1 Predictors of Retention in Care Among HIV+ and At-Risk Youth Sion Kim Harris, PhD Cathryn L. Samples, MD, MPH Peter Keenan, RN, C-PNP Durrell J. Fox,

1

Predictors of Retention in CareAmong HIV+ and At-Risk Youth

Sion Kim Harris, PhD

Cathryn L. Samples, MD, MPH

Peter Keenan, RN, C-PNP

Durrell J. Fox, BS

Maurice W. Melchiono, RN, MS, C-FNP

Elizabeth R. Woods, MD, MPH

Boston HAPPENS Program Participants

Children’s Hospital Boston, Harvard Medical School

Society for Adolescent Medicine Annual Meeting 2002

2

Study Grant Support• Special Projects of National Significance Program

(SPNS) #BRH 970155-05-0, HRSA, DHHS

• Leadership in Adolescent Health Project #5T 71 MC 00009-10, MCH Bureau, HRSA, DHHS

• W.T. Grant Foundation

• Children’s Hospital’s Aerosmith Fund for HIV Care and Prevention

3

Background

• At-risk youth are hard to engage and retain in care

• Ancillary services such as outreach, case management, and mental health services believed to enhance retention in care

4

Study Aims

• Describe retention in care of youth clients of a comprehensive adolescent HIV care program (Boston HAPPENS Program)

• Assess impact of outreach, case management, mental health services on likelihood of retention over time, adjusted for potential confounding factors

5

Program Description

• Boston HIV Adolescent Provider and Peer Education Network for Services (HAPPENS)

• Coordinated network of 8 agencies serving at-risk youth ages 12-24

• Comprehensive care including HIV CTS, outreach, nurse case management, mental health services, and medical care

6

Study Population

• HAPPENS clients seen during data collection period (4.5 years)

• Clients excluded from analysis if had no more than two standard HIV counseling and testing visits

7

Data Collection Method

• Standard forms used by all 10 SPNS programs

• “Contact” form: used at client enrollment or street/phone contact to record demographic and HIV risk info, outreach services provided

• “Intervention” form: used at each health visit to record health care, case management, mental health services provided

8

Data Analyses MethodSurvival analysis:

• Retention measure

– Number of days between first and last visits during data collection period

• Censored cases

– Clients with at least one visit during last program year

9

Key Predictor Variables

• Number of outreach contactsCategories: 0, 1, >2 times

• Number of mental health service visitsCategories: 0, 1, >2 times

• Number of case management visitsCategories: 0, 1-2, >3 times

10

Other Potential Predictors

• Age• Gender• Race/ethnicity• HIV status• Mental health system

involvement• Homeless/runaway• Self-identified sexual

orientation

• Ever pregnant • Ever had STD• Any unprotected sex• Any survival sex, sex with

HIV+ or injection drug using partner (high risk sex)

• Any substance abuse, injection drug use, or needle-sharing

Client Demographic and HIV Risk Factors:

11

• Type of site of initial care– Multi-service outreach agency

– Community health center

– Hospital

Other Potential Predictors (cont’d)

12

Statistical Tests

• All analyses were gender-stratified

• Descriptive analyses:– Kaplan-Meier survival curve analysis

• Bivariate analyses:– Log-rank test to assess group

differences in retention times

13

Statistical Tests (cont’d)

• Multivariate Cox Proportional Hazards (PH) regression modeling– Variables related to retention at p<0.10 entered

into model, including age, HIV status– Adjusted Hazard Ratio (AHR) and 95% CI

• PROC SURVIVAL in SUDAAN® used to account for potential intra-cluster correlation within sites

14

Client Characteristics by Gender (Total N=1426)

Males FemalesGender

differencen % n % p-value

Number of Clients 512 36 914 64 –––

Age (mean years + SD) 20 ± 3 19 ± 2 <0.001

Race/Hispanic ethnicity:

White non-Hispanic 275 57 401 45 0.002

Black non-Hispanic 91 19 207 23

Hispanic 83 17 194 22

Asian/Other 34 7 81 9

HIV positive 21 4 16 2 0.007

Gay, lesbian, bisexual orundecided

112 22 63 7 <0.001

Homeless and/or runaway 77 15 131 14 NS

15

Overall Program Retention

• Retention times ranged from 0-1406 days for males, 0-1577 for females

• About two-thirds of clients returned for at least a second visit (males=64.3%, females=70.0%)

• 25% in both gender groups retained for more than a year

• 10% of males and 15% of females retained for more than two years

16

Probability of Retention Over Timeby Number of Outreach Contacts

MALES FEMALES

>2 contacts*

0

1

>2 contacts*

0

1¥

Number of days between first and last visits

Cum

Pro

babi

li ty

of R

ete n

tion

Log-rank test: *p<0.001; ¥p=0.002

17

Probability of Retention Over Timeby Number of Mental Health Visits

MALES FEMALES

Cum

Pro

babi

li ty

of R

ete n

tion

Number of days between first and last visits

>2 visits*

0

1

>2 visits*

0

1¥

Log-rank test: *p<0.001; ¥p=0.03

18

Probability of Retention Over Timeby Case Management Visits

Cum

Pro

babi

li ty

of R

ete n

tion

Number of days between first and last visits

MALES

>3 visits*

0

1-2

>3 visits*

0

1-2

FEMALES

Log-rank test: *p<0.001

19

Cox PH Modeling Results, MalesAdjustedHazardRatio* (95% CI) p-value

Outreach contacts0 Reference Reference1 1.38 (1.14-1.68) 0.005

>2 0.64 (0.57-0.73) <0.001

Mental health counseling visits0 Reference Reference1 0.83 (0.61-1.13) NS

>2 0.66 (0.54-0.80) <0.001

Case management visits0 Reference Reference

1-2 1.08 (0.92-1.26) NS>3 0.57 (0.46-0.69) <0.001

* Adjusted for age, HIV status, race/ethnicity, sexual orientation, homelessness, MH-system involvement, and high risk sex behaviors

20

Cox PH Modeling Results, FemalesAdjustedHazardRatio* (95% CI) p-value

Outreach contacts0 Reference Reference1 1.32 (1.09-1.61) NS

>2 0.72 (0.61-0.85) 0.002

Mental health counseling visits0 Reference Reference1 0.95 (0.61-1.50) NS

>2 0.59 (0.18-1.92) NS

Case management visits0 Reference Reference

1-2 1.06 (0.75-1.50) NS>3 0.54 (0.34-0.84) 0.013

* Adjusted for age, HIV status, race/ethnicity, sexual orientation, any pregnancy, and any unprotected sex with males

21

Discussion

• Some hard-to-reach, at-risk youth can be retained in care, with intensive efforts

• Increased outreach and case management associated with retention (adjusted for client demographic and risk factors)

• Mental health counseling related to retention of male clients

22

Study Limitations

• HAPPENS clients may not be representative of all at-risk youth

• Visits to providers outside of program not included

• Variability in quality, completeness of data across sites

• Cell sizes for some variables small

23

• Other client and program factors related to retention unmeasured

Study Limitations (cont’d)

24

Implications

• Comprehensive services help retain at-risk youth in care

• Ancillary services such as outreach, case management, mental health services need to be more consistently funded for all at-risk youth, not just HIV+