Mental Health Screening in the Community
Background
The capabilities approach, as articulated by Amartya sen and Martha Nussbaum, focuses on promoting processes that enable a person to be free to do things he or she can. The ‘capabilities’ are a combination of individual, social, educational, economic, and political factors that can help individuals and communities to grow. According to Nussbaum, the capabilities approach is particularly relevant for people in disadvantaged situations who may need more institutional and social support to flourish. Specific attention has been given to people suffering from mental illness. In the mental health field, though the capabilities approach is consistent with the principles of empowerment and recovery models, there are a growing number of intervention and evaluation models. However, the lack of or rarity of consumer-valued participative models is a serious concern. The present study proposes the Achieved Capabilities Questionnaire for Community Mental Health (ACQ‐CMH), a previously developed measure grounded on a collaborative process between researchers and people with mental illness experience.
Study design and population
The study population consisted of 225 adults (18 years to 6 years) mental healthcare service users, and 44% were female. The participants were drawn from the Portuguese community-based mental health response, based on community programs within urban areas. About 95% of the participants were Portuguese, 90% Caucasian, and 7% were from a Portuguese-speaking African country. The majority of the participants (78%) were single and without children (79%). About 74% knew their psychiatric diagnoses; 45% with a schizophrenia diagnosis and 24% with bipolar disorder. Of all the participants, 63% had a prior psychiatric hospitalization, and in this group, about 30% were hospitalized at least once. The total number of hospitalizations varied between 1 and 23 times. About 52% of the participants lived with the family, 18% in group homes, and 30% independently. The sample was divided between those who were not (55%) willing to move to another housing solution and those who were (45%) willing, mostly toward independent housing (71%). About 53% of the participants had high school education, 43% wished to continue studying, and 79% joined educational/training courses. Regarding the professional status of the participants, 78% were professionally inactive, either unemployed, retired, or receiving a social benefit mostly a disability pension (36%). About 22% (n=22) had work experience, but 8% (n=18) were fully employed, while the remaining were trainees or volunteers. In the fully employed group, 78% were supported by the program’s employment services. On average, the participants had not been working for two decades (M= 19.20, SD = 13.75), although about 55% were willing to start a new job. Utilization of employment service time varied from 2 months to 30 years (M= 6.31 years, SD = 6.44 years).
Sampling
Of the 23 services and programs contacted, a total of 11 community-based organizations (9 from the previous study by Sacchetto et al and 2 new) participated in this study.
Measures
The data collection protocol included five research instruments – ACQ‐CMH‐48, WHOQOL‐Bref, K‐6 Distress Scale, the recovery Assessment Scale (RAS‐P), and the Portuguese version of the Empowerment Scale, (ES‐P). In addition, socio-demographic variables and brief questionnaires covering educational, professional, and housing achievements and goals, mental health diagnoses, hospitalizations, and participation in community mental health programs were included.
Data analysis
Psychometric and validity properties of the six-factor and 48-item versions of ACQ-CMH-48 were examined. CFA with maximum likelihood estimation was used to evaluate the model fit. Chi-square test (χ2), comparative fit index (CFI), the Tucker‐Lewis index (TLI), and the root mean square error of approximation (RMSEA) were applied to analyze the model adequacy. For adequate models, χ2/df was found to be between 1.0 and 2.0. For CFI and TLI, indices below 0.85 indicate poor model fit, the range between 0.85 and 0.90 indicates a mediocre but tolerable fit, while 0.90 or above indicates a good fit. Spearman covariate correlations were used to observe convergent and divergent validity of the AQC-CMH with other related measures. Independent samples t-test was applied to test the capacity of the ACQ-CMH and its factors.
RESULTS
- Sampling adequacy was confirmed by Kaiser-Meyer-Olkin (P = 0.89) and Bartlett’s test of sphericity (P < 0.001). The hypothesized six-factor model of the ACQ-CMH presented a poor fit: χ2(225) = 1609.35, p < .001, CFI = 0.84, TLI = 0.84, and RMSEA = 0.051, 90% CI (0.046, 0.056). Hence, a revised model of five factors was considered.
- The revised five‐factor model called Model 2 presented a mediocre fit : χ2(225) = 1310.41, p < .001, CFI = 0.87, TLI = 0.86, and RMSEA = 0.05, 90% CI (0.044,0.055). Though the majority of the items in Model 2 presented good convergent validity (standardized factor loadings of ≥0.60, there was a high correlation between the optimism and affiliation factors (R =0.88). Hence, the hypothesis of a second-order factor was tested.
- The second order model fit showed almost the same outputs than Model 2 (χ2(225) = 1317.77, p< .001, CFI = 0.87, TLI = 0.86, and RMSEA = 0.05, 90% CI [0.045, 0.055]), hence it was rejected.
- Cronbach’s α and CR results of the five‐factor Model 2 showed high internal consistency (0.94 and CR (0.96). Corrected item-total correlations were between 0.31 and 0.69. Family revealed less reliability (α= .68; CR = 0.69), self determination and control (α= .83; CR = 0.83) showed better results than the practical reason (α= .73; CR = 0.73) and self‐sufficiency and determination (α= .79; CR = 0.80) of Model 1.
- The correlations among the overall ACQ-CMH and its subscales were strong (range = 0.52 – 0.86). All the measures in the study protocol supported convergent validity i.e with RAS‐P (r(208) = .41, p< .01), BUES‐P (r(195) = .32, p< .01), and WHOQOL‐Bref (r(171) = .51, p< .01).
- Variables self determination and control showed a strong correlation to the overall ACQ‐CMH (r(225) = .79, p< .01) and positive correlations with both the RAS (r(208) = .25, p< .01) and ES (r(195) = .21, p< .01).
- A significant strong correlation was observed between the Optimism factor and the RAS subscale of personal goals and hope (r(210) = .48, p< .01). However, the RAS Supportive Interpersonal Relationships subscale and the ACQ‐CMH Family dimension were not strongly related (r(209) = .16, p< .05. Regarding ES, there were strong correlations between subscales self esteem and efficacy (r(199) = .25, p< .01) and optimism and control over the future (r(198) = .31,p< .01) with the optimism dimension of the ACQ‐CMH.
DISCUSSION
This paper reports on a consumer-valued research instrument for the evaluation of community mental health interventions. A structural procedure composed of five factors and 43 items revealed a better model fit than that obtained in a previous EFA study. The five‐factor solution, named Model 2, tested in this study proposed a unique factor called self‐determination and control composed of 10 items with a standardized factor loading above 0.40. People with mental illness lack self-determination, power, and control over their lives. Therefore the priority of any intervention should be to promote empowerment and self-determination, giving people with mental illness greater choice and control over their lives. The results indicated Model 2 to be a better solution, considering both psychometric and theoretical aspects. However, the correlation between optimism and affiliation factors was high. Connell et and Gee et al reported that relationships and a sense of belonging are quite closer to the affiliation factor. The findings of the present study support the reliability, sensibility, and both convergent and discriminant validity of using the ACQ‐CMH as a new evaluative measure for community mental health interventions. The ACQ‐CMH offers a consumer‐centric framework with specific dimensions and indicators of capabilities to use in a routine evaluation setting.
Limitations to this study need to be acknowledged. First, future studies are needed to replicate and extend the findings from this initial evaluation considering its smaller size. Second, other psychometric properties such as responsiveness, predictive validity, and sensitivity to change should be analyzed. The effect of programs such as independent housing and supported employment should be examined in the future.
Oncology Related Tools
Other
Latest Research
