Nelson Systematic Reviews to Answer Healthcare, Questions, 2e

70

Chapter 6 • Selecting Studies for Inclusion

BOX 6.2 Reducing Selection Variability with Dual Review When Eligibility Criteria Are Broad Some systematic reviews are exploratory and their intention is to provide a comprehen sive evaluation of evidence. To achieve this goal, the study eligibility criteria need to be framed in broad terms. However, broad criteria can increase selection variability, as illus trated in a review of the effectiveness of intensive primary care programs in reducing hos pital admissions and/or death in high-risk patients with multiple chronic conditions and frequent hospital admissions. 18 The purpose of the review was to assist a health delivery committee in developing a primary care intensivist model for pilot testing. The eligibility criteria are described in Table 6.4.

TABLE 6.4 INCLUSION CRITERIA Population

Patients identified as high risk for hospital admission and/or death, regardless of whether they have a specific disease, such as heart failure

Intervention

Multicomponent, interdisciplinary intensive primary care programs

Comparator

Usual care (without the utilization of an intensive primary care program)

Outcome

All-cause mortality, hospitalization, emergency department use, hospital days

Timing

Studies that include a follow-up period of more than 30 days

Setting

Ambulatory setting

Study design

Systematic reviews, controlled clinical trials, observational studies

Several factors complicated the reviewers’ approach to selecting studies, includ ing the lack of objective criteria for determining level of risk for hospital admission or death; complexity of studies of healthcare delivery systems; and lack of standard taxonomy for describing intensive primary care programs. Accordingly, the review ers anticipated encountering a high level of variation in the literature in the types of programs and patient populations and in the adequacy of reporting relevant PICOTS elements. The team performed dual review, specifically pairing reviewers with and without clinical expertise in primary care. However, this approach resulted in more than typical levels of inter-reviewer variation in judgments about the relevancy of evidence, as outlined in Table 6.5. These discrepancies were discussed and resolved among the review team, ultimately reducing variability while improving the clinical relevance of the report.

Copyright © 2024 Wolters Kluwer, Inc. Unauthorized reproduction of the content is prohibited.

TABLE 6.5 STUDY SELECTION USING DUAL REVIEW OBSERVATION REVIEWERS AGREED

REVIEWERS DISAGREED

No widely accepted disease-nonspecific, objective criteria for determining level of risk for hospital admis sion and/or death exist

Include studies that used indi cators of risk that had high face validity, such as history of persistently high hospital utilization and/or presence of multiple chronic illnesses

Reviewers with clinical exper tise in primary care were more likely to include studies that used less obvious indi cators of risk including need for intensive assistance (eg, impairments in multiple activ ities of daily living, poverty, and age ≥65 years)

Population