The US National HIV/AIDS Strategy (NHAS) emphasizes the use of technology to facilitate coordination of comprehensive care for people with HIV. We examined cost-effectiveness from the health system perspective of 6 health information technology (HIT) interventions implemented during 2008 to 2012 in a Ryan White HIV/AIDS Program (RWHAP) Special Projects of National Significance (SPNS) Program demonstration project.
HIT interventions were implemented at 6 sites: Bronx, New York; Durham, North Carolina; Long Beach, California; New Orleans, Louisiana; New York, New York (2 sites); and Paterson, New Jersey. These interventions included: (1) use of HIV surveillance data to identify out-of-care individuals; (2) extension of access to electronic health records (EHRs) to support service providers; (3) use of electronic laboratory ordering and prescribing; and (4) development of a patient portal. We employed standard microcosting techniques to estimate costs (in 2018 US dollars) associated with intervention implementation. Data from a sample of electronic patient records from each demonstration site were analyzed to compare prescription of antiretroviral therapy (ART), CD4 cell counts, and suppression of viral load, before and after implementation of interventions. Markov models were used to estimate additional healthcare costs and quality-adjusted life-years saved as a result of each intervention. Overall, demonstration site interventions cost $3,913,313 (range = $287,682 to $998,201) among 3,110 individuals (range = 258 to 1,181) over 3 years. Changes in the proportion of patients prescribed ART ranged from a decrease from 87.0% to 72.7% at Site 4 to an increase from 74.6% to 94.2% at Site 6; changes in the proportion of patients with 0 to 200 CD4 cells/mm3 ranged from a decrease from 20.2% to 11.0% in Site 6 to an increase from 16.7% to 30.2% in Site 2; and changes in the proportion of patients with undetectable viral load ranged from a decrease from 84.6% to 46.0% in Site 1 to an increase from 67.0% to 69.9% in Site 5. Four of the 6 interventions—including use of HIV surveillance data to identify out-of-care individuals, use of electronic laboratory ordering and prescribing, and development of a patient portal—were not only cost-effective but also cost saving ($6.87 to $14.91 saved per dollar invested). In contrast, the 2 interventions that extended access to EHRs to support service providers were not effective and, therefore, not cost-effective. Most interventions remained either cost-saving or not cost-effective under all sensitivity analysis scenarios. The intervention that used HIV surveillance data to identify out-of-care individuals was no longer cost-saving when the effect of HIV on an individual’s health status was reduced and when the natural progression of HIV was increased. The results of this study are limited in that we did not have contemporaneous controls for each intervention; thus, we are only able to assess sites against themselves at baseline and not against standard of care during the same time period.
Why was this study done?
- ➢ The Health Resources and Services Administration’s Special Projects of National Significance Program (HRSA/SPNS) funded a 4-year initiative (2007 to 2011) in 6 demonstration sites to enhance and evaluate existing health information electronic network systems for people living with HIV (PLHIV) in underserved communities.
- ➢ Each of the 6 demonstration sites implemented one or more health information technology (HIT) interventions to facilitate comprehensive care and enhance engagement in HIV medical services. These interventions included: (1) use of HIV surveillance data to identify out-of-care individuals; (2) extension of access to electronic health records to support service providers; (3) use of electronic laboratory ordering and prescribing; and (4) development of a patient portal.
- ➢ This study estimates the total costs, cost-effectiveness, and potential cost-savings of these 6 interventions.
What did researchers do and find?
- ➢ We used information on the cost of each intervention and the health status of PLHIV in each setting before and after implementation of each intervention to estimate: (1) changes in the cost of care and other services for PLHIV in each setting; and (2) changes in expected health status (measured as quality-adjusted life-years or QALYs) among PLHIV in each setting. We then used this information to estimate additional healthcare costs and QALYs gained for each intervention.
- ➢ Four of the interventions were associated with lower healthcare costs and better health outcomes (QALYs gained) for PLHIV in each setting. These interventions saved between $6.87 and $14.91 per dollar invested.
- ➢ Two interventions that provided access to medical record information to support service providers were not associated with improved health outcomes for PLHIV in these settings. These interventions were not effective or cost-effective.
What do these findings mean?
- ➢ These results show that HIT interventions that facilitate changes in patient or provider behavior have the potential to improve the health status of PLHIV and reduce healthcare costs. HIT interventions that only provided additional information to support service providers were less successful.
- ➢ This study did not include a contemporaneous comparison group. Therefore, we do not know the degree to which improvements in the health status of PLHIV in these settings were due to changes in the quality of care for PLHIV over the life of the interventions.
Citation: Shade SB, Marseille E, Kirby V, Chakravarty D, Steward W, Koester K, et al. (2021) Health information technology interventions and engagement in HIV care and achievement of viral suppression in publicly funded settings in the US: A cost-effectiveness analysis. PLoS Med 18(4):
Academic Editor: Blythe Adamson, University of Washington, UNITED STATES
Received: August 18, 2019; Accepted: March 25, 2021; Published: April 7, 2021
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: The Ryan White HIV/AIDS Program Services Report (RSR) client level data are not available due to client privacy and confidentiality. HRSA makes aggregate data available in the Ryan White HIV/AIDS Program Annual Client-Level Data Report: https://hab.hrsa.gov/data/data-reports. Requests for aggregate data can be made to: [email protected]
Funding: SBS, EM, DC, WS, KK and JJM were supported by grant number H97HA08477 from the Health Resources and Services Administration (HRSA), HIV/AIDS Bureau’s Special Projects of National Significance (SPNS) Program. The funder did not participate in the design, data collection, analysis or decision to publish, but AC did participate in the preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
antiretroviral therapy; EHR,
electronic health record; HIT,
health information technology; HRSA/SPNS,
Health Resources and Services Administration’s Special Projects of National Significance; NHAS,
US National HIV/AIDS Strategy; PLHIV,
people living with HIV; QALY,
quality-adjusted life-year; RWHAP,
Ryan White HIV/AIDS Program; SPNS,
Special Projects of National Significance; UCSF,
University of California, San Francisco
The United States National HIV/AIDS Strategy (NHAS), first introduced in 2010 and subsequently revised in 2015, has as its goals: to reduce the number of people who become infected with HIV; increase access to care and optimize health outcomes for people living with HIV (PLHIV); reduce HIV-related health disparities; and achieve a more coordinated national response to the HIV epidemic.  The strategy articulates measurable action steps and sets 5-year quantitative targets based on evidence-based approaches. Its stated purpose is to provide a roadmap for responding to the epidemic among public and private stakeholders; its guiding principles include accountability and science-driven decision-making.
The NHAS targets to increase the proportion of all HIV–infected individuals in the US who are aware of their serostatus, linked to care within 3 months of diagnosis, engaged in HIV care continuously and ultimately achieve suppression of the virus in their blood [2,3]. Achieving these targets would contribute to ending the epidemic because suppression of viral load is important for individual patient health and for the prevention of transmission of HIV in the population as a whole .
Unfortunately, there are numerous barriers to “rapid and effective” HIV care [5–12], including structural challenges (e.g., lack of housing), insufficient financing (e.g., lack of insurance) [13,14], personal and cultural characteristics (e.g., beliefs about health system, racism), comorbidities (e.g., mental illness and substance abuse), stigma, fear of confidentiality violations, and healthcare provider attitudes [5–24]. Research suggests that both individual and system-level interventions have the potential to improve care engagement and the quality of HIV medical and support services [3,25].
To date, few studies have explored the extent to which sharing patient information across geographically disparate HIV surveillance, primary care, and support service organizations can enhance linkage to care, retention and adherence to care and treatment, the quality of core and support services, as well as health outcomes for people with HIV. This sharing process, however, is key to achieving rapid and effective HIV care precisely because of the multifaceted impacts on care outcomes listed above. For example, the National HIV/AIDS Strategy: Updated to 2020 (NHAS 2020) goals call for enhanced collaboration among providers, improved assessments and measurements of health outcomes to track progress towards NHAS 2020 goals across populations in a region . Effectively meeting these objectives requires technologies and practices that ensure that providers and the communities in which they are embedded coordinate the services that they deliver to patients.
Health information technology (HIT) interventions, including the electronic transfer of information between organizations (health information exchange), extending access to information with additional providers within organizations, and sharing of information with patients has the potential to enhance engagement in comprehensive HIV care. HIT interventions have been previously implemented in the context of other diseases to link public health surveillance programs to primary care services, laboratory and pharmacies to primary care, and primary and specialty care [27–31]. HIT has the potential to close many of the gaps that lead to suboptimal care for PLHIV by, for example, improving the referral and tracking of patients among services, identifying patients who have missed appointments, and allowing providers to coordinate services to ensure that each individual receives a comprehensive set of services that he/she needs.
To test the potential of HITs in supporting better-coordinated care in publicly funded settings, the Health Resources and Services Administration (HRSA) sponsored an initiative to develop HIT interventions in clinical care sites serving patients eligible for the Ryan White HIV/AIDS Program (RWHAP). The RWHAP was enacted by Congress in 1990 to ensure access to quality HIV care and treatment for those who cannot afford it, provision of support services (e.g., transportation and housing) for those who experience challenges and/or obstacles in entering and remaining in care, and coordination of care (e.g., medical case management) for those who have cooccurring conditions that impact the effectiveness of HIV care [32,33]. RWHAP providers typically deliver a constellation of services to support HIV status awareness, linkage to care, and retention. Each of the demonstration sites in the initiative implemented systems that supported engagement in care by targeting gaps they had identified in their clinical settings.
Description of the initiative
One component of the RWHAP is the Special Projects of National Significance (SPNS) Program, which provides grants to fund innovative models of care and supports the development of effective and innovative delivery systems for HIV care. In 2007, SPNS funded a 4-year initiative in 6 demonstration sites to enhance and evaluate existing health information electronic network systems for PLHIV in underserved communities. These 6 demonstration sites were: the Bronx-Lebanon Hospital Center in Bronx, New York; the City of Paterson Department of Human Services, New Jersey; the Duke University Center for Health Policy in Durham, North Carolina; the Louisiana State University Health Services Center in New Orleans, Louisiana; the New York Presbyterian Hospital in New York, New York; and the St. Mary Medical Center Foundation in Long Beach, California. The University of California, San Francisco (UCSF) Center for AIDS Prevention Studies was funded as the Evaluation and Support Center (hereafter referred to as the “Center”) to conduct a cross-site evaluation of the HIT interventions and provide technical assistance and support to the demonstration sites. This paper estimates the total costs, cost-effectiveness, and potential cost-savings of each intervention in the initiative according to the CHEERS Statement (see S1 CHEERS Checklist).
The protocol for the cross-site evaluation was approved by the Committee for Human Research at UCSF (see S1 Evaluation Protocol). However, this protocol does not include information on how information about the cost-effectiveness of demonstration site interventions will be evaluated as our cost-effectiveness analysis does not include any identifiable data as thus is deemed exempt from UCSF IRB approval.
Each of the 6 demonstration sites implemented one or more HIT interventions to facilitate comprehensive care and enhance engagement in HIV medical services. These interventions and the characteristics of patients within each demonstration site have been previously detailed [34–37] and are summarized here:
- Site 1 developed a structured patient summary within its electronic health record (EHR) that included highlighted alerts to identify needed clinical services. The hospital also provided case managers with access to these patient summaries to facilitate coordination of support services and reinforce engagement in HIV care. This site targeted “high need patients,” defined as patients who had detectable viral loads despite multiple medical and social support service visits across multiple social service providers. This site anticipated that providing access to medical record information to cases managers would facilitate coordination of care and reduction of redundancies in the provision of support services.
- Site 2 provided support service providers access to a regional medical center EHR to facilitate coordination of support services and reinforce engagement in HIV care. This site anticipated that improved coordination of care would result in higher utilization of necessary support services.
- Site 3 linked the state HIV surveillance branch and EHRs in publicly funded health facilities. It created an alert whenever a patient known to the surveillance branch to be out of care for HIV treatment presented for services in an emergency room or other non-HIV healthcare setting. Providers at the care site then acted on the alert to facilitate reengaging the patient in HIV care. This site anticipated that the intervention would improve linkage or reengagement in care among previously out-of-care individuals.
- Site 4 created continuity of care patient summaries. Patients were then given to access this information through a patient portal to facilitate engagement in HIV care. Patients were also able to provide this information to external providers to facilitate coordination of services. This site anticipated that access to information would increase utilization of necessary care and support services.
- Site 5 developed an EHR with summary comparison reports that was shared across all health service providers. This shared record facilitated the development of a quality improvement framework in which activities were implemented to increase targeted prevention services. Health record alerts were also formulated to help providers identify and intervene with patients who had not received needed clinical services. This site anticipated that the intervention would improve the quality of care for HIV–infected patients.
- Site 6 implemented HIT to facilitate electronic prescriptions as well as laboratory work orders to reduce the time needed to access these services and to enhance engagement in HIV care.
The Center collected deidentified quantitative electronic patient record data from each demonstration site for the 6-month period preceding the implementation of the HIT intervention and for each 6-month period thereafter through the end of the project. Demonstration sites were asked to provide a simple random sample of at least 100 patients from each demonstration site; however, some sites chose to provide records for all patients at each time point and some sites chose to submit information for a larger random sample of patients. The present analyses include data for the 6 months prior to the intervention implementation and after full implementation of the intervention (during the last reporting period).
Data elements included as outcomes or effectiveness measures in the present analysis include:
- Prescription of antiretroviral therapy (ART): A binary outcome variable was created to record if the client had been prescribed highly active ART at any time during a reporting period.
- CD4 cell count: A categorical variable was created to record if the client had a low (0 to 200 cells/mm3), medium (201 to 500 cells/mm3), or high (>500 cell/mm3) CD4 cell count.
- Undetectable viral load: For this evaluation, undetectable viral load was defined as less than 1,000 copies/mL because this was the level of detection that was available to the site using data from the HIV surveillance system. A binary outcome variable was created to record if the client had an undetectable viral load at the last test during a reporting period.
Measuring intervention costs
We used standard microcosting techniques to estimate the incremental cost (in 2018 US dollars) of implementation of each site’s intervention from the perspective of the health system . We developed a standard cost data collection tool and protocol. Demonstration site staff completed the data collection tool annually from September 2008 through August 2010. We included additional personnel time, materials, and other resources needed to implement the interventions, whether funded through the SPNS initiative (direct costs) or through other sources (in-kind costs). Sites were granted the same amount annually. Sites used some of these resources to conduct evaluation of their local interventions. These costs were not included in estimation of intervention costs. Costs were adjusted for inflation to 2018 US dollars (https://data.bls.gov/cgi-bin/cpicalc.pl).
Direct and in-kind expenditures were classified in 1 of 4 categories: (i) personnel, including fringe benefits; (ii) other recurring costs consisting of supplies and services; (iii) capital expenditures; and (iv) building space. Capital expenditures such as computers and furniture were amortized over 5 years of expected useful life and assuming no salvage value. Building space was valued at the market rental rate for any space that had previously been utilized for another activity.
In order to understand the activities and associated costs needed to implement each intervention, we worked with demonstration site staff to develop a comprehensive list of personnel activities. These included: HIT design and customization; initial debug-quality control; system monitoring and quality control; ongoing use of system; ongoing maintenance including data backup; system refinement and improvement; user support services; stakeholder sessions; patient recruitment; training of providers; training of case managers; training of patients; and legal and other activities. Demonstration site staff distributed staff time across these 14 activities on a monthly basis and then totaled for the year and multiplied by the compensation rate. The allocation was performed based on the staff member’s knowledge of intervention operations and their understanding of how personnel spend their time between HIT and non-HIT activities. Consultant and other outside costs were entered as direct dollar amounts assigned to each activity such that the total equaled the sum spent for each outside contracted service.
Program-attributable health outcomes were obtained by estimating quality-adjusted life-years (QALYs) gained among patients impacted by the interventions, using information on prescription of ART, CD4 cell count and viral suppression at baseline and, at last, follow-up. We calculated expected QALYs over a 5-year period based on the distribution of patient health states at baseline and follow-up using a Markov health state transition model, literature-derived estimates of transition probabilities and health state utilities, and a per annum discount of 3% (see Tables A-E in S1 Text) . We then compared expected QALYs at baseline and follow-up to estimate QALYs gained.
Measuring healthcare costs
We estimated the cost of healthcare with and without implementation of the interventions by applying literature-derived costs of healthcare (adjusted for inflation to 2018 US dollars) to the distribution of patients by CD4 cell count, viral suppression, and ART before and after implementation of each intervention (see Tables A and E in S1 Text) [41,47–50]. We then estimated the total 5-year cost of healthcare with and without the interventions based on the number of patients exposed to each intervention. For each demonstration site, our outcomes were obtained by taking the difference in costs with and without implementation of the intervention.
We estimated the cost-effectiveness of each intervention using pre-post analysis. We added the difference in modelled 5-year costs of healthcare with and without HIT implementation to the cost of each intervention. This number was then divided by the estimated number of QALYs gained among patients exposed to the interventions to estimate the cost per QALY gained for each intervention. If the difference in 5-year costs of healthcare (with minus without) was greater than the cost of the intervention, then the intervention was considered to be “cost saving.” In this case, cost savings were estimated as the reduction in healthcare costs over the cost of the intervention.
We estimated the percent change in cost per QALY gained when we varied ART and healthcare costs (based on previous literature; see Table E in S1 Text), utilities (200%, 50%) and transition probabilities (200%, 50%) for each intervention in order to examine the consistency of our observed results across participating sites.
Overall, demonstration site interventions cost $3,913,313 over 3 years (range = $287,682 to $998,201; Table 1). This included $3,648,512 in direct costs (range = $224,968 to $889,409) in funding from the HRSA SPNS initiative and $412,289 in indirect (in-kind) costs (range = $24,227 to $199,386) with external funding from demonstration sites and their partners. Personnel effort made up the majority of costs across all intervention sites (overall = 83%; range = 74% to 91%) with a smaller proportion going to recurring goods and services and facility rental (overall = 14%; range = 6% to 26%). Capital equipment (including computers and informatics infrastructure) comprised only a small percentage of costs across all demonstration sites (overall = 3%; range = 1% to 6%).
Table 2 includes the distribution of personnel costs by activity across demonstration sites. Overall, costs associated with personnel effort were evenly divided across conceptualization and design of the interventions (overall = 30%; range = 10% to 55%), ongoing system maintenance and improvement (overall = 30%; range = 13% to 39%), and user training and support (overall = 32%; range = 8% to 55%, although the distribution across these 3 broad categories varied substantially across interventions. Three (3) demonstration sites also incurred legal costs associated with individual and system-level privacy issues (overall = 8%; range = 0% to 19%).
Table 3 includes the distribution of prescription of ART, CD4 cell count, and viral load suppression within sites at baseline and after full HIT intervention implementation. We observed increases in the proportion of patients on ART in Sites 2 (from 81.1 to 90.5), 3 (from 54.0 to 59.0), 5 (from 85.5 to 87.7), and 6 (from 74.6 to 94.2). In contrast, we observed decreases in the proportion on patients on ART in Sites 1 (from 91.0 to 82.9) and 4 (from 87.0 to 72.9). We observed small differences in the distribution of CD4 cell counts within demonstration sites. Of note, the proportion of patients with 0 to 200 CD4 cells/mm3 increased in Sites 1 (from 16.7 to 30.2) and 2 (from 14.8 to 17.8), which may have been indicative of the targeting of out-of-care clients that occurred in these sites. In contrast, we observed decreases in the proportion of patients with 0 to 200 CD4 cells/mm3 in Sites 3 (31.8 to 29.4), 4 (17.4 to 15.2), 5 (20.2 to 11.0), and 6. We observed small changes in the proportion of patients with undetectable viral load in Sites 2 (62.2 to 63.5), 5 (67.0 to 69.9), and 6 (72.2 to 68.0) and large decreases in the proportion of patients with undetectable viral load in Sites 1 (84.6 to 46.0), 3 (34.8 to 9.0), and 4 (39.3 to 24.2).
Table 4 includes information on predicted healthcare costs and QALYs within the exposed population based on prescription of ART, CD4 cell counts, and viral suppression before and after HIT intervention implementation in each site. This information was combined with information about the cost of each intervention to estimate cost-effectiveness and cost savings. Prior to HIT intervention implementation, predicted healthcare costs were high (range = $82,250 to $100,067 per patient per year) and predicted QALYs were low (range = 3.71 to 4.12 per exposed patient) across sites at baseline. We observed a decline in predicted healthcare and intervention costs (range = $68,397 to $94,392 per patient per year; change in costs per patient: Site 1 = $29,244; Site 2 = $8,835; Site 3 = −$15,869; Site 4 = −$20,852; Site 5 = −$36,196; Site 6 = −$4,695) and small changes in predicted QALYs (range = 3.68 to 4.18 per patient; change in additional QALYs per patient: Site 1 = −0.44; Site 2 = −0.01; Site 3 = 0.01; Site 4 = 0.03; Site 5 = 0.08; Site 6 = 0.02) after HIT intervention implementation due to changes in prescription of ART and consequent improvements in CD4 cell counts and viral suppression among patients in the intervention sites.
In 4 sites (Sites 3 to 6), the decline in predicted healthcare costs more than offset the cost of the HIT interventions suggesting that these interventions could be not only cost-effective but also cost saving. In these sites, each dollar invested was associated with a predicted decline in healthcare costs of $6.32 to $12.97. In 2 sites, (Sites 1 and 2), we observed a decline in predicted QALYs among patients exposed to their interventions. Thus, these results suggest that these interventions were not effective and, thus, not cost-effective.
Fig 1 includes results of univariate sensitivity analysis conducted by varying ART costs, healthcare costs, utilities, and transition probabilities. We estimated the percent change in cost per QALY gained under various scenarios. Site 6, which served the largest number of patients, was sensitive to variation in ART and healthcare costs, while Sites 2 and 3, which produced fewer additional QALYs, were sensitive to variation in utilities and transition probabilities. Cost per QALY gained varied less than 20% under all scenarios for Sites 1, 4, and 5. Of note, the status of each intervention as not effective or cost saving was consistent across all almost all scenarios. Site 3 was no longer cost saving when the effect of HIV on a patient’s health was cut in half (utility 50%; cost per QALY gained = $39,397,160) and when the probability of transition from one disease state to another was doubled (transition probability 200%; cost per QALY gained = $3,087,353).
To our knowledge, this initiative is the first to demonstrate that HIT may not only improve engagement in HIV care and clinical outcomes but may also reduce costs associated with HIV care and treatment. Previous research has shown cost savings associated with HIT because it facilitates ordering of laboratory tests, use of bar codes to reduce errors in dispensing of medications in hospitals and distribution, and archiving of picture images [51–53]. This initiative has demonstrated that HIT that focuses on facilitating linkage and retention in HIV care, quality of HIV care, and adherence to ART may also reduce healthcare costs. Four of the 6 interventions we evaluated—including use of HIV surveillance data to identify out-of-care individuals, use of electronic laboratory ordering and prescribing, and development of a patient portal—were not only cost-effective but also cost saving.
In contrast to what we expected, costs associated with these HIT interventions were driven primarily by personnel time and not by costs associated with the purchase or maintenance of equipment and technology. In addition, over 3 years, personnel costs were evenly distributed across conceptualization and design of the HIT interventions, ongoing system maintenance and improvement, and user training and support. Interventions did not have defined start-up periods, but instead conducted iterative improvements over time. These findings correspond to previous research which found that ongoing yearly costs of HIT implementation were approximately 20% of initial costs .
We observed an increase in QALYs and a decrease in expected healthcare costs among patients in 4 of the 6 demonstration sites and a decrease in QALYs and an increase in healthcare costs among patients in 2 demonstration sites. These findings are primarily due to increases in the proportion of patients with 0 to 200 cells/mm3 CD4 counts within Sites 3, 4, 5, and 6 because patients in this subgroup have substantially higher healthcare costs and higher mortality than other subgroups. Although we also observed significant increases in the proportion of patients on ART in Sites 3, 4, and 5, and significant increases in the proportion of patients with undetectable viral load in Sites 1, 3, 4, 5, and 6 , changes in the distribution of these characteristics had substantially less impact on short-term healthcare costs or the probability of mortality among patients with higher CD4 cell counts .
The 4 cost-saving interventions focused on linkage to HIV care (Site 3), improving access to HIV medications (Site 6), improving the quality of HIV care (Site 5), and improving patient access to information about their medical care (Site 4). The 2 interventions that were not cost-effective (Site 1 and Site 2) focused on expanding access to medical record information to support service providers. It is possible that these 2 interventions were less successful than other interventions because they did not result in targeted changes in staff workflow or behavior. In fact, it may well be that the current availability of user interfaces is not adequate to optimize or improve the provider experience at all. As Dr. Robert Wachter has pointed out, “[o]ur iPhones and their digital brethren have made computerization look easy, which makes our experience with health care technology doubly disappointing. An important step is admitting that there is a problem, toning down the hype, and welcoming thoughtful criticism, rather than branding critics as Luddites.” .
These results are relatively insensitive to variation the parameters used to estimate cost-effectiveness. Under 2 scenarios, the Site 2 intervention was no longer cost saving. Furthermore, the additional cost per QALY gained under these scenarios (i.e., $3,087,353 and $39,397,160) dramatically exceeds baseline costs per QALY. This intervention was susceptible to variation in parameters because the distribution in health states was similar before and after implementation of the intervention (i.e., it contributed to a small number of QALYs gained).
These results must be interpreted within a limited context. First, because this initiative consisted of 6 demonstration projects implemented at the health facility or system level, we are not able to isolate the effect of health IT interventions implemented as part of this initiative from other existing interventions at the sites or from temporal trends. We cannot rule out the possibility that interventions outside of the current SPNS initiative contributed to changes in the prescription of ART, the proportion of patients with 0 to 200 cells/mm3 CD4 counts, or the proportion of patients with undetectable viral load. However, provider workflows and procedures in the 4 successful interventions changed substantially as a result of these interventions . Therefore, we have reason to believe that the interventions contributed to meaningful changes in clinical practices and other outcomes. Second, for 3 of the 6 interventions, the population of people affected by the intervention changed over time. Site 1 included “high need” patients who exited the program once they were able to achieve undetectable viral load. Therefore, it is possible that this intervention benefited patients; however, we were not able to observe these patients over an extended period of time and thus were not able to observe substantial changes in CD4 cell count, the factor most associated with the cost of care over time. Of note, this intervention was maintained by Site 1 after completion of the initiative. Similarly, Sites 3 and 4 experienced substantial increases in the number of patients in the intervention over the life of the initiative. Although we have adjusted for differences in patient characteristics over time, it is possible that patients who were new to the site differed from existing patients and these differences explain the observed results. Third, because the care and treatment of people with HIV is complex, we are not able to identify the specific mechanisms through which these HIT interventions effect prescription of ART, the proportion of patients with 0 to 200 cells/mm3 CD4 counts or the proportion of patient with undetectable viral load. However, we are able to hypothesize how each intervention improves the quality of care for HIV–infected patients in each participating site. Fourth, this study was conducted some time ago. In the intervening years, many interventions that employ HIT have been shown to be cost-effective, including those that reduce the time needed for care, reduce duplication of care, and reduce unnecessary visits. However, to date, no studies have been conducted on the cost-effectiveness of the use of HIT in the context of HIV or other chronic diseases .
These findings contribute to a growing literature on the benefits of HIT and counter beliefs that the costs associated with HIT outweigh the potential benefits [53,56–63]. Previous reviews have demonstrated improvements in efficiency and reduced transcription errors. However, data on the effect of HIT on the costs of care and patient outcomes have been limited [59,62–64]. Many authors have suggested that HIT may only be cost effective in large health systems, but other evidence suggests that it may also be cost saving in smaller settings [55,59]. Other findings suggest that the scale of benefits and cost savings may depend on the “fit” of the HIT to the setting, how and how fully HIT is adopted and utilized, and whether HIT is coupled with behavior changes [59,61–64]. Our findings reinforce the role of these factors in that interventions that facilitated targeted changes in provider work flows were found to be cost saving, whereas interventions that facilitated access to information for patients or support service providers were less likely to be effective, cost-effective, or cost saving. The findings presented here suggest that HIT investments may contribute to improvements in effective treatment of HIV to achieve sustained viral suppression, but also may reduce the cost of care for PLHIV.
The authors would like to thank the funders, the collaborators from each of the 6 demonstration sites, and all of the patients in care at participating sites.
The views expressed in this publication are solely the opinions of the authors and do not necessarily reflect the official policies of the US Department of Health and Human Services or the Health Resources and Services Administration nor does mention of the department or agency names imply endorsement by the US Government.
Office of National AIDS Policy. National HIV/AIDS Strategy for the United States: Updated to 2020. ONAP, The White House, July 30, 2015. [cited 2015 July 30]. Available from: https://www.aids.gov/federal-resources/national-hiv-aids-strategy/nhas-update.pdf.
Gardner EM, McLees MP, Steiner JF, Del Rio C, Burman WJ. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clin Infect Dis. 2011 Mar 15;52:793–800. pmid:21367734
Mugavero MJ. Improving engagement in HIV care: what can we do? Top HIV Med. 2008;16:156–61. pmid:19106431
Das M, Chu PL, Santos GM, Scheer S, Vittinghoff E, McFarland W, et al. Decreases in community viral load are accompanied by reductions in new HIV infections in San Francisco. PLoS One. 2010 Jun 10;5:e11068. pmid:20548786
Millman ML. Access to health care in America: National Academies Press; 1993. pmid:11659619
Mugavero MJ, Lin HY, Allison JJ, Giordano TP, Willig JH, Raper JL, et al. Racial disparities in HIV virologic failure: do missed visits matter? J Acquir Immune Defic Syndr. 2009;50:100–8. pmid:19295340
Mugavero MJ, Lin HY, Willig JH, Westfall AO, Ulett KB, Routman JS, et al. Missed visits and mortality among patients establishing initial outpatient HIV treatment. Clin Infect Dis. 2009;48:248–56. pmid:19072715
Tobias C, Cunningham WE, Cunningham CO, Pounds MB. Making the connection: the importance of engagement and retention in HIV medical care. AIDS Patient Care STDs. 2007;21:3–8. pmid:17563287
Torian LV, Wiewel EW. Continuity of HIV-related medical care, New York City, 2005–2009: Do patients who initiate care stay in care? AIDS Patient Care STDS. Feb. 2011;25:79–88. pmid:21284498
Torian LV, Wiewel EW, Liu KL, Sackoff JE, Frieden TR. Risk factors for delayed initiation of medical care after diagnosis of human immunodeficiency virus. Arch Intern Med. 2008;168:1181–7. pmid:18541826
Tripathi A, Gardner LI, Ogbuanu I, Youmans E, Stephens T, Gibson JJ, et al. Predictors of time to enter medical care after a new HIV diagnosis: a statewide population-based study. AIDS Care. 2011 Nov;23:1366–73. pmid:22022847
Tripathi A, Youmans E, Gibson JJ, Duffus WA. The impact of retention in early HIV medical care on viro-immunological parameters and survival: a statewide study. AIDS Res Hum Retrovir. 2011 Jul;27:751–8. pmid:21142607
Bozzette SA, Berry SH, Duan N, Frankel MR, Leibowitz AA, Lefkowitz D, et al. The care of HIV-infected adults in the United States. N Engl J Med. 1998;339:1897–904. pmid:9862946
Brown VB, Smereck GAD, German V, Hughes C, Melchior LA, Huba GJ. Change in perceived barriers and facilitators to treatment among women with HIV/AIDS as a function of psychosocial service utilization. AIDS Patient Care STDs. 2000;14:381–90. pmid:10935054
Cunningham WE, Andersen RM, Katz MH, Stein MD, Turner BJ, Crystal S, et al. The impact of competing subsistence needs and barriers on access to medical care for persons with human immunodeficiency virus receiving care in the United States. Med Care. 1999;37:1270–81. pmid:10599608
Huba GJ, Melchior LA, Smereck GAD, Brown VB, Jean-Louis E, German VF, et al. Perceived barriers to receiving HIV services in groups of traditionally underserved individuals: Empirical models. Home Health Care Serv Q. 2001;19:53–75. pmid:11357465
Palacio H, Shiboski CH, Yelin EH, Hessol NA, Greenblatt RM. Access to and utilization of primary care services among HIV-infected women. J Acquir Immune Defic Syndr. 1999;21:293. pmid:10428107
Bogart LM, Thorburn S. Are HIV AIDS conspiracy beliefs a barrier to HIV prevention among African Americans? J Acquir Immune Defic Syndr. 2005;38:213. pmid:15671808
Feldman RH, Fulwood R. The three leading causes of death in black Americans: Barriers to reducing excess disparity and to improving health behaviors. J Health Care Poor Underserved. 1999;10:45. pmid:9989006
Joyce GF, Chan KS, Orlando M, Burnam MA. Mental health status and use of general medical services for persons with human immunodeficiency virus. Med Care. 2005;43:834. pmid:16034298
Klinkenberg WD, Sacks S. Mental disorders and drug abuse in persons living with HIV/AIDS. AIDS Care. 2004;16:22–42. pmid:15736820
Gifford AL, Collins R, Timberlake D, Schuster MA, Shapiro MF, Bozzette SA, et al. Propensity of HIV patients to seek urgent and emergent care. J Gen Intern Med. 2000;15:833–40. pmid:11119179
Kalichman SC, Rompa D. Functional health literacy is associated with health status and health-related knowledge in people living with HIV-AIDS. J Acquir Immune Defic Syndr. 2000;25:337. pmid:11114834
Bodenlos JS, Grothe KB, Kendra K, Whitehead D, Copeland AL, Brantley PJ. Attitudes toward HIV Health Care Providers scale: development and validation. AIDS Patient Care STDs. 2004;18:714–20. pmid:15659882
Mugavero MJ, Norton WE, Saag MS. Health care system and policy factors influencing engagement in HIV medical care: piecing together the fragments of a fractured health care delivery system. Clin Infect Dis. 2011 Jan 15;52 (Suppl 2):S238–46. pmid:21342913
Office of National AIDS Policy. National HIV/AIDS Strategy for the United States: Updated to 2020, Federal Action Plan, December 2015. [cited 2017 February 22]. Available from: https://www.aids.gov/federal-resources/national-hiv-aids-strategy/nhas-2020-action-plan.pdf
Branger PJ. van’t Hooft A, van der Wouden JC, Moorman PW, van Bemmel JH. Shared care for diabetes: supporting communication between primary and secondary care. Int J Med Inform. 1999;53:133–42. pmid:10193883
Hinman AR, Davidson AJ. Linking children’s health information systems: clinical care, public health, emergency medical systems, and schools. Pediatrics. 2009;123 (Suppl 2):S67–73.
Kaelber DC, Bates DW. Health information exchange and patient safety. J Biomed Inform. 2007;40:S40–5. pmid:17950041
Kaushal R, Barker KN, Bates DW. How can information technology improve patient safety and reduce medication errors in children’s health care? Arch Pediatr Adolesc Med. 2001;155:1002–7. pmid:11529801
Vest JR, Jasperson JS, Zhao H, Gamm LD, Ohsfeldt RL. Use of a health information exchange system in the emergency care of children. BMC Med Inform Decis Mak. 2011 Dec 30;11:78. pmid:22208182
Conviser R, Pounds M. Background for the studies on ancillary services and primary care use. AIDS Care. 2002;14:7–14.32. pmid:12204138
Health Resources and Services Administration. About the Ryan White HIV/AIDS Program. [cited 2019 March 12]. Available from: https://hab.hrsa.gov/about-ryan-white-hivaids-program/about-ryan-white-hivaids-program
Shade SB, Chakravarty D, Koester KA, Steward WT, Myers JJ. Health information exchange interventions can enhance quality and continuity of HIV care. Int J Med Inform. 2015;81:e1–9.
Myers JJ, Koester KA, Chakravarty D, Pearson C, Maiorana A, Shade SB, et al. Perceptions regarding the ease of use and usefulness of health information exchange systems among medical providers, case managers and non-clinical staff members working in HIV care and community settings. Int J Med Inform. 2012;81:e21–9. pmid:22854159
Steward WT, Koester KA, Collins SP, Maiorana A, Myers JJ. The essential role of reconfiguration capabilities in the implementation of HIV-related health information exchanges. Int J Med Inform. 2012;81:e10–20. pmid:22841703
Proeschold-Bell RJ, Belden CM, Parnell H, Cohen S, Cromwell M, Lombard F. A randomized controlled trial of health information exchange between human immunodeficiency virus institutions. J Public Health Manag Pract. 2010;16 (6):521–8. pmid:20885182
Drummond MF, O’Brien B, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. Oxford: Oxford Medical Publications; 1997. p. 305.
Sendi P, Gunthard HF, Simcock M, Ledergerber B, Schupbach J, Battegay M, et al. Cost-effectiveness of genotypic antiretroviral resistance testing in HIV-infected patients with treatment failure. PLoS ONE. 2007;2:e173. pmid:17245449
Elbasha EE, Szucs T, Chaudhary MA, Kumar RN, Roediger A, Cook JR, et al. Cost-effectiveness of raltegravir in antiretroviral treatment-experienced HIV-1-infected patients in Switzerland. HIV Clin Trials. 2009;10:233–53. pmid:19723611
Long EF, Brandeau ML, Owens DK. The cost-effectiveness and population outcomes of expanded HIV screening and antiretroviral treatment in the United States. Ann Intern Med. 2010;153 (12):778–89. pmid:21173412
Bozzette SA, Joyce G, McCaffrey DF, Leibowitz AA, Morton SC, Berry SH, et al. HIV Cost and Services Utilization Study Consortium. Expenditures for the care of HIV-infected patients in the era of highly active antiretroviral therapy. N Engl J Med. 2001;344:817–23. pmid:11248159.
Schackman BR, Gebo KA, Walensky RP, Losina E, Muccio T, Sax PE, et al. The lifetime cost of current human immunodeficiency virus care in the United States. Med Care. 2006;44:990–7. pmid:17063130.
Hutchinson AB, Farnham PG, Dean HD, Ekwueme DU, del Rio C, Kamimoto L, et al. The economic burden of HIV in the United States in the era of highly active antiretroviral therapy: evidence of continuing racial and ethnic differences. J Acquir Immune Defic Syndr. 2006;43:451–7. pmid:16980906.
Simcock M, Sendi P, Ledergerber B, Keller T, Schüpbach J, Battegay M, et al. A longitudinal analysis of health care costs after treatment optimisation following genotypic antiretroviral resistance testing: does resistance testing pay off? Antivir Ther. 2006;11:305–14. pmid:16759046
Haupts S, Ledergerber B, Boni J, Schüpbach J, Kronenberg A, Opravil M, et al. Impact of genotypic resistance testing on selection of salvage regimen in clinical practice. Antivir Ther. 2003;8 (5):443–54. pmid:14640392
Bernard CL, Owens DK, Goldhaber-Fiebert JD, Brandeau ML. Estimation of the cost-effectiveness of HIV prevention portfolios for people who inject drugs in the United States: A model-based analysis. PLoS Med. 2017;14(5):e1002312. Published 2017 May 24. pmid:28542184
Sanders GD, Bayoumi AM, Sundaram V, Bilir SP, Neukermans CP, Rydzak CE, et al. Cost-effectiveness of screening for HIV in the era of highly active antiretroviral therapy. N Engl J Med. 2005; 352 (6):570±85. https://doi.org/10.1056/NEJMsa042657 pmid:15703422
Barnett PG, Chow A, Joyce VR, Bayoumi AM, Griffin SC, Sun H, et al. Effect of management strategies and clinical status on costs of care for advanced HIV. Am J Manag Care. 2014; 20(5):e129±37. pmid:25326927
Paltiel AD, Weinstein MC, Kimmel AD, Seage GR 3rd, Losina E, Zhang H, et al. Expanded screening for HIV in the United StatesÐan analysis of cost-effectiveness. N Engl J Med. 2005; 352(6):586±95. https://doi.org/10.1056/NEJMsa042088 pmid:15703423
Becker SH, Arenson RL. Costs and benefits of picture archiving and communication systems. J Am Med Inform Assoc. 1994;1:361–71. pmid:7850560
Kaushal R, Jha AK, Franz C, Glaser J, Shetty KD, Jaggi T, et al. Return on investment for a computerized physician order entry system. J Am Med Inform Assoc. 2006;13:261–6. pmid:16501178
Maviglia SM, Yoo JY, Franz C, Featherstone E, Churchill W, Bates DW, et al. Cost-benefit analysis of a hospital pharmacy bar code solution. Arch Intern Med. 2007;167:788–794.43. pmid:17452541
Miller RH, West C, Brown TM, Sim I, Ganchoff C. The value of electronic health records in solo or small group practices. Health Aff (Millwood). 2005;24:1127–37. pmid:16162555
Wachter R. 2015. Why Health Care Tech Is Still So Bad. New York Times, March 21. [cited 2016 July 4]. Available from: http://www.nytimes.com/2015/03/22/opinion/sunday/why-health-care-tech-is-still-so-bad.html?_r=1
Sadoughi F, Nasiri S, Ahmadi H. The impact of health information exchange on healthcare quality and cost-effectiveness: A systematic literature review. Comput Methods Prog Biomed 2018 Jul;161:209–232. Epub 2018 Apr 28. Review. pmid:29852963.
Agency for Healthcare Research and Quality. Costs and benefits of Health Information Technology. April 2006. [cited 2019 March 22]. Available from: https://www.ahrq.gov/downloads/pub/evidence/pdf/hitsyscosts/hitsys.pdf
Miller RH, West CE. The value of electronic health records in community health centers: policy implications. Health Aff (Millwood). 2007 Jan-Feb;26(1):206–14. pmid:17211030.
Goldzweig CL, Towfigh A, Maglione M, Shekelle PG. Costs and benefits of health information technology: new trends from the literature. Health Aff (Millwood). 2009;28 (2):w282–293.49. pmid:19174390
Rudin RS, Motala A, Goldzweig CL, Shekelle PG. Usage and effect of health information exchange: A systematic review. Ann Intern Med. 2014;161 (11):803–11. pmid:25437408
Rahurkar S, Vest JR, Menachemi N. Despite the spread of health information exchange, there is little evidence of its impact on cost. use and quality of care Health Aff (Millwood). 2015;34 (3):477–83. pmid:25732499
Hersh WR, Totten AM, Eden KB, Devine B, Gorman P, Kassakian SZ, et al. Outcomes from health information exchange: Systematic /Review and Future Research Needs. JMIR Med Inform. 2015;3 (4):e39. pmid:26678413
Adler-Milstein J, Jha AK. HITECH Act Drove Large Gains In Hospital Electronic Health Record Adoption. Health Aff (Millwood). 2017 Aug 1;36(8):1416–1422. pmid:28784734.
Hincapie A, Warholak T. The impact of health information exchange on health outcomes. Appl Clin Inform. 2011;2 (4):499–507. pmid:23616891