George E. Davis

Predicting Length of Stay for Psychiatric Diagnosis-related Groups Using Neural Networks

Objective : To test the effect of diagnosis on training an artificial neural network (ANN) to predict length of stay (LOS) for psychiatric patients involuntarily admitted to a state hospital. Design : A series of ANNs were trained representing schizophrenia, affective disorders, and diagnosis-related group (DRG) 430. In addition to diagnosis, variables used in training included demographics, severity of illness, and others identified to be significant in predicting LOS. Results : Depending on diagnosis, ANN predictions compared with actual LOS indicated accuracy rates ranging from 35% to 70%. The validity of ANN predictions was determined by comparing LOS estimates with the treatment teams predictions at 72 hours following admission, with the ANN predicting as well as or better than did the treatment team in all cases. Conclusions : One problem in traditional approaches to predicting LOS is the inability of a derived predictive model to maintain accuracy in other independently derived samples. The ANN reported here was capable of maintaining the same predictive efficiency in an independently derived cross-validation sample. The results of ANNs in a cross-validation sample are discussed and the application of this tool in augmenting clinical decision is presented.

Peaks of solar cycles affect the gender ratio

In this study, we report that the gender ratio (GR) at death [where GR=(N(males)/N(males)+N(females))] of those born (and likely conceived) in solar cycle peaks (about a 3-year period occurring on average every approximately 11 years), is inversely related to mean male age at death; e.g., the higher the GR(at death) the lower the mean lifespan, while the GR(at death) of those born in non-peak years has no relation to mean male lifespan. Although changes in the GR are small and may be of little clinical significance, the GR is a sensitive indicator of environmental effects, and therefore is pertinent to epigenetics. This paper supports the hypothesis that solar radiation, probably in the ultraviolet spectrum, by some manner interacts with chromosomal DNA (genes) and produces the genetic variety that not only fosters adaptation, but also produces the diseases that reduce lifespan. This paper also proposes that sunlight is more effective in modifying genomes at the time of conception than later in gestation or infancy. Referring to the work of others, this study also reveals that geographic latitude also affects the GR, suggesting that the variation in light is probably as important as the intensity of light in modifying genomes. This study finds that men sustain more genetic variation, producing 28% more disease than women, as well as a 2% decrease in GR from birth to death, and a shorter life (in Maine) by 7 years.

Short and long term variation in ultraviolet radiation and multiple sclerosis

We examined the role of ultraviolet radiation (UVR) in persons diagnosed with multiple sclerosis (MS) in four different populations, Italians, Danish, White and African Americans. We tested whether variation in UVR as determined by seasons (short term variation) and solar cycles (long term variation) is related to MS birth month and to survival as measured by lifespan. Cases were selected from three Italian MS Case Registries (2,737); from the United States National Center for Health Statistics (56,020); and from the Danish Multiple Sclerosis registry (15,900). Chi-square tests were used to study the pattern of month of birth distribution in patients with MS comparing with general population data. T-tests were employed to study solar cycles association with lifespan. A surplus of births was observed in June for White Americans. A decrease of births in October and November, though not significant after multiple testing correction, was observed in the three populations. In White American with MS overall, males and females, we found that solar cycle is associated with lifespan. We found that season and solar cycles have some role in MS susceptibility and life duration. However, this is an exploratory analysis and further work is needed to discern the association.

Evidence that latitude is directly related to variation in suicide rates.

Objective: To use available suicide-rate data from 20 countries to see patterns and relations more clearly. Method: We obtained raw suicide rates from the Organization for Economic Cooperation and Development (OECD) database from 1960 through 1997 and calculated averages and standard deviations. Results: There is a positive linear relation between the variation in suicide rate and geographic latitude. Conclusions: The variation in light-dark cycles is superimposed upon human mood.

The effect of solar cycles on human lifespan in the 50 United States: Variation in light affects the human genome

This paper describes the Suns effect on the human genome as it relates to lifespan and expands our previous study from the State of Maine to the entire United States and the District of Columbia. In the current study we report that those persons conceived and likely born during the peaks (MAX approximately 3years) of approximately 11-year solar cycles lived an average 1.7years less than those conceived and likely born during non-peaks (MIN approximately 8years). Increased energy at solar MAX, albeit relatively a small 0.1% increase from MIN, apparently modifies the human genome/epigenome and engenders changes that predispose to various diseases, thereby shortening lifespan. It is likely that same energy increases beneficial variety in the genome which may enhance adaptability in a changing environment. This study also reports that living at higher elevations increases exposure to ultraviolet radiation (UVR) and increases the difference between MAX and MIN in the six states at the highest elevations of their population centroids by approximately 13%, further shortening average lifespan about 3 months. How solar energy affects the genome is still not clear. The mechanism could be quantum mechanical (direct effects at a distance) similar to photosynthesis, or mediated by maternal hormones, chemokines or cytokines. The hypothesis is that specific wavelengths of UVR, experienced at critical times in development as at conception or early gestation, and with specific intensity or rate of change, modulates the expression of human diseases. This hypothesis could be readily testable in mice bred to manifest specific diseases.

Indirect evidence that ultraviolet-B radiation mitigates multiple sclerosis in the United States

This article describes on the relationship of the relative prevalence of persons dying with multiple sclerosis with the latitude of the population centroid of those affected in each of the United States. Regression to zero prevalence occurs at the Tropic of Cancer, the latitude where the Sun is at zenith on the summer solstice and where ultraviolet radiation (UVR) is least attenuated. This observation supports UVR as a mitigating force in multiple sclerosis.

Variation in ultraviolet radiation and diabetes: evidence of an epigenetic effect that modulates diabetics’ lifespan

BackgroundPublished research has shown that month-of-birth variations modulate the incidence of adult human diseases. This article explores diabetes type 2 as one of those diseases. This study uses the death records of approximately 829,000 diabetics (approximately 90% were type-2) born before the year 1945 (and dying between 1979 and 2005) to show that variations in adult lifespan vary with ultraviolet radiation (UVR) at solar cycle peaks (MAX, approximately a three-year period) with less at non-peaks (MIN, approximately an eight-year period). The MAX minus MIN (in years) was our measure of sensitivity (for example, responsiveness) to long-term variations in UVR.ResultsDiabetics were less sensitive than non-diabetics, and ethnic minorities were more sensitive than whites. Diabetic males gained 6.1 years, and females 2.3 years over non-diabetics, with diabetic males gaining an average of 3.8 years over diabetic females. Most variation in lifespan occurred in those conceived around the seasonal equinoxes, suggesting that the human epigenome at conception is especially influenced by rapid variation in UVR. With rapidly decreasing UVR at conception, lifespan decreased in the better-nourished, white, female diabetic population.ConclusionsRapidly changing UVR at the equinoxes modulates the expression of an epigenome involving the conservation of energy, a mechanism especially canalized in women. Decreasing UVR at conception and early gestation stimulates energy conservation in persons we consider ‘diabetic’ in today’s environment of caloric surfeit. In the late 19th and early 20th centuries ethnic minorities had poorer nutrition, laborious work, and leaner bodies, and in that environment a calorie-conserving epigenome was a survival advantage. Ethnic minorities with a similar epigenome lived long enough to express diabetes as we define it today and exceeded the lifespan of their non-diabetic contemporaries, while that epigenome in diabetics in the nutritional environment of today is detrimental to lifespan.

A field trial using artificial neural networks to predict psychiatric in-patient length-of-stay

Demands for health care reform will increase service utilization, much of which will fall on a system of expanded primary care providers, many of whom will not be specialists in psychiatry. These providers will need tools to augment their decision-making process. In this paper, we explore the use of Artificial Neural Networks (ANNs) in three different field sites to predict inpatient psychiatric Length-Of-Stay (LOS). This study describes the development and implementation of a runtime system in three different psychiatric facilities. Data was collected at these respective sites using the runtime system, and then this data was used to retrain the networks to determine if site-specific data would improve accuracy of prediction of LOS. The results indicate that ANNs trained with state hospital data could accurately predict LOS in two different community hospital psychiatric units. When the respective ANNs were retrained with approximately 10% new data from these specific hospitals, rates of improvement ranged from 3% to 15%. Our findings demonstrate that an ANN can adapt to different treatment settings and, when retrained, significantly improve prediction of LOS. Prediction rates by the ANN after retraining are comparable to results of a clinical team.

Solar energy at birth and human lifespan

PURPOSE The purpose of this paper is to examine the role of UVR at birth and its relationship to lifespan and determine whether there are significant differential effects on sex and race. We test if variation in UVR, as determined by solar cycles (long-term variation), is related to survival as measured by age at death. METHODS The data used 78 million death records from the National Center for Health Statistics (NCHS) from 1979 to 2013 with accidents, suicides, and war casualties deleted resulted in ~63 million records. Records of persons ≤ 47 years old were also scrubbed because we could not show an effect on lifespan based upon the intensity of solar energy as reflected by sunspot number (SSN). This we hypothesize is due to the protective effect of the hormones associated with growth and reproduction. Also selected were persons afflicted with multiple sclerosis (MS). RESULTS Males of all races born with a UVR intensity as estimated by sunspot number (SSN) ≤ 90 had an average lifespan of 74.4 years, for females of all races, 78.1 years; males born with >90 had an average lifespan of 66.3 years, for females of all races, 70.2 years, resulting in a lifespan decrease of 8.1 years for males and 8.5 years for females. For African-American males born ≤ 90 SSN, 70.8 years and for >90 SSN, 62.5 years, an 8.3-year decrease; similarly, for African-American females ≤ 90 SSN, 75.0, for >90 SSN, 65.4 years, a 9.6-year decrease. Higher solar energy at birth had an adverse effect on human lifespan. We also found that there were twice as many persons with MS born in >80-90 SSN as in the general population. CONCLUSIONS There is a statistically significant inverse relationship between exposure to solar energy at birth and average human lifespan. Solar energy by some mechanism alters the epigenome at birth, but the effect of higher solar energy becomes apparent after the age of natural selection.