A Bayesian functional data model for surveys collected under informative sampling with application to mortality estimation using NHANES.

Functional data are often extremely high-dimensional and exhibit strong dependence structures but can often prove valuable for both prediction and inference. The literature on functional data analysis is well developed; however, there has been very little work involving functional data in complex survey settings. Motivated by physical activity monitor data from the National Health and Nutrition Examination Survey (NHANES), we develop a Bayesian model for functional covariates that can properly account for the survey design. Our approach is intended for non-Gaussian data and can be applied in multivariate settings. In addition, we make use of a variety of Bayesian modeling techniques to ensure that the model is fit in a computationally efficient manner. We illustrate the value of our approach through two simulation studies as well as an example of mortality estimation using NHANES data.

Predicting Profile Soil Properties with Reflectance Spectra via Bayesian Covariate-Assisted External Parameter Orthogonalization.

In situ, diffuse reflectance spectroscopy (DRS) profile soil sensors have the potential to provide both rapid and high-resolution prediction of multiple soil properties for precision agriculture, soil health assessment, and other applications related to environmental protection and agronomic sustainability. However, the effects of soil moisture, other environmental factors, and artefacts of the in-field spectral data collection process often hamper the utility of in situ DRS data. Various processing and modeling techniques have been developed to overcome these challenges, including external parameter orthogonalization (EPO) transformation of the spectra. In addition, Bayesian modeling approaches may improve prediction over traditional partial least squares (PLS) regression. The objectives of this study were to predict soil organic carbon (SOC), total nitrogen (TN), and texture fractions using a large, regional dataset of in situ profile DRS spectra and compare the performance of (1) traditional PLS analysis, (2) PLS on EPO-transformed spectra (PLS-EPO), (3) PLS-EPO with the Bayesian Lasso (PLS-EPO-BL), and (4) covariate-assisted PLS-EPO-BL models. In this study, soil cores and in situ profile DRS spectrometer scans were obtained to ~1 m depth from 22 fields across Missouri and Indiana, USA. In the laboratory, soil cores were split by horizon, air-dried, and sieved (<2 mm) for a total of 708 samples. Soil properties were measured and DRS spectra were collected on these air-dried soil samples. The data were randomly split into training (n = 308), testing (n = 200), and EPO calibration (n = 200) sets, and soil textural class was used as the categorical covariate in the Bayesian models. Model performance was evaluated using the root mean square error of prediction (RMSEP). For the prediction of soil properties using a model trained on dry spectra and tested on field moist spectra, the PLS-EPO transformation dramatically improved model performance relative to PLS alone, reducing RMSEP by 66% and 53% for SOC and TN, respectively, and by 76%, 91%, and 87% for clay, silt, and sand, respectively. The addition of the Bayesian Lasso further reduced RMSEP by 4⁻11% across soil properties, and the categorical covariate reduced RMSEP by another 2⁻9%. Overall, this study illustrates the strength of the combination of EPO spectral transformation paired with Bayesian modeling techniques to overcome environmental factors and in-field data collection artefacts when using in situ DRS data, and highlights the potential for in-field DRS spectroscopy as a tool for rapid, high-resolution prediction of soil properties.

Photoacoustic discrimination of vascular and pigmented lesions using classical and Bayesian methods.

Discrimination of pigmented and vascular lesions in skin can be difficult due to factors such as size, subungual location, and the nature of lesions containing both melanin and vascularity. Misdiagnosis may lead to precancerous or cancerous lesions not receiving proper medical care. To aid in the rapid and accurate diagnosis of such pathologies, we develop a photoacoustic system to determine the nature of skin lesions in vivo. By irradiating skin with two laser wavelengths, 422 and 530 nm, we induce photoacoustic responses, and the relative response at these two wavelengths indicates whether the lesion is pigmented or vascular. This response is due to the distinct absorption spectrum of melanin and hemoglobin. In particular, pigmented lesions have ratios of photoacoustic amplitudes of approximately 1.4 to 1 at the two wavelengths, while vascular lesions have ratios of about 4.0 to 1. Furthermore, we consider two statistical methods for conducting classification of lesions: standard multivariate analysis classification techniques and a Bayesian-model-based approach. We study 15 human subjects with eight vascular and seven pigmented lesions. Using the classical method, we achieve a perfect classification rate, while the Bayesian approach has an error rate of 20%.

Modeling complex phenotypes: generalized linear models using spectrogram predictors of animal communication signals.

A major goal of evolutionary biology is to understand the dynamics of natural selection within populations. The strength and direction of selection can be described by regressing relative fitness measurements on organismal traits of ecological significance. However, many important evolutionary characteristics of organisms are complex, and have correspondingly complex relationships to fitness. Secondary sexual characteristics such as mating displays are prime examples of complex traits with important consequences for reproductive success. Typically, researchers atomize sexual traits such as mating signals into a set of measurements including pitch and duration, in order to include them in a statistical analysis. However, these researcher-defined measurements are unlikely to capture all of the relevant phenotypic variation, especially when the sources of selection are incompletely known. In order to accommodate this complexity we propose a Bayesian dimension-reduced spectrogram generalized linear model that directly incorporates representations of the entire phenotype (one-dimensional acoustic signal) into the model as a predictor while accounting for multiple sources of uncertainty. The first stage of dimension reduction is achieved by treating the spectrogram as an "image" and finding its corresponding empirical orthogonal functions. Subsequently, further dimension reduction is accomplished through model selection using stochastic search variable selection. Thus, the model we develop characterizes key aspects of the acoustic signal that influence sexual selection while alleviating the need to extract higher-level signal traits a priori. This facet of our approach is fundamental and has the potential to provide additional biological insight, as is illustrated in our analysis.

Automated wavelet denoising of photoacoustic signals for circulating melanoma cell detection and burn image reconstruction.

Photoacoustic image reconstruction may involve hundreds of point measurements, each of which contributes unique information about the subsurface absorbing structures under study. For backprojection imaging, two or more point measurements of photoacoustic waves induced by irradiating a biological sample with laser light are used to produce an image of the acoustic source. Each of these measurements must undergo some signal processing, such as denoising or system deconvolution. In order to process the numerous signals, we have developed an automated wavelet algorithm for denoising signals. We appeal to the discrete wavelet transform for denoising photoacoustic signals generated in a dilute melanoma cell suspension and in thermally coagulated blood. We used 5, 9, 45 and 270 melanoma cells in the laser beam path as test concentrations. For the burn phantom, we used coagulated blood in 1.6 mm silicon tube submerged in Intralipid. Although these two targets were chosen as typical applications for photoacoustic detection and imaging, they are of independent interest. The denoising employs level-independent universal thresholding. In order to accommodate nonradix-2 signals, we considered a maximal overlap discrete wavelet transform (MODWT). For the lower melanoma cell concentrations, as the signal-to-noise ratio approached 1, denoising allowed better peak finding. For coagulated blood, the signals were denoised to yield a clean photoacoustic resulting in an improvement of 22% in the reconstructed image. The entire signal processing technique was automated so that minimal user intervention was needed to reconstruct the images. Such an algorithm may be used for image reconstruction and signal extraction for applications such as burn depth imaging, depth profiling of vascular lesions in skin and the detection of single cancer cells in blood samples.

Photoacoustic discrimination of viable and thermally coagulated blood using a two-wavelength method for burn injury monitoring.

Discriminating viable from thermally coagulated blood in a burn wound can be used to profile burn depth, thus aiding the removal of necrotic tissue. In this study, we used a two-wavelength photoacoustic imaging method to discriminate coagulated and non-coagulated blood in a dermal burn phantom. Differences in the optical absorption spectra of coagulated and non-coagulated blood produce different values of the ratio of peak photoacoustic amplitude at 543 and 633 nm. The absorption values obtained from spectroscopic measurements indicate that the ratio of photoacoustic pressure for 543 and 633 nm for non-coagulated blood was 15.7:1 and 1.6:1 for coagulated blood. Using planar blood layers, we found the photoacoustic ratios to be 13.5:1 and 1.6:1, respectively. Using the differences in the ratios of coagulated and non-coagulated blood, we propose a scheme using statistical classification analysis to identify the different blood samples. Based upon these distinctly different ratios, we identified the planar blood samples with an error rate of 0%. Using a burn phantom with cylindrical vessels containing coagulated and non-coagulated blood, we achieved an error rate of 11.4%. These results have shown that photoacoustic imaging could prove to be a valuable tool in the diagnosis of burns.

Assisted living in rural America: results from a national survey.

CONTEXT: Expanding the availability of long-term care (LTC) services and making them more responsive to consumer preferences is an important goal, particularly for elderly people living in rural areas who tend to be older and have greater functional limitations but less access to the range of LTC options available in metropolitan areas. One option that has been growing in popularity is assisted-living facilities (ALFs). PURPOSE AND METHODS: This paper describes rural ALFs and compares them with metropolitan ALFs. Data were collected using a multistage sample design that yielded a nationally representative sample of ALFs. Telephone interviews were completed with administrators of 1,251 ALFs in 1998. FINDINGS: Nationwide, assisted living was largely administered by private payment, and there was an undersupply in rural areas. Compared with metropolitan ALFs, rural ALFs were smaller and less likely to offer the types of services and accommodations associated with the philosophy of assisted living. They were more likely to offer accommodations with little privacy, and while similar in the services they offered, rural ALFs were less likely to have nurses on staff, particularly licensed practical nurses. Moreover, they were less likely to offer a combination of high services and high privacy. Finally, rural ALFs charged lower prices than urban ALFs; however, the average price was still unaffordable for most elderly rural residents. CONCLUSIONS: These findings suggest that assisted living, as currently structured, will make only a marginal contribution to meeting the needs of frail elders in rural areas.

Medicare expenditures for residents in assisted living: data from a national study.

OBJECTIVE: To provide preliminary data on Medicare expenditures for assisted living facility (ALF) residents and to investigate whether ALF characteristics were related to Medicare expenditures for ALF residents. DATA SOURCES/STUDY SETTING: Data from the National Study of Assisted Living for the Frail Elderly conducted in 1998-1999. This analysis was restricted to the 40 percent of ALFs in that sample that adhered to the assisted living (AL) philosophy by offering more than minimal levels of services and privacy. This study involved the approximately 1,200 residents who remained in an ALF from baseline to follow-up data collection. Six months of postbaseline Medicare claims were acquired for 545 of these residents, who did not differ significantly from the larger sample. DATA COLLECTION: Baseline individual and facility data were collected in personal interviews with residents and a combination of personal and telephone interviews with facility staff. Medicare claims data were acquired from the Centers for Medicare and Medicaid Services. STUDY DESIGN: Cross-sectional analyses using logistic and ordinary least squares regression techniques were used to determine the relationships among individual and facility characteristics and Medicare utilization and expenditures. PRINCIPAL FINDINGS: On an annualized basis, AL residents incurred Medicare costs of approximately US dollars 4,800. Just less than 15 percent of AL residents accounted for over 75 percent of total Medicare costs. Both the likelihood of utilizing Medicare-covered services and the intensity of service use were largely unaffected by the characteristics of the ALF in which residents lived. Utilization was largely a function of individual characteristics. The only exception to this general finding was that those individuals who utilized services and resided in smaller ALFs had significantly lower average expenditures than did individuals in larger ALFs. CONCLUSIONS: These preliminary data imply that both the level and distribution of Medicare expenditures among ALF residents were similar to those among the general community-dwelling Medicare beneficiary population. No significant relationships were observed between ALF characteristics and Medicare expenditures, except the effect of facility size. This result may imply that how the AL industry eventually defines itself in terms of services and amenities, other than size, may have little impact on Medicare expenditures for ALF residents. However, this is a single, initial study, so caution must be exercised when considering the implications of these results.

Rurality and nursing home quality: results from a national sample of nursing home admissions.

OBJECTIVES: We examined differences in quality of care among nursing homes in locales of varying degrees of rurality. METHODS: We classified locales into 4 classes according to rurality. We analyzed a 10% sample of nursing home admissions in the United States in 2000 (n=198613) to estimate survival models for 9 quality indicators. RESULTS: For postacute admissions, we observed significant differences in rates of decline for residents in facilities in large towns compared with urban areas, but differences in quality were both negative and positive. Among admissions for long-term or chronic care, rates of decline in 2 of 9 quality areas were lower for residents in isolated areas. CONCLUSIONS: We observed significant differences in a number of quality indicators among different classes of nursing home locations, but differences varied dramatically according to type of admission. These differences did not exhibit the monotonicity that we would have expected had they derived solely from rurality. Also, quality indicators exhibited more similarities than differences across the 4 classes of locales. The results underscore the importance, in some instances, of emphasizing the effects of specific settings rather than some continuum of rurality and of moving beyond the assumption that nursing home residents constitute a homogeneous population.

A national survey of assisted living facilities.

PURPOSE: Throughout the 1990s, assisted living was the most rapidly growing form of senior housing. The purpose of this paper is to describe the existing supply of assisted living facilities (ALFs) and examine the extent to which they matched the philosophy of assisted living. DESIGN AND METHODS: The study involved a multistage sample design to produce nationally representative estimates for the ALF industry. Administrators of nearly 1,500 eligible ALFs were interviewed by telephone. RESULTS: As of 1998, there were an estimated 11,459 ALFs nationwide, with 611,300 beds and 521,500 residents. Nearly 60% offered a combination of low services and low or minimal privacy, whereas only 11% offered relatively high services and high privacy. Seventy-three percent of the resident rooms or apartments were private. Aging-in-place was limited by discharge policies in most ALFs for residents who needed help with transfers, had moderate to severe cognitive impairment, had any behavioral symptoms, or needed nursing care. The industry is largely private pay and unaffordable for low- or moderate-income persons aged >/=75 unless they use assets as well as income to pay. IMPLICATIONS: ALFs differed widely in ownership, size, policies, and the degree to which they manifested the philosophy of assisted living. This diversity represents a challenge for consumers in terms of selecting an appropriate facility and for policy makers in terms of deciding what role they want assisted living to play in long-term care.