Mental Health Disorders and Sexually Transmitted Diseases in a Privately Insured Population

December 1, 2004
David B. Rein, PhD
David B. Rein, PhD

,
Lynda A. Anderson, PhD
Lynda A. Anderson, PhD

,
Kathleen L. Irwin, MD, MPH
Kathleen L. Irwin, MD, MPH

Volume 10, Issue 12

Objectives: To consider whether patients who use mentalhealth services in privately insured settings are also more likely tohave received sexually transmitted disease (STD) or humanimmunodeficiency virus (HIV) diagnoses and whether this relationshipextends to patients with milder mental health disorders.

Methods: Using frequency tables stratified by age and sex, alogistic regression model, and difference of means tests, we examinedthe relationship between mental health claims and STDs in asample of 289 604 privately insured people across the United States.

Results: Patients with mental health claims were more thantwice as likely as other patients to have an STD claim in the sameyear after controlling for confounding factors (odds ratio, 2.33;95% confidence interval, 2.11-2.58). This relationship held forsevere and milder mental health diagnoses, for male and femalepatients, and in each age category from 15 to 44 years. Amongwomen, patients aged 20 to 24 years with a mental health claim hadthe highest predicted probability of STD diagnoses (3.0%); amongmen, patients aged 25 to 29 years with a mental health claim hadthe highest predicted probability of STD diagnoses (1.2%).

Conclusions: In this population, patients with mental healthclaims were more likely to also have claims with diagnoses forSTDs than patients without mental health claims, and this relationshipapplied to severe and milder mental health disorders. Thissuggests that people with mental health disorders in privatelyinsured populations may benefit from routine STD risk assessmentsto identify high-risk patients for referral to cost-effective preventiveservices.

(Am J Manag Care. 2004;10:917-924)

The relationship between mental illness and sexuallytransmitted diseases (STDs) is a potentiallyimportant research focus with implications forprevention programs. A relationship between mentalhealth conditions and STDs may exist because patientswith complex and varied social and sexual interactionsmay experience higher needs for both types of services.Alternatively, patients with mental health disordersmight exert less conscious control over their sexualinteractions because of feelings of worthlessness, restlessness,boredom, or anxiety.1

Considerable attention has been directed at understandingthis relationship within specific subgroups(adolescents, the indigent, and institutionalized populations)whose members are at increased risk for acquiringSTDs, including human immunodeficiency virus(HIV).2-5 Previous studies of publicly insured, institutionalized,or indigent populations have found a relationshipbetween severe mental health disorders (suchas schizophrenia, bipolar condition, and episodes ofpsychosis) and STDs, including HIV infections. Severalstudies1,6-8 have suggested that patients with a range ofbehavioral control disorders may be at greater risk forSTDs. A recent study9 that evaluated a sample of STDclinic patients in Baltimore found that STD clinicpatients with more severe antisocial personality disorderswere at greater risk of STD infection, while thosewith comparatively less severe mood and substanceabuse disorders were not. Rates of STDs may also behigher in privately insured patients who use mentalhealth services, and the relationship between mentalhealth and STDs may extend to milder mental healthconditions in addition to severe ones.

A significant amount of STD morbidity occurs inpatients who seek care from private sources. In 2002,78% of chlamydia cases and 65% of gonorrhea caseswere reported by private clinicians,10 and 62% ofpatients who needed STD care sought it from their privateclinicians.11 If an association is found between theuse of STD services and mental health services, thisinformation could be used to target more efficient STDprevention services to those patients seeking mentalhealth treatment. More efficient targeting is importantbecause significant barriers impede the identificationand proper treatment of STDs in private settings. Thesebarriers include limited STD knowledge among manyprivate-setting practitioners, low priority given to STDdetection and prevention in many private-sector settings,low patient demand for STD screening and riskassessment, reimbursement policy-driven financial disincentivesfor routine sexual risk assessment and STDscreening of asymptomatic patients, and difficulties inoffering care to sex partners who are not enrolled in thesame health plan as the infected index patient.12,13

National survey data demonstrate that private primarycare providers miss many opportunities to provideSTD, HIV, and pregnancy prevention counseling tohigh-risk patients.14 Still, many private providers do notbelieve that STDs are a major issue for the patients theytreat.15 Some physicians (as opposed to nurse practitionersand physician assistants) may be reluctant tooffer STD preventive services because they doubt thecost effectiveness of the intervention.16 Information thatbetter pinpoints which patients are most likely to needSTD preventive services may make providers more willingto offer such services, particularly if it alters theirbeliefs that these services are important and canimprove the cost effectiveness of delivery.

To determine whether administrative data might beused to target STD risk assessments to mental healthpatients, we examined rates of STD diagnoses forpatients who sought mental health services relative tothose who did not, using medical claims data drawnfrom a large national set of private commercial healthplans. We also considered whether higher rates of STDdiagnoses exist in privately insured populations withmental health service utilization for disorders of mildercharacter (specifically, adjustment disorders), and welook at differences in these associations by age and sex.

METHODS

We analyzed a subset of 289 604 patients from the1995 MarketScan17 database, a national commercialdatabase of the private insurance claims of more than 5million people in 1995. The MarketScan database,although not a probabilistic sample, is drawn frominsurance markets across the United States and hasbeen used in previously published studies18,19 of STDcare. To compile the MarketScan database, the Medstatgroup purchases health insurance claims and enrollmentrecords from corporations, not-for-profit organizations,and municipal employers across the UnitedStates; strips personal identifying information (includinginformation about patient ethnicity) from therecords; cleans the records for inconsistencies; andcombines them into data files for research use. Each ofthe plans in this file used a preferred provider or indemnityfee-for-service reimbursement system. Healthmaintenance organizations that did not collectencounter-level claims information were not includedin the data.

We restricted our analyses to enrollees aged 15 to 44years because this age group has the highest risk ofSTDs.2,3 Additional inclusion criteria included an identifiedhealth plan and reported prescription drug coverageinformation.

International Classification of Diseases,

Ninth Revision, Clinical Modification

In this database, a claim represents a request by aprovider to the patient's insurer for payment for a specificservice rendered. Each claim is coded with informationdescribing the type of service or procedurerendered by the physician and the diagnosis code indicatingthe reason this service was required. We identifiedpatients with mental health diagnoses usingpsychiatric diagnosis claims or psychoactive drug prescriptions(codes used todefine mental health conditions and STDs are availableas an appendix from the author). Psychiatric claims fortobacco addiction, autism, or mental retardation werenot considered mental health diagnoses for this analysis.20 Any patient with at least 1 mental health claimwas classified as having used mental health services. Weclassified claims for adjustment disorders as evidence ofa mild mental health condition. Claims for mentalhealth diagnoses are an imperfect measure of servicesrendered because some providers are uncertain of thediagnosis without further evaluation or testing,providers may intentionally miscode to protect patientconfidentiality, and services paid by the patient out ofpocket do not generate claims.

We determined cases of STDs by identifying peoplewith 1 or more claims with a primary diagnosis code ofa bacterial or viral STD. As with mental health claims,any patient with at least 1 claim coded with a relevantSTD diagnosis was considered to have used STD services.A claim only for STD services such as STD laboratorytests (that may yield a negative result) was notsufficient to be classified as an STD diagnosis. In somecases, providers may intentionally not code certainservices with an STD diagnosis because they are uncertainof the diagnosis without further evaluation and testing,thereby preventing services from appearing on amedical bill and enhancing confidentiality. For this reason,claims measures of STD services are imperfectmeasures of actual STD services and demand. Althoughvariations in STD diagnostic codes are likely to bestrongly correlated with variations in the underlyingrates of STDs,18,21 they should not be mistaken for exactmeasurements of STDs and should instead be interpretedsimply as indicators of demand for STD services.

To demonstrate the relationship betweenhaving a mental health claim and an STDclaim, we first calculated rates of STD diagnosisclaims per 100 000 patients, stratified byage group and sex. To test whether this relationshipwas confounded by other variables,we created a logistic regression model usingthe GENMOD procedure in SAS (Version 6.12,SAS Institute, Cary, NC), with a dichotomousmeasure of any STD diagnoses as the dependentvariable. We considered that several variablesmight confound the relationship of STDsand mental health service diagnoses. The variablesage, sex, industry of employment, metropolitanstatistical area, and subject's healthplan were selected as independent variableson the basis of previous research on the determinantsof STDs10,11,22 and the data availablein the analytic file. Industry was used as aproxy variable for socioeconomic status,which was not available. The 8 industriescoded in the data are oil and gas extraction;mining; durable goods manufacturing; nondurablegoods manufacturing; transportation,communications, and utilities; retail; finance,insurance, and real estate; and government(Table 1). The GENMOD procedure analyzeddata represented in a contingency table and fitted linearmodels to functions of response frequencies.23 Thisallowed us to control for the effect of dozens of individualhealth plans and metropolitan statistical areas, in away similar to a fixed-effects model. All potentially confoundingvariables were retained in the model. Themodel was then used to predict the probability of anSTD diagnosis with and without a mental health diagnosisby age and sex, controlling for a patient's healthplan, metropolitan statistical area, and industry.

To assess whether STD diagnosis claims are associatedwith a later increase in mental health claims, foreach individual we calculated the number of mentalhealth diagnosis claims filed before and after the firstSTD diagnosis during 1995 for those patients withclaims for both STD and mental health diagnoses. Alarge increase in the number of mental health serviceclaims following an STD diagnosis claim would suggestthat the STD diagnosis may to some extent stimulatedemand for the mental health services coded with thisdiagnosis. However, if the numbers of mental healthdiagnosis claims before and after an STD diagnosisclaim were roughly equal, this would suggest that anSTD diagnosis claim was not the sole stimulus for thesubsequent use of services associated with mentalhealth diagnoses. Because the monthly timing of individualSTD diagnoses was likely to be randomly distributedacross the 12 months of 1995, we assumed that thetiming of individual STD diagnoses was not likely toaffect the mean number of claims before and after anSTD diagnosis found at the group level. We tested thedifference in the overall mean number of claims beforeand after the first STD diagnosis, using an F test from ananalysis of variance.

Specifically, our methods test the following 3hypotheses: (1) The rates of STD diagnoses per 100 000patients differ significantly between those with andwithout mental health claims, before adjusting forpotentially confounding variables. (2) Controlling forpotentially confounding variables, people with differentmental health diagnoses experience statisticallyhigher probabilities of also having an STD diagnosis.(3) The mean number of mental health claims doesnot differ before and after an STD diagnostic claimamong patients with both STD and mental health diagnosticclaims.

RESULTS

A total of 36 032 patients (12.4%) had at least 1 claimfor a mental health diagnosis. The most prevalent mental health diagnosis claims were related to adjustmentdisorders and prescriptions for antidepressant medications.A total of 1969 patients (0.7%) had at least 1 claimfor an STD diagnosis. The most commonly diagnosedSTDs were genital herpes, human papillomavirus (genitalwarts), genital lice, pelvic inflammatory disease (inwomen), gonorrhea, nongonococcal urethritis (in men),and HIV/AIDS (in men because there were no instancesof HIV/AIDS diagnoses in women in this sample). Fivehundred ninety-seven patients had both mental healthand STD diagnosis claims. In terms of populationbaseduse of medical services, 680 per 100 000 privatelyinsured patients in the overall sample had a claim foran STD diagnosis, compared with 1657 per 100 000 privatelyinsured patients who received at least 1 claim fora mental health diagnosis.

Substantial differences in rates of diagnosis claimsexisted between age groups and sexes (Table 1).Overall, women with a mental health diagnosis claimwere the most likely to have claims with STD diagnoses,followed by women without a mental health diagnosisclaim, men with such a claim, and men without such aclaim. Significant variation in rates of claims with STDdiagnoses existed by age, generally with younger women(15-29 years) and older men (25-39 years) exhibitinghigher rates of such claims.

After controlling for potentially confounding variablesin the logistic model, we found that patients withat least 1 mental health diagnosis claim were more thantwice as likely as patients without such claims to havehad an STD diagnosis claim in the same year (Table 2)(odds ratio [OR], 2.33; 95% confidence interval [CI],2.11-2.58). Patients prescribed antidepressant medicationswere more thantwice as likely to haveSTD diagnosis claimsas patients withoutthese drug claims (OR,2.21; 95% CI, 1.92-2.55). Compared withother teens aged 15 to19 years, teen patientswho had mental healthdiagnosis claims werenearly twice as likelyalso to have had STDdiagnosis claims (OR,1.98; 95% CI, 1.50-2.63).

Patients with claimsfor antipsychotic medications,bipolar condition,and substanceabuse problems had the highest risk of STD diagnosisclaims, but those with milder conditions also had higherrates of STD diagnosis claims than those withoutthese mental health diagnosis claims (Table 3).Patients diagnosed as having adjustment disorderswere nearly twice as likely to have an STD diagnosis aspatients without adjustment disorders (OR, 1.93; 95%CI, 1.64-2.28).

These differences in odds translated into higherabsolute probabilities of STD diagnoses for patients withmental health conditions. No age group of men withoutmental health diagnoses had a population prevalence ofSTD diagnoses that exceeded 0.3%, and no age group ofwomen without mental health diagnoses had a populationprevalence of STD diagnoses that exceeded 1.0%.In contrast, men with mental health claims had a predictedpopulation prevalence of STD diagnoses between0.3% and 1.2%, and women with mental health claimshad a predicted prevalence between 1.2% and 3.0%(Table 4). The association between STD diagnosisclaims and mental health diagnosis claims was strongeramong women than among men but did not vary appreciablyby age. For persons with both mental health andSTD diagnosis claims in the same year, no significantdifference was found between the number of claims formental health diagnoses and drugs before (mean, 3.24)and after (mean, 3.31) a diagnosis of an STD.

DISCUSSION

Our results show a significant relationship betweenclaims for STD diagnoses and mental health diagnoses.This was observed in women and men and in all agegroups within the 15-to 44-year age range. Our studyis consistent with other investigations24 in showing thatthe occurrence of an STD diagnosis claim was not associatedwith a later increase in mental health diagnosisclaims. Furthermore, we found an association forpatients with severe mental health disorders and prescriptions,such as bipolar condition and antipsychoticmedications, as well as for patients with milder conditionsand prescriptions, such as adjustment disordersand antidepressant medications.

This association translated into substantiallyincreased absolute probabilities of STD diagnoses forpatients with mental health diagnoses compared withpatients without such diagnoses. Furthermore, diagnosedSTDs likely indicate a substantial amount ofundiagnosed disease in the community. A population-basedstudy25 of Baltimore, Md, households found thatthe number of prevalent undiagnosed gonococcal andchlamydia infections in adults aged 18 to 35 yearsapproached or exceeded the number of infections thatwere diagnosed and treated annually.

The low prevalence of diagnosed STDs in thosewithout mental health claims is unlikely to attract theattention of most plan managers. However, the 1% to3% prevalence in enrollees with mental health claimsgreatly exceeds the yearly rate of new diagnoses forbreast cancer and prostate cancer (approximately0.3%26 and 0.2%,27 respectively, of the over age 40female and male populations), both of which are routinelyscreened for in private settings. Routine chlamydiascreening of sexually active women 26 and youngerwas ranked as a better prevention value than mammographyin a recent ranking of preventive services basedon burden of disease prevented and cost-effectiveness ofthe service.28

This study may be limited by possible miscoding ofSTDs and mental health claims that was unintentionalor intentional (eg, because the diagnosis was uncertainor to protect patient confidentiality). In addition, theassociation between mental health and STD claims maybe confounded by other variables, specifically race, forwhich data were not available. In addition, because thetiming of STD diagnoses was not retained in the dataanalysis file, we assumed that STD diagnoses were normallydistributed throughout the year. Conclusionsdrawn from the difference in the mean number ofclaims before and after a mental health diagnosis relyon the validity of this assumption.

Substance abuse may play a larger role in driving therelationship between mental health and STD diagnosesthan is apparent from these claims data. Like STDs,substance abuse disorders may not be coded at all ormay be intentionally miscoded as milder mental healthconditions such as adjustment disorders to protectpatient confidentiality. Clinical follow-up studies of thiswork should attempt to collect information on substanceuse behaviors to separate (to the extent possible)the disinhibiting effect of inebriation on the behaviorsthat lead to STDs from the effect of mental health disordersalone. In these data, patients with diagnosedsubstance abuse problems were not driving the relationshipbetween mental health and STD diagnoses.First, only 2.9% of the group with mental health diagnoseshad claims only for substance abuse problemswith no other mental health condition diagnosed.Second, patients with mental health diagnoses otherthan substance abuse also experienced higher rates ofSTD diagnoses (Table 2).

Our findings may not be generalizable to populationsthat tend to be less affluent or experience a higher incidenceof bacterial and viral STDs.9,29 In addition, the1995 data used in this study may not be representativeof populations today because of different clinical andclaims-coding practices for mental health and STD services.Future research should be conducted to verifythese results using more recent data. This researchshould use a longitudinal design to better address issuesof causality between STDs and mental health disordersand should assess the cost effectiveness of targetingSTD/HIV risk assessments to privately insured peoplewho use mental health services.

However, although additional studies on this subjectusing newer data are needed, the theory that mentalhealth disorders may contribute to behaviors that resultin STDs is still valid. Furthermore, our findings are consistentwith more recent nonclaims data showing a similarrelationship in an STD clinic setting.9 Our ability toobserve this relationship may be even stronger todaythan in 1995 for several reasons. If the enactment of theMental Health Parity Act of 1996 resulted in greater privateinsurance coverage of mental health services, itmay have increased the number of people with mentalhealth disorders who can be identified using inexpensive,targeted STD risk assessment. Second, the deliveryof STD services in private settings has likely increasedsince 1995 because of the expansion of Medicaid managedcare, the advent of the State Child Health InsuranceProgram, and the elimination of many publicly fundedSTD clinics; this last reason has likely pushed more peopleto seek care from private providers. These changeshave likely increased the number of patients at high riskfor STDs seen in private settings, thus increasing theneed for and the potential effectiveness of STD riskassessments in private settings as well. Although ideallythis research should be replicated with more recentdata, the core findings seem unlikely to change.

Many private-sector providers still believe that routinesexual risk assessment is a low priority for theirpatients or avoid conducting risk assessments becausethey believe their patients have a low prevalence ofSTDs (a belief that may not be supported by data10,11)because face-to-face risk assessment makes someproviders or patientsfeel uncomfortable orbecause some insurers donot reimburse adequatelyfor risk assessments.

However, our datasuggest that, for patientswith mental health conditions,STD risk assessmentis warranted, andhealth systems andproviders should attemptto identify methods toimplement it routinely.Improvements in primarycare STD/HIV riskassessment can beachieved in outpatientsettings using inexpensiveinterventions30 suchas patient self-assessment forms, nonphysician counselors,education pamphlets, or videos. Examples ofpatient self-assessment tools can be accessed on theInternet.31,32

Routine risk-assessment questions can be integratedinto existing patient forms and questionnaires at negligiblecost to the provider or the insurer. Based on studiesof other interventions to prevent STDs, targetedlow-cost risk assessments are likely to be cost effectivefor the provider, given the proven cost effectiveness oftargeted STD preventive services.

Chlamydia trachomatis

For example, research has found that screening forasymptomatic with a nucleicacid amplification-based test is cost-effective in femalepopulations with prevalences as low as 1.1%.33 Otherresearch has shown that most cost savings associatedwith treating asymptomatic chlamydial infections likelyaccrue within the same year.34 In women, undiagnosedlow-cost bacterial infections may develop into acutecases of pelvic inflammatory disease with costs between$1060 and $1410 per case.35 In the population of roughly23 000 women with mental health diagnosesobserved in these data, 244 cases of pelvic inflammatorydisease were diagnosed, compared with only a combined44 cases of diagnosed gonorrhea and chlamydia,suggesting that targeted bacterial STD screening basedon risk assessments would have likely improved healthoutcomes and generated savings. Treatment of mostSTDs detected early in the course of infection is inexpensive,usually involving 1 to 7 days of antibiotic treatment.Treatment of upper genital tract complications ofuntreated lower genital tract infection such as pelvicinflammatory disease, ectopic pregnancy, and infertilityis more costly; this underscores the potential value ofearly STD detection through risk assessment and referralto preventive screening and other services of highriskpatients. Even costly diseases such as HIV/AIDSare less costly to manage when identified early.Research has demonstrated that early detection andproper management of HIV prevent many of the highcostinpatient admissions associated with acute symptomaticAIDS.36

At the insurance plan level, administrative claimsinformation could be used to target STD preventiveservices to specific high-risk individuals. Insurers couldrecommend sexual risk assessments and referrals toscreening and services for those at risk as elements oftheir standard mental health care. Mental health claimscould also be used to trigger mechanisms (such as confidentialpersonalized communication) that target specificpatients for STD/HIV risk assessment based ontheir use of mental health services. Such policies areethical and legal under current regulations, providedthe information that is already available to the plan'sbenefit managers and participating providers is beingused to enhance patient health and is administeredconfidentially and with sensitivity. Legally, federalHealth Insurance Portability and Accountability Actregulations specifically allow the within-plan use ofpatient administrative health information to target preventivecare (D. Magid, MD, oral communication,2004).37

Sexually transmitted disease screening interventionshave been shown to improve health while savingcosts in several clinical and nonclinical settings.33,34,36The 2 primary determinants of the cost effectiveness ofthese interventions were the disease prevalence andthe intervention cost: the higher the STD prevalenceand the lower the cost of the intervention, the greaterthe benefits of the intervention. Anecdotal reports suggestthat sexual health issues are not generally a concernfor medical groups, psychiatric groups, orhospitals. This is perhaps justified because, before thisstudy, no published literature addressed the question ofwhether the prevalence of STD diagnoses in privatelyinsured populations would economically justify preventivehealth interventions. However, this article demonstratesa prevalence of STD diagnoses in thesubpopulation of privately insured patients with mentalhealth diagnoses or using psychoactive prescriptionsthat exceeds the disease prevalence at which other STDinterventions were found to be cost effective.33,34,36Furthermore, current STD policy makers argue thatmental health systems should implement simple referralsystems to facilitate routine STD care and screening.38 Although our data are from 1995, we have reasonto believe the relationship between mental health diagnosesand STDs remains true.9-11,14,22

Given our results, we suggest that patients withmental health disorders between the ages of 15 and 44years be routinely assessed for STD/HIV risks, particularlyin areas of the country where the overall populationprevalence of STDs is high. This recommendationis likely to be particularly useful in targeting servicestoward women, especially those between the ages of 15and 29. Health systems and providers should considerhow confidential use of claims data could facilitate thetargeting of patients with mental health diagnoses whomay benefit from STD/HIV risk assessment.

Acknowledgments

We thank Terry Chorba for his helpful review and Sharon Barrell forher editorial support.

From the Division of Health Economics Research, RTI (Research Triangle Institute)International, Atlanta, Ga (DBR), and Center for HIV, STD, and TB Prevention, NationalDivision of Sexually Transmitted Disease Prevention, Centers for Disease Control andPrevention (CDC), Atlanta, Ga (LAA, KLI).

Portions of Dr Rein's work were conducted under and supported by the ResearchParticipation Fellowship Program at CDC, administered by the Oak Ridge Institute forScience and Education, Oak Ridge, Tenn, through an interagency agreement between theUS Department of Energy and CDC. Additional funding was supplied by RTI International.

Address correspondence to: David B. Rein, PhD, Division of Health EconomicsResearch, RTI International, 2951 Flowers Road South, Suite 119, Atlanta, GA 30341.E-mail: drein@rti.org.

J Am Acad Child Adolesc

Psychiatry.

1. Brown LK, Danovsky MB, Lourie KJ, DiClemente RJ, Ponton LE. Adolescentswith psychiatric disorders and the risk of HIV. 1997;36:1609-1617.

MMWR Morb Mortal Wkly Rep.

2. Centers for Disease Control and Prevention. 1998 Guidelines for treatment ofsexually transmitted diseases. 1998;47(RR-1):1-116.

Am J Public Health.

3. Richert CA, Peterman TA, Zaidi AA, Ransom RL, Wroten JE, Witte JJ. Amethod for identifying persons at high risk for sexually transmitted infections:opportunity for targeting intervention. 1993;83:520-524.

J Acquir Immune

Defic Syndr Hum Retrovirol.

4. Des Jarlais DC, Friedman SR, Perlis T, et al. Risk behavior and HIV infectionamong new drug injectors in the era of AIDS in New York City. 1999;20:67-72.

Am J Psychiatry.

5. Kelly JA, Murphy DA, Bahr GR, et al. AIDS/HIV risk behavior among thechronically mentally ill. 1992;149:886-889.

Arch Sex

Behav.

6. Kafka MP. Hypersexual desire in males: an operational definition and clinicalimplications for males with paraphilias and paraphilia-related disorders. 1997;26:505-552.

Drug Alcohol Depend.

7. Booth R, Zhang Y. Conduct disorder and HIV risk behaviors among runawayand homeless adolescents. 1997;48:69-76.

Sex Addict Compulsivity.

8. Gordon LJ III, Fargason PJ, Kramer JJ. Sexual behaviors of patients in a residentialchemical dependency program: comparison of sexually compulsive physiciansand nonphysicians with non-sexually compulsive physicians andnonphysicians. 1995;2:233-255.

Sex Transm Dis.

9. Erbelding EJ, Huttun HE, Zenilman JM, Hunt WP, Lyketsos CG. The prevalenceof psychiatric disorders in sexually transmitted disease clinic patients and theirassociation with sexually transmitted disease risk. 2004;31:8-12.

10. Centers for Disease Control and Prevention. Sexually transmitted disease surveillance.Table 2: reported cases of sexually transmitted disease by sex andreporting source: United States, 2002. Available at: http://www.cdc.gov/std/stats/tables/table2.htm. Accessed January 20, 2004.

Fam Plann Perspect.

11. Brackbill R, Sternberg M, Fishbein M. Where do people go for treatment ofsexually transmitted diseases? 1999;31:10-15.

Sex Transm Dis.

12. Rein DB, Anderson LA, Gowda V, Dixon J, Irwin KL. Federally funded sexuallytransmitted disease programs and managed care: a review of current andplanned partnerships. 2001;28:336-342.

MMWR Morb Mortal Wkly Rep.

13. Centers for Disease Control and Prevention. Prevention and managed care:opportunities for managed care organizations, purchasers of health care, and publichealth agencies. 1995;44:1-12.

Pediatrics.

14. Burstein GR, Lowry R, Klein JD, Santelli JS. Missed opportunities for sexuallytransmitted diseases, human immunodeficiency virus, and pregnancy preventionservices during adolescent health supervision visits. 2003;111(pt 1):996-1001.

Sex Transm Dis.

15. Ashton MR, Cook RL, Wiesenfeld HC, et al. Primary care physician attitudesregarding sexually transmitted diseases. 2002;29:246-251.

Sex Transm Infect.

16. Boekeloo BO, Snyder MH, Bobbin M, et al. Provider willingness to screen allsexually active adolescents for chlamydia. 2002;78:369-373.

MarketScan Database User's Guide.

17. MarketScan. Ann Arbor, Mich: MedstatGroup Inc; 1997.

Chlamydia trachomatis.

Jt Comm J Qual Improv.

18. Tao G, Walsh CM, Anderson LA, Irwin KL. Understanding sexual activitydefined in the HEDIS measure of screening young women for 2002;28:435-440.

Obstet Gynecol.

19. Rein DB, Kassler WJ, Irwin KL, Rabiee L. Direct medical cost of pelvic inflammatorydisease and its sequelae: decreasing, but still substantial. 2000;95:397-402.

International Classification of Diseases, Ninth

Revision, Clinical Modification.

20. American Medical Association. Salt Lake City, Utah: Medicode Publications; 1996.

Sex Transm Dis.

21. Ratelle S, Yokoe D, Blejan C, et al. Predictive value of clinical diagnosticcodes for the CDC case definition of pelvic inflammatory disease (PID): implicationsfor surveillance. 2003;30:866-870.

Sex Transm Dis.

22. Wilson SR, Brown NL, Leyden WA, et al. Healthcare utilization by women ina comprehensive managed care population subsequent to diagnosis of a sexuallytransmitted disease. 2002;29:678-688.

SAS/STAT User's Guide, Version 6.

23. SAS Institute Inc. Vol 1. 4th ed. Cary, NC:SAS Institute Inc; 1989.

J Fam Pract.

24. Reed BD, Ruffin MT, Gorenflo DW, Zazove P. The psychosexual impact ofhuman papillomavirus cervical infections. 1999;48:110-116.

JAMA.

25. Turner CF, Rogers SM, Miller HG, et al. Untreated gonococcal and chlamydialinfection in a probability sample of adults. 2002;287:726-733.

MMWR Morb Mortal Wkly Rep.

26. Centers for Disease Control and Prevention. Breast cancer incidence and mortality:United States: 1992. 1996;45:833-837.

JAMA.

27. Potosky AL, Miller BA, Albertson PC, Kramer BS. The role of increasing detectionin the rising incidence of prostate cancer. 1995;273:548-552.

Am J Prev Med.

28. Coffield AB, Maciosek MV, McGinnis JM, et al. Priorities among recommendedclinical preventive services. 2001;21:1-9.

Indicators and Predictors of Health Services Utilization.

29. Aday LA. 2nd ed.Chicago, Ill: Health Administration Press; 1998.

Am J Prev Med.

30. Dodge WT, BlueSpruce J, Grothaus L, et al. Enhancing primary care HIV prevention:a comprehensive clinical intervention. 2001;20:177-183.

AIDS on the College

Campus: A Special Report of the American College Health Association;

31. Wessel LG. Risk assessment form. Extracted from: 1989.Available at: http://www.vpul.upenn.edu/ohe/library/Sexhealth/hiv/riskassesment.htm. Accessed February 10, 2004.

The Handbook for Evaluating HIV

Education.

32. National Center for Chronic Disease Prevention and Health Promotion,Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/HealthyYouth/publications/hiv_handbook/index.htm. Accessed September 30, 2004.

Chlamydia trachomatis

Ann Intern Med.

33. Howell MR, Quinn TC, Gaydos CA. Screening for inasymptomatic women attending family planning clinics: a cost-effectiveness analysisof three strategies. 1998;128:277-284.

N Engl J Med.

34. Scholes D, Stergachis A, Heidrich FE, et al. Prevention of pelvic inflammatorydisease by screening for cervical chlamydial infection. 1996;334:1362-1366.

Sex Transm Dis.

35. Rein DB, Gift TL. A refined estimate of the lifetime cost of pelvic inflammatorydisease [comment]. 2004;31:325.

AIDS Read.

36. Velenti WM. HIV disease management: new technologies improve outcomesand contain costs. 2002;12:64-66.

37. Department of Health and Human Services. OCR HIPAA privacy guidance:marketing. Available at: http://www.hhs.gov/ocr/hipaa/guidelines/marketing.rtf.Accessed January 12, 2004.

Sex Transm Dis.

38. Aral SO. Mental health: a powerful predictor of sexual health? 2004;31:13-14.