A pilot with computer-assisted psychosocial risk –assessment for refugees

Study site

The study was conducted in collaboration with Access Alliance Multicultural Health and Community Services in Toronto. This is a Community Health Centre (CHC) with a multidisciplinary model of care. The staff includes nurses, physicians, psychiatrists, dieticians, social workers, interpreters, peer-outreach workers, and settlement workers. The CHC provides primary health care, community outreach programs (e.g., peer support groups, language classes, and expressive arts programs), and settlement services to immigrants and refugees. The research ethics approval was obtained from the affiliated academic institution.

Intervention

The study intervention was a touch-screen self-assessment survey which Afghan refugee patients completed on a touch-screen iPad in Dari/Farsi language while waiting to see their medical healthcare provider. The Computer-assisted Psychosocial Risk Assessment (CaPRA) survey had questions on psychosocial risks: substance use, exposure to personal violence, depressive symptoms, food and income insecurity, employment, social network, migration status, and coping. To reduce the social sensitivity, the survey also included questions on cardiovascular risks (e.g., physical activity, weight, diabetes, and hypertension) and road and home safety. The eHealth tool generated two tailored print-outs at the point of care. The recommendation sheet for patients summarized their disclosed risks in simple Dari/Farsi language. For the disclosed risks or concerns, tailored messages were printed to enhance patients’ self-esteem (e.g., no one deserves to be hit or there are ways to get your credentials evaluated) and to encourage discussions with the clinician and/or contacts with psychosocial counselor and community services. The risk-report for medical providers summarized patients’ risks with possible referrals. This was attached to the medical chart prior to the consult.

The development of CaPRA content involved multiple phases and a collaborative process. The team first identified the key psychosocial health issues for refugees by a literature review. In absence of specific guidelines for refugee health at that time, the team relied on community engagement process, overseen by an advisory group, to identify priority areas. We sought multiple perspectives by holding brainstorming sessions with healthcare providers (e.g., family physicians, nurse practitioners, social workers, and settlement workers) and representatives of refugee community and organizations serving Afghan refugees. This led to identification of the priority psychosocial risks for which services were also available at the collaborating site. Next, a preliminary version of the multiple-risk survey and messages were developed followed by translation and back-translation [19]. These were further refined for clarity and acceptance by conducting ten qualitative interviews with refugee clients and providers. The final paper-based versions were then converted into a computerized version in both English and Dari/Farsi languages. Prior to the pilot trial reported here, we tested the usability of iPad version with ten refugee clients and providers and made minor adjustments.

Participants and procedures

All site physicians and nurse practitioners received details about the study and provided informed consent. The providers then attended a workshop, moderated by the collaborating clinician, on the risks included in the CaPRA tool. In the resettlement context of newcomer refugees, the discussion focused on the clinical guidelines and best practices for the risk identification, assessment and management. Providers were kept blind to the study’s main focus of intervention’s effect on patient intention to see a psychosocial counselor.

Afghan refugee patients were eligible to participate if they were over 18 years of age, could speak and read Dari/Farsi or English language, were eligible for federal or provincial health care program, and were visiting a participating provider. A bilingual research assistant (RA) approached the potential participants in the waiting room and applied eligibility criteria. If eligible and willing, interested patients received the study details in a separate room. Those who provided informed consent were randomly assigned to the CaPRA group (intervention) or the usual care group (control) with an allocation ratio of 1:1. Before recruitment, the randomization assignment was computer-generated by an off-site biostatistician using varying block sizes for each provider [20]. These patient assignments were sealed in opaque envelopes that were marked on the outside with a physician number and sequence number. The envelopes were opened by the recruiter after patients’ written consent. Afghan patients assigned to the CaPRA group completed the computer survey by using a touch screen iPad. The computer-generated risk-reports were attached to the patient’s medical chart. These patients also received a computer generated recommendation sheet. Patients assigned to the control group continued to receive usual care with no risk assessment before the consultation. Patients in both groups completed a paper-pencil Exit Survey after the visit. Each participant received $30 honorarium and a resource list for community-based services.

Data collection and outcomes

Exit Survey collected information from all participant patients. The section on demographics had questions on age, gender, marital status, education, source of income, English fluency, number of years lived-in-Canada, and use of computers. The section on psychosocial health included questions on patient self-rated health, depression, exposure to violence, purpose of visit, provider discussion on psychosocial issues (i.e., mental health, stress, violence, income, work, language, job, or school), satisfaction with the received care, and patient intention to see a psychosocial counselor. Appointments were made by medical secretary for participants who expressed intention to see a counselor. Questions about patient acceptance of the tool were included for the CaPRA group.

The primary outcome of patient intention to visit a psychosocial counselor was measured by a single item (yes/no). The secondary outcome of patient acceptance of the tool was measured in the CaPRA group by using a previously validated Computerized Lifestyle Assessment Scale (CLAS) with 12 items rated on a 5-point scale (strongly disagree, disagree, not sure, agree, strongly agree) [21]. The three subscales of CLAS are: 1) Benefits (6-item) which represents patient perceived benefits toward the quality of medical consultation and means of achieving them; 2) Privacy-Barrier (3-item) which covers patient concerns about information privacy; and 3) Interaction-Barrier (3-item) which represents concerns about potential interference in the interaction with the healthcare provider. The other secondary outcome of patient satisfaction with the visit was assessed by a 5-point scale (very unsatisfied, unsatisfied, neutral, satisfied, and very satisfied).

Statistical analysis

As this study was a pilot trial, we aimed to recruit 50 refugees based on “Recommendations for Good Practice” by Lancaster et al for pilot randomized trials [22], consistent with others [23].

The Exit Survey data was analyzed using Statistical Package for Social Sciences (SPSS version 18). We executed descriptive statistics (proportions and means) and two-group comparisons using Chi-square and Student t-test. The two-group comparison was not executed for variables where participants could select more than one response (e.g., sources of income, and reason for visit). Some response categories were collapsed due to small sample size (e.g., use of computers).

Participant demographics

Overall, the participants were 37.6 years of age (SD 13.7) and had lived 3.4 years (SD 1.3) in Canada. Majority of them were currently married or in a relationship (69%), had children (73.4%) and reported high school or less education (65%). Most of them were unemployed (84%) and received financial support from social welfare (54%), refugee assistance plan (10%), Ontario disability program (10%), employment insurance (2%), Canada pension plan (4%) and/or family (14%). Participants rated their English language abilities as ‘fair’ on a 5-point scale with a mean of 2.3 (SD .97). Almost one-third were using computers every day (30.6%) while a similar proportion (28.6%) had no previous experience.

Participant psychosocial health

Overall, the participants rated their health as ‘fair’ on a 5-point scale with a mean of 2.3 (SD 1.1). In response to 2-item depression screen, a large number reported feeling down (69.4%) or a lack of pleasure (62.5%) during the last two weeks. Fifty-five percent (27/49) reported both of these symptoms; more among women (19/29) than men (8/20). One fourth of the participants (12/49) had experiences of personal violence during the last 5-years; more among women (9/29) than men (3/20). Participants identified the perpetrator as a stranger (22.4%) or someone they knew (11.9%). The most common reason for their visit on the day of recruitment was to have a routine physical checkup (51%) or a follow-up consult (40.8%). Majority of the participants (69.6%) reported that their provider discussed psychosocial issues during the visit. On comparing the two groups, participants were similar in the self-rated health, depressive symptoms and exposure to violence.

Outcomes

Limitations and challenges

The study findings should be interpreted in light of the design limitations and the context of place and time. The study was a pilot trial with a select group of refugees visiting a single Community Health Centre. This limits the generalizeability to all refugee groups or newcomers. Yet, participants’ response rate of 78% indicates its likely acceptance across vulnerable communities. The interpretation of results warrant caution due to small sample, differences between two groups and volunteer bias of the participant providers. Although statistically significant difference in the demographics of two groups was found only for the number of years lived-in-Canada, we noted some differences in age, gender and level of education. Future research with a larger sample size is needed to allow control of potential confounders. Further, a follow-up component should be included in future designs to assess the impact on patients’ quality of life overtime once they access the referred or suggested psychosocial services.

It is also important to review some field challenges to inform future work. The study applied narrow eligibility criteria (e.g., exclude new patients) and focused on Afghan refugees due to resource limitations in developing a language specific tool. Consequently, few patients were eligible out of nearly two hundred approached in the waiting room. The focus on one sub-group of patients also inhibited the collaborating site, in compliance with the health information and privacy act, to send any pre-visit information letters to patients about the study to encourage early arrivals. Further, a number of practicing clinicians moved to a different clinic along with their patients during the recruitment phase. Although the clinic hired new providers and accepted new patients within a short time, the study recruitment criteria did not allow inclusion of new patients or those seeing new providers. This lengthened the recruitment time. Future research should consider multiple health centers and broader eligibility criteria to address such field challenges. Future models could explore offsite completion of the computer-assisted health risk assessments. However, assessment of socially sensitive issues might not suit non-synchronized approach because provider’s prompt response might be needed. Finally, additional technological advances should also be incorporated, such as linking the printouts to electronic medical records to reduce the documentation burden on the clinician and the use of voice to facilitate reading by patients and to address literacy issues.