Antibiotic resistance genes as emerging contaminants: studies in northern Colorado

This study explores antibiotic resistance genes (ARGs) as emerging environmental contaminants. The purpose of this study was to investigate the occurrence of ARGs in various environmental compartments in northern Colorado, including Cache La Poudre (Poudre) River sediments, irrigation ditches, dairy lagoons, and the effluents of wastewater recycling and drinking water treatment plants. Additionally, ARG concentrations in the Poudre River sediments were analyzed at three time points at five sites with varying levels of urban/agricultural impact and compared with two previously published time points. It was expected that ARG concentrations would be significantly higher in environments directly impacted by urban/agricultural activity than in pristine and lesser-impacted environments. Polymerase chain reaction (PCR) detection assays were applied to detect the presence/absence of several tetracycline and sulfonamide ARGs. Quantitative real-time PCR was used to further quantify two tetracycline ARGs (tet(W) and tet(O)) and two sulfonamide ARGs (sul(I) and sul(II)). The following trend was observed with respect to ARG concentrations (normalized to eubacterial 16S rRNA genes): dairy lagoon water > irrigation ditch water > urban/agriculturally impacted river sediments (p < 0.0001), except for sul(II), which was absent in ditch water. It was noted that tet(W) and tet(O) were also present in treated drinking water and recycled wastewater, suggesting that these are potential pathways for the spread of ARGs to and from humans. On the basis of this study, there is a need for environmental scientists and engineers to help address the issue of the spread of ARGs in the environment.

Probable non-vector-borne transmission of Zika virus, Colorado, USA

Clinical and serologic evidence indicate that 2 American scientists contracted Zika virus infections while working in Senegal in 2008. One of the scientists transmitted this arbovirus to his wife after his return home. Direct contact is implicated as the transmission route, most likely as a sexually transmitted infection.

Informed decision making in outpatient practice: time to get back to basics

CONTEXT

Many clinicians have called for an increased emphasis on the patient's role in clinical decision making. However, little is known about the extent to which physicians foster patient involvement in decision making, particularly in routine office practice.

OBJECTIVE

To characterize the nature and completeness of informed decision making in routine office visits of both primary care physicians and surgeons.

DESIGN

Cross-sectional descriptive evaluation of audiotaped office visits during 1993.

SETTING AND PARTICIPANTS

A total of 1057 encounters among 59 primary care physicians (general internists and family practitioners) and 65 general and orthopedic surgeons; 2 to 12 patients were recruited from each physician's community-based private office.

MAIN OUTCOME MEASURES

Analysis of audiotaped patient-physician discussions for elements of informed decision making, using criteria that varied with the level of decision complexity: basic (eg, laboratory test), intermediate (eg, new medication), or complex (eg, procedure). Criteria for basic decisions included discussion of the nature of the decision and asking the patient to voice a preference; other categories had criteria that were progressively more stringent.

RESULTS

The 1057 audiotaped encounters contained 3552 clinical decisions. Overall, 9.0% of decisions met our definition of completeness for informed decision making. Basic decisions were most often completely informed (17.2%), while no intermediate decisions were completely informed, and only 1 (0.5%) complex decision was completely informed. Among the elements of informed decision making, discussion of the nature of the intervention occurred most frequently (71 %) and assessment of patient understanding least frequently (1.5%).

CONCLUSIONS

Informed decision making among this group of primary care physicians and surgeons was often incomplete. This deficit was present even when criteria for informed decision making were tailored to expect less extensive discussion for decisions of lower complexity. These findings signal the need for efforts to encourage informed decision making in clinical practice.

Electrical wiring configurations and childhood cancer

An excess of electrical wiring configurations suggestive of high current-flow was noted in Colorado in 1976--1977 near the homes of children who developed cancer, as compared to the homes of control children. The finding was strongest for children who had spent their entire lives at the same address, and it appeared to be dose-related. It did not seem to be an artifact of neighborhood, street congestion, social class, or family structure. The reason for the correlation is uncertain; possible effects of current in the water pipes or of AC magnetic fields are suggested.

The incidence and nature of surgical adverse events in Colorado and Utah in 1992

BACKGROUND

Despite more than three decades of research on iatrogenesis, surgical adverse events have not been subjected to detailed study to identify their characteristics. This information could be invaluable, however, for guiding quality assurance and research efforts aimed at reducing the occurrence of surgical adverse events. Thus we conducted a retrospective chart review of 15,000 randomly selected admissions to Colorado and Utah hospitals during 1992 to identify and analyze these events.

METHODS

We selected a representative sample of hospitals from Utah and Colorado and then randomly sampled 15,000 nonpsychiatric discharges from 1992. With use of a 2-stage record-review process modeled on previous adverse event studies, we estimated the incidence, morbidity, and preventability of surgical adverse events that caused death, disability at the time of discharge, or prolonged hospital stay. We characterized their distribution by type of injury and by physician specialty and determined incidence rates by procedure.

RESULTS

Adverse events were no more likely in surgical care than in nonsurgical care. Nonetheless, 66% of all adverse events were surgical, and the annual incidence among hospitalized patients who underwent an operation or child delivery was 3.0% (confidence interval 2.7% to 3.4%). Among surgical adverse events 54% (confidence interval 48.9% to 58.9%) were preventable. We identified 12 common operations with significantly elevated adverse event incidence rates that ranged from 4.4% for hysterectomy (confidence interval 2.9% to 6.8%) to 18.9% for abdominal aortic aneurysm repair (confidence interval 8.3% to 37.5%). Eight operations also carried a significantly higher risk of a preventable adverse event: lower extremity bypass graft (11.0%), abdominal aortic aneurysm repair (8.1%), colon resection (5.9%), coronary artery bypass graft/cardiac valve surgery (4.7%), transurethral resection of the prostate or of a bladder tumor (3.9%), cholecystectomy (3.0%), hysterectomy (2.8%), and appendectomy (1.5%). Among all surgical adverse events, 5.6% (confidence interval 3.7% to 8.3%) resulted in death, accounting for 12.2% (confidence interval 6.9% to 21.4%) of all hospital deaths in Utah and Colorado. Technique-related complications, wound infections, and postoperative bleeding produced nearly half of all surgical adverse events.

CONCLUSION

These findings provide direction for research to identify the causes of surgical adverse events and for targeted quality improvement efforts.

Effect of river landscape on the sediment concentrations of antibiotics and corresponding antibiotic resistance genes (ARG)

The purpose of this study was to quantify antibiotic resistance genes (ARG) in the sediments of the mixed-landscape Cache La Poudre River, which has previously been studied and shown to have high concentrations of antibiotics related to urban and agricultural activities. River sediments were sampled during two events (high-flow and low-flow) from five sites with varying urban and agricultural impact levels. Polymerase-chain-reaction (PCR) detection assays were conducted for four sulfonamide resistance gene families, using newly designed primers, and five tetracycline resistance gene families, using previously published primers. Sul(I), sul(II), tet(W), and tet(O) gene families were further quantified by real-time quantitative polymerase chain reaction (Q-PCR). Resistance to four classes of antibiotics (tetracyclines, sulfonamides, ionophores, and macrolides) was also investigated using a culture-based approach. The quantities of resistance genes normalized to the 16S gene copy number were significantly different between the sites, with higher resistance gene concentrations at the impacted sites than at the pristine site. Total resistant CFUs were over an order of magnitude lower at the pristine site, but differences were less apparent when normalized to the total CFUs. Six tetracyclines and six sulfonamides were also quantified in the sediments and were found to be highest at sites impacted by urban and agricultural activity, with no antibiotics detected at the pristine sit. To the knowledge of the authors, this study is the first to demonstrate a relationship between urban and agricultural activity and microbial resistance in river sediments using quantitative molecular tools.

Adult respiratory distress syndrome: risk with common predispositions

A 1-year survey of patients in three hospitals identified 936 patients who had one predisposition and 57 who had several predispositions to the adult respiratory distress syndrome. From the total predisposed population of 993 patients, 68 subsequently developed the syndrome. An additional 20 patients developed the syndrome from causes other than eight identified predispositions, to bring the total of patients studied to 88. A highly significant difference (p less than 0.0001) was found in the incidence rates of the syndrome between patients with one and several predispositions (5.8 versus 24.6 per 100 patients). Within 72 hours of onset of predisposition, 89.5% of patients who developed the syndrome had been intubated and placed on mechanical ventilation. Fifty-seven of the 88 patients (64.8%) with the syndrome died. By the 14th day 90% of deaths had occurred. There were no age- or sex-specific differences in either incidence or mortality rates. Case fatality rates of the syndrome were high in all predisposed groups.

Validation of self-reported chronic conditions and health services in a managed care population

BACKGROUND

Self-reported data are commonly used to estimate the prevalence of health conditions and the use of preventive health services in the population, but the validity of such data is often questioned.

METHODS

The Behavioral Risk Factor Survey (BRFS) was admin istered by telephone to a stratified, random sample of health maintenanc e organization (HMO) subscribers in Colorado in 1993, and self-reports w ere compared with HMO medical records for 599 adults aged >21. Sensitivity and specificity were calculated for three chronic conditions and use of six preventive services.

RESULTS

Sensitivity was highest for hypertension (83%), moderate for diabetes (73%), and lowest for hypercholesterolemia (59%); specificity was >80% for all three conditions. Sensitivity ranged from 86% to 99% for influenza immunization, clinical breast examination, blood cholesterol screening, mammography, Pap test, and blood pressure screening; specificity was <75% for all preventive services.

CONCLUSIONS

Self-reports are reasonably accurate for certain chronic conditions and for routine screening exams and can provide a useful estimate for broad measures of population prevalence.

Medication errors observed in 36 health care facilities

BACKGROUND

Medication errors are a national concern.

OBJECTIVE

To identify the prevalence of medication errors (doses administered differently than ordered).

DESIGN

A prospective cohort study.

SETTING

Hospitals accredited by the Joint Commission on Accreditation of Healthcare Organizations, nonaccredited hospitals, and skilled nursing facilities in Georgia and Colorado.

PARTICIPANTS

A stratified random sample of 36 institutions. Twenty-six declined, with random replacement. Medication doses given (or omitted) during at least 1 medication pass during a 1- to 4-day period by nurses on high medication-volume nursing units. The target sample was 50 day-shift doses per nursing unit or until all doses for that medication pass were administered.

METHODS

Medication errors were witnessed by observation, and verified by a research pharmacist (E.A.F.). Clinical significance was judged by an expert panel of physicians.

MAIN OUTCOME MEASURE

Medication errors reaching patients.

RESULTS

In the 36 institutions, 19% of the doses (605/3216) were in error. The most frequent errors by category were wrong time (43%), omission (30%), wrong dose (17%), and unauthorized drug (4%). Seven percent of the errors were judged potential adverse drug events. There was no significant difference between error rates in the 3 settings (P =.82) or by size (P =.39). Error rates were higher in Colorado than in Georgia (P =.04)

CONCLUSIONS

Medication errors were common (nearly 1 of every 5 doses in the typical hospital and skilled nursing facility). The percentage of errors rated potentially harmful was 7%, or more than 40 per day in a typical 300-patient facility. The problem of defective medication administration systems, although varied, is widespread.

Physician-patient communication. The relationship with malpractice claims among primary care physicians and surgeons

OBJECTIVE

To identify specific communication behaviors associated with malpractice history in primary care physicians and surgeons.

DESIGN

Comparison of communication behaviors of "claims" vs "no-claims" physicians using audiotapes of 10 routine office visits per physician.

SETTINGS

One hundred twenty-four physician offices in Oregon and Colorado.

PARTICIPANTS

Fifty-nine primary care physicians (general internists and family practitioners) and 65 general and orthopedic surgeons and their patients. Physicians were classified into no-claims or claims (> or =2 lifetime claims) groups based on insurance company records and were stratified by years in practice and specialty.

MAIN OUTCOME MEASURES

Audiotape analysis using the Roter Interaction Analysis System.

RESULTS

Significant differences in communication behaviors of no-claims and claims physicians were identified in primary care physicians but not in surgeons. Compared with claims primary care physicians, no-claims primary care physicians used more statements of orientation (educating patients about what to expect and the flow of a visit), laughed and used humor more, and tended to use more facilitation (soliciting patients' opinions, checking understanding, and encouraging patients to talk). No-claims primary care physicians spent longer in routine visits than claims primary care physicians (mean, 18.3 vs 15.0 minutes), and the length of the visit had an independent effect in predicting claims status. The multivariable model for primary care improved the prediction of claims status by 57% above chance (90% confidence interval, 33%-73%). Multivariable models did not significantly improve prediction of claims status for surgeons.

CONCLUSIONS

Routine physician-patient communication differs in primary care physicians with vs without prior malpractice claims. In contrast, the study did not find communication behaviors to distinguish between claims vs no-claims surgeons. The study identifies specific and teachable communication behaviors associated with fewer malpractice claims for primary care physicians. Physicians can use these findings as they seek to improve communication and decrease malpractice risk. Malpractice insurers can use this information to guide malpractice risk prevention and education for primary care physicians but should not assume that it is appropriate to teach similar behaviors to other specialty groups.

The ecology of the phyllosphere: geographic and phylogenetic variability in the distribution of bacteria on tree leaves

Large populations of bacteria live on leaf surfaces and these phyllosphere bacteria can have important effects on plant health. However, we currently have a limited understanding of bacterial diversity on tree leaves and the inter- and intra-specific variability in phyllosphere community structure. We used a barcoded pyrosequencing technique to characterize the bacterial communities from leaves of 56 tree species in Boulder, Colorado, USA, quantifying the intra- and inter-individual variability in the bacterial communities from 10 of these species. We also examined the geographic variability in phyllosphere communities on Pinus ponderosa from several locations across the globe. Individual tree species harboured high levels of bacterial diversity and there was considerable variability in community composition between trees. The bacterial communities were organized in patterns predictable from the relatedness of the trees as there was significant correspondence between tree phylogeny and bacterial community phylogeny. Inter-specific variability in bacterial community composition exceeded intra-specific variability, a pattern that held even across continents where we observed minimal geographic differentiation in the bacterial communities on P. ponderosa needles.

Campylobacter enteritis: clinical and epidemiologic features

Campylobacter fetus subspecies (ssp.) jejuni has been recently recognized to cause diarrheal disease in man. To assess its importance as an enteric pathogen, we prospectively studied 514 patients with diarrhea. Campylobacter fetus ssp. jejuni was isolated from the feces of 26 patients (5%) and seven of 11 of their symptomatic household contacts. This organism was isolated from the feces of only one of 18 asymptomatic household contacts and not at all from 157 other healthy persons. Seventeen of 20 patients from whom C. fetus ssp. jejuni was isolated from fecal culture showed at least a fourfold rise in specific IgG titers. Review of 35 cases of campylobacter enteritis identified a typical clinical syndrome with acute onset of diarrhea, abdominal pain, fever, and constitutional symptoms. Stool examination revealed blood in 60% and polymorphonuclear leukocytes in 78% of patients. Epidemiologic investigation strongly suggested an external source for the infection in 22 of 35 patients.

Case-control study of childhood cancer and exposure to 60-Hz magnetic fields

Concern with health effects of extremely low frequency magnetic fields has been raised by epidemiologic studies of childhood cancer in relation to proximity to electric power distribution lines. This case-control study was designed to assess the relation between residential exposure to magnetic fields and the development of childhood cancer. Eligible cases consisted of all 356 residents of the five-county 1970 Denver, Colorado Standard Metropolitan Statistical Area aged 0-14 years who were diagnosed with any form of cancer between 1976 and 1983. Controls were selected by random digit dialing to approximate the case distribution by age, sex, and telephone exchange area. Exposure was characterized through in-home electric and magnetic field measurements under low and high power use conditions and wire configuration codes, a surrogate measure of long-term magnetic field levels. Measured magnetic fields under low power use conditions had a modest association with cancer incidence; a cutoff score of 2.0 milligauss resulted in an odds ratio of 1.4 (95% confidence interval (CI) = 0.6-2.9) for total cancers and somewhat larger odds ratios (ORs) for leukemias (OR = 1.9), lymphomas (OR = 2.2), and soft tissue sarcomas (OR = 3.3). Neither magnetic fields (OR = 1.0) nor electric fields (OR = 0.9) under high power use conditions were related to total cancers. Wire codes associated with higher magnetic fields were more common among case than control homes. The odds ratio to contrast very high and high to very low, low, and buried wire codes was 1.5 (95% CI = 1.0-2.3) for total cases, with consistency across cancer subgroups except for brain cancer (OR = 2.0) and lymphomas (OR = 0.8). Contrasts of very high to buried wire code homes produced larger, less precise odds ratios of 2.3 for total cases, 2.9 for leukemias, and 3.3 for lymphomas. Adjusted estimates for measured fields and wire codes did not differ from crude results, indicating an absence of confounding. Limitations to the study are nonresponse (especially for field measurements), differential mobility of cases and controls, and a presumably nondifferential exposure misclassification from the use of imperfect surrogates for long-term magnetic field exposure history. In spite of these concerns, the results encourage further examination of the carcinogenic potential from this form of nonionizing radiation.

Antidepressant pharmaceuticals in two U.S. effluent-impacted streams: occurrence and fate in water and sediment, and selective uptake in fish neural tissue

Antidepressant pharmaceuticals are widely prescribed in the United States; release of municipal wastewater effluent is a primary route introducing them to aquatic environments, where little is known about their distribution and fate. Water, bed sediment, and brain tissue from native white suckers (Catostomus commersoni) were collected upstream and at points progressively downstream from outfalls discharging to two effluent-impacted streams, Boulder Creek (Colorado) and Fourmile Creek (Iowa). A liquid chromatography/tandem mass spectrometry method was used to quantify antidepressants, including fluoxetine, norfluoxetine (degradate), sertraline, norsertraline (degradate), paroxetine, citalopram, fluvoxamine, duloxetine, venlafaxine, and bupropion in all three sample matrices. Antidepressants were not present above the limit of quantitation in water samples upstream from the effluent outfalls but were present at points downstream at ng/L concentrations, even at the farthest downstream sampling site 8.4 km downstream from the outfall. The antidepressants with the highest measured concentrations in both streams were venlafaxine, bupropion, and citalopram and typically were observed at concentrations of at least an order of magnitude greater than the more commonly investigated antidepressants fluoxetine and sertraline. Concentrations of antidepressants in bed sediment were measured at ng/g levels; venlafaxine and fluoxetine were the predominant chemicals observed. Fluoxetine, sertraline, and their degradates were the principal antidepressants observed in fish brain tissue, typically at low ng/g concentrations. A qualitatively different antidepressant profile was observed in brain tissue compared to streamwater samples. This study documents that wastewater effluent can be a point source of antidepressants to stream ecosystems and that the qualitative composition of antidepressants in brain tissue from exposed fish differs substantially from the compositions observed in streamwater and sediment, suggesting selective uptake.

The Graduate Nurse Experience

Graduate nurses experience stress transitioning from student to practicing professional nurse, moving from a familiar educational environment into the workforce, where expectations are to rapidly function as a competent nurse. This study identified the stresses and challenges experienced by cohorts of graduate nurses working in 6 acute care hospitals, during specific timed data periods, to better understand factors that may influence graduate nurse retention. Results report graduate nurses do not feel skilled, comfortable, and confident for as long as 1 year after being hired, highlighting the need for healthcare organizations to provide extended orientation and support programs to facilitate successful entry into practice.

Correlation between upstream human activities and riverine antibiotic resistance genes

Antimicrobial resistance remains a serious and growing human health challenge. The water environment may represent a key dissemination pathway of resistance elements to and from humans. However, quantitative relationships between landscape features and antibiotic resistance genes (ARGs) have not previously been identified. The objective of this study was to examine correlations between ARGs and putative upstream anthropogenic sources in the watershed. sul1 (sulfonamide) and tet(W) (tetracycline) were measured using quantitative polymerase chain reaction in bed and suspended sediment within the South Platte River Basin, which originates from a pristine region in the Rocky Mountains and runs through a gradient of human activities. A geospatial database was constructed to delineate surface water pathways from animal feeding operations, wastewater treatment plants, and fish hatchery and rearing units to river monitoring points. General linear regression models were compared. Riverine sul1 correlated with upstream capacities of animal feeding operations (R(2) = 0.35, p < 0.001) and wastewater treatment plants (R(2) = 0.34, p < 0.001). Weighting for the inverse distances from animal feeding operations along transport pathways strengthened the observed correlations (R(2) = 0.60-0.64, p < 0.001), suggesting the importance of these pathways in ARG dissemination. Correlations were upheld across the four sampling events during the year, and averaging sul1 measurements in bed and suspended sediments over all events yielded the strongest correlation (R(2) = 0.92, p < 0.001). Conversely, a significant relationship with landscape features was not evident for tet(W), which, in contrast to sul1, is broadly distributed in the pristine region and also relatively more prevalent in animal feeding operation lagoons. The findings highlight the need to focus attention on quantifying the contribution of water pathways to the antibiotic resistance disease burden in humans and offer insight into potential strategies to control the spread of ARGs.

Temporal and spatial trends in the occurrence of human and veterinary antibiotics in aqueous and river sediment matrices

The occurrence of 15 antibiotics belonging to three different groups, tetracyclines (TCs), sulfonamides (SAs), and macrolides (MLs), mainly used to prevent or treat illness for humans and also to control disease or to promote the growth for animals was studied in aqueous and sediment matrices. The result of spatial and temporal statistical analysis revealed that measured concentrations of individual antibiotics were significantly different depending on sampling location and time periods for aqueous and sediment samples. High concentrations of human-used antibiotics were detected downstream of a wastewater treatment plant, and animal-used antibiotics were mainly found in a region with significant agricultural activity. Generally, the highest concentrations of antibiotics for both water and sediment samples were measured in winter indicating that low flow conditions and cold-water temperatures might enhance the persistence of these compounds. Furthermore, a pseudo-partitioning coefficient(P-PC) was introduced to provide a better understanding of the partitioning of antibiotics into the sediment. Different P-PC values were found depending on the sorption characteristics of the individual antibiotics. Sediment samples showed a greater detection frequency and a much higher concentration compared to aqueous samples taken at the same site. Since microorganism antibiotic resistance can develop in sediments, the importance of analyzing this matrix is underscored.

Opioid Use and Storage Patterns by Patients after Hospital Discharge following Surgery

INTRODUCTION

Opioid-based analgesic therapy represents a cornerstone of pain management after surgery. The recent rise in opioid sales and opioid overdoses suggests it is important to maximize the safety of opioid prescribing after surgery. Given that patients may live with other family members in the home, safe storage and appropriate disposal of excess opioids after hospital discharge are necessary to prevent unintended secondary exposures. Identifying characteristics of patients who are likely to be prescribed excess opioids after surgery may enable more targeted prescription practices and safety interventions. Our study aimed to elucidate patient-reported opioid use patterns and modes of home storage of opioids among patients discharged home after Cesarean section (C-section) and thoracic surgery. Specifically, we sought to identify characteristics of patients who reported using about half or more versus less of the opioids prescribed to them for use after hospital discharge.

METHODS

For this cohort study, we developed a survey on quality of analgesia following hospital discharge, amounts of opioids taken relative to the amount prescribed, reasons for not taking all prescribed medications, and storage and disposal methods for leftover opioids. Adult patients, who had C-section or thoracic surgery at a tertiary academic medical center, were given a web-based self-administered survey after discharge. Descriptive statistics (means and standard deviations, proportions) were used to describe the study sample and survey results. Comparisons between patients who reported taking about half or more versus less of the opioids prescribed to them for use after hospital discharge were made using unpaired t-tests, Mann-Whitney tests, and Chi-square tests as appropriate.

RESULTS

The majority (53%) of respondents after C-section (N = 30) reported taking either no or very few (less than 5) prescribed opioid pills; 83% reported taking half or less; and 17% of women, reported taking all or nearly all (5 or fewer pills left over) of their opioid prescription. In a cohort of patients after thoracic surgery (n = 31) 45% reported taking either no or very few (5 or less) prescribed opioid pills; 71% reported taking half or less; and 29% of patients reported taking all or nearly all (5 or fewer pills left over) of their opioid prescription. In both cohorts, use of opioids while hospitalized was higher in the group reporting using about half or more of prescribed opioids after discharge. Leftover opioids were stored in an unlocked location in 77% and 73% of cases following C-section and thoracic surgery, respectively.

CONCLUSION

Our findings from surveys in two distinct patient populations at a single academic medical center suggest that current opioid prescribing practices for pain management at hospital discharge following Cesarean section and thoracic surgery may not account for individual patients' analgesic requirements. Excess opioid pills are commonly stored in unsecured locations and represent a potential source for non-medical opioid use and associated morbidity and mortality in patients and their families. Research to develop goal-directed and patient-centered post-discharge opioid prescription practices and encourage opioid safety practices after surgery is needed.

Model averaging and muddled multimodel inferences

Three flawed practices associated with model averaging coefficients for predictor variables in regression models commonly occur when making multimodel inferences in analyses of ecological data. Model-averaged regression coefficients based on Akaike information criterion (AIC) weights have been recommended for addressing model uncertainty but they are not valid, interpretable estimates of partial effects for individual predictors when there is multicollinearity among the predictor variables. Multicollinearity implies that the scaling of units in the denominators of the regression coefficients may change across models such that neither the parameters nor their estimates have common scales, therefore averaging them makes no sense. The associated sums of AIC model weights recommended to assess relative importance of individual predictors are really a measure of relative importance of models, with little information about contributions by individual predictors compared to other measures of relative importance based on effects size or variance reduction. Sometimes the model-averaged regression coefficients for predictor variables are incorrectly used to make model-averaged predictions of the response variable when the models are not linear in the parameters. I demonstrate the issues with the first two practices using the college grade point average example extensively analyzed by Burnham and Anderson. I show how partial standard deviations of the predictor variables can be used to detect changing scales of their estimates with multicollinearity. Standardizing estimates based on partial standard deviations for their variables can be used to make the scaling of the estimates commensurate across models, a necessary but not sufficient condition for model averaging of the estimates to be sensible. A unimodal distribution of estimates and valid interpretation of individual parameters are additional requisite conditions. The standardized estimates or equivalently the t statistics on unstandardized estimates also can be used to provide more informative measures of relative importance than sums of AIC weights. Finally, I illustrate how seriously compromised statistical interpretations and predictions can be for all three of these flawed practices by critiquing their use in a recent species distribution modeling technique developed for predicting Greater Sage-Grouse (Centrocercus urophasianus) distribution in Colorado, USA. These model averaging issues are common in other ecological literature and ought to be discontinued if we are to make effective scientific contributions to ecological knowledge and conservation of natural resources.

Climate change is affecting altitudinal migrants and hibernating species

Calendar date of the beginning of the growing season at high altitude in the Colorado Rocky Mountains is variable but has not changed significantly over the past 25 years. This result differs from growing evidence from low altitudes that climate change is resulting in a longer growing season, earlier migrations, and earlier reproduction in a variety of taxa. At our study site, the beginning of the growing season is controlled by melting of the previous winter's snowpack. Despite a trend for warmer spring temperatures the average date of snowmelt has not changed, perhaps because of the trend for increased winter precipitation. This disjunction between phenology at low and high altitudes may create problems for species, such as many birds, that migrate over altitudinal gradients. We present data indicating that this already may be true for American robins, which are arriving 14 days earlier than they did in 1981; the interval between arrival date and the first date of bare ground has grown by 18 days. We also report evidence for an effect of climate change on hibernation behavior; yellow-bellied marmots are emerging 38 days earlier than 23 years ago, apparently in response to warmer spring air temperatures. Migrants and hibernators may experience problems as a consequence of these changes in phenology, which may be exacerbated if climate models are correct in their predictions of increased winter snowfall in our study area. The trends we report for earlier formation of permanent snowpack and for a longer period of snow cover also have implications for hibernating species.

Obesity-associated hypoventilation in hospitalized patients: prevalence, effects, and outcome

BACKGROUND

Severe obesity is associated with hypoventilation, a disorder that may adversely affect morbidity and mortality. We sought to determine the prevalence and effects of obesity-associated hypoventilation in hospitalized patients.

METHODS

Consecutive admissions to internal medicine services were screened over a 6-month period. In all eligible subjects with severe obesity (body mass index > or =35 kg/m2), we administered a sleep questionnaire, and performed neuropsychological, arterial blood gas, and pulmonary function testing. Hospital course and mortality at 18 months was also determined.

RESULTS

Of 4,332 admissions, 6% (n = 277) of patients were severely obese, of whom 150 were enrolled, 75 refused to participate, and 52 met the exclusion criteria. Hypoventilation (mean [+/- SD] arterial partial pressure of carbon dioxide [PaCO2], 52 +/- 7 mm Hg) was present in 31% (n = 47) of subjects who did not have other reasons for hypercapnia. Decreased objective attention/concentration and increased subjective sleepiness were present in patients with obesity-associated hypoventilation compared with in severely obese hospitalized patients without hypoventilation (simple obesity group; mean PaCO2, 37 +/- 6 mm Hg). There were higher rates of intensive care (P = 0.08), long-term care at discharge (P = 0.01), and mechanical ventilation (P = 0.01) among subjects with obesity-associated hypoventilation. Therapy for hypoventilation at discharge was initiated in only 6 (13%) of the patients with obesity-associated hypoventilation. At 18 months following hospital discharge, mortality was 23% in the obesity-associated hypoventilation group as compared with 9% in the simple obesity group (hazard ratio = 4.0; 95% confidence interval: 1.5 to 10.4].

CONCLUSION

Hypoventilation frequently complicates severe obesity among hospitalized adults and is associated with excess morbidity and mortality.