Incidence data statistical analysis

This report is by Battelle Columbus Division to the Line Pipe Research Supervisory Committee of the American Gas Association. It presents an analysis of statistical data obtained from reports of lea)c or rupture (service) incidents and test failures in natural gas transmission and gathering lines over the 14.5 year period from 1970 through June, 1984. All gas transmission companies were required to notify the Office of Pipeline Safety Operations in the event of a "reportable" incident, as defined by the Code of Federal Regulations. The purpose of the study is to organize the reportable incident data into a meaningful format from which the safety record of the industry can be assessed. 

For other parameters, though, statistical analysis is just one of several considerations that include historical control data and other relevant information about the test agent and related test agents. For example, statistical analysis of a low incidence of... 

Although methods are available for including historical control data in the formal statistical analysis (Tarone, 1982 Dempster et al., 1983 Haseman, 1990), this is usually not done and for good reason. The heterogeneity of historical data requires that they be used qualitatively and selectively to aid in the final interpretation of the data, after completion of the formal statistical analysis. Table 9.3 presents a summary of background tumor incidences for the most commonly employed rodent strains. 

In 1985, the United States Federal Register recommended that the analysis of tumor incidence data be carried out with a Cochran-Armitage (Armitage, 1955 Cochran, 1954) trend test. The test statistic of the Cochran-Armitage test is defined as this term ... 

Although the statistics provided in Section 3.3 concerning the number and severity of reactive incidents are grave, existing sources of incident data are inadequate to identify the number, severity, frequency, and causes of reactive incidents. The following limitations affected CSB analysis of incident data ... 

The results of the CSB incident data analysis are acknowledged as representing only a sampling of recent reactive incident data. This limitation precludes CSB from drawing statistical conclusions on incidence rates or inferring trends in the number or severity of incidents. However, despite these limitations, the data can be used to illustrate the profile and causes of reactive incidents. 

Trend analysis can be confused or invalidated by a sample that is too small. If the charting or analysis is limited only to major incidents, there will often be too few within a period to arrive at meaningful conclusions. For example, a facility with one thousand employees may experience only one or two serious incidents per year, and several years worth of data would be needed to make any meaningful statistical analysis. Minor incidents and near misses can be as useful in trend analysis and preventive prediction as major incidents. All process incidents should be reported, classified, and investigated as appropriate. The severity of an incident is frequently more a function of chance than actual fundamental system differences among accidents and near misses. 

Increases in the incidence of right or reversed ductus arteriosus (9/18 or 50%) and ectopic testicles (3/18 or 33%) were observed in the offspring of unexposed male rats and female rats exposed to 1,742 mg orthophosphoric acid equivalents/m3, 15 minutes/day on gestational days 6-15. Statistical analysis of these data was not presented. No other developmental effects were observed in this study (Brown et al. 1981 Starke et al. 1982). 

Time-correlated single-photon counting (TCSP) has proven to be a much-used method for measuring fluorescence lifetimes. It is highly sensitive in that it requires only one photon to be incident on the detector per excitation cycle, and statistical analysis of the experimental data gives lifetimes with well-defined error limits. Commercial systems are available which allow lifetimes from 50 ps to many tens of nanoseconds to be measured with relative ease and high precision. 

Statistical analysis of familial data by Scott (SI 3) led to the conclusion that the type I and type II silent cholinesterases (referred to by Scott as type O and type T, respectively) are allelic to one another and, by extension, to the Ey gene (S16). As indicated above, and in Section 2.7, some populations have an unexpectedly high incidence of a silent cholinesterase gene (A17, G33, PI, S17). 

The trend identification can be qualitative and performed by graphing the stability data or could be performed by statistical analysis of the collated data. In both cases, the site OOT Standard Operating Procedure (SOP) defines criteria for designating a data deviation from the norm as an OOT incident. The OOT criteria must be set in such a way that all significant OOT incidences are identified, ideally without false positives. 

BLS began collecting additional information on the major injuries in the form of worker and incident characteristics. At that time, BLS also initiated a separate CFOI to review events more effectively than had been possible in the previous survey. The CFOl database can be used to do statistical analysis for fatalities by Standard Industrial Classification (SIC) Codes. The CFOl fatality data is presented in several different categories as shown below ... 

A statistical analysis of much of the published data on myocardial necrosis in male rats fed fats and oils for at least 16 weeks indicated that the heart lesions were negatively correlated to dietary saturated fatty acids (16 0 and 18 0) and linoleic acid (18 2), and positively correlated to dietary linolenic (18 3), oleic (18 1), eicosenoic (20 1), and docosenoic (22 1) acids (Table XXV). Most of the variation in lesion incidence between experiments could be explained by the concentration of 16 0 and 18 3. A plot of the observed incidence of heart lesion versus the lesion incidence predicted based on the concentration of 18 3 and 16 0 in the dietary oil of 23 experiments with over 2000 rats is shown in Fig. 8. Figure 8 shows a continuum of points representing a broad spectrum of fats, oils, and fat-oil mixtures. The more satu-... 

The problem of myocardial necrosis in rats has received the greatest attention in attempts to correlate myocardial necrosis to myocardial lipid changes. Myocardial necrosis has been observed in male rats irrespective of the dietary oil or fat, provided a sufficient number of heart sections (Vies et al., 1976 and 1978) or a sufficient number of rats (Hulan et a/., 1977d) are examined. Dietary fats high in saturated fatty acids (Hulan et a/., 1976 Kramer et a/., 1981 Farnworth et a/., 1982) and linoleic acid (Vies et a/., 1978) are associated with a low incidence of heart lesions, whereas diets which contain linolenic acid (McCutcheon eta/., 1976 Hulan eta/., 1977a Vies et a/., 1978) and erucic acid (Abdellatif and Vies, 1973) are associated with a high incidence of heart lesions. A statistical analysis of much of the heart lesion data clearly showed this correlation of dietary fatty acids to heart lesions (see Chapter 17, Table XXV). 

More serious is the evident fact that Dr. Finkel applies the same acceptance region indiscriminately to the three very different sets of experimental data represented by her curves A, B, and C. It seems likely that the test was designed to handle the life-shortening data (curve A), because a <-test would not be inappropriate to life-span data, since the death-rate function g(t) is (rather crudely) Gaussian. A statistical analysis of the cmTe-C data would be difficult, because the experimental statistic t (20 per cent incidence time) is cumbersome to handle mathematically, as is evident in our discussion. But it is easy to show that Dr. Finkel s acceptance region is entirely inapplicable to the curve-B data ( proportion of animals that survived the latent period of 150 days and then died with osteogenic sarcomas ). 

Some findings of this clinical trial are that daily administration of folic acid, vitamin Bg and vitamin B12 to patients with recent stroke or transient ischemic attack was safe but did not seem to be more effective than placebo in reducing the incidence of major vascular events. These results do not support the use of B vitamins to prevent recurrent stroke. The results of ongoing trials and an individual patient data meta-analysis will add statistical power and precision to present estimates of the effect of B vitamins (VITATOPS Trial Study Group 2010). 

What defines how safe we are Is it aceidents, ineidents, and negative events Lets talk about accidents. The commercial aviation system has the lowest accident rate of any transportation system in terms of miles travelled. That s the good news. However, because there are so few aecidents, they earmot be used for statistical analysis. There are more incidents than accidents, which make incidents a better somce of data, and there are more negative events than incidents. Close calls, near misses, and complaints can all be used as a data somce if they are reported and recorded. However, we still lack knowledge of what really occurs in normal operations. 

---