Costs associated with incorporation of network approaches into STD program activities

Abstract

Background:

Network approaches can be used to study sociosexual partnerships and identify individuals at high risk of infection. Little is known about the cost structure of these services and their association with STD diagnoses.

Method:

We collected costs associated with using a peer network strategy to recruit MSM and transwomen of color in 4 counties in North Carolina: Guilford, Forsyth, Durham, and Wake from February through October 2019. We used a comprehensive costing approach to gather detailed retrospective information on the intervention cost, broken down by category and programmatic activity.

Results:

The sociosexual networks collected consisted of 31 initial seeds (index cases) and 49 peers of those seeds. In peers, 5 cases of HIV and 10 cases of syphilis were identified. The cost per case (HIV or syphilis) identified was $7,325. Personnel costs accounted for 80% of total expenditures, followed by laboratory expenses (12%). Personnel cost was distributed between DIS patient navigators (51%), non-clinical (37%), and management (12%) staff. General administration was the costliest programmatic activity (37%), followed by case management and field services (37%), and study activities (11%). The estimated average cost per patient tested was $2,242.

Conclusions:

Finding positive peer cases in non-clinical settings is costly but may be crucial for limiting the spread of sexually transmitted diseases. The cost of staff was the major driver. This study demonstrates that using a network strategy can be a cost-effective way to identify, test, and refer patients at high-risk of syphilis and HIV infections to care.

Short Summary:

The study demonstrates that a network strategy conducted by DIS patient navigators can be a cost-effective way to identify, test, and refer patients at high-risk of syphilis and HIV infections.

Introduction

HIV and syphilis are costly infections that have had steady or rising rates in North Carolina in recent years.^{1, 2} Partner notification conducted by disease intervention specialists (DIS), a basic approach for identifying individuals at high risk of sexually transmitted infections (STIs) from sexual networks, has served as a core element of syphilis and HIV prevention for decades.³ DIS carry out critical services for patients identified with HIV and syphilis, which include individual risk-reduction counseling, referrals for care, and partner services to prevent further transmission of HIV and syphilis. Although DIS have conducted essential public health interventions for decades, their roles must evolve along with changes in technology and STI epidemiology.⁴ Health departments and community based organizations have previously explored social network strategies to engage persons with HIV infection or at risk for infection to refer or recruit individuals from their social and sexual networks for testing.^{5, 6} In addition, changes in communication technologies have changed the way social and sexual networks form, while biomedical advances, such as pre-exposure prophylaxis (PrEP), have offered new opportunities for HIV prevention. At the same time, cuts in funding to public health departments and concurrent rises in STIs have led to increased caseloads for DIS workers and stagnant pay.⁷ Recently, new federal funding for DIS was announced that has the potential to transform the DIS workforce.⁸

In 2018, the North Carolina Division of Public Health implemented the “Combined HIV and STD Prevention and Care for Vulnerable Men who Have Sex with Men (MSM) and Transgender Women (TGW) via Network Methods” (MATRix-NC) project. MATRix-NC was a longitudinal demonstration project built upon the established infrastructure of the state’s DIS program, which follows a traditional model of field and partner services for HIV and syphilis.^{9, 10} DIS were given enhanced training to engage the social and sexual networks of patients recently diagnosed with syphilis and/or HIV in order to facilitate testing, treatment, and referral to specialized services. The program was conducted in four North Carolina counties with a high incidence of HIV and syphilis: Durham, Guilford, Forsyth, and Wake.^10–12

We specifically collected and analyzed the costs of identifying, testing, and providing treatment and other services to persons recruited from the social and sexual networks of index cases with HIV and/or syphilis enrolled in MATRix-NC from February through October 2019. Little is known about the costs of offering enhanced DIS services and their associated benefits through HIV/STI detection and linkage to care. The objective of this demonstration project was to better understand the costs and potential value of specially trained DIS when a network approach for identifying peers is utilized. We estimated the total costs of operating the program and identified the major drivers of cost.

Methods

Three MATRix-NC DIS were specifically trained to recruit and enroll MSM of color and TGW of color who presented for care in STD clinics and were referred for partner services. Eligibility criteria for initial recruits, or seeds, included patients who were diagnosed with acute HIV or new HIV diagnosis; primary or secondary syphilis; early non-primary, non-secondary syphilis (i.e., early latent); or were diagnosed with late or unknown durations of syphilis with an PR/TRUST titer ≥ 1:32. Priority was given to persons eighteen years or older, had ever had oral or anal sex with a man, and lived, worked, or sought medical care in the participating metropolitan statistical area (Durham, Guilford, Forsyth, and Wake counties).

Following informed consent, seed participants were asked to refer “peers” from their social and sexual networks who could benefit via chain-referral recruitment.¹³ Seeds were asked to refer persons they thought could benefit from STI/HIV testing or prevention services and provide them with referral information for the program. All non-infected persons referred by seeds and tested through the MATRix-NC program were counted as peers for the purposes of this study. Peers who met the seed eligibility requirements with HIV and/or syphilis were further classified as “potential” seeds, or “converted” seeds depending upon whether they declined or agreed to participate, respectively. In addition to providing partner services for seeds and testing of peers, MATRix-NC DIS provided patient navigation to link persons living with HIV to care or at-risk persons for PrEP and other social services.⁹ All participants with confirmed HIV infection were referred to an HIV provider for antiretroviral therapy and those with confirmed syphilis were referred to STD or HIV clinics for treatment as per DIS protocols. All of the MATRix-NC services were provided at no cost for study participants, and none of the seeds or peers were billed for HIV/STI testing at the health departments and referral to other services.

The unit of analysis was the program, and no identifiable patient-level information was collected during this study. A comprehensive approach to gathering detailed information on the costs of the service provided was used. Costs were broken down by 8 cost categories (i.e., personnel, laboratory, travel, contracted services, equipment, supplies, utilities, buildings) and 7 programmatic cost activities (i.e., case management & field services for HIV and syphilis, study activities, general administration, disease & treatment education, case review, training, and other activities). Cost estimates used an ingredients-based approach, whereby a unit cost is multiplied by a resource quantity to generate a total cost. Costs of programmatic activities were prorated based on the proportion of time and/or use of each cost element dedicated to a particular activity. A variety of sources were used to collect data on inputs, outputs, and costs, including visual observation of the facilities, inspection of facility records and registers, clinic observation, expert interviews with clinic and project staff, and retrospective review of project accounts.

Costs per new laboratory-confirmed case of HIV or syphilis were also calculated, using total program costs as the numerator and the number of confirmed cases as the denominator. Since costs per case are very sensitive to the number of cases found, we conducted a sensitivity analysis of how key cost outcomes would change at varying levels of positivity. We estimated 95% confidence intervals for HIV and syphilis positivity using a Wilson score with a continuity correction.¹⁴ Total program costs reflected all costs to recruit and test peers, including travel time, fuel, phone use, office use, as well as monetary incentives issued to seeds for each peer recruited ($15 per peer recruited) and time DIS spent with seeds and peers for these activities. An Excel-based cost data collection tool was used to elicit the cost information and the analysis was done using the statistical package Stata (version 14).¹⁵

Results

From February through October of 2019, 31 initial seeds and 49 peers of those seeds were enrolled into the MATRix-NC program. In peers, 5 new laboratory-confirmed cases of HIV and 10 cases of syphilis were identified. In addition, 2 cases of chlamydia and one case of gonorrhea were diagnosed in peers. Among the 49 peers, there were 3 with laboratory confirmed syphilis and/or HIV that agreed to participate further and were converted to seeds. An additional 10 peers identified with confirmed syphilis and/or HIV declined further study participation as seeds but received routine partner services. Reasons for declining to participate included concerns regarding privacy and confidentiality after diagnosis.

Total costs amounted to almost $110,000 during the study period. Due to the nature of the program which involved a separate administrative structure than traditional DIS services, the largest cost categories were personnel at $87,495 (80% of total) followed by laboratory at $13,097 (12%). The supplies, travel, buildings, utilities, and contracted services cost categories each contributed to 3 percent or less of total expenditures (Figure 1). Personnel included administration, DIS patient navigators, and other non-clinical staff. The DIS patient navigator positions accounted for 51% of personnel costs, “other, non-clinical” positions contributed 37% of the personnel costs, and program directors accounted for the remaining 12% of the cost (Figure 1). Other, non-clinical personnel included a project coordinator and research assistant who assisted MATRix-NC DIS with enrollment procedures. When assessing costs by programmatic activity, “general administration” and “case management and field services” were the most substantial components, with costs of $40,951 (38%) and $40,906 (37%), respectively. “Study” activities (i.e. time spent obtaining written informed consent, recruitment incentives)⁹ contributed $13,248 (11%) to overall costs, while “disease and treatment education”, “case review”, “training”, and “other activities” all contributed to 4% or less of overall costs (Figure 2).

Figure 1:

Costs by Category and Personnel Position, February – October 2019

All costs are in 2019 U.S. dollars.

Figure 2:

Costs by Programmatic Activity, February – October 2019

All costs are in 2019 U.S. dollars.

The cost per new positive of either HIV or syphilis diagnosed was $7,325 (Table 1). In the sensitivity analysis, HIV positivity ranged from 4.0% to 22.2%, and syphilis positivity ranged from 11.3% to 33.8%. Under this uncertainty, we estimated a range of $4,994 to $10,987 per new HIV or syphilis positive (Table 1).

Table 1:

Patient Outcomes and Cost per Outcome

Outcomes		Sensitivity Analysis Range
Patients tested	49	NA
New positive HIV	5	[2 – 11]
HIV positivity	10.2%	[4.0% – 22.2%]
New positive syphilis	10	[6 – 17]
Syphilis positivity	20.4%	[11.3% – 33.8%]
New positive HIV or syphilis	15	[10 – 22]
HIV or syphilis positivity	30.6%	[19.4% – 44.6%]
Costs
Total program cost	$109,873	NA
Cost per patient tested	$2,242	NA
Cost per positive HIV or syphilis	$7,325	[$4,994 – $10,987]

Initial seeds had already been tested and confirmed positive for HIV and/or syphilis. Peers referred by seeds were all tested as part of the MATRix program. All costs are in 2019 U.S. dollars. Confidence intervals for positivity outcomes calculated using the Wilson score with continuity correction and were rounded to the nearest integer.

Discussion

Our cost analysis only covered 9 months of the MATRix-NC program, and Hurt et al. (2021) provide more complete information on the demographics of participants enrolled in the program.⁹ Consistent with data on HIV/STI patients in North Carolina, they found that MATRix-NC participants generally had lower levels of socioeconomic status.^{9, 11, 12} Compared to initial seeds, peers were more commonly recruited from social networks than from sexual networks, and peers were significantly more likely to be heterosexual, cisgender women. Common service referrals for participants included PrEP, Medicaid, housing and food assistance, HIV-care, and non-HIV primary care. We presented estimates of cost per case diagnosed, but service referrals and linkage to care were additional benefits of the DIS activities.

Partner services are very labor-intensive activities. We found that 80% of the program costs were for personnel, and a majority of the personnel costs were for DIS positions. This suggests that the overhead costs for operating a program like MATRix-NC can be relatively low once the DIS receive their specialized training. Furthermore, program administration costs could have been reduced and the efficiency of peer referrals improved over time by integration of the program with traditional services rather than as a demonstration project requiring enrollment procedures. Other cost components, such as study activities, could become a smaller portion of total costs as the program is adopted more widely. Laboratory costs made up most of the non-personnel costs. This cost will necessarily increase with the number of peers enrolled and tested, but it represents an efficient use of resources since the population tested had a relatively high prevalence of undiagnosed of HIV and syphilis. Seventy-five percent of costs went towards general administration and case management and field services activities.

We found costs per new positive of either HIV or syphilis to be $7,325. This cost per case is similar to recent estimates of the direct medical costs and indirect productivity costs saved by treating one case of syphilis.¹⁶ However, because of the high costs of HIV relative to other STIs,¹ the benefits from diagnosing an HIV-positive participant and connecting them to care are much higher than treating a syphilis case. Farnham et al. (2012) found costs per new HIV diagnoses below $27,296 (adjusted for inflation) to be cost-saving from a program perspective.¹⁷ The cost-saving threshold reflects the point at which the cost of identifying a case equals the expected costs averted from diagnosing that case. Thus, even if this intervention had only diagnosed 5 HIV cases and zero syphilis infections, our cost of $21,975 per new HIV positive alone would meet the cost-savings threshold. Figure 3 shows how uncertainty surrounding positivity impacts our cost-savings findings. The dotted line denotes the cost-saving threshold assuming a value of $7,298 per syphilis case diagnosed and $27,296 per HIV case diagnosed. The area above and to the right of the dotted line represents combinations of syphilis and HIV positivity that would be cost-saving given our estimated total program cost. The area below/left of the blue line is not cost-saving, but may still be cost-effective. The triangular point reflects the pair of syphilis and HIV positivity found in the study, and the error bars surrounding this point reflect 95% confidence intervals calculated using the Wilson score with continuity correction. The sensitivity analysis shows that, even at the lower positivity bounds, the intervention would still have been cost-saving.

Figure 3:

The Cost-Saving Threshold at Varying Levels of Syphilis and HIV Positivity

The dotted line denotes the cost-saving threshold assuming a value of $7,298 for each syphilis case diagnosed and $27,296 for each HIV case diagnosed. The area above and to the right of the dotted line represents combinations of syphilis and HIV positivity that would be cost-saving given the estimated program cost. The shaded area below/left of the blue line is not cost-saving but may still be cost-effective. The triangular point reflects the pair of syphilis and HIV positivity found in the study, and the error bars surrounding this point reflect 95% confidence intervals calculated using the Wilson score with continuity correction.

Accounting for all new positives diagnosed, and the service referrals offered by the DIS, the MATRix-NC program was highly cost effective from health system and societal perspectives. That is, the benefits from treating cases, preventing further transmission, and additional linkage to services outweighed the costs of implementing the program even based on its prior model. The networking approach was highly effective in identifying participants at high-risk of HIV infection and/or syphilis who may not have been reached otherwise using traditional DIS approaches. Our study adds to some limited evidence on the cost-effectiveness of referral strategies for identifying new cases of STIs. Golden et al. (2006) used a peer referral strategy to recruit MSM for testing and found a cost of $6,670 (in 2019 dollars) per case of HIV diagnosed,⁵ and Shrestha et al. (2009) found costs of new HIV diagnoses in referred partners ranging from $8,916 – $27,426, depending on location.¹⁸

This study was limited by several factors. First, the MATRix-NC DIS patient navigators were not dedicated to the demonstration project full time. Because their salaries were such an important component of program costs, any mismeasurement of the proportion of time allocated to the intervention may significantly impact cost estimates. The study participants were not billed for any services provided. Billing patient health insurance may increase revenue and offset the costs of public health programs,¹⁹ but implementation of billing is often complicated. Laws governing charging patients for sexual health services, and how revenues from billing can be used, vary by state.²⁰ Because of this, examining the cost of the MATRix-NC program without billing may be more generalizable to other jurisdictions.

Next, the peer network studied in this paper exhibited high HIV prevalence and viremia.¹⁰ The effects demonstrated here were dependent on the local epidemiological context and the social and sexual networks of the seeds and peers, which may not generalize to other settings. Finally, our study used process (number of tests) or intermediate outcomes (number of new diagnoses) to assess effectiveness. Broader benefits of the program might include averted infections from reduced transmission, but estimating longer-term and indirect outcomes was beyond the scope of this study.

Conclusions

We conducted a comprehensive cost evaluation of MATRix to better understand the cost structure of enhanced DIS services directed towards MSM and TGW of color, and their sexual and social networks. Operating a program like MATRix relies heavily on personnel, with DIS playing a particularly important role. The program was successful at identifying a peer population with a disproportionate risk of HIV infection and syphilis, facilitated through chain-referral recruitment from seeds. The benefits of testing, treating, and referrals for this population likely far exceeded the costs of the program given the long-term consequences of undiagnosed HIV and syphilis among these individuals.