Table 2 Summary of methodological strengths and weaknesses
# | First author; Year | Length of follow-up time | Performance evaluation strategy | Patient population | Limitations | Control for biases |
|---|---|---|---|---|---|---|
1 | Cronenwett et al. 2007 [33] | 3 years | Time trend | Clearly defined sample of patients undergoing vascular surgery | -Risk adjustment was not performed. -Only processes of care were evaluated. | None specified in the article. |
2 | Campion et al. 2011 [34] | 4 years | Performance compared between initial and later participants | Sample of end-of-life cancer patients defined by age, sex and tumor type | -Risk adjustment was not performed. -Only processes of care were evaluated. | None specified in the article. |
3 | Stern et al. 2011 [35] | 5 years | Time trend | Clearly defined sample of cystic fibrosis patients. | Limited number of care centers involved | -The performance of each center was analyzed separately -Analysis was age-adjusted for certain indicators |
4 | Hermans et al. 2013 [36] | 1 year | RCT | Clearly defined sample of diabetic patients | -Short follow-up time. -Highly heterogeneous group of care settings involved | -Use of control group. -Differences between patients as well as care settings were accounted for in the analysis |
5 | Merle et al. 2009 [37] | 6 months | Before/after comparison | Clearly defined sample of patients undergoing surgical care for hip fracture. | -Short follow-up time -Small number of hospitals involved. -No use of control group | Analysis performed for each hospital involved. |
6 | Hall et al. 2009 [38] | 3 years | Time trend | Sample of patients undergoing general and vascular surgery | -Self selection of centers, thus the results may not be representative of the population. -The analysis is based on sampling. | Different modelling approaches were used to control for differences between patients. |
7 | Tepas III et al. 2014 [39] | 15 months | Time trend | Sample of patients undergoing general and vascular surgery. | -Short follow-up period. -Little information on patient population. | Risk-adjustment was performed. |
8 | Nuti et al. 2016 [40] | 5 years | Time trend | General population | -Highly aggregated data analysis (regional level) -Use of composite indicator that is based on 14 indicators. | -Population-based study -Data was standardized for age and sex |
9 | Govaert et al. 2016 [41] | 3 years | Time trend | -Population-based -Clearly defined sample of patients undergoing surgery for colorectal cancer. | -Only short-term survival was considered. | -Population-based study -Risk-adjustment was performed to account for differences between patients. -External data validation performed |
10 | Piccoliori et al. 2020 [42] | 3 years | Before/after comparison | Sample of patients with chronic conditions. | -Small-scale study -Results were not adjusted for differences between care providers or patients -Little information on patient population | -Information bias was diminished by removing prevalences from the analysis. |
11 | Qvist et al. 2004 [43] | 1 year | Time trend | Few information on patients characteristics as the focus of the analysis is on the providers | -Short follow-up time period -No risk adjustment was performed. | None specified in the article. |
12 | Nuti et al. 2013 [44] | 4 years | Time trend | General population | -Highly aggregated data analysis (regional level) | -Population-based study -Data was standardized for the population’s health needs |
13 | Van Leersum et al. 2013 [45] | 2 years | Time trend | -Population-based -Clearly defined sample of patients undergoing surgery for colorectal cancer. | - Short follow-up time period | -Population-based study -The data was adjusted for differences between patients. -External data validation was performed |
14 | Margeirsdottir et al. 2010 [46] | 5 years | Time trend | -Population-based -Clearly defined sample of pediatric patients with diabetes. | -No information on non-participants | -Population-based study -Adjustment for patient age and duration of disease was performed. -All measurements were standardized. |
15 | Kodeda et al. 2015 [47] | 18 years | Time trend | -Population-based - Clearly defined sample of patients with colorectal cancer. | -Lack of external data validation -Absence of control group | -Population-based study -Longer follow-up time period. |
16 | Pinnarelli et al. 2011 [48] | 3 years | Time trend | -Population-based - Clearly defined sample of patients undergoing surgical care for hip fracture. | -A number of confounders including patient co-morbidities could not be controlled for in the analysis. | -Population-based study -Risk-adjustment of performance was performed. |
17 | Miyata et al. 2012 [49] | 4 years | Performance compared between initial and later participants | Clearly defined sample of patients undergoing coronary artery bypass graft (CABG) | -Limited number of participants involved | -Risk-adjustment of performance was performed |