all of our data in the article because of the limitations placed on its length by The New England Journal of Medicine . Dr. Fowler argued that, although our results were valid in the spectrum of patients whom we studied, these results might not apply more generally. First, we used the system of Kellgren and Lawrence 1 to classify each patient's study knee according to the severity of osteoarthritis. Our study cohort included thirty patients whose study …

An important determinant of patient outcomes is illness severity, which must be classified to guide clinical decision making and evaluate the effectiveness of diagnostic and therapeutic interventions. Currently, no widely accepted framework for grading illness severity in human immunodeficiency virus—infected patients exists. The best known classification systems for human immunodeficiency virus infection (Centers for Disease Control and Prevention; Walter Reed) are not based on illness severity, and provide stages that are not all inclusive so that a large number of persons cannot be classified (Walter Reed). Although much previous research has focused on individual prognostic factors (oral thrush, CD4 cell count, serum β2-microglobulin), little attention has been given to incorporating these factors into illness severity scales that are easy to use in clinical settings. In addition, despite the progressive functional disability of human immunodeficiency virus—infected individuals, few of the existing approaches to illness severity classification have examined the contribution of functional status. We urge investigators to develop clinically sensible severity scales that are easy to use. Such scales will considerably improve existing approaches that tend to rely solely on the CD4 cell count and do not take into account the known prognostic effects of other variables. (Arch Intern Med. 1993;153:2749-2755)

Aryl hydrocarbon hydroxylase was measured in cultured human lymphocytes induced with benzathracene and in pulmonary alveolar macrophages induced in situ in cigarette smokers. Considered separately, neither lymphocyte AHH nor macrophage AHH levels were distinctly different in either noncancer or lung cancer patients. Considered simultaneously, lymphocyte and macrophage AHH levels are quite different in noncancer and lung cancer patients. The lung cancer patient group was seen to contain a significantly higher percentage of persons with high levels of AHH than did an age-matched group of noncancer patients, (P less than 0.001), when more than one tissue was assayed to determine the individual's enzyme levels.

• A case of Legionella pneumophila infection complicated by thrombotic thrombocytopenic purpura (TTP) was confirmed at autopsy by the demonstration of the organism in lung tissue, and by the finding of widespread intravascular and subendothelial thrombi associated with microinfarctions in all major organs examined. In addition to the typical hematologic abnormalities of TTP, the patient was found to have a low serum C3 level and elevated levels of immune complexes as measured by the liquid phase C1q binding assay. We suggest that the L pneumophila infection caused endothelial damage and/or platelet aggregation, perhaps as a consequence of complement activation, thus initiating the sequence of events leading to extensive microvascular thromboses.

A component capable of binding benzo(a)pyrene was measured in plasma from cigarette smokers and nonsmokers. This plasma fraction was found to have a high specificity of binding to benzo(a)pyrene, bound benzanthracene competitively with benzo(a)pyrene, and was positively correlated (r = 0.861, p less than 0.001) with the capacity of the individual subject's lymphocytes to be induced for AHH activity in culture. An inverse correlation (r = -0.957, p less than 0.001) between the presence of the plasma component in lung cancer patients and the capacity of lung cancer patients' lymphocytes to be induced in culture is unexplained at this time. A benzo(a)pyrene-binding fraction was not found in induced or uninduced cultured lymphocytes from smokers or nonsmokers, or in homogenates of lung excisional tissue from smokers with or without primary lung cancer.

A patient presenting with apparent pulmonary edema in whom transient, large increases in systemic arterial, pulmonary arterial, and pulmonary capillary wedge pressures occurred and spontaneously resolved within a few minutes is presented. This appears to lend support to the pressure and volume overload theory of the pathogenesis of neurogenic pulmonary edema that previously has not been demonstrated in humans.

In a 25 year old man shock and pulmonary edema developed following the intravenous administration of meglumine diatrizoate for an intravenous pyelogram. A pulmonary-capillary wedge pressure of 3 torr and a high protein content in the pulmonary edema effluent confirmed the diagnosis of noncardiogenic pulmonary edema. We suggest that the basis of the increased capillary permeability in this patient may be related to an immunologic reaction to intravenous pyelogram dye.

Aryl hydrocarbon hydroxylase (AHH) activity was measured in pulmonary alveolar macrophages (PAMs) and peripheral blood lymphocytes from cigarette smokers with and without primary lung cancer. Frequency distribution analysis of AHH induction ratios for the two groups revealed an increased number of individuals in the lung cancer patient group with high lymphocyte induction values (P less than 0.05). A similar increase was not shown for high-PAM AHH values in lung cancer patients (P greater than 0.2). When individual PAM and lymphocyte AHH values were compared between noncancer and lung cancer patients, a positive correlation was observed for noncancer patients (r=0.195, P less than 0.001), but no correlation of these values was noted for lung cancer patients. The lung cancer patients were divided into three subgroups of patients showing (I) high PAM and low lymphocyte AHH levels, (II) low PAM and low lymphocyte AHH levels, and (III) low PAM and high lymphocyte AHH levels. When the incidence of family history of cancer was compared for these subgroups, no family cancer history was recorded for persons in subgroup II; however, individuals in subgroups I and III presented family cancer history incidence of 9.5% and 39.3%, respectively. Patients in group III averaged 6 years younger than those in group I. These data suggest that familial factors may be identified among lung cancer patients and that these factors appear to associate as either a cause of an effect with the capacity of pulmonary alveolar macrophages and lymphocytes to be induced for AHH. The data support the hypothesis that high AHH values may be characteristic of lung cancer patients but show that enzyme values determined from a single tissue, either PAMs or lymphocytes, may not be appropriate for showing whether high AHH inducibility is correlated with lung cancer.

We performed a prospective study of 250 men undergoing transurethral resection of the prostate to determine the incidence of perioperative myocardial infarction. The prevalence of coronary artery disease in the study group was 27%. Patients had measurement of total creatine kinase and its MB isoenzyme and electrocardiography preoperatively and on the first three postoperative days. Only one myocardial infarction was diagnosed, an incidence rate of 0.4%. The overall rate of serious post-operative complications was 3.6%. No deaths occurred during the operative hospitalization. We conclude that with transurethral resection perioperative myocardial infarction is a rare event despite the high prevalence of coronary artery disease in this surgical population. Routine postoperative surveillance with electrocardiograms and creatine kinase determinations in asymptomatic patients is not warranted.

A significant portion of internal medicine residency training in the United States today occurs in general medicine sections of Veterans Administration hospitals. The authors studied the demographic, diagnostic, and prognostic characteristics of the patients treated by house staff members rotating through the general medical wards of the Houston, Texas, Veterans Administration (VA) Medical Center. In 2,131 admissions over 13 months, the most frequent primary causes of admissions included congestive heart failure, pneumonia, exacerbation of chronic obstructive lung disease, and malignancy. In 85 percent of the admissions, two or more chronic diseases were present. In 53 percent of admissions, the patients were deemed to be moderately or severely ill on admission. The data indicate that the residents' experience is representative of problems encountered by practicing internists and that VA hospitals make a significant contribution to internal medicine training and thus to the provision of health care for the nation.

• We performed an observational study to determine the prevalence of severe dementia in a general medicine unit, the categories of acute medical care provided to these patients, the process by which treatment decisions are made, and their cost implications. The prevalence of severe dementia was 4.4%. The patients from whom some form of acute medical care was withheld (26 [45.6%] of 57) were more severely ill at admission and had a mortality rate five times higher than those who received full care. Physicians cited family wishes in 75.9% of the decisions to limit care but in only 10.9% of the decisions to give full care. The only differences in charges incurred were due to differential mortality rates in individuals from whom care was withheld. We recommend that hospitals develop and implement protocols for decision making in the care of the severely demented to promote open discussions among providers and families and to increase family contributions to decision making. We believe that the extension of this consultative approach to decisions involving severely demented patients may have the virtue of combining more humane care with more cost-effective care. ( Arch Intern Med 1988;148:1980-1984)