Background
Type of Integration | Definition | Examples |
---|---|---|
Relates to strategies that link similar levels of care | Physicians join existing group practices or multiple groups merge | |
Relates to strategies that link different levels of care | Various health care professionals, such as physicians, nurses, physiotherapists, collaborate with hospitals, universities/medical schools, health plans, etc. | |
Refers to the extent to which professionals coordinate services across various disciplines | Nurse practitioners work with dieticians to provide care for patients with diabetes | |
Clinical integration [38] | Refers to the extent to which care services are coordinated | Maintaining an open communication channel by having dieticians send a consultation report to the family physician after appointment with their patient |
Methods
Criteria for considering systematic reviews
-
Target population included people with multi-morbidities, of which diabetes and/or hypertension was one. Multi-morbidity was defined as having two or more chronic conditions for an individual.
-
Interventions included integration of health care delivery, which was defined as models of care where prevention, diagnosis, or treatment of hypertension, diabetes, or any NCD was combined with the delivery of health care for any other condition (e.g. communicable disease, maternal and child care, mental health, etc.). The integration for these care services may require professional coordination across several disciplines
-
The comparisons included “usual care” as defined specifically by each study and stand-alone models of health care delivery, where care was directed only towards the prevention, diagnosis, or treatment of hypertension or diabetes.
-
Reported outcomes included health outcomes (e.g. all-cause mortality, disease-specific morbidity) and process outcomes (e.g. access to care, retention in care, continuity of care, quality of care, cost of care, user-views of care recipients).
-
Published in English
-
We included systematic reviews which had to include [18, 19]:
-
◦ A clearly stated set of pre-determined objectives with an explicit, reproducible methodology
-
◦ Pre-determined criteria for eligibility
-
◦ A systematic search that attempted to identify all studies that would meet the eligibility criteria through searching through at least two data sources, with at least one of them being an electronic database
-
◦ Performed data extraction and risk of bias assessment
-
Approach for identifying systematic reviews
Selection of systematic reviews and data extraction
Assessing risk of bias of systematic reviews
Assessing reporting
Data analysis
Results
Results of the search and description of included systematic reviews
Review | Search date | Number of studies included | Population and country where study was conducted | Intervention | Comparator | Outcomes |
---|---|---|---|---|---|---|
Atlantis 2014 [23] | August 2013 | 7 | Adults diagnosed with depression and co-morbid diabetes Australia: 1 study USA: 6 studies RCT: 7 Range of number of participants in each study: 58–417 | Integrated care: 2 studies Collaborative care: 5 studies Type of integration: Vertical Clinical | Usual care: 4 studies Enhanced usual care: 3 studies | Depression score outcome • CES-D 20 in 1 study • PHQ-9 in 2 studies • SCL-20 in 4 studies Glycaemic control by HbA1c |
Huang 2013 [24] | 27 March 2013 | 8 | Patients with depression and diabetes One trial included only African Americans and two trials included only old patients (aged over 50 and 60 year old respectively) USA: all studies RCT: 8 Range of number of participants in each study: 58–417 | Integrated management: 2 studies Program to Encourage Active, Rewarding Lives for Seniors (PEARLS): 1 study Multifaceted Diabetes and Depression Program (MDDP): 1 study Collaborative care: 1 study Stepped collaborative care: 1 study An individualized stepped-care depression treatment program: 1 study Improving Mood Promoting Access to Collaborative Treatment (IMPACT): 1 study Type of integration: Vertical Professional Clinical | Normal usual care: 5 trials Enhanced usual care: 3 trials | Depression treatment response and depression remission at the end of follow-up Depression treatment response at 6 and 12 months follow-up Depression remission at 6 and 12 months follow-up Diabetes clinical outcomes (HbA1c values) at the end of follow-up Diabetes clinical outcomes (HbA1c values) at 6 and 12 months follow-up Adherence to medication (including adherence of oral hypoglycemic agents and/or antidepressants) |
Joshi 2014 [25] | 26 May – 13 June, 2013 | 22 total in review; 4 of relevance | Task-shifting for the management of hypertension and cardiovascular diseases: 7 studies Task-shifting for the management of diabetes: 5 studies Cameroon: 3 studies South Africa: 1 study Before-after study: 3 RCT:1 Range of number of participants in each study: 221–1343 | Task-shifting for the management of non-communicable disease Task shifting from: Physicians to health workers: 4 studies Type of integration: Vertical Professional Clinical | Usual healthcare: 4 studies | Process outcomes Disease outcomes Treatment concordance Cost-effectiveness Enablers for task-shifting Barriers to task-shifting |
Smith 2016 [26] | 28 Sept, 2015 | 18 total in review; 6 of relevance | Hypertension and depression: 1 study Depression and diabetes/heart disease: 3 studies Hypertension and diabetes: 2 studies Australia: 1 study UK: 1 study USA: 4 studies Cluster RCT: 1 RCT: 4 RCT (pilot): 1 Range of number of participants in each study: 61–400 | Change to organisation of care delivery through case management or enhanced multidisciplinary team work: 5 studies Patient-oriented, such as educational or self-management support-type interventions: 3 study Type of integration: Vertical Professional Clinical | Usual care: 2 studies Usual care additional services: 3 studies Enhanced primary care: 1 studies | Patient clinical or mental health outcomes Patient-reported outcome measures Utilisation of health services Patient behaviour Provider behaviour Acceptability of the service to recipients and providers, and treatment satisfaction Economic outcomes |
Watson 2013 [27] | 11 June, 2012 | 12 studies; 3 studies of relevance | Adults (mean age = 59 years) and various chronic medical conditions, including 3 studies of relevance focusing on diabetes and depression USA: all studies RCT: 2 studies Preplanned sub-group analysis from a separate randomized controlled trial: 1 study Range of number of participants in each study: 329–417 | All included studies characterized their respective intervention as a form of collaborative care, not another form of a practice-based intervention (such as integrated care) Type of integration: Vertical Professional Clinical | Enhanced usual care: 2 studies (9 articles) Usual care: 1 study (2 articles) | Mental health outcomes Chronic medical outcomes Harms |
Settings
Study design/population
Interventions
Comparison
Outcomes
Risk of bias of included systematic reviews
Study | Domain 1: Study eligibility criteria* | Domain 2: Identification and selection of studies^ | Domain 3: Data collection and study appraisal# | Domain 4: Synthesis and findings+ | Overall risk of bias of review | Justification |
---|---|---|---|---|---|---|
Atlantis 2014 [23] | High | Unclear | High | High | High | D1: Eligibility criteria lacked detail D2: Unclear whether authors searched for unpublished or ongoing studies and whether selection of studies was completed independently or in duplicate D3: Concerns raised about data extraction and risk of bias assessment D4: Risk of bias was not taken into consideration when conclusions were formed |
Huang 2013 [24] | High | High | High | High | High | D1: No published protocol D2: Did not include sources for unpublished reports in search D3: No information on data extraction and risk of bias assessment for two of the studies included in qualitative synthesis D4: Results might have not been reported in light of risk of bias |
Joshi 2014 [25] | High | High | High | High | High | D1: No protocol and lack of specific details on eligibility criteria D2: Did not consider other methods of searching for unpublished literature D3: Lack of information on quality assessment D4: Did not pre-specify methods of data analysis. Risk of bias not adequately assessed or reported with results |
Smith 2016 [26] | Low | Low | Low | Low | Low | No concerns |
Watson 2013 [27] | Low | Low | Low | Low | Low | The review’s eligibility criteria are limited to studies in the English language; however overall there is no concern. After deliberation, all review authors judged this review as a low risk of bias. |
AMSTAR item | Atlantis 2014 [23] | Huang 2013 [24] | Joshi 2014 [25] | Smith 2016 [26] | Watson 2013 [27] |
---|---|---|---|---|---|
1. ‘A priori’ design | No | No | Yes | Yes | Yes |
2. Duplicate study selection and data extraction | Can’t answer | Can’t answer | No | Yes | Yes |
3. Literature search | Yes | Yes | Yes | Yes | Yes |
4. Status of publication | No | Can’t answer | No | Yes | Yes |
5. List of studies | Yes | No | No | Yes | Yes |
6. Characteristics of included studies | Yes | Yes | Yes | Yes | Yes |
7. Scientific quality | Yes | Yes | No | Yes | Yes |
8. Formulation of conclusion | No | No | No | Yes | Yes |
9. Methods used to combine findings | Yes | Yes | Yes | Yes | Yes |
10. Likelihood of publication bias | Yes | Yes | No | Yes | No |
11. Conflict of interest | No | Can’t answer | No | No | Yes |
Quality score | 6 | 5 | 4 | 11 | 10 |
Quality rating | Medium | Medium | Medium | High | High |
Comparison of risk of bias assessment using ROBIS and AMSTAR
Systematic review | ROBIS risk of bias assessment grade | AMSTAR quality assessment grade |
---|---|---|
Atlantis 2014 [23] | High risk of bias | Medium quality |
Huang 2013 [24] | High risk of bias | Medium quality |
Joshi 2014 [25] | High risk of bias | Medium quality |
Smith 2016 [26] | Low risk of bias | High quality |
Watson 2013 [27] | Low risk of bias | High quality |
Comparison of reporting using PRISMA
PRISMA checklist item | Atlantis 2014 [23] | Huang 2013 [24] | Joshi 2014 [25] | Smith 2016 [26] | Watson 2013 [27] |
---|---|---|---|---|---|
Title | |||||
1. Title | Yes | Yes | Yes | Yes | Yes |
Abstract | |||||
2. Structured summary | Yes | Yes | Yes | Yes | Yes |
Introduction | |||||
3. Rationale | Yes | Yes | Yes | Yes | Yes |
4. Objectives | Yes | Yes | Yes | Yes | Yes |
Methods | |||||
5. Protocol and registration | No | Yes | Yes | Yes | Yes |
6. Eligibility criteria | Yes | Yes | Yes | Yes | Yes |
7. Information sources | Yes | Yes | Yes | Yes | Yes |
8. Search | Yes | Yes | No | Yes | Unclear |
9. Study selection | Yes | Yes | No | Yes | Yes |
10. Data collection process | Partly | Partly | No | Yes | Yes |
11. Data items | No | Yes | No | Yes | Yes |
12. Risk of bias in individual studies | Partly | Yes | No | Yes | Yes |
13. Summary measures | Yes | Yes | N/A | Yes | Yes |
14. Synthesis of results | Yes | Yes | No | Yes | Yes |
15. Risk of bias across studies | Yes | Yes | No | Yes | Unclear |
16. Additional analyses | Partly | Yes | No | Yes | Yes |
Results | |||||
17. Study selection | Yes | Yes | Not adequate | Yes | Yes |
18. Study characteristics | Yes | Yes | Yes | Yes | Yes |
19. Risk of bias within studies | Partly | Yes | No | Yes | Unclear |
20. Results of individual studies | Yes | Yes | No | Yes | Yes |
21. Synthesis of results | Yes | Yes | N/A | Yes | Partly |
22. Risk of bias across studies | No | N/A | No | Yes | No |
23. Additional analysis | Yes | N/A | N/A | N/A | No |
Discussion | |||||
24. Summary of evidence | Yes | Yes | Yes | Yes | Partly |
25. Limitations | Partly | Yes | Partly | Yes | Yes |
26. Conclusions | Yes | Partly | Partly | Yes | Yes |
Funding | |||||
27. Funding | Yes | No | Yes | Yes | Partly |
Number of items included (yes) | 19/27 | 22/27 | 10/27 | 26/27 | 19/27 |
Inter-rater reliability of ROBIS and AMSTAR
Overlap of studies included in systematic reviews
Findings on the effects of integrated care
Comorbid conditions | Type of outcome | Findings |
---|---|---|
Diabetes and depression | Risk of all-cause mortality | 6 months: RD 0.00, 95%CI -0.02 to 0.02, 2/7 studies of relevance (Watson 2013) 12 months: RD 0.00, 95%CI -0.02 to 0.01, 2/7 studies of relevance (Watson 2013) |
Depression | Depression scores SMD −0.32, 95%CI -0.53 to −0.11, 7 studies (Atlantis 2014) SMD − 0.41 95%CI -0.63 to − 0.20, 6 studies; 4 of relevance (Smith 2016) Response rate 6 months: RR 1.64, 95%CI 1.28 to 2.10, 4 studies (Huang 2013) 12 months: RR 1.42, 95%CI 1.14 to 1.76, 4 studies (Huang 2013) End of follow-up: RR 1.33, 95%CI 1.05 to 1.68, 4 studies (Huang 2013) Remission 6 months: RR 1.33, 95%CI 1.01 to 1.75, 2 studies (Huang 2013) RD 0.123, 95%CI 0.064 to 0.183, 3 studies (Watson 2013) 12 months: RR 1.20, 95%CI 0.93 to 1.55, 2 studies (Huang 2013) RD 0.077, 95%CI 0.016 to 0.137, 3 studies (Watson 2013) 18 months: RD 0.075, 95%CI 0.013 to 0.136, 3 studies (Watson 2013) 24 months: RD 0.045, 95%CI -0.023 to 0.113, 3 studies (Watson 2013) End of follow-up: RR 1.15, 95%CI 0.87 to 1.52, 2 studies (Huang 2013) Reduction (of at least 50%) in Mental Health score 6 months: RD 0.20, 95%CI 0.14 to 0.26, 4/9 studies of relevance (Watson 2013) 12 months: RD 0.17, 95%CI 0.12 to 0.23, 3/7 studies of relevance (Watson 2013) 18 months: RD 0.12, 95%CI 0.02 to 0.22, 1/3 studies of relevance (Watson 2013) | |
Diabetes clinical outcome (HbA1c level) | 6 months: MD − 0.06, 95%CI -0.24 to 0.12, 4 studies (Huang 2013) MD 0.13, 95%CI -0.22 to 0.48, 3 studies (Watson 2013) 12 months: MD − 0.07, 95%CI -0.28 to 0.13, 4 studies (Huang 2013) MD 0.24, 95%CI -0.14 to 0.62, 3 studies (Watson 2013) End of follow-up: WMD − 0.33, 95%CI -0.66 to 0.00%, 7 studies (Atlantis 2014) MD − 0.13, 95%CI -0.46 to 0.19, 7 studies (Huang 2013) WMD − 0.33, 95%CI -0.66 to 0.00%, 7 studies (Smith 2016) | |
Effect of depression remission on HbA1c | SMD for depression scores were unable to predict the WMD in HbA1c values; p − 0.828, coefficient 0.19, 95%CI -1.93 to 2.31, 7 studies (Atlantis 2014) | |
Symptom improvement | Watson et al. reported greater depression symptom improvement scores in intervention groups at 6 months: MD 0.38, 95%CI 0.24 to 0.51, 3/5 studies of relevance (Watson 2013) 12 months: MD 0.38, 95%CI 0.30 to 0.46, 3/5 studies of relevance (Watson 2013) 24 months: MD 0.18, 95%CI 0.10 to 0.26, 2/3 studies of relevance (Watson 2013) | |
Systolic blood pressure | MD −3.10, 95%CI -7.26 to 1.06, 5 studies (Smith 2016) | |
Access and utilisation of healthcare services | 12 months: range: 42 to 84%; usual care range: 16 to 33%, 3/4 studies of relevance (Watson 2013) | |
Quality of care in terms of mental health treatment satisfaction | 12 months: RD 0.205, 95%CI 0.112 to 0.299, 3/4 studies of relevance (Watson 2013) 24 months: RD 0.14, 95%CI 0.06 to 0.21, 1/3 studies of relevance (Watson 2013) |