Table 2.

The relationship between employment and continuum of care outcomes

Author, year	Location	Study design	Sample (N)	Sampling strategy	Employment measure	Continuum of care measure	Main finding
Diagnosis
Delpierre, 2006	France	Prospective multicenter cohort	Patients from 6 hospitals from 1996–2005 (N = 5702)	Non-probability sample of all patients who sought care from 6 hospitals	Employment status: stable employment, on employment benefits, unemployed or other	Late testing: presenting with symptoms of AIDS or CD4 < 200 mm3 during the year of diagnosis	Among MSM, unemployment associated with late testing (OR = 2.23, 95% CI 1.14–4.36); among women and heterosexuals employment status not associated with late testing
Massari, 2011	France	Prospective cohort	General population living in the Paris metropolitan area, SIRS Cohort (N = 3023)	Multistage, random sample	Employment status: employed, unemployed and inactive, which includes retired or student	HIV testing: Never having been tested for HIV	Among women, unemployment was associated with no history of HIV testing (OR 1.85, 95% CI 1.24–2.77). Among men, a significant association was not observed
D’Arminio, 2011	Italy	Observational Cohort Study	Antiretroviral-naïve patients with a recent HIV diagnosis and CDC + T-cell count (N = 2276)	All patients enrolled in the Icona Foundation Study who met eligibility criteria from 1997 to 2009	Employment status: becoming unemployed, as opposed to retiring	Late diagnosis: patients presenting with symptoms of AIDS or CD4 T-cell count less than or equal to 350/mm3	Unemployment (in comparison to retirement) was associated with late diagnosis (OR = 0.36, 95% CI 0.21–0.61)
Sheehan, 2017	Florida, USA	Surveillance study	Florida Department of Health Enhanced HIV/AIDS Reporting System and American Community Survey (eHARS)(N = 5522)	Reported surveillance cases between 2007 and 2011	Neighborhood employment level: the percent of the population who is 16 of older and are unemployed	Late diagnosis: receiving an AIDS diagnosis within 3 months of receiving their HIV diagnosis	Among Latino men, neighborhood unemployment strata were associated with late testing. By 2nd, 3rd, and 4th strata of unemployment, ORs (95% CIs): 1.37 (1.06–1.78), 1.29 (1.01–1.66), 1.33 (1.03–1.72)
Linkage to HIV Care
Girardi, 2004	Italy	Multi-center cohort	Individuals living with HIV over the age of 18 (n = 4453)	Antiretroviral-naïve patients enrolled from 59 clinical centers from 1997 to 2000	Employment status, coded as stable employment, unemployed, or other	Delayed presentation: greater than 6 months from positive test result to presenting for HIV medical care	Unemployment associated with delayed presentation to care (OR = 2.75, 95% CI 1.66–4.53)
Aaron, 2015	Philadelphia, USA	Retrospective cohort	Individuals with an HIV positive diagnosis who attended a large urban healthcare center (n = 87)	Medical records were used to ascertain HIV positive diagnoses in inpatient and outpatient settings from 2007 to 2011	Employment status: dichotomous variable, where either employed or unemployed	Linkage to care: one visit to an HIV medical provider within 6 months of receiving a positive test result	Unemployment was associated with failure to link to care (OR = 6.50, 95% CI 1.13–37.32)
Nijhawan, 2017	Texas, USA	Retrospective chart review	Individuals who were scheduled for a new patient visit after January 1st, 2014 from two clinics (n = 200; 100 per clinic)	Consecutive sample of patient records from 2014–2016	Employment status, coded as employed, unemployed, or disabled	Linkage to care: missing vs. attending the first scheduled medical visit	Unemployment was associated with missing the first scheduled appointment. (OR = 2.33, 95% CI 1.04–5.24)
Retention in HIV care
Joy, 2008	British Columbia, Canada	Prospective cohort	Treatment-naïve patients accessing HAART between September 1, 1997 and November 30, 2004 (n = 2168)	Enrollment in provincial drug treatment program (the only method by which patients can access HAART in British Columbia)	Neighborhood measurement of percentage of unemployed individuals	Delayed access to treatment: not accessing treatment until patient’s CD4 count was low (< 50 cells/mm)	Unemployment was associated with delayed access to treatment (OR = 1.41, 95% CI 1.14–1.74)
Shacham, 2013	St Louis, USA	Cross-sectional study	All patients who attended the Washington University HIV clinic in 2008 (n = 762)	All patients who attended the Washington University HIV clinic in 2008 were required to complete a behavioral survey as a part of their medical visit. Data was gathered from these surveys	Employment level, measured as a neighborhood-level variable	Receipt of ART prescription:, use of at least three drugs from two different antiretroviral drug classes or the use of at least three nucleoside reverse transcriptase inhibitors	Neighbourhoods with higher rates of unemployment had individuals who were less likely to have a current ART prescription (OR = 1.47, 95% CI 1.05–2.04)
HIV medication adherence
Torres-Madriz, 2011	Boston, USA	Randomized, crossover trial	Patients on ART with a detectable viral load in the greater Boston area (n = 156)	Convenience sample of patients from five patient care settings	Work accommodations dichotomized as any or none	ART adherence: percentage of doses taken correctly during the 30 days prior to each study visit	Patients with accommodations had a 12% higher mean adherence than those without accommodations (Parameter estimate 0.12, 95% CI 0.03–0.20)
Nachega, 2015	N/A – previously conducted trials were used	Meta-analysis	28 articles meeting all criteria were included. This yielded information on 8743 individuals from 14 different countries	Records were identified from Medline, Embase, and Cochrane Central Register of Controlled Trials	Employment was globally assessed across all the included studies	Adherence was globally assessed across all the included studies	In low and high-income countries, being employed was favorably linked to adherence. Low: (OR = 1.85, 95% CI 1.58–2.18);Middle: (OR = 0.94, 95% CI 0.62–1.42); High: (OR = 1.33, 95% CI 1.02–1.74)
Viral load
Dray-Spira, 2005	France	Prospective cohort study	Patients from 66 hospitals across France, enrolled between 1996 and 2002 (n = 319)	Convenience sample of patients recruited through hospitals who were recently diagnosed and ARV-naïve	Employment loss: change from employment to inactivity between two semi-annual visits	High viral load: greater than 10,000 copies/mm3	Both persistent and incident high viral loads were associated with employment loss. Respectively, (OR = 2.4, 95% CI 1.1–5.0); (OR = 3.7, 95% CI 1.0–13.9)