Abstract
We examined the relative efficiency of non-clinical sites to screen for chlamydia in youth and young adults. Chlamydia screening targeting youth (under 30 years of age) was performed at non-clinical sites in high-morbidity neighbourhoods of two California counties. Venues were subdivided into community outreach, schools, parenting centres and drug treatment/correctional facilities. Relative efficiency was estimated with multivariate Poisson regression using incidence of chlamydia per person-hours labour adjusting for strategy and county. Among the 1514 youths screened, the overall prevalence of chlamydia was 5.5%. By venue, the highest prevalence was in drug treatment/correction facilities at 11.1% followed by parenting centres at 6.3%, community outreach at 4.9% and schools at 4.6%. Drug treatment/correctional facilities were the most labour time efficient at 9.9 person-hours per chlamydia case. Schools and parenting centres had the lowest cost per screen at 0.9 person-hours per screen. Adjusted relative labour time efficiency (chlamydia cases per paid person-hour) was significantly higher in schools, 2.0 (95% confidence interval [CI] 1.0–4.2), parenting centres, 3.2 (95% CI 1.6–6.6) and drug treatment/correctional facilities, 2.9 (95% CI 1.0–7.8), compared with community outreach. In conclusion, parenting centres and drug treatment centres and correctional facilities are the most efficient venues for chlamydia screening.
INTRODUCTION
Chlamydia trachomatis infection is common among young men and women in the United States. In the National Longitudinal Study of Adolescent Health for 2001–2002, the prevalence in women and men (ages 18–26) was 4.7% and 3.7%, respectively, with most of those infected having no symptoms. 1 The National Health and Nutrition Examination Survey for 1999–2002 found a prevalence of 4.6% in women and 2.3% in men aged 14–19 years, and 2.5% in women and 3.2% in men aged 20–29 years. 2 Left untreated, chlamydia is associated with pelvic inflammatory disease in women resulting in significant health-care costs related to chronic pelvic pain, tubal pregnancy and infertility. 3
Since most chlamydial infections are asymptomatic, identifying prevalent genital chlamydial infection requires screening sexually active adolescents and young adults. Routine screening of women under the age of 26 is recommended by the Centers for Disease Control and Prevention and the United States Preventive Services Task Force. 4,5 This recommendation is based on reasonable evidence that screening prevents the reproductive health complications of chlamydia infections in women, although it is debated whether there has been an adequate cost-effectiveness evaluation of chlamydia screening programmes. 6
Lack of access to health care has been associated with high chlamydia positivity. 7 Even with health care it has been seen that only 38% of sexually active young women screened who attended health plans. 8 In another study of 5.2 million patient encounters, there was no chlamydia screening performed in 84% of United States outpatient preventive visits by non-pregnant women. 9 Recognizing that many youths are not utilizing clinical services, Sexually Transmitted Disease programmes have endeavoured to identify innovative strategies for reaching this population by chlamydia screening outside of clinical settings using nucleic acid amplification methodologies that allow self-collected urine and vaginal specimens. 10 Because the efficiency of chlamydia screening at different non-clinical sites is not established, we used labour cost to estimate relative cost efficiency of screening at non-clinical venues in two California counties.
METHODS
Design
This was an observational study, conducted from July through October 2002 and June through September 2003, to screen for chlamydia among youth and young adults in geographically targeted communities at non-clinical screening sites. In collaboration with the California Department of Public Health Sexually Transmitted Disease (STD) Control Branch, two California county health departments designed interventions for targeted chlamydia screening in high-morbidity neighbourhoods. The target population included sexually active adolescents and young adults under the age of 30 who lived, worked or participated in local activities in the targeted neighbourhoods of the two California health jurisdictions. End-points for this analysis were labour costs per chlamydia test performed and per chlamydia case identified. Study protocols were reviewed and approved by the Committee on Human Research at the University of California, San Francisco.
Sampling and recruitment
Potential venues for screening were selected based on the geographic location of high chlamydia incidence within the respective counties. A combination of factors was considered in venue selection including feasibility of establishing a collection centre, acceptability of screening within the community and existing partnerships with community agencies. Methods for recruitment varied from active to passive. Active recruitment involved direct contact with the target population to provide education and offer screening. Passive recruitment involved establishing a screening site, distributing flyers and posting signs in the community, and relying on members of the target population to seek screening. The use of volunteers also varied by venue. Participants were offered a variety of incentives, including snacks, soda and movie coupons.
Data collection
Standard project materials in English and Spanish, including consent forms, behavioural questionnaires, specimen collection instructions, fact sheets, laboratory slips and venue log sheets, were developed for all sites. The survey included demographic information, utilization of care and STD risk factors. The project coordinators for the counties collected information for each venue including the type of venue, hours of paid staff time in preparation and implementation, collaborative and outreach strategies, and hours of volunteer staff time.
Urine and self-collected vaginal specimens were collected at each site and transported to local public health laboratories for nucleic acid amplification testing for chlamydia and gonorrhoea. Vaginal swabs were mailed to the University of California San Francisco Chlamydia Laboratory either by health department staff or participants themselves. All testing were performed according to the manufacturers' specifications and test results were reported to the local health jurisdiction STD programmes. Subjects with positive test results were contacted and provided standard treatment and partner management.
Data analysis
We describe the chlamydia rates found at the test sites and compare the cost-effectiveness of the venue types based on labour person-time and the characteristics of the venues. Only those subjects younger than 30 years of age with a specified gender were included. Chlamydia rates were calculated for each venue by gender, age, race/ethnicity, county of testing and access to medical care in the prior year. The chi-square statistic was used to compare the within-group differences.
Sites for chlamydia screening were categorized into four venue types: (1) community outreach, which included neighbourhood settings, mobile home parks, street outreach, community and faith-based agencies, public parks, malls and markets, and health fairs; (2) schools, which included public high schools and alternative schools; (3) parenting centres that served low-income families; and (4) drug treatment/correctional facilities, which included court-mandated and community-based substance abuse treatment centres, juvenile detention and jails.
Screening site characteristics included in the analyses were recruitment methods and the use of volunteers. The methods of recruiting were dichotomized as active or passive based on whether staff actively conducted outreach to recruit potential screening subjects versus relying on individuals to seek testing at an established screening site. For each site, a dichotomous variable was created to reflect whether volunteers were used.
To calculate the labour time per screen and per case identified, we divided the number of hours of total labour, including preparation and operational time for each site, by the number of screens performed and the number of chlamydia cases identified, respectively. Labour cost in person-hours was calculated for each venue type and for venue characteristics.
To compare the relative efficiency of each venue type adjusted for the characteristics of the venues and the county, we used general estimating equation multivariate regression to analyse chlamydia cases as count data at the site level. The use of multivariate general linear models for non-normal distributions for cost data has been described elsewhere. 11,12 Chlamydia counts per site were treated as an over-dispersed Poisson variable with a logarithmic link to the offset of total labour time in hours adjusted for over-dispersion by the Pearson statistic. Two models were used, one with total labour as all person-hours, and one with only paid person-hours that would exclude volunteer time. This derived a parameter estimate for the relative rate of chlamydia cases per labour cost for each covariate used. Statistical analysis was performed with Statistical Analysis System (SAS) version 9.1 (SAS Institute, Cary, NC, USA).
RESULTS
A total of 1632 individuals were screened at 63 different venues in the two counties. A total of 118 subjects were not included in the analysis because they were 30 years of age or older, gender was unspecified or they were tested at a venue for which labour information was not collected. In the remaining 59 screening venues, the number of screening tests performed at each site ranged from 1 to 148 with a median of 11. Of the 1514 subjects included in the analysis, the median age was 17 (range 12–29); 55% were women. The prevalence of chlamydia among those screened was 5.5% and was significantly higher in women compared with men (8.2% versus 3.0%, P < 0.0001), and among African Americans compared with other race/ethnic groups (10.2% versus 4.8%, P = 0.005).
A comparison of the venue types by client characteristics and chlamydia prevalence is shown in Table 1. Community outreach venues screened a larger proportion of men (63% of tests), African Americans (17% of tests) and those reporting no medical care in the previous year (17% of tests). Schools and parenting centres screened a larger proportion of women, 59% and 81% of tests, respectively. By venue, the highest chlamydia prevalence was in drug treatment/correction facilities at 11.1% with the highest rate in women (n = 57, 15.8%). Among men, the highest prevalence was among those screened at parenting centres (n = 64, 9.4%). The prevalence varied by age and race/ethnicity between venues.
Characteristics of the populations screened and prevalence of chlamydia by non-clinical venue type in two California counties
*Comparison of chlamydia prevalence among venues
Overall the median labour time per screen was 1.7 person-hours (range 0.3–42). From the 29 sites that identified at least one chlamydia case, the median labour time per chlamydia case was 17.2 person-hours (range 1.0–113.1). For venue type, labour cost in person-time per screen performed ranged from 0.9 person-hours per screen in schools and parenting centres to 2.3 person-hours per screen in community outreach (Table 2). Labour time per chlamydia case ranged from 9.9 person-hours per screen in drug treatment/correctional facilities to 47.0 person-hours per screen in community outreach.
Labour cost for chlamydia screening by venue characteristics, n = 59 venues
Venues were characterized by recruitment strategy depending on whether or not the intervention involved actively or passively soliciting youth for screening, and whether or not volunteers were used. County 1 was more likely than County 2 to have sites using active recruitment (P = 0.01) and volunteers (P < 0.0001). Both these strategies appeared to be more effective in finding chlamydia cases, even though County 1 had a lower chlamydia prevalence. Volunteers were used in 31 sites equally spread among recruitment strategies. In sites using volunteers, the volunteers accounted for a median of 50% of all site hours (range 4–77%). Volunteer hours did not correlate with total paid person-hours (Pearson r = 0.15).
Adjusting for the association between different characteristics of the venues in multivariate analysis, schools, parenting centres and drug treatment/correctional facilities was significantly more efficient in finding cases per total person-hours of labour compared with community outreach whether volunteer hours were included or if only paid hours were considered (Table 3). Having volunteers increased the effectiveness of finding chlamydia cases per paid labour by 3.2 (95% confidence interval [CI] 1.3–7.8) but if volunteer hours were included in the cost for total labour, using volunteers as a categorical variable did not reach full statistical significance. After adjusting for other factors, active recruitment strategies had no impact.
Relative labour cost-effectiveness of non-clinical site screening strategies for detecting chlamydia cases in two California counties by total and paid labour, n = 59 venues
CI = confidence interval
*Adjusted relative ratio with all covariates in the Poisson regression model
DISCUSSION
This study of chlamydia screening in youth and young adults in non-clinical settings provided useful information to those involved in programme decisions related to chlamydia detection and control. Chlamydia is the most common reportable communicable disease among youth and young adults, but the resource-challenged public health sector needs to be able to target chlamydia screening for maximal yield. Screening asymptomatic youth and young adults at family planning clinics has been suggested to be cost-effective. 13,14 We analysed relative costs at different non-clinical sites using hours of labour time where labour costs comprise the majority of most programme budgets.
Our study demonstrated that non-clinical venues can be used to screen for and identify chlamydia cases. Other publications support our finding that screening at locations frequented by high-risk youth and young adults can yield relatively high chlamydia rates. 15–18 We found that screening at drug treatment/correctional facilities found a high prevalence and was very efficient in personnel time required to detect chlamydia. Schools have been successfully targeted in California and elsewhere to screen large numbers of youths for chlamydia and gonorrhoea. 10,19 Schools were among the most cost-effective screening sites, but the yield of cases per person-hours of labour was less than those among parenting centres and drug treatment/correctional facilities. In our study, parenting centres had the highest prevalence of chlamydia in men with the overall efficiency of parenting centres similar to that for drug treatment/correctional facilities in chlamydia detection. An advantage shared by drug treatment/correctional facilities, schools and parenting centres was the existing infrastructure to incorporate chlamydia screening. If these findings are replicated in other programme evaluations, parenting centres may be considered appropriate venues for extending screening into non-clinical settings.
Other community settings have been shown to effectively identify chlamydia cases. 10,15,20 We grouped community outreach sites into a single group that included diverse locations in neighbourhoods, public parks and community agencies. These sites required significant effort to implement screening evident through the high average labour cost per screen and per chlamydia case. Schools, parenting centres and drug treatment/correctional facilities were all significantly more cost-effective for detecting chlamydia cases than were community outreach settings. Community outreach did appear to more effectively reach groups at particularly high risk, such as those with no regular medical care, Latinos and African Americans, although this may be due to a selection bias of venues chosen.
Using volunteer labour appeared to enhance the ability to increase the number of people screened among target populations and therefore increase detection of chlamydia cases. This appeared to work in two ways: an extension of paid labour at no additional cost and possibly increased recruitment. Many of the volunteers were youth from target neighbourhoods who may recruit peers more effectively than staff. The impact of peer volunteer programmes on increasing the capacity of youth volunteers and improving relationships between health departments and community-based organizations deserves further evaluation.
As an observational study with no randomization of strategies, our ability to make definitive statements on the outcomes is limited. In practice, staff did not strictly adhere to the inclusion criteria and some older individuals were screened. This accounted for only a small number of sites and tests and those individuals were excluded from the analysis. We did not collect data on the non-personnel costs (e.g. recruitment and outreach materials, printing, transportation, laboratory costs and incentives for participants) that may have contributed differentially to screening costs for different sites. Time and cost data were recorded by project co-ordinators, thus may have been over- or under-estimated. In addition, initial costs of establishing non-clinical screening sites and data collection procedures contributed to the high observed costs and it is likely that established screening projects increase in efficiency over time.
This study provides support for non-clinical settings as important chlamydia screening sites for youth and young adults, particularly correctional institutions, drug treatment centres and parenting centres. Schools reach many youths but may not find as many chlamydia cases. Other community outreach sites need to be selected based on local epidemiological data, but in general are more time consuming for the yield of chlamydia cases. STD prevention efforts should continue to include implementing and evaluating innovative methods for reaching unscreened at-risk youths.
Footnotes
ACKNOWLEDGEMENTS
The authors would like to thank the project coordinators at Stanislaus County Health Services Agency, Sandra Rose and Mark Loeser, and the project coordinators at the San Bernardino County Department of Public Health STD Control Program, Crystal Jones-Ramos and Heather Cockerill. In addition, we thank Romni Neiman and Monique Brammeier at the California Department of Public Health STD Control Branch, and Jeanne Moncada and Dr Julius Schachter at the University of California San Francisco Chlamydia Laboratory.