Identifying incarceration status in the electronic health record using large language models in emergency department settings.

Background: Incarceration is a significant social determinant of health, contributing to high morbidity, mortality, and racialized health inequities. However, incarceration status is largely invisible to health services research due to inadequate clinical electronic health record (EHR) capture. This study aims to develop, train, and validate natural language processing (NLP) techniques to more effectively identify incarceration status in the EHR. Methods: The study population consisted of adult patients (≥ 18 y.o.) who presented to the emergency department between June 2013 and August 2021. The EHR database was filtered for notes for specific incarceration-related terms, and then a random selection of 1,000 notes was annotated for incarceration and further stratified into specific statuses of prior history, recent, and current incarceration. For NLP model development, 80% of the notes were used to train the Longformer-based and RoBERTa algorithms. The remaining 20% of the notes underwent analysis with GPT-4. Results: There were 849 unique patients across 989 visits in the 1000 annotated notes. Manual annotation revealed that 559 of 1000 notes (55.9%) contained evidence of incarceration history. ICD-10 code (sensitivity: 4.8%, specificity: 99.1%, F1-score: 0.09) demonstrated inferior performance to RoBERTa NLP (sensitivity: 78.6%, specificity: 73.3%, F1-score: 0.79), Longformer NLP (sensitivity: 94.6%, specificity: 87.5%, F1-score: 0.93), and GPT-4 (sensitivity: 100%, specificity: 61.1%, F1-score: 0.86). Conclusions: Our advanced NLP models demonstrate a high degree of accuracy in identifying incarceration status from clinical notes. Further research is needed to explore their scaled implementation in population health initiatives and assess their potential to mitigate health disparities through tailored system interventions.

Correction: How Does ChatGPT Perform on the United States Medical Licensing Examination (USMLE)? The Implications of Large Language Models for Medical Education and Knowledge Assessment.

[This corrects the article DOI: 10.2196/45312.].

How Does ChatGPT Perform on the United States Medical Licensing Examination (USMLE)? The Implications of Large Language Models for Medical Education and Knowledge Assessment.

BACKGROUND: Chat Generative Pre-trained Transformer (ChatGPT) is a 175-billion-parameter natural language processing model that can generate conversation-style responses to user input. OBJECTIVE: This study aimed to evaluate the performance of ChatGPT on questions within the scope of the United States Medical Licensing Examination (USMLE) Step 1 and Step 2 exams, as well as to analyze responses for user interpretability. METHODS: We used 2 sets of multiple-choice questions to evaluate ChatGPT's performance, each with questions pertaining to Step 1 and Step 2. The first set was derived from AMBOSS, a commonly used question bank for medical students, which also provides statistics on question difficulty and the performance on an exam relative to the user base. The second set was the National Board of Medical Examiners (NBME) free 120 questions. ChatGPT's performance was compared to 2 other large language models, GPT-3 and InstructGPT. The text output of each ChatGPT response was evaluated across 3 qualitative metrics: logical justification of the answer selected, presence of information internal to the question, and presence of information external to the question. RESULTS: Of the 4 data sets, AMBOSS-Step1, AMBOSS-Step2, NBME-Free-Step1, and NBME-Free-Step2, ChatGPT achieved accuracies of 44% (44/100), 42% (42/100), 64.4% (56/87), and 57.8% (59/102), respectively. ChatGPT outperformed InstructGPT by 8.15% on average across all data sets, and GPT-3 performed similarly to random chance. The model demonstrated a significant decrease in performance as question difficulty increased (P=.01) within the AMBOSS-Step1 data set. We found that logical justification for ChatGPT's answer selection was present in 100% of outputs of the NBME data sets. Internal information to the question was present in 96.8% (183/189) of all questions. The presence of information external to the question was 44.5% and 27% lower for incorrect answers relative to correct answers on the NBME-Free-Step1 (P<.001) and NBME-Free-Step2 (P=.001) data sets, respectively. CONCLUSIONS: ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering. By performing at a greater than 60% threshold on the NBME-Free-Step-1 data set, we show that the model achieves the equivalent of a passing score for a third-year medical student. Additionally, we highlight ChatGPT's capacity to provide logic and informational context across the majority of answers. These facts taken together make a compelling case for the potential applications of ChatGPT as an interactive medical education tool to support learning.

Implementation evaluation of a stepped approach to home care assessment using interRAI systems in Ontario, Canada.

In Ontario, new home care clients are screened with the interRAI Contact Assessment and only those expected to require longer-term services receive the comprehensive RAI-Home Care assessment. Although Ontario adopted this two-step approach in 2010, it is unknown whether the assessment guidelines were implemented as intended. To evaluate implementation fidelity, the purpose of this study is to compare expected to actual client profiles and care co-ordinator practice patterns. We linked interRAI CA and RAI-HC assessments and home care referrals and services data for a retrospective cohort of adult home care clients admitted in FY 2016/17. All assessments were done by trained health professionals as part of routine practice. Descriptive analyses were used to evaluate congruency between recommended and actual practice. Adjusted cause-specific hazards and logistic approaches were used to examine time to RAI-HC assessment and being a high-priority client. Of 225,989 unique home care clients admitted to the publicly funded home care program, about three-quarters of clients were assessed with the interRAI CA only (27.9% completed the Preliminary Screener only and 46.6% completed both the Preliminary Screener and Clinical Evaluation). There was substantial agreement between the skip logic and completion of the Clinical Evaluation section (Cohen's kappa = 0.67 [95% CI: 0.66-0.67]). One-quarter of clients were assessed with both the interRAI CA and RAI-HC. As expected, RAI-HC assessed clients were older, reported more health needs, and often received home care services for >6 months. Clients in higher Assessment Urgency Algorithm (AUA) levels were significantly more likely to receive a RAI-HC assessment and be assigned to a higher home care priority level; however, 28.3% of clients in the highest AUA level did not receive a RAI-HC assessment. We conclude that the use of the interRAI CA and RAI-HC balances the investment of time and resources with the information and tools to deliver high-quality, holistic, and client-centred care. The interRAI CA guides the care co-ordinator to screen every client for a broad range of possible needs and tailor further assessment to each client's unique needs. We recommend integrating the AUA into provincial assessment guidelines as well as developing a new quality indicator focused on measuring access to the home care system.

Weather fluctuation can override the effects of integrated nutrient management on fungal disease incidence in the rice fields in Taiwan.

Both weather fluctuation and farming system influence the epidemiology of crop diseases. However, short-term experiments are difficult to mechanistically extrapolate into long-term ecological responses. Using a mechanistic model with Bayesian inference, long-term data spanning 10 years were used to construct relationships among weather fluctuation (temperature, relative humidity, wind, and rainfall), farming system (conventional and low-external-input farming), and crop disease in experimental rice fields in Taiwan. Conventional and low-external-input farming had similar influences on the disease incidence of rice blast. Temperature had a positive influence on the disease incidence only under high relative humidity. Rainfall positively affected the disease incidence until an optimum level of rainfall. Low-external-input farming, with a lower application of fertilizers and other sustainable nutrient management, achieved similar effects on the disease incidence to those achieved by conventional farming. This suggests that weather fluctuation may override the effect of the farming systems on fungal disease incidence in rice fields.

Patterns of home care assessment and service provision before and during the COVID-19 pandemic in Ontario, Canada.

OBJECTIVE: The objective was to compare home care episode, standardised assessment, and service patterns in Ontario's publicly funded home care system during the first wave of the COVID-19 pandemic (i.e., March to September 2020) using the previous year as reference. STUDY DESIGN AND SETTING: We plotted monthly time series data from March 2019 to September 2020 for home care recipients in Ontario, Canada. Home care episodes were linked to interRAI Home Care assessments, interRAI Contact Assessments, and home care services. Health status measures from the patient's most recent interRAI assessment were used to stratify the receipt of personal support, nursing, and occupational or physical therapy services. Significant level and slope changes were detected using Poisson, beta, and linear regression models. RESULTS: The March to September 2020 period was associated with significantly fewer home care admissions, discharges, and standardised assessments. Among those assessed with the interRAI Home Care assessment, significantly fewer patients received any personal support services. Among those assessed with either interRAI assessment and identified to have rehabilitation needs, significantly fewer patients received any therapy services. Among patients receiving services, patients received significantly fewer hours of personal support and fewer therapy visits per month. By September 2020, the rate of admissions and services had mostly returned to pre-pandemic levels, but completion of standardised assessments lagged behind. CONCLUSION: The first wave of the COVID-19 pandemic was associated with substantial changes in Ontario's publicly funded home care system. Although it may have been necessary to prioritise service delivery during a crisis situation, standardised assessments are needed to support individualised patient care and system-level monitoring. Given the potential disruptions to home care services, future studies should examine the impact of the pandemic on the health and well-being of home care recipients and their caregiving networks.

Isotopic assessment of soil N2O emission from a sub-tropical agricultural soil under varying N-inputs.

Nitrogen fertilizers result in high crop productivity but also enhance the emission of N2O, an environmentally harmful greenhouse gas. Only approximately a half of the applied nitrogen is utilized by crops and the rest is either vaporized, leached, or lost as NO, N2O and N2 via soil microbial activity. Thus, improving the nitrogen use efficiency of cropping systems has become a global concern. Factors such as types and rates of fertilizer application, soil texture, moisture level, pH, and microbial activity/diversity play important roles in N2O production. Here, we report the results of N2O production from a set of chamber experiments on an acidic sandy-loam agricultural soil under varying levels of an inorganic N-fertilizer, urea. Stable isotope technique was employed to determine the effect of increasing N-fertilizer levels on N2O emissions and identify the microbial processes involved in fertilizer N-transformation that give rise to N2O. We monitored the isotopic changes in both substrate (ammonium and nitrate) and the product N2O during the entire course of the incubation experiments. Peak N2O emissions of 122 ± 98 µg N2O-N m-2 h-1, 338 ± 49 µg N2O-N m-2 h-1 and 739 ± 296 µg N2O-N m-2 h-1 were observed for urea application rate of 40, 80, and 120 µg N g-1. The duration of emissions also increased with urea levels. The concentration and isotopic compositions of the substrates and product showed time-bound variation. Combining the observations of isotopic effects in δ15N, δ18O, and 15N site preference, we inferred co-occurrence of several microbial N2O production pathways with nitrification and/or fungal denitrification as the dominant processes responsible for N2O emissions. Besides this, dominant signatures of bacterial denitrification were observed in a second N2O emission pulse in intermediate urea-N levels. Signature of N2O consumption by reduction could be traced during declining emissions in treatment with high urea level.

Circulatory Inflammatory Mediators in the Prediction of Anti-Tuberculous Drug-Induced Liver Injury Using RUCAM for Causality Assessment.

BACKGROUND: Anti-tuberculous (TB) medications are common causes of drug-induced liver injury (DILI). Limited data are available on systemic inflammatory mediators as biomarkers for predicting DILI before treatment. We aimed to select predictive markers among potential candidates and to formulate a predictive model of DILI for TB patients. METHODS: Adult active TB patients from a prospective cohort were enrolled, and all participants received standard anti-tuberculous treatment. Development of DILI, defined as ≥5× ULN for alanine transaminase or ≥2.6× ULN of total bilirubin with causality assessment (RUCAM, Roussel Uclaf causality assessment method), was regularly monitored. Pre-treatment plasma was assayed for 15 candidates, and a set of risk prediction scores was established using Cox regression and receiver-operating characteristic analyses. RESULTS: A total of 19 (7.9%) in 240 patients developed DILI (including six carriers of hepatitis B virus) following anti-TB treatment. Interleukin (IL)-22 binding protein (BP), interferon gamma-induced protein 1 (IP-10), soluble CD163 (sCD163), IL-6, and CD206 were significant univariable factors associated with DILI development, and the former three were backward selected as multivariable factors, with adjusted hazards of 0.20 (0.07-0.58), 3.71 (1.35-10.21), and 3.28 (1.07-10.06), respectively. A score set composed of IL-22BP, IP-10, and sCD163 had an improved area under the curve of 0.744 (p < 0.001). CONCLUSIONS: Pre-treatment IL-22BP was a protective biomarker against DILI development under anti-TB treatment, and a score set by additional risk factors of IP-10 and sCD163 employed an adequate DILI prediction.

Review of epidemic, containment strategies, clinical management, and economic evaluation of COVID-19 pandemic.

The spread of the emerging pathogen, named as SARS-CoV-2, has led to an unprecedented COVID-19 pandemic since 1918 influenza pandemic. This review first sheds light on the similarity on global transmission, surges of pandemics, and the disparity of prevention between two pandemics. Such a brief comparison also provides an insight into the potential sequelae of COVID-19 based on the inference drawn from the fact that a cascade of successive influenza pandemic occurred after 1918 and also the previous experience on the epidemic of SARS and MERS occurring in 2003 and 2015, respectively. We then propose a systematic framework for elucidating emerging infectious disease (EID) such as COVID-19 with a panorama viewpoint from natural infection and disease process, public health interventions (non-pharmaceutical interventions (NPIs) and vaccine), clinical treatments and therapies (antivirals), until global aspects of health and economic loss, and economic evaluation of interventions with emphasis on mass vaccination. This review not only concisely delves for evidence-based scientific literatures from the origin of outbreak, the spread of SARS-CoV-2 to three surges of pandemic, and NPIs and vaccine uptakes but also provides a new insight into how to apply big data analytics to identify unprecedented discoveries through COVID-19 pandemic scenario embracing from biomedical to economic viewpoints.

A comparison of 3 frailty measures and adverse outcomes in the intake home care population: a retrospective cohort study.

BACKGROUND: In Ontario, Canada, nearly all home care patients are assessed with a brief clinical assessment (interRAI Contact Assessment [interRAI CA]) on admission. Our objective was to compare 3 frailty measures that can be operationalized using the interRAI CA. METHODS: We conducted a retrospective cohort study using linked patient-level assessment and administrative data for all Ontario adult (≥ 18 yr) home care patients assessed with the interRAI CA in 2014. We employed multivariable logistic models to compare the Changes in Health, End-stage disease and Signs and Symptoms Scale for the Contact Assessment (CHESS-CA), Assessment Urgency Algorithm (AUA) and the Frailty Index for the Contact Assessment (FI-CA) that was created for this study. Our outcomes of interest were death, hospital admission and emergency department visits within 90 days, and assessor-rated need for comprehensive geriatric assessment (CGA). RESULTS: In 2014, there were 228 679 unique adult home care patients in Ontario assessed with the interRAI CA. Controlling for age, sex and health region, being in a higher frailty level defined by any measure increased the likelihood of experiencing adverse outcomes. Among all assessments, CHESS-CA was best suited for predicting death and hospital admission, and either AUA or FI-CA for predicting perceived need for CGA. Previous emergency department visits were more predictive of future visits than frailty. Model fit was independent of whether the assessment was completed over the phone or in person. INTERPRETATION: Frailty measures from the interRAI CA identified patients at higher risk for death, hospital admission and perceived need for CGA. In jurisdictions where the CHESS-CA and AUA are already built into the electronic home care platform, such as Ontario, patients identified as high risk should be prioritized for proactive referral and care planning, and may benefit from greater involvement of primary care and other health professionals in the circle of care.

Dermoscopic assessment of xerosis severity, pigmentation pattern and vascular morphology in subjects with physiological aging and photoaging.

BACKGROUND: Skin aging can be classified as either physiological or pathological aging. Pathological aging is most often due to chronic sunlight exposure (photoaging). Age-dependent changes in dermoscopic features of normal skin have not previously been well defined. OBJECTIVES: We compared the dermoscopic features of skin from non-elderly subjects, elderly subjects with physiological aging, and elderly subjects with photoaging. MATERIALS AND METHODS: We enrolled 40 non-elderly subjects aged 20-60 years, 40 elderly subjects aged >60 years with physiological aging, and 40 elderly subjects aged >60 years with photoaging. Skin from the lower legs of subjects was examined by dermoscopy. RESULTS: Compared with non-elderly subjects, dermoscopic examination of elderly subjects with physiological aging and photoaging revealed various degrees of xerosis (mild: scaling limited to skin furrows; moderate: scaling extending beyond skin furrows with accentuation of skin markings; severe: plate-like scaling extending beyond skin furrows with formation of deep skin fissures). In addition, dermoscopic examination of skin from elderly subjects with photoaging showed increased prevalence of uneven pigmentation (small brown globules, reticular pigmentation, and homogeneous pigmentation in a patchy distribution) and vascular telangiectasia (linear and branching vessels). CONCLUSION: This study provides a novel dermoscopic grading system to evaluate the severity of xerosis and demonstrates the application of dermoscopy for the accurate assessment of subtle morphological changes (including pigmentation pattern and vascular structures) associated with physiological aging and photoaging.

[Distribution Characteristics and Ecological Risk Assessment of Organochlorine Pesticides in Sediments of Zhanjiang Bay].

Sixteen surface sediment samples were collected from the estuary of the Suixi river to the mouth of Zhanjiang Bay and then analyzed for organochlorine pesticides (OCPs) by GC-MS to investigate their distribution and ecological risk. The results showed that the concentrations of OCPs in the sediments ranged from nd to 189.52 ng·g-1 (mean 32.17 ng·g-1), including HCHs (mean 5.81 ng·g-1) and DDTs (mean 26.90 ng·g-1). The distribution characteristics showed that the highest OCPs concentrations were found in the estuary and the main shipping lane areas, and the concentration in the nearshore area was higher than that offshore. Source analysis indicated that the HCHs mainly originated from agricultural applications, while no industrial input was observed. Some "hot-spots" areas occurred in harbors and shipping channels, likely as a result of the presence of paint flakes. Additionally, the concentrations of DDTs were found to be higher than the limits of Chinese Marine sediment quality criteria, and p,p'-DDT was the main type of DDT, presenting inevitable adverse biological effects and high ecological risk. Compared with other bays in China, the concentrations of OCPs in this study were in the upper-median pollution level, especially in harbors and boat maintenance facility areas. High OCPs inputs may occur, and thereby represent a certain ecological risk in Zhanjiang Bay.

Adverse Events in Home Care: Identifying and Responding with interRAI Scales and Clinical Assessment Protocols.

Outcomes of adverse events in home care are varied and multifactorial. This study tested a framework combining two health measures to identify home care recipients at higher risk of long-term care placement or death within one year. Both measures come from the Resident Assessment Instrument-Home Care (RAI-HC), a standardized comprehensive clinical assessment. Persons scoring high in the Method for Assigning Priority Levels (MAPLe) algorithm and Changes in Health, End-stage disease, Signs and Symptoms (CHESS) scale were at the greatest risk of placement or death and more than twice as likely to experience either outcome earlier than others. The target group was more likely to trigger mood, social relationship, and caregiver distress issues, suggesting mental health and psychosocial interventions might help in addition to medical care and/or personal support services. Home care agencies can use this framework to identify home care patients who may require a more intensive care coordinator approach.

Derivation and validation of the Personal Support Algorithm: an evidence-based framework to inform allocation of personal support services in home and community care.

BACKGROUND: Personal support services enable many individuals to stay in their homes, but there are no standard ways to classify need for functional support in home and community care settings. The goal of this project was to develop an evidence-based clinical tool to inform service planning while allowing for flexibility in care coordinator judgment in response to patient and family circumstances. METHODS: The sample included 128,169 Ontario home care patients assessed in 2013 and 25,800 Ontario community support clients assessed between 2014 and 2016. Independent variables were drawn from the Resident Assessment Instrument-Home Care and interRAI Community Health Assessment that are standardised, comprehensive, and fully compatible clinical assessments. Clinical expertise and regression analyses identified candidate variables that were entered into decision tree models. The primary dependent variable was the weekly hours of personal support calculated based on the record of billed services. RESULTS: The Personal Support Algorithm classified need for personal support into six groups with a 32-fold difference in average billed hours of personal support services between the highest and lowest group. The algorithm explained 30.8% of the variability in billed personal support services. Care coordinators and managers reported that the guidelines based on the algorithm classification were consistent with their clinical judgment and current practice. CONCLUSIONS: The Personal Support Algorithm provides a structured yet flexible decision-support framework that may facilitate a more transparent and equitable approach to the allocation of personal support services.

Adult mortality of diseases and injuries attributable to selected metabolic, lifestyle, environmental, and infectious risk factors in Taiwan: a comparative risk assessment.

BACKGROUND: To facilitate priority-setting in health policymaking, we compiled the best available information to estimate the adult mortality (>30 years) burden attributable to 13 metabolic, lifestyle, infectious, and environmental risk factors in Taiwan. METHODS: We obtained data on risk factor exposure from nationally representative health surveys, cause-specific mortality from the National Death Registry, and relative risks from epidemiological studies and meta-analyses. We applied the comparative risk assessment framework to estimate mortality burden attributable to individual risk factors or risk factor clusters. RESULTS: In 2009, high blood glucose accounted for 14,900 deaths (95% UI: 11,850-17,960), or 10.4% of all deaths in that year. It was followed by tobacco smoking (13,340 deaths, 95% UI: 10,330-16,450), high blood pressure (11,190 deaths, 95% UI: 8,190-14,190), ambient particulate matter pollution (8,600 deaths, 95% UI: 7,370-9,840), and dietary risks (high sodium intake and low intake of fruits and vegetables, 7,890 deaths, 95% UI: 5,970-9,810). Overweight-obesity and physical inactivity accounted for 7,620 deaths (95% UI: 6,040-9,190), and 7,400 deaths (95% UI: 6,670-8,130), respectively. The cardiometabolic risk factors of high blood pressure, high blood glucose, high cholesterol, and overweight-obesity jointly accounted for 12,120 deaths (95% UI: 11,220-13,020) from cardiovascular diseases. For domestic risk factors, infections from hepatitis B virus (HBV) and hepatitis C virus (HCV) were responsible for 6,300 deaths (95% UI: 5,610-6,980) and 3,170 deaths (95% UI: 1,860-4,490), respectively, and betel nut use was associated with 1,780 deaths from oral, laryngeal, and esophageal cancer (95% UI: 1,190-2,360). The leading risk factors for years of life lost were similar, but the impact of tobacco smoking and alcohol use became larger because the attributable deaths from these risk factors occurred among young adults aged less than 60 years. CONCLUSIONS: High blood glucose, tobacco smoking, and high blood pressure are the major risk factors for deaths from diseases and injuries among Taiwanese adults. A large number of years of life would be gained if the 13 modifiable risk factors could be removed or reduced to the optimal level.

Cost-effectiveness analysis of population-based screening of hepatocellular carcinoma: Comparing ultrasonography with two-stage screening.

AIM: To assess the cost-effectiveness of two population-based hepatocellular carcinoma (HCC) screening programs, two-stage biomarker-ultrasound method and mass screening using abdominal ultrasonography (AUS). METHODS: In this study, we applied a Markov decision model with a societal perspective and a lifetime horizon for the general population-based cohorts in an area with high HCC incidence, such as Taiwan. The accuracy of biomarkers and ultrasonography was estimated from published meta-analyses. The costs of surveillance, diagnosis, and treatment were based on a combination of published literature, Medicare payments, and medical expenditure at the National Taiwan University Hospital. The main outcome measure was cost per life-year gained with a 3% annual discount rate. RESULTS: The results show that the mass screening using AUS was associated with an incremental cost-effectiveness ratio of USD39825 per life-year gained, whereas two-stage screening was associated with an incremental cost-effectiveness ratio of USD49733 per life-year gained, as compared with no screening. Screening programs with an initial screening age of 50 years old and biennial screening interval were the most cost-effective. These findings were sensitive to the costs of screening tools and the specificity of biomarker screening. CONCLUSION: Mass screening using AUS is more cost effective than two-stage biomarker-ultrasound screening. The most optimal strategy is an initial screening age at 50 years old with a 2-year inter-screening interval.

A population-specific Mini-Nutritional Assessment can effectively grade the nutritional status of stroke rehabilitation patients in Taiwan.

AIMS AND OBJECTIVES: To determine whether a modified version of the Mini-Nutritional Assessment without body mass index could effectively assess the nutritional risk status of stroke rehabilitation patients in Taiwan. BACKGROUND: The Mini-Nutritional Assessment was developed on the basis of clinical data of Western populations. Although widely used, its application to assess stroke rehabilitation patients has been limited. Further, to get best results, population-specific modifications to address anthropometric and lifestyle differences have been suggested, especially for non-Caucasian populations. DESIGN: The study assessed the nutritional status of stroke rehabilitation patients who enrolled in the Long-Term Care Service of Taipei. Strokes who were >40 years old, in the program for >1 month and cognitively able to answer the questions were recruited to participate in the study. METHODS: An on-site in-person interview with structured questionnaire elicited information on personal data, disease history and healthcare use and answers to the Mini-Mental State Examination, the Activities of Daily Living and the Mini-Nutritional Assessment. Patient's nutritional status was assessed with the Mini-Nutritional Assessment in three versions: the original, population-specific (MNA-TI) and population-specific, without body mass index (MNA-TII). RESULTS: The original Mini-Nutritional Assessment rated 24% of patients malnourished and 57% at risk of malnutrition. Similar results, 14 and 64%, respectively, for MNA-TI; and 19 and 57%, respectively, for MNA-TII were observed. CONCLUSION: Both the original and the modified versions of the Mini-Nutritional Assessment can effectively rate the nutritional risk status of stroke rehabilitation patients in Taiwan. Version MNA-TII that adopted population-specific anthropometric cut-values but without body mass index can effectively predict the nutritional status of stroke patients. RELEVANCE TO CLINICAL PRACTICE: The modified scale (MNA-TII) can enhance the application of the tool and timely detection and intervention of undernutrition among stroke rehabilitation patients. It can also help to improve job efficiency of the primary care professionals.

Risk assessment of arsenic-induced internal cancer at long-term low dose exposure.

Previous epidemiological studies have indicated that ingested inorganic arsenic is strongly associated with a wide spectrum of internal cancers. Little is conducted, however, to assess health effects at long-term low dose exposures by linking biologically based mechanistic models and arsenic epidemiological data. We present an integrated approach by linking the Weibull dose-response function and a physiologically based pharmacokinetic (PBPK) model to estimate reference arsenic guideline. The proposed epidemiological data are based on an 8 years follow-up study of 10,138 residents in arseniasis-endemic areas in southwestern and northeastern Taiwan. The 0.01% and 1% excess lifetime cancer risk based point-of-departure analysis were adopted to quantify the internal cancer risks from arsenic in drinking water. Positive relationships between arsenic exposures and cumulative incidence ratios of bladder, lung, and urinary-related cancers were found using Weibull dose-response model r(2)=0.58-0.89). The result shows that the reference arsenic guideline is recommended to be 3.4 microg L(-1) based on male bladder cancer with an excess risk of 10(-4) for a 75-year lifetime exposure. The likelihood of reference arsenic guideline and excess lifetime cancer risk estimates range from 1.9-10.2 microg L(-1) and 2.84 x 10(-5) to 1.96 x 10(-4), respectively, based on the drinking water uptake rates of 1.08-6.52 L d(-1). This study implicates that the Weibull model-based arsenic epidemiological and the PBPK framework can provide a scientific basis to quantify internal cancer risks from arsenic in drinking water and to further recommend the reference drinking water arsenic guideline.