Health Information Technology for Continuous Quality Improvement of HIV Treatment Programs

In this single-center randomized controlled trial, the efficacy of a weekly text message is compared to usual care. Patients in the intervention group will receive a motivational text message reminding them to take their medications. Adherence to ART will be compared between the two study groups.

Patients initiating ART at three clinics in Kenya were randomized into two groups. The first group received weekly text messages that asked, “How are you?” The second group received usual care. Patients receiving the text messages were required to respond within 48 hours. A clinician contacted the patient if the patient responded negatively or failed to respond. Analyses indicate that for patients receiving text messages, self-reported adherence of greater than 95% increased by 12% and viral suppression increased by 9%, as compared to usual care.

Technology websites:
No technology website available; for this research, the client's existing cell phone provider was used.

GIS was used to determine the proportion of the population accessing ART and the geographic variation in ART uptake in northern KwaZulu-Natal, South Africa. Demographic, HIV surveillance, and GIS data sources collected since the initiation of ART roll-out were linked. The mean ART uptake among HIV-positive adults was 21.0% (19.2% men, 21.8% women). Individuals living further from a primary healthcare facility were less likely to access ART (odds ratio  = 0.728, p = 0.002). Remarkably little difference was observed across a series of socio-demographic factors, including level of education, household assets, and urban/rural locale.

The Open Medical Record System was implemented at 16 sites in Rwanda to improve direct patient care. Implementation included focus on data quality, access of laboratory results, clinical decision making, and HIV testing. Ongoing assessment of clinical data has led to the identification of 15 new pediatric HIV diagnoses, a 34% increase in communication of CD4 test results to clinicians, and a 92% decrease in data errors.

GIS was used to assess the distribution of AIDS cases and living conditions in Campinas, Brazil. Maps of the region and the male-to-female ratio of HIV cases indicate a trend toward HIV affecting more women and poorer individuals. In this context, GIS may be used to inform the implementation of prevention and control measures.

A touchscreen EMR was implemented at the point-of-care to support ART program scale-up in Malawi. Upon the initial implementation, feedback from clinicians led to corrective action including a data validation system and modification of software logic. Evaluation of the system concluded that the EMR satisfied patient care and program monitoring objectives.

Text messages and interactive voice response were assessed for monitoring adherence to ART among caregivers of HIV-positive children in Uganda. On a weekly basis, caregivers were asked to respond to one question about their child’s adherence. Data indicate that despite a high level of interest, the rate of response to the adherence questions was low (24%). Participants expressed confusion about the use of personal identification numbers and mobile data collection. Additional patient eduction may be required to use cellular phones to support ART adherence.

EMR-based, patient-specific summaries were developed and implemented for providers caring for HIV-positive adults in Uganda. Provider feedback and workflow analysis informed the development of the summaries. The impact of the summaries was evaluated by a time-motion study, which found an increase in direct care of the patient (2.9 versus 2.3 minutes) and a decrease in the total length of the patient visit (11.5 minutes). Overall provider satisfaction with the clinical summaries was high.

GIS was used to identify clusters of cases of mother-to-child HIV transmission in southern Brazil. Data from national databases was used to assess the differences between social, healthcare delivery, and epidemiologic indicators. Analyses indicate that prenatal care attendance was lower among women with HIV compared to all other women. High birth rate (p < 0.01) and poor socioeconomic conditions (p < 0.05) were associated with higher transmission rates.

The DREAM software was developed to manage patient medical information, support HIV treatment, and facilitate epidemiologic analyses. Software applications include the ability to manage patient appointments, drug storerooms, and language preferences. Special attention was paid to establishing a stable power supply, database development, and data quality and security. User training and feedback were critical to system implementation and sustainability. The DREAM software has increased clinical efficiency and streamlined program level monitoring and evaluation.

An EMR was developed and implemented for the care of HIV-positive children and adults in Haiti. The design and development of the EMR was driven by the importance of providing clinical information during patient care and population-level aggregation of clinical data to support monitoring and evaluation. The system uses a hybrid model, which incorporates structured data entry as well as text fields for maximum utility. A phased approach was used for system implementation to incrementally adjust workflow. Staff, time, and money were challenges to system implementation. Making the system useful for clinical care and monitoring and evaluation required a significant time investment to enter patient data. Physicians developed a workflow for computerized physician order entry.

An HIV-specific EMR was designed and implemented for use in the Philippines. Pilot testing lead to system refinement. Staff were trained to use the EMR, including data entry and coding. Reports generated by the EMR may be electronically forwarded to a central facility for data analysis.

An EMR was implemented to support an HIV treatment program in rural Rwanda. The EMR architecture, OpenMRS, is flexible and easily scalable and does not require a programmer to add new data items or data collection forms. The system includes a laboratory data collection application that can alert clinicians to critical results via text message, reporting tools that support data analysis, and a pharmacy tracking system. Synchronization of data between sites and unreliable Internet connectivity have been challenges to implementation. Incorporation of patient care and reporting functions into a single system reduced costs and the complexity of implementation and maintenance.

Perceptions of the use of cellular phones and the Internet to support adherence to ART was assessed with qualitative interviews of adults with HIV followed at two clinics in Peru. Most patients (74%) were willing to receive text or recorded voice medication reminders. Additionally, most patients (81%) expressed interest in receiving cellular phone–based messages about sexual health. Cellular phones were perceived to be confidential.

A web-based EMR was developed and implemented in rural Haiti to support HIV treatment, clinical communication, data analysis, and drug supply management. The web-based approach supports remote data entry, and an “Offline EMR” facilitates data entry during network connection interruptions. The EMR database may be used to monitor individual patients or the entire clinical cohort via a library of automated queries. This EMR demonstrates that effective information management is possible without modern infrastructure.

A pilot study of an electronic pill box, a wireless device that uses a cellular signal to transmit a record of when the pill box was opened, to monitor adherence to ART was conducted among patients in rural Uganda. Short battery life and signal transmission interruptions were identified as barriers to use of the system. System modifications were implemented and continuous data collection was achieved. Comparison of percent adherence as measured by the electronic pill box (93%, interquartile range  87 to 97%) was similar to adherence measure by unannounced pill count (100%, interquartile range 99 to 100%), self-report (100%, interquartile range 100 to 100%), and prior standard electronic monitoring (92%, interquartile range 79 to 98%).

The use and cost of medication event monitoring systems (MEMS) among pediatric patients on ART was compared to four other measures of adherence: pharmacy refill, a measure of returned medications, caretaker self-report, and a visual analog scale. Adherence to ART was 100% as recorded by the four other measures. MEMS recorded 92% adherence with only 66% of doses taken at the prescribed time. MEMS were more expensive than the other measures.

The use of cellular phone medication reminders was evaluated with a randomized controlled trial among patients initiating ART in South India. Usual care was compared to a weekly automated call with a picture message. Virologic failure, self-reported adherence, and cost-effectiveness was compared between the two groups.

Technology websites:
No technology website available; research based on use of own cell phones.

Task shifting, via community care coordinators using PDAs, was evaluated for the care of HIV-positive patients on ART in Kenya. Care coordinators were selected if they had a secondary education, and 100% self-reported adherence to ART for six months. The coordinators participated in a structured training, including instruction on the use of a PDA with data collection and alert-based decision support functionalities. Intervention patients received a monthly visit from a care coordinator who recorded patient symptoms, vital signs, adherence, and opportunistic infection prophylaxis. As compared the usual care group, there were no significant differences in viral load, CD4 count, change in ART regimen, new opportunistic infection, or pregnancy (p < 0.001). Additionally, the number of clinic appointments was decreased by half, suggesting that the use of mobile technology to support task shifting is safe and effective.

Technology websites:
No specific technology website available; research used PDAs.

GIS was integrated with an HIV database to support an HIV program in Cape Town, South Africa. The GIS was used to identify health centers, transport routes, and basic amenities. The database was designed to manage clinical information including patient health, medications, adherence, appointments, and symptoms. Integration of the GIS and the database allowed visualization of the epidemic in relation to the availability of health services and basic amenities. System users identified concerns about transitioning to a computer-based system and patient confidentiality. This integrated system may support decision making and policy management at the programmatic level.

Mobile phone access and patterns were assessed to ascertain whether phone calls or text messages were a feasible method to improve clinic attendance and adherence among patients followed at an ART clinic in rural Uganda. Of patients surveyed, 64% had access to a mobile phone and all but one were willing to be contacted with appointment reminders. Among patients followed prospectively, 11% of appointments were missed; of those, 79% of patients returned to the clinic, where the mean time to returning was 2.2 days (1.2 days standard deviation). Patient privacy and confidentiality were not considered barriers.

Technology websites:
No technology website available; research subjects used own cell phones.

An EMR was developed and implemented to facilitate HIV care and treatment programs across 18 sites in Kenya. The EMR supported clinical care with functionality to identify patients with missed appointments, support adherence to ART, generate patient summaries, and remind providers to order laboratory tests. Additionally, EMR data are available to support program oversight, quality improvement, and monitoring of performance and productivity.

Adherence to ART, as measured by electronic adherence monitoring, was compared to pill count and self-report among patients followed in a clinic in Blantyre, Malawi. The average adherence was 88.1% for electronic adherence monitoring and 98.6% for pill counting, and 100% of patients reported taking their medications in the previous day according to the self-report questionnaire. Pill counts and self-report were not significantly associated with adherence recorded electronically, suggesting overestimation of adherence by pill count and self-report.

The sustainability and scalability of cellular phone–based applications to improve the quality of care were assessed via interviews with patients and providers in South Africa. The interviews indicated that both patients and providers believe that cellular phone applications can support care due to the prevalence of cellular phones and ability for them to function as service enhancement tools. Participants, however, expressed concern about security and confidentiality and perceived that the cost for adoption and sustainability of the technology should be absorbed by the government or other sponsors.

An information management system was implemented by the Chinese Centers for Disease Control and Prevention during the roll-out of the pediatric ART treatment program in China. The information system supports data collection, reporting, and management. The system is utilized for program development, evaluation, and planning.

A monitoring information system was developed in Myanmar to assess the level of harm reduction services available to curb the spread of HIV among people who inject drugs and to facilitate the evaluation of harm reduction programs. The information system was developed with special consideration given to data standardization, modular structure, system flexibility, system interface, and data analysis and reporting capabilities. Challenges included maintaining system simplicity while balancing the ability to modify the system at a later date and patient confidentiality in the context of collecting data about illegal drug use. Following implementation, data regarding coverage, services, and equipment were collected using paper-based data collection forms that were then entered into the system. The data was then available for analysis to assess process indicators related to service delivery and coverage of harm reduction programs.

A clinical decision support system was developed to support CD4 testing guidelines in western Kenya. The system, which was integrated with an EMR, generated printed summaries and reminders for overdue CD4 cell tests. Comparison of testing rates in a clinic where the system was implemented, to a clinic where the system was not implemented, indicated statistically significant higher rates of CD4 test ordering (odds ratio=1.80, 95% confidence interval 1.34, 3.22, p<.0001). As compared to baseline, there was a statistically significant increase in CD4 test ordering after system implementation (odds ratio=2.32, 95% confidence interval 1.67, 3.22, p<.0001).

Cellular phones were used to deliver weekly adherence reminders using interactive voice response (IVR) and  non-interactive picture-based text messages to patients on ART in Bangalore, India. Study participants were required to report whether they had taken all of their medication in the previous day. Ninety percent of participants reported that the intervention was helpful, and that it was not an invasion of privacy. The majority of participants (87%) preferred IVR over the non-interactive picture-based text messages (11%).

Technology websites:
No specific technology website available; researchers used own cell phones.