Healthcare utilization among persons living with HIV in Manitoba,Canada,prior to HIV diagnosis: A case-control analysis

Abstract

Background

Understanding care patterns of persons living with HIV prior to diagnosis can inform prevention opportunities, earlier diagnosis, and engagement strategies. We examined healthcare utilization among HIV-positive individuals and compared them to HIV-negative controls.

Methods

Data were from a retrospective cohort from Manitoba, Canada. Participants included individuals living with HIV presenting to care between 2007 and 2011, and HIV-negative controls, matched (1:5) by age, sex, and region. Data from population-based administrative databases included physician visits, hospitalizations, drug dispensation, and chlamydia and gonorrhea testing. Diagnoses associated with physician visits were classified according to International Classification of Diseases chapters. Conditional logistic regression models were used to compare cases/controls, with adjusted odds ratios (AORs) and their 95% confidence intervals (95% CI) reported.

Results

A total of 193 cases and 965 controls were included. Physician visits and hospitalizations were higher for cases, compared to controls. In the 2 years prior to case date, cases were more likely to be diagnosed with “blood disorders” (AOR: 4.2, 95% CI: 2.0–9.0), be treated for mood disorders (AOR: 2.4, 95% CI: 1.6–3.4), and to have 1+ visits to a hospital (AOR: 2.2, 95% CI: 1.4–3.6).

Conclusion

Opportunities exist for prevention, screening, and earlier diagnosis. There is a need for better integration of healthcare services with public health.

Keywords

HIV North America epidemiology prevention

Introduction

Despite the advent of combined antiretroviral therapy (ART), many people living with HIV continue to present at an advanced stage of the disease.^1–4 In industrialized countries, late presenters (defined as those who present with a CD4⁺ cell count of <200 cells/mm³ or who progress to AIDS within 1 year of presentation) account for 15–43% of new diagnoses.⁴ In Canada, where universal healthcare exists, 39% of reported new HIV cases present at this late stage,⁴ while in the province of Manitoba, approximately 27% of new HIV diagnoses made in 2016 were in patients with CD4 count <200 cells/mm³.⁵

There are several important implications of delayed recognition of HIV.^6–10 First, the longer patients have untreated disease the more opportunities there are for transmission.^11,12 Second, healthcare costs correlate inversely with CD4⁺ cell counts^13,14; the annual cost for a patient with a CD4 count of <200 cells/mm³ is approximately double that of a patient with a count that is >500 cells/mm³.¹³ Finally, late presentation is known to increase mortality.^15–17

The reasons for late presentation are multiple and include both patient- and provider-related factors. Research has demonstrated that people living with HIV often present to healthcare providers at least once in the years immediately prior to their diagnosis.^18–22 Duffus et al.¹⁹ observed that late presenters sought care more often than those diagnosed early in the disease. They noted that in 33% of prior visits, a high-risk feature, such as injection drug use was present, and yet testing for HIV was not performed. Shaw et al.²³ found that in the 5-year period prior to their diagnosis, serological testing for other sexually transmitted infections (STIs) was higher among individuals living with HIV, compared to HIV-negative controls. A better understanding of healthcare seeking patterns prior to HIV diagnosis may offer insights into opportunities for HIV prevention, as well as opportunities for earlier HIV testing.^23,24

Using population-based administrative and public health databases linked to clinical data, our study aimed to explore the healthcare patterns of individuals living with HIV prior to their initiation into HIV care. By doing so, we aimed to understand missed opportunities for improved screening and earlier diagnosis for those individuals living with HIV, and to identify potential points of contact for prevention and care interventions.

Methods

Study setting and population

We conducted this study as a part of the Missed Opportunities for Diagnoses Epidemiological Study (MODES), a retrospective study using population-based administrative databases linked to HIV clinical data. The two clinical sites of the Manitoba HIV Program (MHP), based in Winnipeg, the capital city of the province of Manitoba, provided care to almost all (i.e., >95%) adults living with HIV in Manitoba. As of 2018, there were approximately 1400 persons living with HIV receiving care through the MHP.⁵ Manitoba Health, Seniors, and Active Living (MHSAL) is the single insurer responsible for payment of health services. Information on physician claims, hospital discharge abstracts and pharmaceutical dispensations is contained within population-based administrative datasets, managed and maintained by MHSAL.

Ethical approval was obtained from the Institutional Review Boards (IRB) at the University of Manitoba and the Winnipeg Regional Health Authority. This study was also approved by the Assembly of Manitoba Chiefs’ Health Information Research Governance Committee and the Health Information Privacy Committee of MHSAL.

Participants

Similar to a previously published case-control study,²³ cases were defined as individuals living with HIV, 18 years of age and older, who were newly enrolled in the MHP between 1 January 2007 and 31 December 2011. Recruitment took place from 1 January 2013 to 31 December 2013. From a potential population of 303 persons enrolled in the MHP within the study period, 208 were eligible for the study. Of these 208, 186 consented to participate in the study (consent rate = 76%). An additional 15 patients who were deceased at the time of study recruitment were added to the 186 patients for a final sample of 201 individuals. Ethics board review waived the need for informed consent for these 15 individuals. Of the 201 individuals, 193 were continuously enrolled in the 5-year period prior to case date and included in the final analysis. Case date was defined as the date the case enrolled in the MHP.

Controls were HIV-negative individuals drawn from the general population who were age-, sex- and region-matched to the cases at a 5:1 ratio. Using a previously published Canadian algorithm,²⁵ controls were deemed “HIV-negative” if they had no existing claims for ICD-9-CM codes [042 or 043] or ICD-10-CA codes [Z20] in physician and hospital claims.

Data sources

Scrambled (i.e., de-identified) personal health identification numbers (PHINs) of cases and controls were used to link the following datasets.

Clinical data

Clinical data were obtained on individuals living with HIV who had provided written informed consent and were abstracted from MHP sites by a research assistant and reviewed by MLB and KK, using a standard data abstraction tool. Information included basic demographic information, date of entry into the MHP, risk factors associated with HIV acquisition, and selected clinical characteristics.

Manitoba health insurance registry

Since 1970, Manitoba Health, Seniors and Active Living (MHSAL) has maintained the Manitoba Health Insurance Registry (“The Registry”) to track individuals registered to receive health services in Manitoba. Date fields for registration, birth, entry into province, and migration in/out of province provide information to track residents for longitudinal analyses, and are updated annually.²⁶

Physician claims and hospital discharge abstract data

Universal health coverage is provided by MHSAL and computerized records of all physician billing claims and hospitalizations are maintained by MHSAL. Each physician claim and hospitalization includes diagnoses coded using the International Classification of Diseases (ICD) system. For outpatient physician claims, only one diagnosis is entered into MHSAL databases, with diagnoses coded up to three digits using ICD-9-CM. For inpatient hospitalizations, multiple diagnoses can be coded per hospital visit; ICD-9-CM was used up to fiscal year 2004/2005, at which point the coding system was changed to ICD-10-CA. For both time periods, ICD codes were coded to the 5-digit level. The accuracy of these administrative health databases has been demonstrated.^27–29

Drug program information network data

Information on all outpatient prescription dispensations is collected in the Drug Program Information Network (DPIN) dataset, which has been online since fiscal year 1995/1996. Within the DPIN, pharmaceutical products are identified by their unique drug identification numbers (DIN).²⁹

Sexually transmitted infections surveillance data

MHSAL maintains the STI Surveillance Database, which contains socio-demographic, diagnostic, and treatment information on all positive cases of Chlamydia trachomatis (CT) and Neisseria gonorrheae (GC) infection for the province, regardless of where testing took place.³⁰ CT and GC infections must be reported to MHSAL by the testing laboratory or healthcare provider. Information relevant for this study includes specimen collection date and type of infection (CT or GC).

Measures

With the exception of bacterial STIs, administrative data were searched for any evidence of healthcare utilization in the 2-year period prior to this date for both cases and controls. A 5-year period was used for bacterial STIs, due to the low number of infections recorded.

Physician visits

Physician visits prior to case date were determined using date of physician visit. Only unique visits, based on visit date and ICD-9-CM code were counted. Diagnoses were categorized according to ICD-9-CM chapters, with binary indicators denoting the presence of a diagnosis for any particular chapter. Any inpatient records (i.e., physician claims during hospital stays) were not counted.

Hospitalizations

Hospitalization prior to case date was defined by hospital admission date, and a binary indicator denoting a hospital visit created, with visits lasting more than 24 h included as a hospitalization.

Pharmaceutical dispensations

Dispensation date was used to determine pharmaceutical dispensations prior to case date; using DIN, and based on the work of Marrie et al. and Shaw et al.,^31–33 the following prescription categories were created: antibiotics, asthma, diabetes, heart disease, hyperlipidemia, hypertension, respiratory illness, anxiety, bipolar disorder, depression, mood disorders, and schizophrenia.

Chlamydia/Gonorrhea infections

Using the MHSAL STI surveillance data, separate binary indicators were created for CT and GC, indicating at least one positive CT or GC test in the 5 years prior to case date. For all measures and to address confounding due to indication bias (i.e., co-occurrence of healthcare utilization with an HIV diagnosis), a washout period of 180 days prior to case date was used. Thus, for both cases and controls, any utilization within a 180-day period prior to the case date was not counted. Additionally, entry into MHP was chosen as ‘case date’ instead of the date of first positive HIV result because not all individuals living with HIV have a previous positive HIV test on record in Manitoba, prior to their entry into MHP. Analysis of individuals with a previous positive HIV test found that the average gap between a positive result and entry into care was less than 60 days.

Statistical methods

Only individuals who were continuously registered (i.e., no evidence of out-migration) and had a first registered date at least five years prior to the study period with the MHSAL registry for the duration of the study period were included in the study. Conditional logistic regression was used to test for the association between HIV status and the likelihood of each measure of healthcare utilization; odds ratios (ORs) and 95% confidence intervals (95% CIs) are reported. All analyses were performed using Stata 13 (College Station, TX).

Results

Of the 208 persons eligible for the study, 186 consented to participate (89.4%). A further 15 individuals who were deceased at the time of study recruitment were included for a final sample of 201 individuals. The IRB waived the need for informed consent for these 15 individuals. However, of these 201 participants, eight were excluded as they did not have evidence of continuous insurance coverage in Manitoba, resulting in a final sample size of 193 individuals in the case cohort (individuals living with HIV) and 965 individuals in the control cohort (HIV negative individuals). Table 1 shows the socio-demographic characteristics of cases and controls. The median age at time of entry into the MHP was 38 years (IQR: 30–46) and 34% of cases were women. Almost all participants lived in Winnipeg (94%).

Table 1.

Selected characteristics of participants.

	Controls (N = 965)	%	Cases (N = 193)	%	Total (N = 1158)	%
Age at case date (median, IQR^a)	38	(30–46)	38	(29–46)	38	(30–46)
Female sex (N, %)	330	34	66	34	396	34
Resides in winnipeg (N, %)	—	—	178	(94%)	—	—
Transmission risk (N, %)^b
Heterosexual sex	—	—	133	(69%)	—	—
MSM^c	—	—	47	(24%)	—	—
PWID^d	—	—	26	(13%)	—	—
Endemic exposure	—	—	28	(15%)	—	—
Infection from a bacterial STI^e in the 5-year period prior to case date
Chlamydia	19	(2%)	11	(6%)	60	(3%)
Gonorrhea	2	(0.2%)	10	(5%)	37	(1%)
Physician visits—2 years prior to diagnosis, by ICD-9-CM chapter^f
Chapter 1: Infectious condition
0	846	87.7	138	71.5	984	85.0
1+	119	12.3	55	28.5	174	15.0
Chapter 2: Neoplasm
0	924	95.8	177	91.7	1101	95.1
1+	41	4.2	16	8.3	57	4.9
Chapter 3: Endocrine
0	825	85.5	161	83.4	986	85.1
1+	140	14.5	32	16.6	172	14.9
Chapter 4: Blood disorders
0	944	97.8	179	92.7	1123	97.0
1+	21	2.2	14	7.3	35	3.0
Chapter 5: Mental health
0	713	73.9	124	64.2	837	72.3
1+	252	26.1	69	35.8	321	27.7
Chapter 6: Nervous system
0	766	79.4	131	67.9	897	77.5
1+	199	20.6	62	32.1	261	22.5
Chapter 7: Circulatory system
0	834	86.4	168	87.0	1002	86.5
1+	131	13.6	25	13.0	156	13.5
Chapter 8: Respiratory condition
0	641	66.4	97	50.3	738	63.7
1+	324	33.6	96	49.7	420	36.3
Chapter 9: Digestive system
0	819	84.9	141	73.1	960	82.9
1+	146	15.1	52	26.9	198	17.1
Chapter 10: Genito-urinary system
0	793	82.2	150	77.7	943	81.4
1+	172	17.8	43	22.3	215	18.6
Chapter 11: Pregnancy-related
0	927	96.1	190	98.4	1117	96.5
1+	38	3.9	3	1.6	41	3.5
Chapter 12: Skin disorders
0	788	81.7	121	62.7	909	78.5
1+	177	18.3	72	37.3	249	21.5
Chapter 13: Musculo-skeletal system
0	716	74.2	133	68.9	849	73.3
1+	249	25.8	60	31.1	309	26.7
Chapter 14: Congenital condition
0	955	99.0	189	97.9	1144	98.8
1+	10	1.0	4	2.1	14	1.2
Chapter 15: Perinatal condition
0	963	99.8	193	100.0	1156	99.8
1+	2	0.2	0	0.0	2	0.2
Chapter 17: Symptoms, ill-defined
0	676	70.1	116	60.1	792	68.4
1+	289	29.9	77	39.9	366	31.6
Chapter 18: Injury and poisoning
0	730	75.6	129	66.8	859	74.2
1+	235	24.4	64	33.2	299	25.8
Chapter 19: Supplementary—health services contacts
0	545	56.5	115	59.6	660	57.0
1+	420	43.5	78	40.4	498	43.0
Any physician visit
0	518	53.7	66	34.2	584	50.4
1+	444	46.3	127	65.8	574	49.6
Hospitalizations
0	894	92.6	164	85.0	1058	91.4
1+	71	7.4	29	15.0	100	8.6
Pharmaceutical dispensations
Antibiotics
0	547	56.6	69	35.8	616	53.2
1+	418	43.4	124	64.2	542	46.8
Medication for asthma
0	851	88.2	162	83.9	1013	87.5
1+	114	11.8	31	16.1	145	12.5
Medication for diabetes
0	913	94.6	175	90.7	1088	94.0
1+	52	5.4	18	9.3	70	6.0
Medication for heart disease
0	953	98.8	190	98.4	1143	98.7
1+	12	1.2	3	1.6	15	1.3
Medication for hyperlipidemia
0	898	93.0	181	93.8	1079	93.2
1+	67	7.0	12	6.2	79	6.8
Medication for hypertension
0	842	87.2	168	87.0	1010	87.2
1+	123	12.8	25	13.0	148	12.8
Medication for respiratory conditions
0	849	88.0	162	83.9	1011	87.3
1+	116	12.0	31	16.1	147	12.7
Medication for anxiety
0	895	92.7	161	83.4	1056	91.2
1+	70	7.3	32	16.6	102	8.8
Medication for bipolar diagnoses
0	948	98.2	188	97.4	1136	98.1
1+	17	1.8	5	2.6	22	1.9
Medication for depression
0	848	87.9	170	88.1	1018	87.9
1+	117	12.1	23	11.9	140	12.1
Medication for mood disorders
0	829	85.9	159	82.4	988	85.3
1+	136	14.1	34	17.6	170	14.7
Medication for schizophrenia
0	921	95.4	183	94.8	1104	95.3
1+	44	4.6	10	5.2	54	4.7

^aIQR: interquartile range.

^bTransmission risk is not mutually exclusive category so does not total up to 100%.

^cMSM: men who have sex with men.

^dPWID: people who inject drugs.

^eSTI: sexually transmitted infection; case date is the day the HIV client entered HIV care; utilization for both HIV-positive individuals and their controls were back-calculated from this date.

^fICD-9-CM: International Classification of Diseases, Ninth Revision, Clinical Modification.

Physician visits

Overall, individuals living with HIV were significantly more likely to have visited a physician at least once in the 2-year period prior to case date as compared to HIV uninfected controls (AOR: 1.7, 95% CI: 1.2–2.8). With respect to specific diagnoses, differences between cases and controls varied by ICD-9-CM chapter, as shown in Table 2. In the 2-year period prior to case date, the largest odds ratios were observed for “Blood and Blood Disorders” (AOR: 4.2, 95% CI: 2.0–9.0), “Infectious Conditions” (AOR: 3.6, 95% CI: 2.3–5.5), and “Skin Disorders” (AOR: 3.1, 95% CI: 2.1–4.4). Examples of “Blood and Blood Disorders” include anemias and hemorrhagic conditions; “Infectious Conditions” include a variety of infectious diseases; while examples of “Skin Disorders” include disorders specific to the skin—cellulitis, impetigo, cysts, and dermatitis. Table 3 shows the top five most frequent diagnoses related to “Infectious Conditions” and “Blood and Blood Disorders.”

Table 2.

One or more visits, by ICD-9-CM^a chapter and hospital visits, cases versus controls, adjusted odds ratios (AOR), and 95% confidence intervals (95% CI) from conditional logistic regression (2 years prior to HIV case date) adjusted for days registered, 180-day washout period.

ICD-9-CM chapter	AOR (95% CI)
Infectious condition	3.59
Infectious condition	(2.34–5.49)**
Neoplasm	2.21
Neoplasm	(1.17–4.20)*
Endocrine	1.31
Endocrine	(0.81–2.10)
Blood and blood disorders	4.21
Blood and blood disorders	(1.97–8.97)**
Mental health	1.76
Mental health	(1.24–2.51)**
Disorders of the nervous system	2.06
Disorders of the nervous system	(1.41–3.00)**
Circulatory system	0.95
Circulatory system	(0.57–1.59)
Respiratory condition	2.44
Respiratory condition	(1.72–3.47)**
Digestive system	2.39
Digestive system	(1.61–3.56)**
Genito-urinary system	1.32
Genito-urinary system	(0.86–2.04)
Pregnancy-related	0.33
Pregnancy-related	(0.10–1.13)
Skin and skin disorders	3.05
Skin and skin disorders	(2.11–4.41)**
Musculo-skeletal system	1.42
Musculo-skeletal system	(0.98–2.05)
Congenital condition	2.76
Congenital condition	(0.76–10.07)
Perinatal condition	0.00
Perinatal condition	(0.00 -.)
Symptoms, Ill-Defined	1.60
Symptoms, Ill-Defined	(1.14–2.26)**
Injury and poisoning	1.78
Injury and poisoning	(1.24–2.54)**
Supplementary health services contacts	0.98
Supplementary health services contacts	(0.69–1.39)
1+ hospitalizations	2.24
1+ hospitalizations	(1.41–3.58)**

*p < .05; ^**p < .01.

^aICD-9-CM: International Classification of Diseases, Ninth Revision, Clinical Modification.

Table 3.

Top five most frequent Chapter 1 and Chapter 4 diagnoses from physician visits, HIV-positive participants (2 years prior to HIV case date), and 180-day washout period.

Diagnosis	Freq (%)
Infectious condition (chapter 1, ICD-9-CM^a)
Other venereal disease	14.5
Pediculosis and pthirus	12.6
Viral infection in other disease/not otherwise specified	12.6
Herpes simplex	11.3
Other viral disease	8.8
All others	40.2
Blood disorders (chapter 4, ICD-9-CM)^a
Iron deficiency anemias	24.1
Other blood disease	20.7
Anemia, not otherwise specified	17.2
Other deficiency anemia	17.2
Aplastic anemia	13.8
All others	7.0

^aInternational Classification of Diseases, Ninth Revision, Clinical Modification.

Hospitalizations

Hospital visits were also significantly associated with HIV status; those with one or more hospital visits in the 2-year period prior to case date were at over twice the odds of being HIV positive (AOR: 2.2, 95% CI: 1.4–3.6; Table 2).

Prescription use

Prescription dispensation was associated with HIV status, with the strength and direction of association dependent on type of pharmaceutical. In the 2-year period prior to case date, prescriptions for the following were associated with being HIV-positive (Table 4): antibiotics (AOR: 3.4, 95% CI: 2.3–4.9), asthma (AOR: 1.8, 95% CI: 1.1–2.9), diabetes (AOR: 2.1, 95% CI: 1.1–4.0), respiratory conditions (AOR: 1.8, 95% CI: 1.1–2.8), and anxiety disorders (AOR: 2.9, 95% CI: 1.8–4.8).

Table 4.

One or more dispensations, by type of prescription, cases versus controls, adjusted odds ratios (AOR), and 95% confidence intervals (95% CI) from conditional logistic regression (2 years prior to HIV case date) adjusted for days registered, 180-day washout period.

	AOR (95% CI)
Antibiotics	3.35
Antibiotics	(2.29–4.90)**
Asthma	1.82
Asthma	(1.13–2.93)*
Diabetes	2.08
Diabetes	(1.09–3.97)*
Heart disease	2.07
Heart disease	(0.52–8.32)
Hyperlipidemia	1.01
Hyperlipidemia	(0.49–2.08)
Hypertension	1.13
Hypertension	(0.66–1.94)
Respiratory	1.77
Respiratory	(1.10–2.84)*
Anxiety	2.92
Anxiety	(1.79–4.76)**
Bipolar disorder	1.43
Bipolar disorder	(0.48–4.25)
Depression	1.28
Depression	(0.77–2.11)
Mood disorder	2.37
Mood disorder	(1.64–3.43)**
Schizophrenia	1.35
Schizophrenia	(0.63–2.90)

*p < .05; **p < .01.

Bacterial Sexually transmitted infections

Persons living with HIV (cases) were more likely to have had at least one chlamydial (AOR: 3.1, 95% CI: 1.6–5.9; 180-day washout period) or one gonorrheal infection (AOR: 11.8, 95% CI: 3.6–13.4; 180-day washout period) in the 5-year period prior to HIV case date, as compared to HIV uninfected controls (Table 5).

Table 5.

Bacterial infection and HIV, odds ratios (OR), and 95% confidence intervals (95% CI) from conditional logistic regression.

		OR (95% CI)
Model I^a	CT^b	3.1 (1.6–5.9)
Model I^a	GC^c	11.8 (3.6–13.4)
Model II^d	CT^b	3.1 (1.5–6.4)
Model II^d	GC^c	10.4 (4.2–25.5)

^a180-day washout period.

^bCT: chlamydia.

^cGC: gonorrhea.

^d365-day washout period.

Discussion

Building on earlier work demonstrating the higher likelihood of serological tests among individuals eventually diagnosed with HIV,²³ our results demonstrate that healthcare utilization was higher among those who were subsequently diagnosed with HIV, compared to their controls. Our analyses also reveal that utilization patterns differ amongst those diagnosed with HIV. Persons living with HIV were more likely to have diagnoses related to infectious conditions, and blood and blood disorders. They were more likely to have received dispensations for antibiotics, and for medications used in the treatment of diabetes and mood disorders. Individuals living with HIV were at higher odds of diagnosed bacterial STIs, with the odds of a GC infection particularly high. These analyses indicate that opportunities exist for earlier diagnosis and screening for those persons living with HIV and suggest that earlier interactions may serve as points of contacts for prevention/intervention opportunities. For example, more active follow-up of novel gonorrhea cases may help prevent subsequent HIV infection.

With respect to the association between diagnoses of blood and blood disorders and subsequent HIV, the MHP gets referrals for patients presenting with lymphadenopathy, anemia and/or thrombocytopenia; disorders associated with HIV.^34–36 A more precise predictive decision-making tool incorporating these disorders in clinical practice may lead to earlier diagnoses of HIV. In describing the co-occurring epidemics of HIV, and substance use, Singer et al. were one of the first to place HIV within a “syndemics” framework,^37–39 recognizing that risk and vulnerability of individuals to HIV occurs in environments of harmful social/structural conditions. Thus, these harmful conditions contribute to poorer health outcomes, including HIV^40,41; the higher use of healthcare services of cases in our study aligns with this perspective.

Our findings can contribute to frameworks used to examine missed opportunities for HIV diagnoses.²³ Several factors have been proposed to explain the heterogeneity in missed opportunities for HIV diagnosis. For example, availability of services has been acknowledged as a barrier to services, such as where socio-economically disadvantaged and marginalized populations are more likely to reside in areas where preventive and screening services are not available.⁴² Conversely, accessibility of services is a factor in situations where various individual, practice-based, or structural barriers deter an individual or certain sub-populations from accessing available services.⁴³ Our study showed that within the context of a publicly-funded healthcare system, HIV-positive individuals were accessing health services at a higher rate than their controls, prior to their diagnosis. Thus, despite the known marginalization of populations that are at risk for HIV,^41,44 barriers deterring this population of HIV-positive individuals from seeking healthcare were overcome. Still unknown is the nature of healthcare sought; whether the care sought was consistent and comprehensive, or a more episodic/“chaotic,” multi-provider type of care.^45,46

More research to understand the structural and practice level issues contributing to gaps in prevention and care services may help to optimize HIV prevention.⁴⁷ For example, in Manitoba, responsibility for follow-up of individuals testing positive for CT or GC, including partner notification, is left to the discretion of the testing practitioner, who may refer follow-up to public health. Although follow-up screening with individuals testing positive for CT or GC is recommended in the Canadian Guidelines on STIs,⁴⁸ there is very little information available on how well follow-up recommendations are implemented. Evidence suggests that placing the onus of follow-up on practitioners has resulted in an uncoordinated and passive system.⁴⁹ More effective coordination between public health and general practitioners can lead to better navigation of the healthcare system, and improved delivery of prevention interventions.⁵⁰ Without better coordination, those at highest risk of STIs may continue to be denied access to testing, and among those that do, a proportion may continue to fall through cracks in the healthcare system.

At the same time, individual-level behavioral factors are also likely at play⁵¹; strengthening community-based prevention efforts, such as outreach to high-risk individuals and other priority populations, and which focus on health promotion, a more enabling environment, and addressing other needs such as housing, mental health and addictions issues may also reduce gaps in prevention.^52,53 Our findings illustrating the higher likelihood of dispensations for mental health conditions among individuals with HIV supports prioritization of mental health services. Community organizations and primary and public healthcare have historically been underfunded in Canada.⁴⁴

Strengths and limitations

Our study had a number of strengths, including the population-based nature of the MHP, as well as empirical data on healthcare utilization. Although we sampled only a portion of the total MHP population available, the socio-demographic characteristics of our sample did not differ substantially from other publications using the entire MHP population.⁵⁴ Our control group allowed for comparisons to the general population.

Our study had some limitations. CT/GC incidence is likely underestimated as data collected by MHSAL is based on a passive surveillance system. However, since populations at higher risk for HIV and other STIs are less likely to access care, this underestimation should have driven differences between cases and controls to the null. Case date was defined as the date of MHP entry; this date was chosen instead of the date of first positive HIV result as not all HIV-positive individuals had a previous positive HIV test on record in Manitoba prior to their entry into MHP. This gap between first positive HIV test and entry into MHP care could have potentially biased rates of CT/GC incidence and testing; however, among those with a previous positive HIV test, the average gap between a positive result and entry into care was less than 60 days. Thus, our 180-day washout period was meant to err on the side of caution. Controls were assumed to be HIV-negative; but are more accurately referred to as not known to be positive.

In conclusion, our results demonstrate distinct patterns of healthcare utilization amongst HIV-positive individuals, prior to their HIV diagnosis. Stronger efforts to integrate the work of public health personnel, community organizations, and general practitioners are needed in the Canadian context to improve HIV prevention efforts, as well as support earlier HIV testing and linkage to care for those who test positive.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Gilead Foundation. SYS is supported by a Canada Research Chair in Program Science and Global Public Health (Tier II).

ORCID iDs

Souradet Y Shaw

Marissa L Becker

References

Krentz

Kliewer

Gill

. Changing mortality rates and causes of death for HIV-infected individuals living in Southern Alberta, Canada from 1984-2003. HIV Med 2005; 6: 99–106.

Palella

Jr. Delaney

Moorman

, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998; 338: 853–860.

Egger

May

Chene

, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002; 360: 119–129.

Girardi

Sabin

Monforte

. Late diagnosis of HIV infection: epidemiological features, consequences and strategies to encourage earlier testing. J Acquir Immune Defic Syndr 2007; 46 (Suppl 1): s3–s8.

Manitoba HIV Program . 2016 Manitoba HIV Program Update. Manitoba: Winnipeg, 2017.

Sterne

May

Costagliola

, et al. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. Lancet 2009; 373: 1352–1363. DOI: 10.1016/s0140-6736(09)60612-7.

Kitahata

Gange

Abraham

, et al. Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med 2009; 360: 1815–1826.

Sellers

Wohl

. Antiretroviral therapy: when to start. Infect Dis Clin North Am 2014; 28: 403–420. DOI: 10.1016/j.idc.2014.05.004.

Williams

Churchill

Anderson

, et al. British HIV Association guidelines for the treatment of HIV-1-positive adults with antiretroviral therapy 2012. HIV Med 2012; 13 Suppl 2: 1–85. DOI: 10.1111/j.1468-1293.2012.01029.x.

10.

Department of Health and Human Services, Panel on Antiretroviral Guidelines for Adults and Adolescents . 2021. Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents with HIV. Department of Health and Human Services. Available at https://clinicalinfo.hiv.gov/sites/default/files/guidelines/documents/AdultandAdolescentGL.pdf

11.

Cohen

Chen

McCauley

, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med 2011; 365: 493–505. DOI: 10.1056/NEJMoa1105243.

12.

Public Health Agency of Canada . HIV/AIDS Epi Update July 2010. Ottawa: Public Health Agency of Canada, 2010.

13.

Krentz

Auld

Gill

. The high cost of medical care for patients who present late (CD4<200 cells/μL) with HIV infection. HIV Med 2004; 5: 93–98.

14.

Krentz

Gill

. The direct medical costs of late presentation (<350/mm) of HIV infection over a 15-year period. AIDS Res Treat 2012; 2012: 757135. DOI: 10.1155/2012/757135.

15.

Fisher

. Late diagnosis of HIV infection: major consequences and missed opportunities. Curr Op Inf Dis 2008; 21: 1–2.

16.

May

Gompels

Delpech

, et al. Impact of late diagnosis and treatment on life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study. BMJ 2011; 343: d6016.

17.

Nakagawa

Lodwick

Smith

, et al. Projected life expectancy of people with HIV according to timing of diagnosis. AIDS 2011; 26(3): 335–343. DOI: 10.1097/QAD.0b013e32834dcec9.

18.

Sudarshi

Pao

Murphy

, et al. Missed opportunities for diagnosing primary HIV infection. Sex Transm Infect 2008; 84: 14–16.

19.

Duffus

Weis

Kettinger

, et al. Risk-based HIV testing in South Carolina health care settings failed to identify the majority of infected individuals. AIDS Patient Care STDS 2009; 23: 339–345.

20.

Burns

Johnson

Nazroo

, et al., on behalf of the SONHIA Collaboration Group. Missed opportunities for earlier HIV diagnosis within primary and secondary healthcare settings in the UK. AIDS 2008; 22: 115–122.

21.

Sudarshi

Pao

Murphy

, et al. Missed opportunities for diagnosing primary HIV infection. Sex Transm Infect 2008; 84: 14–16.

22.

Bernstein

Marcus

Nieri

, et al. Rectal gonorrhea and chlamydia reinfection is associated with increased risk of HIV seroconversion. J Acquir Immune Defic Syndr 2010; 53: 537–543. DOI: 10.1097/QAI.0b013e3181c3ef29.

23.

Shaw

Ireland

McClarty

, et al. Prior history of testing for syphilis, hepatitis B and hepatitis C among a population-based cohort of HIV-positive individuals and their HIV-negative controls. AIDS Care 2017; 29(1): 67–72. DOI: 10.1080/09540121.2016.1200715.

24.

Wejnert

Prejean

Hoots

, et al. Prevalence of missed opportunities for HIV testing among persons unaware of their infection. JAMA 2018; 319: 2555–2557. DOI: 10.1001/jama.2018.7611.

25.

Antoniou

Zagorski

Loutfy

, et al. Validation of case-finding algorithms derived from administrative data for identifying adults living with human immunodeficiency virus infection. PLoS One 2011; 6: e21748. DOI: 10.1371/journal.pone.0021748.

26.

Roos

Nichol

. A research registry: uses, development, and accuracy. J Clin Epi 1999; 52: 39–47.

27.

Blanchard

Ludwig

Wajda

, et al. Incidence and prevalence of diabetes in Manitoba, 1986-1991. Diabetes Care 1996; 19: 807–811.

28.

Bernstein

Blanchard

Rawsthorne

, et al. Epidemiology of Crohn's Disease and ulcerative colitis in a central Canadian province: a population-based study. Am J Epidemiol 1999; 149: 916–924.

29.

Lix

Yogendram

Mann

. Defining and Validating Chronic Diseases: An Administrative Data Approach - an Update with ICD-10-CA. Manitoba Centre for Health Policy, 2008.

30.

Wylie

Jolly

. Patterns of chlamydia and gonorrhea infection in sexual networks in Manitoba, Canada. Sex Transm Dis 2001; 28: 14–24.

31.

Marrie

Leung

, et al. Rising prevalence of vascular comorbidities in multiple sclerosis: validation of administrative definitions for diabetes, hypertension, and hyperlipidemia. Mult Scler 2012; 18: 1310–1319. DOI: 10.1177/1352458512437814.

32.

Marrie

Elliott

Marriott

, et al. Effect of comorbidity on mortality in multiple sclerosis. Neurology 2015; 85: 240–247. DOI: 10.1212/WNL.0000000000001718.

33.

Shaw

Blanchard

Bernstein

. Association between the use of antibiotics in the first year of life and pediatric inflammatory bowel disease. Am J Gastroenterol 2010; 105: 2687–2692. DOI: 10.1038/ajg.2010.398.

34.

Nag

Dey

Nandi

, et al. Etiological study of lymphadenopathy in HIV-infected patients in a tertiary care hospital. J Cytol 2016; 33: 66–70. DOI: 10.4103/0970-9371.182518.

35.

Kaplan

Spira

Fishbein

, et al. A six-year follow-up of HIV-infected homosexual men with lymphadenopathy. Evidence for an increased risk for developing AIDS after the third year of lymphadenopathy. JAMA 1988; 260: 2694–2697.

36.

Passos

Treitinger

Spada

. An overview of the mechanisms of HIV-related thrombocytopenia. Acta Haematol 2010; 124: 13–18. DOI: 10.1159/000313782.

37.

Singer

. A dose of drugs, a touch of violence, a case of AIDS: conceptualizing the SAVA syndemic. Free Inquity Creat Sociol 1996; 24: 99–110.

38.

Singer

Clair

. Syndemics and public health: reconceptualizing disease in bio-social context. Med Anthropol Q 2003; 17: 423–441.

39.

Singer

Erickson

Badiane

, et al. Syndemics, sex and the city: understanding sexually transmitted diseases in social and cultural context. Soc Sci Med 2006; 63: 2010–2021. DOI: 10.1016/j.socscimed.2006.05.012.

40.

Rhodes

. Risk theory in epidemic times: sex, drugs, and the social organization of "risk behaviour". Soc Health Ill 1997; 19: 208–227.

41.

Rhodes

Singer

Bourgois

, et al. The social structural production of HIV risk among injecting drug users. Soc Sci Med 2005; 61: 1026–1044.

42.

Fulcher

Kaukinen

. Mapping and visualizing the location HIV service providers: an exploratory spatial analysis of Toronto neighborhoods. AIDS Care 2005; 17: 386–396. DOI: 10.1080/09540120512331314312.

43.

Rhodes

Kimber

Small

, et al. Public injecting and the need for 'safer environment interventions' in the reduction of drug-related harm. Addiction 2006; 101: 1384–1393. DOI: 10.1111/j.1360-0443.2006.01556.x.

44.

Lewis

A system in name only--access, variation, and reform in Canada's provinces. N Engl J Med 2015; 372: 497–500. DOI: 10.1056/NEJMp1414409.

45.

Kim

Kertesz

Horton

, et al. Episodic homelessness and health care utilization in a prospective cohort of HIV-infected persons with alcohol problems. BMC Health Serv Res 2006; 6: 19. DOI: 10.1186/1472-6963-6-19.

46.

Kendall

Manuel

Younger

, et al. A population-based study evaluating family physicians' HIV experience and care of people living with HIV in Ontario. Ann Fam Med 2015; 13: 436–445. DOI: 10.1370/afm.1822.

47.

Church

Mayhew

. Integration of STI and HIV prevention, care, and treatment into family planning services: a review of the literature. Stud Family Planning 2009; 40: 171–186.

48.

Public Health Agency of Canada . Canadian Guidelines on Sexually Transmitted Infections, http://www.phac-aspc.gc.ca/std-mts/sti-its/cgsti-ldcits/index-eng.php (2015, accessed 5 February 2015).

49.

Shoveller

Johnson

Rosenberg

, et al. Youth's experiences with STI testing in four communities in British Columbia, Canada. Sex Transm Infect 2009; 85: 397–401. DOI: 10.1136/sti.2008.035568.

50.

Thompson

Mugavero

Amico

, et al. Guidelines for improving entry into and retention in care and antiretroviral adherence for persons with HIV: evidence-based recommendations from an International Association of Physicians in AIDS Care panel. Ann Intern Med 2012; 156: 817–833, W-284, W-285, W-286, W-287, W-288, W-289, W-290, W-291, W-292, W-293, W-294. DOI: 10.7326/0003-4819-156-11-201206050-00419.

51.

Boerma

Weir

. Integrating demographic and epidemiological approaches to research on HIV/AIDS: the proximate-determinants framework. J Infect Dis 2005; 191(Suppl 1): S61–S67. DOI: 10.1086/425282.

52.

Mayer

Bekker

Stall

, et al. Comprehensive clinical care for men who have sex with men: an integrated approach. Lancet 2012; 380: 378–387. DOI: 10.1016/S0140-6736(12)60835-6.

53.

Remien

Bauman

Mantell

, et al. Barriers and facilitators to engagement of vulnerable populations in HIV primary care in New York City. J Acquir Immune Defic Syndr 2015; 69(Suppl 1): S16–S24. DOI: 10.1097/QAI.0000000000000577.

54.

Erickson

Becker

Shaw

, et al. Substance use and its impact on care outcomes among HIV-infected individuals in Manitoba. AIDS Care 2015; 27(9): 1168–1173.