Running head: EFFECTIVENESS OF CARE COORDINATION IN TYPE DIABETICS
EFFECTIVENESS OF CARE COORDINATION IN TYPE 2 DIABETICS
Appendix E Evidence Synthesis Table
|Evidence Based Practice Question (PICO):
For adults with T2DM does care coordination in addition to visits with a PCP compared with just seeing a PCP improve HA1c levels?
|Bray, Paul; et. al. (2013)||Chao, Jianqian; et. al. (2014)||Collinsworth, Ashley; et. al. (2013)||Hsu, Chih-Cheng; et. al. (2014)||Yuan, Xiaodan; et. al. (2016)|
|Intervention involved a team-based approach to medication adjustment.||X||X||X|
|Evaluated PCP and care coordinator preferences and opinions.||X|
|Evaluated cultural competency of the care coordinator and its impact on patient outcomes.||X|
|Model for long term outcomes of the care coordination.||X|
|Recommend adopting the care coordination in their patient population.||X||X||X||X||X|
|Recommend research on health economics of care coordination.||X||X|
|Recommend goal setting with the patient as opposed to for the patient.||X||X|
|Recommend longer study to evaluate long term outcomes.||X||X||X|
Appendix C Strengths and Weakness Table
|Evidence Based Practice Question (PICO): PICO Question: For adults with T2DM does care coordination in addition to visits with a PCP compared with just seeing a PCP improve HA1c levels?
|Source (Authors, year)||Strengths||Weaknesses||Level of Quality Rating|
|Bray, Paul; Cummings, Doyle; Morrissey, Susan; Thompson, Debra, Holbert, Don; Wilson, Kyle; Lukusious, Eric; Tanenberg, Robert (2013)||There was no statistically significant difference in the intervention and control.
Homogenous population with all patients being African Americans with new onset of T2DM and HA1c >7.5%.
Only 5% of the patients approached to participate in the study declined
Mortality was mostly due to patients moving from the area
Use of an interdisciplinary team including nurses, pharmacists and dieticians.
|Intervention sights did not include quality improvement training and leadership commitment.
The control sites were selected randomly from practices with similar patient and practice characteristics and from the same regional environment. This decision was driven by national and state initiatives to improve diabetes care among federally funded centers.
Lack of data collected on medication adherence, treatment changes, and patient satisfaction.
The intervention was multidimensional and data collection was not conducted in a way to determine if it was a particular aspect of the intervention caused the results or if it was truly the collaborative team that lead to the results.
|Chao, Jianqian; Zong, Mengmeng; Xu, Hui; Yu, Qing; Jiang, Lili, Li, Yunyun; Song, Long; Liu, Pei (2014)||No study participants were lost to follow-up.
Uses the Markov model to attempt to predict long term outcomes.
Homogeneity between intervention and control groups.
Study approved by Medical Ethics Committee of Southeast University.
The Markov model is able to produce information such as distribution of certain events and decision-making models for complex problems.
All community-based health managers were trained on the same protocol to deliver the intervention.
|Study was relatively short and the Markov is a mathematical predictor not an actual measurement of long-term outcomes.
RCT are not able to run on the long-term.
Economic health considerations were not included.
Data was taken from the UKPDS to determine the transition probability and future studies should include data more applicable to the Chinese population.
|Collinsworth, Ashley; Vulimiri, Madhulika; Schmit, Kathryn; Snead, Christine (2013)||The study was approved by the Baylor Health Care System Institutional Review Board.
The secondary qualitative component gave a sense of the patient, community health workers, and PCP opinions on their feeling on the study.
Observed changes in HA1c were consistently lowered in the five different clinics involved in the study across the city of Dallas.
|Study was relatively short term (1.1 years).
The quantitative data published is the only from the first 18 months of the study.
Informants could have been biased because they knew they were being reported for the study.
Although participants were assured their identity would not be associated with their responses, however, there was a small number of community health worker and PCPs limits the anonymity.
More female participants agreed to participate in the study than males.
15% attrition rate in the study.
|Hsu, Chih-Cheng; Tai, Tong-Yuan (2014)||Large number of participants.
Study conducted over the course of 3.5 years.
Cultural competency of the case mangers was a factor in the study.
Two intervention groups to determine if the cultural company could impact outcomes.
Access to care was improved for all participants in the study.
This study provided public health awareness to diabetes care.
|No evaluation of other determinants of diabetic control (rate of complications, BMI, blood pressure).
This study did not evaluate barriers to care as a determinant of diabetic outcomes.
No evaluation on the cost effectiveness of the intervention studied.
|Yuan, Xiaodan; Wang, Fengmai; Folta Fish, Anne; Xue, Cunyi; Chen, Tao; Liu, Chao; Lou, Qingqing, Loa (2016)||Continued to rack patients in the intervention group even after the intervention ended.
Lack of attrition during the follow-up period.
Care was individualized and patient focused.
Study was approved by the hospital’s institutional review board.
Power analysis of 80%.
Practical protocol that is realistic to implement into practice.
Evaluated patient self-care and empowerment.
|Study only conducted over two years.
Study took place at only one hospital in China.
Study was not blinded.
Study did not analyze health care costs.
Self-care and empowerment was not measured at 24 months.
Simply asking patient questions could have been seen as an “intervention” because it brings awareness to the behavior.
Appendix D Summary Evidence Rating Table
|Evidence Based Practice Question (PICO): For adults with T2DM does care coordination in addition to visits with a PCP compared with just seeing a PCP improve HA1c levels?|
|Level of Evidence||Number of
|Summary of Findings||Overall Quality|
|1) Bray et. al. (2013) This redesigned primary care model worked well in rural African Americans with T2DM to significantly lower HA1c.
2) Chao et. al. (2015) The Markov model is a useful tool when predicting the long-term outcomes in patients with T2DM.
3) Collinsworth et. al. (2013) patients who participated in the study has a statistically significant decrease in HA1c after care coordination. Community health workers and PCPs reported great satisfaction with the care coordination program.
4) Hsu et. al. (2014) This study was able to provide evidence that case management on patients in Taiwan was effective in improving glycemic control for at least three years. Additional research needs to be conducted on the cost effectiveness on care coordination in Taiwan because there a still economic health disparities with single payer health coverage.
5) Yuan et. al. (2016) One year case management in China was an effective intervention at 6 months and continued to be successful at 12 and 24 months as evidenced by lowered HA1c levels and improved patient self-care strategies.
|1) B, well-designed intervention with a collaborative approach to care coordination. Study has poor generalizability due to the specific patient population; however, strong findings in their study population.
2) B, excellent idea to apply a mathematical model to combat a large gap in knowledge. Still need a longer-term study to see if the Markov model was truly accurate.
3) C, many outcomes were measured but only one (HA1c) was truly impacted by the study. Good team approach to care. The qualitative aspect of the study was very informal and could have provided valuable insight if fleshed out further.
4) A, well-designed study which took place over 3 years with a large sample size.
5) A, well-designed study with good outcomes measured.