Treatment of Genital Warts Following HPV Vaccination Program
Treatment of Genital Warts Following HPV Vaccination Program
We conducted a time-series analysis using data from a national registry reporting numbers of patients treated under anaesthesia for genital and anal warts in all private hospitals in Australia.
Medicare is the universal health insurance scheme of Australia and rebates services provided by private doctors and laboratories. Services provided in the public sector are funded by the state and territory governments and are not rebated by Medicare. Each service has a unique item number and aggregated data are publicly available from the Medicare registry (the Medicare Benefit Schedule website).
We extracted data on in-patient treatments under general anaesthesia or regional or field nerve block (excluding pudendal block) requiring admission to a hospital, among 15–44 year olds from 2000 to 2011. The data were aggregated by 6-month time periods, sex, age-group (15–24 years, 25–34 years and 35–44 years), and anatomical site (vulval/vaginal warts - Medicare item numbers 35507, 35508; penile warts - Medicare item number 36815; anal warts - Medicare item numbers 32177, 32180).
In the vaccine period (2007–2011) and pre-vaccine period (2000–2007) we conducted a descriptive analysis of the number of in-patient treatments per year stratified by sex, age-group and anatomical site and calculated percentage change in treatment numbers. We also calculated annual treatment rates per 100,000 populations. Australian population data (derived from the Australian Bureau of Statistics) were used to calculate rates.
We used Box-Jenkins time-series methodology to determine average annual trends in rates of in-patient treatments. Residuals from the time-series models were examined by autocorrelation and partial autocorrelation methods to detect potential serial dependence in the data. When significant serial dependence was detected, Newey-West autocorrelation and heteroscedasticity-corrected standard errors were used in the relevant regression model assuming that the count data (i.e. number of diagnoses per year) followed a Poisson distribution. Estimated models were considered most appropriate if they typically simulated historical behaviour.
Finally, we compared the average annual treatment rates in the pre-vaccine and vaccine periods and describe summary rate ratios along with 95% confidence intervals.
Analyses were conducted and all models were fitted using STATA v12.1 (StataCorp, Texas, US).
Methods
Study Design
We conducted a time-series analysis using data from a national registry reporting numbers of patients treated under anaesthesia for genital and anal warts in all private hospitals in Australia.
Data Source
Medicare is the universal health insurance scheme of Australia and rebates services provided by private doctors and laboratories. Services provided in the public sector are funded by the state and territory governments and are not rebated by Medicare. Each service has a unique item number and aggregated data are publicly available from the Medicare registry (the Medicare Benefit Schedule website).
We extracted data on in-patient treatments under general anaesthesia or regional or field nerve block (excluding pudendal block) requiring admission to a hospital, among 15–44 year olds from 2000 to 2011. The data were aggregated by 6-month time periods, sex, age-group (15–24 years, 25–34 years and 35–44 years), and anatomical site (vulval/vaginal warts - Medicare item numbers 35507, 35508; penile warts - Medicare item number 36815; anal warts - Medicare item numbers 32177, 32180).
Statistical Analyses
In the vaccine period (2007–2011) and pre-vaccine period (2000–2007) we conducted a descriptive analysis of the number of in-patient treatments per year stratified by sex, age-group and anatomical site and calculated percentage change in treatment numbers. We also calculated annual treatment rates per 100,000 populations. Australian population data (derived from the Australian Bureau of Statistics) were used to calculate rates.
We used Box-Jenkins time-series methodology to determine average annual trends in rates of in-patient treatments. Residuals from the time-series models were examined by autocorrelation and partial autocorrelation methods to detect potential serial dependence in the data. When significant serial dependence was detected, Newey-West autocorrelation and heteroscedasticity-corrected standard errors were used in the relevant regression model assuming that the count data (i.e. number of diagnoses per year) followed a Poisson distribution. Estimated models were considered most appropriate if they typically simulated historical behaviour.
Finally, we compared the average annual treatment rates in the pre-vaccine and vaccine periods and describe summary rate ratios along with 95% confidence intervals.
Analyses were conducted and all models were fitted using STATA v12.1 (StataCorp, Texas, US).
Source...