Statistical analysis includes statistical description and inference. The former one describes central and dispersion tendency,sampling error and characteristics of data distribution.The key of statistical description is how to properly select the relevant statistic which primarily depends on three elements:characteristics of sample,type of data and its characteristics of distribution.Secondly,we should pay attention to correct matching and combination use of statistics. Statistical inference includes parameter estimation and hypothesis testing. We usually use interval estimation to do parameter estimation.The correct selection of the methods of hypothesis testing needs the following six factors,research purpose,number of research factors,number of groups or research level,types of experiment design,data types and distribution characteristics,and the sample size.

Objective Constructing a hospital informatization level evaluation model and carry out an empirical research.Methods Based on experts' suggestions and comprehensive scoring method,a evaluation model of hospital informatization level has been formed,and 1,221 Chinese hospitals have been selected randomly in this empirical research.Results The evaluation model and rating scale is reliable,valid and sensitive,and it can be used to evaluate hospital informatization level at larger scope.The difference of informtization level between hospitals in different areas,different type is evident and the development situation in hospitals is imblalace also.Conclusion The hospilal should emphasis on the informatization,and improve the application of information technology.

AIM: To establish the method of fingerprinting analysis onZedoary turmeric oil by GC. Zedoary turmeric oil for the quality control of reference. METHODS: GC method was used to determine the fingerprint and its similarity of Zedoary turmeric oil calculated on Excel and cluster analysis. RESULTS: Seventeen samples of Zedoary turmeric oil indicated that there was a little similarity among them. Based on the cluster analysis, all the Zedoary turmeric oils could be divided into two categories, self-made and market purchasing oil, the latter oils purchased from Liaoning province and Jiangsu province varied considerably. Every category had a common fingerprint mode with clear resemblance. CONCLUSION: When finding the difference between the self-made and the marketpurchasing Zedoary turmeric oil, quality control requires serious consideration.

AIM: To establish the method of fingerprinting analysis on Zedoary turmeric oil by GC.Zedoary turmeric oil for the quality control of reference. METHODS: GC method was used to determine the fingerprint and its similarity of Zedoary turmeric oil calculated on Excel and cluster analysis. RESULTS: Seventeen samples of Zedoary turmeric oil indicated that there was a little similarity among them.Based on the cluster analysis,all the Zedoary turmeric oils could be divided into two categories,self-made and market purchasing oil,the latter oils purchased from Liaoning province and Jiangsu province varied considerably.Every category had a common fingerprint mode with clear resemblance. CONCLUSION: When finding the difference between the self-made and the market purchasing Zedoary turmeric oil,quality control requires serious consideration.

The violence behavior is a kind of aggressive behaviors or attempt to hurt another person psychologically,physically or in other forms.Recently,violent incidents occur more and more frequently,and especially among teenagers.A number of concerns on violence continue to rise,and interpersonal violence is the most concerned type.Perpetrators are violence implementers who determine the occurrence and outcome of violence.Many studies provided the risk factors of interpersonal violence,that expounds the influence of personal level,interpersonal relationship,community background and social factors of interpersonal violence.The establishment of the socioecological risk-factor structural model which focuses on the perpetrators' indi vidual,is of great significance for the effective intervention for interpersonal violence.

Comprehensive evaluation methods are generally used to assess the population data. When we need to estimate the sample data in special situations, the impacts brought by the sampling error should be considered. Due to lack of the accurate measurement for the sensitivity and stability in the comprehensive evaluation methods, sampling errors usually cannot be estimated in the sampling research. Monte Carlo simulation was used in this article to solve the probability of the ordering results, and the matlab programs were presented. Based on the simulated results, we change the conventional “absolute conclusion” of comprehensive evaluation methods to “probability results” for the sample data, and put forward a new sorting and ranking method for the results of comprehensive evaluation.

Objective To evaluate the patient's health care demands in communication,diagnosis,treatment and medical counseling of geriatric BPH outpatients in China.Methods A survey study was carried out in 33 medical centers from 11 different cities in China (North:Lanzhou,Beijing,Shenyang; South:Chengdu,Changsha,Wuhan,Jinan,Nanjing,Hangzhou,Shanghai,Guangzhou).2027 BPH outpatients aged (60 years participated in the survey to measure their health care demands using a questionnaire.Results There were 955 cases (47.1％) thinking communication is necessary before inspection items by a doctor,807 (39.8 ％)thinking communication is necessary before determining the treatment program,389 cases(19.2％) thinking respection items by a doctor is necessary,730 cases(36.0％)the treatment program by a doctor is necessary,849 cases (41.9 ％)thinking health guidance on BPH dignosis is necessary,934 cases (46.1 ％)thinking health guidance on BPH treatment is necessary,The multilevel model showed the demands scores data is similarity among the north and the south.And the patients with more severe lower urinary tract symptoms have more demands on communication and medical counseling.Conclusions It is to get better therapy,considering the patient's health care demands.

Objective To investigate the current situation and influencing factors of overweight and obesity in the elderly from Changsha city.Methods Using stratified sampling,2500 elderly in Changsha were included for the investigation by questionnaire and physical examination.The influencing factors were analyzed with univariate analysis and two-level models.Results Among 2500 persons,the valid questionnaires were 2254,the effective rate was 90.16％.About 39.71％ (895/2254) of Changsha's elderly were in overweight and obesity (BMI≥ 24 kg/m2).Univariate analysis showed that there were significant differences in BMI among different area,education degree,income,marital status,drinking wine,drinking tea,physical exercises and occupation (x2＝ 304.17,18.39,19.35,12.93,13.29,11.77,12.87,12.84,all P＜0.05).The results of two-level models indicated that the independent influencing factors were income,physical exercises,drinking wine,age and occupation.Conclusions It is necessary to take interventional measures against elderly obesity and to encourage the elderly taking physical exercises and keeping reasonable diets.

Objective To synthetically evaluate medical treatment outcomes of two operative methods on benign prostatic hyperplasia (BPH),and to provide evidence for therapeutic regimen selection.Methods Prospective cohort study was applied.BPH patients were treated by two operative methods randomly,follow-up observations were carried out at 3 months,6 months,and 48 months after the treatment.IPSS and BPHQLS were adopted to evaluate prostate and quality of life of patients at different time.The weight coefficients of IPSS score,quality of life score and general hospitalization expense in the comprehensive evaluation were defined by the Delphi Method.Results The synthetic score of both groups showed ascending trend,comprehend evaluation of scores got after treatment for 3 months (OP group was 7.6±0.7,TURP group was 7.4±0.8),6 months(OP group was 6.9 ± 0.9,TURP group was 6.9 ± 1.1) showed no significant differences between two groups(P＞0.05),curative effect of OP group was better than TURP group at 48 moths(OP group was 8.3±0.9 score,TURP group was 7.6± 0.8 score) (P＜0.05).Conclusions The curative effect of OP group was better than TURP group at 48 months.

Objective To compare the efficacy of α1 adrenergic receptor blocker,5α-reductase inhibitor or combination therapy in benign prostatic hyperplasia (BPH),and to explore the application of multi-level statistical model in assessment of BPH treatments.Methods Pragmatic clinical trials (PCT) design was used.BPH patients received drug therapy including α-adrenergic receptor blocker,5α reductase inhibitor,and combination therapy were followed up for 4 years.Multilevel statistical model was used to compare the scores of international prostate symtom score(IPSS) and BPH qality of life scale(QLS) among the 3 therapeutic regimens.Results The intra class correlation coefficients of I-PSS and BPH-QLS were 0.6136 and 0.6946 respectively,which indicated that both data were hierarchical structured.During the follow-up period,scores of IPSS and BPH-QLS were improved along with the drug therapy in 3 regimens.There was curve relationship between treatment time and IPSS and BPH-QLS scores in 3 regimens (all P＜0.05),and there was no significant differences in the trend of curves between the 3 regimens (P＞0.05).Conclusions For BPH patients with moderate symptoms bothering the quality of life,α-adrenergic receptor blocker,5α-reductase inhibitor,or combination therapy are significantly effective in reducing symptoms and improving quality of life.There are no efficacy differences among the 3 drug therapy during the 4-years follow-up.Compared with traditional analysis,multilevel statistical model is a more suitable and precise method for assessment of BPH treatments.