Integration of Patient Reported Outcomes Data among People with HIV in Washington, DC: A sub-cohort of the DC Cohort Longitudinal Study

Abstract

Aim:

Patient reported outcomes (PROs) can help to evaluate gaps and areas for improvement along the HIV care continuum. We sought to describe the methodology and processes of a PROs study within the DC Cohort study population, describe the PROs results to date, report on lessons learned, and describe future directions of the research.

Subject and Methods:

Each study site recruited participants from the DC Cohort, a longitudinal study on people with HIV, to complete the electronic PROs baseline and annual follow-up surveys, which consisted of previously validated measures of social determinants of health, mental health, substance use, medication adherence, and other related measures. The recruitment, enrollment, and data linkage process strengths and limitations were described. Participants’ PROs survey responses were linked to their DC Cohort data to better understand participant’s health overall.

Results:

There have been 1,739 baseline, 610 first annual, and 142 second annual completed surveys between May 2021 and May 2024. Among PROs participants, the most common reported unmet need was food insecurity (33.5%) and the most common mental health condition was generalized anxiety disorder (39.5%). A majority (59.7%) of participants reported being sexually active, but at least half did not use barrier protection.

Conclusion:

PROs surveys complement electronic health record data collected in the DC Cohort study, allowing for examination of health outcomes and factors collected in a longitudinal manner. PROs results could be integrated into near-real time dashboards for use by clinicians and incorporated into routine clinical care to provide a more holistic understanding of the patient.

Background

Patient reported outcomes (PROs) can assess cognitive and physical function, health behaviors, life stressors, stigmatization, adherence to antiretroviral therapy (ART), quality of life (QOL), and many more factors associated with HIV and related comorbidities (Deshpande et al. 2011; Duracinsky et al. 2012; Simpelaere et al. 2016; Porter et al. 2016; Bristowe et al. 2020; Antela et al. 2022). It is important to account for PROs when conducting research involving people with HIV (PWH) as these outcomes can provide valuable insight into aspects of a patient’s life that may not be apparent during routine clinic visits. Capturing PROs longitudinally allows us to gain better insight into how these outcomes may impact other longitudinal measures, such as viral suppression or CD4 counts. However, measuring PROs longitudinally on a large scale comes with obstacles that must be overcome before such data can be effectively used in research studies.

The DC Cohort Study, a longitudinal observational cohort study of PWH receiving care in Washington, D.C., offers an ideal setting to link clinical outcomes obtained through the observational cohort data to individual-level PROs data (Greenberg et al. 2016). The DC Cohort PROs survey includes measures such as substance use, sexual risk behaviors, mental health, QOL, social determinants of health (SDOH), and adherence to ART that can be easily matched to data from participant electronic health records (EHRs).

PROs measures have been implemented in a variety of research settings (Kozak et al. 2012; Antinori et al. 2020; Schulte et al. 2020; Carlsson-Lalloo et al. 2022). However, none have provided a detailed framework for implementing longitudinal PROs measures in a longitudinal cohort of PWH, nor have any provided contemporary PROs data, such as data collected during the COVID-19 pandemic era. Given the potential challenges of implementing a longitudinal PROs survey, we sought to provide a framework for the DC Cohort PROs survey and describe the distribution of PROs from 2021–2024.

Our first objective is to describe the methodology and processes of conducting the PROs study. These processes include participant eligibility, recruitment, enrollment, consent, compensation, linking survey data to the DC Cohort database, and tracking participant follow-up surveys. Additionally, we aim to describe the distribution of PROs results among baseline and annual surveys. Finally, we aim to discuss lessons learned and potential future implementation of PROs research within the DC Cohort study.

Methods

Study Design

The DC Cohort Study is a longitudinal observational cohort study among 14 Washington, DC community, academic, and hospital-based clinics providing HIV care. The cohort began enrollment in 2011 and currently has over 13,000 PWH enrolled. After participant consent or equivalent, patient-level data are collected from EHRs on variables such as socio-demographics, clinic visits, laboratory values, and prescribing information and are imported into a centralized database. Methods of the cohort have been described in detail previously (Greenberg et al. 2016). The PROs substudy began enrollment in May 2021 and requires separate participant consent.

Patient Reported Outcomes of Interest

As of this publication, the PROs survey has nearly 2,000 participants enrolled and consists of a series of previously validated measures of SDOH, mental health, QOL, substance use, adherence to ART, sexual risk behaviors, symptom burden, and internalized stigma. Specifically, SDOH are measured using the Accountable Health Communities (AHC) screening tool (Billioux et al. 2017). The 8-item Public Health Questionnaire (PHQ-8) depression scale was used to measure depressive symptoms (Kroenke et al. 2009). Generalized anxiety disorder (GAD) was measured based on the 7-item GAD scale (GAD-7) (Spitzer et al. 2006). Post-Traumatic Stress Disorder (PTSD) was measured using the primary care PTSD screen (PC-PTSD) (Prins et al. 2003).

Substance use for cannabis, cocaine, amphetamines, inhalants, sedatives, hallucinogens, and opioids was measured using the Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) (2010). The Alcohol Use Disorders Identification Test (AUDIT) was used as the main measure of alcohol consumption (Saunders et al. 1993). QOL was measured using the EuroQol instrument, which measures anxiety/depression, mobility, pain/discomfort, self-care, and completion of usual activities (EuroQol Group 1990).

Adherence to ART was measured using items from the AIDS Clinical Trials Group Adherence Questionnaire (ACTG-AQ) and sexual risk behavior was measured using items from the Sexual Risk Behavior Inventory (SRBI) (Chesney et al. 2000; Fredericksen et al. 2018). Symptom burden was measured using the Memorial Symptom Assessment Scale – Short Form (MSAS - SF) (Sharp et al. 2018). Finally, internalized HIV stigma was measured using the HIV Stigma Framework (Earnshaw et al. 2013).

We used data from the DC Cohort database to categorize basic demographic variables such as race/ethnicity, gender, age, HIV transmission risk factor, time since HIV diagnosis, comorbidities, HIV-related labs, and HIV care encounters per year.

Eligibility and Recruitment

A patient was considered eligible for the PROs substudy if they were 1) a person with HIV receiving HIV care at a DC Cohort clinic site and was actively enrolled in the DC Cohort Study, 2) at least 18 years of age or an adolescent seeking care independently, and 3) able to understand and complete the PROs survey in English or Spanish. A list of active DC Cohort participants was available to each site Research Assistant (RA) to assist with eligibility determination.

The RAs referred to an electronic PROs Consent Log in the password-protected and HIPAA-compliant web-based Research Electronic Data Capture (REDCap), hosted by Children’s National in partnership with George Washington University, prior to approaching a potential participant to determine if they had already been approached and the subsequent outcome of that approach (Harris et al. 2008, 2019). RAs could only access information about participants from their site within REDCap. The PROs Consent Log captured if the participant consented, declined, or was undecided as to whether to participate. Each eligible participant could be approached a maximum of four times if undecided.

Enrollment and Consent

Once an eligible participant indicated interest in the study, the RA provided them with the REDCap PROs survey link and a unique PROs survey ID. The PROs Informed Consent Form (ICF) was embedded in the first page of the survey. The ICF and the survey could be completed by the participant in clinic or remotely.

Linking Surveys

The PROs survey required the participant’s unique PROs ID provided to them by the RA at their site. When a new completed survey with a valid ID appeared in REDCap, the RA added a form in the custom-built Multi-modal Integrated Data Acquisition System (MIDAS) that linked the participant’s unique PROs ID to their DC Cohort ID. This MIDAS form allowed their survey data to be linked to their DC Cohort data.

Invalid Survey IDs

If the PROs ID on a completed survey was invalid due to participant error or a keying error, the DC Cohort Study team collaborated with the RA to determine the correct PROs ID based on the invalid PROs ID, date and time of completion, and demographics from the survey. The survey would either be corrected with the confirmed ID of the participant that completed it, or the survey was not used in the dataset due to inability to link it to a participant.

Compensation

Participants were provided a Visa gift card as a survey incentive, which was sent by mail or given to the participant at their next clinic visit depending on participant preference. The incentive amount increased as the length of the survey increased over time due to the addition of questions, ranging from $15 to $50 with an approximate $1 per minute incentive.

Annual Surveys

Participants were asked to complete the PROs survey annually. RAs accessed a list of participants eligible for their annual survey from the DC Cohort intranet website. The RAs provided participants with their original PROs ID when approaching them for their annual survey. Participants received the corresponding incentive upon completion of annual surveys. Participants may not have annual follow-up surveys in our dataset if they had transferred care sites, withdrew from the study, had not yet reached 335 days since their previous survey completion, or had not responded to outreach from their clinic RA.

Mismatch Demographics

Researchers discovered some inconsistent demographics including discordant sex at birth and discrepant ages reported between baseline, annual surveys, and cohort data with the same PROs ID during data review. DC Cohort study personnel worked with RAs to correct discrepancies as needed. Any unresolved discrepancies resulted in the removal of that survey from the dataset.

Statistical Analyses

We compared participants enrolled in the PROs substudy to active DC Cohort participants who were eligible for PROs but did not consent, were undecided about consenting, or had not yet been approached to participate. Demographics were measured as of the date of the completion of the first PROs survey or May 01, 2024 for non-PROs participants. Chi-Square tests and Wilcoxon rank sum tests were used to calculate p-values for categorical and continuous variables, respectively. Then, we described the distribution of PROs responses on each participant’s first, second, and third annual survey, as applicable. Surveys were considered the next annual survey if they were taken at least 335 days after their last completed survey. All analyses were conducted using SAS 9.4 (Cary, NC) and p-values < 0.05 were considered statistically significant.

Results

Description of PROs Participants

From May 2021 to May 1, 2024, there were 1,739 PROs surveys completed that could be linked to a DC Cohort ID, representing about 22% of 7,805 total eligible participants completing a survey (Table 1). Of the total PROs participants, 610 completed a first annual survey, and 142 completed a second annual survey. In addition, 34 additional surveys were completed that could not be linked back to a DC Cohort ID due to an invalid survey ID entered. Most PROs participants were non-Hispanic Black (1,334; 76.7%), male (1,246; 71.7%), men who have sex with men (MSM) (789; 45.4%), and the mean age at time of first PROs survey was 52 years (standard deviation: 13.1). When compared to the rest of the DC Cohort participants who were eligible for a PROs survey between May 2021 - May 2024, but were not enrolled, PROs participants were younger (54 vs. 56 years of age; p < 0.0001), more likely to be male (71.7% vs. 69.9%; p = 0.0083), have been diagnosed with HIV in the past 5 years (8.5% vs. 4.9%; p < 0.0001), have a CD4 count greater than 500 cells/μL (70.0% vs. 68.3%; p < 0.0001), receive care at an academic hospital (47.5% vs. 33.5%; p < 0.0001), and have more HIV care encounters per year (9 vs. 5; p < 0.0001). PROs participants were also less likely to have been previously diagnosed with AIDS compared to the overall DC Cohort participants (31.6% vs. 40.2%; p < 0.0001) (Table 1).

Table 1.

Comparison of characteristics between PROs participants and overall DC Cohort participants not in PROs substudy

Demographic	PROs Participants N = 1739 N (%)	Non-PROs Participantsa N = 6066 N (%)	p-value
Race/Ethnicity			0.8067
Non-Hispanic Black	1334(76.71)	4627(76.28)
Non-Hispanic White	207(11.9)	729(12.02)
Hispanic	116(6.67)	448(7.39)
Other	24(1.38)	82(1.35)
Unknown	58(3.34)	180(2.97)
Gender			0.0083
Male	1246(71.65)	4241(69.91)
Female	439(25.24)	1680(27.7)
Transgender: male-to-female	40(2.3)	111(1.83)
Transgender: female-to-male	2(0.12)	18(0.3)
Unknown	12(0.69)	16(0.26)
Age [years](Median (IQR))	54 (42, 62)	56 (44, 64)	< 0.0001
HIV Transmission Factor			0.1419
Men who have sex with Men	789(45.37)	2618(43.16)
High Risk Heterosexual (HRH)	508(29.21)	1985(32.72)
Intravenous Drug Use (IDU)	88(5.06)	295(4.86)
Perinatal	38(2.19)	111(1.83)
Other	299(17.19)	1005(16.57)
Unknown	17(0.98)	52(0.86)
Time Since HIV Diagnosis			< 0.0001
0–5 years	147(8.45)	295(4.86)
6–10 years	225(12.94)	709(11.69)
11+ years	1344(77.29)	4985(82.18)
Unknown	23(1.32)	77(1.27)
CD4 [cells/uL] b			< 0.0001
< 200	79(4.54)	380(6.26)
[200,500)	399(22.94)	1518(25.02)
500+	1217(69.98)	4141(68.27)
Unknown	44(2.53)	27(0.45)
Viral Suppression c			< 0.0001
Virally Suppressed	1418(81.54)	5125(84.49)
Not Virally Suppressed	134(7.71)	670(11.05)
Unknown	187(10.75)	271(4.47)
Nadir CD4 [cells/uL](Mean [St.Dev.]	376.1(552.94)	360.43(441.14)	0.3473
AIDS			< 0.0001
AIDS Dx	550(31.63)	2440(40.22)
No AIDS Dx	1179(67.8)	3581(59.03)
Unknown	10(0.58)	45(0.74)
Type of HIV Care Site			< 0.0001
Community	913(52.5)	4029(66.42)
Hospital	826(47.5)	2033(33.51)
Private	0 (0.0)	4(0.07)
HIV Care Encounters per Year (Median (IQR))	9 (2, 14)	5 (2, 9)	< 0.0001

^aParticipants who were still enrolled as of May 01, 2021 and had at least one HIV lab or HIV care encounter from May 01, 2021 to May 01, 2024

^bBased on most recent CD4 lab prior to date of PROs survey completion or May 01, 2024 for non-PROs participants

^cBased on the most recent viral load lab prior to date of PROs survey completion or May 01, 2024 for non-PROs participants, participants were considered virally suppressed if their most recent HIV RNA lab was ≤ 200 copies/mL

Comparison to PROs Refusals

From May 2021 to May 2024, 591 DC Cohort participants have been approached but refused to participate in the substudy, resulting in an estimated refusal rate of 25%. Those who refused to participate were older (56 vs. 54 years old; p = 0.0001), more likely to be female (35% vs. 26%; p < 0.0001), non-Hispanic Black (90% vs. 79%; p < 0.0001), and have public insurance (81% vs. 73%; p < 0.0001).

Description of PROs Responses

Social Determinants of Health and Quality of Life

Based on the results of the baseline surveys, 272 (15.6%) participants reported a transportation need, 195 (11.2%) reported a utility need, and 582 (33.5%) reported having food insecurity (Table 2). Furthermore, 331 (19.0%) reported having depressive symptoms, 686 (39.5%) reported symptoms aligned with GAD, and 372 (21.4%) reported symptoms aligned with PTSD. The most common reported substance use was a history of cannabis use (1139; 65.5%). With regards to QOL, 23.1% of participants reported problems with mobility, 6.7% reported problems with self-care, 21.0% reported problems with usual activities, 44.1% reported problems with pain, and 43.2% reported problems with anxiety and depression. The distribution of these PROs remained constant on each of the annual surveys.

Table 2.

Distribution of participant characteristics and PROs among all PROs participants by completed survey

Patient Reported Outcome	Baseline PROs Survey (N = 1739)	First Annual PROs Survey (N = 610)	Second Annual PROs Survey (N = 142)
Insurance Private Medicare Medicaid Ryan White/ADAP Other Uninsured Unknown	466 (26.8) 323 (18.6) 563 (32.4) 30 (1.7) 315 (18.1) 38 (2.2) 4 (0.2)	136 (22.3) 128 (21.0) 169 (27.7) 5 (0.8) 156 (25.6) 14 (2.3) 2 (0.3)	35 (24.7) 26 (18.3) 42 (29.6) 1 (0.70) 35 (24.7) 3 (2.1) 0 (0.0)
Employment Employed Student Retired Disabled Unemployed Other Unknown	851 (48.9) 50 (2.9) 227 (13.1) 282 (16.2) 290 (16.7) 29 (1.7) 10 (0.6)	272 (44.6) 8 (1.3) 131 (21.5) 104 (17.1) 89 (14.6) 6 (1.0) 0 (0.0)	69 (48.6) 2 (1.4) 30 (21.1) 23 (16.2) 17 (12.0) 1 (0.7) 0 (0.0)
Housing Need Permanent Temporary Homeless Other Unknown	1500 (86.3) 106 (6.1 37 (2.1) 83 (4.8) 13 (0.8)	546 (89.5) 21 (3.4) 12 (2.0) 27 (4.4) 4 (0.6)	133 (93.7) 2 (1.4) 1 (0.7) 6 (4.2) 0 (0.0)
Unmet Needs a Utility Need Transportation Need Food Insecurity	195 (11.2) 272 (15.6) 582 (33.5)	72 (11.8) 95 (15.6) 238 (39.0)	10 (7.0) 22 (15.5) 53 (37.3)
Mental Health & Substance Use
Depression b	331 (19.0)	121 (19.8)	19 (13.4)
Generalized Anxiety Disorder c	686 (39.5)	257 (42.1)	52 (36.6)
PTSD d	372 (21.4)	135 (22.1)	22 (15.5)
History of Cannabis Use e	1139 (65.5)	398 (65.3)	94 (66.2)
History of Cocaine Use e	596 (34.3)	220 (36.1)	42 (29.6)
History of Amphetamine Use e	262 (15.1)	98 (16.1)	16 (11.3)
History of Inhalants Use e	307 (17.7)	100 (16.4)	20 (14.1)
History of Sedatives Use e	134 (7.7)	50 (8.2)	8 (5.6)
History of Hallucinogens Use e	297 (17.1)	97 (15.9)	19 (13.4)
History of Opioids Use e	181 (10.4)	67 (11.0)	11 (7.8)
Alcohol Use Score (Mean [St. Dev.]) f	3.2 (4.7)	3.1 (4.3)	2.6 (3.4)
Reported Quality of Life Problem g,h Mobility Self-Care Usual Activities Pain/Discomfort Anxiety/Depression	344 (23.1) 100 (6.7) 312 (21.0) 656 (44.1) 643 (43.2)	180 (29.5) 54 (8.9) 158 (25.9) 321 (52.6) 292 (47.9)	41 (28.9) 14 (9.9) 38 (26.8) 76 (53.5) 73 (51.4)

ADAP = AIDS Drug Assistance Program; PTSD = Post Traumatic Stress Disorder

^aMeasured with the Accountable Health Communities (AHC) Screening Tool

^bMeasured using the 8-item Public Health Questionnaire (PHQ-8) depression scale

^cMeasured using the 7-item generalized anxiety disorder scale (GAD-7)

^dMeasured using the primary care PTSD screen (PC-PTSD)

^eMeasured using the the Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST)

^fMeasured using the the Alcohol Use Disorders Identification Test (AUDIT)

^gMeasured using the 5-dimensional EuroQol instrument

^hOnly among the 1,489 participants who had access to the QOL on their baseline questionnaire

Symptom Burden

The most commonly reported symptoms experienced in the past week were worrying (46.1%), lack of energy (44.9%), difficulty sleeping (44.9%), and pain (44.1%) (Figure 1). Of those who reported each symptom, 12.7% reported they worry almost constantly, while 9.2%, 16.1%, and 12.5% reported, respectively, that their lack of energy, difficulty sleeping, or pain distresses or bothers them very much (Figure 1).

Fig. 1.

Percentage of reported symptoms in the past week based on baseline PROs surveys (N = 1739)

Adherence and HIV-related Stigma

Nearly all PROs participants (97.3%) reported they were currently taking an ART medication (Figure 2). However, only 79.1% (n = 1339) of those on an ART medication reported adherence to their ART medication in the past 4 days. Of those who reported adherence in the past 4 days, 93.9% (n = 1301) reported adherence in the past 7 days, and reported adherence remained high for all further time points. Reported HIV-related stigma was low overall, with “strongly disagree” being the most common response to each of the four statements. The most commonly agreed or strongly agreed to statement was “I feel ashamed of having HIV” (28%) (Figure 3).

Fig. 2.

Flow chart of adherence to ART among PROs participants based on the baseline PROs survey

Fig. 3.

Distribution of responses to HIV stigma questions based on the baseline PROs survey

Sexual Risk Behaviors

Participants were asked to describe sexual activity and 59.7% reported having at least one sexual partner in the past 3 months (n = 1,038) (Figure 4). Among those who reported sexual activity in the past 3 months, 25.1% (n = 261) said all their partners were PWH and 46.4% (n = 482) reported that none of their partners had HIV. Among those who reported partners with HIV, 60.5% (n = 158) had condomless sex with those partners, but of those reporting condomless sex, only 17.7% (n = 28) reported they were concerned about being exposed to a sexually transmitted infection (STI). Among those who reported none of their partners had HIV, 51.0% (n = 246) reported having condomless sex with those partners and only 11.0% (n = 27) of those participants reported concern with an exposure to an STI during one of these encounters (Figure 4).

Fig. 4.

Flow chart of sexual risk behaviors among PROs participants based on the baseline PROs survey

Discussion

PROs Findings

We evaluated 1,739 PROs surveys among a subset of PWH enrolled in the DC Cohort. Overall, we found that the most common unmet need was food insecurity, reported by one third of our population. This is much higher than the prevalence of food insecurity in the general population of Washington, DC, which was 11% in 2021 (DC Food Council 2021). Similarly, GAD was the most commonly reported mental health condition, with about 40% of participants in our dataset reporting symptoms aligned with GAD. This was further confirmed with the QOL instrument, where just under 40% reported problems with anxiety and depression. This high prevalence of anxiety in our study population is concerning given that prevalence of GAD in the general United States population is only around 3% (Generalized Anxiety Disorder - National Institute of Men tal Health (NIMH) n.d.; Harvard Medical School 2007) and the prevalence of anxiety among PWH in the United States is around 19% (Beer et al. 2019). However, these estimates do align with the prevalence of mood/anxiety disorders among PWH in the Southeastern United States, which was reported to be 39% in a study conducted in 2006 (Pence et al. 2006). These findings show that our population of PWH in the DC area has a substantially higher burden of food insecurity and mental health disorders compared to the general population. Burden of disease, medical bills, stigma, and other HIV-related comorbidities may be contributing to these disparities (Levy et al. 2017; Logie et al. 2018; Remien et al. 2019; Felker-Kantor et al. 2019). Given the unique needs of this population, future research should focus on targeted interventions to improve mental health and food insecurity specifically among populations of PWH.

Most participants reported sexual activity in the month prior to survey completion, among which over half did not use barrier protection. Of note, only 51.0% of participants who had partners without HIV had condomless sex compared to 60.5% of participants who had partners with HIV, suggesting there may be behavioral differences based on the HIV status of their sex partners. However, among those who did not use protection, nearly 30% had a detectable HIV viral load, suggesting potential for HIV transmission to partners without HIV. Despite the high rates of condomless sex, a minority of participants reported being concerned with an STI exposure, regardless of the HIV status of their partners. These findings suggest there is a need for more educational campaigns focusing on sexual risk reduction counseling (Secco et al. 2020; Byrne et al. 2022). Furthermore, among those sexually active, 20% did not know the HIV status of their partner, suggesting participants are often not discussing HIV status with their sex partners.

Finally, despite nearly every participant reporting taking ART, only 79% reported being completely adherent in the past 4 days. Among those who were adherent in the past 4 days, almost all reported being adherent for all further timepoints. This suggests there is a subset of our population that is consistently non-adherent to ART, while those who were recently adherent tend to have consistently high adherence. These findings suggest that despite high uptake of ART, there is still a need for enhanced strategies to improve adherence to ART. Individuals reporting recent non-adherence to ART during routine clinical visits may be consistently non-adherent. Providers can use recent adherence history as an indicator to discuss adherence strategies with their patients. However, it is important to consider that perfect adherence may not be required for viral suppression, and thus minor levels of non-adherence may not be of clinical concern (Miron and Smith 2010; Maggiolo et al. 2022).

Comparison of PROs Use

PROs have been used previously by other cohorts to evaluate various facets of life among PWH and other related infections (Kozak et al. 2012; Antinori et al. 2020; Schulte et al. 2020; Carlsson-Lalloo et al. 2022). The Center for AIDS Research (CFAR) Network of Integrated Clinical Systems (CNICS) uses validated survey instruments to collect information regarding alcohol, substance use, mental health, and ART adherence among PWH using patient reported measures taken every 6 months (Kitahata et al. 2008; Rudolph et al. 2018; Satyanarayana et al. 2021). The Johns Hopkins HIV Clinical Cohort collects similar data elements, in addition to questions regarding the local drug market, intimate partner violence, provider satisfaction, and other site-specific questions (Lesko et al. 2024). These PROs measures provide unique opportunities for research that are not typically available from EHR data. For instance, a cohort study of PWH in the US found that substance use and depression were associated with lower ART adherence from a PROs survey (Kozak et al. 2012). The PROACT pilot study identified a need for more research on interventions involving PROs screening results being conveyed to providers and several studies have researched whether availability of PROs results to providers improved patient-provider communication (Fojo et al. 2021; Jabour et al. 2021; Short et al. 2022). Others have suggested implementing PROs in routine clinical care, wherein the framework for web-based implementation was outlined by two recent publications (Barger et al. 2018; Crossnohere et al. 2024). Despite the wide range of PROs studies available, there are no recent publications providing detailed guidance for the methodological implementation of a longitudinal PROs survey within the framework of a longitudinal HIV cohort. Therefore, other HIV cohorts can use the methods described here to capture PROs data and work towards implementing their own measures into routine clinical care.

Enrollment Successes and Challenges

As of this publication, we have successfully enrolled over 2,000 participants in the PROs survey and recruitment is ongoing. This high recruitment rate is likely due to the inclusive eligibility criteria and ease of enrollment and survey completion. The PROs survey can be accessed electronically and does not require a participant to be seen in-person. This removes a common obstacle that participants, especially PWH, face related to transportation, time away from work, anonymity, and the time and monetary cost of additional healthcare visits.

Even still, there are challenges to enrollment. RAs must contact the participant and provide them a link to the survey and a unique survey ID that the participant must enter when completing the survey. Since many participants complete the survey without RA supervision, this ID could be transcribed incorrectly or forgotten which results in difficulty linking surveys to participants. The effort required for a participant to either retake that survey or for the study team to determine who completed the survey with the invalid ID is considerable. Likely due to our aging DC Cohort population, there are some individuals who choose not to participate in PROs due to technological concerns, even when offered assistance. Furthermore, loss to follow-up and mismatching demographics on follow-up surveys limit our ability to gather longitudinal data on all participants.

Strengths and Limitations

The DC Cohort PROs survey has several limitations. The biggest limitation is the self-reported nature of the responses. Recall bias or social desirability bias may reduce the validity of responses, especially with regards to potentially stigmatizing questions such as substance use and sexual risk behaviors. However, it is important to note that surveys were self-administered so social desirability bias may be limited. Similarly, the subjectivity of responses, such as the severity or frequency of symptoms, may also lead to inconsistent responses between participants and lower the validity or reliability of results. Furthermore, currently only 142 participants completed the survey three times. At this current sample size, we are limited in our ability to conduct complex longitudinal analyses. However, the survey is ongoing and this number will likely reach an adequate sample size in the near future. In addition, the PROs survey is cross-sectional in nature. Therefore, we are unable to determine the temporality of PROs measures that could have a potential causal relationship with each other. Lastly, our PROs survey may be limited by generalizability. As mentioned above, PROs participants differ significantly from eligible DC Cohort participants by age, gender, time since HIV diagnosis, CD4 count, type of HIV care site, engagement in care, and history of AIDS. This may suggest there is something inherently different about DC Cohort participants who choose to take part in the PROs survey and those who do not, potentially lowering the external validity of findings.

Despite these limitations, there are several analytic strengths. First, the longitudinal nature of the study allows for follow-up of participants over time and can be used to evaluate whether changes in PROs responses over time are associated with changes in HIV clinical outcomes or other related factors. Furthermore, the PROs questionnaire allows us to evaluate variables such as sexual risk behaviors, substance use, symptom burden, and QOL that are not always captured, or are captured in less detail, in the typical EHR. The detailed questions about adherence to ART also provide insight into potential gaps in the HIV care continuum. Finally, the ability to link PROs and clinical data allows for a more in-depth understanding of clinical outcomes and behaviors, as well as challenges related to the HIV care continuum.

Future Directions

We anticipate continuing annual PROs survey collection to create a large, longitudinal, comprehensive database for PROs among PWH. Future research will benefit from access to such a database, improving our ability to evaluate these self-reported measures and how they factor into a patient’s overall experience with HIV. The goal of gathering these data is to create tangible change in the local HIV care continuum, further contributing to the goal of ending the HIV epidemic. We believe integration of PROs results into EHRs could help healthcare providers have a more comprehensive picture of their patient’s health. The DC Cohort has an aggregated Data Dashboard that would be ideal to incorporate aggregated PROs results into for provider use. Individual-level PROs data would be even more valuable for more tailored individual care, however also proves more challenging to incorporate into Data Dashboards. The ability for a provider to evaluate treatment effectiveness in regard to a patient’s QOL, symptom burden, or mental health, in addition to routine HIV labs, can help improve patient-provider relationships and positively impact clinical outcomes in PWH. We strive for this framework to allow patients to understand their own PROs and serve as a more effective advocate for their health.

Conclusions

The PROs survey responses in the DC Cohort provided significant findings on quality of life, symptom burden, adherence, stigma, and sexual risk. The combination of annual PROs survey data and ongoing EHR data collected in the DC Cohort provide a comprehensive look into participant health outcomes and factors in a longitudinal manner. Future research could gain valuable insights by utilizing databases that integrate self-reported data and EHR measures, fostering actionable improvements in the local HIV care continuum.

Availability of data, material, and code

Per DC Cohort protocols, data are available upon request and approval of the DC Cohort Executive Committee. Interested parties should email the Principal Investigator at acastel@gwu.edu.