Patient navigators for people with chronic disease: A systematic review

1Chronic diseases, including physical and mental illnesses, are a significant burden to both patients and the health care system. There were an estimated 12.7 million new cases of cancer worldwide in 2008; and diabetes prevalence was estimated to be 6.4%, affecting 285 million adults worldwide in 2010. People with chronic diseases have increased morbidity and consume substantially more health care resources than those without. Adherence to evidence-based recommendations for clinical care is associated with better outcomes and lower resource use for patients with chronic diseases. For example, clinical trials show that in patients with diabetes, tight control of blood pressure, use of statins and achieving good glycemic control improves outcomes and lowers costs. Despite widespread dissemination of practice guidelines, many people with chronic diseases do not receive or adhere to recommended care.

2This difficulty in implementing evidence-based care may be due to a combination of patient, provider and system level barriers. Patient level barriers may include lack of awareness of publicly funded programs (including community-based resources), financial constraints, competing priorities (e.g., family and work), personal circumstances, language and culture (i.e., race/ethnicity); such barriers could make it challenging to follow even seemingly simple lifestyle recommendations. At the provider level, barriers may include lack of clinical decision support systems to implement recommended care, lack of time and knowledge. System level barriers include the inherent complexity of the health care system and suboptimal access to primary or specialty care.

3Patient navigators are trained personnel who help patients overcome modifiable barriers to care and achieve their care goals by providing a tailored approach to addressing individual needs. Navigators may be nurses, social workers or lay health workers, including peers. Patient navigator programs were originally established to reduce gaps in timely cancer care among marginalized populations, and are increasingly in use across the United States and Canada within the cancer field. Patient navigation is also currently used for diabetes, smoking cessation and cancer screening. Depending on the targeted barriers, specific tasks may include one or more of: disease education, health system education, removal of medical system barriers, assistance with insurance coverage, addressing other financial barriers, aiding in care coordination, referral to community resources, and providing emotional support, among others.

4Previous reviews of patient navigators have focused on cancer care, though patient navigators are increasingly being utilized in other areas. While their popularity is growing, and many are touting their benefits, it is not clear whether patient navigator programs are beneficial across a cross-section of chronic diseases. As there is wide variation in the design and implementation of patient navigator programs in various chronic diseases, a systematic review is needed to summarize the characteristics of programs and their effectiveness. In this systematic review, we assess the effectiveness and attributes of patient navigator services, compared with usual care, on patient-oriented outcomes and processes of care in patients with chronic diseases.

5We searched MEDLINE, EMBASE, The Cochrane Central Register of Controlled Trials (CENTRAL), CINAHL, PsycINFO, and Social Work Abstracts up to August 23, 2017 with no language or date restrictions. In our MEDLINE search strategy, we included potential synonyms for patient navigator (case management, care coordination, health coach), terms for the set of chronic diseases of interest, and restricted the search to randomized controlled trials (S1 File). The MEDLINE search strategy was peer reviewed via PRESS. We also systematically searched reference lists of included studies and relevant reviews.

6We included randomized controlled trials (RCTs) evaluating the effectiveness of a patient navigator program compared to usual care. Study population could be adult or pediatric patients, that either had or were being screened for one of the following chronic diseases, as included in the Statistics Canada Canadian Community Health Survey: asthma, arthritis, hypertension, migraine, COPD/emphysema, diabetes, heart disease, cancer, intestinal/stomach ulcers, stroke, urinary incontinence, inflammatory bowel disorder, dementia, mood disorders, anxiety disorders; with the addition of HIV/AIDS, and chronic kidney disease, which includes transplant recipients and patients on dialysis.

7There is currently no standard definition of a patient navigator, thus there is variability between patient navigator programs, as well as overlap with programs under different names. We defined a patient navigator as a person with or without a healthcare-related background that engages with patients on an individual basis to determine barriers to accessing care or following recommended guidelines. The patient navigator also provides information relevant to patients’ specific circumstances to facilitate self-management and access to care. We were flexible in terms of the name of the intervention used by study authors (i.e., patient navigator, community health worker, etc.), as long as program descriptions were consistent with our definition. Studies were excluded if they evaluated programs where patient navigators performed clinical care (i.e., prescribed medication, ordered diagnostic tests, performed physical measurements), or where the role was not formalized (i.e., casual or untrained support).

8Two reviewers independently screened all titles and abstracts of retrieved references. Two authors then applied the full set of inclusion and exclusion criteria to all articles chosen for full text review. Reviewers resolved any disagreements by discussion.

9Data extraction was done by one reviewer, using standard data extraction forms and verified by a second reviewer. For studies with several trial arms, data were collected across each relevant comparison. Data elements included characteristics of the study, outcomes and results, along with details of the navigator program. Outcome measures of interest fell into one of three broad categories: patient-oriented (mortality, health-related quality of life, and complications of disease, e.g., MI, stroke); surrogate outcomes (e.g., achieving target blood pressure or glycemic control); and process measures, including access to appropriate services, and adherence to recommended clinical actions (e.g., cancer screening). Measures of patient experience and patient satisfaction were also collected. The risk of bias criteria suggested by the Cochrane Effective Practice and Organisation of Care Group (EPOC) were used to assess study quality. Items retained from the tool to assess bias were: random sequence generation, allocation concealment, blinding of outcome assessment, incomplete outcome data, group similarity at baseline and intention-to-treat analysis. Risk of bias in each domain was assessed as high, low or unclear.

10Though the goal of the systematic review was to provide a quantitative assessment of the effects of an intervention, we found a heterogeneous group of programs, chronic diseases and outcomes, and we therefore used a narrative approach to data synthesis. To assist in assessing effectiveness across this large number of studies, we tabulated the primary outcome of each of the studies, a summary of the result, and whether the changes were statistically significant. We determined the proportion of studies with positive outcome results (primary or secondary) in each outcome category, stratified by chronic disease. We explored the association between program features and a statistically significant improvement using logistic regression.

11Given the heterogeneity in outcomes, it was not possible to assess publication bias using a traditional funnel plot. To provide an estimate of publication bias, we divided the studies into quintiles of sample size and compared the proportion of studies reporting a positive statistically significant effect across quintiles. All work aligned with a protocol that was developed and published ahead of the review.

12We identified and screened 14,672 potentially relevant abstracts. Seventy-four papers describing 67 unique studies met our inclusion criteria, and were included in the review (Fig 1). Table 1 summarizes the characteristics of the included studies, while Table 2 provides an overview of the individual studies, grouped by disease. The vast majority of studies (90%) were conducted in the United States and sample size varied from 21 to 16,267 participants, with the majority of studies (52%) including between 100 and 500 participants. A summary of the quality assessment is presented in Fig 2 and a detailed assessment by study is presented in the Supplementary Table (S1 Table). Though all studies were RCTs, quality varied, and many studies were lacking information on allocation concealment and blinding of outcome assessment.

13Patient navigation has been tested through RCTs more commonly in the context of cancer care (66%; n = 44) than in any other chronic disease. Of the cancer care studies, the majority were in cancer screening where the patient navigator’s focus was on helping the patient complete the screening test. Other chronic diseases where patient navigators have been studied include diabetes (n = 8), HIV (n = 7), cardiovascular disease (n = 4), chronic kidney disease (n = 2), dementia (n = 1) and multiple chronic diseases (n = 1).

14Most navigator programs (64%) employed lay persons trained for the role. The primary mode of communication was by phone (90%) and over half were based in primary care or the community (57%). Patient navigators were responsible for a wide variety of activities. The most frequent strategy used by patient navigators to address health system barriers was care facilitation (i.e., making referrals, communicating with providers, coordinating care), followed by appointment scheduling (S2 Table). The most common activities used to address patient barriers included addressing patient attitudes and beliefs, appointment reminders, health literacy support and practical assistance (e.g., assistance with transportation, coordination of dependent care, arrangements for financial help or insurance benefits). Patient navigators most often provided education about the tests and treatments required in the form of discussion with patients. Many patient navigators also provided some form of direct psychosocial support to their patients.

15Many studies (n = 26) reported using patient navigators that were culturally aligned, that is, a patient navigator who identified with the patient population in terms of ethnicity or other cultural factors, or included educational materials or communication approaches that were culturally tailored. Frequency of contact between navigators and patients ranged widely from only one contact to ‘as needed’ during the study duration, and duration of navigation varied widely.

16We found significant heterogeneity in primary outcomes. With respect to patient-oriented outcomes, one study included death as a primary outcome, two hypoglycaemia, and five studies assessed quality of life and/or health status as a primary outcome: two in patients undergoing cancer treatment, one in caregivers supporting patients undergoing cancer treatment, one in stroke, and one in patients with multiple chronic diseases. Surrogate outcomes were most often reported in diabetes, where change in A1C levels was reported in seven studies; three studies in HIV reported viral load and one study in CKD reported change in estimated glomerular filtration rate (eGFR). Process outcomes were the most frequently reported primary outcome (n = 50), and they included completion of disease screening and adherence to follow-up procedures. Patient satisfaction or experience was reported as a primary outcome in one study, while hospitalizations and emergency room visits were reported as primary outcomes in three studies. Of the 67 unique studies identified in this review, 45 (67%) reported a statistically significant improvement in one or more primary outcomes. We did not find an association between any program characteristics and the finding of a statistically significant improvement in a primary outcome.

17Secondary outcomes were broader in scope, although many were variations of the primary outcome: for example, diagnostic resolution within a specified time period (where the primary outcome was time to diagnostic resolution). Secondary outcomes more frequently included patient-reported outcomes, including physical and mental health status, quality of life, and psychological distress. Other patient-oriented outcomes were reported as secondary outcomes: diagnostic outcomes of cancer screening and follow-up were reported in three studies, use of acute care was reported in four studies, mortality in two and rate of opportunistic infections was reported in one study Costs were considered in six studies.

18Fig 3 depicts the number of studies that included outcomes within each category of interest (either as primary or secondary) and the proportion of these that demonstrated a statistically significant improvement. Studies were more likely to report positive results for process measures, and less so for surrogate markers, health care utilization, or patient-oriented outcomes. No studies found a negative impact from the patient navigator intervention.

19With respect to publication bias, the proportion of studies reporting a statistically significant improvement in a primary outcome, across sample size quintiles, was 57%, 56%, 68%, 83%, and 69%.

20Healthcare systems worldwide are under tremendous strain to address the needs of patients with chronic diseases. Since improving access to care and adherence to recommended treatment may improve outcomes, there is great interest in navigator programs. Our review of patient navigators for people with chronic diseases identified 54 unique randomized controlled studies. While most studies reported a positive effect of patient navigator programs for their primary outcome, the impact on clinical outcomes remains uncertain. The majority of the outcomes measured in RCTs of patient navigator studies reflect the process of receiving care; few studies assessed patient-oriented outcomes and many studies had short duration of follow-up, with uncertain power to detect an effect on clinical outcomes.

21The variability of patient navigation programs found in the existing literature makes it challenging to make definitive statements about their effectiveness. We propose a consistent definition for a patient navigator, i.e., a person with or without a healthcare-related background that engages with patients on an individual basis to determine barriers to accessing care or following recommended guidelines, and provides information relevant to patients’ specific circumstances to facilitate self-management and access to care. The primary focus is on overcoming barriers, not providing clinical care, and in doing so patient navigators are often a source of social support.

22Most other reviews of patient navigation have been restricted to cancer. The reviews by Wells and Paskett identified 33 studies of patient navigator programs in cancer, not restricted to RCTs. Even when examining patient navigation within one chronic disease (i.e., cancer), there was wide variation in intervention design, study design and study quality. In our review, a higher proportion of studies in cancer prevention or management (n = 32/44, 73%) reported a positive statistically significant effect for one or more primary outcomes versus studies in other chronic diseases (n = 13/23, 56%). Ali-Faisal et al published a systematic review and meta-analysis summarizing the effect of patient navigation programs on health care utilization, including adherence to screening and follow-up care, outcomes that our review classified as process outcomes. Programs were required to use the term navigation or a variant in their description to be included in the review. The authors found that patient navigation was effective in increasing screening rates and improving adherence to recommended care; however, the effect on other health outcomes was less convincing and they noted considerable heterogeneity across studies. Our review of patient navigator programs expanded across chronic diseases and despite similar heterogeneity, echoed many of these previous findings and identified similar limitations.

23Our study had strengths and limitations. Strengths include the thoroughness of our literature search, and our consideration of a broad group of diseases. A quantitative synthesis may have helped identify factors associated with successful patient navigator interventions, however this was precluded by the heterogeneity in both the intervention design and the outcomes reported. This heterogeneity also made it difficult to make definitive statements about the merit of specific patient navigator activities. We were unable to identify the most important elements of patient navigator programs that were associated with an improvement in the primary outcome. Other potential reasons include incomplete reporting (i.e., some program elements may have been present but not reported) or variation in how the individual features were implemented within programs. Though we were unable to summarize our results quantitatively, a descriptive review of these randomized controlled trials provides a comprehensive summary of navigator programs and outcomes reported. Our review was limited to published reports of randomized controlled trials, and therefore, although we noted a trend toward a lower proportion of small studies reporting statistically significant positive results, we could not rule out publication bias.

24Our findings indicate that patient navigator programs improve processes of care, although few studies assessed patient experience, clinical outcomes or costs. The inability to definitively outline successful components remains a key uncertainty in the use of patient navigator programs across chronic diseases. Given the increasing popularity of patient navigator interventions, future studies should use consistent definitions for patient navigator interventions, and in addition to determining which elements of the intervention are most likely to lead to improved outcomes, studies should focus on patient experience and disease-specific clinical outcomes that are important to patients.