الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 79 |  |  |
| | | from | 79 | | |
Abstract
Hepatitis C virus (HCV) is a globally prevalent pathogen and a leading cause of death and morbidity.The most recent estimates of disease burden show an increase in seroprevalence over the last 15 years, equating over 185 million infections total worldwide.Treatment for HCV began with Food and Drug Administration approval of interferon (IFN) in 1991, followed by combined IFN and ribavirin (RBV), with pegylated IFN (peg-IFN) in 2001 and later with pegylated IFN and RBV. Recently, the regimen of SOF and RBV stood as the standard of care.
ARIMA model is one of the most widely used time-series analysis techniques because of its structured modeling basis and acceptable forecasting performance.
Our objectives were to identify ARIMA modelsprediction of total cases seeking treatment, cases eligible for treatment with interferon and early virologic response of hepatitis C patients, estimate the models parameters, asses the models accuracy, use the fitted ARIMA models for forecasts beyond the observed series, determine early response rate to treatment.
Data were analyzed using STATA version (11) and SPSS version (21). The Box-Jenkins methodology was adopted to fit the ARIMA (p,d,q). Time plots of the observed data were done to discover any patterns or irregularities to prepare series for modeling. Original data was first seasonally adjusted so that we have a de-seasonalized series and can thus proceed in forecasting using a non seasonal model. Dickey fuller test was used for testing stationarity. Once the data was made stationary we used ACF and PACF plots to identify the ARIMA model of the appropriate order. The models parameters were estimated and their statistical significancewas tested. An overall check of model adequacy was provided by the Ljung-Box Q statistic. Suitable models were selected and fitted to each different data based on AIC and BIC values where smaller AIC and BIC valuesindicate a better fitted model. MAPE to assess forecast accuracy was recordedwhere a lower MAPE indicate a better fit of the data.
The results of this study are summarized as follows:
1- A total of 105963 hepatitis C cases were registered in Damanhour National Medical Institute from 2009 to 2015. The annual number of registered cases increases from 11651 in year 2009 to 17710 in 2015.
2- The percentage of registered females was higher than males (54.7 % versus 45.3%).
3- Distribution of registered cases by age showed that the majority (67.7%) of cases were in age group 45 to 60 years old
4- The best fit model for registered of HCV patient’s is ARMA (1,1,1).
5- The mean absolute percentage error (MAPE) was used to assess forecast accuracy which was 4.0% for registered cases in 2015.
6- A total of 20306 hepatitis C cases are predicted during 12 months in 2016.
7- The annual number of eligible cases whoresponded to treatment increased from 83.6% in year 2009 to 85% in 2014.
8- The percentage of eligible females for treatment with interferon was higher than males (94.6 % versus 91.3%).
9- Distribution of eligible cases by age showed that the majority (100%) of cases were in age group 30 to 45 years old
10- The best fit model for eligible of HCV cases is ARMA (1,0,0).
11- The mean absolute percentage error (MAPE) was used to assess forecast accuracy which was 11.0% for eligible cases in 2014.
12- A total of 9640 hepatitis C eligible cases are predicted during 12 months in 2015.
13- The percentage of cases who responded to treatment increased from 51% in year 2009 to 67.7% in 2014.
14- The percentage of females who responded to treatment was higher than males (61.5 % versus 59.9%).
15- Distribution of cases according to their response to treatment by age showed that the majority (100%) of cases were in age group 30 to 45 years old.
16- Best fit model for HCV response to treatment is ARIMA (1,0,1) .
17- The mean absolute percentage error (MAPE) was used to assess forecast accuracy which was 10.8% for response to pegylated interferon treatment.
18- Thepredicted response rate to treatment was 63.3% in 2015.
19- A total of 27982 hepatitis C cases were registered in Damanhour National Medical Institute from January to December 2016. The number of responded cases was26890 which constitute 96% of the registered cases.
20- The percentage of registeredfemales was higher than males (58 % versus 42%).
ARIMA is a useful tool for the ministry of health, hospital directors and doctors to plan the necessary needs and develop future studies to treat all diseases, not just the hepatitis C virus.