Clinical Guide > The HIV Clinic > Quality Improvement

Quality Improvement

Author: Bruce D. Agins, MD, MPH
January 2011

Chapter Contents

Quality Improvement: Raising the Bar

Quality improvement (QI) has become a standard of practice for HIV programs in the United States. Quality management is included in contractual requirements for Ryan White-funded programs and has been integrated into training programs. Many clinicians have learned the basics of quality management, and may be participating in or even leading improvement efforts in their clinics. Moving beyond the basics, however, remains a challenge for clinicians who have limited time to participate in activities not related to direct patient care. Yet, focusing on quality can reveal important phenomena in the clinic of which the leadership is unaware or may point out factors that explain why problems have not been easily resolved.

"Quality of care is the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge."

(Institute of Medicine, 1990)

"All health care organizations, professional groups, and private and public purchasers should pursue six major aims; specifically, health care should be safe, effective, patient-centered, timely, efficient, and equitable."

(Institute of Medicine, 2001)

Increasingly, QI work in the HIV field is associated with important health outcomes such as HIV viral load suppression, retention in care, and reductions in costly service utilization such as hospitalizations and emergency department visits. Simple subanalyses of basic performance data may reveal disparities in how care is being provided to different patient groups in the clinic, for example, according to age, gender, or race/ethnicity. At the same time, advances in health information technology have made it easier to generate data for performance measurement, and for performing the simple analyses that can be used for improvement activities.

This chapter will quickly review the basics of QI; an appendix at the end of the chapter illustrates concepts with examples of more advanced improvement work. Chapter Health Resources and Services Administration HIV/AIDS Bureau Performance Measures in this manual presents a national quality initiative developed by the U.S. Health Resources and Services Administration (HRSA) HIV/AIDS Bureau (HAB) for Ryan White HIV/AIDS Program-funded clinics; HRSA's Core Clinical Performance Measures for Adults and Adolescents can be incorporated into the quality improvement programs of local clinics.

The fundamental concepts of QI have evolved over the past century to include the following:

In addition, documentation of QI activities helps to demonstrate institution-wide compliance with accreditation responsibilities and funding requirements, while often giving the clinic or institution an advantage when competing for alliances with purchasers. Data generated from the clinic's QI program showing improvements over time demonstrate to constituents that the program is successful and help to justify its funding.

Finally, monitoring systems that are dependent on single individuals will not last when these key players leave the clinic or are absent for long periods, whereas a fully functioning QI program that involves staff working in teams with a clearly defined infrastructure will keep going when even the most dynamic individuals depart.

This chapter will articulate the core principles and describe activities that easily can be adapted into the HIV ambulatory care setting to implement a sustainable QI program.

Introduction to Quality Improvement

QI includes regular measurement of care processes and outcomes to analyze the performance of the system of care. It involves the implementation of solutions to improve care and the monitoring of their effectiveness, with the goal of achieving optimal health outcomes for patients. Ongoing cycles of change and remeasurement are implemented to test and try different ideas to determine which practices result in improved care. QI activities in clinics can range from a single team focusing on improving one aspect of care to a comprehensive QI program with many teams working on a wide variety of improvement projects, with a well-established plan and an oversight committee.

The methods of QI are based on core principles that are readily translated into a practical approach and integrated into the clinical care delivery system (see Table 1). Successful implementation of QI involves actions at two levels: the QI activities and the HIV program processes that provide the structural backbone for them.

Table 1. Core Principles of Quality Improvement

  • Focus on the customer: improvement activities result in improved patient health
  • Measurement: collect and use data to improve care
  • Emphasis on systems of care: improve processes that link to desired outcomes
  • Involvement of participants: encourage direct participation in teams by those individuals who implement the processes being evaluated

Measurement and Management Are Not Sufficient

Although it is the bedrock for improving care, measurement alone is not sufficient to improve quality. A common pitfall in implementing QI programs is to rely solely upon performance data, the medical or program director's interpretation of it, and one person's decisions about how to make changes. Successful improvements occur most often when staff members from the systems being assessed work together in teams. When they are engaged in the process, staff members are more likely to generate ideas for improvement and to accept changes. Staff review of improvement charts (see Figure 1, below) generates pride and a sense of accomplishment based on members' participation in the QI work. These charts may be posted on bulletin boards in common areas of clinics so that everyone can view them.

Key Components of a Quality Plan

The key elements of a quality plan include a quality statement that describes the purpose, priorities, and goals of the QI program; a description of the organizational systems needed to implement the program, including committee structure and functions, definitions of accountability, roles and responsibilities, the process for obtaining consumer input, core measures, and data collection processes; and a description of how the plan will be evaluated. For further information, see New York State Department of Health AIDS Institute HIVQUAL Group Learning Guide (see "References," below).

QI Personnel

The size of the clinic will determine who participates in quality-of-care activities. In small HIV clinics with a primary care provider, case manager, nurse, and support staff, most of the staff members are involved in all aspects of QI work. Larger institutions usually establish an HIV Quality Committee that includes senior management of the HIV clinic, designated QI staff if there are any, and other key players who work in the clinic. A member of this committee represents the group in the agency-wide QI committee. The Quality Committee identifies the priorities for improvement or agrees to pursue the priorities identified by staff members or patients in the clinic. The Quality Committee also charters improvement teams and identifies potential members who are key stakeholders in the process under investigation. In a small clinic that has only a handful of staff members, all clinic personnel may participate in the quality management program and in QI activities, although perhaps in a less formal way than in larger clinics.

Team Membership and Responsibilities

Teams are formed to address the specific care processes or systems that are targeted for improvement. Team members should be selected to represent the different functions involved in these processes or to represent the components of the system under focus. The size of a team varies according to the size of the clinic and the process under study. In small clinics, the few dedicated HIV program staff members may constitute the project teams, with added representation from different departments as needed (such as from the laboratory, or from other medical disciplines). In larger clinics, teams often include 6-10 members. Membership should include representatives from the different groups in the clinic who are involved in the care process. In addition to the clinical and case management staff, scheduling clerks and medical records personnel often are important participants, especially when follow-up appointments and documentation are important components of the care process or have been identified as areas that need to be improved.

Involving consumers in QI project teams enhances the work of the team. Consumers who are involved in the clinic's community advisory board often are natural leaders and have a good grasp of clinic processes. Their feedback on the experience of care delivery can reveal areas that need improvement. They know the bottlenecks and can inform the staff how long a clinic visit lasts, whether assessments truly occurred, and whether behavioral interventions are effective. Their ideas about what improves care often diverge significantly from those generated by providers and may not even be recognized unless they participate directly in discussions about the system.

Teams are expected to analyze clinical processes, identify areas of change, implement tests of the changes, review data assessing the change, and ultimately make recommendations about which improvements should be adopted in the clinic.

As the project team conducts its work and gains experience, it will become more independent and assume more responsibility for ongoing measurement, data collection, and implementation of steps toward improvement. It should disband when implementation has resulted in sustainable results.

Data Collection

Selecting Indicators

Indicators are measurable aspects of care that can help to evaluate the extent to which a facility provides a certain element of care. Indicators should be based on standards or guidelines, meet the primary goals of QI, and reflect priorities specific to the community and the clinic. In addition, they should represent processes where changes are feasible. For example, in HIV clinics where the population consists of a large number of women, indicators may include rates of routine cervical cancer screening, rates of preconception counseling, or other aspects of care specific to women. In clinics that care for a high volume of patients who have been treated with antiretroviral therapy (ART) for a long time, indicators may focus on rates of virologic suppression, screening for adverse effects of antiretroviral medications, and resistance testing. Some indicators should be selected by soliciting input from patients who attend the clinic (see Table 2). Staff members also often know what aspects of care would benefit from being measured and improved, and they should be consulted to determine priorities. If routine data collection systems already exist in the clinic, data should be reviewed to determine which components of care would be prime candidates for improvement. On a national level, the HRSA HIV/AIDS Bureau (HAB) has developed HIV/AIDS Core Clinical Performance Measures for Adults and Adolescents for monitoring the quality of care provided in Ryan White HIV/AIDS Program-funded clinics. These can be used as a starting point if local priorities have not been established. See chapter Health Resources and Services Administration HIV/AIDS Bureau Performance Measures for these.

Table 2. Methods for Obtaining Input from Patients

  • Suggestion box
  • Surveys
  • Focus groups
  • Consumer advisory board

Ideally, a balanced set of measures should be selected. Different ways to categorize measures might include the following:

Developing and Measuring Indicators

Three major activities constitute the process of indicator development:

The measurement population is defined by determining factors such as the location of care being studied, whether both men and women are eligible, the applicability of the indicator to various age groups, whether any clinical conditions are necessary to determine whether the indicator is applicable, and whether the patient must have been in treatment or visited the clinic more than once.

After the population is defined, the measure needs to be defined. The measure should be objective and should address specific aspects of quality care. It also should have a straightforward, dichotomous answer. For each measure, specific criteria must be developed to define the "yes" response and the "no" response (see Table 3). This often involves deciding the time period during which an activity has been performed. For example, an indicator that measures viral load monitoring must include the frequency with which that test should be performed. One simple way to construct this measure would be to ask, "Was viral load measured within the past 6 months?"

Table 3. Examples of Specific Measures with Definitions of Responses

MeasureDefinition of MeasureYes/No Response
Pneumocystis jiroveci pneumonia (PCP) prophylaxisDid the patient whose CD4 count was <200 cells/µL during the review period receive PCP prophylaxis?Yes: The patient received PCP prophylaxis

No: The patient did not receive PCP prophylaxis
Latent TB infection (LTBI) screening (eligibility: HIV-infected patients without a history of previous TB disease or a history of a test for LTBI)Was a test for LTBI performed (tuberculin skin test or interferon-gamma release assay) and were results documented in the past 12 months?Yes: LTBI testing was performed

No: LTBI testing was not performed (or the test was done but results were not documented)

Several indicators should be measured simultaneously, whether abstracted from medical records or analyzed through administrative databases. Indicators reflecting different aspects of patient management should be selected, as should those involving different populations. Indicators also should be selected to evaluate various components of the health care system, such as the components of the chronic disease model.

The data collection plan includes determining the source of information (e.g., whether medical records or an electronic database will be used), how the data will be recorded, who will record the data, and how a data sample will be selected. A representative sample will allow inferences to be made about the overall clinic population based on observations of the smaller sample. Some form of random sampling should be used, either by using a random numbers table or by selecting every nth record from the list of eligible patients.

A common pitfall at this point is to think of the measurement sample as a research project. For the purposes of QI, a sample needs only to be current, representative, and readily obtained (i.e., sample size calculations and the achievement of statistically significant results are not necessary).

Analyzing and Displaying Data

Data should be reviewed and distributed to all members of the team and others involved in the care process under evaluation. When possible, data should be displayed in graphic format. After data from multiple time periods have been collected (e.g., percentage of patients screened for latent TB), a simple line graph (run chart) can be constructed with each point representing a performance rate (percentage) for a given period of time. This usually is the simplest and most effective way to show performance data (see Figure 1).

Figure 1. Percentage of Patients Screened for Latent TB, by Month

Figure 1. Percentage of Patients Screened for Latent TB, by Month

Identifying Targets and Implementing Improvements

After the project team has reviewed the data, it must decide where opportunities for improvement exist. The first step in this process is to investigate the care process in greater detail. Several techniques are used to accomplish this goal. The simplest is brainstorming, in which key stakeholders offer their suggestions as to which processes are the best candidates for change. Another easy method is flowcharting, in which the group breaks down the process into its components to identify how it is coordinated and how its parts fit together. A fishbone diagram, or cause-and-effect diagram, may aid in exploring and displaying the causes of a particular problem (Figures 2 and 3). It often helps for staff members to consider factors potentially influencing a process that are not obvious, and to help sort out those factors that are external to the clinic and those that are internal. Then, the areas that would be most likely to benefit from improvement are selected for change (see Figure 4) and tested in the Plan-Do-Study-Act (PDSA) cycle (see Figure 5). See "Appendix," below, for examples of QI projects conducted in HIV clinics.

Figure 2. Sample Fishbone Diagram

Figure 2. Sample Fishbone Diagram

Figure 3. Fishbone Diagram: Low Collection Rates of Race and Ethnicity Patient Data

Figure 3. Fishbone Diagram

Figure 4. Sample Flow Chart

Figure 4. Sample Flow Chart

Figure 5. Plan-Do-Study-Act Cycle

Figure 5. Plan-Do-Study-Act Cycle

After a change to a particular step of the process has been selected, a test of the change can be quickly implemented and evaluated. A limited implementation of the proposed change can be tested--perhaps with just a few subsequent patients, or those attending on the following day, or those seen by a particular clinician. If the small change does not work, another change can be selected and implemented quickly. If the change is feasible and improvement is noted, it can be adopted more widely, before formal remeasurement occurs, and a regular period of remeasurement can be adopted. If the change was not successful, then another one can be chosen and tested. Occasionally, multiple changes may be tested simultaneously or on different days of the week.

Often teams become caught up in following a rigid series of steps. Many different approaches to QI exist and can be implemented successfully. However, sometimes just stepping back to refocus on the three basic questions of the model for improvement presented by Langley et al. (see "Suggested Resources," below) can effectively guide QI activities. These are as follows:

Establishing Systems to Support QI

The key to sustaining QI in the clinic is development of an infrastructure that supports ongoing QI activities. The central components of this infrastructure include the following:

Table 4: Tips for Promoting a Culture of Quality Improvement: Integration of Quality into the Regular Work of the Clinic

  • Educate staff members about QI and provide them with the skills to participate in QI activities
  • Consistently articulate the values of QI in staff meetings
  • Display QI data and storyboards (simple statements and visual representations that describe a problem, the evaluation process, the proposed changes and their implementation, and display the results)
  • Celebrate successes
  • Provide opportunities for all staff to participate in QI teams
  • Reward staff members through performance evaluation for their contributions to the QI program

The regular, ongoing work of the QI committee, supported by the clinic leadership, constitutes the backbone of the infrastructure that supports ongoing QI activities. The committee oversees the dynamic process of planning, implementation, and evaluation that involves the following:

These contribute to sustaining the QI program and its activities in the clinic.

Sustaining Improvements

Sustainability is probably the biggest challenge that clinics face in the field of QI. All too often, improvements do not last after initial projects are completed, because the structure and culture to support QI is not present or is not supported. The challenge of sustainability therefore is twofold: to maintain the successes of QI work and its clinical outcomes, and to maintain the systems of QI and to keep the QI program vital. By asking questions about how care systems can be improved and how QI activities are progressing, clinicians play an important role in both catalyzing and supporting QI activities.

Appendix: Case Examples of Quality Management Initiatives

Using QI to Improve ART Management and Virologic Suppression

In one community health center, nearly 10% of patients on antiretroviral therapy (ART) were not virologically suppressed. The clinic had adopted a strategy of developing a specific ART management plan for each patient being treated. Review of the charts of the 45 patients not suppressed showed that only 40% had a plan in the chart. Only 20% had a plan that was executed. Improvement goals were set for each step, to increase from 40% to 90% for documentation and from 20% to 75% for execution. Over the course of a 4-month period, clinicians and support staff were educated about the plan and decision-support tools were created, including an algorithm showing key decision points for plan development and execution with corresponding prompts in the clinic database. Visit forms were revised to incorporate data fields specific to the ART plans. Reminders also were created. All changes were implemented through small tests of change with formal indicator remeasurement in 6 months; this showed that 100% of patients had a plan in their charts and that 71% of them had it executed. Continued monitoring showed a dip in performance 4 months later to 88% and 60%, respectively, but with gains restored 3 months subsequently to 100% and 65%, respectively. Vigilance and reaffirmation of the main steps have been keys to maintaining performance. The fields are now being added to the new clinic electronic medical record (EMR) system with automatic prompts based on changes in viral load values.

Another clinic focused QI efforts directly on improving rates of virologic suppression. Data on the HIV viral loads of each specific provider were generated to show comparative rates as a stimulus for improved performance on the part of individual providers. ART regimens were reviewed for their appropriateness, and renewed education about antiretroviral drug combinations and resistance monitoring was introduced. The clinicians reviewed their own patient lists each day. Patients who were not virologically controlled were contacted by phone by the clinician or a nurse. Nurses eventually were assigned directly to the primary care team to facilitate communication with patients and ensure that specific issues raised during phone conversations were addressed during clinic visits. Adherence problems were particularly common, and were addressed through these multiple contacts. Reasons for adherence lapses were identified, which allowed for more effective targeting of service interventions with specific patients, including substance use and mental health service referrals, regimen switching, and targeted adherence interventions. The individual providers improved their patients' rates of virologic suppression from 45% to 62%. Review of suppression rates also showed that a subset of patients remained controlled and did not require quarterly monitoring. A decrease in visit frequency was possible for this group, reducing overutilization of services and unnecessary costs.

Using QI to Eliminate Disparities

The quality committee in another clinic wanted to determine whether the clinic's performance was consistent across all patient groups and arranged to have the clinic's patient data sorted by race, ethnicity, and primary language spoken. This revealed that data about race, ethnicity, and language spoken were not recorded in a high proportion of patient records. The team invited patients from their community advisory board to attend a staff meeting where a fishbone diagram was developed to identify potential causes of the poor collection rates of these patient data. Potential reasons were identified in all categories: equipment, patients, procedures, and staff (see Figure 3). A flow chart was developed to identify the sequential steps of data collection. A training program for intake staff was developed, resulting in an improvement in collection of these data to 85%.

Subsequent analysis showed that only 54% of African-American patients and 68% of Latino patients had suppressed HIV RNA whereas white patients had a suppression rate of 75%. A focus group with Spanish-speaking patients revealed that these patients were not getting enough information about medication and its side effects. The QI team decided to aim for an improvement in virologic suppression rates to 75%. A number of changes were implemented and tested throughout the clinic including the addition of peer adherence counseling, using teach-back by non-physician staff to facilitate adherence problem-solving, along with medication reconciliation. With these new interventions, virologic suppression rates improved to 71% for African-Americans, 80% for Latinos, and 81% for whites. With ongoing QI activities, suppression rates have subsequently increased for all groups and the gap has narrowed. Ongoing changes aim to narrow the gap even further.

Suggested Resources


HIV-Specific Examples

General QI Websites