An HIV health plan patient navigation program: Engaging HIV

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=whiv20

Journal of HIV/AIDS & Social Services

ISSN: 1538-1501 (Print) 1538-151X (Online) Journal homepage: https://www.tandfonline.com/loi/whiv20

An HIV health plan patient navigation program:

Engaging HIV positive individuals in primary

medical care

Peter Messeri, Maiko Yomogida, Rachel M. Ferat, Lee Garr & Doug Wirth

To cite this article: Peter Messeri, Maiko Yomogida, Rachel M. Ferat, Lee Garr & Doug Wirth

(2019): An HIV health plan patient navigation program: Engaging HIV positive individuals in primary

medical care, Journal of HIV/AIDS & Social Services, DOI: 10.1080/15381501.2019.1699485

To link to this article: https://doi.org/10.1080/15381501.2019.1699485

Published online: 18 Dec 2019.

Submit your article to this journal

View related articles

View Crossmark data

An HIV health plan patient navigation program:

Engaging HIV positive individuals in primary

medical care

Peter Messeri

, Maiko Yomogida

, Rachel M. Ferat

, Lee Garr

, and

Doug Wirth

Department of Sociomedical Sciences, Mailman School of Public Health, Columbia University,

New York, NY, USA;

Mailman School of Public Health, Columbia University, New York, NY, USA;

AMDA Care, New York, NY, USA

ABSTRACT

This paper evaluates a patient navigation component of a

multi-level program that helps HIV-positive members of a New

York City Medicaid health plan sustain engagement in medical

care. A proportional hazard analysis of 856 participants found

that assignment to a patient navigator shortened the time to

a medical care visit by 40%, but was not associated with time

retained in care. These results demonstrate that a health plan

can expedite connection to care through patient navigation

services. They further suggest that to sustain retention in care,

patient navigation may need to be continued after initial con-

nection to care.

ARTICLE HISTORY

Received 17 October 2018

Revised 10 October 2019

Accepted 27 November 2019

KEYWORDS

Entry into care; HIV care;

HIV/AIDS; patient navigator;

program evaluation;

treatment retention

Introduction

Efforts to end the HIV/AIDS epidemic partly rests on early detection and

life-long adherence to effective anti-retroviral medication among infected

individuals. Sustained adherence to medical regimens, in turn, depends on

a life-long engagement in medical care. Sustaining a life-long engagement

in medical care for a chronic health condition, such as HIV/AIDS, is diffi-

cult for any patient population, but it is particularly challenging for the

care of HIV-positive patients, who are concentrated among economically

and socially disadvantage populations.

Patient navigation (PN) is a common intervention in HIV/AIDS that is

intended to support client engagement in medical care (AIDS United, 2015;

Higa, Marks, Crepaz, Liau, & Lyles, 2012; Liau et al., 2013; Mizuno et al.,

2018; Thompson et al., 2012). Navigators support patient decisions to use

medical services by assisting clients in developing care plans, scheduling

CONTACT Peter Messeri [email protected] Mailman School of Public Health, Department of

Sociomedical Sciences, Columbia University, 722 W. 168 St. Room 549B, New York, NY 10027-6902, USA.

Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/whiv.

ß 2019 Taylor & Francis Group, LLC

JOURNAL OF HIV/AIDS & SOCIAL SERVICES

https://doi.org/10.1080/15381501.2019.1699485

appointments, and possibly escorting them to appointments once sched-

uled. PN may target recently diagnosed patients, those weakly connected to

care, patients who are at high risk of dropping out of care, or those who

have dropped out of care.

Drawing strong empirical generalization about PN effectiveness is com-

plicated as features of PN-type interventions described in the research lit-

erature vary in important respects regarding job qualifications, on-the-job

training, scope of navigator tasks, client characteristics, organizational set-

tings in which PN is situated, duration of patient-navigator contact, stage

of disease, and choice of outcomes. Variation along one or more of these

dimensions may have consequences for the differential effectiveness of PN

services. Moreover, because PN is often imbedded in a multi-component

intervention, the PN effect cannot be easily disentangled from the effects of

other components (Higa et al., 2012; Simoni, Amico, Smith, & Nelson,

2010). Inferring PN effects are further weakened in many studies due to

the absence of a non-PN comparison group.

This study reports the results of a PN program implemented by Amida

Care (AC), a New York City Medicaid HIV health plan. The location of

the PN program within a health plan is rare, if not unique, and can be dis-

tinguished from other PN programs that are situated in medical care

organizations or community based organizations. Using administrative and

insurance claims data, we investigate whether assignment to a patient navi-

gator reduced the time to first HIV primary care visit following program

enrollment. We also investigate whether patient navigator assignment had a

downstream effect on improving retention in care following first visit.

Before turning to the empirical findings of the study, we review the existing

literature on patient navigation and HIV care.

Patient navigation theory and concepts

Bradford, Coleman and Cunningham define patient navigation, or using

their term, HIV system navigation, as an “emerging model of care coordin-

ation within HIV care. Navigators assist HIV patients to make better use of

available resources, develop effective communication with providers, sustain

HIV care over time, and navigate the complexities of multidisciplinary

treatment.” (Bradford, Coleman, & Cunningham, 2007:S50) This definition

draws attention to two barriers that obstruct sustained patient engagement

in medical care: a fragmented service system for treating co-occurring con-

ditions such as mental health and substance abuse disorders, and a patient’s

lack of knowledge and communication skills that impede productive

encounters with health care providers. To overcome these barriers to

engagement in care, navigators may extend assistance that includes helping

2 P. MESSERI ET AL.

clients schedule appointments, escorting clients on visits to primary HIV

medical care and other health care services, identifying clients’ resources

and talents so that they can navigate the health care system on their own,

and coaching clients in effective communication with medical providers. A

further function of navigators is to extend moral and emotional support in

a culturally competent manner that reinforces patient self-management

skills (Bradford et al., 2007). The theoretical foundations for training navi-

gators draw upon strength-based perspective from social work (Saleebey,

1997), motivational interviewing (Miller & Rollnick, 2002), and the trans-

theoretical stages of change (Littell & Girvin, 2002).

Empirical literature on HIV PN Effectiveness

Although several published studies in HIV/AIDS have described PN-type

interventions (Andersen et al., 2007; Bradford et al., 2007; Chin, Kang,

Kim, Martinez, & Eckholdt, 2006; Coleman et al., 2007; Craw et al., 2008;

Cunningham et al., 2018; Gardner et al., 2005; Harris et al., 2003; Leider,

Fettig, & Calderon, 2011; Metsch et al., 2016; Molitor et al., 2005; Myers

et al., 2018; Naar-King et al., 2007; Rich et al., 2001; Wohl et al., 2011;

Zaller et al., 2008), the evidence for PN effectiveness is limited. Studies of

HIV/AIDS populations generally find an increase in medical care engage-

ment and an improvement in clinical outcomes post patient-navigation-

type interventions (Andersen et al., 2007; Bradford et al., 2007; Leider

et al., 2011; Mizuno et al., 2018; Rich et al., 2001). However, the absence of

nonintervention comparison groups in the large majority of studies

(Mizuno et al., 2018) weakens attribution of these results to patient naviga-

tion. The published results of four randomized clinical trials of PN-Type

interventions: ARTAS (Gardner et al., 2005), Project Hope (Metsch et al.,

2016), and clinical trials of jail release programs in Los Angeles

(Cunningham et al., 2018) and in San Francisco (Myers et al., 2018) are

mixed. ARTAS tested a brief case management intervention–which per-

formed functions equivalent to what we define as patient navigation–among

newly diagnosed HIV infected individuals that increased initial connection

to care by 40% during a 1-year follow-up period (Gardner et al., 2005). A

more recent multi-site RCT, Project Hope, involved a six-month interven-

tion targeting hospitalized HIV-infected substance users, randomly assigned

to one of three conditions–assignment to a patient navigator with and

without financial incentives and a usual care group. This study found no

significant differences between experimental conditions in viral suppression

at 12-month follow-up (Metsch et al., 2016). The Los Angeles study, LINK

LA, trained patient navigators in a broad range of skills to help HIV-

positive individuals navigate reentry into the community. The San

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 3

Francisco patient navigators offered services similar to the Los Angeles

intervention that were performed in tandem with a case manager. The

usual care condition included jail discharge planning and transitional case

management. Participants in the patient navigation arm of both studies

were more quickly linked to care following community reentry than partici-

pants in the treatment as usual group. There was no difference in retention

in care after 12 months between the experimental and control conditions

for the LINK LA study, but individuals in the PN-intervention group were

more likely to be retained in care in the San Francisco study. Assignment

to the patient navigation arm was associated with better viral suppression

outcomes after 12 months in the LINK LA study but not the San Francisco

study. The mixed results of the four RCT’s underscore the importance of

patient characteristics and organizational setting in determining PN pro-

gram effectiveness.

This study extends outcome studies of HIV/AIDS patient navigation to a

program situated in a Medicaid health plan that serves a socially and eco-

nomically disadvantaged population. The patient navigation program was

one component of ACCESS NY, a multi-level initiative that the Amida

Care health plan designed to promote increased connection to and reten-

tion in medical care among recently enrolled health plan members uncon-

nected to care and to reengage long-time members who had dropped out

of care (Jain et al., 2015). In contrast to PN programs based in medical

care or community-based organizations, the patient navigator-client rela-

tionship based in a Medicaid health plan is less likely to be disrupted if a

member chooses to discontinue care or change providers. Furthermore, the

affiliation of patient navigation with this Medicaid health plan has the

added advantage that patient navigators could assist their clients in con-

necting to their choice of providers from among a large network of over

200 HIV primary medical care providers.

Study hypotheses

Based upon the above literature review, we tested two hypotheses about the

impact of the ACCESS NY PN program on engagement in medical care.

We first hypothesized that assignment to a patient navigator, operating

through visit reminders and escort service, would shorten the duration to

an initial HIV medical care visit following enrollment in ACCESS NY. To

the extent that the navigators improve their clients’ self-management skills,

we further hypothesized that a PN effect should persist after the patient-

navigator relationship ended and would be associated with lengthened

retention in care following first HIV medical care visit post ACCESS

NY enrollment.

4 P. MESSERI ET AL.

Method

Intervention

The ACCESS NY model of patient navigation was designed as a time-lim-

ited intervention that focused on re-connecting clients who had no HIV

primary care visit within six month of enrollment into ACCESS NY. The

program employed patient navigators with both peer and professional qual-

ifications: community health outreach workers (CHOWs) are HIV infected

peers and health navigators (HNs) have a bachelor’s degree in human serv-

ices and experience in working with HIV infected individuals. CHOWs

helped their clients maintain scheduled appointments, understand the

nature of their disease, and manage their care. In addition to the responsi-

bilities of CHOWs, HNs conducted needs assessments, developed care

plans and assisted in care plan implementation.

Upon case assignment, a CHOW/HN conducted up to five outreach

attempts through a combination of an introductory letter, phone calls and

a home visit. Following a successful outreach, ACCESS NY protocol called

for a minimum of weekly contact between navigator and a client. The

CHOW/HN assisted clients in making their initial appointments and

accompanied clients to their first medical appointment to ensure the service

was a good fit and engagement occurred. To ensure continuing connection

to care, the CHOW/HN assisted clients in making further appointments

and facilitated transportation to and from appointments. The CHOW/HN

also helped clients fill medication prescriptions, connect to case manage-

ment and receive lab work. They also follow up with service providers to

assure that members were engaged in services and received the appropriate

level of care from providers who understood and could meet the patients’

special needs. A case remained open for a maximum of three months.

Because of sample size limitations, this study does not separately assess

the effectiveness CHOWs and HNs.

ACCESS NY PN training drew upon the strengths-based perspective from

social work (Saleebey, 1997). CHOWs and HNs underwent a one-week, in-

house training program that covered motivational interviewing, enhanced out-

reach strategies, and cultural diversity. Monthly staff meetings cover topics

that include writing progress notes, field safety and de-escalation skills build-

ing. ACCESS NY protocol called for a minimum of weekly contact between

navigator and a client over a 3-month period. ACCESS NY began enrolling

clients on July 1, 2010, and was closed to new clients on April 15, 2013.

Study participants

AC serves an HIV-positive, Medicaid eligible population that resides in all

five NYC boroughs. During the study period, AC membership rapidly grew

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 5

from 4,000 at the start of the study in 2010 to over 6,000 by its end in

2013. AC members are predominantly African Americans and Latinos.

Besides HIV disease, many AC members struggle with mental illness, sub-

stance abuse, and housing instability. AC members have a choice of over

200 NYC HIV medical care providers affiliated with neighborhood health

centers, ambulatory clinics affiliated with large medical care clinics, medical

care units of AIDS service organizations, and private practices. AC mem-

bers may select services from an even larger network of behavioral medi-

cine and drug treatment programs.

Two groups of AC members were eligible for ACCESS NY services. The

first was composed of AC members who had “dropped out of care.” A

dropout was a long-time AC member, for whom the most recent claim for

HIV primary care visit was 180 days or longer at time of ACCESS NY

enrollment. The second group eligible for ACCESS NY was “unconnected

to care.” They were newly enrolled AC members (within past 179 days),

who had not designated an existing HIV medical care specialist at time of

AC enrollment. Based upon a monthly review of AC membership and

claims data, members who fell into one of the two eligibility groups were

automatically enrolled for ACCESS NY services.

Individuals eligible for this study were those enrolled from the start of

the study in July 2010 through the end of January 2012.

Data sources and study measures

We obtained study data from information AC collects on its members for

administrative purposes and reimbursing insurance claims. De-identified

data from membership files included date of active AC membership, date

of ACCESS NY enrollment, ACCESS NY eligibility status, date of patient

navigator assignment, date of first patient navigator-client contact, age at

ACCESS NY enrollment, NYC borough of residence, and gender. Date of

initial HIV diagnosis and race/ethnicity were incomplete and are presented

for descriptive purposes only. From insurance claims, we obtained dates of

HIV primary care visits. Because AC is a comprehensive health plan for

Medicaid eligible individuals, we are confident that the claims data repre-

sent the complete record of outpatient HIV medical care visits for ACCESS

NY clients. Claims data are complete through July 31, 2013.

The primary study outcome, connection-to-care, was the number of days

that elapsed between enrollment in ACCESS NY and the first post-

enrollment cHIV primary care visit. For study eligible individuals, who

died or resigned from AC before a first visit, we assigned the date of with-

drawal from ACCESS NY as a time of right censoring in the hazard ana-

lysis described below. For the small number of study eligible members,

6 P. MESSERI ET AL.

who remained out of care through the end of the study period, we assigned

a censoring time of July 31, 2013 to coincide with the end of the

study period.

A secondary outcome was the duration of retention-in-care. There is cur-

rently no gold standard for measuring retention-in-care. Instead, the cur-

rent recommendation is to adopt a measure that best matches study

context and data availability (Mugavero et al., 2012). For this study, a nat-

ural metric for retention-in-care is the interval between successive HIV pri-

mary care visits. A study participant was defined as retained in care as long

as the time between successive visits, post ACCESS NY enrollment, was

<190 days. Thus, the time retained in care was the number of days between

the date of the first post ACCESS NY HIV primary care visit and the date

of the medical visit prior to a first 190-day lapse between successive visits.

Based upon this definition, study participants who experienced a lapse in

care of 190 days or more after their first post ACCESS NY medical care

visit were assigned zero days retained in care. A very small number of par-

ticipants (n ¼ 41) experienced multiple lapses in care during the study

period; however, we restricted analysis to the first lapse in care. The dur-

ation of retention in care was right censored as of July 31, 2013.

Research design

The implementation of the ACCESS NY PN program conformed to a nat-

ural experiment in which client assignment to a patient navigator

“experimental group” followed a random-like process. Each month AC staff

generated a list of ACCESS NY clients with no claims for HIV primary care

visits during the previous six months. Starting from the top of this list, AC

staff assigned ACCESS NY clients to navigators until they filled all available

navigator slots. AC staff reported skipping down the list to “correct” for

imbalances in gender and borough of residence. ACCESS NY clients not

assigned a patient navigator formed a “nonintervention” comparison group.

Apart from the ad hoc balancing of patient navigator assignment by gender

and borough, AC staff reported that they used no additional information to

make assignments. Thus, within gender and borough, the monthly CHOW/

HN assignments appear to be “as good as” random.

The primary test of study hypotheses assumed an intention-to-treat

model. Assignment to the experimental or comparison group was based

upon whether or not an ACCESS NY client was assigned a patient naviga-

tor and not on actual contact between study participants and their assigned

patient navigator. We relaxed the intention-to-treat model and tested

whether actual contact between patient navigator and client, a nonrandom

process, strengthened the association beyond mere assignment.

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 7

We fit the t ime to first HIV primar y medical care visit to a propor-

tional hazard model (Kalbfleisch & Pr entice, 1980). Individuals eligibl e

for this analysis met ACCESS NY eligibility criteria and were enrolled in

ACCESS NY through January 31, 2012. We treated PN as a time var y-

ing c ovariate that turned “on” at the date when a study participant was

assigned to a navigator. We als o estimated a second statistical model in

which we distinguished between assignments that did and did not result

in patient contact. To better approximate random assignment, model

covariates included gender, borough of residence, age at ACCES S NY

enrollment, A CCESS NY eligibility status (dropped out of care or

unconnected to c are), and number of days enrolled in AC at time of

ACCESS NY enrollment. To adjust fo r possible effects of staff learning

and the addition of more patient navigat or slots as the study progressed,

we included quarterly fixed effects. Al l covariates conformed to the pro-

portional hazard assumption.

We next fit the length of time ACCESS NY clients were retained in med-

ical care to a proportional hazard model with similar specifications to the

first-visit model. Individuals eligible for this analysis met ACCESS NY eligi-

bility criteria and had at least one primary care visit between July 2010 and

December 31, 2012. Individuals entered the “risk set” on the day of their

first HIV medical care visit post ACCESS NY enrollment. The “failure”

time was the date of the first lapse of 190 days between HIV medical care

visits. For the retention-in-care model, we treated assignment to a patient

navigator as a time invariant variable. We classified individuals as either

assigned or not assigned a patient navigator during the study period.

Under ideal experimental conditions, the navigator assignment would have

preceded the first HIV medical care visit; however, some ACCESS NY cli-

ents were assigned a patient navigator after their first HIV medical care

visit. We tested the patient navigator effect on retention in care, for indi-

viduals assigned before their first primary care visit. Individuals assigned a

patient navigation after their first visit were grouped with individuals never

assigned a patient navigator. Based on the results of the proportional haz-

ard models, we constructed survival functions to estimate median days to

initial visit with and without a patient navigator. Statistical analysis was

performed using Stata 15.0.

This research was approved for human subjects by the Columbia

University Institutional Review Board.

Results

AC enrolled 1,173 of its HIV-positive members into ACCESS NY between July

2010 and January 2012. Following a three-month startup period, monthly

8 P. MESSERI ET AL.

enrollment of new ACCESS NY clients fluctuated between 39 and 110 and was

equal to or >60 in 12 of 16 months of active recruitment. AC staff assigned a

CHOW or a HN to 372 or 32% of ACCESS NY clients. During peak ACCESS

NY activity (April 2011 to July 2012), monthly CHOW/HN assignments fluctu-

ated between 8 and 32, with a median monthly assignment of 18.

After excluding AC members for reasons listed in Figure 1, 856 AC

members formed the study sample: 331 were unconnected-to-care members

new to the health plan and 525 were long-time members who had dropped

out care. ACCESS NY clients were predominantly African American or

Latino males with a mean age of 42. ACCESS NY clients resided in all five

boroughs. The largest concentration of study participants resided in

Brooklyn; substantial numbers reside in the Bronx and Manhattan

(Table 1). Consistent with the assumption that assignment was as good as

random, Table 1 shows a generally balanced distribution for most of the

study participant characteristics between ACCESS NY clients who were and

1,173

Study Exclusion Criteria

ACCESS NY clients enrolled

between July 1, 2010 and

January 31, 2012

947

Unconnected client assigned to PN

within first 30 days of PC

Younger than 18

864

859

Final Study Sample:

856

AC membership lasted < 90 days

post ACCESS NY Enrollment

-226

-83

-5

-3

An HIV primary care visit

occurred less than 180 days prior

to ACCESS NY Enrollment

Figure 1. Study sample determination.

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 9

were not assigned a CHOW/HN before the first medical care visit after

enrollment in the program. For only two comparisons did sample differen-

ces achieve statistical significance. There was an over representation of

Blacks among unconnected-to-care clients assigned a CHOW/PN. Length

of AC membership at time of ACCESS NY enrollment among drop outs

was shorter for those assigned a CHOW/HN.

Table 2 summarizes ACCESS NY patient navigator statistics. A CHOW

or HN was assigned to 286 study sample members, and made contact with

136 or 48% of the assigned clients. The median number of days between

initial client assignment and a successful contact was 15 days, and 90% of

successful contacts were completed within 65 days. The substantial number

of no-contact assignments might have diluted the CHOW/HN effect.

However, the effort of outreach workers to contact ACCESS NY clients

generated a large number of letters and phone calls that may have

prompted a return to care even in the absence of direct contact between

CHOW/HN’s and their prospective clients.

Insurance claims data documented that almost 90% of ACCESS NY cli-

ents (89%) had one or more HIV medical care visit during their enrollment

Table 1. Sample distribution of ACCESS NY clients enrolled between July 1, 2010 and January

31, 2012.

Assigned patient navigator before first medical care visit:

Unconnected to care Dropped out of care

Total No Yes No Yes

N 856 283 48 476 49

Gender (% Male) 70% 72% 81% 68% 66%

Age at PC Enrollment 42 42 41 43 41

Race

White 5% 7% 0%



5% 10%

Black 38% 12% 32% 49% 43%

Hispanic 17% 9% 11% 20% 21%

Other/ Unkn 40% 72% 57% 26% 26%

Education

<H.S. 21% 13% 14% 23% 25%

H.S. 53% 45% 50% 56% 45%

>H.S. 26% 42% 36% 21% 30%

Borough of Residence

Bronx 28% 32% 23% 26% 33%

Brooklyn 37% 34% 31% 40% 33%

Manhattan 21% 21% 40% 18% 24%

Queens 7% 6% 2% 8% 6%

Staten Island 7% 7% 4% 8% 4%

Month since Initial

HIV/AIDS diagnosis

at AC

enrollment (Mean)

149 126 116 160 138

Months AC member

at PC

enrollment

(months)

9.0 <1 <1 15.2 10.8



Significance tests applied to differences between individuals assigned and not assigned patient navigator separ-

ately for unconnected to care and dropped out of care ACCESS clients



p < .05.

10 P. MESSERI ET AL.

in ACCESS NY. Fifty percent of ACCESS NY clients had their first visit

within 99 days of enrollment and 75% within 197 days. A longer time

elapsed between ACCESS NY enrollment and first HIV medical care visit

for dropout enrollees (median time to first visit ¼ 107 days) than uncon-

nected-to-care enrollees (median time to first visit ¼ 86 days).

Connection to care

The top half of Table 3 presents the results of the connection-to-care haz-

ard analysis. The coefficients in Table 3 are the hazard ratios (HRs) for

assignment to a CHOW/HN adjusted for the joint influence of all inde-

pendent variables in the model. HRs measure the relative change in the

“rate” of a first HIV medical care visit following ACCESS NY enrollment.

An HR greater than one indicates an increased rate of first visit, and there-

fore a shorter expected time to first visit. The left-most column presents

HRs estimated for all eligible clients. The middle and right-most columns

present HRs for the unconnected-to-care and dropped-out-of-care clients.

Consistent with the first study hypothesis, the left-most column in

Table 3 indicates a 37% increase in the rate of first HIV medical care visit

following CHOW/HN assignment for the entire study sample. CHOW/HN

assignment shortened the median duration to a first visit by about 22% or

22 days. We further explored whether the CHOW/HN effect was associated

with actual contact with a CHOW/HN. Actual contact between a CHOW/

HN and a client was associated with an 89% increase in the daily rate of

first visit. This is equivalent to a 34% reduction in median time to first visit

from 98 days, in the absence of a CHOW/HN assignment, to 65 days

Table 2. ACCESS NY patient navigator statistics.

Total Unconnected to Care Dropped Out of Care

Total Number of Clients 856 331 525

% (N) of Clients Assigned

Assigned a CHOW/HN 33% (286) 31% (104) 35% (182)

Assigned a CHOW/HN before First HIV

Medical Care Visit

11% (97) 14% (48) 9% (49)

% (N) of Assignments Resulting in

Client Contact

48% (136) 42% (44) 51% (92)

CHOW/HN Assignments before First HIV

Medical Care

Visit Resulting in Client Contact

45% (44) 40% (19) 51% (25)

Days Between All Assignments and

Successful Contact

Median 15 9.5 19



90th Percentile 65 36 66

Days Between Assignments before

HIV Medical Care Visit and

Successful Contact

Median 16.5 9 40



90th Percentile 64.0 34 66

Test for significant difference between unconnected and dropped out of care



p < .05;



p < .01.

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 11

following client contact with a CHOW/HN. By contrast, absent navigator-

client contact, CHOW/HN assignment had no statistically significant effect

on decreasing the time to HIV medical care visit. CHOW/HN assignment

shortened the expected time to first visit for dropouts (p < .05), and with

less confidence for the unconnected-to-care (p < .1). However, contact with

a patient navigator had a significant effect on shortening time to first visit

for both unconnected and drop-out clients.

Retention in care

We identified 673 ACCESS NY clients eligible for the retention-in-care

analysis. These individuals had their first post ACCESS NY HIV enrollment

medical care visit before the end of 2012 and remained insurance plan

members for at least six months following their first post ACCESS NY pri-

mary care visit. Among those eligible for the retention–in-care analysis, 286

or 42% had a lapse in care of 190 days or longer. Through the end of the

study period, July 30, 2013, 25% of eligible clients remained in care for

113 days (3.7 months) or less, whereas 50% of eligible clients remained in

care for at least 475 days (15.8 months). Similar proportions of drop-out

and unconnected-to-care ACCESS NY clients experienced one or more

intervals of 190 days or longer between medical visits.

Results of the retention-in-care analysis suggest that the patient navigator

effect on engagement in care did not persist after the first visit. The pro-

portion of individuals experiencing at least one lapse in care longer than

190 days was similar for ACCESS NY clients who were (45%) and were

Table 3. Adjusted hazard ratio of effect of assignment and contact with patient navigator on

connection and retention in care.

Last date eligible

for analysis All eligible clients

Unconnected to

care clients

Dropped out of

care clients

Connection to

Care:

Assigned PN

ACCESS NY

Enrollment through

January 31, 2012

1.37 (1.11,1.69) 1.38 (0.95,1.98) 1.32 (1.02,1.71)

Assigned but

no contact

0.99 (0.73,1.32) 1.04 (0.63,1.72) 0.88 (0.60,1.27)

Assigned Contact 1.89 (1.46,2.44)

Participants = 826

PCP Visits = 733

1.96 (1.20,3.20)

Participants = 323

PCP Visits = 288

1.91 (1.40,2.61)

Participants = 503

PCP Visits = 445

Retention in

Care: Assigned

First HIV post

ACCESS NY Primary

Care Visit before

January 1, 2013

1.14 (0.83,1.57) 1.17 (0.65,2.08) 1.25 (0.84,1.87)

Assigned but

no contact

0.97 (0.57,1.66) 0.89 (0.35,2.23) 1.10 (0.56,2.15)

Assigned Contact. 1.23 (0.85,1.79)

Participants = 646

Lapses > 189 days

= 275

1.41 (0.70,2.86)

Participants = 258

Lapses > 189 days

= 110

1.33 (0.56,2.15)

Participants = 503

Lapses > 189 days

= 445

The hazard ratios are adjusted for gender, age, borough of residence, length of AC membership at time of

ACCESS NY enrollment and quarter of enrollment.

12 P. MESSERI ET AL.

not (42%) assigned a CHOW/HN prior to a first medical care visit.

Among ACCESS NY clients assigned CHOW/HN before their first primary

care visit, the median time retained in care was 377 days compared to

497 days for clients not assigned a CHOW/HN.

The bottom half of Table 2 presents the statistical test of the retention-

in-care hypothesis. None of the PN assignment coefficients for the

retention-in-care models are statistically significant. Neither CHOW/HN

assignment nor navigator-client contact had an effect on retention in care

for either the unconnected-to-care or dropped-out-care clients.

Discussion

To our knowledge, ACCESS NY is the first Medicaid health plan to offer

its HIV-positive members patient navigation services. Previous studies of

HIV/AIDS patient navigation have either been situated in medical care

organizations, community based organizations, or prison reentry programs

(see Table 4). Study findings indicated a short-term effect of PN, but one

that probably did not persist, once the patient navigator-client relationship

ended. Consistent with the primary study hypothesis, patient assignment to

a CHOW/HN substantially reduced the time to initial primary care visit.

More speculatively, it appeared that the CHOW/HN effect was restricted to

instances in which the CHOW/HN was able to make contact with assigned

clients. Contrary to the second study hypothesis, CHOW/HN assignment

was not associated with the length of retention in care following ini-

tial visit.

Study results may be interpreted with respect to the mixed results of the

four published RCTs of HIV patient navigation. Similar to the results of

this study, ARTAS (Gardner et al., 2005) and Myers et al. (2018) found

that patient navigation increased early linkage to care. The absence of a PN

effect on retention in care for the Myers et al. (2018) study departs from

the positive results of ARTAS (Gardner et al. 2005) but is consistent with

the null findings of LA LINK (Cunningham et al., 2018 ). Patient navigation

was associated with rates of viral suppression in LA LINK (Cunningham

et al., 2018) but not Project Hope (Metsch et al., 2016) and Myers et al.

(2018). An unpublished report of ACCESS NY (Messeri, Yomogida, &

Vardy, 2014, available on request from the lead author), also found no

effect of PN assignment on viral suppression. However, the primary end-

point of ACCESS NY was to connect clients to care, which is several steps

removed from the features of medical care that are most directly related to

viral suppression. The mixed results may be related to the different HIV-

infected populations and settings of these studies. ARTAS focused on a

recently diagnosed population (Gardner et al., 2005), Project Hope focused

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 13

Table 4. Key features of ACCESS NY and characteristics of selected HIV patient navigation programs.

Client target population

Newly diagnosis Unconnected to care Dropped out care

Risk of falling out

care not fully engaged

Recently released

from jail/prison Youth and young adults

Gardner 2005

Leider et al. 2011

Craw 2008

Harris et al. 2003

Cabral 2018

ACCESS NY

Molitor 2005

ACCESS NY

Chin 2006

Kral 2018

Cabral 2018

Bradford 2007

Andersen 2007

Chin 2006

Rich 2001

Cunningham 2018

Kral 2018

Myers 2018

Harris et al. 2003

Patient navigator qualifications

Peer Professional

Incorporate into

other staff Position

ACCESS NY

Bradford 2007

Coleman 2007

Leider et al. 2011

Cabral 2018

Cunningham 2018

Myers 2018

ACCESS NY

Gardner 2005

Chin 2006

Craw 2008

Kral 2018

Andersen 2007

Kral 2018

Duration

time limited Long or

indeterminate

ACCESS NY

Gardner 2005

Leider et al. 2011

Craw 2008

Cabral 2018

Cunningham 2018

Myers 2018

Bradford 2007

Coleman 2007

Andersen 2007

Rich 2001

Kral 2018

Organizational setting

Medical care clinic Social service agency ER Managed care plan Jail re-entry

Gardner 2005

Bradford 2007

Cabral 2018

Kral 2018

Bradford 2007

Coleman 2007

Leider et al. 2011 ACCESS NY Cunningham 2018

Myers 2018

Kral 2018

PN tasks

How to navigate health

and social systems

Accompany clients

to appointments

Develop skills in provider

interactions

Provide social

support

Address different

barriers to care Risk reduction counseling

ACCESS NY

Bradford 2007

Myers 2018

Cabral 2018

Cunningham 2018

Kral 2018

ACCESS NY

Bradford 2007

Cabral 2018

Cunningham 2018

Myers 2018

Kral 2018

Bradford 2007

Craw 2008

Cabral 2018

Cunningham 2018

ACCESS NY

Bradford 2007

Craw 2008

Cabral 2018

Cunningham 2018

Myers 2018

ACCESS NY

Bradford 2007

Craw 2008

Cabral 2018

Cunningham 2018

Kral 2018

Myers 2018

Harris et al. 2003

Cabral 2018

Myers 2018

14 P. MESSERI ET AL.

on hospitalized substance using patients (Metsch et al., 2016); whereas, LA

LINK (Cunningham et al., 2018) and Myers et al. (2018) recruited study

samples from HIV-infected individuals reentering the community following

time in jail.

A strength of this study is that AC staff assigned CHOW /HN blind to

information about potential need or the valu e of the service to individ-

ual members, thus creating a “randomized control” group formed from

clients not assigned a CHOW/HN. However, this is an assumption, as

AC staff did not follow an explicit random assignment protocol. There

is a sli ght possibility that the client list used in assig ning CHOW/H N ’s

happened to be ordered, albeit unintentionally, in ways that resulted in

an unknown selection bias in assign ment of patient navigators that cor-

related with study outcomes. C ontrary to study p rotocol, some AC

members visited a primary care provider within six months of assign-

ment. However, the assignment process ma y still be treated as if it were

random, since AC staff were un aware of such medical visit at time of

assignment, most likely because of delays between a visit and submission

of a claim for AC reimbursement.

A further strength of this study is its strong external validity. The study

sample received care from a larger number of medical providers working in

the broad array of medical care settings that serve NYC Medicaid patients.

Study results may partly be a consequence of how we measured study

outcomes. We chose, for example, a half-year period to identify lapses in

medical care, but retention-in-care has been operationally measured in dif-

ferent ways, for example a minimum of two visits in a six-month period.

Study design limitations and the specific features of the PN model imple-

mented in this study may account for the mixed results. The effects of the

CHOW/HN may have been diminished because we included the early

stages of the program, when AC staff were still learning how best to deploy

CHOW/HNs. Furthermore, termination of client contact with initial pri-

mary care visit may have attenuated the ACCESS NY CHOW/HN’s effect.

A more extended or intensive contact between navigator and client may

have resulted in the development of stronger patient self-management skills

to keep scheduled follow-up visits. However, we did not collect systematic

data on the actual frequency of PN client contact beyond the dates of

patient navigation assignment and termination. Therefore, we could not

directly assess the extent that study outcomes might be related to the inten-

sity of the client-patient navigator relationship. Moreover, we did not find

any published study that linked a variable measure of the intensity of

patient navigation to client outcomes. It remains for future patient naviga-

tion research to incorporate measurement of the intensity of patient navi-

gation-client interactions and its impact on engagement in medical care.

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 15

A further limitation of this study was the absence of measurement of the

similarities and differences in the cultural backgrounds of navigators and

their clients. Although client culture –the beliefs, attitudes and everyday

experiences that shape clients’ health care seeking behavior–was important

in both ACESS NY PN training and assignment of navigators to their cli-

ents, we did not directly measure either the cultural congruence between

navigator and client nor the navigator’s competence to work with culturally

diverse clients. It remains for future research to measure systematically the

cultural dimension of the navigator-client relationship in order to assess

whether navigators can be trained to work effectively with culturally diverse

clients or does optimal patient engagement in care require matching navi-

gators and their clients for cultural congruence.

Conclusion

As noted at the outset of this paper, the research literature applies the

patient navigation label to a range of patient-centered interventions that

vary with respect to (1) client target population, (2) patient navigator quali-

fications, (3) duration, (4) organizational setting, and (5) the tasks naviga-

tors perform. Table 4 illustrates how ACCESS NY and other PN programs

reviewed in this study are located within such a conceptual framework. The

table reveals the considerable diversity of PN programs. Thus, it is not pos-

sible to regard ACCESS NY or any other PN program as representative of

the broad variation in key features of HIV/AIDS patient navigation

interventions.

Since standardization of what navigators do is neither likely nor neces-

sarily desirable, future research should adopt a multi-method strategy that

links a careful delineation of the salient features of a PN intervention to its

impact on client outcomes. A qualitative study component would docu-

ment how a particular PN program maps onto the five dimensions enum-

erated above. The qualitative information would aid in contextualizing

quantitative outcome studies.

Results of this study demonstrate that a PN program based in a

Medicaid health plan can improve access to HIV medical care among dis-

advantaged populations. However, it raises questions for future research

about the continuing need for PN assistance, to retain patents in care once

they are inside the clinic door. This study also recognizes that there is no

unified PN model. PN interventions come in various forms that may con-

dition PN’s impact on patient engagement in medical care. We proposed a

conceptual framework for comparing key similarities and differences in PN

programs that could guide informative generalizations about the

16 P. MESSERI ET AL.

contribution of PN to patient care for HIV/AIDS and other chronic

health conditions.

Acknowledgements

AIDS United through its Positive Charge Initiative supported the research for his study

through a grant to Amida Care with a subcontract for local evaluation to Columbia

University. The findings and conclusions in this paper are those of the authors’ and do not

necessarily reflect the views of AIDS United.

References

AIDS United. (2015). Best practices for integrating peer navigators into HIV models of care.

Unpublished Report. Washington DC.

Andersen, M., Hockman, E., Smereck, G., Tinsley, J., Milfort, D., Wilcox, R., …

Thomas, R. (2007). Retaining women in HIV medical care. Journal of the Association of

Nurses in Aids Care, 18(3), 33–41. doi:10.1016/j.jana.2007.03.007

Bradford, J. B., Coleman, S., & Cunningham, W. (2007). HIV system navigation: An

emerging model to improve HIV care access. AIDS Patient Care and Stds, 21(Suppl 1),

S49– S58. doi:10.1089/apc.2007.9987

Cabral, H. J., Davis-Plourde, K., Sarango, M., Fox, J., Palmisano, J., & Rajabin, S. (2018).

Peer support and the HIV continuum of care: Results from a multisite randomized clin-

ical trial in three urban clinics in the United States. IDS and Behavaior, 22(8),

2627–1539. doi:10.1007/s10461-017-1999-8

Chin, J. J., Kang, E., Kim, J. H., Martinez, J., & Eckholdt, H. (2006). Serving asians and

pacific islanders with HIV/AIDS: Challenges and lessons learned. Journal of Health Care

for the Poor and Underserved, 17(4), 910–927. doi:10.1353/hpu.2006.0119

Coleman, S. M., Boehmer, U., Kanaya, F., Grasso, C., Tan, J., & Bradford, J. B. (2007).

Retention challenges for a Community-Based HIV primary care clinic and implications

for intervention. AIDS Patient Care and Stds, 21(9), 691–701. doi:10.1089/apc.2006.0205

Craw, J. A., Gardner, L. I., Marks, G., Rapp, R. C., Bosshart, J., Duffus, W. A., … Schmitt,

K. (2008). Brief strengths-based case management promotes entry into HIV medical

care: Results of the antiretroviral treatment access study-II. Jaids Journal of Acquired

Immune Deficiency Syndromes, 47(5), 597–606. doi:10.1097/QAI.0b013e3181684c51

Cunningham, W. E., Weiss, R., Nakazono, T., Malek, M. A., Shoptaw, S. J., Ettner, S. L., &

Harawa, N. T. (2018). Effectiveness of a peer navigation intervention to sustain viral sup-

pression among HIV-positive men and transgender women released from jail: The LINK

LA randomized clinical trial. JAMA Internal Medicine, 178(4), 542–553. doi:10.1001/

jamainternmed.2018.0150

Gardner, L. I., Metsch, L. R., Anderson-Mahoney, P., Loughlin, A. M., Rio, C. D.,

Strathdee, S., … Holmberg, S. D. (2005). Efficacy of a brief case management interven-

tion to link recently diagnosed HIV-infected persons to care. AIDS, 19(4), 423–431. doi:

10.1097/01.aids.0000161772.51900.eb

Harris, S. K., Samples, C. L., Keenan, P. M., Fox, D. J., Melchiono, M. W., Woods, E. R., &

Boston, H. P. (2003). Outreach, mental health, and case management services: Can they

help to retain HIV-positive and at-risk youth and young adults in care? Maternal and

Child Health Journal, 7(4), 205–218.

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 17

Higa, D. H., Marks, G., Crepaz, N., Liau, A., & Lyles, C. M. (2012). Interventions to

improve retention in HIV primary care: A systematic review of U.S. studies. Current

HIV/AIDS Reports, 9(4), 313–325. doi:10.1007/s11904-012-0136-6

Jain, K., Holtgrave, D. R., Maulsby, C., Kim, J., Zulliger, R., Massey, M., & Charles, V.

(2015). Improving Access to HIV Care. Baltimore, MD: Johns Hopkins University Press.

Kalbfleisch, J. D., & Prentice, R. L. (1980). The statistical analysis of failure time data. New

York, NY: Wiley

Kral, A. H., Lambdin, B. H., Comfort, M., Powers, C., Cheng, H., Lopez, A. M., …

Lorvick, J. (2018). A strength-based case management intervention to reduce HIV viral

load among people who use drugs. IDS and Behavaior, 22(1), 146–153. doi:10.1007/

s10461-017-19036

Leider, J., Fettig, J., & Calderon, Y. (2011). Engaging HIV-positive individuals in specialized

care from an urban emergency department. AIDS Patient Care and Stds, 25(2), 89–93.

doi:10.1089/apc.2010.0205

Liau, A., Crepaz, N., Lyles, C. M., Higa, D. H., Mullins, M. M., DeLuca, J., … Marks, G.

(2013). Interventions to promote linkage to and utilization of HIV medical care among

HIV-diagnosed persons: A qualitative systematic review, 1996-2011. IDS and Behavaior,

17(6), 1941–1962. doi:10.1007/s10461-013-0435-y

Littell, J. H., & Girvin, H. (2002). Stages of change. A critique. Behavior Modification,

26(2), 223–273. doi:10.1177/0145445502026002006

Messeri, P., Yomogida, M., & Vardy, Y. (2014). ACCESS NY Local Evaluation. Unpublished

report prepared for Amida Care.

Metsch, L. R., Feaster, D. J., Gooden, L., Matheson, T., Stitzer, M., Das, M., … del Rio, C.

(2016). Effect of patient navigation with or without financial incentives on viral suppres-

sion among hospitalized patients with HIV infection and substance use. JAMA, 316(2),

156–170. doi:10.1001/jama2016.8914

Miller, W. R., & Rollnick, S. (2002). Motivational interviewing: Preparing people for change,

2nd ed. New York, NY: Guilford Press.

Mizuno, Y., Hig, D. H., Leighton, C. A., Roland, K. B., Delica, J. B., & Koenig, L. J. (2018).

Is HIV patient navigation associated with HIV care continuum outcomes? AIDS, 32,

2557–2571. doi:10.1097/QAD.0000000000001987

Molitor, F., Kuenneth, C., Waltermeyer, J., Mendoza, M., Aguirre, A., Brockmann, K., &

Crump, C. (2005). Linking HIV-infected persons of color and injection drug users to

HIV medical and other services: The California Bridge Project. AIDS Patient Care and

Stds, 19(6), 406–412. doi:10.1089/apc.2005.19.406

Mugavero, M. J., Westfall, A. O., Zinski, A., Davila, J., Drainoni, M.-L., Gardner, L. I., …

Giordano, T. P. (2012). Measuring retention in HIV care: The elusive gold standard.

JAIDS Journal of Acquired Immune Deficiency Syndromes, 61(5), 574–580. doi:10.1097/

QAI.0b013e318273762f

Myers, J. J., Kang Dufour, M.-S., Koester, K. A., Morewitz, M., Packard, R., Monico Klein,

K., … Tulsky, J. (2018). The effect of Patient Navigation on the likelihood of

Engagement in clinical Care for HIV-Infected Individuals Leaving Jail. American Journal

of Public Health, 108(3), 385–392. doi:10.2105/AJPH2017.304250

Naar-King, S., Bradford, J., Coleman, S., Green-Jones, M., Cabral, H., & Tobias, C. (2007).

Retention in care of persons newly diagnosed with HIV: Outcomes of the outreach ini-

tiative. AIDS Patient Care and Stds, 21(s1), S-40–S-48. doi:10.1089/apc.2007.9988

Rich, J. D., Holmes, L., Salas, C., Macalino, G., Davis, D., Ryczek, J., & Flanigan, T. (2001).

Successful linkage of medical care and community services for HIV-positive offenders

18 P. MESSERI ET AL.

being released from prison. Journal of Urban Health: Bulletin of the New York Academy

of Medicine, 78(2), 279–289. doi:10.1093/jurban/78.2.279

Saleebey, D. (1997). The strengths perspective in social work practice, 2nd ed. New York,

NY: Longman Publishers. doi:10.1093/sw/40.2.280

Simoni, J. M., Amico, K. R., Smith, L., & Nelson, K. (2010). Antiretrivoral adherence inter-

ventions: Translating research findings to the real world clinic. Current Hiv/Aids Reports,

7(1), 44–51. doi:10.1007/s11904-009-0027-5

Thompson, M. A., Aberg, J. A., Hoy, J. F., Telenti, A., Benson, C., Cahn, P., …

Volberding, P. A. (2012). Antiretroviral treatment of adult HIV infection: 2012 recom-

mendations of the International Antiviral Society-USA panel. JAMA, 308(4), 387–402.

doi:10.1001/jama.2012.7961

Wohl, A. R., Garland, W. H., Wu, J., Au, C.-W., Boger, A., Dierst-Davies, R., … Jordan,

W. (2011). A Youth-Focused case management intervention to engage and retain young

gay men of color in HIV care. AIDS Care, 23(8), 988–997. doi:10.1080/09540121.2010.

542125

Zaller, N. D., Holmes, L., Dyl, A. C., Mitty, J. A., Beckwith, C. G., Flanigan, T. P., & Rich,

J. D. (2008). Linkage to treatment and supportive services among HIV-Positive Ex-

Offenders in project bridge. Journal of Health Care for the Poor and Underserved, 19(2),

522–531.

JOURNAL OF HIV/AIDS & SOCIAL SERVICES 19