Description of Synthesis
| Step 1. Identification of the needs by narrative synthesis and meta-analysis | |
| Narrative synthesis (main approach) | Through a narrative synthesis, we developed an integrated interpretation of various primary studies from which conclusions may be drawn.47 This synthesis provides qualitative rather than quantitative data.48 We followed a narrative approach that includes textual description of studies on needs of the patient and caregiver, extraction of the needs and their grouping into categories based on their common features (eg, education on disease), and then transforming data into common rubrics (taxonomy development).37 We divided the identified needs into 3 categories: needs of the patient, needs of the caregiver, and needs common to both (needs of the dyad). |
| Meta-analysis | We performed a meta-analysis on quantitative studies only to determine the prevalence of the needs identified. Starting with the prevalence proportions extracted from each study, we used R 3.1.2 (The R Foundation) to calculate the pooled prevalence proportion and 95% confidence interval for each need.49 We employed random-effects models, since the studies were statistically and methodologically heterogeneous. The I2 statistic was used to measure heterogeneity. Considering the context (needs are generally evaluated in the observational studies) the meta-analysis included studies of different designs (eg, nonrandomized studies and surveys).50 |
| Step 2. Description of case management | |
| Narrative description | We described the main features of CM according to the definition of the Case Management Society of America,13 then mapped the various components of CM to the needs identified in Step 1 that they targeted. |
| Step 3. Evaluation of case management effects by narrative synthesis and meta-analysis | |
| Narrative synthesis (main approach) | We followed the same narrative synthesis approach as in Step 1 to develop a textual description of the effects of CM on the subcategories of needs: mapping the components of CM to the needs identified in Step 1 from the perspectives of the patients and their caregivers.37 |
| Meta-analysis | We then conducted a meta-analysis to explore the potential effects of CM on the identified needs where data were available. At this point, we included only randomized controlled trials at 12 months follow-up. Meta-analyses were conducted on the effects of CM on needs (eg, the need for confidence in caregiving) and on surrogate markers of needs (eg, behavior management as a surrogate for behavior disturbance of the patient and emotional support as a surrogate for depression and burden of the caregiver). To evaluate the intervention effect, we calculated the standardized mean differences between groups along with 95% confidence intervals, since different scales were used to measure the same outcome. To obtain missing data, we contacted the authors. Fixed-effects models were used because of the small sample size (2 in each outcome evaluation).50,51 |
CM = case management.