Trends and emerging directions in HIV risk and prevention research in the Philippines: A systematic review of the literature

Background The Philippines is experiencing one of the fastest growing epidemics globally. Evidence-based public health policies are needed. To describe the public health literature on HIV risk groups and prevention approaches in the Philippines, we reviewed published empirical studies with HIV-related outcomes. Methods Based on an a priori systematic review protocol, we searched PubMed, PsycINFO and CINAHL databases for quantitative studies conducted in the Philippines that reported on HIV risk groups factors and interventions to prevent HIV. The search included studies published as of April 2018. Results We identified 755 records, screened 699 unique titles and abstracts, and conducted full text review of 122 full reports of which 51 articles met inclusion criteria. The majority were cross-sectional studies describing HIV and STI prevalence and risk factors in samples recruited from the Philippines. Four HIV prevention programs conducted in the Philippines were identified, all of which reported improvements on HIV knowledge, attitudes, and behaviors. Overall, female sex workers (FSWs) constituted the primary study population, and few studies reported data from men who have sex with men (MSM), people who inject drugs (PWIDs), and youth. No studies reported on transgender populations. Most studies were focused on examining condom use-related outcomes and STI history, few had biomarkers for HIV, and none addressed biomedical HIV prevention strategies. Conclusion This review identifies an agenda for future HIV research that is needed to address the growing and shifting nature of the HIV epidemic in the Philippines.


Introduction
After the first HIV case was identified in the Philippines in 1984, the country's estimated HIV prevalence had remained low for over two decades [1]. According to the Joint United Nations Programme on HIV/AIDS (UNAIDS)'s surveillance reports, the Philippines' progress towards reaching HIV/AIDS 90-90-90 treatment for people living with HIV and knowing their HIV status (67%), on treatment (32%), and are virally suppressed (29%), is slow as HIV infections rise in the Philippines [2]. National surveillance data showed that the number of new HIV cases in the Philippines started to rise at an alarming rate during the past decade, with an increase from 311 cases identified in 2007 to 8,151 cases identified in 2016 -representing a 26-fold increase in new HIV diagnoses [3]. According to a 2014 national report, 93% of HIV cases in the Philippines were transmitted through sexual contact and were particularly concentrated among youth and young adults [4].
Despite the growing HIV epidemic in the Philippines, there have been challenges in mobilizing local and national HIV prevention, education, and testing programs [5,6]. Evidencebased public health is needed. However, a 2015 report by the World Health Organization (WHO) highlights that the body of HIV research conducted in the Philippines has been limited across all areas, including prevention, epidemiology, evaluation, and behavioral science, which are each essential to developing effective public health strategies [7,8].
Some of the recognized key populations for HIV risk in the Philippines include men who have sex with men (MSM), transgender people, female sex workers (FSW), youth, and overseas workers [1,6,7]. Although the estimated number of people who inject drugs (PWID) in the Philippines has been historically low, there have been anecdotal reports suggesting a growth in this population [9]. The National AIDS and STI Prevention and Control Program for the Philippines has urged for the development of HIV prevention and public health initiatives targeted towards key populations [7]. However, an external review by the WHO of the national response acknowledged multi-level challenges in implementing HIV prevention and treatment activities, including a misalignment of healthcare priorities in the national-and city-level settings, limited healthcare infrastructure and human resources to provide prevention and treatment services to key populations, and a nascent research literature on which to build evidenceinformed strategies [7].
We conducted a systematic review to examine the body of empirical literature on HIV risk and intervention programs in the Philippines. Our specific aims were to synthesize findings about population characteristics associated with HIV status or HIV risk in the Philippines, to identify and describe local HIV prevention interventions evaluated in the Philippines, and to describe the methodological characteristics of this body of research. As a secondary aim, we sought to describe differences in reported HIV risk in studies conducted before 2008 compared with studies published during or after 2008, which represents the year during which national surveillance observed noted increased HIV cases. conducted in the Philippines; 2) sampled Filipino/a participants; 3) published in English; 4) reported quantitative findings on any of the following category of outcomes: biological risk for HIV (biomarkers/biologically confirmed HIV or other STIs); self-reported HIV status or STI diagnoses or symptoms; HIV-related sexual risk behavior (such as condomless sex, commercial sex, sex under influence of drugs/alcohol, sex with partner of HIV-positive or unknown status); injection drug use; knowledge, attitudes and beliefs relating to HIV/AIDS risk and transmission. We did not limit studies according to type of quantitative design (i.e., cross-sectional, intervention trial, etc.) or presence of a comparison or control group.
Electronic searches of PubMed, PsycINFO, and CINAHL were carried out using an a priori search strategy in April 2018. The search strategy included validated MeSH terms for HIV as well as terms related to Philippines. For example, the PubMed search used the following terms: [ This search yielded 755 records (see Fig 1 for a flowchart of the systematic review process). A team of reviewers received training in applying the inclusion criteria to research records; reviewers applied the screening criteria to an initial batch of 100 records and discussed discrepancies until reliability was achieved. Screeners were instructed to apply initial inclusion criteria liberally, such that records were retained in the search process until the team was sure that studies did not meet inclusion criteria. After excluding 56 duplicate records from the database, two reviewers screened the remaining 699 records resulting in a shortlist of 122 records that appeared to meet inclusion criteria based on information in the title or abstract. These articles were retrieved for full text review, which was performed by four reviewers (including the 2 previous reviewers) who identified 51 articles that met inclusion criteria. We excluded 71 articles because they were not based in the Philippines, were not empirical studies (i.e. editorials, commentaries, or reviews), or did not report on outcomes specified for this review. We identified several articles meeting inclusion criteria that reported different findings from the same parent research study; these articles were retained in this review.

Data extraction and analysis
The overarching goal of this review was to summarize trends in the published literature and appraise the methodological quality of identified studies. For all research articles that met inclusion criteria, we extracted information about the year(s) of data collection, study sample and location, sampling method, study design, HIV-related outcome(s), and main findings (see Table 1). Articles that reported data from the same parent study are grouped together in Table 1, with the primary or lead article denoted by superscript "a" and subsequent articles denoted by "b", "c", etc.
Narrative synthesis of studies was conducted based on information extracted into Table 1. Because of heterogeneity in sampling approaches, time, and indicators of HIV risk across studies, meta-analysis was not conducted. We assessed methodological characteristics of each study using a critical appraisal checklist developed by Munn and colleagues [11]. These methodological characteristics included sample representativeness, recruitment strategies, adequacy of sample size, participant drop-out or non-response, description of setting and participant, objective criteria for outcome measurements, reliability of outcome, appropriateness of statistical analysis, accounting of confounding factors, and identification of sub-populations.
Authors were not blind to any aspect of the studies. Funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.  Table 1 presents characteristics of 51 included articles from 30 independent studies. Twentyfive articles reported quantitative findings from independent studies . Two articles reported analyses from a cross-sectional survey of sexual behaviors among adult men and women from Cebu (a large metropolitan city in the Philippines) [37][38]. Fifteen articles reported analyses from a longitudinal, quasi-experimental investigation of a community-based HIV prevention program involving female bar workers and managers in four regions of the Philippines [39][40][41][42][43][44][45][46][47][48][49][50][51][52][53]. Four articles reported findings from a quasi-experimental intervention to prevent HIV and STIs among heterosexuals in southern Philippines [54][55][56][57]. Two articles reported analyses of a survey of people who inject drugs in Manila, Cebu, and Davao cities [58][59]. Three articles reported analyses of a survey of FSW recruited from entertainment establishments in metro Manila [60][61][62].

Characteristics of included studies
Overall, data collection for the included studies occurred between years 1985 to 2015. Of 30 independent studies, most (n = 18) studies reported findings from data collected before 2008 and 10 studies reporting findings from data collected during or after 2008; 2 studies did not specify year of data collection, but both were published before 2008. Sample sizes per study ranged from 62 to 144,000; the latter study involved analysis of blood bank data. Female sex workers (e.g., registered and freelance) constituted the most frequently studied population (n = 11 unique studies). Fewer studies reported data on MSM (n = 3 unique studies), PWID (n = 3 unique studies), youth (n = 4 unique studies), seafarers (n = 3 unique studies), or incarcerated females (n = 1 study). One study reported on overseas worker candidates [35]. No included studies reported data on overseas workers and transgender populations, and no studies reported data on gender expression or gender identity indicators.
Behavioral risk factors. Commonly reported HIV risk factors included self-reported condom use, and substance use behaviors. Of the included articles, thirty-one reported condom use behaviors. Assessment of condom use varied widely across studies, including measures of condomless sex with different partner types (e.g., with sex workers, multiple sexual partners, group sex, regular partners, and casual partners), frequency of condom use (e.g., using condoms always, consistently, inconsistently, and never), condom use according to type of intercourse (e.g., condom use during vaginal or anal sex), and condom use according to specific sexual events (e.g., condom use at first sex, and during last sexual encounter). Assessment of condom attitudes and knowledge also varied, with measurements including knowing that condoms prevent HIV/AIDS, pregnancy, and STI infections, as well as feeling that condoms reduce sexual pleasure or enjoyment, and that using condoms is against religion. Only five studies reported substance use outcome measures [21,34,38,57,62], and these studies varied with regard to assessment about type, frequency, and amount of substance use.
Intervention programs. We identified a total of four independent intervention studies aiming to prevent HIV transmission, with evaluation results reported across 9 reports [17,32,39,41,42,47,48,54,55]. One intervention study conducted in 1995 by Aplasca et al. [17] involved a cluster-randomized trial of a school-based program to improve HIV-related knowledge, attitudes, and behaviors among high school students in Manila. Findings included improved levels of knowledge about HIV biology, transmission and prevention, as well as improved attitudes and compassion for people living with HIV in the intervention versus control groups; no effects were found on intentions to engage in preventative behaviors [17]. Another intervention study conducted in 2013 by Urada et al. [32] involved a human-rights focused HIV intervention for sex workers in Manila. Participants completed a single 4-hour intervention providing HIV and STI knowledge and prevention strategies, and contextualized risk and protective factors in accordance with the laws, systems, and social milieu regarding sex work, violence, and discrimination. In pre-post analysis, participants reported higher levels of knowledge about HIV reproductive health, human rights, research ethics, and intentions to receive an HIV test [32].
Two intervention programs were quasi-experimental studies. First, a 3-year study conducted from 1994-1998 by Morisky et al. [39] used community-based participatory methods to train managers and peer educators on HIV and STI prevention at 130 entertainment establishments in 4 regions in the Philippines. Trained managers and peers then implemented and disseminated information within their establishments. In post-test evaluation analyses, FSW employed in these establishments reported significant improvements in consistent condom use, improvements in HIV testing, and reductions in STI infections [39,42,47,48]. Additional analyses showed that improvements in condom use were strongest in establishments that also instituted condom use policies for employees and patrons [41]. Second, a 3-year quasi-experimental study conducted from 2000-2005 by Morisky et al. [54] used community-based participatory methods to provide HIV prevention education training to peer leaders recruited from six male populations in the southern Philippines: military members, police and firemen, construction workers, taxi drivers, pedicab drivers, and community residents; trainees were then expected to educate 10 or more peer network members on HIV and STI prevention. In posttest evaluation analyses, intervention participants reported lower levels of STI infections as well as improvements in HIV and STI knowledge, attitudes toward condoms, condom use behavior, discussion about HIV with co-workers, and exposure to HIV prevention education, compared with those in the control [54,55].
Finally, another prospective STI treatment study was identified, but was not considered a scalable HIV prevention intervention compared with the four studies described here. In this STI treatment study conducted in 1996-1997 by Aplasca de los Reyes et al. [18], FSWs from Manila and Cebu testing positive for N. gonorrhoeae were randomly assigned to receive a single dose of ciprofloxacin (500 mg) or a single dose of cefixime (400 mg), and were re-evaluated 4-7 days later. Findings showed high rates of treatment failure and resistance in participants who received ciprofloxacin and adequate effects for single-dose cefixime [18].

Methodological appraisal of the included studies.
For selecting participants, most of the study participants were recruited using sampling methods based on non-randomized sampling strategies, including convenience and purposive sampling such as venue-based (e.g., community or establishment) recruitment approaches or from surveillance studies. Common examples of venue-based recruitments were via bars, nightclubs, karaoke TV lounges, massage parlors, and health clinics. One study reported data from a nationally representative youth survey [19], and two studies reported data from respondent driving sampling surveys [34,58]. Although all studies indicated the type of recruitment strategy used (e.g., convenience sampling, etc.), almost 13 did not specifically detail about how participants were recruited. All studies had adequate sample size to perform analysis, and had high participant response rate. Most studies (n = 26) provided adequate description of their participant samples and the study settings.
When considering outcomes, nineteen studies did not use objective standard criteria for measuring outcomes (i.e., validity) such as a biologically-confirmed results. Self-reported outcomes were primarily assessed and there were noticeable inconsistent measurements in condom use and substance use behavioral measures and HIV-related knowledge and attitude measures. Statistical analyses for each study were conducted appropriately based on study design. Of the four intervention studies, one used a cluster randomized control trial design, two used non-randomized comparison groups, and one used a pre-post evaluation design. Intervention studies assessed self-report behavioral outcomes only; none measured biological or objective indictors of HIV risk.

Discussion
Despite the rise in HIV infections documented in the Philippines' national surveillance reports since 2008, this review highlights the limited body of published research on HIV infection and risk factors in key populations, a paucity of research on interventions to promote HIV prevention and testing in the Philippines, and opportunities for improving methodological rigor in future research. Overall, we identified 51 published quantitative papers reporting on HIV-or STI-related biological, behavioral, or social-cognitive findings from 30 unique studies conducted in the Philippines. The majority of papers identified in this review reported on data collected before 2008; only 10 papers (reporting on 7 discrete studies) reported data collected during or after 2008. Moreover, this review identified only 4 HIV prevention intervention evaluations in the published literature.
Prior to 2008, FSWs constituted the population most frequently studied in HIV research conducted in the Philippines. After 2008, a small number of studies included MSM and PWID. Indeed, the papers by Gangcuangco and colleagues [21] and Telan and colleagues [59,60] and are among the only identified published studies which recruited and reported specifically on HIV risk in MSM populations. Findings from PWID were included in four reports-one involving surveillance data from an earlier phase of the epidemic [16], three conducted after 2008 by Verdery and colleagues [34] and by Telan and colleagues [58,59]. Four studies included youth; two were conducted before 2008 [17,19] and two after [28,38]. No identified studies specifically reported on overseas workers and transgender populations.
This review identifies a need to improve the body of knowledge about HIV risk and transmission among key populations in the Philippines. Research targeting MSM, PWID, and transgender populations is needed to understand the transmission risk factors and specific structural, social, behavioral, and epidemiological factors impacting these groups. While there is a considerable body of HIV prevention intervention research focusing on MSM and PWIDs in other parts of the world [63,64], none have been specific to the Philippines. It remains unclear whether existing prevention interventions are adaptable or require distinct design for this national context. Additionally, despite an estimated global HIV prevalence of 19% [65], and anecdotal reports that suggests a growing burden of HIV among transgender women in the Philippines [66], none of the included studies have focused on or included this key population. It is possible that researchers in the Philippines have aggregated transgender populations within the MSM rubric [67]. Given HIV epidemiological trends within the Philippines and evidence from other settings about the disproportionate prevalence of HIV among transgender people, future research must disaggregate transgender and MSM populations and resist the conflation of gender and sexual identities [68]. Additionally, given that UNAIDS surveillance data points to youth and young adults being affected [2], more biological and behavioral research data and interventions are necessary to understand the epidemic among youth and young adults in the Philippines, especially those who are also members of MSM, PWID, or transgender populations.
Most of the studies identified in this review used cross-sectional surveys with convenience samples, involving mostly descriptive measures, which limit the generalizability of the research. While these study designs are useful in exploratory investigations, it is imperative for researchers to increase the rigor of investigations by using longitudinal and experimental studies in order to examine more complex research hypotheses (e.g., testing hypotheses about social-behavioral determinants of HIV infection) and to test interventions [68].
Moreover, as the majority of includes studies were conducted prior 2008, researchers must also examine biomedical factors that might determine or mitigate the growth of HIV in the Philippines-e.g., factors associated with medication adherence and viral load suppression among PLHIV (i.e., treatment-as-prevention), access to and use of biomedical prevention such as post-exposure prophylaxis and pre-exposure prophylaxis, home-based testing, and male-circumcision. Given the political climate regarding drug use in the Philippines, the viability of harm reduction and needle exchange programs for HIV prevention must be carefully considered [69,70].
The literature identified in this review was also noteworthy with regard to the scarcity of investigation into psychosocial and ecological factors that contribute to HIV risk, mental health, personal and community empowerment, stigma, and substance use. While it is important to continue assessing condom use-related measures in key populations, it is imperative that researchers examine contextual conditions and co-morbid health factors associated with low condom use as well as the broader social-structural drivers of HIV risk and infection, in order to understand factors affecting the acceptability and feasibility of bio-behavioral preventative strategies.
Taken together, these findings provide initial insight into the increasing and shifting HIV epidemic in the Philippines, from an initial concentration among FSWs to including populations such as MSM and PWIDs. These findings also suggest the need to develop social-contextual frameworks to prioritize HIV prevention strategies and contextualize HIV risk/ prevention according to the lived experiences of key populations. Assessment of clinical and service provider capacities in responding to HIV risk and infections is also important to guide research and build a stronger public health infrastructure.
Limitations of this review must be considered. First, although this review followed a systematic search protocol and a priori inclusion criteria, it is possible that not all relevant articles were identified. Second, meta-analysis was not deemed appropriate due to heterogeneity in samples, methods, measures, and time of data collection. Third, the review included only English-language publications. Fourth, the generalizability of the review might be limited given the specificity of the context for this review and the ongoing evolution of the Philippines' national epidemic. This is the first known systematic review to provide evidence about and identify gaps in published research about risk groups, risk factors, and intervention approaches addressing the Philippine's HIV epidemic. Studies revealed a nascent body of literature, especially with regard to intervention research, biomedical prevention, and key populations currently impacted by the HIV epidemic such as PWIDs, MSM, and transgender populations. Future HIV research studies in the Philippines must use rigorous research methodologies including purposeful sampling strategies and validated measures (including biological assessments for HIV and STIs, and established measures for substance use, condom use, and mental health). Furthermore, given the evolving epidemic in the Philippines, researchers should capitalize on opportunities to implement and evaluate bio-behavioral intervention strategies including preexposure prophylaxis, treatment-as-prevention, and regular HIV testing with key populations. Use of online data recruitment and data collection approaches can improve access to hard-toreach or remote populations; online research might be especially useful for reaching members of historically stigmatized groups. A multi-disciplinary research agenda for understanding and addressing HIV transmission in the Philippines must span across various domains of research including prevention, epidemiology, treatment, and behavioral and social science.